Charting a Course for Sustainable Hydrological and Meteorological Observation Networks in Developing Countries R. David Grimes • David P. Rogers • Andreas Schumann • Brian F. Day Charting a Course for Sustainable Hydrological and Meteorological Observation Networks in Developing Countries Purpose and Audience The aim of this report is to explore in depth the root causes of why national meteorological and hydrological systems fall into disrepair—even after being modernized—and to share a visionary approach with recommendations to achieve more sustainable outcomes. The report is organized into three parts, with the first providing the larger context and a proposed vision and the other two delving into the more technical aspects of meteorological and hydrological networks, offering guidance and recommendations on their design and sustainable operations over the long term. It builds on successful outcomes in developed nations, which not only offer insights for many developing countries but also help development agencies integrate the principles and conditions of success into the design of their own development projects. The target audience is professionals working on weather and climate investment programs, including National Meteorological and Hydrological Services (NMHS) staff and directors, staff in the core planning and finance ministries, and professionals in development institutions. Charting a Course for Sustainable Hydrological and Meteorological Observation Networks in Developing Countries R. David Grimes • David P. Rogers • Andreas Schumann • Brian F. Day © 2022 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW, Washington DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org Some rights reserved 1 2 3 4 20 19 18 17 This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Nothing herein shall constitute or be considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved. Rights and Permissions This work is available under the Creative Commons Attribution 3.0 IGO license (CC BY 3.0 IGO) http:// creativecommons.org/ licenses/by/3.0/igo. Under the Creative Commons Attribution license, you are free to copy, distribute, transmit, and adapt this work, including for commercial purposes, under the following conditions: Attribution—Please cite the work as follows: World Bank, 2022. Charting a Course for Sustainable Hydrological and Meteorological Networks in Developing Countries. Washington, DC: World Bank. License: Creative Commons Attribution CC BY 3.0 IGO. Translations—If you create a translation of this work, please add the following disclaimer along with the attribution: This translation was not created by The World Bank and should not be considered an official World Bank translation. The World Bank shall not be liable for any content or error in this translation. Adaptations—If you create an adaptation of this work, please add the following disclaimer along with the attribution: This is an adaptation of an original work by The World Bank. Views and opinions expressed in the adaptation are the sole responsibility of the author or authors of the adaptation and are not endorsed by The World Bank. Third-party content—The World Bank does not necessarily own each component of the content contained within the work. The World Bank therefore does not warrant that the use of any third-party–owned individual component or part contained in the work will not infringe on the rights of those third parties. The risk of claims resulting from such infringement rests solely with you. If you wish to reuse a component of the work, it is your responsibility to determine whether permission is needed for that reuse and to obtain permission from the copyright owner. Examples of components can include, but are not limited to, tables, figures, or images. All queries on rights and licenses should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.org. Cover photos: top, Petrovich9; bottom, AscentXmedia Back cover photo: Julia Sudnitskaya    v Contents Foreword.................................................................................................................................... xix Acknowledgments...................................................................................................................... xxi Part 1. A Vision: Charting a Course for Sustainable Hydrological and Meteorological Observation Networks in Developing Countries..................................................................1 Part 1 Abbreviations.....................................................................................................................3 Part 1 Executive Summary............................................................................................................ 5 Purpose and Audience................................................................................................................................. 5 Hydrological and Meteorological Value Chain............................................................................................... 5 Challenges................................................................................................................................................. 6 Why Networks Fail...................................................................................................................................... 6 Vision for a More Holistic and Sustainable Approach..................................................................................... 7 1.1 Setting the Scene................................................................................................................... 9 1.1.1 Introduction................................................................................................................................... 9 1.1.2 Why Sustainable Networks Matter.................................................................................................. 10 1.1.2.1 Assessing Environmental Change........................................................................................... 10 1.1.2.2 Creating Societal and Environmental Benefits...........................................................................11 1.1.2.3 Managing Disaster- and Water-Related Risks........................................................................... 14 1.1.2.4 Framing International Policy Agreements................................................................................ 16 1.1.3 Why Networks and Systems Fail: Framing the Sustainability Problem............................................... 16 1.1.3.1 Affordability.......................................................................................................................17 1.1.3.2 Governance........................................................................................................................ 18 1.1.3.3 Engagement....................................................................................................................... 19 1.1.3.4 Institutional Arrangements................................................................................................... 19 1.1.3.5 Workforce Competencies Challenges....................................................................................... 19 1.1.4 Conclusions.................................................................................................................................. 20 1.2 Pathway to More Sustainable Networks................................................................................. 21 1.2.1 Introduction..................................................................................................................................21 1.2.2 Vision Elements............................................................................................................................ 22 vi    Contents 1.2.2.1 Mapping Context................................................................................................................. 22 1.2.2.2 Understanding TCO and Affordability..................................................................................... 23 1.2.2.3 Decision Considerations....................................................................................................... 23 1.2.3 Understanding Affordability: Fit-for-Purpose and Fit-for-Budget...................................................... 24 1.2.3.1 The TCO Methodology.......................................................................................................... 24 1.2.3.2 Aligning Expectations through Benchmarking.......................................................................... 27 1.2.4 Choosing the Right Fit-For-Purpose Business Model........................................................................ 32 1.2.5 Legal Mandates: Clarity, Authority, and Accessibility...................................................................... 36 1.2.5.1 Legal Frameworks............................................................................................................... 36 1.2.5.2 Data Policies...................................................................................................................... 37 1.2.5.3 Commercial Services and Regulatory Policies........................................................................... 38 1.2.6 Partnering: A System-of-Systems Approach.................................................................................... 39 1.2.7 Innovation and Technology........................................................................................................... 44 1.2.8 Sustaining a Talented Workforce.................................................................................................... 45 1.3 Conclusions and Recommendations.......................................................................................47 Part 1 References....................................................................................................................... 50 Part 1 Boxes Box ES1.1  Key Results from Benchmarking against More Sustainable Network Practices........................................7 Box 1.1.1  How Weather, Water, and Climate Data Support Low-Carbon Footprints................................................ 13 Box 1.2.1  Total Cost of Ownership and Life-Cycle Costing...................................................................................25 Box 1.2.2  Procuring Observation Systems: The Case of India..............................................................................27 Box 1.2.3  Better Hydromet Observations and Services: The Case of Nepal...........................................................34 Box 1.2.4  Regional Cooperation through the South Asian Hydromet Forum.........................................................43 Part 1 Figures Figure ES1.1 Hydromet Value Chain................................................................................................................... 6 Figure 1.1.1  Sample Page from WMO Real-Time Global Data-Processing and Forecasting System Modelling Centre.........................................................................................................................11 Figure 1.1.2  Hydromet Value Chain..................................................................................................................12 Figure 1.1.3  The Value Chain as a Shared Enterprise among the Public, Private, and Academic Sectors...............13 Figure 1.1.4  Summary of Costs of Natural Disasters Worldwide, 1980–2019..................................................... 14 Figure 1.1.5  The Central Role of Water in Sustainable Development................................................................. 16 Figure 1.2.1  Hydrological and Meteorological Development Model................................................................... 22 Figure B1.2.1.1  The Components of Total Cost of Ownership.............................................................................. 25 Figure 1.2.2 Key Elements in Costing Networks............................................................................................... 26 Figure 1.2.3  Annual Indicative Benchmarking Costs per Hydromet Station in Developed Countries, US dollars... 28 Contents   vii Figure 1.2.4  Indicative Costs for Meteorological Stations in Developing Countries Compared with Benchmark Countries....................................................................................................................................31 Figure 1.2.5  Four Different Business Models for Owning and Operating Observation Networks Based on Financial Risk and Technical Readiness ....................................................................................... 33 Figure 1.2.6  Legal and Regulatory Instruments Governing the Provision of Meteorological and Hydrological Services..................................................................................................................................... 36 Figure 1.2.7 System-of-Systems Thinking: Co-Design, Co-Production, and Co-Management................................ 40 Figure 1.3.1 Annual Indicative Benchmarking Costs per Hydromet Station in Developed Countries, US dollars... 49 Part 1 Tables Table 1.1.1  Observation Network Challenges..................................................................................................17 Table 1.2.1  Indicative Costs for Meteorological Stations in Developing Countries Compared with Benchmark Countries................................................................................................................................... 29 Table 1.2.2  Benchmarking Costs for Hydrological Stations in Developed Countries (Canada, Germany, the United States)...................................................................................................................... 32 Part 2. Recommendations for the Design of Sustainable Hydrological Observation Networks in Developing Countries........................................................................................55 Part 2 Abbreviations......................................................................................................... 56 Part 2 Executive Summary................................................................................................. 57 The Problem............................................................................................................................................. 57 Making an Observation Network Fit for Purpose.......................................................................................... 58 Ensuring Alignment of the Costs of the Modernized Observation System with the Budget.............................. 61 Securing Human Resources to Operate a Modern Measurement Network....................................................... 61 Recommendations..................................................................................................................................... 62 2.1 Introduction......................................................................................................................... 64 2.2 The Benefits and Value of Hydrological Services...................................................................67 2.2.1 The Hydrological Value Chain ....................................................................................................... 67 2.2.2 Benefits of Basic Networks............................................................................................................ 70 2.2.3 Benefits of Operational Networks...................................................................................................71 2.2.4 Added Value and the Need for Hydrological Services....................................................................... 72 2.3 Efficient Organization of Hydrological Services....................................................................74 2.3.1 The Basic Structure of a National Hydrological Service.................................................................... 74 2.3.2 Fit for Purpose? Determining the Institutional Structure’s Efficiency................................................ 78 2.3.3 Inter- or Intra-Agency Cooperation with a National Meteorological Service...................................... 80 viii    Contents 2.4 Development of Human Resources........................................................................................82 2.5 Design of Hydrometric Networks.......................................................................................... 86 2.5.1 Requirements for Hydrometric Observations.................................................................................. 87 2.5.2 Specification of the Measurement Program..................................................................................... 88 2.5.3 Stream-Gauging Network Design................................................................................................... 89 2.5.3.1 Network Design for Water Resources Assessment...................................................................... 89 2.5.3.2 The Design of Operational Networks....................................................................................... 89 2.5.3.3 Observations of Water Levels for Lakes and Reservoirs............................................................... 91 2.5.3.4 Stream Gauging in Estuarine River Areas................................................................................ 91 2.5.3.5 Selection of Stream-Gauging Devices...................................................................................... 92 2.5.4 Other Components of Hydrometric Networks................................................................................... 94 2.5.4.1 Operational Precipitation Networks to Support Hydrological Forecasts......................................... 94 2.5.4.2 Snow Course Network.......................................................................................................... 95 2.5.4.3 Water Balance Stations........................................................................................................ 96 2.5.4.4 Evaporation Measurements................................................................................................... 96 2.5.4.5 Sediment Measurements...................................................................................................... 96 2.5.4.6 Groundwater...................................................................................................................... 97 2.5.4.7 Water Quality..................................................................................................................... 97 2.5.5 From Design to Monitoring............................................................................................................ 97 2.6 Modernization of Stream-Gauging Networks........................................................................ 98 2.6.1 Status of the Hydrometric Network................................................................................................ 98 2.6.1.1 Categorization of Levels of Developments................................................................................ 99 2.6.1.2 Planning the Transition between Levels................................................................................. 100 2.6.2 Prioritization of River Basins for Modernization and Technical Re-Equipment................................. 102 2.6.3 Software Requirements and Procurement..................................................................................... 104 2.6.3.1 Software for Data Flow...................................................................................................... 104 2.6.3.2 Software for Hydrological Analyses...................................................................................... 105 2.6.3.3 Options to Provide Necessary Software................................................................................. 107 2.6.4 Ensuring Data Quality................................................................................................................. 108 2.6.5 Boundary Conditions for Success................................................................................................. 109 2.7 How to Estimate Total Costs of Operating a Stream-Gauging Network................................. 110 2.7.1 Methodology of Cost Estimation....................................................................................................110 2.7.1.1 Initial Investments: Capital Costs.......................................................................................... 111 2.7.1.2 Annual Operating Costs....................................................................................................... 112 2.7.1.3 Costs of Maintenance.......................................................................................................... 113 2.7.1.4 Cost of Replacements..........................................................................................................114 2.7.1.5 Application of the TCO........................................................................................................ 115 Contents    ix 2.7.2 Life-Cycle Cost Analysis................................................................................................................ 115 2.7.3 The Problem of Comparability of Costs..........................................................................................119 2.7.4 Cost Assessments of Stream-Gauging Networks in Developed Countries.......................................... 119 2.7.4.1 Initial Investments.............................................................................................................119 2.7.4.2 Annual Operating Costs....................................................................................................... 122 2.7.4.3 Costs of Maintenance..........................................................................................................125 2.7.4.4 Assessments of the Total Costs of Stream Gauges.....................................................................125 2.8 Summary and Conclusions................................................................................................... 129 2.8.1 Ensuring Alignment of the Costs of the Modernized Observation System with the Budget................. 131 2.8.2 Securing Human Resources to Operate a Modern Measurement Network.......................................... 131 2.8.3 Main Conclusions.........................................................................................................................132 Part 2 Annexes.......................................................................................................................... 133 Annex 2.1 Case Study: Valuation and Costing of Stream-Gauging Networks in Canada...................................134 Annex 2.2 Case study: Network Modernization of Russia’s Kuban River Basin...............................................139 Annex 2.3 Case Study: Planning Approach to NHMS Leveling Up in Sri Lanka...............................................143 Part 2 References......................................................................................................................145 Part 2 Boxes Box 2.1.1 A Guide to Key Terminology............................................................................................................65 Box 2.2.1  Key Uses of Historical and Real-Time Hydrological Data.................................................................. 70 Box 2.3.1  A Quick Guide to Estimating Discharges..........................................................................................77 Box 2.5.1  Demand for Discharge Data in a Complex System of a Cascade of Hydropower Plants........................ 90 Box 2.6.1  Software Required for Efficient Data Flow.....................................................................................106 Box 2.7.1  How to Compare the TCO of Two Gauging Stations......................................................................... 116 Box 2.7.2  How to Calculate Life-Cycle Cost for a Typical Gauging Station....................................................... 117 Box 2.7.3  Example of How to Determine the Workload for Discharge Measurements and Estimation of Rating Curves for 15 Gauges..................................................................................................... 124 Part 2 Figures Hydrological Value Chain............................................................................................................ 59 Figure ES2.1  Figure ES2.2 Adaptation of Human Resources Needed for NHS Technological Development................................. 62 Figure 2.2.1 Hydrological Value Chain............................................................................................................ 59 Figure 2.3.1 Typical Scalable Structure of an NHS with One Central Office ........................................................ 75 Figure 2.3.2 Structure of an NHS with Two Regional Hydrological Offices ......................................................... 75 Figure 2.4.1 Adaptation of Human Resources Needed for NHS Technological Development ................................ 83 Figure 2.5.1 Development of Hydrometric Networks as a Cyclic Process ........................................................... 88 Figure B2.5.1.1 Structural Changes Needed to Implement a Cascade of Hydropower Plants .................................90 x    Contents Figure 2.6.1 Criteria to Determine Strategic Importance, Costs, and Operation of Hydrometric Gauges ............ 103 Figure 2.6.2 Data Flow from Gauges to Data Collection, Processing, Storing, and Final Product Preparation .... 105 Figure 2.7.1 Overview of the Total Costs of Ownership Considering Replacement as Part of Life-Cycle Management ............................................................................................................. 111 Figure A2.2.1 Assessment of the Basins of Russia in Terms of Specific Flood Damage ....................................... 140 Part 2 Photos Photo B2.7.2.1 A Standalone Solar-Powered Station with Compact Bubble Sensor, Germany ............................ 117 Photo 2.7.1 and Photo 2.7.2  Two Gauges with Stilling Wells in Germany with Significantly Differing Construction Costs ............................................................................................................... 120 Part 2 Tables Table ES2.1  Range of Costs per Automated Stream Gauge from Examples in Canada, Germany, and the United States.............................................................................................................60 Table 2.2.1  Sectoral Benefits Derived from Water Quantity Information...................................................... 68 Table 2.3.1  Main Applications of Hydrological Products............................................................................. 79 Table 2.3.2  Sectors Using Both Hydrological and Meteorological Services................................................... 81 Table 2.5.1  Hydrometric Equipment for Streamflow and Water-Level Measurements .................................... 93 Table 2.6.1  Levels of Stream-Gauging Networks......................................................................................... 99 Table 2.6.2  Transitions between Levels of Development........................................................................... 100 Table 2.6.3  Personnel Requirements of Transitions to Higher Levels..........................................................101 Table 2.6.4  Required Change of Equipment...............................................................................................101 Table 2.6.5  Different Development Stages of Water-Level Data Collection.................................................. 102 Table B2.7.1.1  Life-Cycle Costs of Station A and Station B..............................................................................116 Table B2.7.1.2  Comparison of Annual Costs for Station A and Station B ..........................................................116 Table B2.7.2.1  Capital Costs (Purchase Costs, Cost of Civil Works) (US$)........................................................ 117 Table B2.7.2.2­  Maintenance Costs (US$/year)................................................................................................ 117 Table B7.2.3 Operation Costs (US$/year).................................................................................................... 117 Table B2.7.2.4  Present Values of Costs, Salvage Value, and the Resulting Life-Cycle-Costs (US$)......................118 Table 2.7.1  Potential Construction Costs: Canada and Germany (US$)....................................................... 120 Table 2.7.2  Costs of In-Situ Recording Water-Level Equipment (US$)......................................................... 121 Table 2.7.3  Costs of a Modern Station (Sensors, Loggers, Solar Panel, GSM Modem) (US$).......................... 121 Table 2.7.4  Costs of Flow and Velocity Measurement Equipment for 15 to 20 Gauges (US$)......................... 121 Table 2.7.5  Costs of Typical Field Equipment for 15 to 20 Gauges (US$)..................................................... 122 Table 2.7.6  Typical Initial Investment Costs Associated with Installation of Stream-Gauging Stations in Canada, Subdivided by Category (US$)............................................................................... 122 Table B2.7.3.1  Specification of the Annual Workload for Discharge Measurements of a Hydrological Technician for an Exemplary Measuring Network of 15 Gauges..................................................124 Table 2.7.7  Lifetime of Selected Components of Hydrometric Networks (Years)...........................................125 Contents    xi Table 2.7.8  Ratios of the Costs of the Initial Investment to the Costs of Direct Operation for Stream Gauges in Some Developed Countries.....................................................................................126 Table 2.7.9  Total Operating Costs, Assessed per Station and Year by the U.S. Geological Survey, by Category (US$) .................................................................................................................126 Table 2.7.10  Total Annual Operating and Maintenance Costs, Assessed per Station by the Water Survey of Canada, by Northern Canada and Southern Canada (US$)..................................................... 127 Table 2.7.11  Total Annual Operating Costs, Assessed per Station, Derived from an Analysis of the Operation Costs for 105 Stream Gauges in Germany (US$)....................................................... 127 Table 2.7.12  Range of Costs per Station: Examples from Canada, Germany, and the United States.................128 Table A2.1.1  Stream-Gauging Network Operation Costs in Northern and Southern Canada (US$)....................138 Part 3. Recommendations for the Design of Sustainable Meteorological Observation Networks and Systems in Developing Countries........................................147 Part 3 Abbreviations.................................................................................................................149 Part 3 Glossary ......................................................................................................................... 152 Part 3 Executive Summary......................................................................................................... 157 Time for a New Approach .........................................................................................................................157 How to Plan a Meteorological Observation System ....................................................................................158 How to Calculate Total Cost of Ownership................................................................................................. 160 Why the Key Weather Observing Systems Matter........................................................................................ 161 Aligning TCO Expectations through Benchmarking .....................................................................................162 Recommendations ...................................................................................................................................163 3.1 Overview of Sustainable Weather Observing Systems......................................................... 165 3.1.1 Introduction................................................................................................................................165 3.1.2 Barriers to Sustainability............................................................................................................ 166 3.1.3 Key Issues.................................................................................................................................. 169 3.1.4 Addressing the Key Issues............................................................................................................ 171 3.2 Network Planning................................................................................................................ 172 3.2.1 Introduction ............................................................................................................................... 172 3.2.2 The Planning Stage ..................................................................................................................... 175 3.2.2.1 Defining the Purpose of the Required Observation Network or System ......................................... 175 3.2.2.2 Identifying the Necessary Meteorological Observations............................................................176 3.2.2.3 Defining Requirement Specifications.....................................................................................176 3.2.2.4 Designing the Observation Network....................................................................................... 177 xii    Contents 3.2.2.5 Developing a Network Data Management Plan.........................................................................178 3.2.2.6 Defining Infrastructure Requirements....................................................................................178 3.2.2.7 Defining Operational Requirements.......................................................................................178 3.2.2.8 Identifying Personnel Requirements......................................................................................179 3.2.2.9 Determining the Most Effective Operating Business Model........................................................ 180 3.2.2.10 Overview of Performing a Cost Analysis.................................................................................181 3.2.3 Recommendations.......................................................................................................................182 3.2.4 References..................................................................................................................................183 3.3 Total Cost of Ownership for a Weather Observing System...................................................184 3.3.1 Introduction............................................................................................................................... 184 3.3.2 Understanding the Total Cost of Ownership...................................................................................185 3.3.3 Key Components of Total Cost of Ownership.................................................................................. 188 3.3.3.1 Initial Investment: Capital Cost ........................................................................................... 188 3.3.3.2 Annual Operating Costs...................................................................................................... 189 3.3.3.3 Annual Maintenance Costs.................................................................................................. 189 3.3.3.4 Replacement Costs............................................................................................................ 190 3.3.4 TCO and Gap Analysis...................................................................................................................191 3.4 Automated Weather Observing Systems ............................................................................. 192 3.4.1 Introduction................................................................................................................................192 3.4.2 The Purpose of the AWOS System..................................................................................................193 3.4.3 Observations in AWOS Systems.................................................................................................... 194 3.4.4 Requirement Specifications for AWOS Systems............................................................................. 196 3.4.5 AWOS Design: Instrument Runways and Precision Approach Runways.............................................197 3.4.6 Site Selection Criteria and Siting Requirements........................................................................... 199 3.4.6.1 Observations for Local Routine and Special Reports and METAR/SPECI....................................... 199 3.4.6.2 Locations for Observation Sites........................................................................................... 200 3.4.6.3 Summary Sensor for Installation Sites.................................................................................. 204 3.4.6.4 Instrument Installation Sites in Practice............................................................................... 205 3.4.6.5 Access to Installation Sites................................................................................................. 206 3.4.6.6 Land Ownership................................................................................................................ 206 3.4.7 AWOS Design for a Single Runway: CAT I, CAT II, CAT III Airports.................................................. 206 3.4.8 Installation Sites........................................................................................................................ 208 3.4.8.1 Weather Station, Met Garden............................................................................................... 208 3.4.8.2 Touchdown Zone and Stop-End Position................................................................................ 209 3.4.8.3 Middle Marker.................................................................................................................. 209 3.4.8.4 Equipment Room or Cloud-Based AWOS................................................................................ 209 Contents    xiii 3.4.8.5 Meteorological Office........................................................................................................ 209 3.4.8.6 Control Tower................................................................................................................... 209 3.4.9 Additional AWOS Equipment........................................................................................................ 209 3.4.9.1 Servers ........................................................................................................................... 210 3.4.9.2 Backup Display Units......................................................................................................... 210 3.4.9.3 Workstations.................................................................................................................... 210 3.4.10 AWOS System Interfaces............................................................................................................. 210 3.4.11 Estimating the Total Cost of Ownership for AWOS.......................................................................... 210 3.4.11.1 AWOS System Costing Introduction....................................................................................... 211 3.4.11.2 Recommended Spares......................................................................................................... 211 3.4.11.3 Optional Items.................................................................................................................. 212 3.4.11.4 Owner Project Costs........................................................................................................... 212 3.4.11.5 Operational Costs..............................................................................................................213 3.4.11.6 Upgrade Strategy...............................................................................................................214 3.4.12 Recommendations.......................................................................................................................214 3.4.13 References..................................................................................................................................215 3.5 Automatic Weather Stations ............................................................................................... 216 3.5.1 The Purpose of Automatic Weather Stations...................................................................................216 3.5.2 Environmental Parameters Measured by AWS................................................................................218 3.5.2.1 Air Temperature.................................................................................................................219 3.5.2.2 Relative Humidity..............................................................................................................219 3.5.2.3 Atmospheric Pressure........................................................................................................ 220 3.5.2.4 Surface Wind................................................................................................................... 220 3.5.2.5 Precipitation..................................................................................................................... 221 3.5.2.6 Global Horizontal Irradiance................................................................................................222 3.5.2.7 Data Loggers.....................................................................................................................222 3.5.3 Determining Geographical Location of Individual AWSs across the Landscape..................................223 3.5.4 Configuring the AWS................................................................................................................... 224 3.5.4.1 Installation Site................................................................................................................ 224 3.5.4.2 Mounting Structures.......................................................................................................... 226 3.5.4.3 Power Supply................................................................................................................... 228 3.5.5 Planning for Operation, Maintenance, and Upgrades of the AWS Network....................................... 228 3.5.5.1 Operational Requirements.................................................................................................. 228 3.5.5.2 Communication Requirements............................................................................................. 229 3.5.5.3 Maintenance of Automatic and Manual Weather Stations......................................................... 229 3.5.5.4 Maintenance of Sensors..................................................................................................... 230 3.5.6 Estimating the Total Cost of Ownership of the AWS Network...........................................................231 xiv    Contents 3.5.6.1 AWS Network Operating, Maintenance, and Life-Cycle Cost Examples..........................................231 3.5.6.2 Personnel.........................................................................................................................233 3.5.7 Recommendations.......................................................................................................................235 3.5.8 References................................................................................................................................. 236 3.6 Upper-Air Systems.............................................................................................................. 237 3.6.1 The Purpose of Upper-Air Systems................................................................................................237 3.6.2 Environmental Parameters Measured during Upper-Air Soundings................................................. 238 3.6.2.1 Air Temperature................................................................................................................ 238 3.6.2.2 Relative Humidity............................................................................................................. 239 3.6.2.3 Upper Wind...................................................................................................................... 239 3.6.2.4 Radiosonde Launch Components.......................................................................................... 239 3.6.3 Determining Where to Place Individual Stations in the Network across the Landscape..................... 239 3.6.4 Configuring and Installing the Upper-Air Station.......................................................................... 240 3.6.5 Planning for Operation, Maintenance, and Upgrades for the Upper-Air Network.............................. 243 3.6.5.1 Operational Requirements.................................................................................................. 243 3.6.5.2 IT Requirements............................................................................................................... 243 3.6.5.3 Maintenance of Upper-Air Stations....................................................................................... 244 3.6.6 Estimating the Total Cost of Ownership of Upper-Air Stations........................................................ 244 3.6.6.1 Procurement Costs............................................................................................................ 244 3.6.6.2 Consumables.................................................................................................................... 244 3.6.6.3 Operator Costs................................................................................................................. 245 3.6.6.4 Maintenance Costs............................................................................................................ 245 3.6.6.5 Administrative Costs......................................................................................................... 245 3.6.6.6 Upper-Air Network Operating, Maintenance, and Life-Cycle Cost Examples.................................. 245 3.6.6.7 Personnel........................................................................................................................ 248 3.6.7 Recommendations...................................................................................................................... 250 3.6.8 References..................................................................................................................................251 3.7 Weather Radar Systems....................................................................................................... 252 3.7.1 The Purpose of Weather Radar Systems.........................................................................................252 3.7.2 Weather Radar Observations....................................................................................................... 254 3.7.3 Determining Where to Place Individual Stations in the Network across the Landscape......................257 3.7.3.1 Configuring the Weather Radar System................................................................................. 259 3.7.4 Configuring the Weather Radar Network....................................................................................... 260 3.7.4.1 Operational Requirements.................................................................................................. 260 3.7.4.2 IT Requirements................................................................................................................261 3.7.4.3 Operational Personnel........................................................................................................261 3.7.4.4 Maintenance.....................................................................................................................261 Contents    xv 3.7.4.5 Upgrades......................................................................................................................... 262 3.7.4.6 Spare Parts...................................................................................................................... 262 3.7.5 Estimating the TCO of Weather Radar Networks............................................................................ 262 3.7.5.1 Procurement Costs............................................................................................................ 262 3.7.5.2 Operational Costs............................................................................................................. 263 3.7.5.3 Weather Radar Operating, Maintenance, and Life-Cycle Cost Examples....................................... 264 3.7.5.4 Personnel Costs................................................................................................................ 266 3.7.6 Recommendations...................................................................................................................... 267 3.7.7 References ................................................................................................................................ 268 3.8 Total Cost of Ownership Exercise........................................................................................269 3.8.1 Introduction............................................................................................................................... 269 3.8.2 Total Cost of Ownership Calculation............................................................................................. 269 3.8.2.1 Header............................................................................................................................. 271 3.8.2.2 Initial Investment (Capital Costs).......................................................................................... 271 3.8.2.3 Total Cost of Annual Operations............................................................................................274 3.8.2.4 Total Cost of Annual Maintenance.........................................................................................275 3.8.2.5 Replacement Costs............................................................................................................ 276 3.8.2.6 Calculate the Total Cost of Ownership Over Lifetime.................................................................277 3.9 Total Cost of Ownership Example Calculation..................................................................... 278 3.9.1 Context for TCO Example..............................................................................................................278 3.9.2 Total Cost of Ownership Example Calculation................................................................................ 280 3.9.3 Total Cost of Annual Operations................................................................................................... 283 3.9.4 Total Cost of Annual Maintenance................................................................................................ 284 3.9.5 Replacement Costs...................................................................................................................... 286 3.9.6 Calculate the Total Cost of Ownership Over Lifetime..................................................................... 286 Part 3 Boxes Box ES3.1 Managing a Meteorological Observing System .............................................................................. 161 Box 3.2.1  ICAO and WMO Key Network Planning References.......................................................................... 175 Box 3.2.2 What Are Systems, Stations, Networks, and Data Centers?.............................................................. 176 Box 3.2.3  Key Elements of a Well-Planned Observation System or Network..................................................... 177 Box 3.2.4  How a Request for Information Works........................................................................................... 181 Box 3.3.1  Possible Goals for Downtimes.......................................................................................................190 Box 3.5.1  Installation of Key AWS Sensors................................................................................................... 227 Box 3.6.1  Performance in Developing Countries Restated.............................................................................. 247 Box 3.6.2  Meteorological Services Canada Contracts Out O&M......................................................................250 xvi    Contents Part 3 Figures Figure 3.3.1  Process of Managing a Meteorological Observing System.............................................................185 Figure 3.3.2  Straight Line Depreciation for a Meteorological Observing System............................................... 186 Figure 3.3.3  Depreciation of a Single Automatic Weather Station....................................................................187 Figure 3.3.4  Overview of the Total Costs of Ownership of a Meteorological Observation System Considering Replacement as Part of Life-Cycle Management.......................................................................... 188 Figure 3.4.1  Schematic of AWOS Functionality.............................................................................................. 194 Figure 3.4.2  Types of Instruments Runways.................................................................................................. 198 Figure 3.4.3  Locations of the RVR and Wind Sensors along a CAT IIIB Runway with More Than 3,000 Meters between the Thresholds................................................................................................. 205 Figure 3.4.4  Example of an AWOS System at an Airport with a CAT I Runway................................................... 206 Figure 3.4.5  Example of an AWOS System at an Airport with a CAT II Runway................................................. 207 Figure 3.4.6  Example of an AWOS System at an Airport with a CAT IIIB Runway.............................................. 208 Figure 3.5.1  Typical Configuration of an AWS Installation............................................................................... 217 Figure 3.5.2  Radiation Shields......................................................................................................................219 Figure 3.5.3  Types of Wind Sensors............................................................................................................... 221 Figure 3.5.4  Precipitation Gauges................................................................................................................. 221 Figure 3.5.5  Solar Radiation Sensor...............................................................................................................222 Figure 3.5.6  Campbell-Stokes Sunshine Recorder ..........................................................................................222 Figure 3.6.1  Manual Launch Site...................................................................................................................241 Figure 3.6.2  Automated Launch Site............................................................................................................. 242 Figure 3.7.1  Beam Height as a Function of Range...........................................................................................257 Figure 3.7.2  Layout and Setup of a Typical Weather Radar Installation........................................................... 259 Figure 3.8.1  Total Cost of Ownership Summary Form..................................................................................... 270 Figure 3.9.1  TCOSF Example 1 Calculation..................................................................................................... 289 Figure 3.9.2  TCOSF Example Calculation (1 and 2 Combined)......................................................................... 290 Part 3 Photos Photo 3.1.1  Images of a Radar Site................................................................................................................ 169 Photo 3.5.1  View of Environment Canada AWS Station near Lacombe, Alberta, Canada (WMO ID: 71242) ...........225 Photo 3.5.2  AWS with 10-Meter Mast and Security Fence.................................................................................227 Photo 3.6.1  Human-Operated Launch Site...................................................................................................... 242 Photo 3.6.2  Automatic Sounding System....................................................................................................... 243 Photo 3.7.1  NEXRAD Located at the National Weather Center (NWC) in Norman, Oklahoma ................................253 Part 3 Tables Table ES3.1  Station Uptimes for Developed and Developing Countries.............................................................. 158 Table ES3.2  Measurement Type versus Accuracy and Cost for AWS...................................................................160 Contents   xvii Table ES3.3  Summary Benchmark Cost for Observation Systems, Developed Countries...................................... 163 Table ES3.4  Summary Costs for Observation Systems, Developing Countries..................................................... 163 Table 3.1.1  SAWS Infrastructure Management Challenges in 2020................................................................... 167 Table 3.2.1  Station Uptimes for Developed and Developing Countries.............................................................. 173 Table 3.2.2  Measurement Type versus Accuracy and Cost for AWS.................................................................... 174 Table 3.4.1  Meteorological Parameters Required for Aviation Meteorological Applications................................ 195 Visibility, RVR, and Decision Height Conditions for Precision Approach Runways............................198 Table 3.4.2  Table 3.4.3  Observation Location Requirements for Local Routine Reports, Special Reports, METAR, and SPECI.......................................................................................................................199 Table 3.4.4  Installation Site Recommendations for RVR Instruments................................................................201 Table 3.4.5  Number of Mandatory and Recommended RVR Sensors per Runway Type.........................................201 Table 3.4.6  Installation Site Recommendations for Wind Speed and Direction Measurements............................202 Table 3.4.7  Installation Distances from the Center Line of the Runway for RVR Sensors and Wind Masts, Frangible and Non-Frangible Mounting..........................................................................................203 Table 3.4.8  Installation Site Recommendations for Cloud Measurements..........................................................203 Table 3.4.9 Installation Locations for Weather Sensors for Non-Precision Approach Runways band CAT I, CAT II, and CAT III Runways........................................................................................................ 204 Table 3.4.10  Single Runway Costing for Each Category Type............................................................................ 211 Table 3.5.1  AWS Annual Network Cost of Operations and Maintenance Data for Example Countries.................... 232 Table 3.5.2  Full-Time Equivalent Staff Numbers for Example Developed and Developing Countries.....................234 Table 3.6.1  Upper-Air Network Cost of Operations and Maintenance Data for Example Countries.......................346 Table 3.6.2  Costs Associated with Balloon Launches.......................................................................................248 Table 3.6.3  Full-Time Equivalent Staff Numbers for Example Countries.............................................................249 Table 3.7.1  Common Specifications of Weather Radar ..................................................................................... 255 Table 3.7.2  Estimated Procurement Costs for Single X-Band, C-Band, and S-Band Weather Radar Systems with 1.0° Beam Widths.................................................................................................................263 Table 3.7.3  Weather Radar Cost of O&M Data for Benchmarking and for Developed and Developing Countries....265 Table 3.7.4  Developing Country Weather Radar Uptime Performance................................................................265 Table 3.7.5  Full-Time Equivalent Staff Numbers for Example Countries.............................................................266 Table 3.9.1  Travel Time Needed for Site Visits................................................................................................279 © Jurkos | istock.com    xix Foreword Hydrological and meteorological processes shape the natural environment and have a pro- found impact on how well people live. Floods, droughts, and extreme weather disrupt social and economic development and must be dealt with within the framework of a changing cli- mate. The World Bank’s development agenda is intimately linked, therefore, to hydrological and meteorological monitoring and forecasting to anticipate threats and opportunities posed by the environment. Application of this knowledge can ameliorate risks by Florida Straits and Grand Bahama Bank. Photo: NASA protecting the vulnerable and reducing exposure. This knowledge depends on the ability to accurately predict the impact of hydrological, weather, and climate events, which in turn requires observations of the entire Earth system from space, land, and sea. It is a highly collaborative venture involving the best efforts of all nations to monitor the environment and share international- ly a core, standardized set of observations to support the forecasting system. Albeit highly beneficial, installing and maintaining surface-based hydrologi- cal and meteorological observation networks is a costly undertaking for many low- and middle-income countries and one that is often underfunded and ne- glected. Data gaps grow and the quality of critical forecasts declines or is never achieved, and the adverse impact on national development is palpable. The World Bank has invested over US$1.5 billion in hydrological and meteo- rological modernization projects over the last two decades. While social and economic benefits are realized, there is growing evidence that the observing systems component of these investments is a weak link that often fails prema- turely. New guidance on how to build and sustain this critical infrastructure is needed. xx    Foreword Commissioned by the World Bank, Charting a Course for Sustainable Hydrological and Meteorological Networks in Developing Countries aims to help nations and development partners design fit-for-purpose networks. The re- port considers both the technical and financial requirements to ensure that all observing systems work to the limits of its design life and do not become inoperable early due to poor maintenance. It introduces the fundamental con- cept of total cost of ownership, which should influence future procurement practices by emphasizing life cycle costs over initial capital costs. The report examines the selection of instrumentation, staffing requirements, and the costs of operations and maintenance. Charting a Course for Sustainable Hydrological and Meteorological Networks in Developing Countries is a new instrument in the toolbox that supports the World Bank’s investment portfolio. We commend it and trust that our devel- opment partners will also find it informative in their efforts to climate-proof their investments. Bernice K. Van Bronkhorst Global Director of Urban, Resilience and Land Practice the World Bank Group    xxi Acknowledgments The motivation for this report is to provide guidance to NMHS managers in developing countries and to the development institutions supporting NMHSs’ modernization efforts to build, operate, and maintain sustainably hydrologi- cal and meteorological observation networks. The report was prepared by a World Bank team led by Vladimir Tsirkunov (Lead Specialist, Head of the GFDRR Hydromet Program, WBG) with the core team comprising Brian F. Day (Lead Meteorological Consultant, WBG; Chair Emeritus of the Association of Hydro-Meteorological Industry, HMEI; VP of Campbell Scientific), R. David Grimes (Lead Meteorological Consultant, WBG; former WMO President and Head of the Meteorological Service of Canada, MSC), David Canal of Castile in Calahorra de Ribas, Spain. Photo: Cesar J. Pollo P. Rogers (Lead Meteorological Consultant, WBG; former Chief Executive of the UK Met Office), Andreas Schumann (Lead Hydrological Consultant, WBG; Senior Professor, Ruhr-University Bochum), and Robert A. Varley (Lead Meteorological Consultant, WBG; former Chief Executive of the UK Met Office). Additional authors contributing to various sections of the report include Marjory-Anne Bromhead (Lead Institutional Consultant, WBG), Christopher Cameron-Hann (Aegaea Limited), Joshua Campbell (HMEI, Campbell Scientific), Jeffrey Kavanaugh (HMEI, University of Alberta), Foeke Kuik (HMEI, Campbell Scientific), Matti Lehmuskero (HMEI, Vaisala), Alain Pietroniro (University of Calgary, former Executive Director, the National Hydrological Service of Canada), Ashish Raval (HMEI, Synoptic Data PBC), Yuri Simonov (Hydrometcentre of Russia), and Pekka Utela (HMEI, Vaisala). The authors are very grateful to Adrian Gerhard (WMO President and President of the German Weather Service, DWD); Bryan Hodge (Chief Engineer and General Manager, Observing Systems & Operations, Bureau of Meteorology, Australia); Michael Staudinger (former President of WMO Region VI and Head of the Central Institute for Meteorology and Geodynamics, ZAMG, Austria); and Stuart Goldstraw, Tim Oakley, and Bruce Truscott (all of the UK Met Office) for their significant contributions to collecting, presenting, and describing baseline data and the main aspects of operations and maintenance of me- teorological observations networks of Germany, Australia, Austria, and the United Kingdom. Valuable assistance in collecting and interpreting data from developing countries was provided by Jerry Lengoasa (former WMO Deputy xxii    Acknowledgments Secretary General and former CEO of the South African Weather Service), Guillermo Navarro (former President of WMO Region III and former Director of Meteorological Directorate of Chile), and Deon Terblanche (former WMO Director of Research). This report benefitted from comments and other inputs provided by Dominique Berod (Head, Earth System Monitoring Division, Infrastructure Department, WMO), Tatiana Dmitrieva (Head of International Relations, FGU Hydrometservice), Jack Hayes (former Head of the US National Weather Service), Fabia Huesler (Deputy Head of Section, Federal Office for the Environment, Hydrology Division, DETEC, Switzerland), Michel Jean (President of WMO Infrastructure Commission), Naohisa Koide (Scientific Officer, Japan Meteorological Agency), K.J. Ramesh (Regional Integrated Multi-Hazard Early Warning Systems for Africa and Asia, RIMES, and former Director General of India Meteorological Department), Anthony Rea (Director, Infrastructure Department, WMO), and S. Srinivasan (RIMES). The team benefitted from the discussions with and contributions from ex- perts representing multiple organizations including the WMO Secretariat, EUMETNET, SAWS, UK Met Office, US NWS, ZAMG, BoM, MSC, and HMEI, among others. The following World Bank peer reviewers provided valuable comments and recommendations, which the team has taken into account: Arati Belle (Senior Disaster Risk Management, DRM, Specialist), Anju Gaur (Senior Water Resource Management, WRM, Specialist), Daniel Kull (Senior DRM Specialist), Makoto Suwa (Senior DRM Specialist), and Noosha Tayebi (WRM Specialist). The team also would like to thank Tetiana Shalkivska (Hydromet Consultant, Global Facility for Disaster Reduction and Recovery, GFDRR), Erika Vargas (Senior Knowledge Management Specialist, GFDRR), Moussa Sidibe (DRM Specialist, GFDRR), and Stephen Miller (Hydromet Consultant, GFDRR), for their help in the preparation of the book. The team thanks Niels H. Nielsen, Practice Manager of the Global Facility for Disaster Reduction and Recovery, who supported this work. The authors are very grateful for excellent editorial support provided by Laura Wallace (substantive editing) and Hope Steele (technical editing). The team is also thankful to ULTRA Designs, Inc. for creative graphic design of the report. This publication significantly benefitted from support provided by the Bureau for Humanitarian Assistance, U.S. Agency for International Development, to the World Bank’s Global Facility for Disaster Reduction and Recovery (GFDRR) under the terms of Award No. AID-OFDA-IO-17-00058. The opinions ex- pressed in this publication are those of the author(s) and do not necessarily reflect the views of the U.S. Agency for International Development.    1 LEAD AUTHORS R. David Grimes Lead Meteorological Consultant, WBG; former WMO PART 1. President and Head of the Meteorological Service of Canada, MSC David P. Rogers Lead Meteorological Consultant, WBG; former Chief Executive of the UK Met Office A Vision: Charting a CONTRIBUTING AUTHORS Brian F. Day Lead Meteorological Consultant, WBG; Chair Course for Sustainable Hydrological and Emeritus of the Association of Hydro-Meteorological Industry (HMEI); Campbell Scientific, Canada Andreas Schumann Meteorological Lead Hydrological Consultant, WBG; Senior Professor, Ruhr-University Bochum, Germany Robert A. Varley Observation Networks Lead Meteorological Consultant, WBG; former Chief Executive of the UK Met Office Vladimir V. Tsirkunov in Developing Lead Specialist, Head of the GFDRR Hydromet Program, WBG Countries 2    Vision for Sustainable Meteorological and Hydrological Observation Networks Weather, climate, and hydrological observation networks are built to last. They are designed to be “fit for purpose” and “fit for national budgets.” They are maintained through effective institutional arrangements and long- term financial and capacity development commitments, thereby contributing to the sustained delivery of high- quality meteorological and hydrological services for the benefit of the whole of society.    3 Part 1 Abbreviations ADCP accoustic Doppler current profiler AI artificial intelligence AMCOMET African Ministerial Committee on Meteorology API application programming interface AWOS automated weather observing system AWS automatic weather station CARICOM Caribbean Community COP26 2021 UN Climate Change Conference of the Parties DHM Department of Hydrology and Meteorology (Nepal) ECMWF European Centre for Medium-Range Forecasting EUMETNET European Meteorological Network EUMETSAT European Organization for the Exploitation of Meteorological Satellites FTE full-time equivalent GBON Global Basic Observing Network GDPFS Global Data-Processing and Forecasting System GFDRR Global Facility for Disaster Reduction and Recovery HMEI Association of Hydro-Meteorological Equipment Industry hydromet hydrological, meteorological, and hydrometeorological ICAO International Civil Aviation Organization ICT information and communication technology IMD India Meteorological Department IoT Internet of Things IPSASB International Public Sector Accounting System Board IWRM integrated water resources management LCC life-cycle costing LDCs least developed countries NHS National Hydrological Service NMHS National Meteorological and Hydrological Services NMS National Meteorological Service NWP numerical weather prediction O&M operation and maintenance OECD Organization for Economic Co-operation and Development OEM original equipment manufacturer OMVS Senegal River Basin Development Organization PFI public finance initiative PFM public financial management 4    Part 1 Abbreviations RSMC Regional Specialized Meteorological Centres RWC Regional WIGOS Centres SAHF South Asia Hydromet Forum SAWS South African Weather Service SIDS small island developing states SIM subscriber identification module SMS short message service SOFF Systematic Observations Financing Facility TCO total cost of ownership UK-EOF UK Environmental Observation Framework UNDP United Nations Development Programme UNEP United Nations Environment Programme WIGOS Regional WMO Integrated Global Observing System WMO World Meteorological Organization All dollar amounts are US dollars unless otherwise indicated. Unless otherwise noted, the sources of all figures are the authors of this publication.    5 Part 1 Executive Summary Purpose and Audience The aim of this report is to explore in depth the root causes of why national meteorological and hydrological systems fall into disrepair—even after being modernized—and to share a visionary approach with recommendations to achieve more sustainable outcomes. The report is organized into three parts, with the first providing the larger context and a proposed vision and the other two delving into the more technical aspects of meteorological and hydrolog- ical networks, offering guidance and recommendations on their design and sustainable operations over the long term. It builds on successful outcomes in developed nations, which not only offer insights for many developing coun- tries but also help development agencies integrate the principles and condi- Mooserboden Reservoir, Austria.Photo: Rusm tions of success into the design of their own development projects. The target audience is professionals working on weather and climate invest- ment programs, including National Meteorological and Hydrological Services (NMHS) staff and directors, staff in the core planning and finance ministries, and professionals in development institutions. Hydrological and Meteorological Value Chain Hydrological and meteorological (subsequently referred to as hydromet) data and services are recognized as key building blocks to support actions that (1) safeguard life and property caused by extreme weather and climate events; (2) build better resilience to the consequences of natural disasters; (3) support more effective planning, preparedness, and robustness for economic develop- ment; and (4) underpin responsible climate action. With population growth, urbanization, and the scale and scope of the changes in climate being ob- “… this report spells out a served and expected, these building blocks are becoming even more valuable. vision … where weather, Economic studies have reinforced the significance of these data and services to national economies, showing benefit-cost ratios ranging from 10:1 to 20:1. climate, and hydrological observation networks are Measurement data are the cornerstone of the hydromet value chain, as de- picted in figure ES1.1. Routine, long-term monitoring of meteorological and built to last … contributing hydrological conditions are foundational to predicting changes over time to the sustained delivery on scales of moments to seasons, and they are equally vital to projecting and understanding impacts of longer-term changes in climate. The impetus of high-quality hydromet that affects the changing patterns of weather, climate, and water resources respect no political borders, and thus international cooperation for the ex- services for the benefit of change, access, and use of these data go a long way toward improving the the whole of society.” 6    Part 1 Executive Summary predictability of these conditions at national and local levels. benefit from more timely early warnings, more efficient eco- Currently, with improved access to global hydrological and nomic decision-making, adapting to more climate-smart built meteorological data and forecasts, all national economies can infrastructures, and more environmentally sensitive policies. FIGURE ES1.1  Hydromet Value Chain Key outcomes and Functions and processes Output services Bene ciaries bene ts Livelihoods of people and communities Community and Climate civil society Informed land-use Water Environmental planning Weather intelligence Resilient communities Monitoring and Decision-makers observation in economic and Health, food, and data Risk mitigation environmental sectors water security Decision- Services making Wise use of natural Governments and Modeling Economic other policy resources and e ciency makers Environmental data analytics stewardship Weather prediction Third-party service E ective Water assessments policies providers Safe and e cient Climate scenarios ampli ers of sustainable transportation development bene ts Energy generation and access Reduced carbon Institutional capacity (resources, people, and partnerships), information and communications technology footprint and mainstreaming advances in innovation and science Performance enablers Note: Observations are the foundation of the hydromet value chain. While there are many other depictions of the hydromet value chain, this one draws together the benefits and institutional capacity to support it. The colors serve only to differentiate the groups of elements. Challenges Why Networks Fail While this report presents an urgent and compelling case While the causes of these failures can be numerous, the pri- for investment to build capacity and hydromet observation mary reasons are insufficient budgetary resources and lim- networks, many countries struggle to modernize or even ited human resources to operate, maintain, and replace as maintain these networks. Over the past 20 years, the World necessary the hydrological and meteorological observing Bank and development agencies have invested billions of US infrastructures and systems. This report analyzes the finan- dollars to rebuild and strengthen hydrological and meteoro- cial data from a sample of 14 upper-middle- to high-income logical observation networks and services in the developing developing countries, where budgets appear to be insuffi- world. But in many instances, these investments have not cient to support their networks. These data are compared to realized their intended benefits, with systems failing prema- those of benchmarking countries (Australia, Austria, Canada, turely. As a result, the return on investment for the country Germany, the United Kingdom, and the United States) that are and the agencies has been suboptimal or lost—or, even worse, following financial and management practices that are sup- these same agencies are re-investing in these countries to re- porting their observation networks over the long term. While store or rebuild some of the very same monitoring networks. not an exhaustive sample from the developing world because financial data are simply not available in many countries, Part 1 Executive Summary    7 the results are indicative and support the conclusion that fi- nancial and human resources are limited to maintain these BOX ES1.1  Key Results from Benchmarking against More networks, particularly for more advanced observing systems Sustainable Network Practices such as radar and radiosonde (upper air) operations (see box ES1.1). This benchmarking analysis shows that while many hy- dromet development projects are well intended, it is often At the outset of a modernization project, the program is not difficult for national governments to maintain the observa- tional infrastructure beyond the early years of operations. always well designed to fit within the national circumstances In particular, national budgets will need to be increased of the country. For example, introducing new stations—with- to support upper-air stations and radar operations, along out paying attention to the existing infrastructure—puts more with hydrological systems. Key findings include: financial stress on operating budgets, which may result in ■ Overall, the performance and operating costs among less available funding for the network’s long-term operations. sample benchmark countries are very similar, providing Poor awareness of the total costs of operating, maintaining, some confidence on the estimated costs of supporting and replacing the infrastructure often leads to a government sustainable network operations. Uptimes for exceeding its financial means and its motivation to support it meteorological systems are generally above 95 percent. in the face of competing priorities. Other challenges include: ■ The operating costs (without labor) for meteorological stations show the scope of national budgets required over the observation instrument/ ■ Insufficient access to spare parts system life cycle, ranging from $6,200 for automatic ■ Greater pressure on staff, given insufficient financial resourc- weather stations (AWSs) to $200,000 for upper-air es to attract and retain the needed scientific, technical, and stations. For hydrometric stations, these costs range management expertise to maintain and sustain networks from $4,000 to $15,700. ■ Lack of clarity about “who does what” and “who supports ■ Total operating costs (including labor) for what” in terms of public good roles, responsibilities, and meteorological stations range from $12,100 for AWSs to $424,000 for upper-air manual stations. resourcing expectations For hydrometric stations, the range is $9,000 to ■ Inadequate institutional structures $24,700. ■ Inefficient management and operations ■ Staff costs for meteorological sites can be calculated ■ Lack of a national framework to leverage capabilities and for any country based on their salary costs per contributions from other contributors at subnational, na- staff full-time equivalent (FTE). In benchmarking tional, or international levels countries, one FTE adds about $5,500 in labor cost per site per year. For meteorological stations, ■ Poor coordination or awareness of existing infrastructures these costs range from 0.08 FTEs for AWSs to 2.8 among development agencies. FTEs for upper-air systems. Sites should be visited ideally four to six times every year to ensure sound Vision for a More Holistic and Sustainable performance and support preventive maintenance. Approach ■ While staffing levels are similar for hydrometric stream gauges, non-salary costs are more variable where site visits are more frequent and where While overcoming these challenges poses big hurdles for locations are often more remote and access more many developing countries, this report spells out a vision to difficult. facilitate a better outcome—one where weather, climate, and ■ Annualized replacement or reinvestment costs hydrological observation networks are built to last; where range from $5,600 for AWSs to $230,000 for radar they are designed to be “fit for purpose” and “fit for national systems. For hydrometric systems, these costs are budgets”; and where they are maintained through effective about $5,500 to $9,000. ■ Investment and capital costs for meteorological institutional arrangements and long-term financial and ca- stations range from $56,000 for AWSs to $2.5 million pacity development commitments—thereby contributing to to $4 million for radar systems. For hydrometric the sustained delivery of high-quality hydromet services for stations, the range is $28,000 to $45,000. the benefit of the whole of society. 8    Part 1 Executive Summary The vision outlines an approach that (1) better accounts for ■ Clarify roles and responsibilities of various actors in the country context (such as limited resources and human re- provision of monitoring and services. Legal mandate and source challenges), (2) sets out a clear understanding of roles policy frameworks provide a level playing field for all and financial and human resource expectations, (3) adopts a concerned by clarifying authorities of public sector insti- more incremental and systematic approach to network mod- tutions, framing expectations for private sector participa- ernization, and (4) fosters more collaborative approaches tion, and ensuring open access to data to further exploit among in-country partners and development agencies. It also their economic value (Rogers et al. 2021b). encourages government leaders, development partners, and ■ Attract and retain scientific, technical, and management directors of NMHSs to: expertise that is vital to support network management, procurements, operations, maintenance, and replacement. ■ Adopt an all-inclusive approach to managing and mod- Competitive compensation and motivating workplace en- ernizing monitoring systems that follows an integrated vironments help attract and retain a skilled workforce. In fit-for-purpose approach by engaging various state actors, addition, enhanced leadership and financial management stakeholders, and development partners; follows best instill confidence in NMHS administrators to make the fi- practices of co-design and co-production (a system-of-sys- nancial case for additional investments. tems approach); and harnesses the principles of effective ■ Promote actions that foster further socioeconomic bene- governance, affordability, and transparent financial ac- fits from national monitoring network operations, thereby counting and management. multiplying the benefits and distributing their costs over a ■ Adopt improved financial and life-cycle management prac- larger number of societal sectors, along with building sup- tices that foster a clearer understanding of the total cost port for ongoing national investments in these networks. of ownership (TCO) of network systems and the long-term financial expectations of these systems on governments. While these best practices are drawn from mostly developed These practices are rooted in life-cycle costing (LCC) and countries, they can be adapted to reflect the developing world activity-based methodology that capture the on-going context, and thereby help guide government decision-making state of the observation equipment, depreciation, residual and implementation. value, and costs of disposal. ■ Benchmark against other countries that have been suc- In sum, this report emphasizes a holistic approach to network cessful at sustaining their hydromet observation networks. management—that is, an enterprise solution—that includes Benchmarking aids in the decision-making process by fol- all facets of hydromet observing systems. This shift away lowing best practices; supporting effective planning; and from a simple procure-and-install mindset is critical to suc- setting realistic financial, human resources, and manage- cessful development of future sustainable networks. The re- ment estimations of the costs over the life cycle of the port presents an urgent and compelling case for investment to infrastructure. build capacity in national hydromet observation networks. It ■ Implement the most appropriate business model decisions also underscores the key conditions for sustainable networks for managing observation networks, which are based on over the long term. These include adopting appropriate pol- national circumstances. This is not necessarily a one-size- icies and arrangements; mobilizing the financial and human fits-all solution as different make-or-buy business models resources; and adopting a whole-of-society approach, which could apply to different observational systems and the requires collaboration across government and the private national circumstance and capacity available to support sector, and from individuals to international organizations. them. ■ Establish priorities that are fit for budget by aligning with financial expectations now and in the future. This also pro- vides for step-by-step investment strategies that provide for making choices that may yield lower operating and downstream capital replacement costs.    9 Setting the Scene 1.1.1 Introduction 1.1 Over the past 20 years, developing countries have invested in upgrading hy- drological and meteorological networks, often with the assistance of devel- Lake Tekapo, New Zealand. Photo: © Helena Bilkova | Dreamstime.com opment partners. In most of these projects, the share of the investment in the modernization of networks has been between 40 and 50 percent of the total project costs. The objectives of these initiatives have been to create reliable analyses, numerical predictions, and forecasts to inform early action, response, and planning across the whole of society. In some countries, monitoring networks have been sustained and improved over the decades. But in others, maintaining them operationally has remained elusive, resulting not only in inoperable or poorly maintained observational infrastructure and systems but also in a failure to realize the intended ben- efits. Why did some succeed where others did not? That is a question that this report tries to answer by exploring the underpinnings of the successes and the possibilities of replicating these successes elsewhere—and thereby contribute to the body of knowledge on observation networks. The report includes: “This report aims to facilitate the development ■ Part 1: A Vision: Charting a Course for Sustainable Meteorological and Hydrological Observation Networks in Developing Countries (this part of the of more strategic and viable report). roadmaps for investments ■ Part 2: Recommendations for the Design of Sustainable Hydrological Observation Networks in Developing Countries. The second part of the report in weather and climate describes the building blocks for the effective design and procurement of hydrological gauging systems. observation networks ■ Part 3: Recommendations for the Design of Sustainable Meteorological where those investments Observation Networks and Systems in Developing Countries. This describes the building blocks for the effective design and procurement of the most are likely to be substantial common land-based observation networks (automated weather observing in the coming decades, as systems (AWOSs), automatic weather stations (AWSs), upper-air systems, and weather radar systems. countries improve resilience to natural hazards and This report aims to facilitate the development of more strategic and viable roadmaps for investments in weather and climate observation networks economies transform in where those investments are likely to be substantial in the coming decades, as countries improve resilience to natural hazards and economies transform response to climate change in response to climate change challenges. challenges.” 10    Setting the Scene The key message of this compendium is that a full under- parameters inform advances in science, prediction, and appli- standing of the financial costs of operation, maintenance, and cations. Weather systems, water cycles, and climate patterns renewal—along with the accompanying necessary institution- respect no geopolitical boundaries, and, as such, no one nation al and human resources capabilities—is central to sustaining can do it alone. Each nation relies on others to systematical- observation networks. ly monitor, assess, and exchange information on the altering states of weather, water, and climate conditions. In doing so, all Part 1, the Vision, begins with a look at why sustainable ob- benefit. Forecasting changes in these states of weather, water, servation networks matter and the root causes of why they and climate require collective understanding; international co- fail (chapter 1.1). A vision for realizing more sustainable net- ordination; and the sharing of measurements, standards, and works is then proposed (chapter 1.2); the Vision concludes protocols for the exchange of observational data.1 with recommendations for guiding and supporting deci- sions and policies of the national governments and National In recent years, there have been remarkable achievements in Meteorological and Hydrological Services (NMHS), the World modeling and predicting changes in weather, supply of water Bank, and other development agencies (chapter 1.3). resources, and climate, thanks to significant progress in high-performance computing, coupled with science and sys- 1.1.2 Why Sustainable Networks Matter tematic observations. Modeling systems offer an interopera- ble platform for data integration and prediction and are now From Copenhagen in 2009 to Paris in 2015 and Glasgow in a more direct gateway to intelligent application programming 2021, world leaders have been gathering in a call for action interface (API) applications,2 supported by data analytics in the face of growing societal consequences of a changing and artificial intelligence (AI). Global and regional enterprise climate—one in which past hydrological and meteorological approaches are bringing new intelligence, capabilities, and (hydromet) extremes are no longer indicative of the severity science tools to developing countries. of what is likely to come. Developed and developing countries alike now recognize the benefits of sustainable hydrological To this end, the World Meteorological Organization (WMO) and meteorological networks; developing countries are ac- Extraordinary Congress in 2021 adopted three key resolu- tively seeking to direct very significant financial support to tions: (1) a Unified Data Policy for International Exchange install and modernize their networks. The challenge for many of Earth System Data (weather, climate, hydrology, ocean, of these countries, which this report seeks to address, is to atmospheric composition, cryosphere, and space weather); ensure that these investments are sustained and improved on (2) the Global Basic Observing Network (GBON); and (3) the over time through a better understanding of and accounting Systematic Observations Financing Facility (SOFF). The goals for the human and financial resources needed to maintain the behind these policy measures are to improve the acquisition, networks. These benefits can be grouped into four areas: (1) access, and exchange of data and information—and thereby assessing environmental change, (2) creating societal and en- improve hydromet predictions—to better support all WMO vironmental benefits, (3) managing disaster and water risks, Members, especially least developed countries (LDCs) and and (4) framing international policy agreements. small island developing states (SIDS). As part of this effort, the WMO has launched a new web portal to the Global Data- 1.1.2.1 Assessing Environmental Change Processing and Forecasting System (GDPFS) to make key me- teorological analyses and forecast products from Members’ Monitoring and observation are fundamental to all sciences numerical weather prediction (NWP) systems more readily but none more than hydrology and meteorology, where regu- accessible (figure 1.1.1).3 lar, consistent measurements of hydrological and atmospheric 1 For meteorology and operational hydrology, this role is fulfilled by the World Meteorological Organization (WMO), a specialized agency of the United Nations (UN). Information about WMO can be found at https://public.wmo.int/en. 2 API comprise a set of procedures that permit the creation of applications by accessing data or features of an operating system, application, or other service. 3 The WMO Information System, coupled with WMO Global Data-Processing and Forecasting System (GDPFS), offer real-time free and open access to hydrological and meteorological data and numerical prediction models from its Members. Setting the Scene    11 FIGURE 1.1.1  Sample Page from WMO Real-Time Global Data-Processing and Forecasting System Modelling Centre Source: Designated GDPFS Centres Web portal for the Global Data-Processing and Forecasting System, https://wmo.maps.arcgis.com/apps/dashboards/7c- 3d45e5003a417988bad63e91ad8748. Note: ECMWF = European Centre for Medium-Range Forecasting; GDPFS = Global Data-Processing and Forecasting System; RSMC = Regional Specialized Meteorological Centres. 1.1.2.2 Creating Societal and Environmental Benefits services) to the decision-making processes by the key actors Hydromet services offer significant benefits for the welfare of (policy and decision-makers, civil society, and advanced ser- people, national governments, and both public and private vice providers in the private sector and academia). The obser- sector institutions, as shown in the hydromet value chain (fig- vations and models provide vital information services (such ure 1.1.2). These services strengthen society’s resilience to as data, forecasts, and early warnings), which lead to risk natural hazards and climate change—and they deliver signif- mitigation strategies; intelligence on the pace of environmen- icant economic and environmental outcomes that are critical tal change and biodiversity; improved economic efficiency; in building new green economies worldwide. These benefits and effective social, economic, and environmental policies. fall into a wide range of areas, including transportation, food Moreover, this chain is a shared enterprise, where public and production, public health, water resources management, eco- private sector entities offer services, often for mutual benefit. system based–adaptation actions, coastal zone management, Nonstate actors can generate significant benefits to support land use planning, and renewable energy. Negotiations at the public sector services such as instrumentation and information 2021 UN Climate Change Conference of the Parties (COP26) and communication technology (ICT) systems, as well as am- made progress toward recognizing the importance of ecosys- plify the value of the infrastructure investments through tai- tem-based actions for delivering benefits for climate adapta- lored services. tion, mitigation (including both sequestration and reduced emissions), and water security. Value is generated through integrated service interfaces, which draw together hydrological and meteorological data with other The hydromet value chain describes the relationship among relevant application information. For example, agriculture the fundamental inputs (observations, modeling, and analytical relies on insights from both meteorology and hydrology for 12    Setting the Scene FIGURE 1.1.2  Hydromet Value Chain Key outcomes and Functions and processes Output services Bene ciaries bene ts Livelihoods of people and communities Community and Climate civil society Informed land-use Water Environmental planning Weather intelligence Resilient communities Monitoring and Decision-makers observation in economic and Health, food, and data Risk mitigation environmental sectors water security Decision- Services making Wise use of natural Governments and Modeling Economic other policy resources and e ciency makers Environmental data analytics stewardship Weather prediction Third-party service E ective Water assessments policies providers Safe and e cient Climate scenarios ampli ers of sustainable transportation development bene ts Energy generation and access Reduced carbon Institutional capacity (resources, people, and partnerships), information and communications technology footprint and mainstreaming advances in innovation and science Performance enablers Note: Observations are the foundation of the hydromet value chain. While there are many other depictions of the hydromet value chain, this one draws together the benefits and institutional capacity to support it. The colors serve only to differentiate the groups of elements. decisions about the best time for planting and the application ■ Hydrological and meteorological measurements support of inputs, predicting crop growth, and assessing the state of water resource management, reservoir planning and oper- soils and markets. Intelligence is created through the integra- ations, flood plain mapping, drought risk assessments, and tion of this information, which supports smart farming deci- regulatory actions (such as water licensing agreements). sions and applications. Together, public and private sectors ■ Long-term local and global climatological records under- can reduce duplication, share costs, improve efficiencies of pin infrastructure design and planning and are used to de- operation, and provide long-term sustainable benefits (figure velop long-term climate projections that support climate 1.1.3). This can be achieved by adopting collaborative practices mitigation and adaptation planning. and principles based on governance, affordability, co-produc- ■ Real-time observations and long-term records support effi- tion, and management (Rogers et al. 2021a). cient economic operations and infrastructure design (such as farming, fishing, renewable energy platforms, and trans- Societal needs strongly influence the priorities for these sys- formative low carbon energy strategies and applications). tems, as do advances in science and innovation. Moreover, the need for these data services is becoming more acute as As part of stepped-up efforts on climate, the type of data society adopts meaningful climate action and builds defenses needed is changing. Many decisions have a climatological to mitigate the impacts of climate change. For example: time scale and require long observation series of historical data, as provided by well-established and sustained hydro- ■ Real-time observations contribute to accurate short-term logical and meteorological networks. Reducing carbon foot- warnings and forecasts of imminent danger to life and prints through fuel switching, electrification, and increasing property from storms and floods. carbon sinks through reforestation and adaptive agricultur- al practices are introducing new requirements and uses for Setting the Scene    13 FIGURE 1.1.3  The Value Chain as a Shared Enterprise among the Public, Private, and Academic Sectors Environment Public sector NMHS Health and Sustainable safety development Science Technology Service Innovation Expertise Capacity Academia Private sector Society Economy Wealth creation meteorological and hydrological data (see box 1.1.1). In ad- BOX 1.1.1  How Weather, Water, and Climate Data Support dition, adaptive management requires consistent environ- Low-Carbon Footprints mental monitoring so that management decisions within a transient climate system can adapt to optimize socioeconom- ic opportunities and minimize risks. Under the Paris Agreement, and reinforced at COP26 in November 2021 in Glasgow, all Parties are expected to navigate to net zero carbon futures—that is, a state in At the same time, environmental data and information are be- which the greenhouse gases going into the atmosphere coming a mainstream commodity in the emergent information are balanced by their removal from the atmosphere. economy and the transition to a low-carbon economy. Two- Reliable and accessible weather, water, and climate data thirds of the global economy is more or less dependent on and prediction over all time scales play a vital part in weather conditions (Munich RE 2021b). Fueled by demands the planning and day-to-day operations of renewable and opportunities driven by the Internet of Things (IoT), ad- energy-based economies, including wind, solar energy, vances in big data analytics, AI, and information technologies, and hydroelectric power. These data also inform other together these advances are transforming decision-making low-carbon economic strategies, including increasing carbon sinks by using improved forestry and agricultural practices. These advances are driving new economic opportu- management practices; improved urban planning, land nities and yielding improved economic performance against use, and land use change; and improved efficient trans- a backdrop of climate change, decarbonization, and disaster portation and commerce supply chain systems. risk. Information from systematic and sustained surface mon- At COP26, increased resources were pledged to support itoring of water, weather, and climate are a vital source of this nature-based solutions and adaptation, particularly to new intelligence, yielding benefits to this emerging economic help more vulnerable countries mitigate climate hazard sector (Alliance for Hydromet Development 2021). risks and emerging food and water security risks. 14    Setting the Scene These new opportunities are creating incentives for greater 2017) estimate the benefit-cost ratio of the United Kingdom participation by the private sector to contribute and serve and Australian meteorological services to their economies to these markets. With government and market strategies to be 14:1 and 11:1, respectively; and their services could pro- monetize carbon to achieve a net zero world, carbon credits vide about $30 billion of value to their economies over the are being actively traded and increasing in value. This “new next decade. In the case of hydrology, an economic assess- currency” is catalyzing private sector engagement to further ment in British Columbia, Canada, stated that their hydro- drive “carbon efficiencies” in economic processes, but these logical networks provide an economic benefit-cost ratio of trading exchanges also require verification and validation of 19:1 (BC Ministry 2003). Thus, while the costs of sustaining decarbonized benefits. Systematic and sustained monitoring monitoring networks over their lifetime can seem onerous, of weather and climate conditions are foundational elements the value they provide in safeguarding societies and nation- in supporting these processes. al economies is far more significant, with benefit-cost ratios ranging from 10:1 to 20:1. What value can be placed on meteorological and hydrological observational data and services? A recent study by the World 1.1.2.3 Managing Disaster- and Water-Related Risks Bank, WMO, and Met Office (Kull et al. 2021) estimates that Today, nations are facing significant sustainability challeng- high-quality and timely forecasts can generate at least $160 es—ones that are likely to persist for decades to come, owing billion per year of global socioeconomic benefits. Forecast to factors such as population growth, urbanization, water and quality depends on access to better-performing NWP sys- food security, and climate (UN DESA, no date). Thus, there tems at the local level, which in turn critically depends on the is an urgent need for systematic monitoring to clarify the availability of real-time observations. Over the years, there rate and scale of environmental change being observed and have been numerous economic studies on the value of weath- expected. In particular, the observed increase in the inten- er, water, and climate services to assess the benefit-cost of sity and frequency of weather, hydrological, and climate ex- NMHS operations. For example, recent studies by London tremes since 1980 (figure 1.1.4) has ratcheted up demand for Economics of the Met Office, UK (London Economics 2015) weather, water, and climate observational data and services and the Bureau of Meteorology, Australia (London Economics (UNDRR and CRED 2020). FIGURE 1.1.4  Summary of Costs of Natural Disasters Worldwide, 1980–2019 Reported disasters Total deaths Total a ected US$ Economic losses 1980–1999 4,212 1.19 3.25 1.63 million billion trillion 2000–2019 7,348 1.23 million 4.03 2.97 billion trillion Source: UNDRR and CRED 2020. Setting the Scene    15 Disaster Risks and ecosystems are all water-dependent and vulnerable to For more than a decade, the World Economic Forum Global water scarcity. Risks Report has rated weather and climate extremes, climate action failures, and natural disasters as the highest risks to Global freshwater use has increased by a factor of six over the economic development going forward (WEF 2022). WEF as- past 100 years; since the 1980s, it has continued to grow at sesses climate change as a clear and present danger, strik- a rate of roughly 1 percent per year (UNESCO 2021) with in- ing harder and more rapidly than expected, with significant creasing population, economic development, and shifting con- consequences for food and water security, human migration, sumption patterns. Restricted spatial and temporal availability and social instability (WEF 2021). The Emissions Gap Report of water limits socioeconomic development in many countries 2021 by the United Nations Environment Programme (UNEP) of the world. Furthermore, climate change—combined with a further highlights those challenges (UNEP 2021). Hydromet more erratic and uncertain supply of water—will aggravate extremes account for 90 percent of the world’s natural disas- the situation of currently water-stressed regions and generate ters and are having devastating consequences and causing water stress in regions where water resources are still abundant political, social, and economic upheaval in many countries today (IPCC 2014). According to the IPCC’s 6th Assessment (Munich RE 2018). Every decade has been more costly than Report (IPCC 2021), continued global warming is projected to the previous one, affecting 50 percent of the world’s popu- further intensify the global water cycle, including its variability lation, killing nearly a million people and totaling more than and the severity of both wet and dry events. Some of the most $4 trillion in economic losses since 1980 (UNDRR and CRED noticeable impacts will be more intense rainfall and associat- 2020). With only 40 percent of these losses insured, respon- ed flooding, as well as more intense drought in many regions. sibilities for the remaining costs of recovery are being borne Several countries are also suffering increasingly severe periods by governments or development agencies (Munich RE 2021a). of extreme heat, windstorms, and forest fires. Flooding accounts for 40 percent of all these loss-related Beyond water scarcity, water quality is of significant concern natural hazards since 1980; they, along with droughts, have where there is a clear relationship between the two. More had the largest societal impacts (WMO 2021). An estimated than 2 billion people lack access to safely managed water and 1.5 billion people are directly exposed to 1-in-100-year flood sanitation services (CDC 2022), and water-borne diseases are events, of which about 89 percent occur in developing coun- one of the most significant factors affecting human health tries (Rentschler and Salhab 2020). River flooding affects 39 in the developing world. This means that effective monitor- million people per year, and the most extreme estimates sug- ing and understanding of water quality and quantity is es- gest this number would rise to 134 million by 2050 (Ward sential to achieve sustainable development at national and and Winsemius 2018). In many cases, the least developed global levels, as illustrated in figure 1.1.5 (PBL Netherlands countries are the most vulnerable and are ill-equipped to be Environmental Assessment Agency 2018). resilient to these events. All these factors reinforce the growing demand for know- However, without the scientific advances made in weather ing the past, current, and future states of water resources and hydrological forecasting and early warning systems, the to support early warning and adaptive management. While harsh consequences on society would have been far great- real-time and long time series hydrological and meteorologi- er—making it even more worrisome that the challenges of cal observation data sets are vital for this purpose, there has sustaining these observing networks are putting these gains been a decline in water monitoring worldwide—likely a con- at risk. sequence of both a lack of resources and monitoring system obsolescence. Water-Related Risks Food security, human health, urban and rural settlements, en- Against this backdrop, international development assistance ergy production, industrial development, economic growth, to water and sanitation and other water-related sectors from 16    Setting the Scene FIGURE 1.1.5  The Central Role of Water in Sustainable to meet their objectives. The Paris Agreement articles lay out Development the requirements for reporting on greenhouse gas emissions Partnerships for the goals No poverty as well as for monitoring the essential climate variables (see Peace, justice and Zero hunger the WMO’s Global Climate Observing System for definitions strong institutions of these variables).5 The Sendai Framework on Disaster Risk Good health Life on land and well-being Reduction has specific targets for establishing early warning systems. In the Dublin Statement of the UN Conference on Life below Quality Water and Environment in 1992, the importance of measure- water education ment networks was emphasized: “Measurement of components of the water cycle, in quantity and quality, and of other charac- Climate Gender equality teristics of the environment affecting water characteristics of action the environment are an essential basis for undertaking effective Responsible water management” (GDRC 1992). All these obligations and Clean water consumption and production and sanitation requirements on “the Parties” of these agreements hinge on reliable, long-term data from hydrological and meteorological Sustainable cities and communities Affordable and clean energy observations. Multi-country River Basin Agreements depend Reduced Decent work and on reliable, collaborative transboundary hydrological moni- economic growth inequalities Industry, innovation and infrastructure toring for their effectiveness, among other things. Group 1 targets: Group 2 targets: Group 3 targets: Strongly related to water related to water indirectly related to water 1.1.3 Why Networks and Systems Fail: Framing the Sustainability Problem Source: PBL Netherlands Environmental Assessment Agency 2018. Even though hydrological and meteorological monitoring development agencies totaled $11 billion in 2017, according networks and their related services are important for public to Organization for Economic Co-operation and Development safety, economic prosperity, and public policy, the reality for (OECD) statistics. More than 72 percent of this support was NMHSs in the majority of least developed and many devel- provided for projects where planning and design depend di- oping countries is that financial means; staff expertise; and rectly on the availability and quality of hydrological data. This access to reliable power, technologies, and adequate tele- high percentage reflects the fact that the greatest economic communication capacity are limited. Table 1.1.1 reports some benefit of hydrologic data comes from traditional engineering of the challenges facing the South African Weather Service applications for the design of hydraulic structures—ranging (SAWS);6 it is one of the more advanced weather services from dams, reservoirs, and water supply systems to hydro- in the developing world, but over recent years budget cuts electric power plants and flood and erosion control facilities.4 have reduced the quality of its maintenance, which has been exacerbated by intermittent power availability, vandalism 1.1.2.4 Framing International Policy Agreements especially at remote stations, staff departures, and long pro- Many international agreements and protocols rely heavily on curement and staff recruitment processes (PMG 2020).7 hydrological and meteorological data and information. Many environmental agreements have references requiring contin- It is difficult for many NMHSs to support the fundamentals uous monitoring and science assessment over various time of their existing hydrological and meteorological value chain scales to support international, regional, and national actions adequately, even without the added pressures of network 4 For further information, see https://www.oecd.org/dac/financing-sustainable-development/development-finance-topics/water-relatedaid.htm. 5 Details about WMO’s Global Climate Observing System can be found at https://public.wmo.int/en/programmes/global-climate-observing-system. 6 Challenges were provided by South African Weather Services with their 2020 financial data on network operations. For more information about SAWS, see https://www.weathersa.co.za/. 7 See PMG 2020 for more information. COVID-19 has further reduced financial resources since revenues from the aviation industry provided up to one-third of revenues. Setting the Scene    17 expansion or modernization. In addition, NMHS manage- of existing and new infrastructure at risk over the medium to ment practices and business models vary widely; some rely long term. heavily on third-party revenues from such sectors as aviation and agriculture to support infrastructure needs and network From these perspectives, the challenges of designing and sus- operations. These approaches often restrict access, limit the taining modernized or rebuilt networks can be attributed to exchange of observational data, and give rise to competition five categories: affordability, governance, engagement, work- with meteorological or hydrological private sector enterpris- force competency, and institutional issues. es for these markets. All these factors put the performance TABLE 1.1.1  Observation Network Challenges Observation Network and Infrastructure Challenges Line South African Weather Service 1 Budget Constraints - Limited Budget to cover all required expensive spare parts for Radar Systems Maintenance 2 Lack of skilled staff in the regions and high turnover of staff, trained staff retention is a problem 3 Recruitment of skilled staff too long and unsupported by lack of budget 4 Procurement process for maintenance material too long impacting badly on turnaround times and uptime of the infrastructure 5 Remuneration and compensation of technical staff risks skill attrition to our competitors 6 Lack of overtime and standby policy impacts on the morale of technical staff and regional technologists 7 Infrastructure vandalism on remote stations impacts uptime and performance figures, lack of security is the factor 8 Lack of universally accessible computerized maintenance management system impacts the operations, monitoring and reporting on infrastructure performance 9 Sustainable and reliable power source as well as back-up power sources impacts the performance figures and uptime of the infrastructure 10 Sustainable and reliable communication infrastructure impacts Meteorological Infrastructure Availability Figures 11 Covid 19 travelling restrictions impacted the uptime and response to maitenance issues on infrastructure including importa- tion of spares and material from abroad Notes Five of the top 10 challenges experienced by SAWS is due to budget Poor design of systems perhaps related to lack of budget HR issue Line 11 related to COVID and considered short term but should be taken into account for the future should a similar situation occur in the future as the weather does not stop nor our need for severe weather prediction so the systems need to remain reliable during these periods Procurement issues need to be taken into account to ensure a steady flow of spare parts to maintain sustainablity of the network/system ensuring data quality. A Met Service must be aware of their government bureaucracy and allow for long lead times for replacement parts - building the key pieces into the capital budget (acquisition at time of tender) might shorten these times Source: Costing data reported by South African Weather Service. Note: This information was provided with costing data from South African Weather Service, 2021. SAWS = South African Weather Service. 1.1.3.1 Affordability is often lacking and different development partners may Unrealistic expectations. NMHSs in developing countries provide different, incompatible technologies. This is very generally operate under significant financial pressure, and challenging for most if not all NMHSs with limited capacity. many have limited opportunities for modernization without Furthermore, international support generally focuses on the support from development institutions. In principle, coun- initial capital and related infrastructure investments rather try governments coordinate aid from different partners. than on the operation and maintenance (O&M) costs, which However, especially where capacity is limited, coordination recipient countries are expected to cover. Unless there is a 18    Setting the Scene full understanding of the longer financial and human resource In practice, although not intended, some NMHSs follow an implications of the initial investments, funds and capacity operate-to-fail approach. This is a big problem, given that may be insufficient for efficient operation over the expected life-cycle management can often accommodate changing re- lifespan of the equipment. This is particularly the case for quirements, less costly infrastructure, and an extension of automated networks and advanced systems such as weather life expectancy of hydromet infrastructures. For example, the radars. onset of AI technologies and methods may offer less costly and more targeted quality assurance practices, and self-cor- Poor awareness of costs of ownership. Without a full un- recting algorithms may reduce the number of revisit periods derstanding of the total cost of ownership (TCO) over the to calibrate weather or hydrometric stations or change the life cycles of hydromet observation systems and a financial requirements for spatial coverage with new data collection means test, governments cannot have full knowledge of the platforms. Waiting until the end-of-life of observing sys- long-term financial liabilities and budgetary expectations, tems can result in a lost opportunity to modernize existing especially those connected with new observation system in- infrastructures. vestments. The primary reason for this is that they have limit- ed financial knowledge of the true costs of operations. Other 1.1.3.2 Governance reasons may relate to inadequate accounting and reporting Competing government priorities. With conflicting practices, where the state and value of aging infrastructures are poorly reported. Introducing new systems through devel- budgetary demands, governments are often not able opment projects can exacerbate the financial pressures on an to allocate sufficient funds for longer-term observa- NMHS, as any budget adjustments are usually not adequate tion network maintenance. Government resources are for properly maintaining network modernization and expan- limited, with many competing demands to support press- sion. As part of this report, the benchmarking analysis (sec- ing societal needs (such as health, education, social tion 1.2.3.2) portrays this issue, where the existing level of protection, transport infrastructure development, and resources needed to support networks in a sample of lower- to secure access to food). In making budgetary decisions, middle- to upper-income countries is inadequate, especially there is often no clear line of sight by government decision- for more advanced systems such as weather radars. Without makers between the contribution and value of hydrological adequate resources for maintenance and life-cycle manage- and meteorological monitoring and their government pro- ment, the performance of network operations decays over grams and priorities. time, leading to premature system failure or obsolescence. Little or no visibility with government decision-makers. Advanced systems are not affordable. Annual NMHS budgets NMHS may lack the legislated mandate that would establish are often too limited to provide the resources needed to ade- it as the statutory institution with the delegated authority quately maintain, upgrade, or replace monitoring equipment to oversee the weather, climate, and hydrological activities throughout the life cycle of hydrological or meteorological within a state (Hodgson 2022). This is not a problem only in networks (especially for advanced systems such as upper-air developing countries. While not a guarantee of financial sta- systems and radars). Yet without systematic replacement or bility, a mandate would help by increasing the visibility of the modernization of hydromet infrastructures, the O&M costs NMHS within government and recognizing the importance of escalate, older equipment becomes more costly to fix, and their activities for society and the economy. Without such a spare parts become less accessible over time. In these cases, mandate, access to government ministers by NMHS directors the financial management methods that consider and value may be limited—resulting in a lack of appreciation by minis- the state of these equipment assets appropriately are not tries of finance or planning (or even the parent ministry) of practiced. the value of the overall enterprise and the need to sustain its operations as the principal service provider and operator Innovation cycles are limited. NMHSs do not generally follow of the national meteorological and hydrological observation a life-cycle approach to modernizing existing infrastructure. networks. Setting the Scene    19 Leadership. Making the case for investments in hydromet and ad hoc design. Limited stakeholder engagement can also monitoring and services falls to the NMHS director(s). In ad- choke the potential for jump-starting innovation and partner- dition, there can be a lack of strong synergies among public ship-enterprise opportunities. good departments, which require climate information from sectors such as agriculture and transport. Leadership and 1.1.3.4 Institutional Arrangements management training is often not given adequate priority to Ineffective organization, planning, and management. prepare senior NMHS staff for these challenges. It is import- Institutional arrangements and structures are important not ant that they be able to make the business case and work ef- only for external engagement but also for effective planning fectively with clients and partners to align the organizational and organization of financial and human resources. A lack needs, outputs, and outcomes of the NMHS with the govern- of good management practices and limited performance ment’s priorities and directions. monitoring leads to inefficiency, poor risk awareness, lim- ited capacity for contingency management, and an erosion 1.1.3.3 Engagement of confidence by senior managers. For most NMHSs (in both Lack of user engagement and outreach. There are often con- developed and developing countries), quality management straints to managing and nurturing client relationships. This and performance monitoring can be an expensive activity. can be due to several factors, including hierarchical issues, However, performance monitoring is key to assessing the budget limitations, insufficient or poorly trained people, or a state of systems. Thus, priorities should be placed on im- limited understanding of the potential client base for NMHS proving monitoring systems as well as on management de- services. The users of these services come from the whole velopment. These skill sets are fundamental in supporting of society, ranging from the public and disaster managers staff performance, risk assessments, and contingency plan- responsible for public safety and security to the ministries ning and in fostering relations and support from partners and responsible for food and water security; urban planners and other stakeholders. Ongoing obsolescence in observing sys- municipalities; and ministries responsible for the environ- tems affects the quality of services to users and attitudes of ment and ecosystem services, tourism, transport, energy, and governments. This can be a vicious circle, where poor perfor- public health agencies. Yet without outreach and communi- mance contributes to eroding government financial support. cation, there is poor awareness among these potential ben- eficiaries of the contribution of the NMHS services to their 1.1.3.5 Workforce Competencies Challenges agenda on the one hand and limits on the potential of the Poor capacity and high staff turnover. An NMHS requires a NMHS to develop targeted information services on the other. high degree of scientific and technical expertise to support This, in turn, undermines the government’s appreciation of its functions, performance, and network operations. However, the value of the service and ultimately undercuts its motiva- workplace issues and maintaining a qualified workforce are tion to properly fund the NMHS. difficult in many developing countries. The required skills are often scarce, and access to education and training oppor- Loss of synergies. Monitoring networks may be distributed tunities can be limited, especially for leadership, scientific, among different organizations (such as other ministries), and engineering competency development. Other challenges universities, or even public-private enterprises (for example, include limited staff time and availability for training; both enterprises for water management and power generation). poor and inadequate workplace conditions; constraints of And often the responsibilities for hydrology and meteorology NMHS compensation (which are often lower than that of pri- are separate. While there may be an awareness or even some vate sector employers or even other government agencies); a informal relationships among these organizations, the lack lack of supervisory skills; and high staff turnover, which re- of formal and regular stakeholder engagement mechanisms sults in the workforce lacking the knowledge or know-how to can lead to a limited understanding of user needs, intend- operate, repair, and replace observing equipment. This is par- ed purposes, and uses; potential contributions from others; ticularly acute when more advanced electronics and complex inefficiency and duplication; and overstated requirements automated observing technologies are introduced. 20    Setting the Scene Thus, key to attracting, developing, and retaining a compe- they target, many developing nations are the most vulnerable tent workforce is allocating adequate financial resources to because of their lack of reliable climate risk early warning support ongoing training and development and offering sal- systems; water storage infrastructures; hydrological and me- aries, benefits, and career structures that are competitive at teorological observations; and the socioeconomic capacity, national and regional levels. Possible ways to develop exper- infrastructure, and resilience to withstand the impacts. This tise include rotating staff among various national NMHSs or new normal is strengthening the case for reliable weather, offering visiting fellow programs or secondments to regional water, and climate data and services on all timescales. and international institutions. And national governments, the World Bank, development agencies, and other organizations As countries move forward with climate action, the require- such as the WMO can help facilitate these opportunities. ments for multi-decadal observations and prediction capabil- ities will be driven by guidance and policy-related decisions 1.1.4 Conclusions on resilience, adaptation, or mitigation. The demand for high-quality hydromet data and information services is grow- Countries all over the world need the services that hydrom- ing significantly—not just for disaster risk management but et observations provide to facilitate decision-making across also for the derivative benefits to national economies. a variety of weather-dependent sectors and to mitigate the risks of severe weather. Day-to-day and long-term decisions Against these realities, many countries in the developing in sectors (ranging from urban planning agriculture, trans- world are facing challenges in adequately operating and port, energy, agriculture, and water resource development) maintaining their hydromet observing infrastructures. Even depend on reliable hydromet information, forecasting, and with support from development agencies, these systems fail warnings of likely changes. The unprecedented challenges of over time because financial resources and expertise are inad- climate change are highlighting the need for these services equate; there is a lack of clarity on roles and responsibilities; as countries become more susceptible to water scarcity and and institutional arrangements fail to optimize the bene- quality issues, along with changes in the frequency and inten- fits and capabilities among local, regional, and even global sity of weather, water, and climate stressors. While these ex- providers. tremes and disasters do not discriminate on where or whom    21 Pathway to More Sustainable Networks 1.2.1 Introduction 1.2 Aqueduct in Outer Los Angeles, California. Photo: © Iofoto | Dreamstime.com Against this backdrop of disappointing outcomes of the investments made in many developing countries, there is a need to rethink the approach to these investments that better accounts for country context (such as limited resourc- es and human resource challenges); sets out a clear understanding of roles as well as of financial and human resource expectations; adopts a more ho- listic, incremental, and systematic approach to network modernization; and fosters more collaborative approaches among the development agencies and in-country partners. Our vision of a pathway forward offers significant potential for securing an outcome where hydrological and meteorological networks are mainstreamed as central to the national interest. Learning from other countries’ approaches where networks are maintained over the long term offers some insights. While these lessons are drawn mostly from developed countries, the solution is not to replicate what works there per se, but rather to use these approaches in a way that reflects the developing world context. “Our vision of a pathway A key theme of this vision is the need to understand the total cost of ownership (TCO) and adopt good financial management methods and practices—such as forward offers significant those being advocated by the International Public Sector Accounting System potential for securing an Board (IPSASB).8 Such an approach will serve to increase the government’s awareness of the overall costs and value of its investments; inform priority outcome where hydrological setting and government budget preparations; and foster better risk-based de- and meteorological cision-making, especially for hydrological and meteorological networks that support good socioeconomic returns. networks are mainstreamed as central to the national This chapter begins with a description of the core components of the vision— which together constitute a holistic approach to modernization—before drill- interest. Learning from ing down into detail. other countries’ approaches where networks are maintained over the long term offers some insights.” International Public Sector Accounting Standards™ (IPSAS™) are a set of accounting 8 standards issued by the IPSAS Board for use by public sector entities around the world in the preparation of financial statements. Its direction is to strengthen public financial management (PFM) globally through increasing adoption of accrual-based IPSAS™. For more details, see https://www.ipsasb.org/. 22    Pathway to More Sustainable Networks 1.2.2 Vision Elements effect, they provide a logic model for a sustainable approach for networks. They can be grouped into three categories: (1) What are the core elements of the vision? As illustrated in fig- mapping context, (2) understanding TCO and affordability, ure 1.2.1, these elements align context and affordability with and (3) decision considerations. the decision-making processes of network modernization—in FIGURE 1.2.1  Hydrological and Meteorological Development Model Vision for sustainable observation networks Mapping context Understanding total cost of ownership and a ordability Decision considerations Fit-for-purpose and t-for-budget alignments Knowing network Capital costs for needs and uses procurement and Stepwise priority setting implementation Renewal costs State and expected life of Appropriate make-or-buy existing monitoring Life-cycle business models infrastructure management Present and future Ongoing operating E cient structures and budget costs Recognizing legal frameworks limitations and depreciation expectations and residual value System-of-systems leveraging of partner Workforce capabilities contributions Potential for network Maintenance costs partnerships Investing in workforce capabilities Note: The colors serve only to differentiate the groups of elements. 1.2.2.1 Mapping Context consideration of other capabilities beyond these institutions. Appreciating the context of the state and performance of a In this regard, knowing the potential contribution of other ob- country’s existing observational infrastructure, the demands servational network providers, both public and private, helps in and uses for the data and information from these networks, right-sizing the requirements for network expansion and mod- and the potential for leveraging existing network partnerships, ernization within the National Meteorological and Hydrological workforce capacities and government financial limitations Service (NMHS). inform what is possible in determining the scale and scope of modernizing hydrological and meteorological networks. For example, NMSs that are transitioning from manual obser- Understanding and prioritizing the data gaps and user require- vations to automated systems are realizing that their work- ments, combined with knowing the utility of existing networks, force skills are inadequate to support sophisticated sensors, sets out the scope of the modernization problem, while exist- data loggers, and weather radars. The same can be said for ing financial and human resource limitations frame the scale NHSs that are adopting more sophisticated modern hydro- of the problem. Modernization initiatives are often designed metric instrumentation such as hydrologic acoustic sounders. and tuned to the specifics of a National Meteorological Service The financial requirement to offer commensurate salaries to (NMS) or National Hydrological Service (NHS) without due attract and retain a more advanced and skilled workforce and Pathway to More Sustainable Networks   23 to provide routine training to stay current with these mod- incremental (stepwise) prioritization of network infrastruc- ern systems adds to the financial demands of operating and ture and systems over time, based on the available budget for sustaining these systems over the long term. Many of these the network at large. systems have additional operating, communications, and power-related costs as well. Gaining access to existing data Make-or-buy business models assessments. These assess- through other network providers may also ease this financial ments inform governments on the acquisition, operation, burden, right-sizing the priority requirements, avoiding du- and maintenance as well as ownership considerations. The plication, and cost-sharing where feasible. make-or-buy models do that by accounting for the cost of op- erating internally versus the cost of contracting out and the 1.2.2.2 Understanding TCO and Affordability accessibility of needed professional and technical expertise. Outsourcing network ownership with a reputable source may Assessing the TCO of the various network options provides an offer more additional benefits in some circumstances. This is important context for evaluating the feasibility of supporting particularly true where an NMHS may lack technical capabil- modernization options over time. The four primary elements ities or budgets are insecure. While the third-party provid- of the TCO are (1) capital procurement and implementation er carries most of the risk, he or she has the opportunity to costs, (2) ongoing operation costs, (3) maintenance costs, source the information to other clients. and (4) life-cycle renewal costs. These costs are not mutually exclusive. For example, radars have widely different capital, Efficient structures and legal frameworks. These frameworks operating, and maintenance costs, depending on their type establish statutory authority and clarity of purpose, along and use. Moreover, the TCO also recognizes the value of un- with offering a stable and potentially financially secure op- derstanding ongoing quality and residual value of network erating environment for NMHS or other network operations. assets over their life cycle. The latter is important for sup- Legal frameworks articulate (1) public good roles and func- porting decision-making about whether to repair or replace tions, (2) division of responsibilities among state actors, (3) and about mid-life sensor system upgrades. While develop- government expectations, (4) scope of investments, and (5) ment agencies provide the upfront capital for procurements, boundaries between public goods and private goods service governments are generally expected to support their ongoing providers. Certain policies determine modes of operations operations and life-cycle replacement. and behaviors—such as open data policies and the role of the private sector—that can significantly multiply the socioeco- 1.2.2.3 Decision Considerations nomic benefits of these government investments. Fit-for-purpose and fit-for-budget. This consideration aligns basic user needs and network requirements with afford- Partnering or “system-of-systems” thinking. This element ability—in effect, a means test. The first step is to specify, offers the significant benefits of sharing costs and filling in the national context, to what end these network systems gaps in observational networks. Engaging all the in-country would be used (such as for early warnings, water resource contributors in a system of systems paradigm offers signifi- management, or flood management), followed by setting key cant potential with the appropriate open data sharing policy priorities for network modernization and design based on frameworks. Network modernization is then considered with- national affordability. This alignment of needs and require- in this broader context, where network costs are distributed ments may mean that some existing stations are no longer and likely more affordable. required or that they need to be relocated. Observing system requirements would be adjusted depending upon their appli- Investing in a well-performing and competent workforce. cation—for example, a climate station or a synoptic weather Investing in improving the workforce is essential for reliable station for international exchange would have more stringent sustainable operations, especially as network system tech- data requirements than other stations needed for routine ag- nologies advance. Even beyond technical education and train- ricultural applications or routine weather watch in-country. ing, investments in leadership and management development Specifying what the systems would be used for allows for provide vital skill sets for financial and asset management 24    Pathway to More Sustainable Networks (including budgeting, accounting, procurement, contracting, expansion or modernization are aligned to national funds avail- and risk contingency processes). It instills more confidence able to support them over their lifetime (fit-for-budget). And it is in NMHS leadership’s ability to make the financial business the TCO that informs a government’s means test decision-mak- cases to their governments, and it supports systemic qual- ing over the expected life cycle of the network components. ity management and continuous improvement. Competitive compensation regimes and motivating workplace environ- 1.2.3.1 The TCO Methodology ments attract and retain a skilled workforce. These invest- A key theme running throughout parts 2 (on hydrology) and 3 ments generate significant value for government investments (on meteorology) of this report is the issue of the sustainability in their NMHS, for which networks represent a major part of of observational networks, which is inherently linked to how their operating expenditures. they are financed. Understanding affordability requires the for- mulation of the TCO and life-cycle costing (LCC) in the decision, 1.2.3 Understanding Affordability: procurement, and operation of observational networks and Fit-for-Purpose and Fit-for-Budget their components (box 1.2.1). While a common practice in the private sector, only a handful of governments routinely apply In many cases, a large proportion of NMHS operating budgets LCC to promote sustainable public procurement—and only a are dedicated to supporting observation systems and data man- agement functions, particularly in least developed countries. small number of low- and middle-income countries are in the Knowing the total expected costs of an observation network early stages of developing sustainable procurement policies. over its life cycle is critical to deciding affordability and mod- ernization priorities. Operation and maintenance (O&M) costs Regardless of the maturity of procurement policy, the TCO is (including the costs of spares) are typically larger than the ini- an essential tool for determining what an observation system tial procurement costs over the life cycle of the observing sys- really costs the organization, beyond the initial capital ac- tems. This is especially true for hydrological networks, where quisition and implementation phase. The TCO can also shed in-situ flow measurements are required and where remote and light on network design and operation and can help manag- fully automated systems are not conducive to monitoring hy- ers, owners, and development partners make better decisions drometric changes in flow, volume, and basin characteristics. about the current state of a meteorological or hydrological network, its sustainability, and its suitability for the evolving Different costing methodologies are practiced for meteorological services supported. But it needs to be couched in a public systems and hydrological ones. Meteorological systems are gen- procurement policy that includes LCC explicitly, or there may erally more straightforward and similar in assessing costs, while be little incentive to look beyond the initial capital acquisi- the costs of running a hydrometric network are more multi-facet- tion. At a minimum, the TCO and LCC should be used to iden- ed. For example, local hydraulic conditions, stream stability, and tify and anticipate the question of sustainability at the design logistics of accessing a site for installation and measurements stage—and this information should be shared with key finan- can vary substantially. To ensure stability of the correlation cial decision-makers, setting an example for broader reform. between water level and discharge (rating curve), a number of field measurements are required to obtain the desired accuracy. Given the early stages of adoption of these practices by most Similarly, travel distances, local road access, and existing infra- economies, the opportunity to test and refine the analysis is structure (such as bridges) can heavily influence the O&M cost of a contribution that NMHSs involved in procurement of obser- a stream gauge. Thus, there are no fixed costings associated with vational networks could make. Such exemplars are necessary any given hydrological station, but rather a costing approach for for governments to adopt the TCO and LCC, as well as accrual a group of stations. Even so, data from various case studies are accounting methods more broadly. Appropriate benchmark available to support true costs of operations. data on hidden costs are difficult to acquire as noted through- out the report. And efforts are needed to fund and main- While observation networks are largely influenced by user tain normalized databases at the national or regional level needs and applications (fit-for-purpose), priorities for network (Perera, Morton, and Perferment 2009). Pathway to More Sustainable Networks   25 BOX 1.2.1  Total Cost of Ownership and Life-Cycle Costing Total cost of ownership (TCO). This is the sum of all costs incurred during the lifetime of owning or using an asset. It includes (1) initial cost; (2) expenses associated with operating and maintaining the asset; and (3) expenses associated with the asset not being able to meet expectations (downtime) minus the residual value of the asset (a measure of the effect of depreciation, quality of maintenance, and operating conditions) (figure B1.2.1.1). The TCO was developed as a tool to assess an acquisition beyond the lowest priced bid to determine the overall best value and least costly asset to procure. FIGURE B1.2.1.1  The Components of Total Cost of Ownership TCO = + + + – Initial cost Operation Maintenance Downtime Residual value Source: Based on Rogers et al. 2022. Life-cycle costing (LCC). This is a technique to establish the TCO that takes into account both the direct costs and indirect costs (also called hidden costs): ■ Direct costs are associated with activities that can be directly attributed to the operation of meteorological or hydrometric sites and can include data acquisition, on-site analysis, quality assurance, site maintenance, energy consumption, site- specific services, and station staff operations. ■ Indirect costs are associated with activities that cannot be attributed. They can include vehicles; common equipment and spare parts; licensing; information and communication technology (ICT) systems; human resources management, training, and development; management, financial, and administrative services; storage and warehousing services; and overall quality management. Depreciation, along with asset replacement and disposal, are also costing elements of LCC, and can be treated as either direct or indirect costs, depending upon the financial accounting methods used. The focus on LCC reveals that, in most cases, the operating and maintenance costs are a significant share of the purchasing authorities’ expenditures (Estevan and Schaefer 2017). Identifying all the potential hidden costs is essential. For example, con- tracted expenses (such as software licenses) can be very large and, if not properly budgeted for, can put an entire observational network in jeopardy if there are insufficient funds left for maintenance. Figure 1.2.2 illustrates costing considerations across the fol- automated weather stations. The same can be said for the lowing four categories for determining the TCO for each type costs of hydrological networks. of network operation: (1) investment and capital costs, (2) operating costs, (3) maintenance costs, and (4) depreciation In reviewing approaches from developed countries, where and replacement costs. It also lists the cost factors under network operations are well supported and maintained, each of these categories (such as staff, training, consumables, many have applied an activity-based costing methodology repairs, and administration). These costs vary depending that incorporates all these considerations. This methodology upon network equipment and operations. For instance, the accounts for those expenses that are directly attributable to costs for acquiring, operating, and maintaining a radar are a specific activity—and takes a prorated approach to attribut- significantly more than the costs for the same categories for ing indirect costs. 26    Pathway to More Sustainable Networks FIGURE 1.2.2  Key Elements in Costing Networks Elements to consider in costing of networks Investment and Operating costs Maintenance costs Replacement costs capital costs  Project proposal  Sta and training  Calibration and  Monitoring equipment preventative care and sensors  Procurement process  Consumables  Unscheduled repairs Vehicles  Equipment and  Utilities and  sensors communications  Spare sensors and  Supporting facilities parts and systems  Construction and  Travel expenditures facilities (weirs,  Supporting facilities infrastructures  Charters and road sheds, and so on) vehicle costs and systems  Procurement process infrastructures and implementation  Vehicles  Administration, materials, shipping  Vehicle repair costs Note: The colors serve only to differentiate the groups of elements. Direct and upfront capital costs for establishing a new station The infrastructure for monitoring networks typically has a are often the easiest and most tractable parts of determining fixed end-of-life timeframe. Ideally, the life expectancy of total cost, but they represent only a portion of the total costs. well-maintained meteorological and hydrological equip- However, operating costs depend upon national circumstanc- ment and sensors ranges from 7 to 10 years, and up to 15 es—such as the NMHS’s scale of operation; the country’s years for radars. For planning and accounting purposes, most spatial extent; and the station’s access (by road, sea, or air), countries adopt a 10-year life cycle for surface observation human resource costs, travel-related expenses, and power. stations.9 The key factors affecting the life cycles of equip- Revisit rates to stations can also differ, depending upon sta- ment are their robustness and impacts of severe or extreme tion security, weather, and climatic conditions. This is par- weather and climate events. Thus, without a regime of regu- ticularly true for hydrometric measurements, where frequent lar maintenance and sensor replacements, life spans can be visits are required because of changes in the relationships shortened, resulting in suboptimal returns on investment and between water levels and flow rates that result from signifi- system obsolescence. The case of the India Meteorological cant changes in hydraulic conditions. In addition, there will Service over the past 15 years offers an example of how to be outlays on other items, such as information and communi- combine a strong maintenance program with promoting the cation technology (ICT), instrument calibration, and preven- growth of indigenous equipment manufacturer vendors as the tive maintenance for vehicles and supporting infrastructures networks of observatories expand (box 1.2.2).10 (such as cableways, weirs, and buildings), which represent a substantial part of the cost of running these networks. 9 This is determined by using the average life cycle of NMHS assets. 10 K.J. Ramesh and G. Srinivasan (RIMES), personal communication with authors. Pathway to More Sustainable Networks   27 1.2.3.2 Aligning Expectations through Benchmarking BOX 1.2.2  Procuring Observation Systems: The Case Benchmarking costs and practices against other NMHSs of India or other organizations that are successful in sustaining hy- dromet observation network aids in informing the TCO. The India Meteorological Department (IMD) was Benchmarking provides a strong basis for governments on formed in 1875. It now operates large networks of which to effectively plan and set national budgets that are observatories, automatic weather stations (AWS), based on realistic financial and human resources costs—along rain gauges, upper-air stations, and radars to sup- port a wide range of services for a large and growing with the best management estimates of the costs over the life number of beneficiaries. cycle of the observation infrastructure. Around 2006, IMD began migrating from manual For that reason, this report examines the hydrological net- to AWS and automated rain gauges; it also took re- sponsibility for maintaining the networks in-house works of three countries (Germany, Canada, and the United after the initial warranty period. It then moved on to States) and the meteorological networks of five coun- outsourcing maintenance as the network expanded. tries (Australia, Austria, Canada, Germany, and the United Now, all new procurements include a 3-year warran- Kingdom) to benchmark costs for operations, maintenance, ty, plus a 7-year annual maintenance contract for a and replacement. These countries were chosen not because total of 10 years support. In addition, for the past 3 they are developed, but because they have a good track re- years, IMD has required any non-indigenous origi- nal equipment manufacturer (OEM) vendor to have a cord of maintaining and sustaining their network operations technology transfer contract with an Indian compa- over the long term. ny, thereby reducing pressure on foreign exchange. This can range from 10 to 40 percent of the contract Key Results of Benchmarking value and allows the indigenous share to grow grad- While parts 2 (on hydrology) and 3 (on meteorology) drill ually. Such an approach ensures that spare parts down into these countries’ costing considerations, this part and expertise exist within the country and are not delayed by import complications. IMD maintains of the report highlights the key results, as shown in figure manual observations, supported by a rate contract 1.2.3. These results reflect appropriate costing for non-salary for procuring needed sensors and accessories. The O&M, staff costs, and annualized replacement costs.11 Staff net result is the development of technical capacity costs, which reflect the domestic cost of labor, are given within the broader economy while ensuring that IMD in terms of the annual level of effort—full-time equivalents gets value for money through its contracts. (FTEs)12—needed to support these networks. The analysis In 2020, IMD’s number of surface AWS was enhanced covers automatic weather systems (AWSs), upper-air oper- to 913 from 15 in 2000, and around 2010, its auto- ations, weather radars, and hydrometric station operations. mated rain gauge network of 1,382 stations was com- missioned. Similarly, its annual budget increased in 2020 to about 4,698 million Indian rupees from The countries in the benchmark sample do not all follow the about 1,386 million Indian rupees in 2001. And start- same business model approach, particularly for upper-air ing around 2010, several state governments began systems and radars. Some have automated upper-air sys- establishing state-level high-density severe weather tems, one contracts out staff operations for these systems, monitoring station networks to support the Weather and some have manual operations. For radars, one country Based Crop Insurance Program of the Government of has a service arrangement to procure data, and some have India, which was launched to protect farming com- munities from extreme weather impacts. C-band rather than S-band systems. And hydrometric stream gauge operations may differ by factors such as location and access. All of these differences affect operating costs. 11 Costs per flight were very close among all the benchmarking countries (followed an averaging approach), but there was more variance for levels of preventive maintenance (calculated on a reasonable estimate based on information from three countries). 12 Full-time equivalent (FTE) is a unit that indicates the workload of an employed person in a way that makes workloads comparable among different institutions and countries. 28    Pathway to More Sustainable Networks This review shows that, while many hydromet development an upper-air manual station. For hydrometric stations, the projects are well intended, it is often difficult for national range is $9,000 to $24,700. governments to maintain the observational infrastructure ■ Staff costs can be calculated for any country using the ex- beyond the early years of operations. In particular, national pected FTE staffing levels multiplied by their average staff budgets will need to be increased to support upper-air sta- costs in their NMHS. In benchmarking countries, one FTE tions and radar operations, along with hydrological systems. adds about $5,900 in labor cost per site per year. For me- Key findings of this review include the following: teorological stations, these costs range from 0.08 FTEs for an AWS to 2.8 FTEs for an upper-air system. Sites should ■ Overall, the performance and operating costs among sam- be visited ideally four to six times every year to ensure ple benchmark countries are very similar, providing some sound performance and support preventive maintenance. confidence on the estimated costs of supporting sustain- ■ While staffing levels are similar for hydrometric stream able network operations. Uptimes for meteorological sys- gauges, non-salary costs are more variable where site tems are generally above 95 percent. visits are needed more frequently and locations are often ■ The operating costs (without labor) for meteorological sta- more remote and access more difficult. tions show the scope of national budgets required over the ■ Annualized replacement or reinvestment costs range from observation instrument/system life cycle—ranging from $5,600 for an AWS to $230,000 for a radar system. For $6,200 for an AWS to $200,000 for an upper-air station. hydrometric systems, these costs are about $5,500 to For hydrometric stations, these costs range from $4,000 $9,000. to $15,700. ■ Investment and capital costs for meteorological stations ■ Total operating costs (including labor) for meteorological range from $56,000 for AWSs to $2.5 million to $4 million stations range from $12,100 for an AWS to $424,000 for for radar systems. For hydrometric stations, the range is $28,000 to $45,000. FIGURE 1.2.3  Annual Indicative Benchmarking Costs per Hydromet Station in Developed Countries, US dollars Investment and capital Operations and maintenance Annualized replacement Networks costs costs costs Meteorological station Non-salary Salary Total cost Staff (FTE) Life cycle (years) Automated weather $56,000 $6,200 $5,900 $12,100 0.08 7–10 $5,600 Upper-air manual $1.5–2.0 million $200,000 $223,000 $424,000 2.8 20 $86,000 Upper-air automated $0.8–1.5 million $200,000 $89,000 $289,000 1.1 15 $89,000 Polarized Doppler radar $2.5–4.0 million $91,300 $85,600 $177,000 1.15 15 $230,000 Hydrometric station Minimum range $28,000 $4,000 $5,000 $9,000 0.1 5 $5,500 Maximum range $45,000 $15,700 $8,700 $24,700 0.15 8 $9,000 Note: The colors serve only to differentiate the groups of elements. FTE = full-time equivalent. How Developing Countries Compare with Benchmark these countries were selected based on access to reliable, Countries available financial data for 2020:13 (1) high-income countries The costs for benchmark countries are compared with those and territories (The Bahamas, Curaçao, Estonia, Uruguay, and of developing countries, along with network uptimes. To that Chile) and (2) upper-middle-income countries (South Africa, end, this review included a sample of developing countries; Panama, the Dominican Republic, Guyana, North Macedonia, The country sample is small and, therefore, may not necessarily be statistically representative of the income country categories. This is because reliable 13 financial data are not available for many countries. Pathway to More Sustainable Networks   29 Surinam, Columbia, Paraguay, and Jamaica). However, the in high-income countries show significant ranges in overall comparative analysis is confined to meteorological systems, operating costs, likely reflecting parallel operations in their as financial data for hydrological station operations was not transition from manual to automated systems. This appears available. to be the case for the level of staffing (FTEs) used in their operations, which is about two to four times the level of Automatic weather stations. See table 1.2.1 and figure 1.2.4. effort seen in the benchmark. However, as these countries While the average uptime is about 80 percent for upper- continue this transition, there is an opportunity to redirect middle-income countries, the analysis suggests that their and re-skill their staff to enable them to move from manual average level of non-salary O&M costs per station ($2,780)— network operations to performing higher-value functions in far below the benchmark for developed countries ($6,170)— their NMHS, such as service delivery. is insufficient to ensure long-term sustainable operation, even accounting for their lower travel costs. Operating costs TABLE 1.2.1  Indicative Costs for Meteorological Stations in Developing Countries Compared with Benchmark Countries Annual O&M/ Annual Sample Aggregation station (non- salary cost/ Total annual Network Category of country (number) functions FTEs/station salary) station O&M/station uptime Automatic weather station benchmarking analysis (US$, thousands) Benchmark developed 5 Average 0.08 $6.17 $6.03 $12.20 97% countries Range 0.04–0.09 6. 3–7.42 3.7–7.3 10–14.7 95–100 High-income countries 5 Average 0.30 $9.38 $8.24 $17.62 89.8% Range 0.1–0.5 3.0–22.0 4.3–19.2 10–44 75–95 Upper-middle income 8 Average 0.19 $2.78 $2.49 $5.27 80.3% countries Range 0.04 –.75 1.0–6.5 0.34–4.75 1.38–10.7 50–95 Upper-air radiosonde station benchmarking analysis (US$, thousands) Benchmark manual 5 Average 2.84 $200 $223 $424 96.1% Range 1.0–3.7 120–270 80–307 212–427 90–100 Benchmark automatic 2 Average 1.07 $200 $89 $283 93.8% Range 0.9–1.1 120–270 80–95 212–292 90–96% High-income countries 4 Average 2.33 $102 $72 $174 92.0% Range 0.3–4.0 50–144 10–160 140–210 75–99 Upper-middle income 6 Average 2.17 $82 $21 $100 85.9% countries Range 0.2–4.0 25–117 6–30 42–140 75–99 Radar Benchmarking Analysis (US$, thousands) Benchmark 5 Average 1.15 $91.30 $90.60 $181.90 93.1% Range 0.5–1.75 65–570 66–107 140–570 80–98 S-band 2 Average 1.1 $114.70 $83.60 $198.30 95.0% C-band 2 Average 1.2 $67.95 $97.60 $165.50 97.8% High-income countries 2 Average 2.75 $123 $105 $228 95.0% Range 1.0–2.9 95–150 30–180 125–330 95 Upper-middle income 6 Average 2.85 $54 $28 $82 86.7% countries Range .4–10 14–200 5.0–57.0 20–250 80–95 Source: World Bank data. Note: Benchmark countries are Australia, Austria, Canada, Germany, and the United Kingdom. FTE = full-time equivalent; O&M = operation and maintenance. 30    Pathway to More Sustainable Networks Upper-air operations. See table 1.2.1 and figure 1.2.4. O&M as less-expensive radars such as X-band ones. Radars often non-salary costs per station in upper-middle-income coun- require mid-life-cycle investments to sustain their perfor- tries are insufficient to support two radiosonde launches per mance over an expected life cycle. But developing countries day, 365 days a year, to comply with World Meteorological often cannot afford to properly operate, maintain, and replace Organization/Global Basic Observing Network (WMO/GBON) them. Given that radars are often acquired through hydromet requirements. These costs, at $82,000 per station, are far development projects, it would be worth considering alter- below the developed country benchmark (for both manual and nate funding and business models, such as co-funding new automated stations) of $200,000. Even though the reported radar operations over their expected lifetime or engaging a uptime is high (86 percent), the standard used by NMSs is third party to operate or own them to offset the lack of ad- likely measured against planned launches being made only vanced technical expertise in NMS to operate and maintain once a day, or even less. With an uptime of 86 percent, the them. level of effort (number of staff FTEs) for manual operations is similar (around 3 FTEs) to the developed country bench- Stream gauges. See table 1.2.2. The benchmarking analysis mark. But as developing countries continue to transition out shows that O&M non-salary costs vary greatly for stream of manual systems, it is striking that the benchmark for auto- gauges—ranging from $4,000 to $15,700 among the sam- matic stations is much lower (around 1 FTE per station). ple of benchmarking countries—depending on their location, access, local amenities (such as bridges), and distance trav- While there is a global benefit to comply with GBON require- eled to service them. Moreover, the frequency of hydromet- ments—better weather prediction models, and thereby bet- ric measurements depends on a variety of environmental ter local forecasts—the operating costs are too high for some factors such as flow rates, flood stage of the water basin, developing countries. Global funding arrangements are now frequency of precipitation. These factors alter the basin being considered to fill this gap. The WMO Extraordinary characteristics and affect the velocity of flow and the water Congress (2021) approved the Sustainable Observations volumes. Generally, significantly more hydrometric stations Funding Facility (SOFF) to support or co-fund the GBON in are required to support water resource assessments or flood/ least developed countries (LDCs) and small island developing drought monitoring in more remote locations, along with states (SIDS), which would support not only surface weather more frequent visits by highly qualified personnel, when com- operations but also upper-air ones.14 pared to stations required to support weather observations. National Hydrological Services (NHSs) typically do not receive Weather radars. See table 1.2.1 and figure 1.2.4. While sufficient additional budget requirements to support modern- uptimes are high (87 percent), O&M non-salary costs for ization or expansion of hydrological stations. Although no re- upper-middle-income countries of $54,000 are about half of liable financial data are available for hydrological networks in what is required, given the developed country benchmark of developing countries, there has been growing concern in the $91,000. FTEs (at around 3) are higher than the developed community about the decline in the number of operational country benchmark (just above 1). Understanding a country’s hydrometric stations in recent years, suggesting that budgets priority requirements for these systems—both S-band and are likely insufficient to support their operations. These is- C-band systems—may lead to more affordable options, such sues are discussed in more depth in part 2 of this report. SOFF is a UN Multi-Partner Trust Fund, established by the Alliance for Hydromet Development and WMO, in collaboration with UNDP and UNEP. At the time 14 of publication, SOFF is being considered only for LDCs and SIDS. Pathway to More Sustainable Networks   31 FIGURE 1.2.4  Indicative Costs for Meteorological Stations in Developing Countries Compared with Benchmark Countries A. Representative costs of automated weather stations (US$, thousands) $20.00 $18.00 Full-time equivalents per station Annualized replacement costs Full Time Equivalents per Station $16.00 Upper-middle .19 Benchmark .08 Annual labor costs $14.00 $12.00 Annual operating costs w/o labor $10.00 $8.00 $6.00 $4.00 $2.00 $0.00 Upper-middle AWS countries Benchmarks B. Representative costs of upper-air operations (US$, thousands) $600 Annualized replacement costs Full-time equivalents per station $500 Upper-middle 2.17 Annual labor costs Benchmarks: $400 Manual 2.84 Annual operating costs Automated 1.07 w/o labor $300 $200 $100 $0 Upper-middle Manual Automated countries Benchmarks C. Representative costs of radar operations (US$, thousands) $500 $450 Full-time equivalents per station Annualized replacement costs $400 Upper-middle 2.73 Annual labor costs $350 Benchmarks: Average 1.15 $300 Annual operating costs w/o labor $250 $200 $150 $100 $50 $0 Upper-middle Average (all) S-band radar C-band radar countries Benchmarks Note: Benchmark countries are Australia, Austria, Canada, Germany, and the United Kingdom. 32    Pathway to More Sustainable Networks TABLE 1.2.2  Benchmarking Costs for Hydrological Stations in Developed Countries (Canada, Germany, the United States) Hydrological network operations benchmarking costs Costing Categories Notes Minimum US$ Maximum US$ Initial investment – capital costs Stationary field equipment, civil works Without measuring footbridges or cableways $23,000 $ 37,000 Mobile field equipment (pro rata costs of flow and Current meter or flow tracker or ADCP or elec- $1,000 $2,700 velocity measuring equipment, in shared use for 15 tromagnetic velocity sensors, data loggers stations) Other mobile field equipment with shared use Truck, boat and trailer, safety gear, tools $3,500 $5,000 Total capital costs $27,500 $44,700 Annual operation and maintenance costs without labor Field work with small maintenance and work up of Without costs of labor $2,000 $8,100 water level and discharge data in offices Business and administrative costs, offices, soft- $1,000 $4,000 ware, computers Annual maintenance costs, spare parts Annualized costs of spare parts and replace- $1,000 $3,600 ment of components Total annual costs (without personnel costs) $4,000 $15,700 Annual total operational costs with labor Labor costs for field and office work $5,000 $8,700 Full-time equivalent (for 15 gauges) 1.5 2.3 Total costs for annual operation and maintenance (without personnel costs) $4,000 $15,700 Total annual costs of operations and maintenance (with personnel costs) $9,000 $24,700 Costs of annual operation and maintenance as a percentage of total capital costs 33% 55% Source: World Bank data. Note: Financial data and benchmarking are based on 2020 data, a period where inflationary pressures were low and the COVID-19 pandemic had just begun. But at the time of publication, inflationary pressures were increasing significantly worldwide—exceeding 5 percent in many developed economies and likely higher in developing ones. Thus, network operational costs may be even higher now, depending on inflationary pressures and access to parts and materials due to supply chain pressures. ADCP = acoustic Doppler current profiler. 1.2.4 Choosing the Right Fit-For-Purpose The key questions to ask are: What business model best re- Business Model flects the national circumstances? What is the internal and external third party capacities to deliver? Who are the consum- As the issue of sustaining networks over their life cycle is ers? What is the market potential? What value can be created? often a consequence of insufficient budget or lack of tech- How can it be done cost effectively? There are four primary nical capabilities, adopting the most appropriate business business models to choose from, based on level of financial model(s) that fit with national circumstances may offer some risk and readiness capacity (figure 1.2.5) (Rogers et al. 2019): added advantages. Given that capital equipment is often pro- vided through external development agencies, it is important ■ The government (NMHS) owns and operates its observa- to account for the capacity of the NMHS to absorb new assets, tional network. This is the most common practice around business models, and asset management from the outset the world. NMHSs are responsible for the ongoing oper- (Rogers et al. 2021b). ations, maintenance, and replacement of the equipment, accounting, and recording for the annual life-cycle depre- ciation. An important consideration for the viability of this business model in sustaining network operations is Pathway to More Sustainable Networks   33 that government agencies follow life-cycle management capabilities, and management competencies. This is the with a phased replacement cycle. As developing country case especially for complex technologies such as weather NMHSs expand their capabilities, this model can become radars in least developed countries and some developing more challenging in terms of budgets, human resources countries. FIGURE 1.2.5  Four Different Business Models for Owning and Operating Observational Networks Based on Financial Risk and Technical Readiness High Government-owned, Government-owned, third-party operated Government-operated A ordability factor Public-public Third-party owned, or Low third-party operated Public-private partnerships Low High capabilities capabilities Readiness factor ■ The government (NMHS) owns its observation networks become untenable or should a more cost-effective alter- and outsources its maintenance. This business model is native arise. This is a viable consideration for expensive not widely practiced. In this case, the agency owns all the operations such as upper-air stations, radar systems, and observation network assets and is responsible for their operations in remote regions. life-cycle management (capital depreciation and replace- ■ The government (NMHS) contracts to third-party service ment). The operations and maintenance of the network are providers who own, operate, and maintain observation- contracted out to a third party—such as an entity in the al networks (the data-as-a-service model). This business private sector or another government agency, as occurs model approach is rarely practiced. However, there are ex- in Nepal (see box 1.2.3). This model is viable in situations amples—such as Bermuda, where the entire meteorological where outsourcing is less costly than in-house services, function is contracted out. In this model, the agency con- where national governments have difficulty attracting and tracts for real-time, accessible, and quality-controlled in- retaining technical expertise and capacity, or where staff formation, satisfying the government’s observational data can be used more effectively on other tasks. It also offers requirements. The third party takes full responsibility for efficiencies of scale, assuming that the third-party contrac- life-cycle management, maintaining and replacing as nec- tor is offering similar services to a larger market. However, essary. The market for such services is growing, encourag- the national government retains ownership and control so ing competition among potential suppliers of observational it can take appropriate action should the contractor costs networks as a service (Rogers et al. 2019). This approach is 34    Pathway to More Sustainable Networks particularly viable for expensive network operations such where locations are determined that bring overall benefit as radars, as is the case in Austria, or in the case of light- to all, and the operations of the network at large are cost- ning detection systems in countries where there is a signif- shared and operated by the NHS on behalf of the partners: icant market potential for data and information services. a co-design and co-production model. ■ Public-public or public–private partnership arrange- ments are put in place to cooperate and maintain obser- Depending upon the national interest, governments can select vational networks. This is rarely practiced. But there are models that are the most appropriate to their network require- some examples, such as the operations of the hydrologi- ments. For instance, consider the example of Nepal in box 1.2.3 cal networks in Canada, where both the national and the (World Bank 2020), where partnering with the private sector provincial levels of governments separately own networks. was a viable approach to improving service delivery. It may However, the governments cooperate on their siting loca- be in a country’s long-term national interest and national se- tions and operations through collaborative arrangements, curity to own and operate their hydrological or meteorological BOX 1.2.3  Better Hydromet Observations and Services: The Case of Nepal Nepal is among the most climate-vulnerable countries in the world, with floods, droughts, and heat and cold waves impacting human lives, agriculture, hydroelectric power generation, and the broader economy. In 2009—at a time when its hydromet system relied on manual data collection with infrequent and unreliable reporting—the government identified better hydrom- et services, including an early warning system and reliable data, as key priorities in its National Strategy for Disaster Risk Management. In 2012, Nepal, with the World Bank assistance, was able to secure funding of $31 million from the Climate Investment Funds Pilot Program for Climate Resilience to improve service delivery (World Bank 2020). The program aimed to: ■ Strengthen the capacity of the Department of Hydrology and Meteorology (DHM) to mitigate climate-related hazards by delivering high-quality weather forecasts and flood early warnings to vulnerable communities; ■ Provide agrometeorology advisory services to farmers; ■ Improve the quality of surface meteorological, hydrological, and agrometeorology observation networks; ■ Disseminate information effectively; and ■ Enhance financial and technical capacity to operate and maintain networks. To date, 70 automatic hydrological stations and 88 automatic weather stations (AWSs) have been installed, and mass dissemi- nation of early warnings to households have been supported through short message services (SMSs), emails, and social media. A key performance indicator was financial sustainability. To that end, DHM has enhanced efficiency by entering into agreements with private sector service providers where DHM capacity is insufficient. Since 2014, it has awarded contracts for the mainte- nance of automated rain gauges and some water level sensors to national private companies, and a local consulting firm has developed a maintenance management platform to facilitate and monitor the performance of maintenance service providers. The project, by demonstrating improved service delivery, has been able to secure more public funding to meet operational and maintenance needs. In addition, DHM has promoted public-private sector partnerships and coordination between partners. Over the past five years, the two main mobile operators have provided free subscriber identification module (SIM) cards to DHM to facilitate trans- mission of AWS data to the DHM data server. This has reduced costs and facilitated operations. Mobile operators also freely transmit flood warnings and alerts through SMS. Furthermore, where aid and private sector organizations have financed the installation of additional hydrological and meteorological stations, the work has been done in coordination with DHM; data generated have been shared in real time with DHM; and some stations have been transferred to DHM on completion. The plan in the coming years is to upgrade all manual weather stations to automatic. Source: Government of Nepal 2009. Pathway to More Sustainable Networks   35 networks, or both. However, it may also be very practical to complex. The range of activities for which private manag- have a third party own and operate others such as their weath- ers can be held to account is much narrower than those er radar networks. Many least developed countries cannot af- for which national governments and public officials can be ford to launch upper-air balloon measurements twice daily; held to account. however, it may be in the long-term interest of the broader ■ Avoid risks to data continuity and contract competitive- global weather modeling community to finance and support ness associated with transferring observing contracts for these functions, based on one of these model considerations. one vendor to another. This can be accomplished to some degree with long-term contracting arrangements. When deciding on models that adopt an outsourcing of re- ■ Avoid situations where a contractor may underfund net- sponsibilities, governments should consider (Alford and work maintenance and capital depreciation to improve O’Flynn 2012, cited in Rogers et al. 2021a): profits, the consequence of which would introduce risks to the longevity of the infrastructure. Third-party non-gov- ■ Are there no strategic reasons for keeping the function in- ernment service providers do not have the same account- house, such as maintaining core competencies? (Strategic ability as governments that act in the public good; their question) motivation is aligned against profit and competitiveness. ■ Can an external provider do the task better and more Clarity of responsibility and transparency in contractual cost-effectively? (Service question) arrangements may go a long way in managing these risks. ■ Are the costs of managing the relationship minimal, is there a competitive market, and is it relatively easy to To mitigate these risks, a long-term contract would be required, specify and monitor the service? (Relationship question) consistent with the expected lifetime of the equipment. This would probably take the form of a public finance initiative It would be advantageous to consider outsourcing the net- (PFI), such as a classical contract—that is, a very formalized work responsibilities if the answer to these questions were and transactional agreement, as opposed to a relational con- all Yes. However, if not, then the government would be better tract, which focuses on the relationship, participation in ex- off to retain these responsibilities in-house. In reality, it is change, reciprocity, or trust (Alford and O’Flynn 2012). Getting most likely that the answers to these questions are not simply these contracts right is challenging; reversing such a contract Yes or No, and, therefore, costs and benefits would need to be to return the network to a government operation would also weighed against the national interest. be challenging not only legally but also because of the loss of technical expertise over time, albeit compensated by great- Deciding on the most appropriate models can be challenging er local capacity outside of the NMHS. Moreover, in choosing and several pitfalls need to be avoided (Rogers et al. 2021a). a model for contracting out, effective contract management Actions to take to avoid these issues include: competencies within the government or NMHS are still needed to keep good oversight on the contract deliverables and con- ■ Avoid handing over to external parties those functions that tract staff to ensure service continuity and performance. This affect the organization’s future capacity to manage risks implies ensuring that some in-house expertise is retained. or externalities. For example, in the case where the obser- vations division of an NMHS outsources the design of its Thus, it is important to determine whether the maintenance networks, the NMHS's ability to meet its mission obliga- and operation of the observing network is a core competency, tions is affected by the performance of the contractor and which should be retained in-house (the strategic question). the capacity of the NMHS to hold this third-party provider From a hydrological perspective, for example, since a serious to account. flood event alters the hydraulic conditions at many hydromet- ■ Avoid situations in which public accountability is under- ric stations in the affected river basin, new rating curves and mined by the third-party contractor. Outsourcing removes a high number of discharge measurements would be required the NMHS’s direct control over service performance, yet after such events. Such a challenge is not predictable and the NMHS remains responsible for any failings or conse- cannot be planned, and it would be a reason for retaining the quences. Accountability structures and relationships are core competency with the agency. 36    Pathway to More Sustainable Networks 1.2.5 Legal Mandates: Clarity, Authority, and framework for the provision of meteorological and hydrolog- Accessibility ical services (Rogers et al. 2021b) (figure 1.2.6). While this framework covers all aspects of service provision, it is partic- International agreements and national laws in the form of ularly important for the production, use, and sharing of mete- laws, policies, and regulations form the legal and regulatory orological and hydrological observations. FIGURE 1.2.6  Legal and Regulatory Instruments Governing the Provision of Meteorological and Hydrological Services International Agreements International conventions, treaties, directives, agreements, and resolutions Water Resource Disaster Meteorological Competition Management Law Management Law Law Law National Laws, Policies and Regulations Data Policy Revenue and pricing policy of Regulations Regulations Regulations NMS services Regulations on water use governing governing governing services hydrological services disaster and commercial Regulations emergency services governing operations meteorological services Procedures Operational Public Operating Public water use Public hydrological Commercial forecast and meteorological services services services warning services services Source: Rogers et al. 2021b. Note: Pale blue = international agreements; orange = public law; tan = laws governing private sector and procedures governing commercial services; gray = policies; dark blue = regulations; and green = public services. 1.2.5.1 Legal Frameworks agreement is the Convention of the WMO, which provides the International law is the body of law that regulates the legal legal basis for the establishment and functioning of the orga- relationship between states and international organizations, nization. The convention does not, however, seek to set out such as the WMO, recognized under international law. As for minimum standards for the provision of meteorological and international law related to meteorology, the principal legal hydrological services. Instead, such standards are set out in Pathway to More Sustainable Networks   37 technical regulations.15 However, the standard practices and an important component of open data. In Europe, one of six procedures described by these technical regulations and re- thematic data categories is meteorological, which is recog- lated documents do not impose binding obligations on WMO nized as having important socioeconomic worth and high Members. Rather they are highly encouraged expectations value for the economy and society. In situ meteorological with the aim of meeting these requirements, qualified by data from instruments and weather forecasts are identified “Members shall do their utmost to implement the standard as high-value data sets. practices and procedures.” One exception is the meteoro- logical services for international aviation, adopted under the In contrast to the United States and Europe, many developing auspices of the Convention on International Civil Aviation, countries have low scores on the open data barometer pub- which is implemented by the International Civil Aviation lished by the World Wide Web Foundation.16 This poor showing Organization (ICAO). In this case, the observations required appears to reflect the unwillingness of governments to allow to support meteorological services for international aviation data to be used without restrictions, and this lack of openness are regulated and legal standards are established. spills over into meteorological and hydrological data. 1.2.5.2 Data Policies A critical component of open data is meaningful engagement between the government and business users, which allows All economies are beneficiaries of an expanding digital eco- countries to benefit from investments in open data policies system in which data are generated, collected, organized, and and infrastructure. The World Bank has identified a number of exchanged for the purpose of deriving value from the accu- issues that must be addressed for open data policies to work mulated information; large gains accrue from data being used effectively—among them, advocacy, transparency, legislation broadly (Rogers and Tsirkunov 2021). Adopting national open including amending national security legislation, amending data policies would have a positive impact on the develop- practices on collecting charges for electronically available ment of hydromet services. There is a tendency to underes- data, implementing policies on copyright and licensing, under- timate the social and economic benefits of a well-developed standing the high social welfare associated with sharing gov- hydromet value chain, which has the consequence that po- ernment data, engaging with users of data, and recognizing the tential benefits are unrealized and regulatory policies focus value of information for the economy (World Bank 2017). mainly on the effect of the value chain on public institutions rather than the wider economy. National data policies that Some benefits from free and unrestricted access to hydromet support the free, unrestricted use and re-use of data have data include (Stewart 2015): significant economic benefit in general (European Data Portal 2020), increasing opportunities for the provision of bespoke ■ Better quality and a greater variety of hydromet products services for manufacturing, agriculture, water management, and services, especially on a global and regional basis; telecommunications, insurance, transport, energy, health ■ Improvements to the numerous other services at a nation- care, and other sectors. al level that are predicated on hydromet data and informa- tion and result in substantial social and economic benefits Competitive advantage has long existed in the United States and lives saved; because of the value of the public good provided by govern- ■ Enhanced national and international research into hy- ment data through its policy of open access that complements dromet systems and a better understanding of the hydro- data available to the private sector. Growth in the value-add- logical system, leading to improved decision-making to ed of government data has been achieved by the ability to benefit society; directly link government data sources through application ■ A better understanding and appreciation of the impor- programming interfaces (APIs) and integrate them with busi- tance of hydromet data, likely leading to more support for ness data through advances in data analytics. Technology is the observing systems; WMO sets out technical regulations and data standards. These can be found in Documents No. 2 and No. 49 in WMO-No. 2. 15 More information about the World Wide Web Foundation can be found at https://webfoundation.org/. 16 38    Pathway to More Sustainable Networks ■ Developing good regional hydromet products with the In many cases, the NMHSs do not benefit directly from the same quality and accuracy over the region; sale of either observational data or tailored services. Instead, ■ Promoting education in hydrology and meteorology, lead- these funds are often returned to the treasury or the parent ing to a better understanding of our environment, includ- ministry (Rogers et al. 2021b). Consequently, the financial ing its systems and changes; and impact of an open data policy may be minimal, but the per- ■ Promoting and strengthening collaboration between pro- ception that the NMHS may be disadvantaged compared with viders and users of hydromet data and products. potential rivals—including other NMHSs and the private sec- tor—that may make better use of open data for financial gain Despite the high value of meteorological and hydrological is real. This problem is most acute in NMHSs that receive in- observations, only a small fraction of publicly funded meteo- sufficient funding from the government to carry out their pub- rological and hydrological data is openly available nationally lic task fully and therefore associate open data with a threat and internationally. While countries agree to share a subset to their ongoing existence as a public body. Given the power of their data with the WMO, often the larger fraction of their of open data to society, these concerns must be resolved. data is withheld for commercial reasons. Improvements in forecasts and warnings are increasingly focused on provid- 1.2.5.3 Commercial Services and Regulatory Policies ing reliable information on street-level scales that put much greater temporal and spatial requirements on observational Few countries have a regulatory framework governing the data. At the same time, new sources of data from Internet of entire hydromet enterprise. The Power of Partnership report Things (IoT) sensors and shared personal data are contribut- (World Bank 2019) describes the desirability and even neces- ing to forecast improvements. sity of there being a national independent regulator. While private companies may articulate a preference for minimum Since nearly all, if not all, data created by an NMHS are fi- regulation, there is widespread agreement that regulation nanced by government, it should be in the interest of the can create clarity, transparency, and stability. Regulations government to maximize the use and benefit of those data. would specify minimum technical requirements, including When data are not shared widely, there are costs associat- staff qualifications and observational systems for all service ed with the inefficiencies that arise from the input not being providers. They could also set out a licensing system under used at the appropriate scale. Given the nature of meteoro- which specific services can be provided only by a licensed logical phenomena, this scale ranges from national to global; entity, although this may not be the optimal approach if it maximum efficiency and benefit are achieved by sharing data imposes an unreasonable regulatory burden on potential ser- as widely as possible to help many. Regional data sharing in- vice providers. creases the utility of observations, and the co-production and dissemination of forecasts and warnings improves the reli- A further issue that arises in regulating the sector is that, in ability of information across multiple countries.17 practice, the only body that is likely to have the necessary technical skills to act as a regulator is the NMHS itself. There Developing open data policies is relatively straightforward; is the potential for a significant conflict of interest if the however, there may be a loss of revenue in some NMHSs that NMHS is also a competitor with the other service providers. would be hard to replace quickly without additional govern- However, it is possible for representatives of public and pri- ment financing. This is more often the case in more advanced vate sectors to work together to define service norms, for ex- NMHSs that have developed business models around the sale ample, in the development of national observation networks. of data and the use of revenue within their service. When open data policies are introduced, governments generally ad- Considerable debate continues about the efficacy of the pub- just budgets accordingly, although this not a universal prac- lic sector providing commercial services. However, since an tice (Rogers et al. 2021b). increasing number of NMHSs are providing paid-for services See, for example, EUMETNET at https://www.eumetnet.eu/; ECMWF at https://www.ecmwf.int/; and the EUMETNET Meteoalarm 2.0 system at https:// 17 preparecenter.org/wp-content/uploads/2020/09/cap-workshop-2020-meteoalarm-.pdf. Pathway to More Sustainable Networks   39 that incorporate data from their observation networks, it is agriculture or navigation)—likely resulting in fewer benefits essential to understand how these services can be provided than would occur if the networks were designed to meet all without distorting or preventing the development of a do- user needs within a country. Similarly, it would be highly un- mestic market for services. This requires an understanding likely or even affordable for a government’s NMHS to operate of the potential applicability of competition or antitrust law a single national observational network to meet everyone’s to public NMHSs if they operate as undertakings with the needs. Thus, there is a case for viewing existing observational basic purpose of promoting or maintaining fair competition infrastructure, networks, and systems in country as a single (UK Gov 2019). Applying these rules, data paid for by gov- collaborative and integrated national asset to broaden bene- ernment in support of the public task of the NMHS are shared fits collectively for all users. openly and the NMHS generates non-public task revenue by providing commercial services in competition with the pri- Three fundamental reasons to consider public good networks vate sector. Effective business-oriented mindsets in both pub- as a single, collaborative, and integrated asset are shown lic and private sector institutions should have as a primary below: ambition that competitive relationships on one topic do not exclude cooperation on others. Noncompetitive relationships ■ Planning, designing, and operating new water manage- between actors in the public, private, and academic sectors ment facilities (including reservoirs, water withdrawals have considerable advantages and enable value creation in for irrigation or municipal water supply, or hydroelectric the enterprise. Cooperation, collaboration, and coproduction power plants) depend on a complete understanding of the are ways in which value can be created in a noncompetitive changing hydrological conditions in the whole water basin. environment (Rogers et al. 2021a). In some situations, the co-design of the basin’s hydrologi- cal and meteorological networks would be more cost-effec- 1.2.6 Partnering: A System-of-Systems Approach tive and would better inform decision outcomes. ■ In much of the developing world, agriculture plays a sig- Key component of fit-for-purpose and fit-for-budget network nificant role in supporting food security and the economy. design is leveraging partnering opportunities to share costs Gaining farmers’ input into the information they value and further stretch capacity and capabilities. A system-of- would yield better support for irrigation and other farm-re- systems approach entails a holistic view of network design lated decisions—as would greater cooperation among and management, garnering contributions from other network NMSs, NHSs, agrometeorological services, and other net- and data contributors. It follows four key principles: engaging work operators within a region. network players, leveraging what exists to fill gaps, enabling ■ Hydrological and meteorological information, historical partnerships, and co-design/co-management/co-production trends, and models are critical inputs to flood plain map- of networks (figure 1.2.7). ping to increase awareness of urban flooding risks caused by urbanization and climate change. Urban planning re- Principle 1: Engaging the Network Players and Their quirements and flood and water management regimes Contributions influence the siting and specifications for operators in all Public good networks. At the country level, policy advice and hydrometeorological networks within a water basin. services are based on authoritative, accurate, accessible in- formation that flows from hydrological and meteorological These examples illustrate the importance of user-producer observation networks. Contextualizing the need for these relationships and how cooperation among the operators of networks in terms of the national agenda and user require- networks improves relevance and cost-effectiveness. Both ments is an essential step for ensuring their long-term finan- outcomes can politically justify financing and advocacy for cial support. But, in most cases, they are often designed to monitoring network performance and longevity. Even though address a single purpose—such as air transport, reservoir these networks usually have independent operators, princi- management, or a specific socioeconomic sector (such as ples of co-design and co-production offer significant benefits. 40    Pathway to More Sustainable Networks FIGURE 1.2.7  System-of-Systems Thinking: Co-Design, Co-Production, and Co-Management 1. Engage other network players 6. Evaluate 2. Leverage performance what exists to and adjust ll gaps Co-design Co-production Co-management 5. Co-design and cooperate 3. Build and when it makes leverage sense partnerships 4. Co-identify user requirements Principle 2: Leveraging What Exists to Fill the Gaps First that monitor weather, water, and climate patterns that con- While NMHSs are the primary public good institutions in a tribute to NMHS missions. Today, there are more than 450 en- country for meteorological and hydrological information, le- vironmental satellites. Advanced meteorological and climate veraging the contribution from other providers—before em- agencies have developed scientific methods to assimilate barking on adding observational infrastructure—can go a long satellite information into weather, climate, and hydrological way in filling the data gaps in support of user applications. As models. Access and infusion of the information from satellites a first step, investigating mechanisms and establishing coop- complement surface-based observational systems. erative agreements with the other network and information providers in country and beyond often offsets the costs for ex- Surface observed data. Surface-based data are highly valued pansion, reduces duplication, and can result in sharing costs, by weather, water resources, and climate models; they are thereby lowering overall the costs of operations. also used to validate satellite measurements. Nontraditional sources of information—such as atmospheric measurements Satellite information. Leveraging existing monitoring net- being acquired through smart devices and other IoT applica- works with additional information offers opportunities to tions (crowd sourcing) can also be helpful for the now-casting prioritize. The past 50 years have seen the development, and weather watch functions as well as other applications. growth, and use of Earth-sensing environmental satellites Moreover, while governments have been the primary actors Pathway to More Sustainable Networks   41 in furnishing standardized observing networks and satel- Private sector contributions. In recent years, the private sec- lites, the private sector has become more active in owning, tor has become more involved in the weather enterprise, as operating, and selling environmental data over the past de- demands for its services are increasingly needed to inform cade. Importantly, this emerging private sector role may have low-carbon or less-carbon-intensive economic growth, to downstream impacts on satellite mission continuity and data safeguard built infrastructure, and to de-risk other financial access policies. International and national data policies are and economic assets from climate change. The private sector weighting the importance of satellite information to public builds on its experience engaging users to develop and offer safety and climate action mandates, and their significance solutions through services, and on its strong track record in to global/regional weather, climate, and water resources rapidly bringing technological innovations and applications modeling. This complementarity between satellite and sur- to markets—such as products and services that leverage the face-based measurement further underscores the importance use of smart devices, IoT, automation, and AI. Private sector of sustaining existing surface-based observing networks over providers are now active in offering capabilities, products, the long term. and services in the whole value chain from observations to analysis and prediction to services. And it, too, relies on free Academic sector contributions. In some parts of the devel- and open access to publicly funded national networks, mak- oping world, the role and capacity of the academic sector to ing it a strong advocate for governments maintaining these support research objectives is likely to be limited. However, networks over the long term. However, it is more difficult for the potential of the academic community to contribute the private sector to provide access to its own data, given its weather, water, climate, and related scientific knowledge and proprietary issues. But making more affordable access to their understanding to support national interests is significant. The networks, while protecting their proprietary interests, would science community develops innovations and builds capacity offer broader benefits to the whole country. that support NMHS operations, other public good functions, and private sector development. This sector also amplifies Principle 3: Leveraging Partnerships: Key to Sustainable the benefits from free and open access to data from public Network Management sector–operated hydrological and meteorological networks, Enabling partnerships and stakeholder engagement through and in some cases from private sector providers. Moreover, national arrangements and institutional structures offers researchers are active in designing and operating their own significant collaborative benefits, including scientific and networks that respond to specific research requirements and technical capabilities that can shore up in-country capaci- development initiatives. ties. Strengthening relationships at the global, regional, and national levels offer access to expertise, training, know-how Economic sector contributions. In developing countries, the opportunities, and products. Typically, global and regional role and contribution of the private sector in providing hy- working groups lack ongoing representation from LDCs and drological and meteorological services is varied. For some SIDS, largely owing to limited financial and human resources. countries, private sector hydrometeorological enterprises are Reversing this trend by assigning increased priority to these in their infancy; for others, services are provided by foreign functions would not only increase the value of the available private entities; and for still others, private sector enterpris- extramural products and services but also strengthen scien- es are making significant contributions. However, for some tific and technical skills capabilities through participation economic sectors, the private sector is active in operating its in learning events. This would be an asset for maintaining own observational networks for its specific purposes (such as high-performing observational networks and systems. commercial farming, road and rail operations, water-related industries, municipal drainage, and energy production and Global partnerships. The Alliance for Hydromet Development distribution). and the WMO, in collaboration with the United Nations Development Programme (UNDP) and UNEP, established a 42    Pathway to More Sustainable Networks UN Multi-Party Trust Fund18 —the Systematic Observations some WMO Members and by academia.20 Learning mod- Financing Facility (SOFF)19—to better support over the long ules cover a variety of topics, including technical guides. term the networks of countries with the largest capacity ■ Regional WMO Integrated Global Observing System gaps. The intention is to expand global coverage and extend (WIGOS) Centres provide support to NMHSs for network the longevity of observing networks by providing leveraged implementation and operations in the form of a collabora- funds to supplier government budgets to supplement gov- tive framework for advancing project implementation and ernment budgets. The SOFF is aimed at shoring up technical capacity development aims. Guidance is provided in the capacity and funding to support the national components of WMO WIGOS Manual (WMO 2019b), the WMO Technical the Global Basic Observational Network (GBON)—filling data Guidelines for Regional WIGOS Centres (WMO 2018), and gaps and sustaining network operations—in LDCs and SIDS. the WMO Guide to the WIGOS (WMO 2019a). Technical assistance is being provided by peers from more ■ WMO regional associations also facilitate collaboration advanced NMSs. GBON sets out the requirement for surface and support for less-developed Members and some nation- and upper-air meteorological stations data to improve the al Members have their own forums to support these objec- performance of global numerical weather prediction models tives—such as Spain, which sponsors the Iberian American and better national and local forecasts and early warnings. Forum, and the Caribbean Meteorological Organization, which coordinates services for the Caribbean Community The cross-border nature of weather, water, and climate phe- (CARICOM).21 nomena requires greater commitment and closer cooperation ■ The African Ministerial Committee on Meteorology among all nations, especially those that share regional prox- (AMCOMET)—a ministerial forum—coordinates meteo- imity. Like a chain, the meteorological and hydrological sys- rological development in Africa and is supported by the tem performance is only as strong as its weakest link. Through African Union. One of their priorities includes filling in the the WMO, nations are committed to collaborating to strength- data gaps. NMHSs actively participate and benefit from en global public goods. This approach not only supports the these coordinated plans and actions, which are being sup- less-developed NMHSs but also brings benefits home through ported by the African Development Bank as well as other better, more reliable understanding and predictions of long- development donors. term weather, climate, and hydrological conditions. The free ■ Although most European countries are generally more sci- and open exchange of global and regional weather models, entifically and technologically advanced, their approach informed by local weather observations, are providing high- to coordination serves as a good example for others. They er-quality forecasts to support local early warning systems and have created several successful collaborative structures forecasts for all. And a more comprehensive monitoring regime —such as European Centre for Medium Range Weather over a lake or river basin yields better insight and foresight for Prediction (ECMWF), European Organization for the water resources management for all who live in the basin. Exploitation of Meteorological Satellites (EUMETSAT), and European Meteorological Network (EUMETNET)—which Regional cooperation. This cooperation offers benefits that have established governance mechanisms to coordinate range from sharing to learning to capacity building. WMO re- financial resourcing, policies, and program development gional structures and donor organizations facilitate this co- for numerical weather prediction, satellite, and land-based operation. For example: observations. ■ The South Asia Hydromet Forum (SAHF)—which in- ■ The Global Campus is a library of online learning tools cludes Afghanistan, Bangladesh, Bhutan, India, Maldives, and resources on hydrometeorology that are supported by Myanmar, Nepal, Pakistan, and Sri Lanka—is enhancing the 18 Details about the Alliance for Hydromet Development can be found at https://alliancehydromet.org. 19 Details about SOFF are available at https://alliancehydromet.org/soff/. 20 See the Education and Training node at the WMO website at https://public.wmo.int/en/resources/training. 21 For more information about the Caribbean Community (CARICOM) see https://caricom.org/. Pathway to More Sustainable Networks   43 national capacity of its members through the development The IWRM is recognized as the accepted paradigm for effi- of shared regional capabilities (box 1.2.4). cient, equitable, and sustainable management of water re- sources. According to the Dublin Statement of 1992, the Water basin–wide cooperation. In the water arena, basin- principle “Water development and management should be wide governance and cooperation are key to understanding based on a participatory approach, involving users, planners hydrological processes and support integrated water resourc- and policy-makers at all levels” requires a clear understand- es management (IWRM). There are 263 transboundary lakes ing by those who are involved in and affected by the deci- and river basins covering nearly 50 percent of the world’s land sion-making (GDRC 1992).22 Decisions are increasingly made mass and affecting 145 nations (UN Water, no date). Success by consensus and involve all relevant stakeholders. Thus, it is contingent on common approaches and measurement stan- is imperative that reliable, spatially pertinent water quality dards, exchange of relevant data and information, and the and availability information be accessible in a timely, open existence of governance bodies to oversee the shared monitor- manner. This necessity also holds true for decision-making ing and management of water resources. These institutional that relies on meteorological information, where free and instruments also promote an opportunity for network efficien- open access would strengthen science-based policy develop- cy and optimization within the basin, providing sustainable ment and decision-making, optimize return on investments and cost sharing benefits for all participants. Such structures by governments in hydrological and meteorological networks, exist for many significant lake and river basins—such as the and greatly enhance the socioeconomic benefits through the Senegal River, which is governed through the Senegal River data’s reuse. Basin Development Organization (OMVS). BOX 1.2.4  Regional Cooperation through the South Asian Hydromet Forum In the past two decades, over 50 percent of South Asians, or more than 750 million people, have been affected by at least one natural disaster, ranging from floods, droughts, and thunderstorms to heatwaves, landslides, and cyclones. Efforts to strength- en disaster early warning systems and weather services require national-level modernization efforts but also have a regional dimension. That is why the South Asia Hydromet Forum (SAHF) was launched in fall 2018. Co-hosted by the World Bank and the WMO, it aims to provide a framework for greater cooperation among NMHSs in the region. It also enjoys the support of nu- merous other development partners, such as the European Union and the Global Facility for Disaster Reduction and Recovery. The goal of SAHF is for South Asia to adopt collaborative regional strategies to increase the production of weather, climate, and water-related information; and to strengthen the delivery of user-oriented advisory services to support the development of sustainable climate-resilient economies. A key aspect of this strategy is the pooling of data, resources, and information to strengthen the capabilities of all NMHSs in the region, regardless of their size or indigenous capacity. At this point, national investments have facilitated significant progress in the quality of weather and climate services, but the demands of beneficiaries exceed the ability of countries acting alone to provide the required level of services. South Asia is now embarking on a regional investment strategy to build shared facilities to provide the most cost-effective and sustainable means to exploit the best science and technology and to complement continuing national investments. The strategy takes a regional approach to the design of meteorological observational networks; this approach includes all the elements of the total cost of ownership and the means to sustain the infrastructure by pooling human and financial resources across region. South Asia plans to be among the first group of developing economies to exploit shared cloud computing resources to support observational networks, numerical prediction, and analytical needs. The Dublin Statement was a major agreement amount world experts in 1992 at the International Conference on Water and the Environment. See also UN 22 Water, no date. 44    Pathway to More Sustainable Networks Principle 4: National Collaborative Model: national network approach based on practices of co-design A Co-Management Approach and co-production offers significant financial and human re- For most countries, the roles and responsibilities for vari- source benefits. This results in many cases in an overall cost ous sectors—including agriculture/crop production, water reduction in their O&M across the collective national ob- resources, transportation, early warning, and disaster man- servational networks. In these cases, a coordination forum agement—are divided among public sector institutions at helps to ensure data interoperability, standards, and quality national and subnational levels. A co-management/co-pro- assurance. duction approach among these primarily public entities of- fers increased synergy and cost-effectiveness in designing Depending on the national circumstance, this integrated na- and operating various networks. Establishing an appropriate tional cooperative could be extended to private sources of governance board affords better visibility to national gov- data. In these cases, all parties—including the private sec- ernment decision-makers, along with greater efficiency and tor—recognize that more value can be realized by acting functionality. An operations committee made up of the vari- together than acting alone. Success of co-production is de- ous contributing entities handles the day-to-day management pendent upon factors of building partnerships based on trust and decision-making while sharing recommendations and ad- and confidence; meaningful engagement of key actors; estab- vice with the governance board. Such an approach also facil- lishing common or related goals; and institutional arrange- itates wider access to data, and clients (paying or otherwise) ments that provide for coordinated design, development, and are managed more holistically. Given the pressures of budget in some cases production. In most cases, the investment in and capacity, better coordination among network operators coordinated action provides a positive return on investment. can contribute to the long-term sustainability of these obser- vational systems. 1.2.7 Innovation and Technology In the case of water resources management, roles and re- The meteorological and hydrological communities generally sponsibilities are generally fragmented among different lev- follow a guarded approach to adopting new technologies. This els of government—for example, water use and rights can approach is generally warranted, given the mission of public both be managed at subnational levels. However, there are safety. Even the most financially secure institutions take time examples of good practices to increase efficiency and coop- to adopt new processes and technologies. The introduction eration. Some countries, such as Canada, have established of new technologies is considered against the institutional intergovernmental agreements along with co-management or appetite for risk and the success of previously adopted in- shared decision-making governance structures. These insti- novations. For assessing technology readiness levels and tutional arrangements can harmonize cost-sharing principles the applicability to individual observing systems, a rate of for hydrological stations, thereby meeting the needs for all innovation framework could be developed to determine the jurisdictions. Such mechanisms can also provide benefits to readiness capacity of an NMHS to adopt new and emerging meteorological operations. technologies. Such a framework would enable realistic goals for modernization to be established in terms of financial ca- Co-management/co-production approaches also avoid du- pacity, capacity building requirements, and the willingness plication, share costs and risks, and provide potential for to adopt revised standard operating procedures. However, network rationalization. For example, the United Kingdom even some of the most advanced NMHSs have faced challeng- established such an approach in 2008: the Earth Observing es with the speed of adoption of the automation of surface Forum.23 Moreover, inviting academic and private sector par- observing and, to a lesser degree, radiosonde automation. ticipants—as clients or providers or both—in an advisory ca- Their experiences offer lessons learned and best practices to pacity would refine service offerings and strengthen advocacy further develop such a framework. for sustaining network operations. Adopting an integrated For more information about the UK Environmental Observation Framework (UK-EOF), see https://www.ukeof.org.uk/. 23 Pathway to More Sustainable Networks   45 Many breakthrough technologies have made possible the 1.2.8 Sustaining a Talented Workforce application of more nontraditional observations—such as estimating rainfall from attenuation of signals between The WMO prescribes, through its technical regulations, the cell phone towers, citizen purchase and deployment of requisite education and training necessary to lead, man- commercial surface sensors, IoT devices, and smartphone age, maintain, and operate observation networks. Ongoing sensors. There are a few cases of high-capacity NMHSs education and refresher training is required to keep staff that have purchased data from these sources; these pur- skills current in accordance with best practices,24 and as chased data are quality-controlled and applied to the oper- new technologies are introduced. Beyond technical exper- ational forecasting. Such an approach offers observational tise, leadership and scientific capacities are also needed to coverage in areas, such as urban areas, that are present- support effective network design, management, and oper- ly sparsely covered. But these new data sources present ation. This is particularly true for hydrology, where strong challenges in terms of observational quality, data access, contributions from science and engineering are a prereq- and volumes, as well as privacy and ethical concerns. Such uisite to fulfilling its core functions. challenges can be solved only through more systematic partnerships, founded on solid observations and associat- Recruiting and retaining a highly skilled scientific and ed research. technical workforce to operate and maintain observation networks is vital for ensuring the longevity and continuity Low-cost weather stations offer options to fill gaps within of data and information that remains authoritative, accu- a meteorological observation network. And better remote rate, and accessible to support the hydrometeorological observing systems (such as autonomous drones, crowd- value chain and, ultimately, decision-makers. However, sourced observations, or application of IoT-based systems) most countries are experiencing some degree of difficul- will have a significant impact on global forecast quality at ty. The following strategies can, to some extent, mitigate higher resolutions, as well as on the forecasting of extreme these challenges: events. These technological innovations have often been delivered by the private sector, which has become increas- ■ Invest in leaders. Leadership training would better ingly proficient in providing observational data, telecom- equip directors or heads of NMHSs to direct, guide, and munications, and hardware—and thereby extending the administer meteorological and hydrological organiza- possibilities for national and international agencies to tions. For those who were promoted from within, top consider procuring integrated data services. areas of emphasis would be on strategic visioning, finan- cial planning and business case development, human AI and machine learning offer great potential for improve- resource management, performance assessment, and ment in all elements of the hydromet value chain. For ex- effective communications. For those leaders who were ample, AI-based systems are already competitive with very appointed from outside, the top focus for recruitment short-range limited-area model forecasts. Long time series would be understanding the business and its function. of observational data—together with reanalyses and oper- ■ Build stronger links with local academic institutions. ational forecasts—is essential to derive and improve intel- This type of leveraging is part of a strategy that supports ligent analytics methods for operational use, especially for investments in science and engineering disciplines, phenomena that require further understanding of physical along with building partnering schemes to create recog- processes. nized pathways from education into careers. ■ Expand education programs. Given that in many devel- oping countries the number of post-secondary institu- tions that offer hydrological or meteorological programs are limited or nonexistent, development agencies For information about the WMO Capacity Development Strategy, see WMO 2015. 24 46    Pathway to More Sustainable Networks should consider expanding their programs to finance guest ■ Heighten appreciation of network monitoring functions. professors, lecturers, international student fellowships, There is a need to encourage senior policy makers and de- and even program chairs at colleges or universities. Such cision-makers (and those at the political level, where ap- an approach would also bolster the number of socioeco- propriate) to recognize the high value of these monitoring nomic applications from hydrological and meteorological functions within a country and to collaboratively champi- services within a country. on better compensation for operational support. ■ Develop better competency-based profiles. These profiles ■ Invest in development and career training in supervisory should meet workforce requirements commensurate with staff. Supervisors play a vital role in guiding, on-the-job agency-level education and training plans. Learning plans training, and mentoring operating personnel; staff trust for individuals afford the opportunity to adopt self-paced and look to them for counsel and support. Therefore, on- (with measurable milestone setting) learning objectives going training, enhancing working conditions, and attrac- to grow and develop within the workplace. Also, the WMO tive compensation regimes for supervisors are essential Global Campus and Regional Training Centres are encour- in supporting retention strategies and building the overall aged to develop priorities and curricula and to facilitate caliber of the entire workforce, staff access to in-situ and remote learning resources. ■ Outsource functions that are more suited to delivery by ■ Define attractive job roles with a career path. The goal a contracted partner. Note that such a strategy depends is to retain high performers and create an overall work on developing effective contract negotiation and manage- environment that creates positive energy; generates moti- ment skills. vation, trust, and loyalty; and visibly rewards staff achieve- ments. These programs should recognize and reward high Given the vital role of staff at all levels in ensuring the long- performers in the workplace with further training qualifi- term success of investments in observation networks, it is cations and promotion as an incentive in return for loyalty. essential that governments—facilitated by projects support- ■ Create a work culture of sharing within the workplace. ed by development partners as needed—allocate sufficient There needs to be an environment in which experienced priority to capacity development of human resources. This staff and junior personnel can work together to strengthen includes leadership development, training and education, the transfer of experience and know-how—which is par- management systems, and building a wider network of sup- ticularly important where local conditions play a pivotal port both locally and through international institutions. influence, as is the case for hydrology.    47 Conclusions and Recommendations Accessible observational data and information bring significant societal 1.3 value and are foundations of the hydromet services value chain, by which weather, climate, and water services strengthen resilience to natural hazards and climate change and contribute to improving the economic performance. International funding institutions have invested billions of US dollars into hydrological and meteorological projects in developing countries over the past two decades. However, the returns on such investments have often been suboptimal. Part 1 of this report explores the root causes that observing networks in many developing countries fall into obsolescence. These include the following constraints: Mini weather station in China. Photo: xxwp ■ There are limited resources within National Meteorological and Hydrological Services (NMHS) to support modernization. ■ There is a poor awareness of the total cost of operations and unrealistic expectations in terms of the capacity of national governments to afford and support new or upgraded installations—especially the more advanced systems such as weather radars or upper-air systems. ■ There are competing government priorities and conflicting demands for limited resources, along with a poor awareness of the value of these net- works. Instead, decision-makers are drawn to investments in service deliv- ery functions such as warnings, which have greater visibility. ■ There is a lack of clear mandates, roles, and responsibilities among state “There are competing actors—especially for an NMHS, as this gap contributes to the problem of lack of government profile and access to adequate financial resources. government priorities and ■ There is a lack of strong advocacy among the user communities for network conflicting demands for services, caused in part by a lack of open data policies and user community engagement. The result is a loss of synergy among various public and pri- limited resources, along vate organizations that participate in observational network operations. with a poor awareness of ■ There is limited human resource capacity to support the networks over the long term, owing to limited education and training opportunities and high the value of these networks. staff turnover. Instead, decision-makers Parts 2 and 3 of this report delve more deeply into the technical aspects of the are drawn to investments in proper and sustainable design of hydrological and meteorological networks, including their operational costs. service delivery functions such as warnings, which have greater visibility.” 48    Conclusions and Recommendations Overall, this report, with its three parts, suggests a pathway ■ Partnership for global public good. High-quality, acces- for sustainable investments in hydrological and meteorologi- sible hydrological and meteorological data serve not only cal infrastructures based on the following principles: the national interest but also the broader regional and global public good. Building on the Alliance for Hydromet ■ Clarity, authority, and accessibility. A legal framework Development—which was created at COP25 in Madrid to establishes statutory authority and clarity about expecta- enable better collaboration among development agen- tions, roles, and responsibilities of NMHSs, private sector cies—can optimize capacity and contributions for design- hydrological and meteorological service providers, data ing and sustaining networks and utilizing instruments such policies, financial accountability, and service provision. as the Systematic Observations Financing Facility (SOFF) It anchors the importance of hydromet observing data in least developed countries (LDCs) and small island de- and services in the broader national interests, including veloping states (SIDS). It is in the long-term interests of disaster management, climate action, and economic de- the global hydromet enterprise to support these observa- velopment plans. It can also provide the framework for tions, and development agencies could serve as conduits public-private sector collaboration and co-production of for richer countries to support these systems as part of the data and services. A data policy that supports free and global public good delivery commitment. open public good hydrological and meteorological data and information enables significant co-benefits to the How should governments calculate the TCO? The report out- economy, distributing the cost of sustaining observation- lines an approach that allows governments to better deter- al networks over several functions and decision-support mine what they can afford and to set priorities for network systems. modernization, often with the assistance of development ■ Fit-for-purpose, fit-for-budget. Understanding the total agencies. It encourages them to make right-sized choices in cost of ownership (TCO) supports smart decisions on establishing priorities for modernization, adopting the most network design, modernization, and implementation. appropriate business models, and creating an appropriate Benchmarking costs of operations, preventative mainte- budget to sustainably support these networks. nance, human resources, and mid-life replacement informs the TCO at the beginning of the planning phase. This fa- The benchmarking analysis in figure 1.3.1 highlights the fact cilitates systems and infrastructures that are affordable that annual costs of non-salary operations can account for and well maintained, and attract a competent workforce. at least 10 to 15 percent of the initial capital investment for In addition, adopting an appropriate financial business automated weather stations and 20 to 30 percent for hydro- model—options range from the NMHS taking full charge metric stream gauge stations. But these costs are contingent of the network operations and maintenance to various de- on location and access; in remote locations, costs can double grees of contracting out these functions—offers significant or even triple, especially for hydrological networks. Human potential benefits to a government. resources costs are given in staff full-time equivalents (FTEs), ■ System-of-systems approach to network design. This in- as labor costs vary greatly among countries and depend on volves partnering where feasible and engaging with other network systems; costs are affected by level of technical public sector institutions and private sector organiza- competence needed and on-going proficiency training re- tions—all supported by open data policies that maximize quirements such as is required for engineers and local staff economic impact. compensation regimes. Conclusions and Recommendations    49 FIGURE 1.3.1  Annual Indicative Benchmarking Costs per Hydromet Station in Developed Countries, US dollars Investment and Annualized Networks Operations and maintenance costs capital costs replacement costs Meteorological station Non-salary Salary Total cost Staff (FTE) Life cycle (years) Automated weather $56,000 $6,200 $5,900 $12,100 0.08 7–10 $5,600 Upper-air manual $1.5–2.0 million $200,000 $223,000 $424,000 2.8 20 $86,000 Upper-air automated $0.8–1.5 million $200,000 $89,000 $289,000 1.1 15 $89,000 Polarized Doppler radar $2.5–4.0 million $91,300 $85,600 $177,000 1.15 15 $230,000 Hydrometric station Minimum range $28,000 $4,000 $5,000 $9,000 0.1 5 $5,500 Maximum range $45,000 $15,700 $8,700 $24,700 0.15 8 $9,000 Note: The colors serve only to differentiate the groups of elements. FTE = full-time equivalent. It is recommended that governments—and development part- establishments to build necessary skills, provide in-ser- ners, as needed—take the following actions: vice training to take advantage of new technologies, re- ward staff adequately and ensure career paths, and work ■ Create a roadmap for developing their observation net- with private sector providers as appropriate. Furthermore, work, based on country context and critical data require- development agencies should consider integrating these ments. This should include a long-term financial plan to requirements as part of their project support. guide investments from the government and development ■ Consider building partnerships at every level. At the partners. local level, this includes partnerships with beneficiaries of ■ Prepare a financial plan that includes the TCO of the com- hydromet services that can also provide local-level data, ponents of the modernized observing network to ensure including municipalities and other government agencies that—at each stage of development, implementation, and that may be delivering targeted hydromet services. For operation—the network is affordable and reliable. This example, this could relate to disaster risk, agriculture, or may require an iterative approach to the design, where shipping agencies. It includes private sector providers, the scale of network operations aligns with financial academic institutions, and ministries of planning and fi- expectations. nance. Outreach and communications are needed at every ■ Consider the most appropriate business model for network level—ranging from international development partners to operations—that is, one that accounts for risks and oppor- regional and global hydromet observation networks and tunities consistent with national interest, risk profiles, the World Meteorological Organization (WMO). staff competencies, and financial affordability. Outsourcing more technically advanced systems such as radars, where The bottom line is that, as extreme weather and climate staff expertise is limited, may be advantageous. conditions become more frequent—and more threatening to ■ Consider innovative financing schemes that would result lives, property, and the economy—the global need has never in the long-term sustainability of observation networks been greater for accurate and reliable data, forecasts, and and systems, particularly those that are cost prohibitive warnings. This report concludes that a key step in building re- (such as weather radars). silience to natural hazards and climate change is to optimize ■ Prioritize building human resource capacity in hydrom- investments in sustainable meteorological and hydrological et management, operations, maintenance, and plan- observation, with the focus on networks that are built to last ning. NMHSs should work with educational and training and deliver global, national, and local goods. 50    Part 1 References Students launching a weather balloon. Photo: © Kevin M. Mccarthy | Dreamstime.com Alford, John and Janine O’Flynn. 2012. Rethinking Public Service Delivery: Managing with External Providers. London: Palgrave Macmillan. Alliance for Hydromet Development. 2021. Hydromet Gap Report 2021. https:// alliancehydromet.org/gap-report/. BC Ministry (British Columbia Ministry of Sustainable Resource Management). 2003. Hydrometric Business Review. Victoria and Vancouver, BC: British Columbia Ministry of Sustainable Resource Management, Resource Information Department. https://www.obwb.ca/newsite/wp-content/uploads/2013/09/water_quantity_mon- itoring_in_bc_business_review.pdf. CDC (Centers for Disease Control and Prevention). 2022. Global Water, Sanitation, & Hygiene (WASH). Last reviewed February 16, 2022. Content source: Centers for Disease Control and Prevention, National Center for Emerging and Zoonotic Infectious Diseases (NCEZID), Division of Foodborne, Waterborne, and Environmental Diseases at CDC. https://www.cdc.gov/ healthywater/global/wash_statistics.html#access. Estevan, Helena and Bettina Schaefer. 2017. Life Cycle Costing State of the Art Report. SPP Regions. https://sppregions.eu/fileadmin/user_upload/Life_Cycle_ Costing_SoA_Report.pdf. European Data Portal. 2020. Analytical Report 9: The Economic Benefits of Open Data. Luxembourg: Publications Office of the European Union. https:// data.europa.eu/sites/default/files/the-economic-impact-of-open-data.pdf. GDRC (Global Development Research Center). 1992. The Dublin Statement on Water and Sustainable Development. International Conference on Water and the Environment, Dublin, Ireland, January 1992. https://www.gdrc.org/uem/ water/dublin-statement.html. Government of Nepal. 2009. National Strategy for Disaster Risk Management, 2009. Unofficial Translation. http://extwprlegs1.fao.org/docs/pdf/nep143046. pdf. Hodgson, Stephen. 2022. Meteorological Law and the Global Weather Enterprise. World Bank, Washington DC (in preparation). References   51 IPCC (Intergovernmental Panel on Climate Change). 2014: climate-change-and-natural-disasters/natural-disasters/ Climate Change 2014: Impacts, Adaptation, and Vulnerability. floods/interview-han-seung-soo.html. Part A: Global and Sectoral Aspects. Contribution of Working Group II to the Fifth Assessment Report of the Intergovernmental Munich RE. 2021a. Risks Posed by Natural Disasters: Losses Panel on Climate Change, edited by C.B. Field, V.R. Barros, D.J. Are Trending Upwards. https://www.munichre.com/en/risks/ Dokken, K.J. Mach, M.D. Mastrandrea, T.E. Bilir, M. Chatterjee, natural-disasters-losses-are-trending-upwards.html. K.L. Ebi, Y.O. Estrada, R.C. Genova, B. Girma, E.S. Kissel, A.N. Levy, S. MacCracken, P.R. Mastrandrea, and L.L. White. Munich RE. 2021b. Weather Risks: Volatile Weather Can Cambridge, United Kingdom and New York City: Cambridge Significantly Affect Your Profits. https://www.munichre.com/ University Press. https://www.ipcc.ch/report/ar5/wg2/. en/risks/weather-risks-can-significantly-affect-your-profits. html. IPCC (Intergovernmental Panel on Climate Change). 2021. Climate Change 2021:The Physical Science Basis: Contribution PBL Netherlands Environmental Assessment Agency. of Working Group I to the Sixth Assessment Report of the 2018. “Increasing Water Risks Are a Barrier to Sustainable Intergovernmental Panel on Climate Change, edited by V. Development.” PBL News Release, April 19, 2018. https:// Masson-Delmotte, P. Zhai, A. Pirani, S. L. Connors, C. Péan, S. www.pbl.nl/en/news/2018/increasing-water-risks-are-a-bar- Berger, N. Caud, Y. Chen, L. Goldfarb, M. I. Gomis, M. Huang, K. rier-to-sustainable-development. Leitzell, E. Lonnoy, J.B.R. Matthews, T. K. Maycock, T. Waterfield, O. Yelekçi, R. Yu, and B. Zhou. Cambridge University Press. In Perera, Oshani, Barbara Morton, and Tina Perfrement. 2009. “Life Press. https://www.ipcc.ch/report/ar6/wg1/. Cycle Costing in Sustainable Public Procurement: A Question of Value.” International Institute for Sustainable Development. Kull, Daniel, Lars Peter Riishojgaard, John Eyre, and Robert https://www.iisd.org/system/files/publications/life_cycle_costing. A. Varley. 2021. The Value of Surface-Based Meteorological pdf. Observation Data. World Bank/ WMO/ and British Crown, Met Office. https://openknowledge.worldbank.org/bitstream/han- PMG (Parliamentary Monitoring Group). 2020. “South African dle/10986/35178/The-Value-of-Surface-based-Meteorological- Weather Service & SANBI 2020/21 Annual Performance Observation-Data.pdf?sequence=1&isAllowed=y. Plans; with Ministry; COVID-19 Update: Meeting Summary.” May 7, 2020. https://pmg.org.za/committee-meeting/30158/. London Economics. 2015. Met Office: General Review – Economic Analysis: Final Report. London: London Economics. https:// Rentschler, Jun and Melda Salhab. 2020. “People in Harm’s londoneconomics.co.uk/wp-content/uploads/2016/03/MET- Way: Flood Exposure and Poverty in 189 Countries.” Policy Office-General-Review-Final-Published.pdf. Research Working Paper No. 9447. Washington, DC: World Bank. http://hdl.handle.net/10986/34655. London Economics. 2017. A Study of the Economic Impact of the Services Provided by the Bureau of Meteorology. A Report Rogers, David P., Anna-Maria Bogdanova, Vladimir Tsirkunov, by London Economics for the Department of the Environment and Makoto Suwa. 2021a. Public and Private Engagement and Energy, Commonwealth Government of Australia. in Hydromet Services: From Rivalry to Coproduction in London: London Economics. https://londoneconomics.co.uk/ Meteorological and Hydrological Service Delivery. Public and blog/publication/study-economic-impact-services-provid- Private Engagement in Hydromet Services Washington, DC: ed-bureau-meteorology-july-2017/. World Bank Group. https://documents.worldbank.org/curated/ en/229941619193500042/Public-and-Private-Engagement- Munich RE. 2018. “Topics Online: Floods – Water and in-Hydromet-Services-From-Rivalry-to-Coproduction-in- Disasters. Interview with Dr. Han Seung-soo.” March Meteorological-and-Hydrological-Service-Delivery. 13, 2018. https://www.munichre.com/topics-online/en/ 52    References Rogers, David P. and Vladimir Tsirkunov. 2021. “Open Data: A UNESCO (United Nations Educational, Scientific and Path to Climate Resilience and Economic Development in South Cultural Organization). 2021. The United Nations World Water Asia?” Technical Note. Washington, DC: World Bank. https://doc- Development Report 2021: Valuing Water. Paris: UNESCO. uments1.worldbank.org/curated/en/342901614809536537/ https://unesdoc.unesco.org/ark:/48223/pf0000375724. pdf/Open-Data-A-Path-to-Climate-Resilience-and-Economic- Development-in-South-Asia.pdf. UN Water (United Nations Water). No date. Water Facts: Transboundary Waters https://www.unwater.org/water-facts/ Rogers, David P., Vladimir Tsirkunov, Haleh Kootval, Alice transboundary-waters/. Soares, Daniel Kull, Anna-Maria Bogdanova, and Makoto Suwa. 2019. Weathering the Change: How to Improve Hydromet Ward, Philip J. and Hessel Winsemius. 2018. “River Flood Services in Developing Countries? Washington, DC: World Bank. Risk.” Institute for Environmental Studies, Vrije Universiteit https://openknowledge.worldbank.org/handle/10986/31507. Amsterdam. https://www.pbl.nl/sites/default/files/downloads/ pbl-2018-the-geography-of-future-water-challenges-river- Rogers, David P., Vladimir Tsirkunov, Alan Thorpe, Anna- flood-risk_3147.pdf. Maria Bogdanova, Makoto Suwa, Haleh Kootval, Stephen Hodgson, and Michael Staudinger. 2021b. “The Level Playing WEF (World Economic Forum). 2021. The Global Risks Field: Business of Weather, Water, and Climate Services.” Report 2021, 16th Edition. Geneva: World Economic Forum. Washington, DC: World Bank. https://openknowledge.world- https://www3.weforum.org/docs/WEF_The_Global_Risks_ bank.org/handle/10986/35670. Report_2021.pdf. Stewart, Blake. 2015. “Measuring What We Manage: The WEF (World Economic Forum). 2022. The Global Risks Importance of Hydrological Data to Water Resources Report 2022, 17th Edition. Geneva: World Economic Forum. Management.” Proceedings of the International Association https://www3.weforum.org/docs/WEF_The_Global_Risks_ of Hydrological Sciences 366: 80–85. DOI: 10.5194/ Report_2022.pdf. piahs-366-80-2015. WMO (World Meteorological Organization). 2015. WMO Capacity UK Gov (UK Government). Department for Business, Energy Development Strategy and Implementation Plan. WMO-No. 1133. & Industrial Strategy. 2019. Met Office Framework Document https://library.wmo.int/doc_num.php?explnum_id=7871. 2019. London: Crown. https://assets.publishing.service.gov. uk/government/uploads/system/uploads/attachment_data/ WMO (World Meteorological Organization). 2018. Technical file/786668/met-office-framework-document-2019.pdf. Guidelines for Regional WIGOS Centres on the WIGOS Data Quality Monitoring System. WMO-No. 1224. 2018 edition. UN DESA (United Nations Department of Economic and Social https://library.wmo.int/doc_num.php?explnum_id=5681. Affairs). No date. Sustainable Development. The 17 Goals. https://sdgs.un.org/goals. WMO (World Meteorological Organization). 2019a. Guide to the WMO Integrated Global Observing System. WMO-No. 1165. UNDRR and CRED (UN Office for Disaster Risk Reduction and 2019 edition, 2021 update. https://library.wmo.int/doc_num. Centre for Research on the Epidemiology of Disasters). 2020. php?explnum_id=11137. Human Cost of Disasters: An Overview of the Last 20 years 2000-2019. Released October 2020. https://reliefweb.int/report/ WMO (World Meteorological Organization). 2019b. Manual world/human-cost-disasters-overview-last-20-years-2000-2019. on the WMO Integrated Global Observing System: Annex VIII to the WMO Technical Regulations. WMO-No. 1160. 2019 edi- UNEP (United Nations Environment Programme). 2021. tion, 2021 update. https://library.wmo.int/?lvl=notice_dis- Emissions Gap Report 2021: The Heat Is On – A World of Climate play&id=19223#.YmiLoO3MKUk. Promises Not Yet Delivered. Nairobi: UNEP. https://www.unep. org/resources/emissions-gap-report-2021. References   53 WMO (World Meteorological Organization). 2021. WMO Atlas World Bank. 2019. The Power of Partnership: Public and Private of Mortality and Economic Losses from Weather, Climate and Engagement in Hydromet Services. Washington, DC: World Water Extremes (1970–2019). WMO-No. 1267. Geneva: WMO. Bank. http://hdl.handle.net/10986/34615. https://library.wmo.int/doc_num.php?explnum_id=10989. World Bank. 2020. Project “Building Resilience to World Bank. 2017. World Bank Support for Open Data 2012- Climate Related Hazards.” https://projects.worldbank. 2017. Washington, DC: World Bank. https://openknowledge. org/en/projects-operations/project-detail/P127508. worldbank.org/handle/10986/28616. Manual discharge measurement in a small watercourse. Photo courtesy of U. Koehler, Germany.     55 PART 2. LEAD AUTHOR Andreas Schumann Lead Hydrological Consultant, WBG; Senior Professor, Ruhr-University Bochum, Germany Recommendations CONTRIBUTING AUTHORS Christopher Cameron-Hann Senior Hydrological Consultant, WBG; Director, Aegaea Limited, UK Alain Pietroniro Lead Hydrological Consultant, WBG; Professor, University of Calgary, Canada; former Executive for the Design of Sustainable Director, National Hydrological Service of Canada Yuri Simonov Senior Hydrological Consultant, WBG; Head of Hydrological Hydrological Forecasts, Hydrometcentre, Russia Observation Networks and Systems in Developing Countries 56    Part 2 Abbreviations ADCP acoustic Doppler current profiler CResMPA Climate Resilience Multi-Phase Programmatic Approach DBMS database management system EWS early warning system GISs geographical information systems GSM Global System for Mobile Communications GPRS General Packet Radio Service HPPs hydropower plants ICT information and communication technology IPS information processing systems ISO International Organization for Standardization IT information technology IWRM integrated water resources management LCC life-cycle costing LCCA life-cycle cost analysis NDMA National Disaster Management Agency NHP National Hydrometric Program NHS National Hydrological Service NMS National Meteorological Service NWP numerical weather prediction OECD Organisation for Economic Co-operation and Development QA quality assurance QC quality control QMS Quality Management System QPFs quantitative precipitation forecasts RRR Rolling Review of Requirements SDGs Sustainable Development Goals SOPs standard operating procedures TCO total cost of ownership UCAR University Corporation for Atmospheric Research (United States) UNESCO United Nations Educational, Scientific and Cultural Organization USGS United States Geological Survey WIGOS WMO Integrated Global Observing System WMO World Meteorological Organization WRA water resources assessment WSC Water Survey of Canada XML Extensible Markup Language All dollar amounts are US dollars unless otherwise indicated. Unless otherwise noted, the sources of all figures are the authors of this publication.    57 Part 2 Executive Summary A meteorology garden in Bali, Indonesia. Photo: © Pande Putu Hadi Wiguna | Dreamstime.com The Problem The limited availability of water in its temporal and spatial distribution re- stricts socioeconomic development in many countries. Combined with a more erratic and uncertain supply of water, climate change will aggravate the sit- uation of currently water-stressed regions and generate water stress in re- gions where water resources are currently secure. Under these circumstances, hydrological data and derived information are indispensable to enable users of water resources to make wise decisions affecting the security of life and property, ensuring economic growth, and protecting environmental quality. Water resources cannot be managed unless we know where they are, their quantity and quality, and how variable they currently are and likely to be in the foreseeable future. Despite an ever-increasing need for data, there is growing evidence that hydrological observation networks are in decline in many parts of the world (Mishra and Coulibaly 2009; Rodda 1995; Rodda et al. 1993). In recognition of their importance, the World Bank Group, along with other development partners, has invested hundreds of millions of US dollars over the past two decades to rebuild and strengthen hydrological and meteoro- logical observation networks and services in developing countries. However, the return on such investments has been suboptimal because project funding is time-limited and many developing countries do not have the financial re- sources, management systems, and expertise needed to operate, maintain, and eventually replace these networks after funding expires. In general, an expansion and modernization of the measurement networks is “… the World Bank associated with the introduction of new technologies for sensors, data trans- mission, data storage, and data analysis. However, the introduction of new Group, along with other technologies cannot overcome the existing problems of hydrological services development partners, in developing countries or even exacerbate existing deficits. These problems arise mainly from: has invested hundreds of ■ A fragmented policy environment in which users of hydrological data and millions of US dollars over information pursue particular and often temporary interests related to sin- the past two decades to gle issues (irrigation, hydropower, flood protection, water supply); ■ Inadequate budgets that no longer meet the higher requirements for hy- rebuild and strengthen drological data provision and result in insufficient capacities to maintain hydrological and hydrological infrastructure; ■ Inability to recruit, train, and retain qualified personnel—which becomes meteorological observation an increasing problem with the elimination of manual tasks as a result of networks and services in the digitalization of observation networks, the use of automated systems, and computer-based data processing and analysis methods; developing countries.” 58    Part 2 Executive Summary ■ Poor relationships with users whose data needs can be es- capacity required to operate a modernized hydrological moni- timated only on a case-by-case basis and do not serve as a toring network, the following steps are necessary: basis for strategic planning of observation networks; and ■ Unsatisfactory delivery of services that fall short of digital 1. Identify hydrological data and information needs. The data acquisition capabilities, including, for example, the net worth of the data derived from a hydrological obser- provision of hydrological forecasts with integration of re- vation network depends on their subsequent uses and al-time weather data. applications. This value has to be assessed in cooper- ation with users, but also with an awareness of exist- Many networks have been designed without adequate con- ing knowledge deficits about hydrological conditions. sideration of local operating realities, making sustained use Special emphasis has to be given to ongoing and expect- throughout their life cycle difficult. There is also often insuffi- ed hydrological changes. In the hydrological value chain cient consideration of the need for the National Hydrological (figure ES2.1), the two categories of hydrological data Service (NHS) to adapt to new technological requirements to (real-time data and historical data) are shown together these realities. This concerns, in particular, the provision of with the main applications: planning and operation of the necessary personnel and financial capacities for the op- water management facilities, disaster risk management, eration of a modernized observation network. To ensure that and design of hydraulics structures. the outcome of the modernization is fit for purpose and does 2. Plan the modernized observation networks. Any mod- not exceed the budget, the planning process must take into ernization or new installation of a hydrometric network account the direct and hidden costs of operation as well as requires comprehensive planning, which should include: the increased need for qualified personnel. (1) the selection of gauge locations according to the po- tential uses of the data to be collected; (2) the selection This part of the report is a practical guide for designing sus- of technical equipment according to local conditions; (3) tainable hydrological networks. It is intended to guide future an assessment of the human and technical capacity re- decision-making by the World Bank, its clients, and develop- quired to operate the modernized observation network; ment partners. The target audience is professionals working and (4) an estimation of future staffing and funding on hydrometric investment programs—such as NHS staff and needs of the NHS to ensure continued operation of the directors, staff in key planning and finance ministries, and pro- modernized observing network and to provide data and fessionals in development institutions. Hydrometric activities information of the required quantity and quality. are used to determine various water properties in rivers, lakes, 3. Evaluate the required human resources. The modern- and reservoirs, such as water level, flow, sediment transport ization of networks is often associated with a change in and deposition, water temperature and other physical proper- technologies. This shifts the need for technical person- ties of water, characteristics of ice cover, and chemical proper- nel to the field of electronics, communications technolo- ties, but the focus here is primarily on the planning and design gy, and information technology (IT) services. Educating of hydrometric stations to estimate water levels and derived and training personnel in these areas is essential for discharges. The determination of discharges is the main objec- coping with technological change and adapting the qual- tive here, since quantitative statements about discharges form ity of hydrological information to increasing demands. the basis of all hydrological analyses and forecasts. According 4. Estimate the total cost of ownership (TCO) of the mod- to a survey by the World Meteorological Organization (WMO), ernized network. The TCO looks at the total costs (costs there is a lack of action, especially in the complex determina- of initial investment, operation, and maintenance) of a tion of outflows, as only 63 of 101 countries that participated piece of equipment or a service from the time of installa- in the survey collect these data (WMO 2020). tion or the starting point of the service till the time when it should be replaced. Here the equipment consists of Making an Observation Network Fit for Purpose stream gauges and the service entails the provision of hydrological data and information. A modern network’s An overly optimistic number of new or refurbished stations is total costs result from capital costs; infrastructure costs often considered in a development project, yet the longer-term (for example, the costs of road construction); operat- ability to operate them properly is not. To determine the ing costs of stations, information and communication Part 2 Executive Summary   59 FIGURE ES2.1  Hydrological Value Chain Network design Methods of observations Equipment Network operation Data transmission Primary and secondary data processing and storage Real-time data Data analysis Hydrological forecasting Historical data, statistics Forecasts Water management Disaster risk Water resources Design data system operation management planning Water management, water use and protection against adverse water conditions Public Water Power Agriculture Navigation Ecology safety supply production Source: Based on WMO 2008. 60    Part 2 Executive Summary TABLE ES2.1  Range of Costs per Automated Stream Gauge from Examples in Canada, Germany, and the United States Cost category Minimum Maximum Remarks Initial investment: Capital costs Stationary field equipment, civil works $23,000 $37,000 Not including footbridges and cableways Mobile field equipment (pro rata costs of flow and $1,000 $2,700 For example, current meter or flow tracker or ADCP or velocity measuring equipment, in shared use for electromagnetic velocity sensors, data logger units of 15 gauges) Other mobile field equipment with shared use $3,500 $5,000 For example, trucks, boats and trailers, safety gear, tools Total capital costs $27,500 $44,700 Annual operating costs Field work with little maintenance and with the $2,000 $8,100 Not including labor costs workup of water level and discharge data in offices Business and administrative costs $1,000 $4,000 Including offices, software, computers, but without labor costs Annual maintenance costs Annualized costs of maintenance $1,000 $3,600 Including spare parts Total costs for annual operation and maintenance $4,000 $15,700 Not including personnel costs Labor costs Labor costs for field and office work per gauge $5,000 $8,700 None Full-time equivalent (for 15 gauges) 1.5 FTEs 2.3 FTEs 15 gauges are a feasible number for 1 hydrological tech- nician plus 0.5 co-workers Total costs in dollars and percentages Total costs of annual operation and maintenance, $9,000 $24,700 Including personnel costs including labor costs Costs of annual operation and maintenance as a 33% 55% None percentage of capital costs Data source: USGS 2007; personal communication, Alain Pietroniro, WSC (Canada); data collection of Andreas Schumann, (Germany). Note: ADCP = acoustic Doppler current profiler; FTE = full-time equivalent; WSC = Water Survey of Canada. technology (ICT), data storage and archiving, and man- regard to the intended long-term operation of hydrological power; costs of maintenance and replacement of spare observation systems, which are essential for hydrological parts; and business operating and administrative costs. data for design purposes, the life-cycle costs (LCCs) also have Because these costs depend on the specific conditions of to be considered. The LLCs include the costs of replacement the network (for example, the distances between stream when equipment reaches its end of lifetime. Consideration gauges and their accessibility, the need for frequent of annual depreciation and the remaining residual value in discharge measurements, regional climatic conditions, life-cycle costing provides the opportunity to decide about the cost of maintenance, and other local factors), it is the economic sense of replacements of single components difficult to define a generally valid cost rate. Table ES2.1 with a shorter lifetime in more complex equipment. shows the range of TCOs for a typical stream gauge using examples from industrialized countries. Annual opera- Before starting the process of modernization, beneficiary tion and maintenance costs range from 33 percent to 50 governments should be aware of all financial, technical, and percent of capital costs. human resource requirements for sustainable operation and commit to comply with those requirements. The estimated TCO Obviously, the estimations of TCO and of the required human is essential for the network design. It has to be focused on pri- resources are two essential aspects of the planning process orities that ensure an outcome that is consistent with financial that are decisive for the sustainability of modernization. With expectations and affordable both currently and in the future. Part 2 Executive Summary   61 Ensuring Alignment of the Costs of the considered that data from such stations, located for exam- Modernized Observation System with the ple in mountainous areas with insufficient infrastructure, Budget may be more important than data from easily accessible and thus “cheaper” stations. Thus, in such cases, several Both network planning and funding for operations and main- objectives must be weighed against each other. tenance should ensure that the network can be maintained and operated over the long term. In many cases, this will re- Securing Human Resources to Operate a Modern quire an increase in the budget to ensure the long-term goal Measurement Network of modernization, or a reduction in the scope of moderniza- tion so as not to overburden the NHS. If a sufficient increase At its core, a strong NHS consists of two basic structures: first, of the budget is impossible, an upgrade of the levels of devel- the data collection and management side in the form of gauges, opment of stream-gauging networks often can be performed monitoring equipment, ICT infrastructure, and other hardware only gradually, owing to limited resources. In such cases, and software that are required to run a complex data-inten- only some gauges can be automated, and remaining stations sive monitoring program; and second, the human resources should be operated in a safe manual mode. Any gradual up- with the skills necessary to operate this technical system and grade of stream-gauging networks will require the prioritiza- to ensure data collection, transmission, processing, and data tion of gauges based on their importance to socioeconomic analysis (figure ES2.2). Modernization campaigns within a conditions, gains in hydrological information, and the ability short period of time can overwhelm an NHS because of the di- of the NHS to operate them in a sustainable manner. Any pri- vergence between the requirements of new technologies and oritization of new gauges to be installed or the modernization the existing staff skill levels. It is imperative that short-term of existing gauges requires consideration of the following targeted improvements in technical capabilities be combined multiple objectives and constraints: with the development of a medium- to long-term strategy of institutional support of human resource development. In this ■ Strategic importance. Determining strategic significance context, a fundamental task is to train operators of modernized requires assessing the importance of a particular site’s hydrological monitoring systems to be able not only to collect runoff data to water management needs in both the pres- reliable hydrological data, but also to process these data into ent and the future. The ongoing changes of demand and information that meets both the potential of the monitoring supply must be considered as well as the relevance of the systems in use and the requirements of the users. data for disaster risk management. In addition, there may be different hydrological priorities for gauges of the basic The recruitment of qualified personnel to operate, for exam- network in different regions of a country, determined by ple, more complex measuring instruments and ICT equip- the characteristics of the available water resources. ment, or to develop software solutions for combining sensors ■ Operational data. Here the need to get water-level data in and databases, is made more difficult by the fact that the NHS near-real time must be weighed. The necessity of receiv- is in competition with other sectors of the economy whose ing data in an operational mode depends on the specific wage structures are more attractive. The increase of the conditions of the watersheds, the need for early warnings, wage level required to overcome this difficulty is often not and/or the requirements to operate water management considered sufficiently in modernization projects. The NHS is systems. challenged to assess its educational and training needs, to es- ■ Costs. Here the recurrent costs for the operation, mainte- tablish strong links with local academic institutions with the nance, and replacement of each station have to be con- aim of organizing permanent education and training oppor- sidered together with the capital costs of the investments tunities and of installing partnering schemes to create recog- for the station and equipment. The investment and running nized pathways from education into careers. If investment is costs of a gauge depend on the local conditions. There is not made in skilled personnel, all the benefits of the technolo- a tendency to close the most expensive stations in order gy will be lost because there will be no personnel to maintain to reduce the overall cost of the network, but it must be the equipment, process the data, and make the most of it. 62    Part 2 Executive Summary FIGURE ES2.2   Adaptation of Human Resources Needed for NHS Technological Development An Effective NHS Higher requirements for human resources Hardware and Reliable, timely Human resources technological hydrometric data improvements for ■ People and hydrological Requirements ■ Education long-term Gauges + ■ of users ■ Skills information, ■ ICT targeted toward ■ Training ■ Data stakeholders and ■ Methodological management users developments ■ Data analysis Qualified operation and maintenance Note: ICT = information and communication technology; NHS = National Hydrological Service. Recommendations data storage, and data analysis. Such a project funding is time-limited, but many developing countries lack the 1. In the coming decades, water problems worldwide on-going financial resources, management systems, and will be exacerbated by advancing climate change and expertise necessary to support the long-term operation, growing populations. Devastating droughts and floods maintenance, and eventual replacement of these net- already threaten socioeconomic development in many works. If rising costs for operation, maintenance, and re- developing countries and are forcing people to migrate. placement of a modernized network have to be offset by Adaptation, protection from hazards caused by too much shrinking budgets, such a network cannot be financed in or too little water, and wise water management are in- the long run. Modernization campaigns with extensive conceivable without knowledge of the spatial and tempo- technical installations within a short period are likely ral variability of water resources. Most NHSs in the least to overwhelm an NHS as there is divergence between developed countries and many developing ones lack the technical advancement and the human resources. financial resources, staff expertise, and access to mod- 2. In terms of affordability, there are two options: to im- ern technologies to adequately address their growing prove the financial and human capacities of an NHS in a missions. International support is therefore essential to short period of time so that the transition to a completely support them, but this support is mostly provided in the new technological system can be mastered, or to adapt form of projects to expand and modernize the observa- the modernization projects to the existing capacities of tion networks and projects associated with the introduc- the NHS, which are increased stepwise. Although the tion of new technologies for sensors, data transmission, second option requires the operation of an only partially Part 2 Executive Summary   63 modernized measurement network over a certain period p The costs of operation would be better known by the of time, it is more promising in many cases: NHS and the future requirements for funding could be p By selecting the gauges to be upgraded as a priori- assessed in a realistic way. ty, the highest specific socioeconomic benefit of the 3. However, such a gradual modernization would need funding would be gained. to be progressively extended to a country’s entire NHS p The divergence between technical advancement and to ultimately achieve a largely homogeneous network human resources would be reduced as the time re- structure and improve the quality and quantity of hydro- quired for training, implementation in existing struc- logical information provided. An extension of the proj- tures, and testing of the new technology in parallel ect duration and greater effort needed to implement the with existing structures would be extended. modernization would be unavoidable, but this would be more than justified by the increase in sustainability and longevity of the modernized hydrological network. 64    2.1 Introduction Water is at the core of sustainable development and is critical for socioeco- nomic development, healthy ecosystems, and human survival itself. It is cen- tral to the production and preservation of a host of benefits and services for people. It is also at the heart of the climate adaptation issue, serving as the crucial link between the climate system, human society, and the environment. But water is a finite and irreplaceable resource that is fundamental to human well-being. It is renewable only if well managed (UN DESA 2015). However, as María Garcés, President of the 73rd Session of the UN General Assembly, Nepal. Photo courtesy of V. Tsirkunov, WBG. noted, “that which cannot be measured, cannot be managed” (Garcés 2018). In recognition of all this, the World Bank Group, along with other development partners, has invested hundreds of millions of US dollars over the past two decades to rebuild and strengthen hydrological and meteorological obser- vation networks and services in developing countries. After all, hydrological data and information are the core element for the wise utilization of water re- sources and for the protection of society from water-related hazards (floods, droughts, diseases). However, these significant investments have not realized their intended benefits because, in many cases, these networks and systems have become inoperable or poorly maintained. At the root of the problem is that many donor-supported networks have not been designed with local operating realities in mind and thus they cannot ensure sustainability over their life cycle. This guide is intended to support the planning and design of sustainable hy- drological observation networks (see box 2.1.1). Such a process requires a re- view of the financial viability of operating these monitoring networks, based on a realistic estimate of costs for their operation, maintenance, and replace- ment. Given small and shrinking budgets, the efficiency of network design and operation that determines these costs becomes particularly relevant. The requirements of local operating conditions must be considered, as well as the available state-of-the-art measurement technology with its varying degrees of complexity and the available human resources needed to use it. The re- sulting operating costs should be compared to the economic benefits of the hydrological information, such as avoiding damage from adverse hydrological “…water is a finite and conditions. irreplaceable resource that This document is intended to inform international institutions and organiza- is fundamental to human tions as well as national authorities cooperating toward the aim of establish- ing or developing hydrological services. The main purpose of this cooperation well-being.” is to build sustainable and affordable observing networks with the goal of Introduction   65 better aligning the outputs and outcomes of these networks with the priorities and needs of national governments and BOX 2.1.1   A Guide to Key Terminology stakeholders. Hydrology: The science of water on and under the Earth’s The structure of the guide follows the general procedure of surface and subsurface, its occurrence and movement, modernizing a hydrometric network: (1) analyzing hydrologi- its physical and chemical properties, and its interactions with the environment. cal data and information needs, (2) estimating conditions for network operation (organizational structure and available Hydrometric network (also hydrological observation net- resources), (3) designing and planning a network, and (4) work): A set of hydrometric stations that are designed assessing the feasibility to operate the network sustainably. and operated to measure spatial and temporal distribu- tions of hydrological properties to characterize the re- gional hydrological conditions. Chapter 2.2 focuses on the value of hydrological data and information and the way they should be assessed in cooper- Hydrometric station: A station at which data on water in rivers, lakes, or reservoirs are obtained on one or ation with data users. These are national requirements that more of the following elements: stage, streamflow, sed- determine how modernized networks will be used and which iment transport and deposit, water temperature and functions they will need to fulfill. other physical properties of water, characteristics of ice cover, and chemical properties of water (WMO Technical Chapter 2.3 outlines the organizational structure required Regulations, Volume III: Hydrology, 2008). to operate a hydrometric network and provide hydrolog- Hydrometry: The monitoring of the components of the hy- ical information from observed data. The modernization of drological cycle (rainfall, evaporation, runoff, storage of networks is often associated with a change in technologies, water on or under the Earth’s surface) as well as water which shifts the need for technical service to the areas of elec- quality and flow characteristics of surface waters. tronics, communications technology, and information tech- Meteorology: The science of the Earth’s atmosphere as nology (IT) services. Extended observation networks require it relates to short-term weather and longer-term climate decentralized structures to operate distributed observations variations. efficiently. There is also an ongoing need to strengthen hy- Modernization of hydrometric networks: The introduc- drometric capacities, technical service, and hydrological sec- tion of new techniques and methods for the acquisition tions for analysis and forecasts. of hydrological data and their evaluation by greater use of automated data collection and transmission devices/ Chapter 2.4 highlights the requirements for human resources instruments. needed to operate and maintain the observation network and National Hydrological Service (NHS): An institution whose use the provided data efficiently. In most cases, hydrologi- core business is the provision of information about the cal information is derived from the classification of current water (or hydrological) cycle and the status and trends of a country’s water resources. observations in the climatological setting (for example, to assess hydrological extremes). But a real-time setting for hy- National Meteorological Service (NMS): An institution drological forecasting also matters. whose core business is the provision of information about the atmosphere and its phenomena and especially with weather and weather forecasting. Chapter 2.5 explains the main aspects of network design for specific regional conditions and the utilization of data. The Stream gauge: A subcategory of hydrometric stations overall network design has to be reviewed and further devel- where streamflow (water levels and discharges) are ob- served and measured. These data characterize the water oped in tandem with socioeconomic development goals. The availability, floods, and droughts and are essential for demand for hydrological data depends on the extent to which water management issues and public safety. it is needed for decision-making. Stream-gauging network: The entirety of stream gauges at surface waters, designed and operated to characterize water quantities in a region (river basin) of interest. 66    Introduction Chapter 2.6 explains the planning steps required to modern- Chapter 2.8 summarizes the steps required to modernize ize a stream-gauging network. Such modernization must be a monitoring network, by: (1) identifying hydrological data based on an assessment of the starting point, a specification and information needs, (2) planning the modernized moni- of objectives, and a prioritization of tasks to reach the goal. toring network, (3) assessing the human resources required, In the process, setting priorities and taking an incremental and (4) estimating the total cost of ownership for the mod- approach to modernization should avoid overwhelming a ernized network. In the concluding remarks, the two options National Hydrological Service (NHS). The affordability of for ensuring the affordability of a modernized hydrological operations, maintenance, and replacement is an essential observation network are presented: improving the financial component of an assessment of the sustainability of any and human capacity of an NHS in a short period of time to modernization. If these conditions cannot be met, it would cope with the transition to a new technological observation be necessary to refocus efforts on the most important gauges. system or adapting the modernization project to increase the capacity of the NHS only gradually. In the annexes, three in- Chapter 2.7 focuses on the total costs of operation of formative case studies from different NHSs are presented to stream-gauging networks, with an emphasis on taking a demonstrate: life-cycle management approach for replacing core equip- ment. These include: (1) ongoing costs of operation and ■ The way of cost assessments (Annex 2.1, Canada), maintenance, over and above capital investment costs; (2) ■ The adaptation of hydrometric networks to socioeconomic the cost of stream-gauging instruments and indispensable developments (Annex 2.2, Russia), and infrastructure components (such as vehicles, computers, and ■ The need to consider national boundary conditions in offices, including some overhead); and (3) indirect operating modernization projects (Annex 2.3, Sri Lanka). costs (such as equipment for discharge measurements, hy- drometric software, and vehicles not attributable to a single monitoring site) distributed over the entire network. The var- ious cost components are listed in a cost model, which can be adapted to specific regional conditions.    67 The Benefits and Value of Hydrological Services 2.2 2.2.1 The Hydrological Value Chain Accurate information on the condition and trend of a country’s water resourc- es is required for economic and social development and for the maintenance of environmental quality. A National Hydrological Service (NHS) provides the basis for the environmental stewardship of a country’s most valuable natural resource and the advancement of a country’s social and economic conditions. Almost every sector of a nation’s economy uses water information for its plan- ning, development, or operational purposes. Water is of inestimable value, and as competition for water increases in most parts of the world, the value of water information grows. Nepal. Photo courtesy of V. Tsirkunov, WBG. The value of water and climate observations for economic prosperity under- pin all sectors of society over a range of timescales. An overview about the sectoral benefits derived from water quantity information is given in table 2.2.1 (Gardner, Doyle, and Patterson 2017). It is clear that population growth, agricultural water management, resource extraction, and urbanization in a changing climate are multiplying the demands for water-related data and information. In-situ observations are truly the backbone of the information chain; these provide the important data for decision-making and model de- velopment required for the wide range of necessary water resources applica- tions and decisions. Despite an ever-increasing need for data, there is growing evidence that hy- drological observation networks are in decline in many parts of the world (Lins 2008; Mishra and Coulibaly 2009; Rodda 1995; Rodda et al. 1993; World Bank 2018). International projects funded by a variety of donors are attempting to halt and even reverse this process. A major issue for such in- vestments is the longer-term sustainability of the modernized observation network. After the installation of gauges, the operational demands that must be met to maintain a credible network require an ongoing funding mecha- “Despite an ever-increasing nism. However, these ongoing costs are far outweighed by the ensuing bene- fits, especially in a changing and highly variable world. need for data, there is growing evidence that The net worth of the data derived from an observation network is a function of the subsequent uses and applications that are made of them. The hydro- hydrological observation logical value chain, as illustrated in figure 2.2.1, is complex, with two different networks are in decline in requirement-profiles for hydrological data and information that have to be accounted for (see box 2.2.1 and figure 2.2.1): many parts of the world.” 68    The Benefits and Value of Hydrological Services TABLE 2.2.1  Sectoral Benefits Derived from Water Quantity Information Sector/type of information Water quantity information (meteorological and hydrological data) Reduce costs from uncertainty by increasing managers’ ability to plan more efficiently and respond All sectors more quickly to water availability and use issues Agriculture (irrigated and Increase revenue by improving planting and irrigation decisions to increase yields or product quality non-irrigated) Reduce costs and increase revenue by improving planning and production decisions (e.g., by knowing Energy production the amount of energy needed to transport water to meet demand) Increase revenue by improving planning and operations decisions about planting and harvesting (e.g., Forestry information about groundwater/soil moisture helps managers project biomass growth and operate efficiently) Reduce costs by optimizing the design of hydropower facilities Hydropower production Reduce costs by improving reliability and avoiding operations shutdowns Reduce costs and increase production by optimizing planning and operations on the basis of water Mining availability information and flood/weather forecasting Reduce costs by optimizing planning for shipping/logistics industry Transportation Reduce costs by optimizing barge operation Increase revenue through using information (e.g., lake/stream levels, snowpack for skiing) to improve Tourism and recreation planning and attract tourists to recreation Reduce costs by using information to improve risk planning for floods and other disasters Reduce costs by optimizing operations and disaster planning with improved weather and flooding forecasts Water service provider Reduce costs and increase revenue by monitoring water delivery and consumption to optimize supply/ demand operations decisions, improve reliability, understand user behavior, and plan water and waste- water treatment (improved projections reduce spending on excess capacity) Source: Gardner, Doyle, and Patterson 2017. ■ How a country’s water resources vary by time and space. There is a variety in the products that an NHS might provide, Water resources cannot be properly managed unless we such as: know where they are, their quantity and quality, and how variable they are likely to be in the foreseeable future. The ■ A runoff monitoring service designed to provide specific data collected are essential for planning and designing data or information at a particular location for a particular water management facilities for municipal, agricultural, or client. industrial use and for flood risk management. The analyses ■ The provision of flood and drought forecasts and warnings. that have to be derived from these data require long obser- ■ Hydrometeorological observations and interpretations vation series (historical data), which originate mainly from obtained from an ancillary or partner observing network. a hydrometric basic network. ■ Current and near-future hydrological conditions. This infor- Although “raw” data (for example, daily rainfall totals) are mation, which comes from real-time data, forms the basis often supplied, hydrological database management sys- for operating water management systems, assessing ongo- tems make it possible to provide statistics (such as daily, ing droughts (for example, to manage water demand), and monthly, seasonal, and annual means or maxima) that are taking situation-based flood protection measures. more useful to clients and water resources practitioners. The Benefits and Value of Hydrological Services    69 FIGURE 2.2.1  Hydrological Value Chain ■ Network design Methods of observations Equipment Network operation Data transmission Primary and secondary data processing and storage Real-time data Data analysis Hydrological forecasting Historical data, statistics Forecasts Water management Disaster risk Water resources Design data system operation management planning Water management, water use and protection against adverse water conditions Public Water Power Agriculture Navigation Ecology safety supply production Source: Based on WMO 2008. ■ Water-related information—such as a comprehensive ■ Climate change assessment for local water resources. assessment of national water resources; statistics of the ■ Advice and decision support, where information and knowl- magnitude, frequency, and duration characteristics of flood edge are turned into recommendations for a response to events; and maps of spatial/temporal trends in groundwater conditions, such as advice on appropriate responses to a quality. contaminant spill on a major river or an evolving drought. ■ Increasingly sophisticated hydrological database manage- ment systems, such as time series analyses, geographical Data and information are of value when they are used to information systems (GISs), and hydrological modeling make a decision—either a better one than otherwise would technologies. These systems are continually extending the have been made, or one that is made with a greater level of boundaries of the information that an NHS can provide. confidence than otherwise would have been possible. Hence, ■ Coupled weather, climate, and hydrological services, based an NHS provides increased confidence and reduced risk to on numerical weather system and climate models. its clients as they make water-related decisions. The primary 70    The Benefits and Value of Hydrological Services BOX 2.2.1  Key Uses of Historical and Real-Time Hydrological Data Historical data. The most important application of hydrological data and derived information is in traditional civil engineering. Here the longest possible observations of hydrological conditions are required to design hydraulic structures—such as dams, reservoirs and water supply facilities, hydroelectric power plants, bridges and other river crossings, wastewater treatment plants and waste disposal facilities, mining tailings ponds and treatment facilities, flood and erosion control facilities, drainage and irrigation systems, and in-stream fisheries. They are also used to validate our hydrological understanding to assess the reliability of forecasting and climate change models. Real-time data. The usefulness of these data for operating flood warning systems depends largely on hydrological conditions, the need for runoff forecasting, and the possibilities of applying hydrological models. For large rivers, conditions for effective forecasting are often good, but for smaller, fast-responding mountain rivers, the lead time for forecasts is limited. The same holds for urban catchments. The potential to avoid flood damage through flood risk management (for example, by temporary flood protection measures such as evacuation or flood barriers) must be constantly reviewed and evaluated to ensure an effi- cient use of the data. Real-time data also help with managing water-based utilities (such as hydropower plants, waterworks, urban drainage, and wastewater treatment plants) and with adapting operations to hydrological conditions in the forestry, agriculture, and tourism sectors. role of hydrometric observation networks is, then, to provide water information will be needed, so a well-designed network this critical information on the status of the resources to all will allow for data extraction or extrapolation to ungauged stakeholders and decision-makers. Such information may be rivers or parts of a basin where information may be need- required for the following purposes (WMO/UNESCO 1991): ed. A basic WRA requires data on the variation in time and space of the flows along with physical, chemical, and biolog- ■ Assessing a country’s water resources (quantity, quality, ical characteristics of the entire water cycle (precipitation, distribution in time and space), the potential for water-re- evaporation, air humidity, runoff, river, lakes, snow and ice, lated development, and the ability of the supply to meet soil moisture, groundwater). Long time series of these data actual or foreseeable demand; are essential for specifying statistical characteristics of flow ■ Planning, designing, and operating water projects; and their temporal variabilities to assess long-term changes ■ Assessing the environmental, economic, and social impacts caused by climate variabilities, land-use, or climate change. of existing and proposed water resources management practices and planning sound management strategies; Typically, hydrological gauges are installed to provide in- ■ Providing security for people and property against wa- formation on discharge conditions at a particular profile of ter-related hazards, particularly floods and droughts; a watercourse. The observed discharge originates from the ■ Allocating water among competing uses, both within the catchment runoff that drains the basin through this profile. country and cross-border; and This catchment should be representative of a natural region ■ Meeting regulatory requirements. not only to characterize its water resources but also to obtain regional information on runoff conditions, as well as their de- To fulfill these tasks, which relate to different spatial and pendencies on precipitation and hydrological processes. For temporal scales, two categories of stream-gauging networks basic networks, the sites should be selected in a way that cov- exist: the basic network and the network of stations for oper- ers the variety of natural hydrological regimes and different ational tasks. ranges of spatial scales. 2.2.2 Benefits of Basic Networks When properly designed, a basic network provides a level of hydrological information at any location within its region A basic network provides hydrological information for water of applicability that would contribute in a meaningful way resources assessments (WRA) that are currently unanticipat- to any water resources assessment. To meet these require- ed but are nevertheless required for future decisions about ments, several pre-conditions have to be fulfilled: the resource. In reality, it is often difficult to predict where The Benefits and Value of Hydrological Services    71 ■ A mechanism must be available to transfer hydrological and facilitating the interpretation of hydrological information. information from the sites at which the data are collected Therefore, it is sometimes difficult to sustain a basic nation- to any other comparable site at the river network within a al network in many countries because doing so requires a hydrologically similar region. Such mechanisms can be de- longer-term view and its benefits are not always immediate. rived from generalizations beyond individual catchments Although the basic network data provide the knowledge base by learning from differences among many catchments. for a country’s hydrological situation, monitoring efforts are ■ A means for estimating the amount of hydrological infor- often more focused on specific applications to meet immedi- mation (or, conversely, the uncertainty) at any site must ate needs—at the expense of more holistic national coverage. also exist. This amount of information depends on the This means that, at a time when global warming may be exac- basic knowledge about the variability of the runoff regime erbating weather extremes and water shortages, hydrologists in time and space. are less able to understand water supply changes, predict ■ The derivation of water management decisions must in- droughts, and forecast floods than they were several decades clude the ability to collect more data before a final decision ago. Thus, it is vital to be able to show that the NHS can provide is made. So-called project gauges, operated in parallel with the long-term benefits of a comprehensive, integrated NHS, in base gauges for a limited time (several years), can be used which broadly based data collection is more economical—both to compare the flow regimes at both sites and derive a meth- in the operation of monitoring networks and in the assessment od to extend the short series of observations at the project of water resources. site with the transfer of a long data series from base gauges. 2.2.3 Benefits of Operational Networks It is important to provide a long time series for monitoring ongoing changes in hydrological conditions that are evident Operational stream-gauging stations are intended for the re- as a result of climate change and increased human activities. al-time monitoring and forecasting of hydrological phenom- Given the importance of the global hydrosphere to the global ena, making them essential in water management and flood energy balance, it is essential to integrate national baseline forecasting. These data support water-use operations (such networks into a more integrated "Earth system approach" that as irrigation, reservoir management, river or lake navigation, considers the planet as a whole, linking the atmosphere, ocean and power generation) and are used to ensure compliance and hydrosphere, terrestrial realm, cryosphere, and even the with licensing requirements—particularly for allocating water biosphere. The World Meteorological Organization (WMO)'s resources to competing uses; assessing the environmental, efforts in this area—particularly the activities to implement the economic, and social impacts of water management practices WMO Integrated Global Observing System (WIGOS)—should be (especially under critical hydrological conditions); and en- taken into account in network design and supported as part of suring the safety of people and property from water-related national data policy. Unfortunately, when a baseline network hazards (such as floods and droughts). Hydrological data in has operated for decades, stations with long-term records are near real-time are a precondition of early warning systems often abandoned because it is felt that no additional hydrologic (EWS) that are based on combined weather and water data. information can be obtained for planning purposes. However, In Europe, it can be estimated that hydrometeorological infor- given the climate variabilities and the long cyclic processes mation and early warning systems save several hundred lives in the atmosphere, such an assumption is often unfounded. per year, avoid between 460 million and 2.7 billion euros in Indeed, analyses of long observation series in many parts of disaster asset losses per year, and produce between 3.4 and the world show long-term changes in hydrological phenomena, 34 billion euros of additional benefits per year through the as the occurrence of low-flow and high-flow periods is dynamic optimization of economic production in weather-sensitive and quite variable. A short observation series simply cannot sectors (Hallegatte 2012). represent such variabilities and often leads to poor water man- agement decisions. Operational networks are also referred to as control and steer- ing networks. The control functions are not limited to the oper- In the past, the fundamental activity of an NHS was design- ators of water management facilities. They are also part of the ing and operating a basic network of observing stations, but regulatory function of the water administration, particularly today that role has expanded to being an information service when water is a scarce resource whose multiple utilizations 72    The Benefits and Value of Hydrological Services require state regulation. Because the government is the public task, the NHS should be reimbursed for the costs it incurs for custodian of water resources, it has the final authority about services it provides to a water user. rights of usage. This means that water resources are often owned, controlled, regulated, and allocated by the state. Water Institutional arrangements. Institutional arrangements resources should be managed in the context of a national water for dealing with water resources are extensively covered in strategy that reflects the nation’s social, economic, and envi- water legislation. Such arrangements designate one or (often) ronmental objectives, and that is based on an assessment of more government agencies with the ultimate responsibility the country’s water resources. This assessment has to be based for water resources (such as the allocation and supervision on hydrological data and information. As a public trustee of of water rights; the preparation of plans, programs, and poli- the nation’s water, the state is entrusted with the authority for cies; and enforcement provisions). Because the NHS provides regulating the use, flow, and control of all types of water in the the database for most water-related decisions, it should par- country. The most important obligations of a state to manage ticipate in these arrangements. As for decentralizing water water resources, which have to be incorporated into water leg- management activities—which is based on the principle that islation, (Salman and Bradlow 2006) are: “nothing should be done at a higher level of government that can be done satisfactorily at a lower level”—basin manage- Regulation of water use and allocation. The state is entrust- ment authorities responsible for developing water manage- ed with the responsibility for ensuring that water is allocated ment plans for a specific river basin should be established. equitably and beneficially, as well as with the authority for This requires a decentralized structure of stream-gauging regulating the use, flow, and control of all types of water in capacities in hydrometric branches. It offers the best op- the country. It issues a permit or license in accordance with portunity to work with water user associations, which are clear and transparent criteria and procedures before anyone responsible for operating and maintaining district water fa- can use water or construct water infrastructure. Such a permit cilities (such as treatment plants, reservoirs, and irrigation would describe the types of water use permitted, the quantity systems) and for collecting water fees. of water that can be used, and the water standards (which define, for example, the quantity and quality of the water dis- 2.2.4 Added Value and the Need for charged after use) with which the permit holder must comply. Hydrological Services The stream-gauging network is needed to monitor users’ com- pliance with these criteria, especially when distributing the Given that water is the core element of all socioeconomic allocation of water among competitive users. developments, it is not surprising that funding water-related activities comprises a substantial part of international de- Regulation of water infrastructure. In most countries, legis- velopment aid. The Organisation for Economic Co-operation lation endows the government with the responsibility for the and Development (OECD) reports that, in 2017, official de- use, protection, capital investment, and safety of all state- velopment assistance to water supply and sanitation and owned water infrastructure, while making private owners other water-related sectors totaled $11,321.5 million1 – with responsible for their own infrastructure. The government has 72 percent of this support going to projects in sectors that the authority to inspect privately owned water infrastructure depend directly on hydrological data. The biggest users are and ensure its compliance with standards, permits, and reg- water transport (25.2 percent) and agriculture water resourc- istration requirements (such as requirements for an environ- es (21.4 percent). mental impact assessment), as well as its proper functioning and safety. For this purpose, the owner can be obliged to It is difficult to imagine how these funds can be applied effec- self-monitor the firm’s water management activities, which tively without sufficient information on hydrological conditions, often include the installation and operation of gauges. The and the need for such will continue to increase. An important NHS should be responsible for the professional supervision reason why is the international community’s adoption of the of self-monitoring. If the water user is not able to fulfil this 2030 Agenda for Sustainable Development, which defines 17 These data come from the OECD’s Aid to the Water and Sanitation Sector (database) (accessed March 21, 2022), https://stats.oecd.org/qwids/. 1 The Benefits and Value of Hydrological Services    73 Sustainable Development Goals (SDGs), one of which centers on water and sanitation (Goal 6) and links to all the other goals. Moreover, many of the 169 targets in the Agenda cannot be achieved without also meeting the targets under Goal 6, and vice versa. For example, many of the targets related to social and economic development depend on and support a sustain- able, reliable water supply of adequate quality and quantity. In addition, with both synergies and potential conflicts to be man- aged in order to meet the SDGs, accurate and historical as well as current hydrological data and information will be critical. Integrated water resources management (IWRM) provides a framework for addressing many of the links between social and Taking samples of groundwater. Photo: BartCo economic development and water resources systems by bal- ancing the needs of sectors and stakeholders. Understanding the interplay within the food–water–energy–ecosystems nexus and a determination of how to balance the positive or negative impacts of new water management activities are unimaginable without hydrological data. However, although the importance of hydrological informa- tion and data is obvious, it is difficult to estimate its bene- fits, which fall into three areas: (1) design-related benefits, (2) flood warning and prevention benefits, and (3) sustain- inadequacy of the network of at least $63 million annually able resource management benefits. In all of these cases, the in capital and operating inefficiencies and losses throughout value of the hydrological network is determined by the need the resource sector of British Columbia. And it concludes that for the information it provides and by its ability to provide these loses could be mitigated by doubling the network size. accurate and representative data to meet that need—which, in turn, depends on both the structural infrastructure and the How about at the sectoral level? The study finds that in the hydrological conditions of the respective country. transportation sector—one of 15 sectors assessed—the eco- nomic benefits of hydrometric data primarily relate to cost One way to view the issue is to consider what the absence of savings in the design of stream-crossing structures. This oc- these data or poor-quality data would mean for planning and curs because, in the absence of adequate data, bridge designs design decisions. Here the answer is simple: it would result in tend to be more conservative, with the height and length of suboptimal or completely misguided decisions, which might the bridge typically increased to address the hydrological un- affect the reliability, security, and resiliency of infrastructure. certainty, thereby ratcheting up costs. In this Canadian cost But how much would this cost a country? Attempts have been study, a saving of 15 percent of the costs of these structures made to both estimate the intrinsic value of the network and has been assumed to be the maximum potential hydrometric do a cost-benefit analysis. One of these attempts is a 2003 benefits arising from use of hydrometric data for the design study of the stream-gauging network in British Columbia, of engineered bridges, trestles, and culverts. In 2003, it was Canada (BC Ministry of Sustainable Resource Management estimated that over the next few years, investment of $77 mil- 2003) (see annex 2.1). This study estimates the total annual lion annually would be made for the construction of engineered benefits (in 2003 US dollars) generated by the then-existing stream crossings, both in the public transportation and forest hydrometric program of 461 stations at $65 million, and the sectors. Assuming a maximum benefit of 15 percent, and that cost of operating and maintaining the existing network at hydrological data are the determining factor of design in 50 about $3.5 million annually—resulting in a cost-benefit ratio percent of the cases, the benefit of the current hydrometric sys- of about 1:19. It also estimates an opportunity cost due to the tem for this investment would be $5.8 million annually. 74    2.3 Efficient Organization of Hydrological Services A National Hydrological Service (NHS) has to provide high-quality hydrolog- ical data and information at the national level in a cost-effective way (WMO 2006). This demands systematic hydrological observations and data process- ing, which are the core of the development of databases. The processing of these data delivers information and knowledge required for the effective man- agement of water resources and protection from damage caused by floods or droughts. Despite major advances in hydrometric technology, hydrological monitoring networks are in decline in many parts of the world. This decline is reflected in a reduction in the number of gauges and a deterioration of data quality caused by increased uncertainty of rating curves and a loss of knowl- edge about secure handling and storage of data. Many parts of the information base required for integrated water resources management are outside the areas in which NHSs have traditionally tended to work. This implies that an Photo courtesy of M. Falcone, Water Survey of Canada. NHS needs proper institutional development to meet new challenges, which in turn entails a rethinking of how NHSs are structured. 2.3.1 The Basic Structure of a National Hydrological Service The main tasks of an NHS are: ■ To design and operate observation networks ■ To collect, process, and preserve data ■ To develop standards and quality assurance programs ■ To assess user requirements for water-related data and information and to provide such data and information (such as hydrological forecasts and water resources assessments). According to these tasks, the hydrological service can be divided into admin- istrative, technical, and hydrological structures. A typical structure of an NHS with a single central office is shown in figure 2.3.1. Here the technical struc- tures consist of a measuring instruments service and an ICT service. Further “Despite major advances regional structural units can be formed, as shown in figure 2.3.2, if the size of the country and the density of the measurement network require that. in hydrometric technology, hydrological monitoring The tasks of the structural units of an NHS can be characterized as follows: networks are in decline in Administrative unit. This unit carries out organizational tasks, manages eco- many parts of the world.” nomic resources, and develops human resources. Efficient Organization of Hydrological Services    75 FIGURE 2.3.1  Typical Scalable Structure of an NHS with One Central Office National Hydrological Service Gauge Administration Gauge Hydrological analysis Hydrology department modeling Gauge Measuring instruments service Hydrometric service Gauge ICT service Note: ICT = information communication technology; IT = information technology. FIGURE 2.3.2  Structure of an NHS with Two Regional Hydrological Offices National Hydrological Service Gauge Gauge Hydrological field Administration offices Gauge Hydrological analysis Gauge modeling Hydrology department Gauge Measuring instruments Hydrological field Gauge Hydrometric service service offices Gauge Gauge ICT service Note: ICT = information communication technology. 76    Efficient Organization of Hydrological Services Technical units. These units provide the measuring instru- water-level data rather than measured regularly (see box mentation service and information and communication tech- 2.3.1). This requires establishing a relationship between nology (ICT) service. The instrumentation service carries out the two variables, which has to be adapted to changing work on the technical equipment of the monitoring network flow cross-sections with repeated discharge measure- and ensures its proper operation and maintenance (includ- ments for a wide range of water levels. ing its calibration). It is responsible for the installation and ■ Hydrological analyses and modeling. In-situ water-level ob- commissioning of measuring equipment and the provision of servations are the backbone of hydrological services, but methodological and technical guidance for the proper oper- a wide range of necessary water resources applications ation of the equipment. The ICT service plays a major role and decisions are based on derived discharge information. in ensuring the timeliness of data provision and dissemina- This unit is responsible for: tion of hydrological information (such as status reports and p Providing derived hydrological information, which forecasts). In addition, a centralized data center oversees the are based on time series (statistics, expert opinions, continuous data transfer, processing, analysis, archiving, and design values, and water balances) dissemination of all types of data and hydrological informa- p Cooperation with users by adapting the hydrological tion. It is tasked with (1) the procurement and installation products to their requirements of the full set of the hardware and software required for a p Technical contributions to the justification of water modern forecast and warning system, and (2) the installation management decisions (for example, by analysis of and technical support of a service delivery platform and ap- ongoing changes and impact assessments for regional plications (such as data management and dissemination of and national water balances) hydrological information). p Development and application of hydrological mod- els and the provision of hydrological forecasts and warnings. Hydrology department. This department has the overarching responsibility for all hydrological activities. It aggregates the A direct link between monitoring activities and a provision of hydrological data to hydrological information; ensures the hydrological information at the regional level is beneficial for quality assurance of an NHS’s products and services; verifies the quality of services. The hydrological service can fulfil client that standard procedures are being followed (for example, requests, which may vary with actual hydrological conditions, by independent checks on flow-rating curves or field work); and it can ensure direct quality control of hydrological data validates that archived data meet defined standards (for ex- with short reaction times to fill gaps or ensure an adaptation of ample, by cross-comparison between neighboring stations); rating curves to changing conditions of flow cross-sections. In and ensures that all aspects of the system are being consis- the hydrological field offices, at least one engineer-hydrologist tently monitored. Within the hydrological department, there should be allocated to provide derived hydrological informa- are two subunits: tion and to supervise the hydrological technicians. ■ Hydrometric service. This operates the hydrometric net- work, provides hydrological data, and (if applicable) su- The hydrometric technician’s main duties are to conduct pervises local hydrometric activities of the hydrological discharge measurements and respond to the difference be- field offices. It is scalable, depending on the network’s size. tween what is found and what was expected should the dis- If the service is organized in a decentralized way by local charge measurement fall off the stage-discharge curve (see hydrological offices, its activities should not be limited to box 2.3.1). The quality of the derived flow data is highly hydrometry but include the provision of derived informa- dependent on the ability to specify the temporal variabil- tion about regional hydrological conditions, in cooperation ity of this relationship. Here the frequency and the quality with and under the supervision of the unit for hydrological of actual discharge measurements that are used to confirm analyses at the central office. The provision of hydrologi- the need of stage-discharge changes is a very important fac- cal data is a complex process because the most important tor—one that determines the hydrometric workload. Because parameter is discharge, which is calculated from observed of this complexity and cost, typically only a limited number Efficient Organization of Hydrological Services    77 BOX 2.3.1  A Quick Guide to Estimating Discharges In-situ flow estimation is based on well-established procedures consistent with international practices (WMO 2010). A common misconception for most stream-gauging systems is that they measure hourly or daily discharge in streams. In reality, hourly or daily discharge is a derived result. Time-based water level, or stage, is measured continuously at each stream-gauging station. Rating curves that are relationships between stages and discharges are applied to transform these data into discharges. These curves are site-specific and are a temporal variable, as the cross-section is changing. Periodic discharge measurements are required to set up rating curves, to control them, and to update them. The discharge is measured indirectly by measuring flow velocity (v) and cross-sectional area (A). Thereafter, discharge (Q) is calculated by mul- tiplying the flow velocity with the cross-sectional area flowed through: Q (m3/s) = v (m/s) × A (m2). This velocity-area method requires punctual measurement of the flow velocity over the cross-section by current meters. Modern ultrasonic Doppler profile flow meters are able to integrate the flow velocities over the area of the cross-section and to simul- taneously estimate the cross-sectional area. Discharge measurements are indispensable and cannot be replaced by hydraulic modeling. It has to be considered that the velocity depends on the flow cross-section and vice versa, and that the cross-sec- tion is formed by hydraulic conditions, which are determining the flow velocity and the variation of riverbeds. An after-stage discharge relationship (“rating curve”) is established once sufficient discharge measurements have been made to control and adapt it to ongoing changes of the hydraulic conditions. Over time, the measured discharge value may not fall on the predetermined open water curve because other natural factors influence the hydraulic capacity of a stream. These temporal changes can be subdivided into two components: (1) seasonal variabilities caused by aquatic vegetation growth and die-off or icy conditions in wintertime, and (2) variabilities resulting from aggradation/degradation of the channel. Periodic discharge measurements and interpretation of the temporal variability of the discharge-stage relationships based on local knowledge are used to adjust rating curves. As a result, hydrometric technicians have to continue with periodic discharge measurements even after a curve has been established to know when changes to the rating curve happen. In order to allocate sufficient capacities to a hydrometric unit, the engineering hydraulic principles, local conditions, stream stability, and many other factors need to be understood. The stability of the rating curve dictates the number of field measurements required to obtain a desired accuracy. Daily, hourly, or even higher resolute discharge values (intervals of up to 5 minutes in the minimum are possible) are derived from the water-level record, based on the rating curve and manual or computerized interpretation. of monitoring locations provide discharge data where such adjustment or new derivation of a rating curve is often neces- data are of absolute necessity (such as for water resources sary because of changes in channel morphology resulting from assessment, flow estimation for a certain user, or hydrolog- the flood. Even during non-flood events, the growth of aquatic ical modeling). vegetation or the deposit of sediment will result in a change to the flow cross-section, and the rating curves are often sea- The workload to operate a stream-gauging station and to de- sonally variable and have to be changed accordingly. Thus, the rive good-quality flow data depends heavily on the local con- required capacity for hydrometric service depends not only on ditions at the gauging site. Because the stage-discharge curve the number of monitoring sites but also on the variability of must cover the entire range of discharge, from low flow up to gauging profiles according to the local hydraulic conditions. extreme floods, the stream-gauging activities are unevenly dis- tributed during the year. In periods with high temporal fluctua- The tasks of hydrometry are also spatially distributed among tions in runoff, a higher frequency of discharge measurements the different river basins of a country. The density of oper- is required. This is especially important during flood sea- ational stations needed depends on the degree of water sons, when the number of discharge measurements have resource utilization and on the risk of adverse impacts of to be increased significantly. Notably, after a flood event, an hydrological extremes on the society. In the design of the 78    Efficient Organization of Hydrological Services stream-gauging network, the heterogeneity of the hydrolog- data users and its capabilities to fulfill demand. However, rel- ical conditions within a country has to be considered. If the evance, efficiency, and effectiveness strongly rely on client/ natural conditions do not vary much, characterization of the user engagement and feedback mechanisms. regional hydrological conditions is less challenging. Often the stream-gauging stations are unevenly distributed within A basic tenet of any NHS is that it is a business and should be a country. The spatial distribution of gauges—which defines run in a business-like way. This does not imply that an NHS the accessibility of the sites and the travel time to reach these must become a profit-making business, or that it may ignore stations—has to be taken into account while planning hydro- its public service responsibilities. It does imply, though, that metric service’s activities. It often makes economic sense NHS managers should: to set up regional hydrological field offices (figure 2.3.2) to reduce travel times and keep in close contact with regional ■ Recognize that their prime reason for existence is to serve users of hydrological data and information. their clients, the bulk of whom are the general public; ■ Adopt or develop (within the constraints set by their par- The hydrometric unit obtains the data from the stream-gaug- ent organization or government policy) administrative, ing stations; it then performs a quality control check; charac- management, and leadership practices that produce the terizes the current hydrological situation in daily, weekly, and best possible results in terms of efficiency, effectiveness, monthly reports; organizes public information; and provides and responsiveness to clients; and open hydrological data to users. On demand, it offers a mon- ■ Aim to ensure that the NHS is a growing concern whose itoring service, designed to provide very specific data or in- assets are not depreciating and whose prospects for future formation at a particular location for a particular client. Given business are expanding rather than contracting. the heterogeneity of the observation network density, a de- centralized hydrometric service could be more effective to re- A forward-looking strategy of an NHS should meet the fol- duce travel times and to allocate hydrometric capacities. This lowing aims (this list is based on WMO 2009; this and other is especially true in states with large geographical extents WMO recommendations and guidelines to hydrological topics and time-varying requirements for discharge measurements. are currently being revised and updated): 2.3.2 Fit for Purpose? Determining the ■ Identify applications areas of hydrological data and infor- Institutional Structure’s Efficiency mation and current and potential clients of the NHS and maintain and update a client database. To evaluate the efficiency of an NHS, its products have to be ■ Identify the products and services required by the clients evaluated in relationship to the available capacities in the country. Within any NHS, there must be clear alignment and that the NHS can provide. understanding of the impacts and benefits of improving the ■ Identify the most suitable mode or place of delivery of the hydrometric system. This can be particularly challenging if product or service to the client—such as using the internet there are responsibilities for water management issues dis- to provide access to real-time data; emails and short mes- tributed among multiple departments or ministries with sage service–based warnings; and hydrological cell phone responsibility for flood forecasting, irrigation, hydropower, applications with the possibility of subscribing to alerts. water supply, and disaster risk management. Often the de- ■ Determine a pricing policy for products and services, and mand for the hydrological data and information is outside the for clients. areas in which hydrological services have traditionally tend- ■ Specify the types of people involved in delivering the prod- ed to work. This implies that an NHS must develop new capa- uct or service. bilities or establish partnerships or strategic alliances with ■ Characterize the processes of product or service delivery complementary agencies. The efficiency of the data transfer, according to the needs of the clients. the practical applicability, and, as a result, the visibility of ■ Promote the NHS where potential clients can be clearly the NHS are all determined by the degree of cooperation with identified and contacted directly. Efficient Organization of Hydrological Services    79 To adapt its products to new requirements, the NHS should or rehabilitating aquatic ecosystems. Reliable data on water continually monitor trends in demand for water, national and quantities are the backbone of international cooperation in provincial policies and development goals, trends and events transboundary watersheds (Dixon et al. 2020). in various economic sectors, and international agreements. By implementing a Rolling Review of Requirements (RRR) Depending on applications, the NHS has to provide: process, the NHS should ensure that user requirements for observations are compared regularly with the capabilities of ■ Data in real time and situation reports about current hy- present and planned observing systems (WMO 2021a). drological conditions, along with forecasts of discharges in the near future (hours, days, weeks) The main applications of hydrological data and informa- ■ Long time series of observed discharges and their statisti- tion are related to water management activities—which is cal analyses an element of several sectors including agriculture, power ■ Seasonal forecasts, especially for droughts and reservoir production, and water supply (for industry and the public) management (table 2.3.1). In addition, the design of nearly all elements ■ Water balances for river basins to characterize the hydro- of infrastructure (roads, bridges, urban drainage systems) logical conditions in the past and future (impact assess- requires hydrological data and information, as does envi- ments of changes). ronmental monitoring for water pollution and for protecting TABLE 2.3.1  Main Applications of Hydrological Products Application of hydrological data and Real-time data Derived hydrological analyses information Client needed and expert opinions Public safety National disaster management Flood statistics, design floods, flood ■ Flood management, risk assessment agency, communities forecasts, and warnings and management X ■ Flood EWS ■ Flood defense Water resource assessment and planning State agencies; International Water balances for river basins, of water projects cooperation in transboundary verification of compliance with watersheds international regulations Planning, design, and operation of Private companies, state authorities, X Safe yield estimations, design reservoirs municipal utilities, water boards floods, statistics of discharges Irrigation and drainage Agricultural authorities and agencies X Water availability, seasonal drought (e.g., irrigation associations) forecasts Hydropower and energy-related projects State agencies, private companies X Design floods, statistics of discharg- es, environmental flow requirements Navigation and river training State agencies X Statistics of discharges Urban drainage Communities X Statistics of discharges Sediment transport and river channel State agencies Statistics of discharges morphology Water quality and hydro-ecology State agencies X Statistics of discharges, low flow indicators, water temperature Design of nearly all elements of State agencies, supply and disposal Flood statistics infrastructure companies, construction companies Source: World Bank. Note: EWS = early warning system. 80    Efficient Organization of Hydrological Services The users of the data and information from the NHS are the strong links between the terrestrial and atmospheric ele- distributed among various ministries and institutions (for ments of the hydrological cycle, almost all hydrological prod- example, the country’s ministry of environment, ministry of ucts rely on climatological and meteorological data. construction, ministry of energy, ministry of agriculture, state water resources management administration, disaster man- Meteorological services are usually divided into two broad agement, river basin commissions, water boards, agricultural classes: weather services and climate services, based on the irrigation boards, and municipalities). While some hydrolog- characteristic time scales of weather (minutes to weeks) and ical products are region-specific and have to be coordinated climate (months to centuries). Hydrological applications do with their users (for example, flood forecasts, water balances not necessarily make this distinction, and cooperation is re- for river basins, forecasts for reservoirs, or navigable water- quired at both temporal scales. For example: ways), others can be standardized (for example, statistical analyses of discharge series at gauges). ■ Water resources assessments require water balances for river basins. These balances are quantitative relationships Notwithstanding the great importance of hydrological data between precipitation and evaporation to estimate the and information for economic and social development and climatic availability of runoff. In addition to precipitation for disaster management, the public visibility of an NHS is data, long time series datasets of temperature, humidity, low and often visible only during critical hydrological situa- and wind speed are needed to derive evaporation. tions. To better meet the needs of users, it would be helpful ■ Hydrological forecasts are based on the timely provision of to establish a customer service area in the hydrological de- measured or predicted precipitation data. partment to receive inquiries, advise users on how to use the products, and work with users to initiate new or improved The application of hydrological models requires state-of-the-art products. To increase the awareness of the benefits of hydro- applications for both tasks, and with the recent advancements logical services, the following steps are recommended: in modeling, the demand for meteorological and climatological data in hydrology has risen significantly. Despite these advanc- ■ Conduct a market analysis for needed services; work close- es, none of the hydrological models produce reliable results ly with stakeholders from the state, municipal, and private without a calibration of model parameters and their validation. sectors; and provide the needed products. For these purposes, long time series of meteorological param- ■ Analyze existing deficits (for example, to ensure the up- eters are required. The operational use of the parametrized date of rating curves) and needs for updating products models for forecasting also depends on the availability of actu- and explain to users the existing limitations and future al meteorological and climatological data and forecast. opportunities. ■ Set up a program of priorities to gradually expand the Take the case of the design and development of a flood forecast- range of services in terms of space and content. ing system. Here there is a need for historical and real-time pre- ■ Ensure that frequently requested products (for example, cipitation information that is highly dependent on orographic flood statistics) can be provided online. conditions and land use, along with a spatial distribution of rain gauges that is suitable for this purpose, with high temporal res- 2.3.3 Inter- or Intra-Agency Cooperation with a olution of the measured data and online data transmission to National Meteorological Service the NHS. Moreover, other components of the meteorological ob- servation system are not always directly useable for hydrolog- NHSs provide information on the past, present, and future ical purposes. Examples are weather radars, where the rainfall state of rivers, lakes, and other inland waters, with a spe- rate is related to the median diameter of rain drops or the water cial emphasis on streamflow. These services focus mainly equivalent of the snow cover, which can be interpolated through on the surface component of the water cycle, through which a combination of in-situ gauges and radar data. In such cases, the rainfall over a catchment is partitioned between storage, cooperation between the NHS and the National Meteorological runoff, and evaporation back into the atmosphere. In view of Service (NMS) is essential to provide meteorological data Efficient Organization of Hydrological Services    81 relevant for hydrological analyses and forecasts. The benefits ■ Level 3. There is a moderately well-functioning relationship are not one-directional; many NHSs collect ancillary hydrolog- between the meteorological, hydrological, and water re- ical information (precipitation, snow courses) that can provide sources communities, but considerable room for formalizing valuable insights to an NMS as well. the relationship and standard operating procedures (SOPs). ■ Level 4. The meteorological, hydrological, and water re- Hydrological applications require meteorological and clima- sources sectors have a high-level formal agreement in place tological data that are often customized to the hydrological and an established working relationship, and data sharing service. For example, public weather forecasts are usually in- takes place, but institutions still tend to develop products adequate to meet the hydrological prediction needs related to and services in isolation. spatial and temporal variability of precipitation. Quantitative ■ Level 5. The meteorological, hydrological, and water re- precipitation forecasts (QPFs) derived from numerical weather sources sectors have strong SOPs and agreements in place, models close the information gap with a high practical rele- allowing them to work closely together in developing new vance for an NHS. Good cooperation with an NMS is essential and improved products and providing seamless and ad- for ensuring the provision of sound hydrological products. One vanced services. example of this cooperation is seen in meteorological ensem- ble forecasts, which specify the uncertainties of QPF and tem- Close cooperation is also very useful for collaborating with peratures, based on current meteorological conditions. Here stakeholders and clients. This is especially true for those in- the resulting hydrological products (forecasts) depend on com- volved directly in the water sector, but also true for many sec- munication processes that influence the ability of hydrologists tors where both hydrological and meteorological information to interpret or ingest the meteorological data and information are valued (table 2.3.2). for further application. This requires the awareness on behalf of the meteorological service on the criticality of its products Higher levels of cooperation could be reached more easily by and services to an NHS’s value cycle (WMO 2021b). merging hydrological and meteorological services. The goal is to produce specific benefits from synergies through coop- The level of cooperation can be assessed by the following erative arrangements that minimize duplication and facilitate criteria: the sharing of technology. For example, a shared data center could improve the exchange of data and information between ■ Level 1. There is no or very little meteorological input in hy- hydrological and meteorological departments while increas- drology and water resource management. ing the effectiveness of the use of newly installed IT systems. ■ Level 2. Meteorological input in hydrology and water re- Despite these advantages, most NHSs (69 out of 101 coun- source management happens on an ad hoc basis and/or tries responding to a 2020 WMO survey) are administratively during times of disaster. separate from National Meteorological Services (WMO 2020). TABLE 2.3.2  Sectors Using Both Hydrological and Meteorological Services Economic sectors Public safety Natural resources ■ Water supply ■ Tourism ■ EWS ■ Water resources assessments ■ Manufacturing ■ Agriculture ■ Defense ■ Natural resources management ■ Energy ■ Transportation ■ Emergency management (forests, coasts, terrestrial, and marine ■ Insurance and ■ Construction ■ Health ecosystems) finance ■ Mining ■ Transportation safety Source: WMO 2015. 82    2.4 Development of Human Resources At its core, a strong NHS consists of two basic structures: first, the data col- lection and management side in the form of gauges, monitoring equipment, ICT infrastructure, and other hardware and software that is required to run a complex data-intensive monitoring and production system; and second, the human resources with the skills necessary to operate this technical system and to ensure robust data collection, transmission, processing, and data anal- ysis (figure 2.4.1). Photo courtesy of M. Falcone, Water Survey of Canada. The sustainability of modernized hydrological observation networks is crit- ically dependent on a highly skilled scientific and technical workforce. A modern National Hydrological Service (NHS) demands highly trained and educated personnel with specific competencies that range from technical abilities to engineering and scientific hydrology. The requirements for the qualification of employees changes with the adoption of new technologies and stream-gauging equipment. Whereas some decades ago measurement technology was based on precision mechanics and data were registered on paper and sent by post, now electronic measurement systems and modern (often wireless) data transmission dominate. Installation and maintenance “Hydrological observing require specially trained staff who can not only replace and repair hardware but who can also adapt software for these systems. Recruiting qualified per- systems have evolved with sonnel to operate more complex measuring instruments and ICT equipment, the aid of scientific and for example, or to develop software solutions for combining sensors and data- bases, is made more difficult by the fact that the NHS has a poor starting po- engineering assessments. sition in terms of competition with other sectors of the economy whose wage New technologies are structure is more attractive. This requires the NHS to increase the wage level for highly qualified experts—a cost that is often not sufficiently considered in becoming available for modernization projects. various aspects of network Hydrological observing systems have evolved with the aid of scientific and engineering assessments. New technologies are becoming available for vari- operations (such as ous aspects of network operations (such as instruments, transition, and pro- instruments, transition, and cessing software), but training systems are still needed to maintain staff skill levels in adopting new equipment and for other operational technologies. An processing software), but NHS should be able to assess its educational and training needs, as well as to training systems are needed provide and access training within the country to support staff development. Failure to adequately address these challenges is a significant risk factor not to maintain staff skill levels only in network maintenance over the long term but also in project design and in adopting new equipment implementation. and for other operational Any project planning for modernizations should assess the existing capac- technologies.” ity of the hydrometric organization at an early stage to understand what Development of Human Resources    83 FIGURE 2.4.1  Adaptation of Human Resources Needed for NHS Technological Development An Effective NHS Higher requirements for human resources Hardware and Reliable, timely Human resources technological hydrometric data improvements for ■ People and hydrological Requirements ■ Education long-term Gauges + ■ of users ■ Skills information, ■ ICT targeted toward ■ Training ■ Data stakeholders and ■ Methodological management users developments ■ Data analysis Qualified operation and maintenance Note: ICT = information and communication technology. resources are available and what change is required. A sus- applied engineering discipline with a strong emphasis on tainable solution for staff resourcing must be found early on design values to a special field of geosciences while still in any funding stream, as the lead time to develop and train maintaining a focus on real-world applications. Developing staff could be between three and five years. In the short term, countries with low levels of resilience are faced with serious gaps in capacity can be met by using local and internation- direct and indirect consequences of changing climate and al consultants. But it is important that all international and land use. A growing demand for water and greater exposure expert consulting staff have a requirement within their proj- to floods and other hydrometeorological extremes require in- ect specifications to train and develop local staff. Their goal tegrated hydrological services to help manage these realities. should be to “make themselves redundant” over the course of Given the non-stationarity of hydrological conditions, analy- the project. In Sri Lanka, the Irrigation Department suggested ses of observed time series are no longer sufficient to answer that each international expert be shadowed and assisted by the question: What happens when? This realization has been two key local experts from the department whose objective particularly transformative in hydrological science. was to learn and take over these roles. For long-term success, there must be a clear plan on how to train, recruit, and devel- As in many other disciplines, the specialization of experts op staff that will be able to maintain, operate, and improve in hydrology is an ongoing process. In the United States in the systems that are put in place. 1958, roughly half of practicing hydrologists had a civil engi- neering background—and the vast majority had received their As for water resources assessments and hydrological mod- hydrology training through field practice and applied appren- eling, hydrology has evolved from a mainly problem-driven, ticeship, often within governmental resource management 84    Development of Human Resources agencies rather than through formal university study (Ruddell Given that in many developing countries the number of and Wagener 2015). Only in the 1960s did formal hydrolo- post-secondary institutions that offer hydrological programs gy education begin in several parts of the world, and not are limited or nonexistent, providing at least temporary learn- until the early 2000s did advances in science, technology, ing and training capacities to foreign guest lecturers, and engineering, and mathematics education begin to benefit even program chairs at colleges or universities, could close the hydrological community. At that point, education in hy- this gap. International student fellowships to qualify spe- drology became interdisciplinary, systems-oriented, applied cialists in subareas (for example, in hydrological modeling) but rooted in science, both broad and specialized, and inter- in international institutions (for example, at the Institute for departmental, with an added emphasis on political, social, Water Education in Delft)2 could be a way to boost the quali- legal, and economic aspects of water resources management. fications of NHS hydrological experts. Unlike in past decades, What does this mean for education and training in develop- however, today’s graduates from these programs are being ing countries? The complete adoption of courses in hydrology offered more employment choices, such as in the fields of en- and water management from highly developed industrialized gineering, computing, and ICT. There are also many interna- countries in their entirety is usually not possible—a limita- tional and online training modules available. The Hydrology tion that applies also to small, developed countries in Europe, and Water Resource Program of the WMO makes available whose higher education in hydrology is constrained by the support for the hydrological community. Similarly, the United small number of students and limited opportunities to work Nations Educational, Scientific and Cultural Organization in this field. (UNESCO) offers online and direct learning through its in- ternational hydrology program. The University Corporation What can a developing country’s NHS do against this back- for Atmospheric Research (UCAR) in the United States also drop? Priorities would be to: provides a series of free online training courses in hydrolo- gy through its COMET portal (https://www.comet.ucar.edu/). ■ Assess its educational and training needs in light of cli- There are numerous other examples, and many universities matic conditions and the requirements of their particular have recently moved to online learning, affording new oppor- country’s users of hydrological information. tunities for developing countries. ■ Create education and training opportunities in the form of basic courses with high application potential in hydrology Hydrologists require local knowledge and experience, on top (for example, statistics, geoinformatics, programming soft- of their scientific and technical expertise, to place hydrolog- ware for sensors, and remote sensing) by local (not neces- ical data into the context of prevailing watershed conditions sarily hydrological) professionals. and to evaluate the hydrological regime. Addressing hydrolog- ■ Organize a program of training in hydrology with foreign ical issues often requires adapting methods to the specifics of experts, given regional hydrological conditions and the re- particular site types. Knowledge of local and regional charac- quirements of users. teristics is also essential for building and using hydrological models to make predictions that are useful for site-specific Hydrology education is a challenging, complicated, and valid decisions. Large staff turnover demands repeated invest- area of scholarship (Ruddell and Wagener 2015). In most hy- ments in education and training. Furthermore, if many experi- drology courses, generic methods are taught that can more or enced hydrologists leave at the same time, a very uneven age less be applied anywhere as long as some basic criteria are demographic leads to loss of expertise and local knowledge met. However, hydrological systems are not that simple, and of the regional hydrological conditions. At the same time, an generic equations are not easily translated into local solu- obsolete workforce and a lack of young employees may result tions. Hence, experience with specific hydrological systems in a low willingness to apply new technologies, develop and (for example, semiarid or mountainous) is crucial. IHE Delft Institute for Water Education is the largest international graduate water education facility in the world. Over 23,000 water professionals from 2 more than 190 mainly developing countries and countries in transition have been educated at the Institute since 1957. IHE Delft works under the auspices of UNESCO. Details about the Institute are available at https://www.un-ihe.org/. Development of Human Resources    85 apply new methods, or work in an interdisciplinary environ- and building access to international education and train- ment to solve new problems. ing programs; ■ Developing competency-based profiles that meet work- Senior technical staff and supervisors require the domain ex- force requirements commensurate with agency-level edu- pertise (such as people/process management and meteoro- cation and training plans. Learning plans for individuals logical/hydrological skills) to ensure the continuing effective afford the opportunity to adopt self-paced learning ob- operation of the network. Supervisors are responsible for en- jectives that can be measured by milestones to develop suring the competence and effectiveness of staff within their within the workplace. There is also a need to encourage teams. Performance management is a key aspect—one that the development of priorities and curricula for regional includes defining job scope and targets, performance mon- training centers and to facilitate staff access to in-situ and itoring, building on strengths and addressing weaknesses, remote learning resources; and dealing appropriately with disciplinary and inefficiency ■ Establishing career development programs that recognize cases. Also critical are strategic and operational planning, and reward high performers in the workplace with further human resource management and succession planning, training qualifications and promotion as an incentive to budget preparations, business case development, and com- stay; munications to senior government officials. However, access ■ Creating a work culture of sharing within the workplace to to skilled management staff who possess the needed back- strengthen the transfer of experience and know-how—es- ground and knowledge of hydrological systems and opera- pecially given that local conditions play a pivotal influence tions is often in short supply. in hydrology; and ■ Investing in development and career training for supervi- The bottom line is that the development of human resources sory staff, who play an important role in guiding staff, on- requires: the-job teaching, and mentoring operating personnel. ■ Establishing strong links with local academic institutions Finally, if the funding for the modernization of the NHS is (for example, supporting investments in science and en- placed within the framework of single projects, it would be gineering disciplines), building partnering schemes to better to extend the funding over longer life cycles or estab- create recognized pathways from education into careers, lish development funding for ongoing support. 86    2.5 Design of Hydrometric Networks In the past, hydrometric networks were installed and extended mainly accord- ing to the need for obtaining information about the water resources available for specific water management projects (such as the installation of large hy- dropower systems in the former Soviet Union), or to improve flood protection for specific sites. As the demand for data for water management and the flood risks change with time, these networks have to be adapted according to new requirements. Thus, it is essential that a network be adapted with time to the hydrological conditions, the needs of its users—including the regulatory functions of the state in the management of water resources—and the available financial and human resources of the National Hydrological Service (NHS). The determining factors for network sustainability are the number of stations, the regional hy- drological and the local hydraulic conditions, and the technical issues around Reservoir in Asturias, Spain. Photo: ilia-art measuring equipment. Also important are the required costs for capital invest- ment, operations, and maintenance, as well as the personnel needed to do so. Any modernization or new installation of a hydrometric network requires a planning approach, referred to here as design. Critical to its design are (1) choosing the locations of the gauges according to possible uses of the data to be obtained, (2) selecting the technical equipment according to the local con- ditions, (3) assessing the required human and technical capacities to operate the network, and (4) comparing the capabilities of the service to provide the required capacities to operate the network in a sustainable way. “… it is essential that a These criteria, which have to be applied in the design phase, are interrelated. For example, it is important to consider the different requirements for the network be adapted with data from the stakeholder or NHS perspective. Not all data must be continu- time to the hydrological ously available at a high temporal resolution. Sophisticated equipment that provides data with a resolution of minutes but require high service capacities conditions, the needs of and expensive spare parts will probably have greater downtime than a sim- its users—including the ple system of manual observations by observers who manually read out data twice per day. regulatory functions of the state in the management of However, for both sophisticated and less-sophisticated equipment, the re- cords from all stations have to be of high quality. After all, hydrometric re- water resources—and the cords may be of little value if the stations are not operated correctly and if the reliability of data is low. This implies that the long-term quality of the available financial and human data must be taken into consideration when designing a network to maintain resources of the NHS.” a sustainable and viable system. Here quality depends on (1) the site of the Design of Hydrometric Networks    87 gauge; (2) the installed technical equipment; (3) the capacity The design of a hydrometric network requires a multitude of to operate, maintain, and replace it; (4) the data transmission considerations to be effective and to deal with the changing system; (5) the quality control of incoming data; and (6) the realities faced by the many local stakeholders who rely on data storage capacity. the information. One of these considerations is the need to distinguish between real-time needs (for flood forecasting or Every design for a hydrometric network should be based on reservoir safety) and climatic data needs (long time series a specification of the demand for hydrometric observations for planning and design purposes). Often the importance of and the specification of the measurement program. Such a high-quality data depends on its use. For analyses of flood network can be subdivided into two parts: its stream-gauging frequencies and the provision of design floods, high accuracy component (gauges to estimate water level and flow at sur- is required in the determination of the peak discharges, but face waters) and ancillary hydrometric stations for specific during a flood, timeliness of data reporting is often the pri- purposes (such as climatological and meteorological stations, ority, whereas accuracy can often be reviewed and adjusted snow observation sites, and groundwater observations). The after the event. focus here is on the stream-gauging network, but recommen- dations are also given for snow courses, sediment, water Despite the complex influencing factors, difficult local hy- quality stations, and climatological stations. draulic conditions, and elaborate needs analyses required to justify where stream-gauging devices should be placed, it is 2.5.1 Requirements for Hydrometric often possible to meet many requirements for data in a single Observations system design. Modern hydrometric production systems en- able both real-time and detailed archival analysis, but they The main driver in establishing a hydrometric system is the also present unique maintenance and operational challenges. demand for data. This demand has several facets that have to be considered: The main product of stream-gauging services—the runoff or flow—is the result of different hydrological processes within a ■ For water resources assessment, it is necessary to repre- watershed. Hydrometric observations are essential to under- sent the spatial and temporal variability of the hydrologi- stand and quantify these processes. However, in order to put cal conditions, which depends on natural conditions (such these data into a spatial and climatic context, further sources as climate, geomorphology, and soils). Here, it is the rep- of information are needed (WMO 2012): resentativeness of the stations that determines the ability to transfer the derived hydrological information from ob- ■ Characteristics of watersheds, such as drainage area, to- served to unobserved sites. pography, soil, geology, land use, and natural vegetation, ■ To control water management activities; assess their en- preferentially provided with a geographical information vironmental, economic, and social impacts; and modify or system (GIS) optimize the water management strategies, stations have ■ Climate data, such as precipitation, temperature, humid- to be allocated in accordance with the water management ity, wind speed, cloudiness, synoptic information, and significance of human interventions in the hydrological forecasts and alerts; as well as medium- and long-range cycle. weather forecasts ■ For flood forecasting, it is necessary to have actual runoff ■ Socioeconomic data, made necessary by urbanization and data upstream as an input into forecasting systems, togeth- water withdrawal from surface and groundwater, water ab- er with stations close to the most vulnerable areas where straction, and consumption. flood hazards are concentrated to assess the tendencies of rising or falling water levels ongoing during flood events. To ensure these data at the location of interest, often a data exchange between operators of different observing systems is required. 88    Design of Hydrometric Networks 2.5.2 Specification of the Measurement observation sites for the minimum network, it is vital that the Program records at all stations be of good quality, as they form the baseline for development. When designing the stream-gauging network, it is often the case that requirements for information from different poten- Hydrometric network development is an evolutionary pro- tial water users or about the hydrological regime are still un- cess, starting with a minimum number of observing stations known (or unknown in sufficient detail). The concept of an and increasing gradually (as necessary) until an optimum net- optimum network—the ideal situation—can be adapted to work is achieved. This occurs when the amount and quality of achieve a realistic near-optimum network design based on data collected and information processed are economically surrogate measures, objectives, and other criteria. According justifiable and meet the stakeholders’ needs. The density of a to the World Meteorological Organization (WMO 2008), cre- hydrometric network should provide reliable information on ating a full-scale and complete network is either impossible the hydrological characteristics within a certain region. This or impractical. Therefore, various surrogate approaches are minimum network is intended as a first step toward satisfy- used to derive flow information and make inferences about ing the most serious gaps in water resources assessment and the hydrology. Based on the experience of NHSs around the gaining an overview of the hydrological characteristics within world, WMO advocates that an optimum network should not a country. The further extension of a network must be a grad- be attempted until a minimum number of stations has been ual process, which should be reconsidered from time to time established. This minimum network is intended as a first step with respect to emerging new water uses and with specific toward addressing the most serious observation gaps for needs for hydrological information, as well as evolving needs water resources development. But given the small number of of existing users (figure 2.5.1). FIGURE 2.5.1  Development of Hydrometric Networks as a Cyclic Process Information utilization and decision-making Hydrometric network Information design based on user dissemination requirements NHS Data synthesis and analysis Data sensing and (including modeling, recording forecasting) Data validation and archiving Note: NHS = National Hydrological Service. Design of Hydrometric Networks    89 2.5.3 Stream-Gauging Network Design information for flood forecasting models. Frequency of data required (and thus frequency of observations) depend on the The five primary elements that make up the design of a reactiveness of river basins and on their size. Typically, for stream-gauging network are discussed below. large plain rivers with slow formation processes, hydrological forecasts are issued with a 12- to 24-hour temporal resolu- 2.5.3.1 Network Design for Water Resources Assessment tion. For small mountainous flashy catchments, observation data of high temporal frequency are required—in this case, an The basic network, intended for water resources assessment, hourly to sub-hourly frequency of observations is vital. One should satisfy spatial and linear interpolation of hydrologi- of the most devastating hydrological phenomena are flash cal elements (runoff characteristics) so it can continuously floods, which annually cause human causalities and lead to represent fields of hydrological elements with sufficient significant economic damage. Flash flood warning systems quality. Statistical approaches and regression analyses have have specific requirements for the observing gauges and the widespread application in network design. There are also a monitoring program: number of good practices in terms of having sufficient net- work density in different natural conditions (WMO 2009) and ■ Upper reaches (flashy basins) should be equipped with au- a number of criteria to use to identify the number of gauges tomated stations. that is sufficient in particular parts of a river basin (see Nixon ■ Data from automated meteorological stations (especial- 1996, for example). To specify the required density of a net- ly rainfall gauges) should be integrated into hydrological work of stations, a common substitution is to maximize infor- forecasting systems directly (that is, establishing a data mation content in lieu of optimizing the economic value of exchange). the data. If information is used properly, it can be expected to ■ There should be a larger number of stations on small foot- contribute to the economic worth resulting from a decision. hill rivers. 2.5.3.2 The Design of Operational Networks For large river basins, the flood propagation in middle- and The design of operational networks is strongly affected by low-river course has to be simulated with hybrid models, the informational needs of the respective water management based on hydrological and hydraulic components. Their oper- tasks. For flood monitoring and forecasting, the location of ation requires the data from a chain of hydrological stations stream-gauging stations depends on the location of inundat- located along the main river reach. In addition to these gaug- ed areas and damage that could be reduced by early warnings es at the main river, several gauges are needed at tributaries based on timely data, the types of floods (steep or flat flood to provide reliable flood forecasts in large river basins. waves), and the hydrological modeling approaches used (or Water managers require both observed streamflow data in that are planned to be used in the future). Distributed flood a timely manner and forecasts to control water resources forecasting models are particularly sensitive to network den- systems. The complexity of these systems and the targets sities. Mountainous regions susceptible to flash floods re- of operation determine the need for hydrological data. The quire a higher density of gauges within the network, since the interrelationship between water management and hydrolog- lead time of forecasts is short and the differences between ical services is often complex and requires an exchange of neighboring watersheds could be significant. But large rivers data and information. Hydrological data (for example, the respond more slowly and their hydrological data can be more inflow into a reservoir) are the basis for situation-related spatially representative. water management decisions. With this decision (for exam- ple, the release or storage of water from or in the reservoir) The flood modeling system input information requires stream- the hydrological conditions downstream are changed. Here flow data, thus stage-discharge relationships (rating curves) the hydrological forecasts, provided by the hydrological ser- have to be developed and well maintained at the gauges that vice, depend on the knowledge of the water management are intended for model-based flood forecasting. The optimal decisions. A good example is what occurs with a cascade of design of the operational network should provide sufficient hydropower plants (see box 2.5.1). 90    Design of Hydrometric Networks BOX 2.5.1  Demand for Discharge Data in a Complex System of a Cascade of Hydropower Plants The accuracy of flow accounting in the case of large cascades of hydropower plants is ensured if 80–90 percent of inflow in the reservoirs is covered with stream-gauging measurements. In other, less critical cases, as well as in sparsely populated areas, a decrease in the share of runoff-controlled fraction of total inflow by observations can be allowed. In each specific case, this fraction must be determined, taking into account the specific operating conditions of the power system. Take the case of a hydroelectric power station cascade development in a river system, as shown in figure B2.5.1.1. The existing backbone network of 11 stream gauges and 4 level gauges can be considered sufficient for an overall assessment of water re- sources. However, the implementation of the cascade of hydroelectric power plants will require some changes in the structure of the stream-gauging network. These changes include additional streamflow gauges that are established to control the inflow; four gauges fall into the backwater zone and should be either moved or removed; and the flow metering should be transferred to the hydroelectric power station. After that, the stream gauging will cover around 70 percent of the total volume of the lateral inflow into the planned cascade. FIGURE B2.5.1.1  Structural Changes Needed to Implement a Cascade of Hydropower Plants a. Increase in catchment area Catchment area, km2, thousands 27.7 25 20 15 10 5 50 100 150 191 Length of main river, km b. Hydrographic scheme of a river basin with existing and planned cascade of reservoirs Existing gauges (level and stream ow; only level) Planned gauges Gauges to be closed Gauges to be reallocated Hydropower plants Existing reservoirs Planned reservoirs Design of Hydrometric Networks    91 The spatial distribution of hydrometric networks for water The choice of a representative gauge within the designated management depends on the water management issues and areas can be made while taking into account the following the spatial distribution of water resources. For example, if a factors: river is used to supply water for irrigation, a stream-gauging control of the entire flow is necessary. Gauges have to be al- ■ The longest series of observations located not only when rivers leave the mountains (this is a ■ Typical morphometric, water-balance, and physical-geo- prerequisite for forecasts of water availability) but also at the graphic characteristics of a lake and its drainage basin beginning of the fan, which is usually formed by mountain ■ The highest correlation coefficients of the series of obser- rivers entering the flat part of the basin. Going further down- vations of the certain gauge with the series of observations stream along the river, gauges are required in the sections of at another gauge large water intakes not only from surface water but also from ■ The agreement of the observation data at the selected groundwater. gauge with the general conditions of the lakes in the area as a function of the expected average long-term amplitude 2.5.3.3 Observations of Water Levels for Lakes and of water-level fluctuations in terms of physical-geographi- Reservoirs cal and morphometric factors. Lakes and reservoirs are essential components of water re- 2.5.3.4 Stream Gauging in Estuarine River Areas sources systems in many countries. The monitoring of the limnological conditions (water level only) is essential for When examining the river mouth area, it is necessary to have the regulation of lakes and reservoirs. By compensation of at least three reference gauges that are equipped with auto- the temporal fluctuations of runoff, lakes and reservoirs are matic recorders. One of the gauges has to be located on the essential to reduce the impact of hydrological extremes. The river’s mouth nearshore, much more seaward than the mouth basic principle for planning the network of gauges on lakes bar.3 Another gauge should be installed at the mouth of the and reservoirs in modern conditions should be a minimum river, within a well-pronounced sea-level impact. The third sufficiency of the network. The main tasks of the NHS in the should be located near the river boundary of the mouth area, service of the state economy in this regard are the provision where the influence of marine factors is little or completely of current and forecast information on the quantity and qual- absent (fluctuations in the water level due to tides and surg- ity of water resources and the provision of the longest pos- es during low water should not exceed ±5 centimeters). This sible series of observations of the water level at lakes and gauge can measure the water-level according to the program, reservoirs for the assessment of climate trends. A larger num- which accounts for the river water runoff at the upper bound- ber of lakes often occur clustered in certain natural areas. ary of the mouth area. All water-level measuring instruments used at level gauges in the river mouth area must have a per- There are several approaches to planning an observing net- mitted error of no more than ±1 centimeter. work on lakes and reservoirs: ■ A practical approach, based on experience or on the sat- In the presence of a branched system of watercourses of the isfaction of an emerging need in solving specific problems river delta, the support gauges are placed on several large ■ An object-by-object approach, which is used for designing branches—not just one—and at the top of the delta. the network on large lakes and reservoirs ■ Zoning a lake or reservoir’s basin, taking into account the Temporary gauges could be installed for special operation- physical and geographical conditions, as well as the water al tasks—for example, to study the transformation of surge, regime, that makes it possible to transfer the characteris- tidal waves, spreading of flood waves, floods, releases, the tics of the hydrological regime obtained in one or several formation of ice-drifting-jamming level maxima, and flooding lakes or reservoirs to others in the selected area. of the delta. A mouth bar is an element of a deltaic system, which typically refers to mid-channel deposition of the sediment transported by the river channel at the river 3 mouth. 92    Design of Hydrometric Networks The spatial frequency of the gauge network in the river mouth ■ Device environment. Achieving proper operation and lon- area is determined by checking the consistency of water-lev- gevity of devices under different environmental conditions el fluctuations in adjacent locations. If the correlation of the is an important selection factor. Water measurement de- series of the corresponding water levels (such as full and low vices that depend on electronic devices and transducers water, surge peaks, flood peaks) at two adjacent posts of the must have appropriate protective housings for harsh en- estuary network does not meet the requirement of the con- vironments. Improper protection against the site environ- sistency criterion, it may be necessary to equip additional ment can cause equipment failure or loss of accuracy. gauges between the control locations. ■ Adaptability to site conditions. The selection of a mea- surement device must consider the site of the proposed For the practical implementation of the above principles for measurement. Several potential sites may be available for optimizing the stream-gauging network in the estuarine areas a given measurement; the selection of a device depends of large rivers, it is recommended to organize a field office upon the exact site chosen. Branched and meandering riv- (subdivision), which provides constant monitoring of hydro- ers and gauging sites with high water-level fluctuations or meteorological processes and methodological guidance of heavy sediment loads require a specific design of gauges the observation network in these areas. and installed equipment. ■ Vandalism potential. Instrumentation located near public 2.5.3.5 Selection of Stream-Gauging Devices access is a prime target for vandalism. Often valuable and Equipment is often the main concern when it comes to mod- generally usable components (such as solar panels or data ernizing a stream-gauging system. It is relatively easy to pro- loggers) are stolen. Where vandalism is a problem, instru- cure and can be undertaken with clear goals. But to select mentation that can be easily protected is preferable. the appropriate measuring device, numerous considerations must be taken into account: (1) local conditions at the mea- NHS Data Requirements and Financial and Technical suring point; (2) data requirements; (3) the NHS’s technical Capacities capacities; and (4) the country’s infrastructural, institutional, and financial situation. Table 2.5.1 provides an overview of ■ Utilization of data. The requirements for timely hydro- the most common stream-gauging equipment for streamflow logical data depend on the type of the network (basic or and water-level measurements. The main factors that should operational) and the operational use of data. Timely data influence the selection of a measuring device can be catego- are essential for hydrological forecasting and operation of rized as follows: water management facilities—but not for water resources assessments and design floods, which are based on long Local Conditions historical time series. ■ Requirements of operation. The operation of gauges re- ■ Range of flow rates. The smaller the catchment area, the quires the control of their functionality, including changes higher the variability of the runoff. This dependency has of operating resources (batteries), calibration of sensors, impacts on the variability of cross sections and the veloc- and control of the cross-section. Some of these activities ities of discharges. The selected device must be able to can be performed by observers, while others require the measure over the range of hydrological conditions encoun- visit of specialists. tered. Many measurement methods have a limited range of ■ Maintenance requirement. Costly equipment provided by flow conditions for which they are applicable. Large errors international vendors often requires technical capacities in measurement can occur when the flow is outside this and a skilled workforce to fix sensors and components range. Generally, the device should be selected to cover that are not available within an NHS or even in a country. the range desired. Choosing a device (for example, a cur- Spare parts that have to be imported, or the need to send rent meter) that can handle a larger flow rate could result measuring equipment outside the country to be repaired, in the elimination of measurement capability at lower flow increase the risk of temporary or even permanent failures rates, and vice versa. of gauges. Design of Hydrometric Networks    93 ■ ICT capacities. Prior to developing an equipment purchase purchase of automatic reporting stations if there are sig- list, a clear understanding is needed of the institutional nificant gaps in technologies that would preclude them setups, communications capacity, and ICT capacity with- from functioning. in the country. There is little point in recommending the TABLE 2.5.1  Hydrometric Equipment for Streamflow and Water-Level Measurements Vendors Measured Status of required Costsa, and variable Equipment type Advantages Concerns training (US$) modelsb Current meter Low cost, reliable, widely Low speed of mea- Intermediate: 1,500– applicable (e.g., under surements (lower than Training on the 12,000 the ice), well-developed using ADCP profiler by conventional of methods, low cost and an order of magnitude), streamflow wide availability of spare preliminary organization measurement with parts, tools and accesso- of the control section is mechanical current ries easier maintenance required (on small rivers: meter Different girder or suspension models and bridges, remote stream vendors gauging installations hanging cradles; on medium-sized rivers: cable-boat crossings) Acoustic High speed of measure- The high cost of the Advanced: Training 30,000– Rio Grande Doppler current ment production; requires device (price varies on the ADCP oper- 65,000 Stream Pro profiler (ADCP) fewer observers. On small and depends greatly on ations, basics on SonTek and medium-sized rivers, its characteristics, it IT and functioning RiverRay does not require a water- is cheaper for shallow of GPS craft; safe measurements rivers, requires more Water (on mountain rivers); more complex and expensive discharge accurate than traditional maintenance, requires measurements of veloci- highly qualified observer, ties at single points; does not applicable during not require the arrange- periods of ice cover or ment of measuring bridges ice drift) Electromagnetic Sustainable; wide range of Quality of measurement, Intermediate: 3,000– OTT MF pro, velocity sensors measurement velocities; which depends greatly Similar to conven- 5,000 EM Flow applicable under high tur- on the correct position of tional meter train- Meter bidity values; applicable the sensor in relation to ing but sensors are to the shallowest streams the stream different (down to 2 cm depth) Floats Low cost; safety of staff is Poor quality Basic: Tracking Low cost, ensured when measuring of time of floats – approach- ice drift and debris drift surface velocity, ing 0 recalculation into average velocity Volumetric Safe; accurate; possible to Limited applicability Basic: How to pre- Low cost, Valley line method + flow measure on the smallest (only for the smallest pare solute, points approach- Measuring meter streams rivers) where to inject and ing 0 canal measure continued 94    Design of Hydrometric Networks (Table 2.5.1 continued) Vendors Measured Status of required Costsa, and variable Equipment type Advantages Concerns training (US$) modelsb Staff gauge Basis for reading water Often requires struc- Basic: No specific 100–500 Different levels by observers tural measures for training required models and when other sensors fail; installation; dependent vendors required to calibrate other on regular readings by sensors observers Radar Not affected by possible Need to be installed Intermediate:  1,500– Seba negative effects of water; vertically over the water Training required 3,000 hydrome- easy installation surface (at an outrig- on the IT side: trie Radar ger or a crossing of how to program Sensor the river); has limited sensors—e.g., Sebapuls applicability during the change frequency ice period of measurements Hydrostatic Easy and reliable instal- Complex installation Intermediate: 4,000– DST-22 Seba sensors lation; secure sensor; process, water tempera- Training required 6,000 hydrome- Water possible to vandal-proof ture is measured at the on the IT side: trie, 36XW level the design (invisible); depth of the sensor, the how to program Keller AG low power consumption; level gauge is exposed sensors—e.g., relatively low cost of the to water and riverbed change frequency sensor processes of measurements Bubble sensor No pressure sensor or Expensive sensor; instal- Intermediate: 7,000– PS-Light electronics in the water lation is complex; not Training required 10,000 2 Seba vandal-proof; high power on the IT side: hydrometrie consumption how to program sensors—e.g., change frequency of measurements Float-operated Inexpensive; easy to Need to install a calming Basic: No specific 1,500– Surfloat- shaft recorder install; the device is not level gauge well (heated training required 3,000 Sensor exposed to negative water in winter period in mod- 2 Seba effects erate climate) hydrometrie Source: World Bank. Note: ADCP = acoustic Doppler current profiler; GPS = Global Positioning System; IT = information technologies. a. Numbers given are an approximate range; b. The vendor examples provided are not comprehensive or representative for the entire market. 2.5.4 Other Components of Hydrometric for evaporation from water, soil, and snow. Observation data Networks from such networks serve as the basis for the development of national regulatory materials for calculating runoff, elements Depending on national requirements and conditions, other of heat and water balances for small river catchments and parameters besides water levels and discharges have to be marshes, and calculation methods for determining evapo- measured at hydrometric gauges, including: precipitation, ration from different land uses by energy balances and soil water temperature, ice phenomena (if applicable), sediment moisture measurements. loads of runoff, groundwater, and selected parameters of water quality (if required). 2.5.4.1 Operational Precipitation Networks to Support Hydrological Forecasts Hydrological stations for specific purposes include water For flood forecasting purposes, the operation of a meteoro- balance and marsh stations as well as observation stations logical network is vital (operated by either the NMS or the Design of Hydrometric Networks    95 meteorological part of the NHS). It is essential to make sure to select appropriate locations for snow courses because of that informational basis (a database management system) en- complex terrain and significant snow drifts. sures data flow from the NMS, or the NHS, which collects pre- cipitation information to a full extent. The NHS should have The location of snow courses should be sufficient for repre- all the necessary agreements with the agency (or agencies) senting important areas of the river basin where snow dy- responsible for the meteorological data collection process. namics play a relevant role in the overall water balance of a river basin. In the case of mountainous parts of the basin with The location of meteorological stations tends to be in low- snow accumulation, it might be challenging to select appro- lands and in populated places—near the damage centers, for priate locations for snow courses because of complex terrain which flood forecasts and warning are generally required. But and significant snow drifts. for modeling and forecasting, it is crucial to have rainfall in- formation in headwater areas. Thus, when modernizing the Snow courses should be conducted by special teams network, the allocation of rain gauges should reflect the re- equipped with instruments for sampling snow and for deter- quirements of the hydrological forecasting system. If neces- mining snow depth, weight, and water equivalent of snow sary, additional rain gauges have to be installed. cover. A full range of elevations, slope exposures, and vege- tation should be considered in the basin. In terms of vegeta- Requirements for the frequency of observations are driven by tion, traditionally field and forests courses should be covered. the hydrological modeling and forecasting system, which re- According to some sources (such as the WMO), the density quires observed precipitation data for running a rainfall-run- of 1 snow course per 5,000 square kilometers is reasonable off model at a certain time step. Thus, temporal resolution for relatively homogeneous plain regions, while 1 course per of rainfall data varies from sub-hourly in mountainous areas 2,000–3,000 square kilometers is reasonable for less homo- for flash flood forecasting purposes to sub-daily in more plain geneous areas (for example, mountainous ones). basins. The location of these snow courses should account for me- 2.5.4.2 Snow Course Network teorological stations in the area, as meteorological stations can indicate the value of total snow accumulation and can In territories where the regime of a river is affected by snow include snow measurement equipment. Snow redistribution, cover accumulation during a cold period of the year, snow densification and sublimation will impact what remains on measurements are important for hydrological forecasts, es- the ground. During melt, understanding the basin’s snow pecially in terms of long-range lead times. A critical piece of water equivalent available for melt is crucial for both water information is the maximal values of depth, areal extent, and management and flood forecasting. equivalent of snow cover prior to the start of melting (at the dates close to the maximum accumulation). These data are Snow measurements in mountains provide critical in- needed for the provision of medium- to long-range hydro- formation for flood forecasts during snowmelt periods. logical forecasts and water resources management. On the Measurements should be organized either as snow courses ground snow measurements are arranged as snow courses. at the dates close to the dates with the expected maximum A snow course is an established survey line, usually several accumulation of snowpack, or with measurements from air- hundred meters, traversing representative terrain in a region craft of graduated snow stakes located in representative sites of appreciable snow accumulation. Along this course, mea- (their representativeness should be checked by comparison surements of snow depth and density are made to determine with different sites, and the surroundings of the site should its water equivalent. The location of snow courses should be be protected against trespassing). sufficient for representing important areas of the river basin where snow dynamics play a relevant role in the overall water Snow sampling equipment is relatively simple and affordable. balance of a river basin. In the case of mountainous parts of It includes a ruler, snow tube (metal or plastic), a snow cutter, the basin with snow accumulation, it might be challenging a weighting apparatus for determining the weight of the snow 96    Design of Hydrometric Networks cores, and a wire cradle for supporting the tube, along with 2.5.4.4 Evaporation Measurements other tools for operating the snow sampler. The main users of The network of evaporation stations from soil and snow (if ap- snow course information are hydrologists, given that the data plicable) provides information for the study of the evaporation feed into hydrological forecasts. Thus, the organization of snow regime and its determining factors in various natural and ar- surveys often lies with an NHS. Meteorological stations that are tificial affected conditions. Observation stations are typically located in snow-affected areas should be equipped with snow located in agricultural areas. measurement equipment to measure snowfall rates and snow- depth on the ground. This information can also be used for hy- The main principle of design is to maximize observation cov- drological informational and forecasting products. erage of all the main natural zones and typical landscapes of a country. This approach makes it possible to observe the 2.5.4.3 Water Balance Stations regularities of the evaporation process in various natural con- Water balance stations perform complex observations of the ditions of a country, develop methods for its calculation in a main elements of the water balance and factors that cause particular area in the absence or inadequacy of observation their changes in various natural zones. Such observations are data, and perform reliable spatial generalizations that make it carried out on small representative and experimental river possible to characterize the features of the evaporation change catchments. To come up with a real value of observation data over the territory. from the station, observations should be carried out for a long period of time (50–60 years). This allows changes in all com- Observations and calculations on the following elements are ponents of the water balance to be assessed over the observa- typically done at a network of specialized stations: evaporation tion period. from the water surface using an evaporation pan and an evap- orating pool with an area of ​​ 20 square meters, dug into the Typically, there are seven elements to be measured at the ground; temperature of water, soil, and air; precipitation; wind station: the water levels on rivers; sediment discharge; water speed (at a height of 2 meters and at the height of a weather quality; soil moisture; freezing and thawing of the soil (if appli- vane); water vapor pressure at the level of the evaporating sur- cable); ground water levels; and meteorological observations, face and at a height of 2 meters; and air humidity deficit. including actinometrical and thermal balance observations. To measure the total evaporation from the soil and the evap- Underlying the design of a network of water balance stations is oration under the vegetation cover, as well as the evaporation the landscape-hydrological approach, which implies the place- from snow, the station has to be equipped with two weighable ment of stations in the main natural zones and landscapes of evaporators (one for soil and one for snow). Measurements are a country. At the same time, the station’s location should be based on the determination of evaporation from the difference representative of a given natural zone or landscape, so that ob- between two weightings of soil (snow) monoliths on special servations reflect the processes of heat and moisture exchange scales at the beginning and the end of the billing period. for this natural zone. 2.5.4.5 Sediment Measurements Observation objects (for example, catchments, drainage, and Sediment discharge plays a significant role in the design, man- meteorological sites) of a water balance station should re- agement, and operation of reservoirs, as well as in river basin flect the spatial heterogeneity of hydrogeological conditions. management and planning activities. For that reason, measure- Therefore, the best option for the location is the “nesting” de- ments of the sediment transported at the hydrometric gaug- sign, with the choice of a site close to the main watercourse es provide insight into an important characteristic of a river. upstream and sites in the alluvial catchments of its tributaries. Observations of sediment transport and deposit are included in the observing program of the basic hydrometric network. Observation points are typically concentrated in the area of​​ Sediment is measured at the gauges where water discharge is the meteorological site and on small, thoroughly studied river measured, to estimate the total load of sediments for specific catchments. locations of particular interest (for example, estuarine areas, Design of Hydrometric Networks    97 river gauges near the inflow into reservoir, or other engineering 2.5.4.7 Water Quality hydrological objects). Hydrochemical observations on hydrometric gauges network are important for environmental protection. These include In its Guide to Hydrological Practices, the WMO recommends calculations of the removal (transfer) of pollutants with river that sediment discharge should be measured at 15–30 percent runoff; calculations of background concentrations of chemi- of hydrological gauges, which form a minimum network (de- cals in water streams; and calculations of mixing of polluted scribed above) (WMO 2008). water masses in rivers, aimed at making predictive calcula- tions (modeling) changes in the content of pollutants along In general, the requirements of measuring sediment dis- the length of watercourses and others. charge—the network design and monitoring program—should be governed by the needs of a specific issue of water resources The design of water quality stations should consider the fol- management and depend on the quantity of sediments and its lowing locations: temporal variability. For example, measurements may be made only in the flood season at some of the basic network gauges. ■ Those with the most stressed ecological state ■ Those having an important fishery and nature conserva- 2.5.4.6 Groundwater tion value In many countries, groundwater is one of the most valuable nat- ■ Those located near the state border and the borders of the ural resources, yet it is often overexploited, and the conditions of constituent entities of a country groundwater recharge are changing with the climate. This means ■ Those where the population, the number of wastewater that it will be increasingly important to take a more holistic view sources, and the level of pollution have changed signifi- of interactions between surface and groundwater—and that will cantly over the past years. require NHSs to cooperate closely with agencies responsible for groundwater monitoring. Effective management of groundwater The optimization of observation programs focuses on the resources depends on groundwater monitoring and data acquisi- widespread use of mobile hydrochemical laboratories, which tion. Such management is often a task performed by authorities can solve a variety of water quality network issues. These with specific regulatory functions, with groundwater observa- include remoteness of water bodies from laboratories; the tions mostly done by geological services. Only in a few countries need to send water samples, which often leads to a violation is this currently a task of an NHS. of the shelf life according to the requirements of regulatory documents; and the impossibility of determining a number Water-level measurements from observation wells provide of indicators at the site of water sampling by specialists who the most fundamental indicator of the quality and quantity of have the necessary qualifications. this resource and are critical to meaningful evaluations of the quantity and quality of groundwater and its interaction with 2.5.5 From Design to Monitoring surface water. But given that this information is available only at known points of data, groundwater hydrology is an inter- In sum, hydrometry is a core business of an NHS. From the enu- pretive science—requiring interpolation and extrapolation to meration of the different types of hydrological monitoring net- provide a more holistic view on the status of aquifers. Besides works, it is clear that an NHS may be responsible for monitoring capturing dynamic data that changes over time (groundwater a variety of components of the hydrological cycle—including level and quality monitoring), a groundwater network depends precipitation (preferably in cooperation with the country’s NMS), on static data that show no significant variation with time (for groundwater properties, water quality, and flow characteristics example, land use inventories, geological maps, results of well of surface waters. Hydrometric activities are determined by the and aquifer pumping tests). particular hydrological conditions, user requirements, and insti- tutional structure of each country. The stream-gauging network is the basic component of any hydrological monitoring system, a topic considered in the next chapter. 98    2.6 Modernization of Stream- Gauging Networks Status of the Hydrometric Network 2.6.1 Any modernization of a stream-gauging network should start with an assess- ment of the status of the existing network, including an inventory of existing stations. Critical elements to consider are the length of observation period and the number of discharge measurements, along with rating curves developed within the last 10 years, and the current and near-term future use of the data. This should be followed by an assessment of the need to modernize these stations or install additional gauges and can be completed by asking: Ethiopia. Photo courtesy of V. Tsirkunov, WBG. ■ Which hydrological variables need to be observed? ■ Are they required for better hydrological knowledge (a basic network)? ■ Where do they need to be observed? ■ How often do they need to be observed? The answers to these questions are essential for estimating the scale of the modernization during the planning phase. A key set of considerations is ask- ing “Where are we now?” “Where do we want to be?” “How long do we have to achieve our goal?” With sufficient time and money, it is possible to create “Modernization campaigns a stream-gauging network that is geographically complete and highly precise, but it comes at a cost in terms of both capital and operational capacities that with extensive technical may simply be unrealistic. installations in a short time A more realistic approach, therefore, is to seek incremental improvements can overwhelm an NHS in the and find the best benefit-cost ratio—whether that is through organizational changes, equipment purchases, or both. Sustainability comes from incremen- planning phase and also in tal improvements to existing systems. maintaining the operation Modernization campaigns with extensive technical installations in a short after the installation. It is time can overwhelm a National Hydrological Service (NHS) in the planning imperative that short-term phase and in maintaining its operation after the installation. It is impera- tive that short-term targeted improvements and the development of a mid- to targeted improvements long-term strategy be weighed as the path to sustainability. Not only are there issues in terms of existing workload but there are often significant regulatory and the development of a and human resource constraints as well. mid- to long-term strategy be weighed as the path to sustainability.” Modernization of Stream-Gauging Networks    99 2.6.1.1 Categorization of Levels of Developments stream-gauging monitoring equipment that relays information The following classification system of five categories of de- to a hydrometric/hydrological team. There are some effective velopment can be used to categorize the development of a gauges and ratings, but there is still significant improvement country’s hydrometric service (table 2.6.1): potential. The responsible institutions are building toward improving their risk management and are undertaking flood Level 0: Undeveloped. The recipient country currently un- forecasting with a view to producing flood warnings. dertakes no stream-gauging data collection or hydrological analysis but relies solely on historical evidence and experi- Level 3: Developed. The country operates a largely automat- ence. This lack of development requires the completely new ed network of monitors and has some good analysis and un- installation of an observation network and the creation of the derstanding of the network, its benefits, and its limitations. basic stream-gauging structure of the NHS. The country is looking to increase the utility of the system for flood forecasting and warning as part of a hazard risk reduc- Level 1: Development initiated. The country undertakes basic tion program. stream-gauging functions and has a sparse network of gaug- es from which readings are collected manually and recorded Level 4: Advanced. The country has in place a fully distrib- in hard copy and transcribed to digital format. Rudimentary uted automatic gauge reporting system and undertakes de- ratings curves and analysis are undertaken. Flood warnings tailed ratings assessments and hydrological analysis. This may come largely from experiential assessments and reviews might be considered a “finished” system, although contin- of historical events. uous monitoring and improvement is in effect. The country operates a hydrometeorological forecasting system capable Level 2: Development in progress. The country undertakes of parameterizing risk and uncertainty. a mix of Level 1 activities, but also has some automated TABLE 2.6.1  Levels of Stream-Gauging Networks Level 1: Level 2: Level 0: Development Development in Level 3: Level 4: Criteria Undeveloped initiated progress Developed Advanced Status of No hydrometric data Sparse network of Level 1 activities Largely automated Fully distributed stream-gauging collection gauges; readings are plus some network of moni- automatic gauge network collected manually, automated tors and has some reporting system; recorded in hard monitoring good analysis and undertakes detailed copy, and tran- equipment relaying understanding of the ratings assessments scribed to digital information to a network, its benefits, and hydrological format hydrological office and its limitations analysis Discharge Absent Insufficient number Ensured for most Ensured for nearly As Level 3 measurements relevant gauges all gauges Rating curves Absent Often outdated; Sufficient for Ensured according As Level 3 available only for a selected gauges; to the variability of few selected gauges not available for the the cross-section for majority of gauges nearly all gauges Data transfer Absent In paper format, Automatic for some Automatic for nearly As Level 3 transcribed to digital stations; manual for all stations format manually others with GSM or GPRS Source: World Bank. Note: GSM = Global System for Mobile Communications; GPRS = General Packet Radio Service. 100    Modernization of Stream-Gauging Networks 2.6.1.2 Planning the Transition between Levels Such a planning exercise must account for funding, econom- The modernization of networks is usually carried out in either ic viability, sustainability, and the complexity of the network a single project or a sequence of projects that are designed design requirements. This means clearly identifying the mod- for improving basic hydrological systems. Although the goal ernization limitations and ensuring that effective strategies for all countries is to reach Level 4 as quickly as possible, it is are in place to move from one level to another. The investment not always possible to develop an entire network to this level in resources must be targeted with an eye for what works now within a single project life cycle funded by the government or and what will work in the future. Immediate needs should be development partner. Thus, project planning must consider met first to develop essential components of the system. At appropriate steps that can be taken in the short term within the same time, guided investment should seek to address the the project, while maintaining medium- and long-term aspira- long-term impediments to further modernization. Such an tions for Level 4 completion (table 2.6.2). approach requires multiple timescales but will lead to signifi- cantly more successful and sustainable development. TABLE 2.6.2  Transitions between Levels of Development Existing level Desired level Project focus Level 0 Level 1 A full hydrological review of the country will be required from the ground up to understand catchments, risk areas, river types, and potential issues. The project setup will require higher levels of investment in the planning stage, but this would bring about a significant return on investment by marginalizing mistakes in the setup. The project should focus on in-country staff and training—undertaking a capacity and needs assessment and developing a full plan of action to develop the NHS. An equipment plan should also be produced and reviewed. Level 1 Level 2 A stream-gauging review of the country and the existing gauging setup should be undertaken. The aim would be to understand the good, bad, and missing parts of the system. The focus should be on analyzing the existing system first, before expanding the network. The priority should be a comprehensive understanding of limitations in equipment, geographic coverage, and staffing. Level 2 Level 3 A project might focus on a gap analysis for equipment, training, and resources, and work with experts to identify areas where the NHS system might improve. The improvements will likely be incremental. Good questions for NHS staff would be: What is going well? What could be improved? What would you like to be able to do but cannot? From answers to these questions, a project plan that prioritizes the country’s needs can be implemented. Training and simulation of real-world events may help expose weaknesses in forecasting and monitoring setups, which may also lead to an improvement. Source: World Bank. Note: The transition from Level 3 to 4 is not discussed here as it requires medium- to long-term activities, which are out of scope of the modernization projects that are discussed here. The realization of planned transitions to a higher level re- (increasingly) data scientists, the system will be neither ef- quires qualified personnel (table 2.6.3). Staffing and per- fective nor sustainable. To ensure sustainability of approach, sonnel resourcing is often a pinch point in developing more there must be a clear plan to develop over two timeframes: a effective hydrometric networks through their modernizations. short-term plan that sees staff being required immediately to It is clear that, even with the best technology, without effec- fulfil a need, and a long-term plan that sees a need for a well- tive trained and professional observers, hydrologists, and staffed hydrometrics team capable of managing the system. Modernization of Stream-Gauging Networks    101 TABLE 2.6.3  Personnel Requirements of Transitions to Higher Levels Existing level Desired level Personnel improvement Level 0 Level 1 Develop staff understanding and training in hydrometrics and hydrology. The focus should be on the value of data collection, principles of hydrometric techniques, and core concepts related to understand- ing the catchments. Staff will likely require ongoing international support as part of the project, and there may be a prolonged period of engagement and systems monitoring to ensure that progress is being made. Level 1 Level 2 Substantial increase of the training of staff is required, along with the need to improve applications of hydrometric concepts, rating reviews, and hydrometric best practices. Staff should be trained in the latest methods, perhaps in conjunction with local or international universities and partners. Staff should begin training to: (1) develop the ability to run their system independently; (2) plan changes, updates, and maintenance of their system; and (3) become responsible for forecasts and flood response. Moving from Level 1 to Level 2 may benefit from international assistance and periodic reviews of systems and approaches. Level 2 Level 3 At this point, the country is unlikely to still be a recipient of funding from development partners. However, the transition could be managed by the reviews (from Level 2) being tied to development goals that are set by the country and compared with international best practice. The focus should be on implementing staff development and improvement schemes that previously may have been ad hoc. Staff should now be highly proficient, but they may still benefit from highly specialized training and development. Source: World Bank. Note: The transition from level 3 to 4 is not discussed here as it requires medium- to long-term activities, which are out of scope of the modernization projects that are discussed here. Leveling up with equipment is largely a function of cost—both restrictions. Equipment spending should be targeted and fo- in the initial outlay and in ongoing maintenance and repair. cused on how to make the greatest improvements over short The selection of equipment also relies on the infrastructure timeframes, while investing in the long-term development of within the country such as telecommunications, informa- the system (table 2.6.4). tion and communication technology (ICT), and geographical TABLE 2.6.4  Required Change of Equipment Existing level Desired level Change in equipment Level 0 Level 1 With little or no equipment, the largest benefits will come from collecting data by whatever method is appropriate, affordable, and achievable. For example, the application of board gauges and a system of field assistants taking daily readings might be appropriate, but consideration could also be given to floats or radar if funding and infrastructure are available. There should also be more frequent readings in critical areas, or use of automatic reporting gauges (e.g., current meters or acoustic Doppler current profilers) in identified critical areas (e.g., upstream from the main population centers with regard to flood risk management). Digitization of data should be undertaken if computerized assets are in place and affordable. Level 1 Level 2 The focus should be on greater spatial resolution of high-quality gauging information, wider application of automatic reporting stations, and use of hydrological field assistants to provide widespread check readings and help monitor equipment. To improve flood forecasting for the most populated areas, the network of automatic stations should be further extended upstream of cities. Level 2 Level 3 At this point, automated reporting stations are standard across much of the country, and high-quality reporting instrumentation is being checked, calibrated, and installed as appropriate throughout the country. Countries are also considering developing a hydrometric network in some representative catchments at the research level and developing their monitoring networks accordingly. Source: World Bank. Note: The transition from Level 3 to 4 is not discussed here as it requires medium- to long-term activities, which are out of scope of the modernization projects that are discussed here. 102    Modernization of Stream-Gauging Networks 2.6.2 Prioritization of River Basins for operated in a safe manual mode. Table 2.6.5 shows the evolu- Modernization and Technical Re- tion of stream gauges, starting with staff gauges and manual Equipment recording of water levels and ending with automatic stations with combined redundant sensors, remote data access, and An update of the levels of development of stream-gauging complex data transfer systems. Any gradual updating of networks can often be done only gradually, owing to limit- stream-gauging networks requires a prioritization of gauges ed resources. In such cases, only the most relevant gauges based on their relevance for socioeconomic conditions and should be automated, and remaining stations should be the NHS’s ability to operate them in a sustainable way. TABLE 2.6.5  Different Development Stages of Water-Level Data Collection Data capture Data collection Sensing Recording Frequency Transmission Digitization Staff gauge Manual reading, recorded Once or twice per day Manual: Field observers, postal Manual in hard copy service, telephone Float-operated mechani- Continuous with a writ- Typical temporal resolu- Postal service (weekly, monthly) Manual cal recorders ing device on a registra- tion: hourly tion drum on paper Float-operated shaft Every 5 minutes with an Typical temporal resolu- GPSR, GSM, satellite Automatic encoder electronic recording tion: 15 minutes Multiple combined water- Every 5 minutes with an Every 5 minutes is pos- GPSR, GSM, satellite Automatic level sensors: radar; hy- electronic recording sible. Typical temporal drostatic, bubble sensors, resolution: 15 minutes webcams Source: World Bank. Note: GPRS = General Packet Radio Service; GSM = Global System for Mobile Communications. Any prioritization of gauges that have to be newly installed or ■ Operation. Here the need to get water-level data in modernization of existing sites requires the consideration of near-real time must be weighed. The reality is that such multiple objectives and restrictions (see annex 2.2 for an ex- rapid information is not required in every case. For water ample of prioritization of gauges at the scale of a river basin). resource assessments of large rivers with slow temporal These criteria can be subdivided into three groups of criteria variability of water levels, one observation per day might and several sub-criteria (figure 2.6.1): be sufficient. In contrast, fast-responding watersheds in mountainous regions need at least hourly resolution of ■ Strategic importance. Here the issue is the importance data during flood events to ensure the reliability of flood of flow data from a particular site for the requirements of warnings and forecasts. The need to receive data in an op- water management in the present and forecasted in the fu- erational mode depends on the specific conditions of the ture (as far into the future as can be estimated). The ongo- watersheds, the need for early warnings, and the require- ing changes of demand and supply have to be considered ments of operating water management systems. as well as the relevance of the data for disaster risk man- ■ Costs. Here the recurrent costs for the operation, mainte- agement. In addition, there may be different hydrological nance, and replacement of each station must be consid- priorities for gauges in the basic network in different re- ered together with the capital costs of the investments for gions of a country, determined by the characteristics of the the station and equipment. The investment and running available water resources. Moreover, the spatial require- costs of a gauge depend on several factors (such as loca- ments for hydrological information change with time and tion, hydrological conditions, climatic conditions, and ex- so must be kept up to date. isting infrastructure; see chapter 2.7). There is a tendency Modernization of Stream-Gauging Networks    103 FIGURE 2.6.1  Criteria to Determine Strategic Importance, Costs, and Operation of Hydrometric Gauges Agriculture Hydropower Water management issues Operational Industrial supply Water supply Public supply Flood forecasting Strategic importance Representativity for a natural space Hydrological basic network Potential relevance for future water management issues Climate change Evaluation of a gauge Capital costs Costs Recurrent costs Accessibility Ease of maintenance Availability of appropriate staff at required levels Operation Length of series Water level Quality of record Potential quaility of the flow record for stations that are the most expensive to be closed to is important to consider the level of economic development of reduce the total costs of the network. It must be consid- the regions. This needs to consider the risks and needs for hy- ered that data from such stations—located, for example, in drological informational products to prevent negative water mountainous areas with insufficient infrastructure—could impacts on the population and on the national economy, of be more relevant than data from “inexpensive” stations. In which floods are considered the most dangerous hydrological such cases, multiple objectives must be balanced. phenomena. Flood risk maps can be used to estimate poten- tial damage from basin to basin, or, in their absence, histori- The decisive factor when choosing priority basins for modern- cal information on past floods and their negative impacts can ization and technical upgrade should be their socioeconomic be used. With industry and agriculture, which have intensi- benefits—that is, avoided damages and the number of people fied their water use in the last decades, the socioeconomic affected by the negative impacts of water, or the expected need importance of droughts has risen in many parts of the world. to develop water resources in the future. On the other hand, Limitations of water availability require water supply regula- the ability of the service to operate the stream gauges in a sus- tions for nonprioritized users and other integrated water re- tainable way also must be assessed. Thus, the design of a mod- sources management (IWRM) measures (section 2.2.4)—and ernized network requires a cyclic process to reach a balance a control of these measures by stream gauging. between the need for data, the way these data are provided, and the NHS’s capacity to operate the modernized network. The main criteria for determining the sequence and priority of river basins for technical refurbishment and moderniza- When choosing the criteria for determining the priority of tion of the hydrometric network uses the following charac- river basins for technical re-equipment and modernization, it teristics, which reflect the degree of economic use of water 104    Modernization of Stream-Gauging Networks resources, the population density of the river basin, and the convert water-level information into flow, which is critical for number of users of hydrological information: operational hydrology. Consistent time series analysis is re- quired for real-time plausibility checks, error detection, data ■ Population density in the territory of the river basin cleaning, data flagging, and automatic bias corrections. ■ The number of water facilities in the territory of the river basin The software requirements can be subdivided into two cate- ■ The number of users of hydrological information gories: (1) software to ensure the data flow from sensors to ■ The number of tasks to be solved because of improving the end users of data and (2) software to support the processing quality and efficiency of providing users with hydrologi- of data and hydrological analyses. cal information and preventing the risks of negative water impacts 2.6.3.1 Software for Data Flow ■ Information on catastrophic floods, damage, and number The software to manage the flow of data is a key element of of victims in past periods. the operation of automated networks. A hydrometric observ- ing system includes three interconnected components: data 2.6.3 Software Requirements and Procurement acquisition, data and metadata management, and provision of hydrological data and their aggregation into hydrological Any modernization of a hydrometric network will result in information (that is, flood or low-flow statistics and hydrolog- more digital data, as computation systems are an integral ical status reports). A modern hydrological software system part of any hydrometric system and an NHS. Water-level in- provides all stages of receiving, collecting, transmitting, and formation needs to be converted to flow information in real processing hydrological information, along with preparing time, while being quality assured for archiving as a contin- the relevant information products and issuing them to users. uous time series. These data make up part of an important It ensures the fulfillment of NHS obligations to monitor water- real-time and historical record and are subject to rigorous courses and surface water bodies in a country and to provide quality assurance for final review or approval. Converting the public good hydrological data and information. An exam- water level to flow requires constant measurement in the field ple of the data flow from sensors to end users is shown in and constant upgrading of the rating curve. It must also in- figure 2.6.2. clude a software platform capable of handling large volumes of data in a coherent and user-friendly environment. Building These tasks require specialized software solutions to orga- rating curves can be a complex task; a rating for a specific nize a seamless flow of hydrological data and information river reach contains measurements, shifts, blends, and peri- from the observation system through a data management ods of applicability that describe a channel as it changes over system to workstations of hydrologists and as condensed in- time. In addition, data ingestion in real time is required to formation to end users and to the public (box 2.6.1). Modernization of Stream-Gauging Networks    105 FIGURE 2.6.2  Data Flow from Gauges to Data Collection, Processing, Storing, and Final Product Preparation NHS users’ workstation Monitoring, Data processing and Forecasting, storing Regime products Visualization, data and product access Data QA/QC, processing, via web application DBMS, Operational and regime data and products, Metadata management, Map server, User interface Interaction with observations acquisition End users, decision-makers system Data Software collections and data Control center(s) scheme commands updates Observations Observations in in XML XML or text files Photo Manual measurements: observer, streamflow measurements Automatic measurements from water-level sensors, photo cameras, other Note: DBMS = database management system; QA = quality assurance; QC = quality control; XML = Extensible Markup Language. 2.6.3.2 Software for Hydrological Analyses coefficients and adapt calculation methods to different peri- ods within a year. A second is that it must allow manual cor- Besides the general data validation, analysis, and dissemina- rection of the rating curve, if necessary. A third is that it must tion processes, an NHS requires specific software tools for a compare measurements of water levels, using standard and range of activities: automated measuring instruments. Transfer of water-level data into streamflow data. Operational databases for model-based hydrological fore- Discharges are derived from water-level data by using a rat- casts. Hydrological forecasting models often require inputs ing curve based on discharge measurements for different of observed hydrometeorological data in near-real time from stages. There are several methods and tools available to fit different sources. For this purpose, a specific database man- rating curves with mathematical functions. To apply them, a agement system (DBMS) has to be developed to handle hy- suitable software solution is required. drometric data, synoptic data (at least air temperature and precipitation data), forecasts from numerical weather pre- Automated calculation of streamflow from water-level data. diction (NWP) models, and snow water equivalent and snow Besides the transfer of water-level data into discharges with cover data from snow courses or satellite-based systems in the rating curve, several functionalities of the software are tight cooperation with the NMS. Observed data and hydrolog- essential. One is that it must consider seasonal variable ical forecasts must be stored and processed for further visu- aquatic vegetation or icy conditions in winter with correction alization and dissemination. 106    Modernization of Stream-Gauging Networks Reporting and visualization. State-of-the-art reporting and BOX 2.6.1  Software Required for Efficient Data Flow visualization of data and information implies geographical in- formation system (GIS) and web-based technologies. A number The software required to ensure the data flow consists of of web-based GIS software tools are available to link with a several components, each of which solves a specific task: DBMS. Web-based applications provide users with a graphical interface to give them the opportunity to visualize and analyze Data collection center hydrometeorological information on demand and to provide ■ Data collection from stream gauging (both automatic data access in different forms. But developing a web-based GIS and manual) application with open access to operational data and products ■ Control of mode of work of automatic sensors, requires expert knowledge. Such specialists are rarely avail- monitoring of data sources, generation of reports on the functioning, and signalization in case of able in an NHS. Only in developed NHSs with an existing infor- malfunctioning mation technology (IT) department can the essential support ■ Interaction with data collection platforms from other be ensured. In other cases, an outsourcing of the development organizations (for example, collection of operational and updating should be given serious consideration, including data from reservoirs, weather data), using the cloud-based services. Since access to dissemination services protocols and formats of these systems. such as web or other standard dissemination systems highest Data processing and storage during critical hydrological situations, support for such appli- cations must be robust and assured. ■ Quality assurance and quality control of incoming data Setup and maintenance of long-term archives for large net- ■ Automatic computation of derived data products (discharges) and validations with neighboring works. The setup of a national hydrometric database is the pre- gauges condition for an analysis of long data series to provide reliable ■ Storage of controlled data in a national hydrometric design floods, as well as for information about climate-induced database. changes. For these purposes, it is essential to digitize histori- cal paper-based records and update the database with current Metadata management quality-assured observations. This database should also con- ■ Administration of the characteristics (metadata) tain long-term storage of observational information that are of stream-gauging stations—such as technical obtained and derived from various sources. This could include equipment, date of the last equipment calibration/ field information, details around high-water event or notes check, list of observations and discharge measurements, frequency of observations and data around equipment malfunction or changes. transmission, description of the location of the gauge, maps, photographs, and hydraulic profile. Software for statistical analyses. The required hydrological information for design purposes is often derived from statis- Data synthesis and analysis tical analyses of long time series. For this task, special soft- ■ Automatic reporting system for hydrological situation ware packages that may be procured although open-source monitoring systems that perform statistical analysis on hydrometric re- ■ Interactive analysis of data and synthesis, cords are increasingly becoming available. data aggregation into information, reports by hydrologists. Software for hydrological modeling. Hydrological models Dissemination are indispensable for deterministic forecasts in hydrology, and existing models can also be applied for planning purpos- ■ Provision of user-specific information es, especially to provide impact assessments. Many different ■ Public web portal for hydrological data and reports (manual enquiry system). types of models exist with different costs, functionalities, de- grees of complexity, and demand for data. A detailed analysis of the requirements and opportunities to operate such mod- els is essential before models are introduced. Modernization of Stream-Gauging Networks    107 2.6.3.3 Options to Provide Necessary Software Within the framework of projects to modernize hydromet- In-house development by the NHS. Software development of ric networks, the use of sophisticated hydrological software the informational system is challenging, as several protocols packages—which ensure data flow, data processing, and data need to be combined. It demands highly qualified IT special- analysis—is often favored because individual programming ists (in software development, database administration, and with a similar functionality would be very expensive (more web-based GIS applications) to implement the various soft- than $1 million). In addition, an enormous amount of prepa- ware components into a single data and products chain (see ratory and planning activities is needed to develop technical, figure 2.6.2). The functionality of the result requires the su- IT, and detailed software concepts; together these account pervision of hydrological specialists. The IT specialists can be for at least one-third of the software package creation costs. either part of the internal IT department of the NHS or from Software procurement is simply more cost-effective if there is specialized software companies. an off-the-shelf package available that can be used under the existing IT infrastructure. That is why, in recent years, most Application of several single software components (free- modern NHSs have migrated to commercial off-the-shelf soft- ware or commercial software). Freeware requires signifi- ware packages. These packages provide complete solutions cant IT efforts to be robustly implemented into a system, and with specific hydrological tools (such as organizing the utili- often support from the software developers to interlink it is zation of updated rating curves). indispensable. For the software components listed above, both freeware and commercial software tools are available. Even so, individual adaptation is always required, at least to Typically, freeware solutions demand more effort from IT spe- the existing IT infrastructure. Therefore, it is often not suf- cialists to be implemented, but are more flexible. Continuous ficient just to buy software. The cost of software purchase technical support is required on every step of the data and and customization together are in the range of $500,000 ± products flow as data collection, database management, hy- $200.000, depending on the circumstances and individual drological calculations and modeling, and dissemination of requirements. Due to the rapid developments that occur at GIS-web products. Such support must be included in the fu- the operating system level, updates are required at relatively ture NHS budget to secure sustainability of the hydrological short time intervals. A complex version change of such a soft- information system. ware system could cost an additional $100,000 ± $25,000 (for example, the user could request this if the ICT platform Comprehensive hydrological software platforms. These changed) depending on the number of tools and licenses the packages are available to cover the total range of software software package contains. Thus, the design and specification requirements of a hydrometric network in the form of a hy- for hardware and operating systems must be considered in drological computational production system. They are an in- tandem with the commercial software platform—particularly tegral part of any hydrometric system and NHS, given that for data storage, data backup, and network configurations. the package considers the specific demand for hydrological computations. Water-level information needs to be convert- Moreover, there are often challenges associated with main- ed to flow information in real time. If the rating curve is to taining ongoing coherence between operating system, hard- be adapted because of new discharge measurements, these ware, and commercial software solutions, underscoring the discharge data have to be updated to reflect the changing rat- need for strong and consistent IT support. To ensure fail-safe ing curve. This process must be detailed and quality assured operation, a service contract for maintenance and support for archiving as a continuous time series. These data make contract is necessary. For off-the-shelf software, these costs up part of an important real-time and historical record and are in the range of $20,000 per year, but they can increase are subject to rigorous quality assurance for final review or immeasurably (> $100,000 per year) for individual develop- approval. To achieve this in an effective way, an NHS requires ments. Leasing models that include services are also avail- hydrometric production software and data ingestion capabil- able, with a price tag in the range of $200,000 per year. ities that require standard telecommunication platforms and server-based systems. 108    Modernization of Stream-Gauging Networks Besides this more traditional approach—which requires hard- centers in a more brick-and-mortar approach. But this is mis- ware, infrastructure, computer operating system installa- guided for a number of reasons: tions, and a local data center—cloud-based technologies are now available to ensure stability, expandability, and rigor to ■ There are many hidden costs in domestic data center ap- maintain these production systems. These new technologies proaches that are rarely considered—such as the hidden offer several advantages: costs of maintaining on-site software, which include server costs, the costs of operating system software, and data- ■ There is no requirement for IT infrastructure, servers, serv- base software costs and upgrade costs. er maintenance and upgrades, or the implementation of ■ Labor for maintaining IT hardware and software infrastruc- software or operating system upgrades. ture can be expensive. IT infrastructure requires constant ■ These systems are often International Organization for attention, including applying patches, updates, and up- Standardization (ISO) compliant. grades to the operating systems and application software. ■ Software updates are automated, as are updates to bug- Moreover, as application software and production systems fixes and compatibility with operating system changes. are released, they need to be uploaded in a timely manner. ■ Cloud-based solutions often incorporate a full back-up and And if server operating systems and applications software disaster recovery plan. are not continuously upgraded by local IT personnel, there ■ Initial capital infrastructure costs for computer hardware could be complex and difficult data migration challenges. are minimal, and upgrades to hardware are not required. ■ Internal IT systems are increasingly at risk of cyberattacks, This results in an overall lower total cost of ownership. including ransomware attacks. Some examples have highlighted a 25 percent cost savings over a three-year period. At this juncture, the best way forward may be to focus on ■ No big upgrade and data migrations are required, as the cloud-based data production and archival systems. They hardware/software solution is continuously updated by IT are easy to implement, require no upfront hardware cap- professionals. ital expenses, and minimize IT labor needs and hardware ■ These systems are accessible from anywhere with an inter- costs. Moreover, these systems are highly secure, reliable, net connection. expandable, and accessible. At this time, there are a num- ■ These systems are often scalable, limiting the requirement ber of choices available for cloud-based production system for resizing the IT hardware requirements. platforms (such as Aquatic Informatics® in Canada or Kisters ■ According to Alain Pietroniro, former Executive Director of Group in Germany). NHS Canada, global experience for these cloud-based sys- tems have shown a 99.9 percent reliability on uptime—a 2.6.4 Ensuring Data Quality score that is simply unachievable in domestic systems. When switching from manual measurements at stream-gaug- In recent years, Canada and the US Geological Survey have ing stations (for example, traditional measurements at 8.00 adopted cloud-based solutions in the hopes of better focus- and 20.00 local time) to measurements with automatic in- ing on the engineering and science challenges associated struments, it is necessary to organize parallel observations with an NHS, as opposed to the time-consuming task of main- with traditional (manual) and new measuring instruments taining software and keeping IT systems operational. Another for two to three years to develop methods for extending the more recent implementation is used in New South Wales, observation series and preserving their homogeneity when Australia, for telemetry, groundwater, and surface water data switching to new measuring instruments. management. The main tasks to be solved because of parallel observations However, development projects historically, and even recent- are the observance of the principle of homogeneity and com- ly, have not embraced cloud-based solutions, largely because parability of observation data and the determination of the there is a sense of ownership and pride associated with data optimal frequency of control measurements of hydrological Modernization of Stream-Gauging Networks    109 characteristics after a complete transition to an automated 2.6.5 Boundary Conditions for Success method of their measurement. Modernization usually means the transition from analog to For transitioning to the use of automated measuring instru- digital data acquisition. This process brings a number of ments, it is recommended to: advantages: ■ Adjust water-level observation programs—transition from ■ Higher temporal resolutions of the data become possible. manual observations to continuous ones with the determi- ■ Quasi-continuous data flows from gauges to the hydrolog- nation of the optimal interval between measurements and ical data storage and management system can be ensured; the optimal interval of averaging instantaneous measure- this reduces errors in data transmission. ments performed by the sensor. ■ Automated data quality control can be performed. ■ New possibilities for the operational use of hydrological ■ Change the terms—the form of transmission, the frequen- data are provided; users and the public can be better in- cy, or any specific method—of data transmission to data formed about the current hydrological situation. collection centers. ■ Assign threshold values of the water level at which a However, these advantages can be realized only if the bound- change of measurements and transmission frequency is ary conditions discussed in the following chapter can be required. provided. In addition to hardware and software, a digitized ■ Develop and optimize visual observation programs information chain requires highly qualified and specialized and data transfer procedure using an automated photo personnel and an institutional framework that ensures the recorder. system’s long-term operation. Since the rate of software re- ■ Development programs for parallel observations of water newal in particular is high, this requires a detailed needs and levels with new instruments and using standard equipment. cost assessment. 110    2.7 How to Estimate Total Costs of Operating a Stream-Gauging Network 2.7.1 Methodology of Cost Estimation The total cost of ownership (TCO) looks at the total cost of a product or ser- vice from the time of purchase or the starting point of the service. Here the product is hydrological data and information—the service that the National Hydrological Service (NHS) provides. Its total costs result from purchase costs of equipment; infrastructure costs (for example, costs of constructions); operating costs of stations, information and communication technology (ICT), data storage and archiving, and labor costs; costs of maintenance and re- placement; and business operating and administrative costs. A hydrological Taking samples of groundwater. Photo: BartCo observation network should be “fit for budget.” In order to make appropriate trade-offs in network requirements, taking into account available funding, a comprehensive understanding of the TCO of the hydrological monitoring in- frastructure over its lifetime is essential for assessing its affordability. The TCO concept, as illustrated in figure 2.7.1, has four main cost categories: (1) initial investment (capital costs), (2) annual operating costs, (3) mainte- nance costs, and (4) replacement. In each category, there are several subcat- egories, where the costs can be specified in detail. It is essential to estimate and to compare the costs by categories to find ways to reduce these costs and “In order to make to assess the affordability of the modernized network. appropriate trade-offs in network requirements, taking into account available funding, a comprehensive understanding of the TCO of the hydrological monitoring infrastructure over its lifetime is essential for assessing its affordability.” How to Estimate Total Costs of Operating a Stream-Gauging Network    111 FIGURE 2.7.1  Overview of the Total Costs of Ownership Considering Replacement as Part of Life-Cycle Management Total cost of ownership Initial investment - capital Annual operating costs Maintenance costs costs Replacement (life- cycle management) Capital cost of equipment Labor cost Labor cost Direct operating costs Preventative maintenance Field equipment Cost of all field quipment Stationary equipment at Field operating costs gauge site(s) Vehicles Stationary equipment at Information processing guage site(s) Mobile field equipment systems Mobile field equipment Discharge measurements IPS computers and software Cost of information processing Leveling Corrective maintenance systems (IPS) Other operating costs Cost of civil works Field equipment Land lease Supporting structures Labor cost Costs of utilities Adaptive maintenance Cost to procure land Consumables Stationary equipment at Cost of site preparation Indirect operating costs guage site(s) Installation of utilities Mobile field equipment Construction of major supporting structures Business costs IPS computers and software Cost of security Costs of information processing Agency costs to install operational systems systems Labor cost Operating of web-based information systems Project management Software licenses Cost of logistics Service level agreements Calibrations of measurement instruments Office rent, furniture, travel costs, vehicles Administrative costs 2.7.1.1 Initial Investments: Capital Costs Capital Cost of Equipment Capital costs are one-time expenses incurred for the pur- Total cost of field equipment. This category is the best- chase of land, construction, and equipment used to bring a known part of the cost estimations. It covers not only field stream-gauging network to an operable status and in the ren- equipment at the site of the gauge (such as a new water-level dering of hydrological services. The modernization of hydro- sensor system) but also field equipment required to operate metric networks is often carried out as part of single projects, several gauges. A shelter, footbridge, or cable-crane or cable- where these initial investment costs are considered to be the way are costs of a specific station, whereas, for example, a main cost factor. truck or discharge measuring devices are used to operate sev- eral gauges in a region. Making this distinction, capital costs should be broken down into two subcategories: construction 112    How to Estimate Total Costs of Operating a Stream-Gauging Network and equipment costs specific to a site are associated with indi- Agency Costs to Install Operational Systems vidual stations, while other capital costs can be shared across Agency costs relate to internal or consulting costs required to a network. Thus, the total cost of field equipment is determined ensure that any site construction meets all the permitting and by the equipment to be installed at the gauges (stationary access requirements needed to operate the system. Typical in- equipment) and the mobile field equipment required to oper- ternal costs are related to human resources required to man- ate a group of gauges. age the project, carry out field inspection, or carry out some of the work. Project management and construction logistics are Total cost of information processing systems (IPS). These sys- two important considerations. tems are essential for data retrieval, quality assurance/quality control, data storage, data processing, and disseminating de- Project management. This role should be carried out by an rived information. Their cost depends on the necessary equip- ment, including computer servers, communication interfaces, NHS staff member (if experienced in project management) or backup systems, and software. It also includes the cost of ICT, a consultant hired to act on behalf of the agency. The project which links stream gauges and hydrological offices. manager is responsible for developing the resulting tender documentation, coordinating and scheduling equipment deliv- Cost of Civil Works ery with the vendors, supervising civil works and installations, The operation and longevity of the various components of overseeing the training of staff, and testing to ensure that the stream gauges depend on the correct installation of the tech- project runs smoothly and is completed on time. The time re- nical equipment, which must be adapted to local conditions. quired for project management can be as little as six months River gauges must be located on riverbanks in such a way or, depending on scope and complexity, as long as one year or that they can capture all water-level fluctuations and are not more. damaged by extreme floods. In some cases, existing structures (such as bridges) can be used for installations; in other cases, Cost of logistics. This includes NHS staff travel to the site(s) to specific infrastructure has to be installed. The total costs to complete the installation; time on site to complete the instal- prepare the location of a stream gauge depend on: lation; and time of staff at a vendor facility for the site accep- tance test, training, travel days, and cost of travel. ■ Cost to procure land. This includes the cost of properties in floodplains, which are often inexpensive but not suitable for operating gauges under all hydrological conditions or re- 2.7.1.2 Annual Operating Costs quire special construction provisions. Operating costs are the annual expenses associated with the ■ Cost of site preparation. This includes the cost of leveling operation of stream gauges and the provision of hydrological and/or clearing the land and constructing roads to facilitate services. But these costs are often underestimated. For an NHS, installing, operating, and maintaining stream gauges. these costs can be subdivided into direct and indirect operat- ■ Installation of utilities. This includes the cost of power con- ing costs. nections and communications. ■ Construction of major supporting structures. This includes Direct Operating Costs the cost of items (such as instrument shelters, buildings, Direct operating costs. These are costs that relate directly to abutments, embankments, stairs, stilling wells and in- the operation of gauging stations and the production and pro- take systems, foundations of cableway systems, and light- ning-protection systems) required for installing stream vision of data and information from these gauging stations. gauge components (such as recording devices, staff gauges, They include the collection, compilation, calculation, and ap- cableways). proval of data; the verification and transmission of data to the ■ Cost of security. This includes the cost of perimeter fenc- national archive; the regional archiving of data and derived in- ing, cameras, and alarms for security to prevent theft and formation; and the provision of data, information, and services vandalism. to users. How to Estimate Total Costs of Operating a Stream-Gauging Network    113 Field operating costs. This category covers a variety of items. p Costs of service-level agreements to service complex One is the cost of discharge measurements. Since discharges equipment such as servers (which reduce training must be derived from water levels by applying a relationship costs and salaries of highly qualified technical staff). between water level and discharge (rating curve), regular ■ Calibrations of measurement instruments (such as rotor measurements are essential to establish and control these current meters) needed to ensure data quality rating curves. These can only be done in-situ and require ■ Analysis and planning of changing data demand over time staff to present at the gauge to carry out the measurement. ■ Office rent, costs of furniture, travel costs, and vehicles. There is simply no feasible solution otherwise. These mea- surements typically determine most of the operating costs of Administrative costs. These include day-to-day program, stream-gauging networks and depend heavily on personnel financial, and human resource management; ordering re- and travel costs. The required frequency of measurements at placements of failed equipment; warehousing and invento- each station (often between four and nine times per year) de- ry control; occupational safety and health; and ISO/Quality pends on regional hydrological and geographical conditions. Management System (QMS) management. Another item is the cost of leveling. Besides automated record- 2.7.1.3 Costs of Maintenance ing gauges (flow), a reference staff gauge is an indispensable Depending on the technical installations, the maintenance component of a stream-gauging station. It is used to adjust of field equipment (such as measuring devices, shelters, and the recording systems and deliver manual observations if it shafts) is required at regular intervals or on demand (for ex- fails. The elevation of the zero point of the staff gauge has to ample, if gauges are damaged by floods). Certain components be fixed and controlled at least once per year by leveling to of an automated station have a shorter lifetime than the sta- ensure the comparability of water-level data over years. tion as a whole and require one or more replacements over the system’s life expectancy. One example is cellular commu- Yet another source of direct operating costs includes (1) con- nications devices, which have a lifetime of 5 years as a com- sumables, (2) land lease costs for use and access to land not ponent of a station with a lifetime of 15 years. An estimate owned by the government, and (3) cost of utilities (such as of the cost and frequency of replacement of spare parts and power connection and communication costs). components is required for life-cycle costing (LCC) to enable the NHS to plan future expenditures and make procurement Indirect Operating Costs decisions. To distinguish between regular and unscheduled Indirect operating costs are business operating and admin- maintenance and to obtain data on the reliability, vulnerabil- istrative costs that are attributable to the collection and pro- ity, and robustness of the observation network, three types duction of data and information that are not attributable to a of maintenance must be considered: preventive, corrective, given monitoring site and distributed over the entire network. and adaptive. Business operating costs. These include all the costs on top Preventive Maintenance of the direct operating costs that are necessary to operate Taking preventive steps is a significant cost for all observa- the NHS. Here “necessary” means that the expenditures help tion systems and can have two or more levels of technical carry out all the activities that are required for an NHS to re- training required: alize its missions. These costs include: ■ Cost of local observers. Observers who are located near the ■ Costs of IPS site, with minimal training, can handle jobs such as ensur- p Costs to develop and operate new data products (such ing the site is secure, the grass is cut, and the staff gauge is as web-based information systems) clean. Hiring them reduces maintenance costs for the less p Software licenses that provide various levels of up- technical work, because they command a lower pay scale grades and support with no logistical costs. p Costs of cloud storage, computers, and ICT 114    How to Estimate Total Costs of Operating a Stream-Gauging Network ■ Cost of full-time equivalent (FTE) workers for hydrologi- the effectivity of maintenance, equipment reaches the end of cal technicians (NHS staff). This is the annual cost of the its lifetime and has to be completely replaced. Replacing the FTE staff to carry out maintenance work on gauges. Field entire piece of equipment (such as an automated station or technicians require a higher degree of technical education an acoustic Doppler current profiler, or ADCP) should not be and training to do routine preventative maintenance. The confused with replacing spare parts or components, which is frequency (as per vendor guidelines) may be increased part of maintenance. The costs of equipment replacement are or decreased because of site environmental conditions. regularly recurring costs for high-value items for which a spe- Maintenance typically includes equipment checks and re- cific life cycle can be defined, and which consequently have to calibration, periodic cleaning and lubrication, and replace- be replaced on a regular basis. ment of components and upgrades. These activities can be combined with the regular discharge measurements. The longer a piece of deteriorating equipment is operat- ■ Costs of maintenance by external personnel. Some compo- ed the more maintenance it requires to maintain efficiency. nents (such as lightning protection, data transfer facilities, Furthermore, the longer such equipment is kept, the less facilities for occupational safety) must be maintained and value it retains and the more likely it is to be made obsolete controlled (often by law) by external experts. by new equipment with new technologies. If the equipment is replaced frequently, however, investment costs increase. Corrective Maintenance Thus, the problem is to determine when to replace such items An effective program must have adequate supplies, spares, and how much maintenance (particularly preventive) to per- and trained electronic and other maintenance personnel form so that the sum of the operating, maintenance, and in- readily available. Because corrective maintenance is the re- vestment costs is minimized. pair of unanticipated equipment failure or damage from nat- ural events, vandalism, or theft, it is difficult to calculate an Often an expected lifetime exists to support the planning of exact value of these costs. Thus, there needs to be a register replacements. But the real lifetime also depends on the phys- of all costs and their temporal and spatial distribution to en- ical and economic lifetime: sure that the NHS can develop its own understanding of these costs through the operation of the system. ■ Physical lifetime of a piece of equipment. This is the amount of time before its ability to function drops to zero. Adaptive Maintenance This depends on design, manufacturing quality, and mate- This is the planned support and upgrades of the entire value rials used, as well as operating conditions and the quality chain, from measurement to data processing to the state- of maintenance. It also depends on whether preventive of-the-art system. Typically, upgrades (such as software up- maintenance and timely replacement of essential compo- grades) are needed so that the NHS can take advantage of nents with shorter service life are ensured. That said, the innovations that occur during the system’s lifetime that physical lifetime should be based on the manufacturer’s provide more value from the data. Although adaptive main- recommendations. For example, a complete solar-powered tenance is typically not performed annually, the NMS should self-sufficient station with radar level sensor or compact consult with the vendor on the typical frequency and associ- bubbler sensor may have a lifetime of 15 years with prop- ated cost. Often annual service-level agreements provide for er maintenance, which includes the replacement of short- updates and upgrades to systems. er-life components such as batteries and moving parts that have a lifetime of 5 to 7 years. 2.7.1.4 Cost of Replacements ■ Economic lifetime of a piece of equipment. This refers to Replacement problems involve items that degenerate with use the point where maintaining a product is more expensive and with the passage of time, depending on the environmen- than replacing it—and its economic lifetime could be dif- tal conditions. This also involves items that fail after a certain ferent from its actual physical lifetime. To determine the amount of use or time, as their lifetimes are limited for tech- economic lifetime, it is necessary to record and analyze nical reasons. Depending on environmental conditions and the costs of maintenance, spare parts, and replacements of How to Estimate Total Costs of Operating a Stream-Gauging Network    115 components and to compare these costs with the residual dispose of an object or process, when it is equally appropri- value of the equipment. Moreover, the economic lifetime ate to implement each of them. TCOs are a substantial part of can be reduced by technical progress if new technical de- such analyses—but the scope of LCCA is more extended, given velopments replace existing technologies. For example, that indirect (“hidden”) costs and the replacement of equip- most desktop computers can be expected to work for at ment over a long period of operation are considered. LCCA least three years with the current software. However, the includes costs associated with staffing and training, upgrades requirements of the new software mean that these com- of technical systems, energy consumption, licensing, spare puters must be upgraded with new components to then points, instrument failures, depreciation, and asset disposal. continue to be used for five to eight years. A hydrological observation network consists of many differ- 2.7.1.5 Application of the TCO ent items (such as stationary and mobile field equipment, ve- hicles, IPS facilities). Each item has a specific TCO. Life-cycle The TCO refers to the sum of all costs incurred throughout the costing (LCC) is a method of adding up all the costs associat- lifetime of owning or using an asset. In general, TCO analy- ed with an asset, starting from its initial cost of replacement. sis brings out the "hidden" costs of asset ownership and in- Of course, it is quite difficult to estimate the TCO of a network cludes all the costs related to system acquisition, along with based on the life-cycle cost of each component, but LCC is the labor costs for those using or supporting these systems. very useful for comparing products with the same function- The TCO enables decision-makers to include operating and ality, albeit a different distribution of cost categories during maintenance costs, along with procurement costs, in the de- their lifetime. cision-making process to choose between competing offers. It combines initial costs and expenses associated with operat- The LCC calculation involves adding six types of costs: (1) ing and maintaining the equipment: purchase costs, (2) maintenance costs, (3) operational costs, (4) financing costs, (5) depreciation costs, and (6) end-of-life TCO = Initial Costs + Costs of Operation + Costs of costs and replacement costs. However, not all these types of Maintenance − Residual Value costs are relevant for all equipment. One example is that of hydrological measurement equipment, for which the disposal Generally, public services tend to buy products with a lower after lifetime (end-of-life costs) is often not very costly but upfront cost. However, with time, the recurring costs (main- could be high for other assets (such as a building). Another tenance costs, operating costs) add up. Over its total lifetime, example involves depreciation. For private industry, the de- one product can be much more expensive than one that has a preciation of assets allows them to write off a portion of the higher upfront cost but a lower recurring cost. Money saved purchase cost each year over the expected lifetime of the at the time of purchase can be easily exceeded by operations asset for both tax and accounting purposes. In the public ser- and maintenance costs. In practice, the cost of purchase and vices, depreciation is not relevant for taxation. Regardless, the costs of operation and maintenance are often itemized accounting for the residual value of assets is still essential separately. The former is booked as a capital expenditure for decision-making about the reasonableness of replacing while the latter is part of running costs.  A comprehensive components, evaluating maintenance by comparing the ex- analysis of the TCO unifies both items (see box 2.7.1). pected lifetime with the real lifetime, and planning the future replacement of measurement systems. 2.7.2 Life-Cycle Cost Analysis An example of LCC for a stream gauge with a typical wa- Life-cycle cost analysis (LCCA) is a tool to determine the ter-level recording system is given in box 2.7.2. It shows that most cost-effective option among different competing al- operation and maintenance accounts for by far the largest ternatives to purchase, own, operate, maintain, and, finally, share of total costs. 116    How to Estimate Total Costs of Operating a Stream-Gauging Network BOX 2.7.1  How to Compare the TCO of Two Gauging Stations A National Hydrological Service (NHS) got two offers for different complete solar-powered self-sufficient gauging stations. These stations differ in most cost categories: the capital cost of the water-level recording system, the lifetime cost, the cost of annual maintenance, and the lifetime of an essential component (batteries). But the cost of replacing these batteries is equal: $1,500 per change (table B2.7.1.1). TABLE B2.7.1.1  Life-Cycle Costs of Station A and Station B Cost of annual Lifetime of the Cost of Capital cost of Lifetime of maintenance (without installed lithium replacement Cost of annual Station equipment equipment replacement of batteries) batteries of batteries operation Station A $15,000 15 years $850 5 years $1,500 $800 Station B $11,000 12 years $1,200 3 years $1,500 $900 To estimate and compare the annual costs of owning, operating, and maintaining both stations, the equivalent annual costs have to be estimated. The capital cost has to be translated into a capital annuity—a series of annual payments to repay this equipment over its lifetime. Furthermore, the replacements costs have to be annualized, which means accounting for the dif- ferent times at which these costs are incurred. This is done by discounting these future costs into their present values with the following formula: n PR = R1 + R2 (1 + i)1 (1 + i)2 +…+ Rn = Rk ∑ (1 + i)n k=1 (1 + i)k Where PR is the present value of all replacements over the lifetime of the gauging station, Rk is the cost of replacement of a com- ponent at the end of the fiscal year k (for Station A $1,500 at the end of year 5 and 10, and for Station B $1,500 at the end of the years 3, 6, and 9). The value n stands for the total lifetime of the station in years (15 or 12 years) and i for the discount rate (in this example, 0.05). The present value of the replacements also has to be translated into an annuity. This is done by transferring the capital costs C (or the present value of all replacements PR) into an annuity A with the following formula: i (1 + i) n .C A= (1 + i) n ­— 1 The annuity of the capital costs C of Station A are $1,445 and the annuity of PR $202; for Station B the respective values are $1,241 and $382. Now the equivalent annual costs of both stations can be compared (table B2.7.1.2). TABLE B2.7.1.2  Comparison of Annual Costs for Station A and Station B Cost Station A Station B Annuity of capital cost $1,445 $1,241 Annuity of the replacement of batteries $202 $382 Cost of annual maintenance $850 $1,200 Cost of annual operation $800 $900 Total equivalent annual costs $3,297 $3,723 By standardizing the annual cost, the NHS in charge of a capital budgeting decision where cost is the only issue would select Station A because it has a total equivalent annual cost that is $426, or 11 percent lower than that of Station B. Source: Original calculations for this publication. How to Estimate Total Costs of Operating a Stream-Gauging Network    117 BOX 2.7.2  How to Calculate Life-Cycle Cost for a Typical Gauging Station PHOTO B2.7.2.1  A Standalone Solar- Powered Station with Compact Bubble Sensor, Germany The equipment: A water-level recording system with an automated, standalone solar-powered station with radar level sensor or a compact bubble sensor (Photo B2.7.2.1). Here it is critical to understand that this example is directed toward estimat- ing water level or “stage,” not streamflow, which is the derived parameter. The estimation of discharges from recorded water levels requires a relationship between water level and discharge that has to be calibrated by discharge mea- surements. Not considered here are the resulting capital costs for field equip- ment, or operational costs for discharge measurements, data retrieval, quality assurance/quality control, and data storage and processing. The main categories of TCO can be specified as shown in tables B2.7.2.1 to B2.7.2.3: Source: A.Schumann, World Bank. TABLE B2.7.2.1  Capital Costs (Purchase Costs, Cost of Civil Works) (US$) Cost category Amount Purchase costs of the water-level recording system with logger, uninterruptible power supply, $16,000 and control panel Installation and parameterization $1,000 Civil works (foundation, installation of a staff gauge) $3,000 Total initial investment costs $19,500 TABLE B2.7.2.2­   Maintenance Costs (US$/year) Cost category Amount Minor maintenance by observers (control, cleaning, cutting of vegetation) $500 per year Maintenance by qualified technicians (required equipment checks and recalibration, periodic $750 per year cleaning and lubrication, and replacement of components and upgrades) Replacement of rechargeable batteries and moving parts $500 every 5 years TABLE B2.7.2.3  Operation Costs (US$/year) Cost category Amount Control of incoming water-level data, quality assurance/quality control, data storage and $1,000 per year processing (without discharge measurements) continued 118    How to Estimate Total Costs of Operating a Stream-Gauging Network (Box 2.7.2 continued) All of the costs except the initial investment cost have to be discounted to the present. Since they are recognized over time, a discount rate must be applied to the cost (or salvage value). Discounting is the process of determining the present value of a payment or a stream of payments. The reference point is always the time when the first costs were incurred. The period is re- ferred to the respective end of an operating year. The discount rate should reflect the opportunity cost of using public resources for other projects or programs. The starting point for this opportunity cost is the inflation rate. The present value of the costs of the two replacements of batteries in this example ($500) can be estimated as follows: MCR5 MCR10 PMCR = + (1 + i) 5 (1 + i)10 With an assumed discount rate i = 0.05, this results in a present value of the two replacements in the framework of preventative maintenance of PMCR = $392 + $307 = $699. This discounting, which has to be done for all costs except initial costs, results in the following formula: T T T LCC = C0 + ∑ OCt (1 + i)t + ∑ MCt (1 + i)t + ∑ MCRt (1 + i)t – SVt (1 + i)T t=1 t=1 t=1 where LCC = life-cycle costs over lifetime T, C0 = initial investment costs, OCt = operating costs in year t, MCt = costs of maintenance in year t, MCRt = costs of exchange of components with shorter service life than T (including spare parts), and SVT = salvage value at end of lifetime T. Here T stands for the expected lifetime of 15 years, t is the year of operation where the costs incurred, C0 for the initial invest- ment, OC for the operation costs, MC for the costs of maintenance (labor costs), MCR for the costs of spare parts (which can be easily added to the annual labor cost of maintenance and which are here handled separately for didactic reasons), and SV for the salvage value. If we assume that the costs of labor (or maintenance) will be increased by inflation with the same discount rate i, discounting and annual growth cancel each other. Special emphasis must be given to the salvage value: The expected lifetime of the station is determined by the water-level re- cording system. Construction items (foundation, staff gauge) have a lifetime of 30 years. After 15 years, the salvage value is 50 percent ($1,500) of the construction cost with a present value of $347. The LCC of this typical water level gauging station is given in table B2.7.2.4: TABLE B2.7.2.4  Present Values of Costs, Salvage Value, and the Resulting Life-Cycle Costs (US$) Costs and salvage value Present values Initial investment $19,500 Operation costs $15,000 Maintenance (labor) $18,750 Maintenance (spare parts) $699 Salvage value (gain) $347 Life-cycle costs $53,602 It is essential to note that the capital costs are only 36 percent of the life-cycle costs while operation and maintenance are 64 percent ($2,250 per year). It should be noted that the provision of discharge data is associated with significantly higher oper- ating costs (discharge measurements, rating curves, data conversion). Source: Original calculations for this publication. How to Estimate Total Costs of Operating a Stream-Gauging Network    119 2.7.3 The Problem of Comparability of Costs Based on this information, the entire network must be filtered into groups of categories of station type to specify the costs of It is difficult to make a general estimate of the total cost of operation, maintenance, and replacement within these groups ownership of a stream-gauging network. Although the various and to estimate the total expected cost of the network. The cost categories are the same, there are significant differenc- relevance of such an approach is demonstrated in annex 2.1, es in ownership costs both between NHSs and within NHSs where the annual operating costs of hydrological networks in among gauges. Several factors may result in differences be- southern and northern Canada are compared. Because salary tween these ownership costs (Hamilton 2012): costs vary and reflect national conditions, we discuss the re- quired labor in full-time equivalents (FTEs), together with the ■ Differences in capital costs, caused mainly by different associated personnel costs, although these costs are country costs for construction and mobile field equipment specific. ■ Frequency requirements of discharge measurements and maintaining services Despite this complexity, it is useful to compare data on the ■ Unscheduled field visits to repair or replace equipment average costs of setting up and operating watercourse net- (corrective maintenance) works between NHSs in order to show the relationships be- ■ Time and effort spent on corrections and post-processing tween the various cost categories and, based on this, the total of the data costs of ownership. ■ Data lost due to sensor failure ■ Amount of data degraded by high uncertainty 2.7.4 Cost Assessments of Stream-Gauging ■ Costs for supplies of spare parts and accessories. Networks in Developed Countries Within a country, NHSs are faced with the challenges of het- It is difficult to obtain accurate cost estimates for hydro- erogeneous natural conditions, such as the need for stream metric networks from NHSs. In some cases, total costs are gauges on rivers of various sizes. The problem is that vari- provided, but not at the required level of detail. Here, the able stream cross-sections, very pronounced seasonal runoff three cost categories—initial investment, annual operating regimes with high floods and extreme dry periods, poor traffic costs, and maintenance costs—are specified using data from and communication routes, and strong climatic fluctuations Canada, Germany, and the United States. In this section, each lead to high demands on hydrological monitoring systems. To category is considered and then the total costs are evaluated. give a full picture of costs, it would be necessary to specify Because of the different hydrologic conditions in these coun- the different characteristics of every station in the network. tries, the results show the relationships between the costs of The required information includes: the different categories and the range of costs depending on the different natural conditions. ■ The site and its location ■ The station purposes 2.7.4.1 Initial Investments ■ The type of operation (such as water level only, mechanical Construction costs depend on local accessibility issues (in- recording without automatic data transfer, or full automat- cluding the availability of electric lines, communication lines, ed stations) and shelters). In average conditions with road network acces- ■ Stationary field equipment at stations sibility and electricity (modern stations are often supplied by ■ Mobile field equipment required to operate the site solar power), the installation costs (including construction) ■ Field time required to operate the site to standard may vary from about $10,000 to as high as $100,000 in the ■ Travel distances areas with an insufficient transport infrastructure. These high ■ Office time required to run stations costs are for very remote or inaccessible sites. Depending on ■ Life-cycle management plan. the nature of the installations, some typical construction sta- tion costs in Canada and Germany vary in a broad range, as shown in table 2.7.1. 120    How to Estimate Total Costs of Operating a Stream-Gauging Network TABLE 2.7.1  Potential Construction Costs: Canada and Germany (US$) Canada Germany Typical construction costs Low High Low High Standard shelter and concrete pad 8,000 25,000 10,000 20,000 Staff gauge (with/without stairs) n.a. n.a. 500 8,000 Measuring footbridge n.a. n.a. 30,000 50,000 Cableway system, manned (foundations, structure 30,000 60,000 n.a. n.a. winches, cables) Cableway system, unmanned (cablefox, foundation, 30,000 60,000 20,000 80,000 working platform) Connection to electricity and telephone n.a. n.a. 5,000 30,000 Reinforcement of access road, slope stabilization n.a. n.a. 2,000 40,000 Source: Personal communication, Alain Pietroniro, WSC (Canada); data collection of Andreas Schumann (Germany). Note: n.a. = not available. The site-specific construction costs depend on: ■ Site-specific installations (such as gauge house and gauge shaft, foundation of the staff gauge, embankment stabiliza- ■ The location of the measuring point (access to the con- tion, and riverbed design; installation of a measuring weir; struction site, access to infrastructure, electricity, tele- installations of a measuring bridge or a cable crane sys- phone, pavement of the access road, and flood protection) tem; and demand for redundant remote data transmission) ■ Whether land acquisition is required (photos 2.7.1 and 2.7.2 provide examples of two different ■ Watercourse structure (width, gradient, slope, bed stilling well constructions in Germany that differ in dura- characteristics) bility and construction cost). ■ Installed equipment PHOTO 2.7.1 AND PHOTO 2.7.2  Two Gauges with Stilling Wells in Germany with Significantly Differing Construction Costs Source: A. Schumann, World Bank. How to Estimate Total Costs of Operating a Stream-Gauging Network    121 In addition, higher construction costs result from national TABLE 2.7.3  Costs of a Modern Station (Sensors, Loggers, regulations for occupational health and safety, nature con- Solar Panel, GSM Modem) (US$) servation, construction, and electro-technical installations. Expense type Cost For example, railings, stair access, fall protection, ground- Procurement (e.g., OTT-CBS or RLS, plus 17,000 ing, working platforms, compensatory measures for changes logger, UPS, panel, controller) of natural conditions, and closed seasons for wildlife can all Installation costs (foundation; setup; parame- 4,000 boost costs. Another factor is the expected longevity of the terization by technician, but without adjusting new station and the personnel costs of construction workers. the staff gauge) Costs for connecting to the wireless network 2,000 Local equipment costs depend on the purpose of the station if signal reception is not sufficient and radio relay is required (not always necessary) and the decisions around local in-situ equipment and accu- Total 23,000 racy requirements. The World Meteorological Organization (WMO 2010) highlights several instruments that could be Source: World Bank data. Note: GSM = Global System for Mobile Communications; OTT-CBS = OTT- used, depending on the infrastructure available (including compact bubbler sensor; OTT-RLS = radar level sensor; UPS = uninterruptible nonrecording gauges such as staff gauges, and automated power supply. systems such as data loggers). Table 2.7.2 provides a list of costs of recording water-level instrumentation typically used Field equipment costs result mainly from the need for regular in most installations. discharge measurements. As water-level data are insufficient for quantitative hydrological analyses, further hydrometric TABLE 2.7.2  Costs of In-Situ Recording Water-Level Equipment field equipment is essential for these measurements. A list of (US$) stream-gauging field equipment required for field operations, Equipment type Price (approximate) as shown in table 2.7.4 with their costs, are typically used to operate several gauges (between 15 and 20 gauges) in a Staff gauge 500 network. Other costs can include tools for repair, chain saws Radar sensor 1,500 to remove debris, ice-augurs for drilling through ice, first-aid Hydrostatic sensor 5,000 equipment and other general safety equipment. Bubble sensor 9,000 Floating 1,500  TABLE 2.7.4  Costs of Flow and Velocity Measurement Data logger 5,000 Equipment for 15 to 20 Gauges (US$) Source: World Bank data. Measuring equipment type Price (approximate) Current meter 1,500–12,000 In addition, installation costs and other costs that depend Flow tracker 15,000 on local conditions and the purpose of the station must be Acoustic Doppler current profiler 35,000 considered. This is demonstrated in table 2.7.3 with the ex- (ADCP) ample of the costs of a complete solar-powered self-sufficient Electromagnetic velocity sensor 4,000 station with radar level sensor or compact bubbler sensor in Data logger 5,000 Germany. Similar automated stations are used in the other countries. Source: Personal communication, Alain Pietroniro, WSC (Canada); data collection of Andreas Schumann (Germany). The successful operation of a flow gauge typically requires 4 to 10 field visits per year. The number of gauges that can be visited per day depends on the travel distance between them. Typically, 2 or 3 gauges can be visited per day. These visits are primarily needed for discharge measurements but 122    How to Estimate Total Costs of Operating a Stream-Gauging Network also to control gauges after critical hydrological situations. TABLE 2.7.6  Typical Initial Investment Costs Associated with Typically, one fleet vehicle is required per cluster of 10 to 15 Installation of Stream-Gauging Stations in Canada, Subdivided by Category (US$) gauges, if the vehicle would also carry with it all of the capi- tal equipment required for flow measurements (including the Total cost for boat and trailer, if necessary, for taking measurements). The Cost category 15 stations Cost per station field unit would also carry the required flow measuring equip- Initial construction 300,000 20,000 and shelter ment (ADCP, flow meter), safety gear, and other field equip- In-situ equipment 150,000 10,000 ment. Table 2.7.5 shows the required additional equipment Flow and velocity mea- 15,000 1,000 for field measurements (beside the measurement equipment) surement equipment and their costs. Other field equipment 88,000 5,870 Total 553,000 36,870 TABLE 2.7.5  Costs of Typical Field Equipment for 15 to 20 Source: Data collection of Alain Pietroniro, WSC (Canada). Gauges (US$) Field equipment costs Price (approximate) 2.7.4.2 Annual Operating Costs Truck and canopy 40,000 Knowing the cost of operating a streamflow gauge is ex- Boat and trailer 40,000 tremely important for an NHS as it prepares a budget after Safety gear 5,000 modernization and re-fitting the network’s equipment. Here Tools 3,000 the requirement to measure streamflow on a regular basis to Source: Data collection of Alain Pietroniro, WSC (Canada). update the rating curve and estimate flow plays a prominent role. The bulk of the costs are tied to field visits for mainte- To characterize the relationships between each category of nance and velocity measurements, which occur on average initial investment and capital cost, a typical facility in Canada about five to seven times per year. is used here (table 2.7.6). The following assumptions are The annual operating costs result from equipment operating made for specifying initial costs per station (including initial costs (direct operating costs of gauges) and business operat- construction and shelter, in-situ equipment, unloading equip- ing and administrative costs (indirect operating costs), which ment, and field equipment): are related to producing and delivering data and information from those sites. They are based on the human resources and ■ In-situ requirements are for a shelter, standard data log- consumables required to carry out field visits, computations, ger, pressure transducer system, and minimal construction and maintain office and computer infrastructure. Additional costs. overhead costs, depending on the administrative structure of ■ No cableway or footbridge is required. a NHS, can be estimated as 20 percent of the related human ■ Field equipment can accommodate the servicing of 15 resources costs. in-situ water-level sites. ■ Truck and boats can service 15 water-level sites. The costs for discharge measurements and the setup of rating curves are significant and often dominate the direct operation costs—with human resources likely the biggest component. Because of the need to take reasonably frequent velocity mea- surements in a river to maintain an acceptable rating curve, it is necessary to allocate sufficient personnel for this purpose. Similarly, to carry out the computation required for estimat- ing flow, there is a requirement for manual intervention and interpretation of the data for accurate flow estimation. How to Estimate Total Costs of Operating a Stream-Gauging Network    123 The direct operating costs to estimate discharges are highly frequencies of discharge measurements. Hydrological gauges dependent on the gauge’s location. Here the key elements are that are located on rivers with a stable bed (such as no me- density, variability, and technology. andering, single channel), or on rivers where only one or at most two seasonal variable rating curves must be considered, Density of the stream-gauging network. The staffing capaci- require 20 to 30 percent fewer discharge measurements than ty required for stream gauging is directly related to the num- gauges where the profile is modified permanently by erosion ber of gauges within a region. The demand for hydrological or with a seasonal high variable cross section. observations depends on the natural conditions, water man- agement needs, and exposure of the population and econom- The measurement equipment. The efficiency of discharge ic values to water-related risks. measurements depends on the applied measuring technology (such as ADCP or mechanical current meters) and local in- ■ Case I: In less populated regions, the economic value of stallations (such as the availability of a cable crane system or stream gauging is often low in relationship to the efforts measuring footbridges, or the stabilization of the riverbed by of providing hydrological data. This is mainly the result of construction measures). The opportunities to apply different a low level of infrastructure development. If the distances techniques for discharge measurements depend on the river’s between gauges are long, the travel time of the hydromet- width and depth, its water quality, and other local conditions. ric service to ensure sufficient operation of these gauges is high. The operation of these gauges requires more re- Operating a network is labor intensive. Operating a small sources than average. If there is no need to expand the network of 15 gauges requires in the best case an annual hydrological observation network further, the number of labor effort of at least 90 hours for a qualified hydrologist, gauges should not be extended. another 780 hours for a hydrological technician, 450 hours ■ Case II: In regions where there is a need to develop the for a skilled worker, and 143 hours for an unskilled worker. hydrological observation network substantially, but with The resulting costs depend on national wage levels. The hy- not enough available transport infrastructure, the required drometric technician’s main duties are to conduct streamflow capacities to operate gauges are determined by high- measurements and respond to the difference between what er-than-average travel times, which can be partially com- is found and what was expected, along with maintaining the pensated for by a higher density of the network. site. Nearly 50 percent of the working time is required for dis- ■ Case III: If there is a high-density observation network charge measurements. Depending on these criteria, a quali- and an average availability of transport infrastructure, the fied hydrological technician could handle possibly 15 gauges. further extension of the stream-gauging network requires Under unfavorable conditions, the technician could handle fewer additional resources for operating new stations than only 5 to 10 gauges, but in a very favorable case up to 20. in the two other cases. Here it is assumed that other people support that technician in taking discharge measurements and station maintenance. To specify the relationships of required resources among the Two people are needed to handle both the equipment for dis- three cases, discussed above, if the required resources in charge measurements and ensure occupational health and Case III were equal to 100, in Case II it would be equal to safety considerations. The costing of labor for the operation 125, and in Case I equal to 200. of a network assumes that one technician can maintain field visits and data production for 15 stations and that someone The variability of flow cross-section geometry. The derivation in a supervisory role can oversee these efforts for up to five of a relationship between water level and discharge and their technicians (75 stations). continuous updating is essential to provide flow data of good quality. Given that the flow cross-section is changing from The operating costs of a single gauge (excluding the cost geomorphological processes such as erosion and sedimen- of labor and maintenance) are in the range of $2,000 to tation—but also from seasonal varying impacts (such as ice $10,000. This wide range implies that there are no fixed cost- and vegetation)—the local conditions determine the required ings associated with any given type of station, but rather a 124    How to Estimate Total Costs of Operating a Stream-Gauging Network costing approach that needs to be developed on a regional or case is given in box 2.7.3. It demonstrates how the total work- national basis to determine the true costs of setting up and load depends on the size of the cross-sections, the travel time operating a stream-gauging network. involved, and the number of measurements needed per year. This assessment can easily be applied to other cases. One example of how to specify the workload for discharge measurements and rating curves derived from it in a typical BOX 2.7.3  Example of How to Determine the Workload for Discharge Measurements and Estimation of Rating Curves for 15 Gauges Basic assumptions: ■ The 15 gauges are located at rivers with different widths. ■ A mechanical flow meter is used for discharge measurements. ■ Six measurements are done per gauge and per year, plus one leveling of each staff gauge. ■ The travel time per day with discharge measurements is assumed to be 5 hours (small rivers are closer together than large ones, discharges at two gauges can be measured per day at small rivers but just one at a large river). The annual workload, required for discharge measurements at these 15 gauges, results from a combination of different activities that are dependent on the structure of the network as shown in table B2.7.3.1 TABLE B2.7.3.1  Specification of the Annual Workload for Discharge Measurements of a Hydrological Technician for an Exemplary Measuring Network of 15 Gauges Duration Visited Workload measurement Number of Workload gauges per Total number travel (5 Total River width (hours) gauges (hours/year) day of days hours/day) workload ≤25 m 1.5 6 54 2 18 90 144 >25 m and 2.5 6 90 2 18 90 180 ≤100 m >100 m 4 3 72 1 18 90 162 Staff gauge 1.5 16 24 24 leveling Source: Assessment by Andreas Schumann (Germany). The analyses of these 90 measurements (6 measurements per gauge, 15 gauges) are connected with an office time for the hy- drological technician of an additional 180 hours, the setup of new rating curves (assumed for a required change every second year) with a further 24 hours. The monthly quality control of data requires another 270 hours/year. Thus, in this case the total workload of a hydrological technician would be 984 hours per year. To specify annual FTEs, the basic divisions of a technologist’s activities include leaves (as per contractual agreement) and work time have to be differentiated into operational activities training. It is estimated that non-operational activities (meet- and non-operational activities. Operational activities can be ings, testing and maintaining equipment, purchasing replace- further divided into hydrometric work (field and office), meet- ments for worn parts, and other administrative work) account ings and committees, and office activities (administrative du- for about 20 percent of an annual work cycle, leaving 80 per- ties, equipment maintenance and testing). Non-operational cent for hydrometric work. These stream-gauging activities How to Estimate Total Costs of Operating a Stream-Gauging Network    125 include station data analysis (50 percent) and field activities A vehicle may have a life cycle of 10 or 12 years, depending on (50 percent). Office activities include computation time, ar- its use. Similarly, hydro-acoustic sensors typically have a life chiving of data. cycle of about 8 years, while in-situ water-level equipment and data loggers may have a much shorter life cycle, given Several administrative activities can be attributable to the their exposure to the environment and the possibility of dam- collection and production of data but are distributed over the age from floods, ice, debris, or other factors. entire network (that is, they are not attributable to a given monitoring site). These activities include day-to-day program Ongoing or amortized capital costs of equipment required financial and human resource management, warehousing and to measure discharge can be estimated according to the as- inventory control, ISO/QMS management, and training. Over sumed lifetime (in table 2.7.7, it is assumed that the equip- and above the single field technologist, there are require- ment is used for a group of about 15 stations). ments for supervision, office staff, maintenance staff, and ad- ministration staff. These indirect human resources costs are TABLE 2.7.7  Lifetime of Selected Components of Hydrometric considered to be typical overhead costs for maintaining the Networks (Years) infrastructure and are estimated to require about 20 percent Equipment Lifetime of a person year. In-situ field equipment 5 In total, operating 15 sites requires 1.2 person years of ef- Acoustic Doppler current profiler (ADCP) 8 fort from hydrological technicians plus 0.3 person years Truck 12 from a worker/driver. Boat 20 Cableway 25 Salary cost can be determined by the level of effort, field Source: Data collection of Andreas Schumann (Germany). time, and office time required on a per station basis as shown above. Determining how many stations a technician can oper- ate then determines the number of staff required to maintain 2.7.4.4 Assessments of the Total Costs of Stream Gauges the network. The upfront capital cost of establishing a new station is often the easiest and most tractable part of establishing the cost, 2.7.4.3 Costs of Maintenance but it represents the smaller portion of the true cost. Any The maintenance costs depend on the qualifications of staff modernization project requires an assessment of the TCO of and the repair-friendliness of the equipment. In the best case, the network. Given that these costs can be broken down into NHS employees can perform the needed maintenance, but in several components, it is relatively easy to specify costs of the worst case, this requires external specialists. The costs of office equipment, computers, and software that are essen- maintenance are closely related to the equipment’s life cycle. tial to run a modern hydrometric service and are compara- Good maintenance can prolong the life cycle. If capital equip- ble all over the world. Also costs of the field equipment are ment is required for each site, then it needs to be amortized also comparable all over the world. The costs of construction accordingly. Shelters, weirs, and other structures must also could differ significantly regarding the required efforts (which be maintained. depend on climate conditions, location of stations, longevity) and the costs of local labor. However, it is difficult to specify It is often necessary to replace worn parts to achieve the the costs of maintenance and discharge measurements, as rated service life of a station. Automated water-level gauging these depend heavily on installed equipment, location, and stations have a lifetime of 15 years, but some components local conditions. Moreover, expectations of reliability and re- (cellular data transmission or rechargeable batteries) should silience of hydrometric networks must be considered. be replaced after 5 years. The long-term availability of these components and, if they are purchased from abroad, the cost Regardless of these uncertainties, this report tries to shed of their importation must be considered. light on the average cost of procurement and installation of 126    How to Estimate Total Costs of Operating a Stream-Gauging Network modern stream gauges (sensors, loggers, data transfer) and of very unfavorable locations, the annual operating costs the average costs for operation and maintenance in some de- range from 30 to 50 percent of initial investments (table veloped countries. The results show that, with the exception 2.7.8). TABLE 2.7.8  Ratios of the Costs of the Initial Investment to the Costs of Direct Operation for Stream Gauges in Some Developed Countries Annual costs for operation and maintenance, including discharge measurements and Ratio of costs of operation to Country Initial investment (US$) data management, (US$) costs of installation Republic of Ireland $18,000 $6,000 0.33 United States $35,000 $10,000–$15,000 0.3 to 0.4 Canada $20,000–$30,000 $10,000–$20,000 0.3 to 1.0 Germany $20,000–$30,000 $10,000 0.33 to 0.5 Source: Data collection of Andreas Schumann (Germany). But such general assessments tell us little about the various between the northern and southern parts of the country. In categories of costs, which help determine what approach table 2.7.10, the costs of both regions are summarized. to take to estimate per-station operating costs. An example (table 2.7.9) from the USGS (USGS 2007) illustrates how the Another example, based on a cost analysis of a stream gauging annual operational costs per station are assigned to different network in Germany consisting of 105 stations, distinguishes categories. Here the total costs are higher than indicated in between office and field work. Here the minor maintenance table 2.7.10 when the administrative costs are considered. is contained in the field work. Only for maintenance by ex- ternal parties is separate information on costs provided. The Another detailed cost analysis is given for Canada in annex 2.1. summarized costs of spare parts are combined with life-cycle There the costs of stream-gauging stations differ significantly management costs (table 2.7.11). The costs of administration are not given. TABLE 2.7.9  Total Operating Costs, Assessed per Station and Year by the U.S. Geological Survey, by Category (US$) Category Cost per stream gauge per year Building and utilities 1,100 Field equipment 1,500 Vehicles 490 Travel 190 Data management and delivery 1,200 Total annual cost per typical continuous stream gauge 4,500 Labor costs for field and office work 5,300 Administrative costs 4,200 Total costs 14,000 Source: USGS 2007. Note: Data are based on survey data from Colorado, Texas, and the State of Washington. How to Estimate Total Costs of Operating a Stream-Gauging Network    127 TABLE 2.7.10  Total Annual Operating and Maintenance Costs, Assessed per Station by the Water Survey of Canada, by Northern Canada and Southern Canada (US$) Cost per stream gauge per year Category Southern Canada Northern Canada Operating costs Total operating costs including field visits, lease, power, telemetry 2,012 9,406 Costs of office operation (furniture, supplies, phone) 146 1,297 Hydrometric software 159 169 Total operating costs (not including personnel costs) 2,317 10,872 Labor costs for field and office work 5,081 8,466 Total operating costs (including personnel costs) 7,398 19,338 Maintenance costs Station maintenance, repair, and building maintenance 319 775 Life-cycle management (depreciation) For equipment worth less than $10,000 (operating) 144 266 For equipment worth more than $10,000 (capital) 438 610 Total costs of maintenance 901 1,651 Total costs of operation and maintenance 8,299 20,989 Administrative costs 1,650 2,954 Total annual costs 9,949 22,292 Source: Cost analysis from Alain Pietroniro, WSC (Canada), 2013. Note: Number of stations: 573 in Southern Canada, 129 in Northern Canada. Number of staff: 61 in Southern Canada, 20 in Northern Canada. TABLE 2.7.11 Total Annual Operating Costs, Assessed per Station, Derived from an Analysis of the Operation Costs for 105 Stream Gauges in Germany (US$) Category Cost per stream gauge per year Operating costs Costs of field work: Discharge measurements and minor maintenance (field work, including salaries) 2,200 Costs of office operation (analyses of measurements, assessing of rating curves, quality control of 4,300 data, including salaries) Costs of electricity 200 Total operating costs (including personnel costs) 6,700 Maintenance costs Cost of maintenance by external personnel 1,800 Cost of inspections by external personnel (e.g., to ensure occupational safety and health) 600 Cost of spare parts and annualized life-cycle costs (depreciation of station, vehicles, and measure- 1,200 ment instruments) Total maintenance costs (costs of externals and spare parts) 3,600 Total costs 10,300 Source: Data collection of Andreas Schumann (Germany). 128    How to Estimate Total Costs of Operating a Stream-Gauging Network These three examples differ in their consideration of individ- summarizes the range of capital costs and annual costs for ual cost shares. It should be considered that the operating operating and maintaining a modern stream gauge. These costs depend on the level of wages, but the administrative are approximated values, which may result in different total costs are also very different. Considering these differences, costs, especially in the different combinations of favorable the total cost of operating stream gauges can be only ap- and unfavorable conditions. However, it is clear that the on- proximated. Based on detailed costing estimates from the going costs for operation and maintenance significantly ex- Water Survey of Canada, the United States Geological Survey ceed the capital costs after only a few years. (USGS), and a German hydrometric network, table 2.7.12 TABLE 2.7.12  Range of Costs per Station: Examples from Canada, Germany, and the United States Category Remarks Minimum Maximum Initial investment: Capital costs Stationary field equipment, civil works without measuring footbridges or cableways $23,000 $37,000 Mobile field equipment (pro rata costs of flow and Current meter or flow tracker or ADCP or elec- $1,000 $2,700 velocity measuring equipment, in shared use for 15 tromagnetic velocity sensors, data logger stations) Other mobile field equipment with shared use Truck, boat and trailer, safety gear, tools $3,500 $5,000 Total capital costs $27,500 $44,700 Annual operating costs Field work with minor maintenance and work up of Without costs of labor $2,000 $8,100 water-level and discharge data in offices Business and administrative costs, offices, software, $1,000 $4,000 computers Annual maintenance costs Annualized costs of maintenance, including spare $1,000 $3,600 parts Costs for annual operation and maintenance (without personnel costs) $4,000 $15,700 Labor costs for field and office work $5,000 $8,700 FTE This FTE is required to operate an average of 1.5 2.3 15 gauges Total costs of annual operation and maintenance (with personnel costs) $9,000 $24,700 Costs of annual operation and maintenance as a percentage of capital costs 33% 55% Source: USGS 2007; data collections of Alain Pietroniro, WSC (Canada) and Andreas Schumann (Germany). Note: ADCP = acoustic Doppler current profiler; FTE = full-time equivalent. Although these costs were collected in developed countries, of use, the requirements for operation and maintenance will they are applicable to developing countries because of the probably lie within this range. hydrometric equipment used. Depending on the conditions    129 Summary and Conclusions The restricted availability of water in both time and space limits socioeco- 2.8 nomic developments in many countries. Combined with a more erratic and uncertain supply, climate change will aggravate the situation of currently water-stressed regions and generate water stress in regions where water resources are still abundant today. Under these circumstances, hydrological data and derived information are indispensable to enable users of water re- sources to make wise decisions affecting the security of life and property, ensuring economic growth, and protecting environmental quality. Water resources cannot be managed unless we know where they are, their quantity and quality, and how variable they are likely to be in the foreseeable future. Knowledge about the ever-changing spatial and temporal availability of water, the impact of climatic- and human-induced changes on water re- sources, and the risks caused by hydrological extremes requires functional Vietnam. Photo courtesy of V. Tsirkunov, WBG. National Hydrological Services (NHSs). In recognition of their importance, the World Bank Group, along with other development partners, has invested hundreds of millions of US dollars over the past two decades to rebuild and strengthen hydrological and meteorological observation networks and ser- vices in developing countries. But the introduction of new technologies cannot overcome the existing prob- lems of NHSs in developing countries, which result from (1) fragmented and challenging policy environments; (2) insufficient budgets; (3) an inability to attract, train, and retain qualified staff; (4) insufficient capacities to maintain hydrological infrastructures; (5) poor connection with users; and (6) unsatis- factory service delivery (World Bank 2018). Moreover, such problems will be “Knowledge about the exacerbated in the years ahead because of the following factors: ever-changing spatial and ■ The new technologies will place new demands on the qualifications of temporal availability of personnel. water, the impact of climatic- ■ Additional measuring devices and increased data streams will require ad- ditional hardware and software. and human-induced changes Hydrometric equipment (gauges, vehicles) will need to be expanded. ■ on water resources, and the ■ Spare parts and interchangeable components from abroad will be needed. ■ Many modernized stations will have to be replaced at the same time in the risks caused by hydrological future, according to their life cycle. extremes requires functional Further complicating matters, many networks have been designed without National Hydrological accounting for local operating conditions, making it difficult to ensure sus- tainable use over their life cycle. Three reasons can be cited: Services (NHSs).” 130    Summary and Conclusions Inadequate planning of the observation networks. This oc- network requires comprehensive planning, which should curs because the specifics of the area of operation are not include selection of gauge locations according to the sufficiently accounted for. These specifics range from local potential uses of the data to be collected; selection of operating conditions and the vulnerability of technical solu- technical equipment according to local conditions; as- tions to required human resources and discharge measure- sessment of the human and technical capacity required ments that must be taken at appropriate frequencies and with to operate the modernized observation network; and es- a particular required quality. timation of future staffing and funding needs of the NHS to ensure the continued operation of the modernized Underestimation of operation, maintenance, and replace- observing network and to provide data and information ment costs. An overly optimistic number of new or refur- of the required quantity and quality. bished stations is often considered in a development project, 3. Evaluate the required human resources. The modern- yet the longer-term ability to operate them properly is not. ization of networks is often associated with a change in Modern stream-gauging stations are based on electronic technologies. This shifts the need for technical person- equipment that provide digital data. The application of these nel to the field of electronics, communications technolo- technologies requires qualified staff, redundant systems gy, and information technology (IT) services. Educating (such as the opportunities to switch to manual observations and training personnel in these areas is essential to if the automated systems fail temporarily), a warehousing of cope with technological change and to adapt the quality spare parts, and opportunities to repair equipment within the of hydrological information to increasing demands. country. It is insufficient to operate hydrological gauges to 4. Estimate the total cost of ownership (TCO) of the register water levels without discharge measurements (often modernized network. The TCO results from capital underestimated) and rating curves to transfer water levels costs; infrastructure costs (for example, costs of road into streamflow. constructions); operating costs of stations, informa- tion, and communication technology (ICT), data storage Inadequate processing, storage, and analysis of water level and archiving, and personnel; costs of maintaining and and discharge data. To better meet users’ needs, besides replacing spare parts; and business operating and ad- qualified personnel, there are technical prerequisites (hard- ministrative costs. Because these costs depend on the ware and software), which, in turn, have operation, mainte- specific conditions of the network (for example, the den- nance, and replacement requirements. sity of the observation network, the need for frequent discharge measurements, regional climatic conditions, What can be done to turn this situation around and estab- the cost of maintenance, and other local factors), it is lish or modernize a hydrological observation network? The difficult to establish a generally applicable cost rate, following steps are recommended: but experience has shown that annual operation and maintenance costs range from 33 percent to 50 percent 1. Identify hydrological data and information needs. of capital costs. Regarding the intended long-term op- The net worth of the data derived from a hydrological eration of hydrological observation systems, which are observation network depends on the subsequent uses essential for hydrological data for design purposes, the and applications that are made of them. This value must life-cycle costs have to be considered. These include be assessed in cooperation with users, but also with an the consideration of the costs of replacement when an awareness of existing knowledge deficits about hydro- equipment reaches the end of its lifetime. The consider- logical conditions. Special emphasis has to be given to ation of annual depreciation and the remaining residual ongoing and expected hydrological changes. The main value in life-cycle costing gives the opportunity to de- applications of hydrological data are the planning and cide about the economic sense of replacements of sin- operation of water management facilities, disaster risk gle components with a shorter lifetime in more complex management, and the design of hydraulics structures. equipment. 2. Planning the modernized observation networks. Any modernization or new installation of a hydrometric Summary and Conclusions   131 Before starting the modernization project, beneficiary gov- early warnings, and the requirements for operating water ernments should be aware of all financial, technical, and management systems. human resource requirements for sustainable operation and ■ Costs. Here the recurrent costs for the operation, mainte- commit to comply with them. The estimated TCO is essential nance, and replacement of each station have to be con- for the network design. It must be focused on priorities that sidered together with the capital costs of the investments ensure an outcome that is consistent with financial expecta- for the station and equipment. The investment and run- tions and affordable both at present and in the future. ning costs of a gauge depend on local conditions. There is a tendency to close the most expensive stations in order to 2.8.1 Ensuring Alignment of the Costs of the reduce the overall cost of the network, but it must be con- Modernized Observation System with the sidered that data from such stations—located, for example, Budget in mountainous areas with insufficient infrastructure— may be more important than data from easily accessible Both network planning and funding for operations and main- and thus “cheaper” stations. Thus, in such cases, several tenance should ensure that the network can be maintained objectives must be weighed against each other. and operated over the long term. In many cases, this will re- quire an increase in the budget to ensure the long-term goal 2.8.2 Securing Human Resources to Operate a of modernization, or a reduction in the scope of moderniza- Modern Measurement Network tion so as not to overburden the NHS. If a sufficient increase of the budget is impossible, an upgrade of the levels of devel- Modernization campaigns within a short period of time can opment of stream-gauging networks often can be done only overwhelm an NHS because of the divergence between the gradually, because of limited resources. In such cases, only requirements of new technologies and existing staff skill lev- the most relevant gauges should be automated, and remain- els. It is imperative that short-term targeted improvements ing stations should be operated in a safe manual mode. Any in technical capabilities be combined with the development gradual upgrade of stream-gauging networks will require pri- of a medium- to long-term strategy of institutional support of oritization of gauges based on their importance to socioeco- human resource development. In this context, a fundamental nomic conditions, gains in hydrological information, and the task is to train operators of modernized hydrological moni- ability of the NHS to operate them in a sustainable manner. toring systems to be able not only to collect reliable hydro- Any prioritization of new gauges to be installed or modern- logical data, but also to process these data into information izations of existing sites requires consideration of multiple that meets both the potential of the monitoring systems in objectives and constraints: use and the requirements of the users. The recruitment of qualified personnel to operate, for example, more complex ■ Strategic importance of data. The importance of a partic- measuring instruments and ICT equipment, or to develop ular site’s runoff data to water management needs in the software solutions for combining sensors and databases, is present and future has to be assessed. Ongoing changes made more difficult by the fact that the NHS is in competition of demand and supply have to be considered as well as with other sectors of the economy whose wage structure is the relevance of the data for disaster risk management. In more attractive. The increase of the wage level required to addition, there may be different hydrologic priorities for address this issue successfully is often not considered suffi- base network gauges in different regions of a country, de- ciently in modernization projects. The NHS is challenged to pending on the hydrologic characteristics of the available assess its educational and training needs and to establish water resources and their current or planned uses. strong links with local academic institutions with the aim of ■ Operational data. Here the need to get water-level and organizing permanent education and training opportunities discharge data in near real time must be weighed. The and of installing partnering schemes to create recognized need to receive data in an operational mode depends on pathways from education into careers. If investment is not the specific conditions of the watersheds, the need for made in skilled personnel, all the benefits of the technology 132    Summary and Conclusions will be negated because there will be no personnel to main- existing capacities of the NHS, which are increased stepwise. tain the equipment, process the data, and make the most of it. Although the second option requires the operation of an only partially modernized measurement network over a certain 2.8.3 Main Conclusions period of time, it is more promising in many cases: In the coming decades, water problems worldwide will be ■ By selecting the gauges to be upgraded as a priority, the exacerbated by advancing climate change and growing pop- highest specific socioeconomic benefit of the funding ulations. Devastating droughts and floods threaten socio- would be gained. economic development in many developing countries and ■ The divergence between technical advancement and force people to migrate. Adaptation, protection from hazards human resources would be reduced because the time re- caused by too much or too little water, and wise water man- quired for training, implementation in existing structures, agement are inconceivable without knowledge of the spatial and testing of the new technology in parallel with existing and temporal variability of water resources. structures would be extended. ■ The costs of operation would be better known by the NHS Most NHSs in the least developed and many developing coun- and the future requirements for funding could be assessed tries lack the financial resources, staff expertise, and access in a realistic way. to modern technologies to adequately address their grow- ing missions. International support is therefore essential to However, such a gradual modernization would need to be pro- support them, but this support is mostly provided in form of gressively extended to the entire NHS to ultimately achieve a projects to expand and modernize the observation networks largely homogeneous structure of the network and improve and associated with the introduction of new technologies for the quality and quantity of hydrological information provid- sensors, data transmission, data storage, and data analysis. ed. An extension of the project duration and a higher effort for Such project funding is time limited, but many developing the implementation of the modernization would be unavoid- countries lack the financial resources, management systems, able, but this would be justified by the increase in sustain- and expertise necessary to support the long-term operation, ability and longevity of the modernized hydrological network. maintenance, and eventual replacement of these networks. If rising costs for operation, maintenance, and replacement The bottom line is that sustainable hydrometric networks de- of a modernized network have to be offset by shrinking pend on joint efforts of international donors and NHSs. Both budgets, such a network cannot be funded in the long run. groups of stakeholders are challenged to develop solutions Modernization campaigns that rely on extensive technical in- that are feasible, affordable in the next decade, and effi- stallations within a short period are likely to overwhelm an cient to create added value to socioeconomic developments NHS, as there is divergence between technical advancement through improved NHSs in developing countries. The need and the available human resources. for consideration of country-specific conditions cannot be overstated. A holistic planning approach that takes into ac- In terms of affordability, there are two options: to improve count the totality of the respective circumstances is essential the financial and human capacities of an NHS quickly so that to find sustainable solutions. Planning here determines the the transition to a completely new technological system can success of the outcome. be mastered, or to adapt the modernization projects to the    133 Part 2 Annexes Congo River, preparation of discharge mesurement with ADPC. Photo courtesy of V. Tsirkunov, WBG. Overview Two of the following three annexes are case studies from hydrological obser- vation networks; the third one refers to an interesting planning approach for modernizing a hydrometric network in Sri Lanka. Annex 2.1 originates from the Water Survey of Canada (WSC). The WSC is the national authority responsible for the collection, interpretation, and dissem- ination of water resource data and information in Canada. This case study shows how the direct and indirect costs of operating hydrometric networks depend on natural conditions. For this purpose, the results of cost calcula- tions for hydrometric networks in Canada in the north and south are com- pared. It finds that the average total operational costs (including field visits) per station in the north are nearly five times higher than they are in the south ($9,400 to $2,000). This example demonstrates how adverse climatic condi- tions and inadequate infrastructure determines costs of hydrological obser- vation networks. A rough estimate of the total costs of operation is therefore problematic. Annex 2.2 describes the evolution of a hydrometric network in the Kuban River Basin in the Russian Federation. It highlights the need to increase net- work density to match operational data demand and considers how the in- terplay between hydrological conditions and human requirements for flood safety and hydropower determines network design. Such a needs assessment is essential to ensure a cost-effective expansion of hydrological monitoring systems that best meets the needs of users. Annex 2.3 examines how National Hydrological Services (NHSs) can best en- sure the long-term success of institutional and technical solutions by looking at Sri Lanka, where a new planning approach to modernize its hydrometric network is under way. This approach is based on an analysis of the weakness- es of similar projects in the past (for example, the problem with obtaining imported costly equipment, or insufficient human capacities) and integrates solutions of these problems in the planning process. This project establishes a new approach to planning, installs new ways of procuring modern technol- ogies domestically, and gives special emphasis to the development of human capacities. The long implementation period provides an opportunity for inter- im assessments and adjustments by the project’s steering committee. 134    Annex 2.1 1 Case Study: Valuation and Costing of Stream-Gauging Networks in Canada Canada is the second-largest country in the world in terms of area (behind Aero Cable car suspended on Niagara Whirlpool. Photo: Eileen Tan Russia), with the Water Survey of Canada (WSC) operating nearly 2,800 hydrological stations. This backdrop offers a unique opportunity to discuss the cost assessment methodology and to compare the differences between costs under favorable conditions in the south with the costs in the more dif- ficult northern part of the country. Hydrometric observations in Canada offer a way of quantitatively under- standing the water cycle, especially since they encompass an extremely wide range of phenomena. The mitigation of economic and human losses caused by floods and droughts is simply not possible without stream gauging. Important rainfall-runoff phenomenon and statistical analysis of past hydrological data cannot be achieved without stream-gauging observations. It is also extremely important to collect hydrological data over long periods of time in order to evaluate the natural states of systems and understand how these systems may evolve as a result of land-use or climate change. Disaster mitigation mea- sures and water supply planning for industrial development are impossible without measurements. The socioeconomic development of a country and the well-being of its citizenry are all fundamentally tied to the existence and sus- tainability of hydrometric observation networks. In Canada, the WSC—in partnership with the provinces, territories, and other agencies—operates a network of 2,783 stations and maintains a database con- taining historical data for an additional 5,577 discontinued stations through- out the country. Data from the 5,577 inactive stations are stored with the active station data in the national hydrometric database. Most of the active stations are located in the southern half of the country, where the population and economic pressures are greatest. The WSC has a formal Quality Management System (QMS), certified by the International Organization for Standardization (ISO), to guide the planning and management of its operations. Data collected by the National Hydrometric Program (NHP) are housed in two databases maintained by WSC: Case Study: Valuation and Costing of Stream-Gauging Networks in Canada    135 ■ HYDEX contains inventory information pertaining to active The ultimate beneficiary may be a member of the and inactive water-gauging stations in Canada, including public who benefits from flood warning or a farmer their locations, equipment, and type(s) of data collected. who is allowed a license to abstract water from a ■ HYDAT contains all of the water data collected through the river. NHP for all active and inactive stations listed in HYDEX. These data include daily and monthly mean flow, water Despite these difficulties, the intrinsic value of the network level, and sediment concentrations for stations across and a benefit-cost analysis have been attempted. An assess- Canada. ment of the stream-gauging network in British Columbia ■ Key federal legislation, arrangements, and initiatives sup- Canada (BC Ministry of Sustainable Resource Management ported by long-term hydrometric monitoring information 2003) estimated total annual benefits (in 2003 US dollars) would benefit from a legislative scan to ensure that man- generated by the existing stream-gauging program of 461 sta- dates and legal obligation are being met. In Canada the tions at $65 million. The cost of operating and maintaining federal legislation focuses on a series of 11 or more federal this network, at the time, was estimated at about $3.5 mil- acts. Two of these are highlighted as examples: the Canada lion annually, resulting in a benefit-cost ratio of about 19:1. Water Act from 19854 and the Fisheries Act, 1985.5 In addition, in examining the network design, they estimated a further opportunity cost due to the inadequacy of the net- A stream-gauging network costing and its value over the long work is estimated to result in at least $77 million annually term is a function not only of its effectiveness but also of in capital and operating inefficiencies and losses throughout its sustainability, its cost, its ease of use, and its personnel. the resource sector of British Columbia. They concluded that This annex examines considerations and requirements with- these losses could be mitigated by doubling the network size. in the Canadian context for ensuring sustainable operations. Lessons here can be applied to other parts of the world. Despite their benefits, governments are reluctant to spend and invest in national networks. In a federal evaluation of the Quantifying Benefits for Canada Canadian hydrometric program (Environment Canada 2014), the adequacy of the network was highlighted as bounded and The sustainability of stream-gauging networks is inevitably not meeting existing standards. The audit reported that find- tied to the design of the system, the cost of its operation, the ings from the literature, in combination with feedback from ability to manage it over a life cycle, and the recognition of key informants, indicated that the current network of gaug- the value that its observations provide. It is the last point ing stations in Canada was unable to fully meet the intended that has proven difficult for even the most robust and well-es- objective of satisfying Canadian needs for hydrometric data. tablished services. The economic benefit associated with hy- drometric data is a very difficult question to answer. Water This same report also noted that the rationale for public sector resources managers are often required to defend spending on involvement in the collection of hydrometric data is strongly hydrometric monitoring, yet there are no known or published supported (Hanemann 2006). Hydrological information has methods to assess the value. As noted by Walker (2000), significant economic value, and once produced, it can have many users and be readily shared. Private sector producers The benefits of data are often ill defined and may of such information encounter significant challenges in re- not be realized for many years until a representa- ceiving payment from all users for the cost of producing this tive record has been collected or until a decision is information. The existence of such non-payers means that made which could not have been anticipated, e.g., market forces are unlikely to produce the optimal amount of to build a new reservoir in a catchment. The ulti- hydrometric information. This failure of market incentives to mate use of data is often not foreseen. … Many of produce the best quantity of this type of information is the the benefits are often qualitative and intangible. … core rationale for government provision of hydrometric data. 4 The Canada Water Act can be found at https://laws-lois.justice.gc.ca/eng/acts/c-11/FullText.html. 5 The Fisheries Act can be found at https://laws-lois.justice.gc.ca/eng/acts/f-14/FullText.html. 136    Case Study: Valuation and Costing of Stream-Gauging Networks in Canada The value to society of water information and weak incen- filtered network as direct costs. The station input information tives for private sector providers to satisfy the demand for includes: such information establishes a sound basis for the public sec- tor to provide water information as a public good. ■ The site and its location ■ Type of operation (for example, water-level only, data Understanding the Cost of a Stream-Gauging transfer in real time) Network ■ Funding sources ■ Station purpose In order to understand the cost of running a stream-gauging ■ Office time required to run stations network, the engineering hydraulic principals, local condi- ■ Field time required to operate the site to standard tions, stream bed stability, and many other factors need to ■ Equipment required to operate the site be understood. The stability of the rating curve dictates the ■ Travel distances number of field measurements required to obtain a desired ■ Equipment on site accuracy. This implies that there are no fixed costings asso- ■ Life-cycle management plan. ciated with any given station, but rather a costing approach needs to be developed on a regional or national basis to de- This information forms the basis for determining the per-sta- termine the true cost of setting up and operating a national tion operating costs and the level of effort (time) required for hydrometric network. The hydrometric technician’s main du- a station visit. The operational cost is calculated from the ties are to conduct streamflow measurements and respond total input factors multiplied by a base per-station allocation. to the difference between what is found and what was ex- Air charter costs (if required) are calculated from separate pected—that is, does the discharge measurement fall off the input factors. stage-discharge curve and how will the data be adjusted. Capital allocations are based either on the number of staff In Canada, a stream-gauging costing model was developed or the number of stations dependent on what that item is at- to determine the costs of running existing networks, with full tached to and an assigned amortization period for each item. life-cycle analysis costing and while maintaining a quality-as- For example, a truck used to service 15 stations will typically sured data workflow. It is important to differentiate between have a 10-year life cycle. If the costing model requires each budget and expenditures in this context, since costing models technician to maintain a truck and measurement equipment, present the idealized situation while expenditures may vary then the life-cycle management costs should be tied to the from year to year, depending on conditions, staffing levels, circuit the technician operates and the number of staff. If the floods, and drought conditions. For example, planning a life capital equipment is required for each site, then it needs to cycle for a truck may be an average 10-year operating cycle, be amortized accordingly. but there may be good reason to extend the life cycle to 14 years, depending on its use. Costing models represent aver- Lastly, salary cost can be determined by the level of effort, age condition, full life-cycle replacement, and fully staffed field time, and office time required on a per-station basis. offices. This is rarely possible in real-world situations, and Determining how many stations a technician can operate expenditures are often lower than what a costing model may then determines the number of staff required to maintain the dictate for a variety of reasons. Nonetheless, a costing frame- network. work is critical to set an upper limit for costing of a properly resourced and quality-assured ISO organization. Once these costs are defined, current network operations can be broken down into direct costs and indirect costs. These A stream-gauging costing model takes input information for are defined as: every station in the network; allows for the full network to be filtered into clusters of categories of station types; and ■ Direct operating costs. These are costs for activities con- then allocates salary, operating, and capital resources to the sidered to relate directly to the operation of stream-gauging Case Study: Valuation and Costing of Stream-Gauging Networks in Canada    137 sites and the production and delivery of data and informa- distances, and other factors play an important role in deter- tion from those sites. These activities include the collec- mining costs for a southern versus northern location. This tion, compilation, computation, and approval of data; the table also indicates the percent contribution of costs catego- submission and verification of data to the national archive; ries to the total cost of running a network under different cli- regional archiving of data and information; the provision matic conditions. In the north, the average total operational of data, information, and services; and minor monitoring costs (including field visits) per station are nearly five times site maintenance. higher than they are in the south ($9,400 to $2,000). Here it ■ Indirect operating costs. These are costs for activities is important to note that the number of staff and the number considered to be attributable to the collection and produc- of stations operated in the northern and southern examples tion of data and that are distributed over the entire net- below are quite different. In the more southern regime, where work (that is, not attributable to a given monitoring site). most sites are accessible by vehicle, the operating costs rep- These activities include day-to-day program financial and resent 20 percent of the overall costs to run the program, human resource management, warehousing and inventory while in the north, they represent 39 percent. However, sal- control, ISO/QMS management, and training and computa- ary costs are actually comparable between the north and the tional software costs. south when normalized by the number of staff. Table A2.1.1 shows two different costing model estimates for the north and the south of Canada. Clearly, logistics, 138    Case Study: Valuation and Costing of Stream-Gauging Networks in Canada TABLE A2.1.1  Stream-Gauging Network Operation Costs in Northern and Southern Canada (US$) Southern Northern Southern Northern Canada (% of Canada (% of Cost type Canada (US$) Canada (US$) Item total costs) total costs) Direct costs 2,774,802 1,033,658 Total technician and supervisor salary and 48 33 benefits 1,153,003 1,213,419 Total operating costs including field visits, leases, 20 39 power, telemetry, shipping 148,492 58,430 Total data management (hydrological service) 3 2 salary and benefits 83,786 167,293 Office operation (furniture, supplies, phone, 1 5 occupational safety and health) 182,676 99,996 Station maintenance, repair, district freight, and 3 3 WSC building maintenance 91,126 21,753 Hydrometric software 2 1 102,755 24,233 Training 2 1 5,460 20,280 Direct travel for meetings 0 1 82,476 34,377 Direct life-cycle management for goods worth 1 1 less than $10,000 (operating) 250,851 78,640 Direct life-cycle management for goods worth 4 3 more than $10,000 (capital) Total direct 4,875,427 2,752,079 84 88 costs Indirect costs 869,392 313,337 Indirect salaries and benefits (day-to-day 15 10 program financial and human resource manage- ment; warehousing and inventory control; ISO/ QMS management; training and computational software costs) 3,900 7,800 Staff, relocation 0 0 12,413 4,532 Office furniture, supplies, phone 0 0 12,785 4,668 Hydrometric software 0 0 7,800 Other travel 0 0 47,000 42,900 Indirect travel 1 1 Total indirect 945,490 381,036 16 12 costs Total costs 5,820,916 3,133,115 100 100 Source: Data collection of Alain Pietroniro, WSC (Canada). Note: In southern Canada, there are 573 stations, 61 staff, and 7 visits per station on average. In northern Canada, there are 129 stations, 20 staff, and 5 visits per station on average. ISO = International Organization for Standardization; QMS = Quality Management System; WSC = Water Service Canada. Case study: Network Modernization of Russia’s Kuban River Basin    139 Case study: Network Modernization of Russia’s Kuban River Basin 1 Annex 2.2 The modernization of an extensive hydrological observation system often requires enormous investments that exceed a country’s economic capabili- ties. In such cases the prioritization of river basins and stream gauges with- in these basins becomes necessary. This case study provides an example of how the consideration of affordability requires a cost-benefit analysis based on knowledge of hydrological conditions and socioeconomic needs. Kuban river basin. Photo: Garmasheva Natalia In the last century, the former Soviet Union developed an exemplary hydro- logical observation system. The state of Russia’s hydrometric network used to be near to the optimum in the early to mid-1980s, but it began to deteriorate in the 1990s. By the early 2000s, the network’s level of instrumentation and software no longer reflected the modern state, scoring a Level 2 on the 1–4 scale that was introduced in chapter 2.6. Since 2005, two National Hydromet System Modernization Projects have been developed and implemented in Russia, supported with loans of $80 million and $60 million, respectively, from the World Bank with the objectives of modernizing the measurement, transmission, and processing equipment and the software of the Russian Federal Service for Hydrometeorology and Environmental Monitoring. Training programs have been developed and implemented as well. Because the territory of the country is large and would require huge invest- ments to modernize the entire network, it was decided to prioritize the river basins (see section 2.6.2) and modernize them a few at a time. This prioriti- zation was driven by socioeconomic needs for hydrological data and infor- mation. It is an example of how considering the affordability of modernized networks requires a cost-benefit analysis based on knowledge of hydrological conditions and human needs. Within the framework of these modernization projects, the hydrometric network of the Kuban River Basin was also modern- ized. The Kuban River begins in the northern slope of the Caucasus Mountains in southern Russia and flows north and west for 870 kilometers to the Sea of Azov. The basin was chosen for modernization after an analysis of the specific flood damage per square kilometer (figure A2.2.1). Map A2.2.1, which shows the Kuban River Basin (with over 57,900 square kilometers), pinpoints the locations of manual gauges before the modernization and automatic gauges that were installed during RosHydromet-1. 140    Case study: Network Modernization of Russia’s Kuban River Basin The Kuban River Basin was selected as a top-priority basin areas) and riverine floods in the middle reach of the rivers. because of extensive water sector development, frequent In the middle and low reaches of the Kuban River there are hydrological adverse impacts due to river flooding, and a water resources management facilities (for example, series of relatively high population density. The basin is affected by reservoirs, irrigation canals). significant flash floods in upper catchments (mountainous FIGURE A2.2.1  Assessment of the Basins of Russia in Terms of Specific Flood Damage Sakhalin and Kamchatka rivers Amur Baikal basin rivers Eastern Siberian rivers Region/river basin Western Siberian rivers Ural Northern Caucasus rivers Kuban Don Volga North and north-west rivers $0 $1,000 $2,000 $3,000 $4,000 $5,000 $6,000 $7,000 Mean annual speci c ood damage, in US$/km Source: Adapted from Dobroumov and Tumanovskaya 2002. Project planning began with an evaluation of the current status Upper part of the basin (region 1 in map A2.2.1). This area is of the hydrometric system and of ways to develop and modern- mountainous and subject to flash floods. The main user of hy- ize it, given the key structural, technical, and target indicators. drological information and products is the National Disaster Management Agency (NDMA). Stream-gauging stations in the The optimum location of the basic network was decided area are intended to provide data to flood forecasting and well in advance (through regression analysis), based on the early warning system’s rainfall-runoff models to issue flood requirements of spatial and linear interpolation of the hy- warnings. To support flash-flood early warnings, the following drological regime characteristics. Taking into account the modernization activities were implemented: various water users in the basin, the optimum location of the operational stream-gauging network was studied to meet ■ A total of seven gauges were established in the outlets modern users’ requirements. Data from operational gauges of small foothill rivers with an elevation of more than were intended for flood warning and forecast services, as well 1,500 meters (in locations where gauges were previously as for the operation of numerous reservoirs and canals. closed). The locations of the gauges were selected based on the flood generation areas in upper catchments, taking To tackle the operational network redesign, the whole river into account the requirements of the rainfall-runoff model. basin was divided into several parts according to the hydro- Furthermore, a number of meteorological, snow, and pre- logical and geomorphological conditions, economic develop- cipitation measurement sites were established in areas of ment, and water sector demand. high elevation. Case study: Network Modernization of Russia’s Kuban River Basin    141 ■ These newly installed gauges, as well as all manual gaug- flood forecasts determined by the lag time of a flood wave. es (which existed before), were equipped with automatic Stream-gauging stations were also set up to meet the objec- water-level sensors, intended to measure the water level tives of water-level monitoring and control, thereby ensuring with variable frequency—from 5 to 10 minutes during reliable calculations and the forecasting of inflow reservoirs. heavy rainfall events to several times a day during low- Furthermore, additional automatic gauges were installed flow periods. along the main river reach, and mobile laboratories (trucks ■ At the same sites, automatic meteorological stations were and minivans with acoustic Doppler current profiler equip- installed. ment) for discharge measurements were introduced. ■ High-altitude stations were equipped with satellite or radio communication facilities to ensure timely data pro- Lower part of the river basin (region 3 in the map). This area vision to data collection and processing centers. is intended for irrigation and some navigation. The hydrologi- cal regime is determined by discharges from the dam, as well Middle part of the basin (region 2 in map A2.2.1). This area is as estuarine tidal and storm surge processes. The estuarine actively used by hydropower, water supply, and irrigation for area of the river is equipped with three estuarine gauges and agricultural needs. In terms of flood forecasting, flood routing a hydrometric estuarine office to control and support estua- models were applied. Additional gauges along the river reach rine measurements. were installed to meet the requirements of the lead time of MAP A2.2.1  Modernization of the Kuban River Stream-Gauging Network Legend Operational gauges in 2005 New gauges installed in 2010- 2012 as part of RosHydromet Modernization Project - 1 Lakes and reservoirs Rivers Source: Original map for this publication. 142    Case study: Network Modernization of Russia’s Kuban River Basin Other modernizing activities included monitoring rationaliza- The bottom line is that, as a result of the modernization tion programs for optimized hydrometric network and mod- project, the observation system gained features of Level 4 ernized equipment as well as automation of the collection (according to section 2.6.1). Looking ahead, further improve- and transmission of hydrological data, data processing, and ments are needed in the flood forecasting system, which has dissemination to users, applying GIS-web technologies. to be based on modern data and information. Case Study: Planning Approach to NHMS Leveling Up in Sri Lanka    143 Case Study: Planning Approach to NHMS Leveling Up in Sri Lanka 1Annex 2.3 A planning approach that reflects some World Bank guidelines for leveling up national hydrometeorological services (NHMSs) is the Climate Resilience Multi-Phase Programmatic Approach (CResMPA) for Sri Lanka. It takes into account lessons learned from previous projects, and it establishes a new approach to planning, installs new ways of procuring modern technologies domestically, and stresses developing human capacities. The long imple- mentation period provides an opportunity for interim assessments and ad- justments by the program’s steering committee. Sri Lanka. Photo: Artfoliophoto Within the framework of a CResMPA, the inception phase for planning Sri Lanka’s hydrological program officially began in 2019. It focused on deliver- ing a broad understanding of the current institutional and technical situation within Sri Lanka and determining how best to develop a flood forecast and water management monitoring network and operations center. This phase also looked at the institutional, legislative, and operational challenges that would need to be overcome for the scheme to succeed. The program will cover 25 basins across Sri Lanka—over 75 percent of the country (whose total area is 65,610 square kilometers)—with an initial pilot scheme for five basins. The first phase provided a detailed assessment of the potential for the implementation of the flood forecasting and water manage- ment operations center. One of the main requirements of the government of Sri Lanka was the long-term success and the sustainability of the institutional and technical solutions. The main reasons that previous projects in this area had failed to achieve this goal were technical, financial, and a based on the limited staff skill structure. Problems included: ■ Costly equipment procured from international vendors ■ Lack of management planning and repair planning for the sensors after they had been deployed ■ Incomplete or unfinished software for monitoring ■ Inconsistent approach to influencing the government (the “Treasury”) for greater funding ■ Lack of adequate staff training and post-project planning (a “set-and-for- get” attitude) 144    Case Study: Planning Approach to NHMS Leveling Up in Sri Lanka ■ Failure of project implementation to adequately address The government is now reviewing the results and recommen- Annex 2.3 the training and development gaps between the project dations in terms of how the reforms would work on a five-ba- and the day-to-day operational requirements sin pilot study before approving them for extension to the ■ Recruitment and resourcing issues. overall system. Given Sri Lanka’s position as a lower-middle-income country, For equipment purchases, Sri Lanka is placing a special em- the government has requested that future programs and re- phasis on the use of national suppliers. One of its key consid- views focus on eliminating these weak points. This requires erations is to ensure there would be an opportunity to get the establishing partnerships with domestic manufacturers to new equipment and software from local suppliers and reduce avoid costly imports of equipment and ensure national repair reliance on international vendors. A previous project on res- and maintenance; setting up intergovernmental taskforces to ervoir and dam safety had involved the installation of some share data and improve cooperation; using open source in- 20 monitoring stations bought on the international market at formation, software, and development; implementing “citizen substantial cost (some of which was an importation cost). The science” philosophies to engage the broader communities in equipment functioned well, but there was little provision for taking ownership of the observation system; and providing operational repair and maintenance within the contract—and ongoing education and capacity building of staff, along with the cost of repairs and spare parts was so prohibitive that succession planning for the operations center. those monitoring stations (when they failed) could not be repaired. The CResMPA highlighted that the following steps would in- crease the viability of the flood forecast and water manage- The use of local companies to provide the equipment would ment operation center and associated systems: have several benefits for Sri Lanka from a skills and longevity point of view: the producers would be directly available for ■ Designing a modernized network. This should be based on technical and sales requests and able to offer training and a root-and-branch review of the existing stream-gauging support in the local language; training, development, and network (including re-survey and re-rating). In an effort to fault resolution could be achieved faster and with lower over- avoid duplication and gaps, the expansion of the network head costs; and it would enable the growth of a new industry, should not begin until the existing system is verified. There which would offer increased technical capacity (and over the should also be a risk/cost-based review of the observation long term, would promote competition between suppliers). network to prioritize areas of high risk/importance for The incoming offers demonstrated that the provision of such more costly installations. loggers and equipment from a local supplier was substantial- ■ Developing human resources. This should be done by im- ly cheaper than international comparisons. It is worth noting plementing a staff recruitment and development plan for that the acoustic Doppler current profiler (ADCP) compo- the next 10 years based on an in-depth review of available nent—as well as the radar sensors for these systems—would staff and skills matrix reports, establishing partnerships be imported, so there is a reliance on availability of parts that with local universities, and setting up degrees and certifi- are then assembled within the country. cation programs for staff learning. ■ Implementing of new technologies. Here the focus is With regard to Sri Lanka’s turbulent, fast-flowing, sedi- on data acquisition and management by developing sen- ment-laden watercourses, the development of its own mon- sors in cooperation with local suppliers; expanding the itoring stations is an ambitious project. But if the testing of Irrigation Department’s Materials and Resources division; prototypes is successful, it would signal substantial progress and implementing an open source policy for data and soft- in developing cost-effective hydrometric systems for devel- ware, where appropriate—supported by in-house expertise oping countries. and the release of funds for targeted software platforms to try to be budget neutral.    145 Part 2 References BC Ministry of Sustainable Resource Management. 2003. Hydrometric Business Review April 2003, Resource Information Department. http://www.geosci- entific.com/technical/tech_references_pdf_files/Water%20Quantity%20 Monitoring%20in%20BC.pdf. Dobroumov, B. M. and S. M. Tumanovskaya. 2002. “Inundations in Russian Rivers: Their Formation and Zoning.” Meteorologia i Gidrologia 12: 70–78. Dixon, Harry, Sophia Sandström, Christophe Cudennec, Harry F. Lins, Tommaso Abrate, Dominique Bérod, Igor Chernov, Nirina Ravalitera, Daniel Sighomnou, Vrangfoss staircase locks, Norway. Photo: naumoid and Florian Teichert. 2020. “Intergovernmental Cooperation for Hydrometry – What, Why and How?” Hydrological Sciences Journal, Special Issue: Hydrological Data: Opportunities and Barriers 2020. https://www.tandfonline.com/doi/full/10 .1080/02626667.2020.1764569. Environment Canada. 2014. Evaluation of the Hydrological Service and Water Survey: Final Report, April 2014. https://www.canada.ca/en/environment- climate-change/corporate/transparency/priorities-management/evaluations/ hydrological-service-water-survey.html. Garcés, María. 2018. Statement by H.E. Mrs. María Fernanda Espinosa Garcés, President of the 73rd Session of the UN General Assembly. December 3, 2018. https://www.un.org/pga/73/2018/12/04/cop24-global-leaders-forum/. Gardner, John, Martin Doyle, and Lauren Patterson. 2017. “Estimating the Value of Public Water Data.” Working Paper NI WP 17-05, Durham, NC, Duke University. https://nicholasinstitute.duke.edu/content/estimating-value-public-water-data. GFDRR and World Bank (Global Facility for Disaster Reduction and Recovery and World Bank). 2018. Assessment of the State of Hydrological Services in Developing Countries. Washington, DC: World Bank. https://www.gfdrr.org/sites/ default/files/publication/state-of-hydrological-services_web.pdf. Hallegatte, Stephane. 2012. “A Cost Effective Solution to Reduce Disaster Losses in Developing Countries: Hydro-Meteorological Services, Early Warning, and Evacuation.” World Bank Policy Research Working Paper 6058, World Bank, Washington, DC. https://openknowledge.worldbank.org/handle/10986/9359. Hamilton, Stuart. 2012. “The 5 Essential Elements of a Hydrological Monitoring Programme.” WMO Bulletin 61 (1): S26–32. Hanemann, W. Michael. 2006. “The Economic Conception of Water.” In Water Crisis: Myth or Reality? Edited by Peter P. Rogers, M. Ramon Llamas, and Luis Martinez-Cortina, 61–91. London: Taylor and Francis. Lins, Harry F. 2008. “Challenges to Hydrological Observations.” WMO Bulletin 57 (1). 146    Part 2 References Mishra, Ashok K. and Paulin Coulibaly. 2009. “Developments in WMO (World Meteorological Organization). 2008. Guide to Hydrometric Network Design: A Review.” Reviews of Geophysics Hydrological Practices, Volume I Hydrology: From Measurement 47 (2): S1491. DOI: 10.1029/2007RG000243. to Hydrological Information. WMO-No. 168, 6th edition. Geneva: WMO. https://library.wmo.int/index.php?lvl=notice_ Nixon, S. C. 1996. ”European Freshwater Monitoring Network display&id=21815#.Yw5rO-zMIqt. Design.” Topic Rep. No. 10/1996, European Environment Agency, Copenhagen. WMO (World Meteorological Organization). 2009. Guide to Hydrological Practices, Volume II: Management of Water Rodda, John C. 1995. “Guessing or Assessing the World’s Water Resources and Application of Hydrological Practices WMO-No. Resources?” Water and Environment Journal 9 (4): 360–68. 168, 6th edition. Geneva: WMO. https://library.wmo.int/index. DOI: 10.1111/j.1747-6593.1995.tb00953.x. php?lvl=notice_display&id=543#.YijSyYnMKUk. Rodda, John C., Serge A. Piens, Naginder S. Sehmi, and WMO (World Meteorological Organization). 2010. Manual on Geoffrey Matthews. 1993. “Towards a World Hydrological Cycle Stream Gauging, Volume 1, Field Work. WMO-No. 1044. Geneva: Observing System.” Hydrological Sciences Journal 38 (5): S. WMO. https://library.wmo.int/doc_num.php?explnum_id=219. 373–78. DOI: 10.1080/026266693099492687. WMO (World Meteorological Organization). 2012. Climate and Ruddell, Benjamin Lyle and Thorsten Wagener. 2015. “Grand Meteorological Information Requirements for Water Manage- Challenges for Hydrology Education in the 21st Century.” ment. WMO-No. 1094. Geneva: WMO. https://library.wmo.int/ Journal of Hydrologic Engineering 20 (1): A4014001. DOI: index.php?lvl=notice_display&id=12760#.YjiDajUxnql. 10.1061/(ASCE)HE.1943-5584.0000956. WMO (World Meteorological Organization). 2015. Valuing Salman, Salman M. A. and Daniel D. Bradlow. 2006. Weather and Climate: Economic Assessment of Meteorological Regulatory Frameworks for Water Resources Management: and Hydrological Services. WMO-No. 1153. Geneva: WMO. A Comparative Study. Law, Justice, and Development. https://library.wmo.int/index.php?lvl=notice_display&id=17225#. Washington, DC: World Bank. https://openknowledge. YxjZWuzMIqt. worldbank.org/handle/10986/7054 License: CC BY 3.0 IGO. WMO (World Meteorological Organization). 2020. Hydro UN DESA (United Nations Department of Economic and Dashboard Survey. https://hydrohub.wmo.int/en/hydrology-dashboard. Social Affairs). 2015. International Decade for Action “WATER FOR LIFE” 2005–2015. UN WATER. https://www.un.org/ WMO (World Meteorological Organization). 2021a. waterforlifedecade/water_and_sustainable_development. Manual on the WMO Integrated Global Observing System, shtml. Annex VIII to the WMO Technical Regulations, WMO-No. 1160, 2021. Geneva: WMO. https://community.wmo.int/ USGS (United States Geological Survey). 2007. Qualitative rolling-review-requirements-process. Comparison of Streamflow Information Programs of the U.S. Geological Survey and Three Non-Federal Agencies. USGS WMO (World Meteorological Organization). 2021b. Country Open-File Report 2007–1426. https://www.usgs.gov/publi- Hydromet Diagnostics: Prototype for Road-Testing. August 6, cations/qualitative-comparison-streamflow-information-pro- 2020. Geneva: WMO. https://library.wmo.int/index.php?lvl=no- grams-us-geological-survey-and-three. tice_display&id=21906#.YQEQkkBCTyQ. Walker, Susan. 2000. “The Value of Hydrometric Information World Bank. 2018. Assessment of the State of Hydrological in Water Resources Management and Flood Control.” Services in Developing Countries. GFDRR. https://www.gfdrr. Meteorological Applications 7 (4): S. 387–97. DOI: 10.1017/ org/en/publication/assessment-state-hydrological-ser- S1350482700001626. vices-developing-countries. WMO (World Meteorological Organization). 2006. Guidelines on the Role, Operation and Management of National Hydrological Services. Operational Hydrology Report No. 49, WMO-No. 1003. Geneva: WMO. https://library.wmo.int/doc_ num.php?explnum_id=4852.     147 PART 3. LEAD AUTHOR Brian F. Day Lead Meteorological Consultant, WBG; Chair Emeritus of the Association of the Hydro- Meteorological Industry (HMEI); Campbell Scientific CONTRIBUTING AUTHORS R. David Grimes Lead Meteorological Consultant, WBG; former WMO Recommendations for the Design of President and Head of the Meteorological Service of Canada, MSC Jeffrey Kavanaugh Sustainable AWS chapter HMEI; Associate Professor, University of Alberta Foeke Kuik Meteorological AWOS chapter HMEI Member – Campbell Scientific Pekka Utela Observation Networks Radar chapter HMEI Member – Vaisala Matti Lehmuskero Upper-Air chapter and Systems in HMEI Member – Vaisala David P. Rogers Lead Meteorological Consultant, WBG; former Chief Developing Countries Executive of the UK Met Office Joshua Campbell HMEI Member – Campbell Scientific Ashish Raval Senior Meteorological Consultant, WBG; HMEI Member – Synoptic Data PBC Uganda. Photo courtesy of V. Tsirkunov, WBG. 148    References    149 Part 3 Abbreviations AC alternating current AFTN aeronautical fixed telecommunications network AMHS Aeronautical Message Handling System AT air temperature ATC air traffic control AWOS automated weather observing system AWS automatic weather station BKN broken clouds BL background luminance BUFR Binary Universal Form for the Representation CAPPI constant altitude plan position indicator CAT I Category I CAT II Category II CAT III Category III CB cumulonimbus CONOPS concept of operations DH decision height DT dewpoint temperature ECP Engineering Change Proposal FAT Factory Acceptance Test FDCU Field Data Collection Unit FEW few clouds FO fiber optics FTE full-time equivalent GB gigabyte GBON Global Basic Observing Network GFDRR Global Facility for Disaster Reduction and Recovery GPS Global Positioning System GTS global telecommunication system HMEI Hydro-Meteorological Industry hPa hectopascal ICAO International Civil Aviation Organization ICT information and communication technology IFIs international financial institutions 150    Part 3 Abbreviations ILS instrument landing system IP internet protocol IPS information processing systems ISO International Organization for Standardization IT information technology JICA Japan International Cooperation Agency KVM keyboard/video/mouse L luminance LAN local area network LCD liquid-crystal display LDCs least developed countries MB megabyte METAR Meteorological Terminal Aviation Routine Weather Report MG met garden MHz megahertz MiD mid-point position of the runway MLS microwave landing system MOR meteorological optical range MSC Meteorological Services Canada MSL mean sea level MSS message switching system MWS manual weather station NGOs nongovernmental organizations NMHS National Meteorological and Hydrological Services NMS National Meteorological Service NMs network managing system NPA non-precision approach NWP numerical weather prediction O&M operations and maintenance OFZ obstacle-free zone OS operating system OSCAR Observing Systems Capability Analysis and Review Tool P pressure PC personal computer PSTN public switched telephone network PW present weather QA/QC quality assurance/quality control QMS Quality Management Systems QNH/QFE sea-level pressure/pressure at aerodrome elevation Part 3 Abbreviations   151 RCI replacement cost with installation RFI request for information RH relative humidity RVR runway visual range RWY runway SAT site acceptance test SAWS South African Weather Service SIDS small island developing states SLA service-level agreements SOFF Systematic Observations Financing Facility SR solar radiation STOP-END stop-end position of the runway TCO total cost of ownership TCOSF Total Cost of Ownership Summary Form TCP Transmission Control Protocol TCU towering cumulus TDZ touchdown zone TS thunderstorm UNDP United Nations Development Programme UNEP United Nations Environment Programme UPS uninterruptable power supply UTC Coordinated Universal Time VIS visibility WD wind direction WIGOS WMO Integrated Global Weather Observing System WIS WMO Information System WMO World Meteorological Organization WS wind speed ZAMG Austrian Meteorological Service All dollar amounts are US dollars unless otherwise indicated. Unless otherwise noted, the sources of all figures are the authors of this publication. 152    Part 3 Glossary Term/Abbreviation Description Aerodrome A defined area on land or water (including any buildings, installations, and equipment) intended to be used either wholly or in part for the arrival, departure, and surface movement of aircraft. Aeronautical meteorological A station designated to make observations and meteorological reports for use in station international air navigation. Aeronautical fixed The aeronautical fixed telecommunications network is a worldwide system of aeronautical telecommunications network fixed circuits provided, as part of the aeronautical fixed service, for the exchange of (AFTN) messages and/or digital data between aeronautical fixed stations having the same or compatible communications characteristics. Air temperature The temperature indicated by a thermometer exposed to the air in a place sheltered from direct solar radiation (degree Celsius, °C). Altitude The vertical distance of a level, a point, or an object considered as a point, measured from mean sea level (MSL). Atmospheric pressure Pressure (force per unit area) exerted by the atmosphere on any surface by virtue of its weight; it is equivalent to the weight of a vertical column of air extending above a surface of unit area to the outer limit of the atmosphere (hectopascal, hPa). CAPPI The constant altitude plan position indicator, better known as CAPPI, is a radar display that gives a horizontal cross-section of data at constant altitude.  Ceilometer A ceilometer is an instrument for measuring the height of the base of a cloud layer, with or without a recording device. Measurement is done by calculating the return time of laser light pulses reflected by the cloud base. Cloud amount This is the fraction of the sky covered by the clouds of a certain genus, species, variety, layer, or combination of clouds. Cloud base The cloud base is the lowest level of a cloud or cloud layer (meter, m, or foot, ft). Elevation Elevation refers to the vertical distance of a point or a level, on or affixed to the surface of the Earth, measured from mean sea level. Forecast A forecast is a statement of expected meteorological conditions for a specified time or period, and for a specified area or portion of airspace. Forward scatter sensor This is an instrument for estimating extinction coefficient by measuring the flux scattered from a light beam by particles present in the atmosphere. Frangible Able to be broken into fragments. Ground truthing Ground truthing and the collection of ground-truth data on location enables calibration of remote-sensing data and aids in the interpretation and analysis of what is being sensed. Hot-standby Hot-standby is a term used for a computer that is ready to take over the operation of another computer within a short time period (for example, within seconds). Lightning detection network This is a network of lightning detectors transmitting in real time to a central computer that locates lightning flashes by combining information received from each detector. Part 3 Glossary   153 Term/Abbreviation Description Luminance (L) (photometric Luminance refers to the luminous intensity of any surface in a given direction per unit of brightness) projected area (candela per square meter, or cd/m2). METAR METAR is the name of the code for an aerodrome routine report; this is defined in WMO-306. It is a format for reporting weather information. A METAR weather report is predominantly used by aircraft pilots, and by meteorologists, who use aggregated METAR information to assist in weather forecasting. Meteorological information This refers to a meteorological report, analysis, forecast, and any other statement relating to existing or expected meteorological conditions. Meteorological office This is an office designated to provide meteorological service for international air navigation. Meteorological report A meteorological report is a statement of observed meteorological conditions related to a specified time and location. MID This is the MID position of a runway. Meteorological optical range The meteorological optical range is the length of the path in the atmosphere required (MOR) to reduce the luminous flux in a collimated beam from an incandescent lamp, at a color temperature of 2,700 K, to 0.05 of its original value, the luminous flux being evaluated by means of the photometric luminosity function of the International Commission on Illumination (CIE) (meter, m, or kilometer, km). The relationship between meteorological optical range and extinction coefficient (at the contrast threshold of ε = 0.05) using Koschmieder’s law is: MOR = − ln (0.05)/σ ≈ 3/σ. MOR = visibility under certain conditions (see Visibility). Observation (meteorological) This refers to the evaluation of one or more meteorological elements. Obstacle-free zone (OFZ) An obstacle-free zone is the airspace above the inner approach surface, inner transitional surfaces, and balked landing surface and that portion of the strip bounded by these surfaces, which is not penetrated by any fixed obstacle other than a low-mass and frangibly mounted one required for air navigation purposes. Okta In meteorology, an okta is a unit of measurement used to describe the amount of cloud cover at any given location. Sky conditions are estimated in terms of how many eighths of the sky are covered in cloud, ranging from 0 oktas (completely clear sky) through to 8 oktas (completely overcast). In addition, in the SYNOP code there is an extra cloud cover indicator “9” indicating that the sky is totally obscured (that is, hidden from view), usually due to dense fog or heavy snow. Precipitation intensity Precipitation intensity indicates the amount of precipitation collected per unit time interval. It is expressed as light, moderate, or heavy. Each intensity is defined with respect to the type of precipitation occurring, based on rate of fall. Present weather Present weather is the weather existing at a station at the time of observation. Present weather sensor This sensor measures physical parameters of the atmosphere and calculating a limited set of present weather, always including present weather related to precipitation. 154    Part 3 Glossary Term/Abbreviation Description Prevailing visibility The prevailing visibility is the greatest visibility value, observed in accordance with the definition of visibility, which is reached within at least half the horizon circle or within at least half of the surface of the aerodrome. These areas could comprise contiguous or non- contiguous sectors (meter, m, or kilometer, km). Note: This parameter may be assessed by human observation and/or instrumented systems. When instruments are installed, they are used to obtain the best estimate of the prevailing visibility. QFE Atmospheric pressure at aerodrome elevation (or at runway threshold) (hectopascal, hPa). QNH Altimeter subscale setting to obtain elevation when on the ground (hectopascal, hPa). Reliability This refers to the ability to perform with correct and consistent results (such as mean time between failure for equipment). Repeater A repeater is a device that links two computer networks on Open Systems Interconnection (OSI)-layer 1 (physical layer). Report A report is a meteorological code message in a standardized format. It is defined in the WMO Manual on Codes (WMO-No. 306, available at https://library.wmo.int/index.php?lvl=notice_ display&id=13617#.YnM4kNrMIuU). Runway A runway is a defined rectangular area on a land aerodrome prepared for the landing and take-off of aircraft. Runway strip A runway strip is a defined area that includes the runway and stopway, if provided, intended: ■ to reduce the risk of damage to aircraft running off a runway, and ■ to protect aircraft flying over it during take-off or landing operations. Runway visual range (RVR) The runway visual range is the range over which the pilot of an aircraft on the center line of a runway can see the runway surface markings, or the lights delineating the runway or identifying its center line (meter, m). SPECI An Aviation Selected Special Weather Report (SPECI) is issued when there has been a significant change from the METAR such as significant changes to wind direction, visibility, weather phenomena, ceiling height, or severe weather conditions. It is defined in the WMO Manual on Codes (WMO-No. 306). SYNOP Surface Synoptic Observations. Defined in WMO 306. Transmission Control A Transmission Control Protocol (TCP) is a transport protocol that is used on top of internet Protocol (TCP) protocol (IP) to ensure reliable transmission of packets. It includes mechanisms to solve many of the problems that arise from packet-based messaging (such as lost packets, out-of- order packets, duplicate packets, and corrupted packets). Touchdown zone (TDZ) The touchdown zone is the portion of a runway, beyond the threshold, where it is intended that landing airplanes first contact the runway. Threshold The threshold is the beginning of that portion of the runway usable for landing. Transmissometer A transmissometer is an instrument that takes a direct measurement of the transmittance between two points in space—that is, over a specified path length or baseline used for measuring visibility or the transmission of light in the atmosphere. Part 3 Glossary   155 Term/Abbreviation Description Visibility (VIS) Visibility for aeronautical purposes is the greater of: 1. The greatest distance at which a black object of suitable dimensions, situated near the ground, can be seen and recognized when observed against a bright background; 2. The greatest distance at which lights in the vicinity of 1,000 candelas can be seen and identified against an unlit background. Note: The two distances have different values in air of a given extinction coefficient, and the latter (2) varies with the background illumination. The former (1) is represented by the meteorological optical range (MOR). Weather radar This is an adaptation of radar for meteorological purposes. The scattering of electromagnetic waves, at wavelengths of a few millimeters to several centimeters, by raindrops and cloud drops is used to determine the distance, size, shape, location, motion, and phase (liquid and solid), as well as the intensity of the precipitation. Another application is in the detection of clear-air phenomena through scattering by insects, birds, and so on, and fluctuation of the refractive index. Australia. © tracielouise | istock.com    157 Part 3 Executive Summary Time for a New Approach It has long been recognized that quality meteorological observation systems are the backbone of accurate weather forecasts by National Meteorological Services (NMSs) in both the developed and developing worlds. Data from the four most common ground-based systems that are purchased or acquired by Norway. Photo: © Richard Semik | Dreamstime. developing countries—(1) automated weather observing systems (AWOSs, also referred to as AWOS systems); (2) weather stations—both automatic weather stations (AWSs) and manual weather stations (MWSs); (3) upper-air systems; and (4) weather radars—are critical in preparing for the effects of severe weather events by providing advanced warnings that allow govern- ments to manage the risks to life and infrastructure. Such data also benefit many sectors (including agriculture, forestry, hydrology, transportation, and com_170027224 renewable energy) that contribute to economic growth and social welfare. A major problem for developing countries is that these systems are costly to purchase and maintain over their operational lifetimes. For that reason, the World Bank and other international financial institutions (IFIs) have pro- vided loans and grants that total in the hundreds of millions of US dollars collectively over the past two decades. The focus has been on reducing the “Data from the four most high capital cost of these systems. But, too often, the investment of capital common systems that are has failed to yield the anticipated benefits, and the acquired system has fall- en into disrepair, owing to a shortfall of annual operations and maintenance purchased or acquired by (O&M) funding. As a result, these systems are not reaching their operational developing countries … lifetime goals of 10 to 20 years in developing countries. are critical in preparing This part of the publication is intended to be an aid to the financial deci- sion-making process for governments in developing countries and their re- for the effects of severe spective NMSs, the World Bank, and other IFI staff, and potential suppliers of weather events by providing equipment and services. It highlights the need to understand the total cost of ownership (TCO) over the life cycle of these four systems, gives recommen- advanced warnings that dations for the design of sustainable networks for each of these systems, and allow governments to provides background information on estimating the TCO prior to the country making the capital investments. It also adds value by drawing on data from manage the risks to life and five leading NMSs from developed countries to create benchmark values on infrastructure. Such data procurement, O&M, and consumables in order to set costing expectations of these activities at an appropriate level. also benefit many sectors … that contribute to economic growth and social welfare.” 158    Part 3 Executive Summary How to Plan a Meteorological Observation failure and lag time to repair, its availability or uptime is System 6,400/8,000 = 80 percent. For a newly acquired system, an NMS needs to commit to uptimes of 90 percent; as they gain Prior to beginning the planning process, the NMS must set experience, this should be increased to 95 percent or better. and commit to a goal for the desired level of operation over A system that is operating at 90 percent can be considered to the system’s lifespan. One of the most crucial and simplest be marginally sustained. Uptimes below 90 percent are not metrics to measure sustainability that is applicable for all yielding the necessary data flow to support the design goal of observation networks is called uptime or availability—that is, improved routine or severe weather forecasts. the percentage of time that a system is fully operational. It is common for many mature NMSs to reach 95 percent to Throughout the report we discuss operational effectiveness 98 percent (table ES3.1a). For a developing country, however, of a system or network in terms of its uptime or downtime. it may be more realistic to target a minimum of 90 percent Downtime is the reciprocal of uptime, so if a system has an and commit to increasing the percentage of uptime as person- uptime of 95 percent its downtime is considered to be 5 nel become more proficient at maintenance and repairs and percent. as the agency’s knowledge base and operational experience grows. It should be noted that, while some developing coun- For example, if an AWS is expected to deliver 8,000 data tries are achieving uptimes of 90 percent or above, many are values (total from all parameters) over a 30-day period but falling far short of this benchmark and, therefore, are not re- instead provides only 6,400 values because of component ceiving the full benefit from their systems (table ES3.1b). TABLE ES3.1  Station Uptimes for Developed and Developing Countries   AWS Radar Upper Air Sites Sites Sites Country (number) Uptime (%) (number) Uptime (%) (number) Uptime (%) a. Developed Countries: Station Uptimes Reach 95–98% Australia 701 99% 34 98% 38 98% Austria 270 100% 5 80% 1 100% Canada 585 96% 30 95% 30 90% Estonia 107 95% 2 95% 2 95% Germany 997 98% 18 97% 10 99% Hungary 292 90% 2 95% 4 95% Norway 256 97% 11 97% 6 96% United Kingdom 300 97% 15 98% 7 90% b. Developing Countries: Station Uptimes Fall Far Short of Even 90% Argentina 14 50% 7 50% 3 95% Colombia 365 50% 4 80% 8 80% Ecuador 105 20% 5 30% 3 30% Jamaica 96 95% 1 95% 1 95% Paraguay 80 90% 1 60% 1 — South Africa 236 88% 10 92% 11 92% Source: Data provided by NMS to the GFDRR of the World Bank. Note: Austria targets 100 percent uptime for radar (20 percent due to one radar system failure in 2020). Columbia targets 80 percent uptime for AWS (50 percent in 2020 due to COVID issues). AWS = automatic weather station; GFDRR = Global Facility for Disaster Reduction and Recovery; NMS = National Meteorological Service; — = not available. Part 3 Executive Summary   159 To ensure that meteorological observation systems are sus- topography, and staffing at the stations (for regular tainable in developing countries, planning the systems prior weekly or semiweekly station maintenance). to acquiring them is a critical first step. The 10 elements of 5. Developing a network data management plan. The plan the planning stage include: should include protocols for data backup, data securi- ty, quality assurance/quality control (QA/QC), analysis, 1. Purpose. The first and most important task is to clearly transmission/distribution, archival storage, and the define the purpose of the weather observation network sharing of data globally. This should be done prior to or system, as this determines what information is need- construction. Possible questions might be: Will data ed, when and how often, in what site locations, and to storage take place in a physical location managed by the what level of uncertainty. A key question might be: Will NMS or by an online (cloud) data service? Will multiple it increase public safety by providing more complete data centers in different geographical locations be used weather information to aviators, or by giving earlier for data collection and backup of systems? warning of approaching severe weather? Once the pur- 6. Defining infrastructure requirements. The plan must pose is identified, the NMS should perform a gap analy- define infrastructure requirements (civil works) includ- sis to determine what stations or systems are needed to ing utilities, site preparation, station mounting plat- complement or replace existing infrastructure. forms, necessary buildings or other structures, land 2. Identifying the necessary meteorological observa- acquisition, and security fencing. Additionally, consider- tions. Understanding the data needs clearly will inform ation should be given to the size and type of the central the type of measurements required. In some cases, nec- data processing systems to receive, process, transmit, essary weather information might include near-surface and store the data. air temperature, relative humidity, barometric pressure, 7. Defining operational requirements. Regardless of ini- precipitation, and wind speed and direction at a number tial cost or specification, all weather monitoring equip- of locations; in other cases, weather radar observation ment requires periodic maintenance, recalibration, and might be necessary. replacement. Thus, a well-defined maintenance strategy 3. Defining the requirement specifications. Define the is critical to ensuring the sustainability of an observa- required specifications based on global standards that tion network. It is recommended that operational re- ensure the data are of sufficient quality and fit for pur- quirements be planned for the duration of the expected pose. Specifications have a large impact on the capital system or network life cycle. The plan should include cost of a system (see table ES3.2), as the more accurate regularly scheduled (preventive), unscheduled (correc- the system or measurement, the more expensive. More tive), and upgrade (adaptive) maintenance plans. importantly, specifications set the foundation for qual- 8. Identifying personnel requirements. No network will ity data fit for purpose that support the purpose of the function properly without the efforts of dedicated, qual- observation system. In every case, reliable and accurate ified personnel. Personnel install, maintain, and repair observations provide the foundation upon which all in- stations; they manage, maintain, and perform QA/QC on terpretations and forecasts are built and why the initial incoming data; and they ensure that the gathered data capital investment will be made. are made available for analysis and to produce data 4. Designing the observation network plan. This includes products. consideration of optimal spacing of stations, number 9. Determining the most effective business model to of stations, location of stations, design of the station, support network operations. There are three business site accessibility, and reliable data communications. For models that NMSs may consider employing to achieve example, spacing depends on the type and purpose of long-term operation of their network based on factors the network (what to measure and where) and several such as internal personnel technical capabilities and other factors—such as accessibility (which depends on available budget: roads or if the stations are far from the NMS or airports), p Government-owned and government-operated net- works. The NMS owns and operates the equipment in 160    Part 3 Executive Summary the network. This is considered the traditional meth- p Privately owned and privately operated networks od for an NMS to obtain the data it requires to support (also known as data as a service). The contractor owns its monitoring needs. In this model, the NMS is fully and operates the equipment in the network supply- responsible for the flow and quality of data. ing data to the NMS. This model should be considered p Government-owned and privately operated net- by an NMS if it does not have the capital finances to works. The NMS owns the network, but a large majori- acquire the network or the necessary in-house skill ty of its stations are operated by an outside contractor. set to carry out the activities required to sustain the This model should be considered by an NMS if it has network. the capital finances to acquire the network but might 10. Performing a cost analysis. Once all the components of not have the necessary in-house skill set to carry out the system are well understood and specified, the TCO the activities required to sustain the network, thereby can be developed that will inform on the required capi- ensuring quality data flow. tal and annual operational budgets. TABLE ES3.2  Measurement Type versus Accuracy and Cost for AWS Accuracy Class Description Type and cost A Measurement meets the WMO required measurement un- Reference climatological or research stations certainty and stated achievable measurement uncertainty, Higher that is Annex 1.A. B Measurements have a wider uncertainty than class A. Synoptic or AWOS stations Threshold C Specifications more relaxed than class B. Well-maintained public weather stations Lower D Wider than class C or no information is available. Crowd sourced weather stations Source: Based on Hartley, B. 2018. Data Collection Network Modernisation: What You Need to Know. Meteorological Service of New Zealand Limited. Note: Annex 1.A is in WMO 2018; see https://library.wmo.int/doc_num.php?explnum_id=10616. AWOS = automated weather observing system; AWS = automatic weather station; WMO = World Meteorological Organization. How to Calculate Total Cost of Ownership The TCO includes not only the initial capital investment of equipment but also an understanding of the annual costs The most common reasons that weather observing systems required to sustain the operation of the system throughout do not achieve their expected operational lifetimes involve its expected lifetime. Whether the meteorological observa- unrealistic expectations that result from a poor understand- tion system is being replaced because of its end of life or ing of the system’s true operating cost or the lack of a sustain- because a new system is being contemplated to enhance an ing budget and qualified staff to run a complex observation NMS’s capabilities, the principles of costing are similar. The infrastructure. The simplest way to minimize this risk is for four phases of managing an observing system—design; pro- the NMS to have as complete a knowledge as possible of all cure and install; operate, maintain, update; and end of life the costs associated with the capital investment and opera- (box ES3.1)—are also the same for both scenarios. Different tional funding. And the best way to do that is to calculate the approaches to the network design and operation may need TCO, which looks at the long-term affordability and financial to be weighed with a goal of minimizing the TCO while also requirements of acquiring, implementing, and operating an minimizing downtime to ensure consistent data flow. The TCO observation system or network designed in support of the can inform on decision-making when compared to available NMS’s data requirements. capital and operation budgets that may reduce the design of the system to fit with available funding. When no budget ex- ists, an analysis of the TCO can aid an NMS on what financial resources are required. Part 3 Executive Summary   161 Automated Weather Observing Systems (AWOSs). An AWOS BOX ES3.1  Managing a Meteorological Observing System is a complete and highly regulated system that includes au- tomatic weather stations, other sensors, and software on a Simplistically, there are four phases in the management server that produces and disseminates real-time meteoro- of an observation system that factor into the total cost of logical observations and reports during the period when an ownership (TCO). airport operates (based on arrivals and departures). While ■ Phase 1: Design. Starting with a comprehensive most of the flying public is typically unaware of their exis- system design (plan) is critical for the development tence, AWOS systems are critical to flight operations such as of the TCO and the sustainable operation of an landing approach or take off direction at an aerodrome, and observation system. to determine whether aircraft operations need to cease until ■ Phase 2: Procure and install. Each country has its conditions improve. own procurement procedures, and the installation of a system is unique to the circumstances and capabilities of a National Meteorological Service While AWOS systems can be complex in nature, based on the (NMS). However, every major acquisition project number of runways and runway configuration, both the World requires strong project management that takes it Meteorological Organization (WMO) and the International through all the steps of procurement to the ultimate Civil Aviation Organization (ICAO) impose a strict design commissioning of the station. structure. This results in a marked difference between AWOSs ■ Phase 3: Operate, maintain, update. Depending on and automatic weather stations, which are intended for gen- the business model the NMS chooses, the operate, maintain, and update (life cycle) phase is the eral weather forecasting and typically have more latitude in most expensive, given a lifetime of 10 to 20 years, design and specification. AWOS networks are part of a regu- depending on the observation system. For example, lated service mandated by ICAO, and AWOS maintenance is competitive staff remuneration and the ongoing costs typically supported through cost recovery from aircraft land- of training, spare parts, logistics to service sites, ing fees. regular maintenance and calibration of sensors, consumables (upper-air), and life-cycle management are often underestimated. When overlooked or Automatic Weather Stations (AWSs). An AWS provides re- not budgeted for properly, they cause a significant liable, accurate, and continuous observations of surface reduction in life expectancy of the system to the weather conditions without the need for human observers. To point where the system is no longer a benefit to the accomplish this, it employs sensors that convert meteorologi- NMS. cal conditions to electronic signals, which are then measured ■ Phase 4: End of life. Years before the observation and recorded by a data logger. Data from weather stations are system approaches the end of its useful operational life, the NMS will need to begin the design phase for used for a wide range of sectors—including real-time weath- its replacement. While the end-of-life phase is based er observations; forecasting of general weather and extreme on developing a new design, it does have financial weather events; agriculture; hydrology; climatology; and implications—such as the cost of disposing of the road, marine, and aviation transportation safety. old system, as well as other costs that need to be included in the TCO. In low-income countries, weather observations might be made through a combination of AWSs and manual weather stations (MWSs), as determined by cost, capacity, capability, Why the Key Weather Observing Systems Matter and government policy. Whether a given station is manually operated or automated can also depend on its application. The four key meteorological observations systems that this For example, a country might use a network of AWSs to sup- part of the publication explores in depth are those most com- port real-time synoptic forecasts and WMO data exchange re- monly purchased or acquired by developing countries. quirements but employ human observers to gather weather observations for agricultural operations or long-term records aimed at understanding climate variability and change. 162    Part 3 Executive Summary Upper-Air Systems. These systems produce observations Canada, Germany, and the United Kingdom) that operate suc- (also known as soundings or upper-air soundings) which are cessful, well-run networks with a good track record in main- considered a primary measurement of atmospheric condi- taining and sustaining their network operations over the long tions aloft. They provide a three-dimensional representation term. These NMSs do not all follow the same business model of the atmosphere for forecasters to evaluate the vertical approach, particularly for upper air and radars. Some have distribution of temperature, humidity, and wind—all critical automated upper-air systems, one contracts out staff opera- information for forecasting severe thunderstorms and torna- tions, and some have manual operations. For radars, one NMS does in the summer and winter storms in the winter. has a service arrangement to procure data, and some have C-band rather than S-band. Over land, vertical profiles of atmospheric conditions are typically obtained using radiosondes launched from dedicat- Our results show that, overall, the operating costs for the ed ground stations and carried aloft by gas-filled balloons. benchmark countries are very similar, providing some con- Radiosondes contain sensors for measuring upper-air envi- fidence about the estimated costs of supporting sustainable ronmental parameters, electronics for converting the sensor network operations (see table ES3.3). Furthermore, these output into digital signals, and a radio transmitter. Radio costs (without labor) indicate the scope of national budgets signals are received by an antenna and radio receiver at the required over the observation instrument/system life cycle— ground station, decoded, and converted to meteorological ranging from $6,200 for an AWS to $200,000 for an upper-air observations, which are compiled into sounding reports that station. The infrastructure for monitoring networks typically are transmitted to the NMS. Upper-air data are part of the has a fixed end-of-life timeframe. Ideally, the life cycle for WMO Global Basic Observing Network (GBON) and are shared well-maintained meteorological equipment and sensors rang- with the international community following WMO protocols es from 7 to 10 years, and up to 15 years for radars. The key and standards. factors affecting the life cycles of equipment are their robust- ness and the impacts of severe or extreme weather and cli- Weather Radar Systems. Over the past 30 years, the need mate events. Without a regimen of regular maintenance and for weather radars has grown significantly. While radar infor- sensor replacements, life spans can be shortened, resulting in mation is now beginning to be incorporated into numerical suboptimal returns on investment and system obsolescence. weather prediction (NWP) models to improve weather predic- tions, NMSs primarily use radars for nowcasting applications As for the developing countries, our results show that they and early warnings for rainfall intensity, flash flooding events, spend significantly less than developed countries on AWS, and dangerous thunderstorms. In colder climates, radars are upper-air, and weather radar systems (see table ES3.4). In important for distinguishing between rain events and snow fact, their outlays on O&M are about half of the developed events. They are capable of assessing, monitoring, and pre- countries’ benchmark. While some countries meet or exceed dicting characteristics of wind and shear zones. And they pro- the 90 percent uptime, on average, developing countries are vide localized, highly detailed, timely, and three-dimensional not attaining this minimum performance level, with some as sensing and observing capability that no other meteorologi- low as 50 percent—an indication that the NMS is not receiv- cal monitoring system can provide. ing value for the capital investment made. Aligning TCO Expectations through While the data presented in tables ES3.3 and ES3.4 do not Benchmarking include the full TCO over the observation system’s lifetime, it is clear that the differences in O&M spending result in a So, what does this part’s benchmarking exercise tell us about significant increase in how NMSs perform in developed coun- how much developed and developing countries are current- tries versus those in developing countries. If the full value ly spending on the most common observation systems? For from the capital investment in observation programs in the the developed countries, the exercise draws on information developing countries is to be realized, the first step needs to provided by five developed-world NMSs (Australia, Austria, be an understanding of the TCO. Part 3 Executive Summary   163 TABLE ES3.3  Summary Benchmark Cost for Observation Systems, Developed Countries Lifetime cost (O&M times expected Expected Annual O&M lifetime plus RCI) Sites Performance lifetime RCI (without labor) (without labor) Observation system (number) (uptime, %) (years) per site (US$) per site (US$) per site (US$) AWS 1 96% 10 $56,000 $6,200 $118,000 Upper-air             Manual 1 95% 20 $1,000,000 $200,000 $5,000,000 Automatic 1 95% 15 $715,000 $200,000 $3,715,000 Weather radar             S-band 1 95% 15 $3,900,000 $105,000 $5,475,000 C-band 1 97% 15 $3,000,000 $87,000 $4,305,000 Source: Data provided by NMS to the GFDRR of the World Bank. Note: Benchmark data are from 2020 and calculated based on information provided by Australia, Austria, Canada, Germany, and the United Kingdom. Labor costs have been omitted from this summary, as these costs are highly variable between the developed and developing countries—although they need to be considered when operating an observation system over its expected lifetime. O&M includes annual costs for spare parts, life-cycle management, logistics for site visits, and other costs. AWS = automatic weather station; EL = expected lifetime; GFDRR = Global Facility for Disaster Reduction and Recovery; NMS = National Meteorological Services; O&M = operations and maintenance; RCI = replacement cost with installation. TABLE ES3.4  Summary Costs for Observation Systems, Developing Countries Performance Expected lifetime Annual O&M (without labor) Observation system Sites (number) (uptime, %) (years) per site (US$) AWS 1 80% 10 $2,800 Upper-air 1 86% 20 $80,000 Weather radar 1 87% 15 $47,000 Source: Data provided by NMS to the GFDRR of the World Bank. Note: Data are from 2020 and averaged based on information that was considered to be the most reliable. Data for AWS are from eight upper-middle-income developing countries: the Dominican Republic, Guyana, Jamaica, North Macedonia, Panama, Paraguay, South Africa, and Surinam. Data for upper-air from five upper-middle-income developing countries: Colombia, the Dominican Republic, Panama, Paraguay, and South Africa. Data for weather radar are from six upper-middle-income developing countries: Guyana, North Macedonia, Panama, Paraguay, South Africa, and Surinam. No data were provided on replacement cost including installation from the developing countries. AWS = automatic weather station; GFDRR = Global Facility for Disaster Reduction and Recovery; NMS = National Meteorological Services; O&M = operations and maintenance. Recommendations improve in-country weather observation capabilities by sizing the system or network to match the NMS’s capa- In sum, although country needs will vary depending on local bilities and budget. One outcome of this approach would economic and geographical conditions, this part of the publi- be the development of a financial plan that provides an cation puts forward a number of recommendations for NMSs estimate of the TCO over the network’s expected lifetime that cut across the four key systems: that has the long-term support of the government and development partners. 1. Develop a full network plan up front. The government 2. Adopt a TCO approach. Prior to purchasing a new ob- and development partners should encourage and sup- servation system or network, the NMS, government, port the NMS in preparing a full network plan to promote and development partners need to consider not just the a thorough understanding of the TCO. This will help capital acquisition cost but also costs associated with 164    Part 3 Executive Summary installation, a human resource plan, and long-term oper- 4. Develop operational processes. Key to the ongoing suc- ation (including upgrades and equipment maintenance). cess of any observation system is long-term planning While this can represent a considerable effort for an NMS for operation processes that include regular training of pre-purchase, the benefits of understanding the TCO and staff, spare parts, maintenance and calibration of sen- the requirements for the system or network management sors routines, scheduling of consumables (upper-air) provide the foundation for long-term success. and spare parts purchases, upgrades to computer equip- 3. Appoint a project manager. Project management is crit- ment, software, and analysis of the incoming data as ical to the successful implementation of any meteoro- well as life-cycle management. Development and imple- logical observation system. The project manager is the mentation of processes that support the full life cycle of key focal point of communications between the supplier the observation system will ensure the successful oper- and the NMS and must have decision-making authority ation over its life expectancy to the long-term benefit of over the implementation phase. the NMS.    165 Overview of Sustainable Weather Observing Systems 3.1.1 Introduction 3.1 For capital investment in the four most common meteorological observation systems that are purchased or acquired by developing countries—(1) automat- ed weather observing systems (AWOSs, also referred to as AWOS systems); (2) automatic weather stations (AWSs), including manual weather stations (MWSs); (3) upper-air systems; and (4) weather radars—the National Meteorological Service (NMS) should consider fully the annual operational costs over the life cycle of the equipment. When this is not done, the network often fails to yield the anticipated benefits from investments supported by the national institu- tions, the World Bank, or other international financial institutions (IFIs), as the ongoing financial and staff resources allocated by the NMS are insufficient to maintain the equipment. Ensuring that networks are “right-sized” for both data requirements and annual operational budget constraints is a critical step to achieving sustainability. There is a strong need for a practical guide outlining United States. Photo: milehightraveler the scope and composition of major tasks and processes required to support the sustainable operation of weather monitoring networks. This part of the publication is intended to be an aid to the financial deci- sion-making process for governments in developing countries and their re- spective NMSs, the World Bank and other IFI staff, and potential suppliers of equipment and services. It highlights the need to understand the total cost of ownership (TCO) over the life cycle of these four systems, gives recommen- dations for the design of sustainable networks for each of these systems, and provides background information on estimating the TCO prior to the country making the capital investment. “This … publication is intended to be an aid to the As a first step, it is recommended that this chapter be read in its entirety to understand why meteorological observation systems and networks in many financial decision-making developing countries are not being maintained or operated sustainably over their expected lifetimes. Each of the four systems share common design prin- process for governments ciples (see chapter 3.2) that should be understood prior to their acquisition— in developing countries notably, the basic process of planning, designing, and operating a weather observation network that provides quality data “fit for purpose” over its ex- and their respective NMSs, pected lifetime. the World Bank and other Second, the reader should review chapter 3.3, Total Cost of Ownership, for a IFI staff, and potential summary of general cost considerations for a weather observation system or suppliers of equipment and network. The chapter includes a comprehensive design review to ensure that (1) data collected are fit for purpose and of quality suitable for their intended services.” 166    Overview of Sustainable Weather Observing Systems application that supports user requirements; (2) there is a airport must have four to six hours of accurate weather obser- plan for operating, maintaining, and growing the system or vations in advance of travel, as well as information on current network over its lifetime; and (3) the annual budget is suffi- weather conditions for the destination airport and at least one cient to ensure long-term sustainability. This holistic under- alternate destination prior to arrival. Failure to maintain this standing of what is required over the lifetime of the network is level of service can have negative impacts on the country’s key to sustainable operations. Thus, we strongly recommend economy and, ultimately, its public safety record. AWOS is working through the Total Cost of Ownership Exercise (chap- unique and more successful than the other three observation ter 3.8) prior to finalizing a network design and proceeding systems because of its high degree of regulation and the rev- with acquisition to ensure that the NMS derives the greatest enue it generates through landing fees levied against arriving value from its meteorological observation system(s). aircraft. As a result, the airport authority, or the NMS, priori- tizes operating AWOS systems to international standards and After reading chapters 3.1, 3.2, and 3.3 in this volume, the ensuring that the systems are adequately and actively fund- reader should review the chapter that discusses the meteo- ed, maintained, and upgraded by trained staff. In contrast, rological observing system(s) that the NMS or government AWS, upper-air, and weather radar systems and networks are is considering for introduction or modernization (AWOSs, not regulated and are developed using technical guidelines AWSs, upper-air systems, or weather radar systems). This or recommended best practices that typically do not outline chapter begins with a brief look at the main barriers to sus- the financial aspects of operations. Furthermore, these three tainability and then explores the key issues that cut across systems lack readily available cost recovery mechanisms, in- them and possible ways to address them. stead relying on government budget allocations for the long- term funding required to maintain sustainability. 3.1.2 Barriers to Sustainability Key takeaway: For AWOS, the combination of a high de- What are the main barriers to sustainable operation of weath- gree of regulation and ongoing cost recovery funding cre- er observation systems and networks? The following three ates an environment of more sustainable operations than examples point to different aspects of the problem for both seen in AWS, upper-air, and weather radar systems. The developed and developing countries. concept of a regulated service should be explored further to see how this concept could apply to other observing Example 1: Automated Weather Observing Systems systems. Automated weather observing systems (AWOSs) typically op- erate reliably over timescales that meet or exceed their ex- Example 2: South African Weather Service pected lifetimes—unlike AWS, upper-air observation systems, and weather radar systems, which often fall into disrepair well The South African Weather Service (SAWS) is one of the more before the end of their expected lifetimes. The longevity and advanced weather services in the developing world. It owns performance difference prompts an examination as to why. and operates a network of 236 AWS sites, 11 upper-air sta- tions, and a mix of X-, C-, and S-band weather radars (see In most countries (including least developed countries and table 3.7.1) totaling 15 systems. To improve the performance small island developing states), there is a strong economic of its weather observation platforms, it has identified several dependence on maintaining international aviation operations challenges that most affect its ability to maintain data flow for the transport of goods and services that are critical for from the networks it operates. Table 3.1.1 shows the 10 most economic success. To support international flight operations, significant challenges to sustainable operations. This list pro- a country’s airport and transportation authorities must meet vides insights into a range of issues faced by SAWS and most strict regulations defined by the International Civil Aviation other developing countries, and it can be distilled down to Organization (ICAO) and follow World Meteorological a single root cause: reduction of ongoing financial support. Organization (WMO) requirements for aviation weather ob- SAWS traditionally receives a third of its funding from its cost servations and reporting. Aircraft destined for an international recovery services to aviation, but this source was significantly Overview of Sustainable Weather Observing Systems   167 TABLE 3.1.1  SAWS Infrastructure Management Challenges in 2020 Item Infrastructure management challenges Reason 1 Budget Constraints – Limited Budget to cover all required expensive spare parts for Radar Systems Budget Maintenance. 2 Lack of skilled staff in the regions and high turnover of staff, trained staff retention is a problem. Budget Remuneration and compensation of technical staff risks skill attrition to our competitors. 3 Recruitment of skilled staff is too long and unsupported by lack of budget. Budget 4 Sustainable and reliable power source as well as back-up power sources impacts the performance figures Budget and uptime of the infrastructure. 5 Sustainable and reliable communication infrastructure impacts Meteorological Infrastructure Availability Budget Figures. 6 Procurement process for maintenance material is too long impacting badly on turnaround times and uptime Process of the infrastructure. 7 Lack of overtime and standby policy impacts on the morale of technical staff and regional technologists. Process 8 Lack of universally accessible computerized maintenance management system impacts the operations, mon- Process itoring and reporting on infrastructure performance. 9 Infrastructure vandalism on remote stations impacts uptime and performance figures, lack of security is the Security factor. 10 Covid 19 travelling restrictions impacted the uptime and response to maintenance issues on infrastructure Process including importation of spares and material from abroad (in 2020). Source: South African Weather Service. impacted during 2020 due to COVID (Item 10 in the table), on top of government cuts. As a result, it is now allocating Key takeaway: The challenges faced by SAWS highlight funds from the capital budget to maintain operations—an un- the importance of proper long-term stable financing and desirable and unsustainable situation, as funding for replace- planning in the success and sustainability of weather ment parts and life-cycle management of the systems is now observation systems. inadequate. The lack of consistent financial support makes it difficult for SAWS to maintain its systems at its pre-2020 Example 3: Developing Country NMS level of service. Another example centers in a developing country, where the World Bank is helping to modernize the country’s whole me- The SAWS example also points to the issues of retention and teorological system. The barriers to sustainable operations recruitment (Items 2 and 3 in table 3.1.1) of skilled staff. These faced by this country are similar to those generally seen all issues are shared throughout developing world NMSs. While over the world. The data presented here are based on a review partially budget related, the developing world does not have by the World Bank prior to being published and so the NMS can- the same pool of skilled talent to draw from that the devel- not be identified. The summary information as presented is in oped world has. Lack of skilled staff reduces the capacity of line with general industry knowledge of the barriers to sustain- the NMS to perform its mandate. To resolve this issue, NMSs able operations. in the developing world need to offer remuneration packages at the in-country market rate for skilled staff that are compet- The developing country NMS has operated five S-band ra- itive in order to recruit and retain employees. dars that were originally donated by the Japan International Cooperation Agency (JICA) for about 15 years. All radars were functional for the first 10 years of operation and were finan- cially supported and maintained by the NMS. Unfortunately, 168    Overview of Sustainable Weather Observing Systems after 10 years, the five radars deteriorated to the point that, no ability to analyze radar data now exists. This lapse in 2021, only one radar system was fully operational, two in human resources indicates the need for a long-term were partially functional, and two were nonfunctional. Of the recruitment plan with ongoing contracts to support a partially functional radars, one had data integrity issues and training program for technical and scientific staff. the other had a damaged power supply. The developing coun- p An absence of civil works maintenance. This problem try NMS clearly supported the radars for the first 10 years, is illustrated in the four panels of photo 3.1.1, which ensuring that there was a sufficient budget to carry out the show clear signs of corrosion and damage to the basic operations and maintenance of the systems. So why did the infrastructure of a radar site. radars not remain functioning for the full 15 years and start p Lack of upgrades to the radar software, analysis, to degrade 10 years into their operational life? A number of and data products. Current usage of the radar sys- reasons stand out: tems is limited to basic image visualization. Radars cost millions of dollars to purchase and operate, and ■ Lack of replacement parts. A radar should have a sustain- maximum value can be realized only by taking advan- able lifetime of 15 years and may be extended to a peri- tage of advancements in data analysis and products. od of 20 years if regular maintenance and upgrades are Planning over the 15-year lifetime or longer is vital if performed (see chapter 3.7 on Weather Radar). But a lack the true value of a radar system is to be realized. of replacement parts from the manufacturer severely hurt p Lack of consistent electricity and data connections. the developing country NMS's ability to maintain the ra- The operations of the radar(s) are considerably ham- dars—in turn, this led directly to the premature failure of pered by the lack of consistent power mains and data the systems. Manufacturers will discontinue the support connections. The process to “switch on” the radars is of existing systems because of a lack of component parts agreed between the National Storm Warning Centre or a change in technology. Thus, NMSs need to be aware and the regional radar site. But, owing to limited of the longevity of the systems they purchase or are gifted radar analysis and processing software, the radar im- and plan for part obsolescence to ensure continuous oper- ages are uploaded manually to the internet for view- ations over the lifetime. ing by the forecaster—a process that has a significant chance of introducing further delays and errors into The lack of parts and maintenance cascaded into: the severe storm monitoring and alerting process. Moreover, the integrity of the radar data is question- p A loss of staff and capability. There has been a lack of able. The radar software should be capable of export- ongoing recruitment and training of current staff be- ing the radar data into third-party products (such as cause many of the technical and scientific staff have nowcasting systems, forecaster workstations, warn- left the NMS through retirement. As a result, little or ing systems). Overview of Sustainable Weather Observing Systems   169 PHOTO 3.1.1  Images of a Radar Site a. Radar Pedestal and Antenna inside Radome b. Radome Panels c. Corrosion on Radome Base Ring Structure d. Water Leakage inside Radome Source: Photos provided by the World Bank. Key takeaway: To realize the full value of a weather radar 3.1.3 Key Issues system or network, an NMS should develop a comprehen- The benefits of establishing a weather observation network in sive operational plan for a minimum of 15 years or longer. a developing country brings value beyond public safety and The plan should include annual budget planning for ongo- extends well past the country’s borders—such as filling in the ing maintenance and upgrades to the system(s), stocking global data gaps that exist primarily in the developing world. spare parts, and human resource management (such as As a consequence, although the World Bank and other IFIs ongoing training). The plan should be regularly updated provide developing countries with funding to cover primar- every 5 years for the radar program to realize its full poten- ily the capital costs of purchasing equipment for increasing tial lifetime. Ongoing financial and manufacturer support weather forecasting or extreme weather warning capabilities, will be critical. 170    Overview of Sustainable Weather Observing Systems it is clear in cases where systems fail that NMSs do not plan as a necessary step, this can leave the NMS overwhelmed and convey the amount of financing required to maintain and by the amount of technology its staff needs to learn; the operate the newly established networks over their expected number of systems that require support; and the amount life cycles to the government finance department. Even in of data that needs to be collected, analyzed, stored, and the cases where governments promise to support these new compiled into weather forecasts and reports. It is often investments in their NMS (as in Example 2, SAWS), shifting unable to hire or train enough qualified staff to operate or priorities and challenges make these pledges difficult to keep. maintain the observation networks properly, or to manage the data. The lack of long-term financial support—which for AWS and ■ Uncoordinated external financing. The concept of siloed AWOS might mean 10 years, for weather radar 15 years, and or fragmented networks may be unintentionally supported for upper-air systems 15–20 years—is one of the major causes by external financing. There are many instances where, for of the lack of sustainability. This fact was prevalent in each of example, a developing country seeks support from various the three examples of barriers provided: developing partners. One IFI will support the NMS in ac- quiring a meteorological network for improving weather ■ Insufficient budget. The NMS does not have sufficient prediction; another will fund a hydrological monitoring funding in its regular budget to perform regularly sched- network for flood forecasting; and a third will support an uled site visits for system and infrastructure maintenance, AWS network for improving agricultural yields, leading to which leads to the deterioration and ultimate failure of a duplication of observations over small regional areas. equipment and subsequent loss of data. Key to the main- While there is overlap, these separate networks will have tenance component is securing staff to carry out the work different sets of equipment, making requisite maintenance and reducing employee retention issues by remunerating extremely challenging for the limited and often insuffi- them at competitive rates. The NMS is unable to perform ciently qualified staff. Typically, data in these networks regular sensor calibration, which is necessary to ensure are not shared among entities within the country, let alone that the data stream is of the required quality. It is un- outside the country—and, given the siloed approach, it is able to purchase needed spare parts that support both usual that these networks are not sustainable in the long the maintenance and life-cycle management programs if term. A countrywide monitoring strategy might help IFIs operations are to be maintained—such as for radar, which and donor countries to work together to solve not just the is a significant expense. It cannot purchase the necessary initial procurement of the network equipment but also to consumables required to carry out the program (such as develop the long-term funding that is required to support radiosondes and launch-support materials). If a budget sustainable operations. is available, internal procurement processes have long ■ Inadequate planning. There is a significant amount of lead times, resulting in parts and consumable shortages planning required prior to receiving an observation system and adding significantly to observation system downtime. so that the NMS understands how to implement and take Additionally, the NMS is often unable to cover the cost of full advantage of the technology over its operational life. electricity to power the systems or pay monthly communi- This needs to be done in partnership with the manufactur- cations fees. er, requiring discipline to implement scheduled planning ■ Sheer scope of tasks. The scope of tasks the NMS is ex- reviews. The clear goal of planning is to ensure that the pected to perform with or without additional financing NMS, in partnership with the manufacturer, has ongoing is significant. External resources in the form of loans or technical support, upgrades, and replacement parts avail- grants to strengthen the capabilities of the NMS to im- able over the system’s lifetime. prove its ability to forecast hazards are usually received by the government. This often involves upgrading obser- Together, these issues underscore that current development vation networks to support very short range forecasts, practices are not necessarily yielding the results expected at and it necessarily involves some automation because the the time of investment—and that it would be unwise to assume timescales are too short for manual operations. While seen that a different result would be achieved by continuing on the Overview of Sustainable Weather Observing Systems   171 same path. New strategies will require all parties (starting purposes and provides the basis, for example, for informing with governments, their respective agencies, and the IFIs) to hydrological forecasts, renewable energy project planning, rethink how programs are constructed and funded to develop agriculture and forestry, transportation safety, and basic sci- monitoring programs that are sustainable over the long-term entific inquiry. These additional uses would represent signifi- and give the countries more value-added. Governments and cant added values—potentially at little additional costs to the their agencies will need to take a more integrated approach to NMS for data homogenization and long-term quality control. monitoring to yield the highest data value in return for their For this reason, it is recommended that the governments ex- investments, and IFIs will need to provide longer-term financ- plore and facilitate partnerships with other governmental ing to support a country’s aspirations of better forecasts. agencies, nongovernmental organizations (NGOs), universi- ties, and industries that require the same (or similar) data 3.1.4 Addressing the Key Issues sets to maximize the value of data collected by its observa- tional systems. The Alliance for Hydromet Development set the creation of the Systematic Observations Financing Facility (SOFF) as its Care should be taken to ensure that weather observation first priority. SOFF is a multi-partner United Nations initiative systems are designed to meet, where possible, the data and (established by the WMO; the United Nations Development accuracy requirements of all desired applications, as speci- Programme, or UNDP; and the United Nations Environment fied during consultations with partners. Although the cost of Programme, or UNEP; and including major development part- building, operating, and maintaining a system of systems that ners) focused on supporting the implementation of the Global meets the needs of multiple entities will likely be higher than Basic Observing Network (GBON). Although in its infancy, a system built for a single purpose, the overall cost will likely SOFF takes a holistic view of the funding of meteorological be lower than that required for several independent systems observation networks. The financing mechanism addresses built for narrower (and commonly overlapping) purposes. the gap, supplementing resources needed to support surface In many cases, the additional cost can be offset somewhat weather and upper-air networks in small island developing by cost-sharing agreements with the various partners. Less states (SIDS) and least developed countries (LDCs). duplication of efforts will also reduce costs associated with information technology (IT), data management, equipment On top of capital investments that support setting up or mod- life-cycle management, and field and support staff. Similarly, ernizing GBON-compliant AWS and upper-air networks, SOFF for weather radar and upper-air systems, which are expen- will provide long-term support beyond project life cycles and sive to operate and maintain, it is recommended that the NMS cover a significant share of the costs of network maintenance explore forming a consortium with neighboring countries to and operations. As part of the SOFF funding, a recipient coun- distribute the financial burden that these systems represent try must comply with the GBON requirements for internation- while benefitting from the significant data value they provide. al exchange of data through the WMO Information System These collaborative approaches hold significant promise in (WIS). This provides the benefit of filling today’s major data providing more value than the more common siloed or frag- gaps, which is critical for global numerical weather predic- mented approach, which is driven by more narrow institu- tion and short-term and seasonal global forecasts and also tional priorities and their competition for limited national helps regional and local observing networks that provide ben- funding. efits for specific sets of users. One way to increase the value and sustainability of AWS, up- per-air, and weather radar systems is to explore avenues for more broadly sharing the observational weather data they produce. Consider a set of data that is collected for a single use, such as weather forecasting. Once the forecast is issued, the continued value of the data set lies in its use for climate 172    3.2 Network Planning 3.2.1 Introduction Prior to beginning the planning process of a sustainable weather observation network, the National Meteorological Service (NMS) must set and commit to a goal for the desired level of operational performance over its lifespan. One of the key and simplest metrics to measure that is applicable for all observation networks is called uptime or availability—the percentage of time that a system is fully operational. For example, if an automatic weather station (AWS) is expected to deliver 8,000 data values (total from all parameters) over a 30- day period but instead provides only 6,400 values due to component failure and lag time to repair, its availability or uptime is 6,400/8,000 = 80 percent. Uptimes relate only to the availability and capture of data from a station or network versus its total potential. They do not speak to the quality or accura- Photo: © Tomo Jesenicnik | Dreamstime.com cy of the data based on standard quality assurance/quality control (QA/QC) processes or to regular maintenance of the station, respectively. Keep in mind that this key goal is a guide to gauging how well the NMS main- tains the observing network at a fully functional state and, when not met, the NMS must commit to reviewing the reasons for this shortfall in order and take the necessary steps to improve network performance. Uptimes of 95 percent or better should be the goal to minimize socioeconomic losses and increase productivity of weather-dependent sectors—and it is common for many ma- ture NMSs to reach 95 percent or higher (table 3.2.1a.). For a developing country, however, it may be more realistic to target a minimum of 90 per- cent and commit to increasing the percentage of uptime as personnel become more proficient at maintenance and repairs and as the agency’s knowledge base and operational experience grows. It should be noted that, while some developing countries are achieving uptimes of 90 percent or above, many are “The suggested process falling far short of this benchmark and, therefore, are not receiving the full benefit from their systems (table 3.2.1b.). for planning a sustainable weather observation network works equally well whether a new network is required or an existing network needs to be expanded or updated.” Network Planning   173 TABLE 3.2.1  Station Uptimes for Developed and Developing Countries AWS Radar Upper Air   Sites Sites Sites Country (number) Uptime (%) (number) Uptime (%) (number) Uptime (%) a. Developed Countries: Station Uptimes Reach 95–98% Australia 701 99% 34 98% 38 98% Austria 270 100% 5 80% 1 100% Canada 585 96% 30 95% 30 90% Estonia 107 95% 2 95% 2 95% Germany 997 98% 18 97% 10 99% Hungary 292 90% 2 95% 4 95% Norway 256 97% 11 97% 6 96% United Kingdom 300 97% 15 98% 7 90% b. Developing Countries: Station Uptimes Fall Far Short of Even 90% Argentina 14 50% 7 50% 3 95% Colombia 365 50% 4 80% 8 80% Ecuador 105 20% 5 30% 3 30% Jamaica 96 95% 1 95% 1 95% Paraguay 80 90% 1 60% 1 — South Africa 236 88% 10 92% 11 92% Source: Data provided by NMS to the GFDRR of the World Bank. Note: Austria targets 100 percent uptime for radar (20 percent reduction due to one radar system failure in 2020). Columbia targets 80 percent uptime for AWS (50 percent in 2020 due to Covid issues). AWS = automatic weather station; GFDRR = Global Facility for Disaster Reduction and Recovery; NMS = National Meteorological Service; — = not available. The suggested process for planning a sustainable weather process should be repeated until the estimated total cost is observation network works equally well whether a new net- brought in line with the long-term financial resources. work is required or an existing network needs to be expanded or updated. In either case, the process facilitates estimation So how can the TCO be lowered? Several strategies stand out: of the total cost of ownership (TCO) of the network—which includes both capital cost (procurement) and costs related Choose a less ambitious network design. This can be done to installation, operation, maintenance and upgrades, data by reducing the number of individual stations or systems or management, and personnel for all types of weather stations the number of different types of stations or systems. These (manual, automatic, aviation) and upper-air and weather reductions will likely lower costs in other areas as well, in- radar systems. It is important to note that this process is cluding land acquisition, site preparation, installation, and intended to be iterative: if the initial TCO estimate exceeds operating costs (such as power and communications); spare the potentially available budget, the network design should components; maintenance and upgrades; and the number of be reviewed to see where costs can be reduced to make the personnel required to maintain and operate the network. system or network affordable over its expected timeline. The 174    Network Planning What is not recommended is to cut costs with measures that to consider whether lower-specification alternatives will meet lessen data availability (increase downtime) or reliability; the data requirement needs or require more frequent mainte- adopt strategies that reduce maintenance frequency, neces- nance or replacement than higher-specification instrumenta- sary spare components, or planned upgrades; or eliminate tion, which could reduce or eliminate potential cost savings. personnel who are required for maintaining and operat- ing the (reduced) network. It is also important to consider Using AWSs as an example, Hartley (2018) discusses four whether less-expensive capital items—such as instrumenta- measurement quality classes that quantify the measurement tion that does not meet the needs of the data requirement or uncertainty. Although the World Meteorological Organization instrumentation that is not robust enough to stand up to the (WMO) has not adopted these classes, they can be used to rigors of the environment—might require significantly more illustrate the concept of specifications versus cost, as done frequent maintenance or replacement, which could reduce, in table 3.2.2. In this case, an NMS considering a network of negate, or exceed potential cost savings. synoptic AWSs should do so citing specifications above the threshold line. Equipment below the threshold of required Accept greater measurement uncertainty. Specify lower-cost uncertainty can be used as infill stations but not as the pri- or lower-specification instrumentation, or the number of envi- mary source of data. ronmental parameters being observed. Again, it is important TABLE 3.2.2  Measurement Type versus Accuracy and Cost for AWS Accuracy Class Description Type and Cost A Measurement meets the WMO required measurement un- Reference climatological or research stations certainty and stated achievable measurement uncertainty, Higher that is Annex 1.A. B Measurements have a wider uncertainty than class A. Synoptic or AWOS stations Threshold C Specifications more relaxed than class B. Well maintained public weather stations Lower D Wider than class C or no information is available. Crowd sourced weather stations Source: Based on Hartley 2018. Note: Annex 1.A refers to the annex in WMO-No. 8 (WMO 2018). AWOS = automated weather observing system; AWS = automatic weather station; WMO = World Meteorological Organization. Use a mix of manual and automated systems. When the TCO Share the costs. Consider cost-sharing agreements that help or the technology exceeds the capabilities of the NMS, con- support the ongoing operations of a network or system. These sider less expensive or alternative methods of developing or could involve: extending the network. For example, the NWS could ensure the stable operation of those parts of the existing manual ■ Renewable energy providers that require observational weather station network that meet the data requirements surface weather data rather than replacing them with more expensive AWSs. Over ■ Infrastructure, hydroelectric providers that require prod- time, and as the NMS gains experience, the NMS could gradu- ucts derived from radar ally replace manual stations with AWSs. ■ Countries that require observational weather data and network design help. It could be a multinational regional Incorporate alternative data sources. Consider other sourc- planning and design agreement aimed at prioritizing key es of observational data that might be available in-country technologies (such as weather radar). that meet requirements at little or no cost. For example, a long-standing practice is to incorporate aviation or agricul- Implement gradually. Build a subset of the planned sta- ture data for use in public forecasting. tions/systems in the initial network phase—and delay full Network Planning   175 implementation until the initial phase is complete and operat- 3.2.2 The Planning Stage ing smoothly or when additional funding becomes available. This approach, which can be done even if the full network The planning stage is crucial for setting the foundation for plan is economically feasible, is a way to reduce initial costs, sustainability of the network or system. Each of the 10 ele- personnel needs, and maintenance requirements, thereby ments of the planning stage are discussed below. better positioning the network for long-term sustainability. 3.2.2.1 Defining the Purpose of the Required These strategies work for different types of networks, whether Observation Network or System they are composed of AWS, manual weather stations, AWOS, The first step in the planning process is to clearly define the upper-air stations, weather radar systems, climate observa- purpose of the weather observation network or system, as tion stations, or a mixture of these types. This chapter iden- this determines what information is needed, when and how tifies the network planning process from start to finish and often, in what site locations, and to what level of uncertainty concludes with some recommendations. For key International (see box 3.2.2 for key meteorological planning terms). Key ex- Civil Aviation Organization (ICAO) and WMO documents on ample questions to ask include: network planning, see box 3.2.1. ■ Is a network needed to replace or augment an existing BOX 3.2.1  ICAO and WMO Key Network Planning observation network? It is important to do a detailed gap References analysis between the existing and the new network to un- derstand the added value to the NMS. Observational and operational standards can be found in ■ Is there a higher frequency of observations required to the World Meteorological Organization (WMO)'s Guide to support nowcasting? WMO defines nowcasting as  a de- Instruments and Methods of Observation (WMO-No. 8) and tailed analysis and description of the current weather and in documentation related to the WMO Integrated Global then forecasting ahead for a period from 0 to 6 hours. Observing System (WIGOS). ■ Will it be used to improve short-term weather forecasting WMO has defined a metadata standard in WIGOS Metadata in a heavily populated area? Standard (WMO-No. 1192), which is designed to meet a ■ Will it increase public safety by providing more complete broad range of WMO operational requirements and scien- weather information to aviators, or by giving earlier warn- tific requirements and to share information (collectively ing of approaching severe weather? referred to as metadata) about weather observation sta- ■ Will it allow better monitoring of the changing environ- tions and systems (including locations, station types, and classifications), observational parameters measured, and ment, thus informing our understanding of the impacts of the frequency and uncertainties of measurements. climate change in the region? ■ Will it fulfill some combination of these purposes? A web-based tool, the Observing Systems Capability Analysis and Review Tool (or OSCAR), available at https:// ■ Will it allow for ground truthing of other data—for example, oscar.wmo.int/surface/#/, has been developed to manage an automated rain gauge network for calibration of radar these requirements. data or a network of AWS to calibrate satellite data? For automatic weather stations (AWS), WMO, together with the International Civil Aviation Organization (ICAO), Once the purpose(s) of the network has been identified, the specifies requirements for weather observations neces- NMS should perform a gap analysis to determine what sta- sary for flight operation (see the Manual on Automatic tions or systems are needed to complement or replace exist- Meteorological Observing Systems at Aerodromes, Doc ing infrastructure. 9837 AN/454, 2nd ed., 2001). The product of every network or system is data. It is from these data that we draw insights about the environment, such as the current and future states of weather, hazards to 176    Network Planning navigation, infrastructure, or populations. It is, therefore, pressure, precipitation, and wind speed and direction at a critical that the data generated are complete, accurate, and number of locations; in other cases, weather radar observa- consistent enough to support the end purpose of the network tions might be necessary. In every case, reliable and accurate or system—and that there is a plan for how to share the data observations provide the foundation upon which all interpre- with the broader meteorological community. tations and forecasts are built and, thus, the fundamental jus- tification for the initial capital investment. BOX 3.2.2  What Are Systems, Stations, Networks, and It is important to note that measurements made to high- Data Centers? er-than-necessary uncertainty (lower-than-necessary accura- cy) can result in unreliable analyses and forecasts and will A system comprises one or more components, generally not yield the value of the network. In contrast, those made at a single location, that gathers sufficient information to lower-than-necessary uncertainty (higher-than-necessary to produce a complete weather data product. For exam- ple, an automated weather observing system (AWOS) at accuracy) can significantly increase costs without increasing an airport consists of a number of different sensors and value. Setting appropriate specifications for the measured instruments for providing continuous, real-time informa- parameters allows the network to perform its intended pur- tion, and reports on weather conditions at the airport. pose, while simultaneously providing the best value for the In contrast, a station (which is also composed of one or investment cost. In the end, a well-planned observation sys- more components situated in a single location) does not tem or network should include a financial plan, a business typically provide enough information to produce a com- plan, a business model, a concept of operations (CONOPS), plete weather data product on its own; a single automatic and an operating plan (see box 3.2.3). weather station (AWS) provides a good example. Because of the limited information they provide, it is 3.2.2.3 Defining Requirement Specifications common to link multiple stations together into a network The third step in the process is to define the required spec- in which all data are shared with a data center. The data center facilitates data management, including quality ifications based on global standards. These include the level control/quality assurance, storage, and analysis, and the of measurement uncertainty, the frequency of measurement, production of data reports. and other requirements (such as those related to instrument Note that systems can also contribute information to a mounting and siting). WMO has developed observational and data center. For example, even though the primary func- operational standards to ensure that measurements are made tion of an AWOS is to provide weather information to sup- at sufficient frequency and uncertainty for almost all types of port flight operations at the specific airport where it is weather observations and to promote the integration and shar- installed, it might also send observational data to a data ing of observation data between different countries’ NMSs. center to improve regional forecasts. For some applica- tions, a standalone weather observation system might suffice; for others, a network of weather observation sta- The price of an instrument is determined largely by its set tions will be more appropriate; for others still, a combi- of specifications, which include measurement uncertainty, nation of systems and networked stations will provide the measurement range, operational temperature and humidity best solution. range, any internal processing (often called smart sensors), and overall ruggedness. While both the initial purchase and ongoing operational costs of higher-specification feature-rich 3.2.2.2 Identifying the Necessary Meteorological instruments are often greater than lower-specification alter- Observations natives, these greater costs can be justified in applications The second step is to identify the specific observations nec- where low uncertainty or high reliability are paramount. For essary for fulfilling the network’s purpose, and for this and all example, marine environments and sites that are especially following steps any monitoring partners should be included. windy, dusty, cold, hot, or insect-rich, as well as high latitude In some cases, necessary weather information might include or altitude sites, all pose challenges to instruments and sta- near-surface air temperature, relative humidity, barometric tions and should be specified and built accordingly. Network Planning   177 BOX 3.2.3  Key Elements of a Well-Planned Observation System or Network The financial plan should consider the total cost of ownership (TCO) of the components of the observing network to ensure that at each stage of development, implementation, and operation the network is affordable and reliable. The business plan should describe the value derived from the observing network; support an increase in financing to support the capital purchase, installation, and operations and maintenance of the network; and consider the impact of different financial scenarios on the network’s performance based on a business model. The business model should be used to define how the observing network is supported, covering all business processes and policies. It should be used to address staffing and expense issues by considering the merits of in-house or external contracting to support observations. And it should include any international or national data-sharing policies and agreements. The various approaches can be tested using the concept of operations (CONOPS). The CONOPS should be used to create consensus on how the system will operate, reduce risks prior to procurement, and in- corporate quality improvement by leveraging existing and new infrastructure to boost the observing network’s performance. It should describe the scope and characteristics of the proposed system and how the system will be used. The operating plan should describe how the observing system is used—including operating procedures and routine mainte- nance schedules, unforeseen maintenance, acquisition of spare parts, life cycle, and repair schedules. It provides input to the financial plan by capturing actual expenditures. Higher-specification instruments are usually more expen- network according to the optimal spatial distribution if the sive and can be more complicated to maintain, support, and resulting network cannot be fully supported operationally. calibrate. On the other hand, less expensive instruments are ■ Calculating the number of stations to be installed in the commonly less robust, and, therefore, require more frequent network. This is based on the optimal spacing guidelines maintenance and (potentially) replacement—meaning that and practical considerations where stations can be in- in some cases, their total costs can approach or even exceed stalled. If the network is to include several different types those of higher-specification instruments. Thus, the right bal- of stations (for example, a mix of automatic weather sta- ance between specifications, costs, and maintenance/support tions and manual weather stations), the number of each capabilities of the network personnel must be assessed to station type should be determined. guarantee that the owner can keep the system operational for ■ Locating the individual stations. This should be done 10 to 20 years or longer, depending on the observing system. according to the optimal spacing guidelines and site specifications. Consideration should be given to suitabil- ity with respect to land cover and geological characteris- 3.2.2.4 Designing the Observation Network tics; whether the proposed sites are publicly or privately The fourth step is to design the overall observation network owned; proximity to public infrastructure, such as power plan. This includes: and communication lines; site accessibility; and security. ■ Specifying station design. This should be done in accor- ■ Determining the optimal spacing of stations. This spacing dance with the requirements for its specific set of instru- depends on the type and purpose of the network (what to ments, mounting system, power, communications, and measure and where) along with several other factors, such security requirements. as accessibility (which depends on roads or whether the ■ Categorizing stations (if applicable) as primary, second- stations are far from the NMS or airports), topography, and ary, or tertiary. Categorizing should be based on criteria staffing at the stations (for regular weekly or semiweekly defined by the NMS or WMO (such as climate stations). station maintenance). It makes no sense to implement a In a synoptic network, primary stations might perform 178    Network Planning measurements of parameters (such as soil temperature for suitable computer hardware, software, and storage de- or lightning detection) in addition to traditional measure- vices will need to be assessed and their costs calculated. ments (such as air temperature, relative humidity, baro- ■ Will multiple data centers in different geographical loca- metric pressure, surface wind, precipitation, and global tions be used for data collection and backup of systems? radiation). Using two or more data centers that monitor the status of ■ Estimating the site’s accessibility. This estimation should the others allows for uninterrupted data flow and backup include travel times and logistic costs associated with a should one center go offline. visit to the station’s location. ■ Who will require access, and for what purpose(s)? What ■ Planning for data communications. This should be done to software and hardware will be required to support individ- ensure reliable and secure transmission of data to the data ual user needs? center. If, for example, data are to be transmitted via a cel- ■ Will data be processed at each station (known as edge lular network, it is critical to ensure that there is sufficient processing) or centrally at the NMS data center or cloud coverage at the proposed station locations. If only satellite services? Processing at individual stations should be done communication is available, it might be efficient to process only where data communications are limited or expensive, data at the station and transfer only a small amount of pro- or where there is a local need for the data (such as at an cessed data. Where possible, redundancy will enhance the aerodrome). Processing data centrally is preferred for successful transmission of data back to the central server maintenance and upgrade considerations, as it is easier to or cloud server. perform modifications at one central location than it is at multiple stations. Regardless of system type, a site survey should be performed at each selected location to ensure that the system meets its 3.2.2.6 Defining Infrastructure Requirements intended purpose and is constructed to international stan- The sixth step is to define infrastructure requirements. These dards. The site survey should address representativeness of include civil works (such as power and communications), site observations made; availability, reliability, and cost of con- preparation, station mounting platforms, necessary buildings necting power, testing of communications link(s), and other or other structures, and security fencing. In addition, the NMS required utilities (for example, water); and site security. If a must install servers for the central data processing systems; site survey is not possible, a list of questions should be made this requires an information and communication technology on the above topics for use by the NMS’ project manager, (ICT) infrastructure—such as data communication equip- given that a baseline set of information is required to prepare ment, equipment rooms, information technology (IT) security for the site installation to avoid delays and higher costs. processes, and ICT support staff. 3.2.2.5 Developing a Network Data Management Plan 3.2.2.7 Defining Operational Requirements The fifth step is to develop a network data management plan. The seventh step is to define operational requirements. The plan should include protocols for data backup, data secu- Regardless of initial cost or specification, all weather moni- rity, QA/QC, analysis, transmission/distribution, and archival toring equipment requires periodic maintenance, recalibra- storage. The plan should be in place prior to network con- tion, and replacement. Thus, a well-defined maintenance struction and answer the following questions: strategy is critical to maintain an observation network. It is recommended that operational requirements be planned for ■ Will data storage take place in a physical location man- the duration of the expected system or network lifetime. The aged by the NMS or through an online (cloud) data service? plan should include: In either case, calculations should be made regarding the rate at which data will be communicated to the data center ■ A daily to weekly station inspection and maintenance (to determine bandwidth and costs) and total volume of plan. Several sensors, especially those with optical win- data that will need to be stored over the network’s lifetime dows—such as global radiation sensors, visibility/present (to determine storage capacity). Similarly, requirements Network Planning   179 weather sensors, and ceilometers—need daily to weekly ■ An unscheduled corrective maintenance plan. Besides cleaning (to remove elements such as dust, debris, bird scheduled maintenance, plans and procedures should be droppings, spiderwebs). Grass must be mowed every two put in place for any required unscheduled maintenance, weeks in the growing season. Rain gauges must have leaves which can be necessitated by sensor or station failure, and other items removed to ensure measurement integrity. unexpected instrument degradation (as revealed by data A local person near the station could be contracted and QA/QC), vandalism, and so on. Unscheduled maintenance trained to handle these activities, reducing the logistical should be performed within the data criticality timeframe. costs of NMS personal traveling to the site and performing ■ An adaptive maintenance plan. This supports upgrades of these actions. the entire value chain, from measurement to data process- ■ A scheduled preventive maintenance plan. Visits should ing. Typically, upgrades (software, for example) are need- typically be planned every 3 to 12 months (depending ed so that an NMS can take advantage of innovations that on local conditions, location, and system requirements). provide more value from the data or because of regulatory Because weather monitoring equipment is constantly changes (for example, by ICAO in the specification of aero- exposed to harsh environmental conditions (such as in- drome operations or AWOS system configuration). tense UV radiation, temperature extremes, dust, and high winds), they must be periodically maintained if they are to The task of adhering to a comprehensive maintenance plan deliver accurate and reliable measurements. For all of the can be made easier by using service-level agreements. These network equipment to receive preventive maintenance at a agreements are contracts whereby instrument manufactur- suitable interval, careful planning should include: ers or third-party contractors agree to perform services on p Scheduled station visits at a frequency determined behalf of the NMS for the ongoing operation, calibration, and by each station’s data criticality, environmental con- maintenance of the system(s). ditions, and accessibility. p Clearly defined maintenance procedures for each 3.2.2.8 Identifying Personnel Requirements instrument, which meet manufacturers’ recommen- The eighth step is to identify personnel requirements. No dations for the specific instrumentation used at each network will function properly without the efforts of dedi- station, including: cated, qualified personnel. Personnel install, maintain, and p A calibration plan. This is for cases where instru- repair stations; they manage, maintain, and perform QA/QC ment maintenance requires recalibration or refur- on incoming data; and they ensure that the gathered data are bishment to limit downtime (such as an instrument made available for analysis and to produce data products. rotation schedule). It is a good idea to invest in suffi- Each of these roles require people with suitable qualifica- cient spare sensors in the initial purchasing project tions and training, underscoring the need to make training to be able to set up a calibration procedure based on and retention top priorities, along with paying competitive swapping sensors when calibrations expire. Spare commercial market rates. As the geographical extent of a net- components, hardware, and sensors will cost 10 work increases, so do travel-related expenses and human re- percent to 15 percent or more of the initial capital source costs. In both cases, larger networks typically require cost of the project, depending on the system. more personnel. p A life-cycle management plan. Sensors and hard- ware should be replaced before they reach their As for personnel categories, field technicians (also called end of life to ensure the continuous flow of quality system operators or system observers) perform the technical data and limit the downtime of a measurement sta- tasks of maintenance, repair, and upgrades to the meteoro- tion or system. logical systems. In some countries, they also perform regular p An upgrading plan. Instruments or complete sta- upkeep and minor maintenance of the sites (such as mow- tions should be upgraded as they become obsolete, ing grass, trimming vegetation, and cleaning dust and debris or as more suitable options become available over from sensors and structures). In other countries, these duties their life cycle. are performed by separate site maintenance personnel. Field 180    Network Planning technicians should possess a degree in a related technical To this end, the government or the NMS should appoint a discipline: in Canada, a technical degree is required; in some project manager to work closely with the vendors’ project countries (such as Norway and Estonia), engineers perform manager. If this individual is within the NMS, the NMS should these functions. Field technicians typically receive additional consider contracting a systems integrator to manage the training from the NMS or the manufacturer of the meteorolog- project. A project plan should be prepared that outlines the ical systems. Given that the tasks performed by site mainte- responsibilities of both parties to ensure the project is deliv- nance personnel generally require much less skill or training, ered on time and on budget. these positions can commonly be filled by members of nearby communities. 3.2.2.9 Determining the Most Effective Operating Business Model The NMS will also require dedicated support personnel for The ninth step is to determine the most effective business data management and analysis, maintenance planning and model to support network operations. There are three busi- scheduling, maintaining an inventory of spare parts, and ness models that NMSs may consider employing to achieve life-cycle management. Additional support personnel include long term operations of their network: but are not limited to: ■ Government-owned and government-operated networks. ■ Specialists/scientists to analyze, interpret, and apply radar Here the NMS owns and operates the equipment in the net- information work. This is the most widely practiced method by NMSs ■ IT specialists supporting data ingestion, QA/QC functions, to obtain the data needed to support their monitoring re- storage, and the flow of data and products to forecast quirements. In this model, the NMS is wholly responsible models for the acquisition, distribution and exchange, and quality ■ Network management and planning specialists, including assurance of their data. those who work in service, incident, change and process ■ Government-owned and privately operated networks. In improvement management, and life-cycle support this arrangement, the NMS owns the network, but a high ■ Operational support 24/7 from the IT service desk to log majority of its stations are operated by an outside contrac- field site, communications, or server failures to ensure re- tor. The Meteorological Services Canada (MSC) uses this pairs are affected in a timely manner. type of business model for operating its upper-air network: MSC owns the network hardware but uses third-party con- To bring the network plan to reality, it is important that the tractors to perform the balloon launches and rudimentary observational and operational specifications developed in maintenance tasks. Canada has chosen this business model, the planning exercise are met; the necessary station compo- as it is less expensive given its geographical size. This model nents and spares are purchased; the stations and data center should be considered by an NMS if it has the capital finances are constructed properly and on schedule; and the necessary to acquire the network but might not have the necessary in- personnel are hired, trained, and provided with adequate re- house skill set to carry out the activities required to sustain sources. The building and delivery of observation systems in the network, thereby ensuring quality data flow. projects can be described in a general manner for most types ■ Privately owned and privately operated networks (also of observation systems. Through strict project management, known as data as a service). This is where the contrac- a full range of design, reviews, testing, installation, commis- tor owns and operates the equipment in the network. sioning, training, and services is provided, which together en- This model is used by the Austrian government for its sure that the network is installed correctly and on schedule. weather radar network (although the partner is a govern- The network should be fully operational when transferred to ment-owned company, rather than a true private entity): the NMS, so that it can be sustainably operated for its expect- the Austrian Met Service pays a fee and receives the data ed lifespan. without having to acquire and operate the network of ra- dars. This model should be considered by an NMS if it does not have either the capital finances to acquire the network Network Planning   181 or the necessary in-house skill set to properly operate and p Power and communication maintain complex observation systems such as radars. p Personnel, including salary/benefits, training, and logistics p Schedule replacement of computers, sensors, and The NMS should consider these models and adopt the most other equipment. cost-effective one that will serve the national interest over the long term. In particular, models in which maintenance If the capital costs are not known, the NMS could use a re- and operation are outsourced to third party should be strong- quest for information (RFI) process prior to releasing the ten- ly considered in cases where the NMS has concerns that re- der (see box 3.2.4). cruiting and retaining qualified personnel might be an issue. It is important to note that the choice of business model can strongly impact the operational requirements and personnel BOX 3.2.4  How a Request for Information Works requirements, and, therefore, the overall cost of the network. As such, these sections should be considered to be strongly Under a request for information (RFI), the National interdependent components of a successful network design. Meteorological Service (NMS) documents the purpose When developing a business model, the NMS should consult of the network, types of measurements required (with with meteorological equipment manufacturers, NMSs with specifications), an understanding of the required spare parts and life-cycle management costs, and operational significant experience, or entities with comparable expertise. requirements to support the system to ensure an expect- ed level of performance. The vendors will respond with 3.2.2.10 Overview of Performing a Cost Analysis general comments on the system design, pricing informa- The tenth, and final, step in the planning process is to perform tion, and descriptions of currently available instruments and technologies that they would provide to satisfy the a cost analysis. This can be done once all of the components described needs of the NMS. of the network are well specified. What follows is a quick overview of a costing exercise (the topic of chapter 3.8). This Neither party is obligated to proceed, which provides a safe way to gather and exchange information. The NMS process depends on understanding of the costs, and vendors has a chance to learn how to better design and build its can provide preliminary quotations. The cost analysis should weather observation system, and vendors have a chance include: to ask questions and clarify their understanding of the NMS’s monitoring goals and requirements. If the pricing ■ Procurement (purchasing) project costs, including all returned is too high, the NMS can either look at a different costs associated with: business model to achieve its monitoring goals or scale back the network size (the number of stations or the pa- p The purchase of all components of the stations rameters measured) and technical characteristics—for p Spare instruments and components example, it can review whether an X-band radar or an p Shipping the components from vendors and transpor- S-band radar will serve the NMS’s needs to better fit the tation to the location of the stations. available budget. In this way, the NMS has a better pros- ■ Installation costs, including all costs associated with: pect of success when it officially releases the tender, as it p Site preparation and station construction has had a chance to confirm that it can support the net- work in a sustainable fashion. p Providing power and communication services p System testing and validation p If a centralized data center is to be employed, building construction/renovation, equipment, setup, and testing When planning to implement a new network of stations or p Personnel, including recruitment, salary/benefits, and systems, the NMS should start with a small initial acquisition, training. armed with a plan to expand the network over several phases, ■ Annual operating costs, including all costs associated rather than procuring all of the equipment and attempting to with: build the full network in a single push. This phased approach p Network maintenance, repair, and upgrades offers the NMS time to gain experience with new technolo- p Replacement of broken sensors and equipment gies and to develop a base knowledge of how to operate and 182    Network Planning maintain the systems—before being tasked with building, partners, and should include a long-term financial plan, operating, and maintaining a full-scale national observation along with other items, to support implementation (box network. It also assists in long-term life-cycle management 3.2.3). and the eventual replacement of the network at the end of 2. Develop a full network plan up front. The government its operational life, given that it distributes the timing and and development partners should encourage and sup- financial burdens of maintenance and replacement activities port the NMS in preparing a full network plan to promote over a longer period of time. a thorough understanding of the TCO. This will help im- prove in-country weather observation capabilities by Prior to performing the cost analysis (the Total Cost of sizing the system or network to match the NMS’s capa- Ownership Exercise provided in chapter 3.8), the NMS bilities and budget. One outcome of this approach would should review the relevant chapter or chapters covering be the development of a financial plan that provides an the meteorological observing system or network it seeks to estimate of the TCO over the network’s expected life that implement. These chapters include chapter 3.4, Automated has the long-term support of the government and devel- Weather Observing Systems; chapter 3.5, Automated Weather opment partners. Stations; chapter 3.6, Upper-Air Systems; and chapter 3.7, 3. Explore innovative long-term financial models. The Weather Radar Systems. government and development partners should work together to develop new long-term financial models to Each of the chapters highlights key points for an NMS to con- ensure that investments in capital projects promote sider, along with data specific to life-cycle management costs, sustainability. The current, primarily capital invest- operation maintenance costs, staffing requirements, and so ment–based, approach is not achieving the desired or on. The data were provided by NMSs in both developed and expected outcomes. Greater attention needs to be paid developing countries, following Global Facility for Disaster to boosting the NMS’s financial stability and operations Reduction and Recovery (GFDRR)/World Bank requests. It to enable the weather observation system to perform should be noted that each country operates and accounts for more reliably over its expected life-cycle. its activities differently, and we have used our best efforts to 4. Improve the quality of network operational data. homogenize and analyze this information. Useable data from Governments and development partners should encour- the developing country NMSs were often sporadic and incon- age NMSs to record and provide access to financial infor- sistent and thus should be considered indicative only. NMSs mation on network operations. A significant challenge from five developed countries—Austria, Australia, Canada, in creating this part of the report was the sparse and Germany, and the United Kingdom—submitted comprehen- inconsistent availability of specific operational, human sive data that were used as a basis for the benchmark esti- resource, and financial data in developing countries. As mates for AWS, upper-air systems, and radar systems. a consistent data set is built and analyzed, it can be used to better assist NMSs in developing and planning future 3.2.3 Recommendations projects. These data should include but perhaps not be limited to: The bottom line is that government agencies, NMSs, and de- p Capital cost of the system/network that clearly identi- velopment agencies (such as the World Bank) have many op- fies the type of equipment being acquired; if a net- portunities now to increase the likelihood of the long-term work, include the quantity of stations, measurement success of the observational programs that they support. The parameters, options, and so on in general terms. following recommendations may prove useful in this regard: p Cost of infrastructure should include site preparations, land acquisition, building/tower costs, ongoing or es- 1. Create a roadmap for developing an observation net- timated power and communications, and so on. work. This can be a standalone document or part of a p Staffing costs and personnel/full time equivalent (FTE) larger one covering the entire service. It should be used counts should be included for field technicians and to guide investments from government and development Network Planning   183 support staff, which includes a summary of duties and 3.2.4 References responsibilities. p Annual expenditures related to operation, mainte- Hartley, B. 2018. Data Collection Network Modernisation: What nance, consumables, and life-cycle management. You Need to Know. Meteorological Service of New Zealand 5. Adopt a TCO approach. Prior to purchasing a new ob- Limited. servational system or network, the NMS, government, and development partners need to consider not just the ICAO (International Civil Aviation Organization). 2011. Manual capital acquisition cost but also costs associated with on Automatic Meteorological Observing Systems at Aerodromes, installation, a human resource plan, and long-term oper- Document 9837 AN/454, 2nd edition. Montreal: ICAO. http:// ation (including upgrades and equipment maintenance). www.icscc.org.cn/upload/file/20190102/Doc.9837-EN%20 While this can represent a considerable effort for an NMS Manual%20on%20Automatic%20Meteorological%20 pre-purchase, the benefits of understanding the TCO and Observing%20Systems%20at%20Aerodromes.pdf. the requirements for the system or network management provide the foundation for long-term success. WMO (World Meteorological Organization). 2018. Guide 6. Freely and openly sharing NMS data. Beyond their own to Instruments and Methods of Observation, Volume I: data requirements, NMSs should partner with other Measurement of Meteorological Variables. WMO-No. 8, 2018 agencies and entities that will expand their observa- edition. Geneva: WMO. https://library.wmo.int/doc_num. tional reach and the value of their data collection. NMSs php?explnum_id=10616. should openly and freely share all data through WMO’s Integrated Global Weather Observing System (WIGOS)/ WMO (World Meteorological Organization). 2019. WIGOS WIS program and the GTS to support public safety obli- Metadata Standard. WMO-No. 1192, 2019 edition. https://li- gations, forecasting and warning capabilities nationally brary.wmo.int/doc_num.php?explnum_id=10109. and internationally, and a global understanding of the world’s weather and changing climate. WMO OSCAR (World Meteorological Organization Observing Systems Capability Analysis and Review Tool). https://www. wmo-sat.info/oscar/. 184    3.3 Total Cost of Ownership for a Weather Observing System 3.3.1 Introduction Many weather observing systems do not achieve their expected operational lifetimes for a variety of reasons, but the most common are either an unre- alistic expectation that is the result of a poor understanding of the system’s true operating cost or the lack of a sustaining budget and qualified staff to run complex observation infrastructure. The simplest way to minimize this risk is Vietnam. Photo courtesy of V. Tsirkunov, WBG. for the National Meteorological Service (NMS) to have as complete a knowl- edge as possible—prior to going to procurement—of all the costs associated with the capital investment. And the best way to do that is to calculate the total cost of ownership (TCO), which looks at the long-term affordability and financial requirements of implementing and operating an observation system or network designed in support of the NMS’s data requirements. The TCO includes not only the initial capital investment of equipment but also an annual budget that allows the sustained operation of the system through its expected lifetime. Whether the meteorological observation system is being replaced because of its end of life or because a new system is being “The key factors to contemplated, the principles of costing are similar. Different approaches to the network design and operation may need to be weighed with a goal of min- successfully operating imizing the TCO while also minimizing downtime to ensure consistent data a sustainable weather flow. This exercise helps the various parties make informed decisions about the sustainability and suitability of the services they anticipate acquiring and observation system include supporting. (1) a composition and The key factors to successfully operating a sustainable weather observation structure that meets the system include (1) a composition and structure that meets the country’s ob- servational needs; (2) a well-developed operational, maintenance, calibra- country’s observational tion, and life-cycle management process; and (3) a stable and well-trained needs; (2) a well-developed workforce. But the cornerstone to sustainability is ensuring that financial re- sources are dedicated to support the system over its expected lifetime—and operational, maintenance, that makes critical the need to understand the TCO of a weather observation calibration, and life-cycle system over its expected life cycle. management process; The TCO can vary significantly from system to system and from country to and (3) a stable and well- country for many reasons. Each country poses a unique set of challenges because of its geography, climate, monitoring needs, and NMS capabilities. trained workforce.” Despite this, all countries have two things in common: (1) the need for quality Total Cost of Ownership for a Weather Observing System    185 data that is “fit for purpose” to support mandates such as FIGURE 3.3.1  Process of Managing a Meteorological public safety, and (2) the need to justify the total cost of the Observing System observation system that will provide the data. This chapter offers an overview of the TCO, starting with why it is such an important tool for an NMS and then detailing its key Design components. 3.3.2 Understanding the Total Cost of Ownership End of Procure life and install To understand the value of the TCO, we need to answer the question “Why does an NMS exist?” While this has been dis- cussed many times previously, it bears repeating. At its most Operate basic, it exists to develop forecasts to better prepare the pop- Maintain ulation when extreme weather is about to occur and to be an Update integral part of getting the economy running again when the weather turns good. In recent decades, expectations for an NMS, as well as its mandate, have grown to include issues such as climate adaptation and economic prosperity, but its basic reason for existing is public safety. How the NMS car- It is also critical to understand that the TCO serves two im- ries out its basic function is by using scientific measurements portant purposes in supporting the constant flow of data: to characterize a chaotic environment from an uninterrupted supply of data. The concept of sustainable observation sys- ■ Financial assessment guide. The TCO attributes the costs tems is rooted in the premise that there is a constant flow of of an observation system over its lifetime, which promotes data from the field. What the NMS does is to provide accu- the understanding of the need for a sustaining budget (see rate forecasts using measured scientific data that support its section 3.3.3). primary function of public safety and support more efficient ■ Asset management. The TCO informs on how to manage an weather-dependent sectors. An NMS cannot successfully de- asset through its entire life with the goal of doing so in a liver forecasts if the weather observing systems it operates fiscally responsible manner. are not sustained over the lifetime of the asset. Without the constant flow of quality data, the NMS will be unable to deliv- The TCO financial assessment guide is a process that enables er its basic function to the level required. NMSs to collect data that support decision-making through to operations, repair, and replacements. The collection and The basic function of an NMS is long lived and, therefore, so is recording of financial data that support the decision-making its need of a constant flow of data. In comparison, the obser- is an iterative process, based on learning and experience. As vation systems it operates are short lived, up to 10–25 years. understanding grows, so does the NMS’s knowledge of what Figure 3.3.1 shows this constant process of designing, procur- financial data are required to manage the asset that supports ing and installing, operating, maintaining, and updating, until future budget development. the end of life. When the life of the asset is deemed to be at an end, the observation system will need to be replaced and One financial tool that informs asset management is the value the process will repeat. But if the flow of data is to be main- of an asset, which is determined by the degree of its depreci- tained, the NMS must begin the planning of the replacement ation—that is, the value lost over the period of its life cycle, system years prior to the current system’s end of life in order either due to wear and tear or when manufactured replace- to complete the designing, procuring, and installing phases ment parts are no longer accessible. While not part of the well before. financial assessment calculation, asset management is a 186    Total Cost of Ownership for a Weather Observing System valuable tool in assessing the financial health of an obser- lubrication is not carried out on the recommended schedule, vation system, understanding the scale of financial liability the moving parts will wear out prematurely or seize in place, (assuming the observing asset will be eventually replaced), not allowing the antenna to spin. This scenario can occur and supporting the decision-making process on when to re- within a year of procurement and installation of the system if pair or replace the asset. With the straight line depreciation the required maintenance is not performed. The result is that method, the value of an asset is reduced uniformly over each the radar’s value to the NMS as a source of data is rendered fiscal period until it reaches its financial end of life, at which zero without an expensive overhaul costing several thou- point its residual value is assumed to be zero (figure 3.3.2). sands of dollars—even though the residual value or financial worth of the radar may be more than $2.5 million. Clearly, FIGURE 3.3.2  Straight Line Depreciation for a Meteorological the radar has both financial life and value to the NMS, mak- Observing System ing it an easy decision to carry out the repairs. In retrospect though, it would have been less costly to have carried out the routine maintenance. It also would have created value by en- suring that the asset’s rate of depreciation remains constant, and the radar continues to be a valued data source. Case 2: Older weather radar Value of asset ($) If the same radar were 12 years old and failed, would it be worth doing repairs on a depreciated item worth $600,000? The answer would depend on several factors, such as: ■ Is the radar still a valued source of data that is critical to Number of periods (years) the NMS basic mandate of public safety? ■ Does the vendor still support and manufacture critical Note: This kind of depreciation represents the reduction in value over the parts for the radar? system’s suggested lifetime. ■ Has the radar maintained a high level of performance, or has it required significant maintenance to sustain its operations? While depreciation is a calculated value of the asset’s reduced ■ What are the costs of the repairs and ongoing maintenance worth over time, the rate of depreciation can be accelerated of a 12-year-old system versus the cost of replacement? as a result of lack of routine maintenance and life-cycle man- ■ If the radar is repaired, how many more years of service agement. If an outside agency were to review and inspect an could the NMS receive from the radar? observation site, how would they value the investment? They ■ If the NMS were to purchase a replacement, what would would look at how well the site is maintained, the operation- the cost be? How long before the system could be oper- al functionality of the equipment, and the consistent flow of ational? Is it even possible to obtain the capital funding? data (downtime). ■ Should the NMS consider replacing the radar through a data-as-a-service contract to reduce its capital require- The three cases below provide some context on the deci- ment for replacement? The answer may be based on sion-making to repair or continue to maintain based on the whether the NMS has the annual budget available to sup- assets’ current value. port such a contract. Case 1: Newish weather radar Each of the above questions and others will need to be an- A weather radar is normally depreciated over its 15-year life- swered to come to a fiscally responsible solution. time. But if a simple routine maintenance procedure such as Total Cost of Ownership for a Weather Observing System    187 Case 3: Automatic weather station depreciation ■ Is the core technology of the data logger still viable? An automatic weather station (AWS) has a lifetime of 10 or ■ Does the NMS have plans to increase the number of sen- more years, but batteries have only so many charge cycles sors measured? Does the data logger have spare capacity? available before they fail, usually every 5 to 6 years. If the ■ Should the NMS look at phasing new equipment into its NMS does not replace the batteries within the recommend- network and use the existing equipment as spares? ed lifetime, complete station failure may occur at 5 years ■ Is the core technology of the communications device still as shown by the orange line in figure 3.3.3. Similarly, some viable? sensor components also require scheduled life-cycle manage- ■ Does the vendor still support, repair, and calibrate the core ment, and when not replaced based on manufacturer’s rec- measurement and communications technologies? ommendations, they become unreliable and eventually fail. ■ What is the cost of full station renewal? Is it even possible Again, with no data, the AWS provides no value to the NMS to obtain the capital funding for this? until replacements parts are installed. ■ What is the cost to the NMS of learning the new technology of the replacement system? Other equipment—specifically relative humidity probes, wind sensors, and barometric sensors as well as the batteries—has Each of the above questions and others will need to be an- a useful life of 5 years before replacement. The white line in swered to come to a fiscally responsible solution. figure 3.3.3 shows a life-cycle investment at Year 5 that in- creases the asset value. With no further investment, the asset The cases above provide a sense of the decisions required is completely depreciated by Year 10. However, a significant to operate a network. The questions an NMS asks and the portion of the asset exceeds the lifetime set for the AWS and answers it determines are contextually based on the impor- if a second life-cycle investment is made in Year 10, the as- tance of the data an observation system provides, individual set’s life can be extended an additional 5 years, further delay- circumstances, and the NMS’s fundamental mandate. At each ing the need for full station replacement (dark blue line). The step of the process, the NMS must find a balance between its NMS must decide whether to replace the AWS at 10 years or basic mandate and its available budget—a task that is aided do another life-cycle investment and operate the system for by examining the TCO. another 5 years. At Year 10 (dark blue line), there is still sig- nificant value in the depreciated AWS. Again, the NWS needs to ask some critical questions: FIGURE 3.3.3  Depreciation of a Single Automatic Weather Station $50,000 5 Years 10 Years $40,000 Depreciated value 15 Years $30,000 $20,000 $10,000 $– 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Years 188    Total Cost of Ownership for a Weather Observing System 3.3.3 Key Components of Total Cost of 3.3.3.1 Initial Investment: Capital Costs Ownership The first category is initial investment (capital costs) which includes all costs to purchase and install the observational The TCO for a meteorological observation system can be di- system, ranging from capital equipment and civil works to vided into four major categories, as shown in figure 3.3.4: (1) supplier services and NMS project costs. initial investment (capital costs), (2) annual operating costs, (3) annual maintenance costs, and (4) replacement costs. The Capital equipment. This covers the cost of all field equip- number and how the individual items in the costing catego- ment to be installed, including supporting hardware (such ries are recorded will be based on the procedures created by as tools, test equipment, vehicles) and information process- the NMS and the respective government finance department. ing systems (IPS) (such as ICT; data collection centers; data FIGURE 3.3.4  Overview of the Total Costs of Ownership of a Meteorological Observation System Considering Replacement as Part of Life-Cycle Management Total cost of ownership Initial investment: Capital costs Annual operating costs Maintenance costs Replacement costs (life- cycle management) Capital cost of equipment Business costs Cost of preventative maintenance Cost of life-cycle management Administrative costs Cost of all field equipment Field staff (labor) Field equipment System Support staff (labor) Logistics Stationary equipment at Mobile field equipment field site(s) Replacement parts IPS computers and Mobile field equipment Other operating costs IPS computers and software software Cost of information processing Mobile field equipment systems (IPS) Cost of full system replacement Calibration of field equipment Cost of civil works Cost of corrective maintenance Consumables Procure land Software licenses Field equipment Site preparation Supporting structures Service-level agreements Installation of utilities Construction of major supporting structures Land leases Cost of adaptive maintenance Security Utilities Stationary equipment Offices and server room Asset management Mobile field equipment Cost of supplier services Calibration of field equipment IPS computers and software Installation of utilities Factory acceptance testing Site acceptance testing Training Prepaid calibrations Internal project costs Project management Staff (labor) Logistics Total Cost of Ownership for a Weather Observing System    189 processing to final products; data storage, management, and Standardization/Quality Management Systems (ISO/QMS) archiving). These costs are derived from the necessary hard- management, and training. Also business operating costs, ware (such as computer servers, communications interfaces, administrative costs, and other operating costs should be re- backup systems) and software (such as data retrieval from corded in this category. the field; quality assurance/quality control, or QA/QC; data storage, archiving, and dissemination). ■ Business operating costs. These include land and build- ing leases, utilities, fuel for backup generators, equipment Civil works. These costs center on procuring land and build- (servers, computers, vehicles), inventory costs, insurance, ings and infrastructure. One component is the cost of site and computational software costs. preparation, which might include clearing and/or leveling the ■ Administrative costs. These include the costs of personnel land and constructing roads to facilitate the installation and expressed in full-time equivalent workers (FTEs) or mone- repair of existing field equipment. Another is cost of the con- tary terms required to carry out the work of the day-to-day struction of major supporting structures (such as buildings, financial and human resource management, scheduling tower bases, towers) for observing systems such as weather site maintenance, warehousing and maintaining inventory radar and upper air. Others include the cost of installing util- control of spares, ordering replacements of failed equip- ities (such as mains power, water, sewer, communications), ment, reviewing network health, development of mainte- the cost of security (such as perimeter fencing, cameras, nance and repair processes, and life-cycle planning and alarms), and the cost of setting up offices and server areas. scheduling. ■ Other operating costs. These costs range from calibration Supplier services. This covers a variety of services. One is of field equipment and consumables to software licenses factory acceptance testing, which provides an opportunity and service-level agreements. Consumables, which are for the NMS to travel to the manufacturer’s facility to see a needed to ensure that the necessary parts for both mainte- working setup of the equipment where the NMS, when satis- nance and operations are available to sustain the system, fied, provides approval prior to the equipment being shipped. can carry a high price tag. For example, this cost is very Another is site acceptance testing, which occurs when the significant for upper-air systems, which require consum- NMS contracts the installation of the field equipment to en- ables for each launch twice per day (see table 3.6.2). sure that all systems are operating to specification. Other ser- vices include installation of equipment, prepaid calibration 3.3.3.3 Annual Maintenance Costs of sensors, and training on the new equipment and systems. The third category is maintenance costs, which primarily cov- ers the cost of personnel and travel to sites (logistics). There Internal NMS project costs. These costs refer to project man- are three main types: preventive, corrective, and adaptive. agement, staff, and logistics. Project management should be carried out by an NMS staff member (if experienced in this Preventive maintenance. This is a significant cost for all ob- area) or by a consultant hired to act on the NMS’ behalf. The servation systems. One category is the cost of personnel, re- time required can be as little as six months or, depending on corded in FTEs. Field technicians can have two or more levels scope and complexity, one year or more. Staff costs can be sig- of technical training required: nificant should the NMS choose to carry out the installation. ■ The first requires minimal training, and includes ensuring 3.3.3.2 Annual Operating Costs the site is secure, cutting grass, and performing minor The second category is annual operating costs. These are maintenance (such as cleaning the radiation shield); this costs of activities attributed to the collection and produc- can be carried out by individuals located near the site. tion of data, and which are distributed over the entire net- ■ The second requires a higher degree of technical ed- work (that is, not attributable to a given monitoring site). ucation and training to carry out routine preventive Examples include network design and enhancements, de- maintenance, following vendor guidelines on frequency velopment of data products, International Organization for (adjusted, as needed, for site environmental conditions). 190    Total Cost of Ownership for a Weather Observing System The maintenance typically includes required equipment uptime) and maintain the flow of critical data (see box 3.3.1). checks and recalibration, periodic cleaning and lubrica- For example, an AWS may have a lifetime of 10 years, but a tion, and replacement of components and upgrades. relative humidity sensor needs to be replaced every 5 to 7 years, or at least once in the system’s lifetime. Another category of preventive maintenance is the cost of logistics to travel to sites, including vehicle or special equip- BOX 3.3.1  Possible Goals for Downtimes ment rental, costs of public transportation, hotel accommo- dations, and meal allowance. Downtime is a key metric of equipment operations Corrective maintenance. An effective program must have an and should be measured by the NMS. It is expressed availability of adequate supplies, spares, and trained elec- as a percentage and should be as small as possible. In tronic and other maintenance personnel. Because corrective the following chapters on weather observing systems, maintenance is the repair of unanticipated equipment failure downtimes of 5 percent or less are presented, as these or damage from natural events, vandalism, or theft, it is diffi- cases indicate that the agencies are realizing the value cult to calculate an exact value. It is suggested that a value of of the significant investment, which helps the NMS 15 percent to 20 percent of preventive maintenance costs— meet its obligations. including the cost of equipment repairs or replacement when no longer serviceable—be used until the NMS has gained For developing countries that start operating a new enough experience to develop its own cost understanding. automated system, a downtime of less than 10 percent may be acceptable for an initial period (two to three Adaptive maintenance. This covers planned support and up- years) while NMS staff learn how to operate and devel- grades of the entire value chain, from measurement to data op the new system. But after that initial period, down- processing to training. Typically, upgrades (such as software) time should begin to fall as NMS experience grows, are needed so that an NMS can take advantage during the with a committed target goal of reaching 5 percent or system’s lifetime of innovations that provide more value from less. the data. While adaptive maintenance is usually not per- formed annually, the NMS should consult with the vendor on the typical frequency and cost associated, including whether Equipment lifetime. This depends on its absolute physical annual service-level agreements provide for updates and up- and economic lifetimes: grades to systems. ■ Absolute physical lifetime of equipment. This is the 3.3.3.4 Replacement Costs amount of time before the equipment’s ability to function The fourth category is replacement, which addresses the costs drops to zero. It depends on its design, manufacturing of replacing equipment. There are the life-cycle management quality, and materials used as well as the specific operat- costs of replacing short-lived component parts and sensors ing conditions and quality of the maintenance. It can be over the operational life of system, as well as the cost of full prolonged by preventive maintenance and replacement of system replacement when a system reaches its end of life. components. Physical lifetime should be based on recom- Both need to be addressed, carefully planned, and managed mendations from the manufacturer. For example, a single if the NMS is to have a continuous flow of data over decades. weather station may have a lifetime of 10 years with prop- er maintenance, which includes the replacement of short- Life-cycle management. This is critical to ensuring the sus- er-life components (such as a relative humidity sensor or tainability of a weather observing system. Each component of batteries that have a lifetime of 5 to 7 years). a system has a lifetime and must be replaced when it reach- ■ Economic lifetime of a piece of equipment. The econom- es that point, which means that the NMS must allocate the ic lifetime of a piece of equipment can be different from needed financial resources to minimize downtime (maximize its actual physical lifetime. Thus, the equipment can be in Total Cost of Ownership for a Weather Observing System    191 adequate physical condition but may not be economically Regardless, the NMS will eventually have to replace the ob- useful as the costs of failures and costs of maintenance serving system when it reaches the end of its useful life. accumulate over a certain time span and exceed the costs Planning the transition will be time-consuming, and it will of a total renewal of the equipment. For example, a vehicle need to take into account the design, procurement, and im- owned by the NMS may reach its economic lifetime before plementation of the new system as well as transitioning out its anticipated physical lifetime due to the high cost of the old system. maintenance and replacement parts and time lost due to repairs. As a result, it makes economic sense to replace 3.3.4 TCO and Gap Analysis the vehicle as the cost of replacement is less than the cost of maintenance and the value cost of downtime. Economic The TCO provides a basic understanding of the financial ob- lifetime can also occur with technology changes—such as ligations required to operate a successful and sustainable when a cellular modem for transmitting data from the field system prior to procuring a system. However, it is not just a site is no longer supported by the service provider. costing tool. It also provides a decision framework to support NMSs. TCO informs timely action, processes, and procedures Total system replacement. This is a decision that occurs be- to ensure that uptime is maximized to 95 percent or more. fore the end of the lifetime expectancy is reached for an ob- servation system or network. It is at this point that the NMS The NMS should also weigh its current capabilities against must decide whether it is financially more cost-effective to the new ones it will require to successfully operate the new bear the capital cost along with the cost of installation and re- system. This can be done with what is known as a gap analy- place the entire equipment suite or to extend the system’s op- sis. This enables the NMS to understand the space (the gap) erational life through another round of life-cycle investments. between “where they are today” or the present state and “where they want to be” or the desired state. Understanding The NMS will need to confront this question by determining the gap in capabilities will aid in developing the business several financial metrics: plan and the concept of operations. ■ The cost of maintaining the existing system versus the cost Note to reader: See chapter 3.8 for guidance on how to cal- of replacing it culate the TCO and chapter 3.9 for a case example of such a ■ Whether the system is technically viable calculation. ■ Supplier commitment to support the existing system ■ The cost of training and staff adoption of the new system ■ Depreciated or residual value of the system. 192    3.4 Automated Weather Observing Systems 3.4.1 Introduction An automated weather observing system (AWOS) produces real-time obser- vations during the period when an airport operates (based on arrivals and departures), and it monitors a very specific set of meteorological parameters for generating and disseminating meteorological reports. While most of the flying public is typically unaware of their existence, AWOS systems are criti- cal to flight operations at an aerodrome such as landing approach or take off direction and to determine whether the aircraft operations need to cease until conditions improve. The International Civil Aviation Organization (ICAO) has defined three types of precision approach runways: CAT I, CAT II, and CAT IIIB. All international aerodrome operations use these types of AWOS systems in various combina- tions, ranging from the use of a single CAT type (I, II, IIIB) to a mixture of CAT Aircraft display. Photo: olyniteowl I, II, and IIIB types for multiple runways. While AWOS systems can be com- plex in nature, based on the number of runways and runway configuration, both the World Meteorological Organization (WMO) and ICAO impose a strict design structure. This results in a marked difference between AWOSs and automatic weather stations (AWSs), which are designed for general weath- er forecasting and typically have more latitude in design and specifications. While manufacturers of AWOSs understand the nuances of the regulations, users should have a basic knowledge of these specifications and how they “While most of the flying relate to the equipment that they are purchasing. Key elements include the following: public is typically unaware AWOS networks are part of a regulated service mandated by ICAO. The legal of their existence, AWOS framework created by the Convention on International Civil Aviation (the systems are critical to flight Chicago Convention) includes meteorological services for the safety of civil aviation, permitting states to charge for the providing of weather services operations at an aerodrome to aircraft overflying their national territory for air navigation on a nondis- such as landing approach criminatory basis and to the use of airports. The standards for these mete- orological services are contained within Annex 3 of the Chicago Convention or take off direction and and are identical to WMO Technical Regulations (WMO-No. 49) Volume II. to determine whether the The Chicago Convention does not require states to provide meteorological services for air navigation, but if they decide to do so, they must apply the aircraft operations need standards set out in Annex 3, and the level of such charges for their use and the way they are calculated must be submitted to ICAO (Hodgson 2022, in to cease until conditions process). improve.” Automated Weather Observing Systems    193 A National Meteorological Service (NMS) is not necessarily the airport for both local and international use. In principle, responsible for aviation weather safety in a given country. there are two types of AWOS systems: As a result, data from these AWOSs are not necessarily shared with the NMS. When data from AWOSs are not readily shared, ■ Manned. This type of AWOS has a meteorological observer the NMS should obtain an agreement for AWOS data to be operating the system. provided in real time through a sharing arrangement for use ■ Automated. This type is unmanned, has no observer pres- in NMS weather forecasting. Note that AWOS data are not ent, and is operated in a fully automated manner. generally suitable for climatology-related applications, given the very specific focus of measurements. The manned AWOS uses the same or similar electronic record- ing instruments as those used in a fully automated AWOS; Airports registered at ICAO must follow ICAO and WMO stan- the difference is that in a manned system, the observer is re- dards. Countries can determine their own observational stan- sponsible for generating the meteorological observations and dards and requirements for domestic airports, which means reports. that these airports might not meet ICAO CAT (Category) spec- ifications. However, should an airport choose to register with WMO is responsible for setting specifications for the mete- ICAO, it must follow ICAO standards. orological instruments, measurement procedures, and data processing methods, and for coding the meteorological re- Governments value well-maintained AWOSs highly. This ports. There are two key documents: the Guide to Instruments high importance reflects the value that open and operating and Methods of Observations, WMO-No. 8 (WMO 2018a), airports have on their economies, in particular from tourism and Technical Regulations, Basic Documents No. 2, Volume II: and trade. Although operations are not necessarily 24/7, this Meteorological Service for International Air Navigation, WMO- same emphasis on performance is not shared by the other key No. 49 (WMO 2018c). observing networks: AWSs; upper-air systems, and weather radar systems. A good way to understand the functionality of AWOS systems is from the workflow—starting with data from meteorologi- AWOS maintenance is typically supported financially cal sensors through data processing to output products (re- through cost recovery from aircraft landing fees. This type ports) and users who receive the data and products on their of financial support does not apply to other weather systems, displays: such as AWSs, upper-air systems, and weather radar systems. ■ Collecting of meteorological data (input). The main input One of the most important functions of AWOS is report gen- of data into an AWOS comes from meteorological sensors, eration. The combination of system electronic and manual with additional input provided by the meteorological ob- observations and AWOS software provides continuous, re- server. In some situations, additional data are contributed al-time weather condition information and reports official by weather radar, satellite observations, a lightning detec- observations for aviation that support aerodrome operations. tion network, and so on. ■ Processing and storage of AWOS data (internal). This This chapter will guide the NMS through some of the key function is performed by a system consisting of one or more points to consider when building a sustainable AWOS and computer servers. All meteorological data are processed concludes with some recommendations. for quality control and archived before being compiled into weather reports and other meteorological products. Much 3.4.2 The Purpose of the AWOS System of the processing is automatic and requires no human in- teraction. In addition, metadata, which provide informa- An AWOS system is a meteorological observation system that tion on the operational status of the system components provides real-time monitoring of weather conditions at an and overall system health, are recorded. airport and generates and disseminates weather reports for 194    Automated Weather Observing Systems ■ Generation and dissemination of AWOS products (output). functionality of AWOS systems is illustrated in figure 3.4.1. AWOS systems generate reports from the meteorologi- Starting with the observations (top left), all data are pro- cal data collected which is shared with a variety of users. cessed and stored on the AWOS data processing and storage Within the airport, the reports are disseminated to various servers. Local routine meteorological reports are used inter- internal users (observers, for air traffic control or ATC, fire nally at the airport itself, and others (METAR and SPECI re- department, airport authorities, air lines, and others). How ports) are disseminated internationally. Most users receive the reports and data are transmitted is based on the commu- meteorological information in the form of display-only data. nications infrastructure at the airport. Internationally, data Only the meteorological observer is interactively involved are disseminated as the Meteorological Terminal Aviation with the data in generating the meteorological reports. The Routine Weather Report (METAR) and SPECI reports via network management system is used for technical monitor- aeronautical fixed telecommunications network (AFTN), ing of the AWOS system. The AWOS software also contains Aeronautical Message Handling System (AMHS), or mes- a module for the configuration of the system by a system sage switching system (MSS) communication platforms. The administrator. FIGURE 3.4.1  Schematic of AWOS Functionality Input Processing Output • AFTN Meteorological International • AMHS observations AWOS meteorological reports Sensor and eld • MSS data processing stations & storage Local Radio Technical data meteorological reports PTSN Local routine reports System administration data Technical data System Technician Observer ATC Other administrator Technical monitoring Weather monitoring Weather monitoring Weather monitoring System administration of the AWOS Meteorological reports Users Source: Foeke Kuik, personal communication. Note: Green items are related to meteorological data, orange to technical data, yellow to stage and data processing, pale blue to system administration/ management data. Air traffic control (blue) and Other (gray) receive data for monitoring the weather conditions and for taking decisions based on weather conditions (for example, which runways will be used). AFTN = aeronautical fixed telecommunications network; AMHS = Aeronautical Message Handling System; ATC = air traffic control; MSS (GTS) = message switching system (global telecommunication system); PSTN = public switched telephone network. 3.4.3 Observations in AWOS Systems comprised from a combination of observations (M/D). The blue parameters at the bottom of the table (lines 16–24) are AWOS systems cover a large variety of meteorological param- measurements that are regularly requested in AWOS systems, eters that are needed for aviation meteorological applica- but not required by WMO or ICAO. These measurements are tions. These are listed in table 3.4.1, along with the types of typically made when the NMS responsible for the operational sensors required to obtain the observations in an aerodrome. AWOS system desires surface synoptic observations (SYNOP) The table also shows which parameters are included in the reports from the aerodrome location (which requires addi- METAR (and other reports), and whether these parameters tional observations); they are not used in METAR. are directly measured (M), derived (D) from algorithms, or Automated Weather Observing Systems    195 TABLE 3.4.1  Meteorological Parameters Required for Aviation Meteorological Applications Meteorological Number parameter Sensor (commonly used) M/D In METAR Remarks Required 1 Pressure Capacitive/frequency M QNH QNH/QFE derived from the measured pressure. 2 Dew point Capacitive/dew point M/D DT Calculated from the RH and air temperature. temperature mirror Relative humidity Capacitive M No Often combined with temperature sensor. RH is not required, but it is used to calculate dewpoint. 3 Air temperature PT100 M AT Often combined with RH sensor. 4 Background Background luminance M No Background luminance is required for assessing Luminance sensor visibility in case it is derived from lights and for RVR assessment. 5 Cloud height Ceilometer M Up to 3 layers Layers with CB and TCU are always reported. 6 Cloud amount Ceilometer/observer M Cloud amount coded in METAR as FEW (1–2 oktas), SCT (3–4 oktas), BKN (5–7 oktas), OVC (8/8 oktas), NSC, NCD (auto-METAR), derived. 7 Cloud type Observer D CB or TCU Cloud type is not possible in a reliable way with instruments or cameras yet. Only reported in METAR when cumulonimbus (CB) or towering cumulus (TCU) is present at the airport. 8 MOR (Visibility) Transmissometer/forward M Prevailing Background luminance is required for calculat- scatter/observer visibility ing Aeronautical Visibility based on MOR and background luminance. Runway Visual Transmissometer/forward M/D RWY/RVR RVR is derived from measurements of visibil- Range scatter/observer ity, background luminance, and runway light settings. Runway light setting is external or manual input to the AWOS system. Reported for RWYs in use (max. 4). 9 Precipitation inten- Tipping bucket, forward M In present In mm per hour. In weather code, coded as +, −, sity (liquid or solid) scatter sensor weather section or no intensity. 10 Precipitation inten- M In present In mm per hour. In weather code, coded as +, −, sity (solid) weather section or no intensity. 11 Precipitation type Forward scatter sensor D In present None. weather section 12 Wind gust Ultrasonic, cup anemome- M Conditional None. ter, propeller 13 Wind speed Ultrasonic, cup anemome- M Yes None. ter, propeller 14 Wind direction Ultrasonic, vane, M Yes None. propeller 15 Sea surface PT100 M In supplemen- For platform or airports at a seaside. temperature tary information – groups Optional 16 Precipitation Tipping bucket, forward M No Over a certain period of time. Not in METAR. amount scatter sensor 17 Snowfall depth Ultrasonic, laser M No Only measured in regions with much snowfall. continued 196    Automated Weather Observing Systems (Table 3.4.1 continued) Meteorological Number parameter Sensor (commonly used) M/D In METAR Remarks 18 Snow cover depth Ultrasonic, laser M No None. 19 Lightning Lightning antenna M Can be used in Lightning at and in the vicinity of the airport present weather is one of the present weather codes. Although desirable, lightning data are not always includ- ed in AWOS systems. Lightning data can be obtained from a lightning sensor/antenna at the airport, or it can be purchased as a service, and then used in AWOS systems and inserted in the reports. 20 Global radiation Pyranometer M No Sometimes present. 21 Sunshine duration Sunshine duration record- M/D No Sometimes present. er or pyranometer 22 Soil temperature Sub-surface PT100 M No Sometimes. Of interest in cold climate for infor- mation on runway surface icing conditions. 23 Freezing rain Frequency measurement M No Only in cold conditions. Source: Foeke Kuik, personal communication. Note: The sea surface temperature only applies for aerodromes at a seaside. Items 16 through 23 are not required for the METAR/SPECI products, but can still be required, depending on local conditions and if needed for safe operations. AT = air temperature; AWOS = automated weather observing system; BKN = broken clouds, 5–7 oktas; CB = cumulonimbus; D = derived from algorithms; DT = dewpoint temperature; FEW = few clouds, 1–2 oktas; METAR = Meteorological Terminal Aviation Routine Weather Report; M = measured directly; MOR = meteorological optical range, NCD = nil clouds detected; NSC = nil significant clouds, OVC = overcast, 8/8 oktas, QNH/QFE = altimeter setting/station pressure, RH = relative humidity; RVR = runway visual range; RWYs = runways; TCU = towering cumulus. 3.4.4 Requirement Specifications for AWOS ■ Dissemination of international reports. Many aerodromes Systems apply aeronautical fixed telecommunication network (AFTN) systems, but some use the global telecommunica- ICAO provides specifications and regulations on what param- tion system (GTS) through WMO via a message switch, es- eters must be observed and at what frequency, where sen- pecially when the NMS is responsible for the AWOS system. sors and equipment are to be installed, and how data should The client who procures the AWOS (usually an NMS or ATC, be presented to observer. Overall, it regulates the function- sometimes an airport owner) must specify what system is al aspects of AWOS. WMO is responsible for the functional used, as the correct interface must be provided in the AWOS meteorological aspects of AWOS, including requirement system. specifications for the sensors, acceptable uncertainties, the ■ Design of the field stations. If an aerodrome has only one definition of algorithms used to process data, and the content runway, the station design is relatively simple. If there of meteorological reports. are multiple runways, deciding where field stations and sensors are installed is more complex, especially when However, several aspects of AWOS systems are not defined runways are parallel or cross each other. Sensors may be by either ICAO or WMO; notably, those related to engineering combined sited at a single site to serve multiple runways. design. For example: ■ How to build the field stations. A field station can consist of a simple serial-to-internet protocol (IP) converter with ■ Data communications between the field stations and the a digital subscriber line (DSL) module, forwarding all data AWOS data processing system. In many aerodromes, cop- from the sensor in an IP format directly to the AWOS pro- per cable or fiber optics is available between these loca- cessing system. Other configurations use a full data logger, tions; where no cable connection exists, radios may be used. which might or might not be required. Data loggers archive Automated Weather Observing Systems    197 data in onboard memory, and since all data are used in real sensors and the best measurement sites and conditions, (2) time, “old” data might not be relevant. breakthrough is acceptable for general application, and (3) ■ Installation of redundant sensors. Redundant sensors are threshold is the minimum acceptable uncertainty. not required but should be installed as a best practice. ■ Battery backup. This might be required but might not be 3.4.5 AWOS Design: Instrument Runways and feasible for sensors that require heating. Precision Approach Runways ■ Transient and surge protection. This is not required, but as expensive sensors and equipment are vulnerable to ICAO distinguishes among different runway categories (see damage from lightning surges, it is highly recommended ICAO 2018a, section 4.2.8.3) based on a combination of de- to install such devices at the sensors and field stations, cision height (DH)–that is, the minimum height at which a especially when there are long (copper) cable runs. pilot must initiate a missed approach if the required visual ■ Architecture for processing systems. WMO defines the algo- approach reference has not been established, visibility (VIS), rithms for data processing, but it does not define an architec- and runway visual range (RVR) restrictions. Because the re- ture for the processing systems. Traditionally, observational quirements for VIS and RVR observations in aviation weather data from field stations are sent to an AWOS server, where systems depend on the category of the runway by which it is they are processed. Data are then forwarded to the user dis- installed, the appropriate AWOS design depends on the cat- plays and observer user workstation, where observations egory of runway by which it will be installed. The different are completed and prepared for dissemination via reports. categories of runways and their dependence on DH, VIS, and ■ Installation of AWOS servers. The AWOS servers may be RVR are shown in figure 3.4.2, which details the two types of installed in a hot-redundant configuration, or there could instrument runways: (1) a precision approach runway, which be a “cold” spare server. uses an instrument landing system or a precision approach radar with vertical (glide path) and lateral guidance (course AWOS systems typically have many sensors, field stations, deviation); and (2) non-precision approach runways, which servers, and other components that require monitoring, use air navigation facilities with only lateral guidance. management, and configuration. Those that combine such functionality are often referred to as network management Both the non-precision approach and precision approach run- systems (NMs). An NMs allows system administrators to con- ways (see ICAO 2018a, section 1.1, Definitions) are equipped figure users, passwords, sensors and backup sensors, screen with “visual aids and non-visual aid(s),” which are a combi- layouts, and so on. Because the NMs monitors all the compo- nation of VIS and RVR instruments and instrument landing nents and parts in the AWOS system, technicians can use the equipment such as the instrument landing system (ILS) and NMs to monitor the status of all components. the microwave landing system (MLS). In aviation weather systems, only VIS and RVR instruments are relevant; ILS/MLS As for required specifications for AWOS sensors, these are pub- systems are not part of aviation weather systems but are in- lished by ICAO, WMO-No. 8, and the WMO Integrated Global stead handled by air traffic control (ATC). Weather Observing System/Observing Systems Capability Analysis and Review Tool (WIGOS/OSCAR). The conditions for precision approach runways are shown in table 3.4.2. Usually, the category of the runway is specified In principle, the WMO requirement specifications (WMO from CAT I to CAT IIIB, and matching instrumentation for VIS 2018a) are the highest-level specifications (of the three en- and RVR measurements are required. CAT IIIC runways re- tities) for (automated) weather observations; these are typ- quire RVR measurements down to 0 meters. However, most ically used in many tenders. WIGOS/OSCAR provides the RVR sensors will only report RVR down to 50 meters (CAT specifications in terms of uncertainties for three levels: (1) IIIB) and are capable of reporting lower values with the re- goal is the highest level, which can be achieved with the best quired uncertainty for RVR measurements. 198    Automated Weather Observing Systems FIGURE 3.4.2  Types of Instruments Runways Instrument Runway ICAO Annex 14, 1.1 Type A: DH > 75m (250 ft) Type B: DH ≤ 75m (250 ft) Non-Precision Approach runway Precision Approach runways CAT I CAT II CAT IIIA NPA Runway • DH < 30 m or no DH restriction if RVR ≥ 175 m • DH ≥ 60 m • 30 m ≤ DH ≤ 60 m • Vis > 1,000 m • RVR ≥ 175 m • VIS > 800 m or RVR ≥ 550 m • RVR ≥ 300 m CAT IIIB • DH < 15 m or no DH restriction if 50 m ≤ RVR ≤ 175m • 50 m ≤ RVR ≤ 175 m CAT IIIC • No DH limitations • No RVR limitations Source: ICAO 2018c, Section 1.1. Note: Orange refers to NPA runways; gray to PA runways. DH = decision height; NPA = non-precision approach; PA = precision approach; RVR = runway visual range; VIS = visibility. TABLE 3.4.2  Visibility, RVR, and Decision Height Conditions for Precision Approach Runways Category Visibility RVR Decision height (DH) NPA RWY > 1,000 m NA > 75 m (250 ft) CAT I > 800 m ≥ 550 m ≥ 60 m CAT II Undefined ≥ 300 m < 60 m but no less than 30 m CAT III A Undefined > 175 m < 30 m, or no DH with RVR ≥ 175 m B Undefined 50 m ≤ RVR ≤ 175 m < 15 m, or no DH with RVR < 175 m C Undefined No limitation No limitation Source: ICAO 2018c. Note: DH is not relevant for aviation weather systems but listed for completeness. DH = decision height; NPA RWY = non-precision approach runway. Automated Weather Observing Systems    199 3.4.6 Site Selection Criteria and Siting WMO provides guidance on all meteorological parameters Requirements required for the MET REPORT and SPECIAL (WMO 2018a, Section 4.6, Recommendations), as summarized in table The siting requirements and criteria that are addressed in this 3.4.3. For RVR measurements (WMO 2018a, Section 4.6.3.4), section are very well documented by WMO. WMO says that RVR assessments shall be representative of: 3.4.6.1 Observations for Local Routine and Special ■ The touchdown zone of runways for non-precision or CAT I Reports and METAR/SPECI instrument approach and landing operations ■ The touchdown zone and the mid-point of runways for CAT As one of its main tasks, the AWOS generates reports for local II instrument approach and landing operations use at the airport and for international dissemination. These ■ The touchdown zone, the mid-point, and stop-end of include: runways for CAT III instrument approach and landing operations. ■ Local routine and local special reports. These reports pro- vide meteorological information for the airport for local When landing, the touchdown zone (TDZ) is the location on use only and are not intended for international sharing. the runway after passing the threshold, where airplanes’ They give the information that the meteorological users wheels first touch the tarmac. The mid-point position (MID) (observers) see displayed on their screens. The data go is about the middle position of the runway, but it should not into the local routine report called MET REPORT and the be further than 1,500 meters from the threshold. The end po- local special report called SPECIAL, which are for local use sition, or STOP-END, is the far end of the runway, located at at the airport only and contain the complete meteorolog- the position of the TDZ if the runway is used from the other ical information for all runways. There are separate MET end. Some meteorological parameters are required for specif- REPORTs and SPECIALs for arriving and departing aircraft. ic locations for departing aircraft, while others are required ■ METAR and SPECI reports. These are similar to the MET at specific locations for arriving aircraft; still others are rep- REPORT and SPECIAL, but they are intended for interna- resentative of the airport and vicinity. tional dissemination. The meteorological information is generally representative of the airport as a whole, whereas RVR is for only the runway in use. TABLE 3.4.3  Observation Location Requirements for Local Routine Reports, Special Reports, METAR, and SPECI For local routine and special reports Parameter Arriving aircraft Departing aircraft METAR/SPECI WS/WD TDZ Along RWY Whole RWY or RWY complex Visibility TDZ Along RWY Aerodrome RVR ■ Non precision and CAT I: TDZ Not specified RWY in use ■ CAT II: TDZ & MID ■ CAT III: TDZ, MID, STOP-END Clouds Threshold of RWY in use Threshold of RWY in use Aerodrome and vicinity Present Weather Aerodrome Aerodrome and vicinity AT/Dewpoint Whole RWY complex Whole RWY complex Pressure (QNH & QFE) Whole RWY complex Whole RWY complex Source: WMO 2018b. Note: AT = air temperature; STOP-END = stop-end position of the runway, MID = mid-point position; RVR = runway visual range; RWY = runway; TDZ = touchdown zone; WD = wind direction, WS = wind speed. 200    Automated Weather Observing Systems 3.4.6.2 Locations for Observation Sites siting should be done correctly. In general, many locations in Observations include non-runway-specific measurements of, an aerodrome should be suitable. In most installations, instru- ments measure temperature and relative humidity (RH), and for example, temperature, relative humidity, barometric pres- dewpoint is calculated from these two values. sure, and precipitation. In general, sites should be chosen to be representative for the runway complex, which means Locations for pressure sensors. From barometric pressure they should not be close to large structures (such as build- measurement(s), the QFE (atmospheric pressure at aero- ings) that can influence the measurements. They also should drome elevation or at runway threshold), and the QNH (atmo- not be on elevated structures (such as rooftops), but rather at spheric sub-scale setting to obtain elevation—or, at runway ground level, and should be fully exposed to meteorological threshold, hectopascal, hPa) are calculated. Barometric pres- conditions. Sensors should be installed far enough apart that sure does not vary much across an aerodrome, so the mea- they do not affect each other. surements are not runway specific, and sensors are usually located in the met garden. Because barometric pressure mea- Often sensors for non-runway-specific observations are surements are important for airports, barometers are com- grouped together and installed in a meteorological garden monly installed in redundant configurations. In addition to (met garden) located somewhere within the runway complex, QFE and QNH, there are two other pressure parameters that but not necessarily close to a runway. If 10-meter wind masts are used for airports at higher altitudes: pressure altitude and are included in a met garden, they should at least be 106 density altitude. meters (frangible 10-meter mast) or 220 meters (non-frangi- ble 10-meter mast) away from runway center lines. Visibility Locations for visibility sensors. (Prevailing) visibility is and present weather (forward scatter) sensors are usually required at the TDZ and along the runway for local routine installed on 2.5-meter masts, and if sufficiently far from the and special reports, and it should be representative for the runway center line (> 192 meters), they do not have to be in- whole aerodrome for the METAR/SPECI. Visibility and RVR stalled on frangible masts. Top lights might also be required. can change considerably across an aerodrome when mist or fog is present. Therefore, several visibility sensors are typi- Sensors for measuring temperature and relative humidity are cally installed in one or multiple positions along the runway. typically installed in a radiation shield at 1.5–2 meters high Forward scatter sensors and transmissometers can measure (non-frangible) masts. A rain gauge is usually installed on a visibility and are used for RVR but require the addition of a concrete slab with the orifice at approximately 1 meter above background luminance sensor. Most forward scatter sensors ground level. Best results for rain gauges require wind shield- also provide present weather data. ing around the rain gauge, the top of which should be at the same level as the top of the rain gauge. Barometric pressure Locations for present weather sensors. Present weather (pre- sensors can be installed inside the enclosure that is used for cipitation type, intensity) is measured with forward scatter the data logger and communication equipment. These barom- sensors, sometimes with some additional sensors. Therefore, eters require an open connection to outside of the enclosure, the present weather observations are representative of the and best results are obtained by using a static pressure port locations where the forward scatter sensors for visibility and that minimizes the effect of wind flow on the pressure mea- RVR are installed; sometimes, visibility and present weather surements. In some airports, backup sensors are installed in sensors are installed in the met garden to be representative the met garden, or sometimes a fully instrumented backup for the whole airport (METAR/SPECI). Visibility/RVR sensors met garden is installed. installed near a runway are considered representative of only that runway. Forward scatter present weather sensors can Locations for temperature and dewpoint sensors. These mea- also provide precipitation intensities and amounts (totals), surements must be representative of the whole runway com- but the uncertainties in such measurements are often fairly plex for all reports. They can be located in the met garden or in high. Precipitation amount and intensity are typically mea- a field station near a runway. Temperature and dewpoint mea- sured using tipping bucket rain gauges, which are installed in surements should be representative of a larger area, and thus the met garden as they are not considered runway dependent. Automated Weather Observing Systems    201 Locations for RVR observations. RVR sensors should be in- more than 1,000–1,500 meters from the (TDZ) threshold. The stalled on 2.5-meter-high frangible masts. The minimum dis- RVR at the STOP-END position should be no more than 300 tance from the center line of the runway for these installation meters from the STOP-END threshold. Table 3.4.5 shows the sites is 83 meters, the maximum distance is 120 meters—and number of RVR sensors required by runway type for the local the sites should not penetrate the inner transitional surface routine and special reports, which varies from one to three, (table 3.4.4). The RVR at TDZ should be no more than 300 with some mandatory and others just recommended. meters from the threshold; the RVR at the MID position, no TABLE 3.4.4  Installation Site Recommendations for RVR Instruments Typical equipment Transmissometer and/or forward scatter meter Typical dimensions of equipment Usually two units: transmitter and receiver. In the case of transmissometer, they are separated over baseline (length of the order of 20 m depending on range of visibilities to be assessed). Height of units approximately 2.5 m (7.5 ft) above the runway. Solid foundation plinths required. Operational area for which element is Up to four transmissometers or forward-scatter meters per runway (that is, runways for which to be representative RVR is required), for the touchdown zone, the mid-point, the second mid-point (if required), and the stop-end of the runway. Siting provision in WMO 2014 Not more than 120 m laterally from runway center line. For the touchdown zone, the mid-point, and the stop-end, units should be 300 m, 1,000 m, and 1,500 m along runway from threshold, respectively. Remarks Should be sited within 120 m laterally from runway center line but should not infringe the obstacle-free zone (OFZ) (that is, inner transitional surface) for precision approach runways. Should be a frangible structure; for example, tubular supports and shearing bolts at foundation. Source: ICAO 2017, Appendix 2, Table A2-3. Note: OFZ = obstacle-free zone; RVR = runway visual range. TABLE 3.4.5  Number of Mandatory and Recommended RVR Sensors per Runway Type Number Runway type Location of RVR sensors Requirements NPA and CAT I TDZ 1 Recommended and not Mandatory CAT II TDZ and MID 2 Mandatory STOP-END 1 Recommended and not Mandatory CAT III TDZ, MID and STOP-END 3 Mandatory If the distance between the two thresholds is more than 3,000 m, two MID positions are required, meaning that four RVR units must be installed. Note: Sometimes RVR units are installed in redundant configurations, which can lead to six or more RVR units per runway. Source: ICAO 2017, Appendix 2, Table A2-3. Note: CAT = category; MID = mid-point position; NPA = non-precision approach, RVR = runway visual range; RWY = runway; STOP-END = end position, TDZ = touchdown zone. 202    Automated Weather Observing Systems In addition, background luminance—the amount of light com- special reports, wind measurements are runway specific, ing from the sky—must be measured to assess the RVR as it whereas for METAR/SPECI, wind measurements that are rep- is an input parameter in the algorithm for calculating RVR. resentative for the RWY complex are required. For CAT II and It can be measured anywhere in the runway complex, and a CAT III runways, at least two wind measurement positions background luminance sensor is typically installed on a cross are recommended, with one at the TDZ and one at the STOP- arm of a forward scatter sensor, or on a pole of a transmis- END position. Sometimes a third wind mast is installed in the someter, to measure the light coming from a fixed direction MID position as well. in the sky. RVR sensors usually are located in the TDZ, MID, and STOP-END positions, and this is where the background WMO recommends wind sensors to be installed on 10 m luminance sensor is usually installed as well. At least one masts. However, ICAO also accepts 6 m masts. The masts can background luminance sensor is required, for example, to be be frangible or non-frangible (fixed). When they are frangible, installed at the TDZ RVR unit, and often a second unit is in- they can be located between the inner transactional surface stalled as a backup sensor at the MID or STOP-END position and the transactional surface (see table 3.4.6). RVR. Because the masts should not infringe on the transactional Locations for surface wind speed and direction observa- surfaces, the closest installation distances from the center tions, wind masts. For non-precision approach runways and line of the runway differ for the four types of masts. The masts CAT I runways, one wind mast with wind speed and wind di- are allowed to infringe the obstacle-free zone (OFZ), but only rection measurements usually suffices. For local routine and under exceptional conditions. TABLE 3.4.6  Installation Site Recommendations for Wind Speed and Direction Measurements Item Description Typical equipment Anemometer and wind vane, ultrasonic wind sensor. Typical dimensions of equipment Usually mounted on tubular or lattice mast 10 m (30 ft) high. Single tube mast for both instruments appropriate in proximity to runways. Operational area for which ele- Conditions along the runway and TDZ in local routine and special reports; conditions above the ment is to be representative whole runway (complex) in METAR and SPECI. Where prevailing wind varies significantly at different sections of the runway, multiple anemometers are recommended. Siting provision in WMO 2018a No specific provisions so long as observations are representative of relevant operational areas. Remarks Siting will be governed by obstacle limitation surfaces and local prevailing surface wind regime. Generally speaking, if the wind field over the aerodrome is homogeneous, one strategically sited anemometer may suffice, preferably sited so as not to infringe transitional surfaces. However, depending on local conditions, it may be necessary to locate a frangible and lighted mast within the runway strip. Only in exceptional circumstances should the mast infringe the OFZ (that is, inner transitional surface) for precision approach runways. In the latter case, the mast must be frangible, lighted, and preferably shielded by an existing essential navigation aid. The site must not be affected by buildings, and so on, or by aircraft operations (for example, jet efflux during taxiing). Source: ICAO 2017, Appendix 2, Table A2-3. Note: OFZ = obstacle-free zone; TDZ = touchdown zone; WMO = World Meteorological Organization. Automated Weather Observing Systems    203 TABLE 3.4.7  Installation Distances from the Center Line of the Runway for RVR Sensors and Wind Masts, Frangible and Non- Frangible Mounting Installation distance Frangible mast Non-frangible mast Sensor Mast length (m) Minimum Maximum Minimum Maximum RVR 2.0 81 120 n.a. n.a. 2.5 83 120 n.a. n.a. Wind 6.0 93 UD 192 UD 10.0 105 UD 220 UD Source: ICAO 2018a, Table 4-1. Note: The angles between the inner transitional surface (ITS) and the transitional surface (TS) and the runway surface are 18.4o and 8.1o, respectively. The ITS and TS are 75 m and 150 m from the center line of the runway, respectively. The green cells are the preferred options for RVR and wind masts installations. n.a.= not applicable; RVR = runway visibility range; UD = undefined. Wind sensors should be installed on 10-meter-high frangible installations are at the middle marker position (extended di- masts. The minimum distance from the center line of the run- rection of the runway), or somewhere within the runway com- way for these installation sites is 105 meters for a frangible plex (such as at the TDZ and STOP-END field station positions) mast and 220 meters for a non-frangible mast (table 3.4.7). A (table 3.4.8). Cloud cover usually does not vary much over the 6-meter frangible mast can be installed 93 meters from the airport area for most cloud types, although this is dependent center line of the runway, a non-frangible 6-meter mast from on the cloud type. For example, cumulus clouds or towering 192 meters. The wind masts (10-meter frangible masts) can cumulonimbus clouds can differ markedly on small spatial be installed in the same location as the RVR at TDZ and STOP- scales. Because ceilometers sense the cloud base in only a END, although they have to be a little farther out from the small portion of the sky above the sensor, localized clouds center line of the runway. might not be detected by the ceilometer. In the processing of ceilometer measurements, it is common for averaging over Locations for ceilometer installations. Measurements of the 30-minute intervals to be applied, which decreases the risk of cloud base should be representative of the approach area missing clouds. Another option is to use multiple ceilometers for local routine (and special) reports and for the airport and a spatial averaging algorithm. region for METAR and SPECI. Suitable locations for these TABLE 3.4.8  Installation Site Recommendations for Cloud Measurements Item Description Typical equipment Ceilometer. Typical dimensions of equipment Usually less than 1.5 m (5 ft) high but rather solid structure including foundation plinth. Operational area for which element is Generally representative of the approach area in local routine and special reports and of the to be representative aerodrome and its vicinity in METAR and SPECI. Siting provision in WMO 2018a At a distance of less than 1,200 m (4,000 ft) from the landing threshold. Remarks May be located at the middle marker site or within the runway strip but preferably not infringing the OFZ (that is, the inner transitional surface) for precision approach runways. Source: ICAO 2018a. Note: OFZ = obstacle-free zone; WMO = World Meteorological Organization. 204    Automated Weather Observing Systems Ceilometers should be installed at the middle marker site of position will be further than 1,500 meters from the MID po- the instrument landing system or at a distance of less than sition. The bottom panel shows a situation where the TDZ is 1,200 m (4,000 ft) before the landing threshold. The instru- on the east side, reversed from the situation in the top panel. ments should not infringe on the OFZ. If a middle marker is If the runway is used from both sides, it can be seen that two not possible, they should be installed within the runway com- MID positions are required, as they do not coincide for run- plex (such as in a met garden or at a sensor group near a TDZ). way use from the east and the west. If the distance between thresholds is no more than 3,000 meters, one MID position 3.4.6.3 Summary Sensor for Installation Sites suffices. An overview of potential sensor locations for runway specific sensors is shown in figure 3.4.3, which illustrates the situa- In addition, table 3.4.9 provides an overview of installation tion for a runway with the distance between the thresholds sites for all categories runways. This is a theoretical overview, of more than 3,000 meters. The top panel shows a situation as there will always be specific requirements from the airport where the TDZ is on the west side. For CAT II and CAT III on instrument installation locations, local climatological con- runways, the third RVR sensor is located at the STOP-END po- ditions matter, and buildings and other structures in an air- sition. The RVR at the MID position should not be further than port can cause deviations from the best theoretical solution. 1,500 meters from the threshold. The RVR at the STOP-END TABLE 3.4.9  Installation Locations for Weather Sensors for Non-Precision Approach Runways band CAT I, CAT II, and CAT III Runways NPA and CAT I CAT II CAT III Parameters/sensor TDZ MID END MG TDZ MID END MG TDZ MID END MG WS/WD 1 1 1 1 1 AT & DT 1 1 1 Pressure 1 1 1 Ceilometer 1 1 1 1 1 Visibility 1 1 1 1 1 RVR 1 1 1 1 1 1 BL 1 1 1 1 PW 1 1 1 1 1 1 1 1 Source: Foeke Kuik, personal communication. Note: Complies with ICAO and WMO recommendations and requirements for best instrument installation locations. The colors are used to make it easy to distinguish between the different categories. The table shows the number of sensors recommended in the different locations. Blank cells indicate that there is no recommendation to install a sensor there. Recommendations for non-precision approach runways and CAT I runways are the same. AT & DT = air temperature & dewpoint temperature; BL = background luminance; CAT = category; END = stop-end position of the runway; ICAO = International Civil Aviation Organization; MG = met garden; MID = mid-point position; NPA = non-precision approach; RVR = runway visual range; TDZ = touchdown zone; WD = wind direction; WS = wind speed. Present weather sensors are usually combined in the visibili- operational use, the other as a backup sensor. The best loca- ty sensor. In that case, the present weather sensors are in the tion for a ceilometer is at the middle marker position, but this same locations as the visibility and RVR sensors. Sometimes is not always possible. Installation(s) at TDZ and/or STOP- a separate present weather sensor is installed in the met END positions are good alternatives. garden. One background luminance sensor is installed for Automated Weather Observing Systems    205 FIGURE 3.4.3  Locations of the RVR and Wind Sensors along a CAT IIIB Runway with More Than 3,000 Meters between the Thresholds Source: Foeke Kuik, personal communication. Note: If the distance between the thresholds is 3,000 m or less, one MID position is sufficient; if it is more than 3,000 m, two different MID positions are required. The maximum distance between the MID and STOP-END position (for RVR instruments) is not defined by ICAO. The installation distances of 83–120 m from the center line of the runway for RVR instruments is for 2.5-m frangible masts, the > 105 m for wind masts is for 10-m frangible masts. For CAT I and CAT II runways, only one or two RVR sensors may be present at TDZ and TDZ/MID positions, respectively. Ceilometers are recommended at the middle marker position, and if this is not possible, at a distance of less than 1,200 m (4,000 ft) before the landing threshold. CAT = category; ICAO = International Civil Aviation Organization; MID = mid-point position; RVR = runway visual range; STOP-END = end position; TDZ = touchdown zone. 3.4.6.4 Instrument Installation Sites in Practice ■ If one set of wind, visibility, and RVR measurements is re- Many aerodromes have one runway, and for them, the above quired near the TDZ or STOP-END position, often all sen- guidance can be applied. But other considerations often arise sors are installed at that location for convenience. for installing sensors or field stations with groups of sensors. ■ Ceilometers can be installed at the middle marker po- For example, an aerodrome might have multiple crossing sition—a good place for cloud measurements in the ap- runways, in which case a design for the instrument field sta- proach area of the runway. Moreover, alternating current tions must be made based on how the runways are operated. (AC) power is generally available there, as are data com- Often, field station equipment can be shared between run- munication cables. Alternatively, ceilometers can be locat- ways. However, the presence of objects (such as buildings) ed at the TDZ and STOP-END field station positions. can make it impossible to obtain high-quality meteorological ■ Field stations require AC power. Many instruments use measurements, requiring alternative locations for the field sensor heating, blowers, and so on. As a result, solar pan- stations. els cannot generate sufficient power, unless many large solar panels are combined with large-capacity batteries. If Although each case must be evaluated on local conditions, in AC power must be brought to new field station locations, practice, the following installation patterns emerge: cost can be an issue. 206    Automated Weather Observing Systems ■ Data communication cables and fiber optics are preferred 3.4.6.6 Land Ownership over radio communication, as cables and fiber optics are In general, the land on which an airport is located is owned more reliable than wireless communication methods. by the airport. Installation of sensors or equipment can be Radios in general also need a clear line of sight, which may done with permission of the landowner, considering the re- be an issue at aerodromes where the vertical tails of air- strictions for obstacles as described in section 3.4.6.2. planes are known to block sightlines. 3.4.7 AWOS Design for a Single Runway: CAT I, 3.4.6.5 Access to Installation Sites CAT II, CAT III Airports As can be seen from the sections above, sensors in aero- dromes are installed close to the runway(s) and taxiways. CAT I runway. This type of runway has equipment only on Technicians in airports usually receive training and a permit the TDZ side (figure 3.4.4). Often, all the non-runway specific to be allowed to drive on runways and taxiways. In general, sensors are installed here as well, so that only one data con- such technicians can perform the regular maintenance of the nection is required from a field station to the AWOS server. It sensors and field station equipment. Access for others can be is always recommended to install barometers and other criti- a problem, unless accompanied by a technician with a permit. cal instruments in a redundant configuration—with the client deciding the level of redundancy required. FIGURE 3.4.4  Example of an AWOS System at an Airport with a CAT I Runway Source: Foeke Kuik, personal communication. Note: AT = air temperature; ATC = air traffic control; FO = fiber optics; KVM = keyboard/video/mouse; LAN = local area network; MID = mid-point position; P = pressure; Precip = precipitation; PW = present weather; RH = relative humidity; RVR = runway visual range; SR = solar radiation, STOP-END = end position; TDZ = touchdown zone; TS = thunderstorm; UPS = uninterruptable power supply; VIS = visibility; WAN = wide area network, WD = wind direction, WS = wind speed. Automated Weather Observing Systems    207 CAT II runway. This type of runway has wind measurements garden in this design. Often CAT II runways also include a at both ends, but only the TDZ and MID position have RVR. In third RVR at the STOP-END position if the runway is used for figure 3.4.5, the runway-independent equipment is combined arrivals from both ends. with the RVR at the MID position; there is no separate met FIGURE 3.4.5  Example of an AWOS System at an Airport with a CAT II Runway Source: Foeke Kuik, personal communication. Note: AT = air temperature; ATC = air traffic control; FO = fiber optics; KVM = keyboard/video/mouse; LAN = local area network; MID = mid-point position; P = pressure, Precip = precipitation; PW = present weather; RH = relative humidity; RVR = runway visual range; SR = solar radiation; STOP-END = end position; TDZ = touchdown zone; TS = thunderstorm; UPS = uninterruptable power supply; VIS = visibility; WAN = wide area network; WD = wind direction; WS = wind speed. CAT III runways. This type of runway has wind measurements with forward scatter sensors, all RVR positions can also pro- and RVRs at the TDZ/STOP-END positions (figure 3.4.6). The vide present weather measurements. MID position also has an RVR sensor. Since RVR is measured 208    Automated Weather Observing Systems FIGURE 3.4.6  Example of an AWOS System at an Airport with a CAT IIIB Runway Source: Foeke Kuik, personal communication. Note: This configuration also applies to CAT IIIA and CAT IIIC runways. AT = air temperature; ATC = air traffic control; FO = fiber optics; KVM = keyboard/video/ mouse; LAN = local area network; MID = mid-point position; P = pressure; Precip = precipitation; PW = present weather; RH = relative humidity; RVR = runway visual range; SR = solar radiation; STOP-END = end position; TDZ = touchdown zone; TS = thunderstorm; UPS = uninterruptable power supply; VIS = visibility; WAN = wide area network; WD = wind direction; WS = wind speed. In figure 3.4.6, the non-runway-specific measurements are 3.4.8 Installation Sites installed in the met garden. This can be a separate location, or it can be combined with any of the three field stations near There are several possible installation sites for sensors, the runaway. The met garden in figure 3.4.6 has sensors for which are discussed in detail in this section. temperature/relative humidity, pressure sensors (dual cell), a thunderstorm (TS) sensor for lightning detection, a rain 3.4.8.1 Weather Station, Met Garden gauge, and a global radiation sensor for the measurement The sensors in an airport met garden are usually non-run- of solar radiation. The ceilometers are shown at the mid- way-specific sensors such as those measuring baromet- dle marker position. Ceilometer data can be transferred to ric pressure, air temperature, relative humidity, dewpoint the server by cable (if present, copper or glass fiber) or by (calculated), precipitation, lightning detection, and soil radio. A connection to the runway light system is required temperature. Often met gardens also include an extra wind (not shown) to include runway light intensities into the RVR measurement position. calculations. For time synchronization, a GPS clock can be connected to the server (not shown), or a network timeserver can be used. Automated Weather Observing Systems    209 The meteorological sensors are connected to sensor interfac- system ingests data from all of the sensors, performs all me- es and data communication equipment. Redundant commu- teorological calculations and derivations, stores data, and nication equipment can be applied, and an uninterruptable performs system administrator functions. All data communi- power supply (UPS) can be installed, along with cabinet heat- cation equipment is installed in the equipment room. A UPS ing. AC power is the best option, as solar power is usually should be installed to filter AC for power surges and to shut insufficient to support cabinet and sensor heating. down the AWOS servers in case AC power is going to be un- available for a longer time. For data communication, copper cables, fiber optics, or radio can be used. Note that for radio communication, restric- If support for ICT infrastructure for the servers is likely to tions may apply for frequency bands and power. Also, free be difficult, it is possible to employ a cloud-based AWOS. In sightlines are required for radio communication, although such a system, computer resources (including hardware, data repeaters can be used if sightlines are obstructed. Data com- storage, and software) are located remotely and accessed via munication devices should be installed in field cabinets as internet connection. Observational data from field stations complete operational systems. are forwarded to the cloud-based AWOS, where they are pro- cessed, stored, and made available to users, who also access 3.4.8.2 Touchdown Zone and Stop-End Position these data via the internet. RVR should be measured near the TDZ for a CAT I precision approach, and at TDZ, MID, and END positions for CAT II and 3.4.8.5 Meteorological Office CAT III A and B, according to ICAO. It is standard practice to The meteorological observer usually works in the met of- use a forward scatter sensor for this measurement. For RVR fice, where the observer workstation(s) will be installed. The calculations, both a background luminance sensor and an observer connects to the server (or the cloud-based AWOS) interface to the runway lighting system are needed. For the through a web browser or client application to access AWOS latter, a serial or TCP/IP connection is preferred. But if this is data. The observer is responsible for the METAR/SPECI, local unavailable, a field station can measure the physical current MET REPORTs, SPECIALs, and so on, which are generated by through the runway light cables, so long as it has AC power the AWOS software. In case of an unmanned operation, an and a data connection to the AWOS server. observer workstation should be installed to allow monitoring of the automatic operation of the AWOS system. Wind sensors are to be installed at runway TDZ and STOP- END positions, unless a site survey concludes that these po- 3.4.8.6 Control Tower sitions are strongly influenced by topography, buildings, or In the control tower, one or two user displays are usually in- other disturbing factors. stalled, as required. The ATC displays are display-only sys- tems: ATC users can view meteorological information, but 3.4.8.3 Middle Marker they cannot interact with the system. These displays can be Ceilometers for runways with an instrumented precision ap- standard 24- to 27-inch monitors, although sometimes con- proach should be installed at the middle marker position, but sole mounted touch screen displays are preferred. A backup they can be installed anywhere as long as measurements are display is also typically provided if the sensor and station representative of the approach. The middle marker usually configurations permit this feature. is a convenient position because power and data cables are usually already present there. Field stations at the TDZ and 3.4.9 Additional AWOS Equipment STOP-END positions are good alternatives. In addition to the measurement equipment located at various 3.4.8.4 Equipment Room or Cloud-Based AWOS sites, AWOS systems also require a base amount of computer In the equipment room, the server rack will be installed hardware to ingest, process, and display the data; this is dis- with either a single or dual server configuration. This server cussed in the three sections below. 210    Automated Weather Observing Systems 3.4.9.1 Servers ■ Administrative workstation. This is configured for system The AWOS server system is installed in a (rack-mounted) sin- monitoring, maintenance tasks, and system administration. gle server or dual hot-standby configuration, as required. The ■ Technician workstation. This is configured for technical latter enables a fast and automatic changeover in the case (system) monitoring, such as assessing whether sensors, of failure. A full AWOS server system includes a 19-inch rack computers, data communication equipment, and so on are with front (glass) and back door, shelves, integrated key- working correctly. board/video/monitor (KVM) with screen, keyboard, mouse, ■ Display workstation. This is configured with display-only network switches, and terminal servers. If the users already functionality. have rack space available or there is already network equip- ■ Tower data presentation systems. These are display-only ment, it is recommended to reuse whatever is available if systems configured for ATC users. these parts are still in a good condition for reliable operation. 3.4.10 AWOS System Interfaces 3.4.9.2 Backup Display Units AWOS systems usually have several interfaces with other sys- These are units that can be installed in consoles (tower), tems to allow the exchange of data. The servers, workstations, racks, and so on; they are intended as a backup. The backup and displays all connect via a LAN infrastructure, which is to display system receives its input from the sensors directly, be provided by the user, unless it is included in the project not from the server. Even when both servers fail and thus the deliverables. Standard Ethernet TCP/IP socket communica- client workstations do not receive data from the servers, the tions is used for all networking functionalities. Ethernet TCP/ backup display will still show sensor data. The backup display IP and associated LAN hubs that are part of the scope of sup- usually is a small computer with a (touch screen) liquid-crys- ply are proven, robust, fast, and reliable. Most AWOS systems tal display (LCD) screen. It has a processor, a local area net- include a standard AFTN/AMHS interface that supports WMO work (LAN) connector, a serial input, USB connection, and so messages. All WMO messages are automatically forwarded to forth. The sensors that are displayed can be chosen (all wind an AFTN/AMHS bulletin. sensor, pressure, cloud base and amount, temperature, rela- tive humidity, and so on). The unit also performs some data Data output to remote or external locations can be either processing, so that both raw sensor readings can be displayed via serial ports or via LAN to other local computers. Such along with other values that require limited processing, such non-standard interfaces must be specified by the users, and as 2-minute and 10-minute averages. the supplier can develop and configure these interfaces as required. The AWOS server interfaces with several other sys- 3.4.9.3 Workstations tems, some of which are listed in the sections above. Not all These can be provided for several users and purposes. A typ- systems will be present in every AWOS system, but the server ical workstation will be a personal computer (PC) running a should be configured to include what the user requires. Windows or Linux operating system (OS) and will be connect- ed to the server locally through a LAN or remotely through a 3.4.11 Estimating the Total Cost of Ownership leased line, serial line, or modem. In most cases, workstations for AWOS will be protected with a username and password and have a monitor, keyboard, pointer (mouse, touchpad, or trackball), To aid in understanding the total cost of ownership (TCO) and printer (if appropriate). Workstations can include: over a 10-year period, chapter 3.8 provides a guide in calcu- lating the TCO. Although this costing exercise will require an ■ User workstations. These provide user access to meteoro- effort on the part of the owner, working through the calcula- logical data and reports. tion will provide a basic understanding of the system being ■ Observer workstation. This is configured for preparing purchased, its TCO, and the financial and human resources METAR, SPECIAL, TREND, SYNOP, and other reports. required to ensure sustainable operations over the 10-year life cycle of the system. It is important to note that both the Automated Weather Observing Systems    211 client and the vendor have individual responsibilities for the shipping is commonly included in this category, which can successful outcome of the AWOS project, and that they will be considerable because of the bulky and heavy 10-meter need to collaborate on some tasks. wind masts. The cost of factory services is a significant portion of the total capital costs, and the value of these Based on the information provided in previous sections, it services should not be underestimated. When executed is clear that there are many complexities based on the local properly and according to strict procedures and project needs of the aerodrome that must be resolved to successfully management, a new AWOS will have a much greater likeli- implement an AWOS. The process of building out a reason- hood of functioning properly after installation than if these able estimate of project costs rests on four elements: (1) the services are rushed or incomplete. capital cost for system, spares, and options; (2) client project ■ Field services. These services include the final steps of the costs; (3) annual maintenance costs; and (4) life-cycle man- installation and commissioning process and are performed agement costs. The result is the TCO. at the airport. AWOS installation times are typically 1.5 to 3 weeks, but they can take longer depending on the amount 3.4.11.1   AWOS System Costing Introduction of equipment to be installed and whether the site prepara- Even though CAT I, CAT II, and CAT III runways differ in terms tion is complete when the installers arrive on site. of the number and types of observations made, they share much of the same equipment, software, and reports, and they The total capital costs are shown in table 3.4.10, broken down follow the same structure, as outlined below: into a single runway for each category type. Actual costing can range from +/− 20 percent or more, depending on runway ■ Field station equipment. This includes items such as configuration, options selected, and local requirements. The sensors, data loggers, communication equipment, and total costs range from $185,000 for a CAT I to $245,000 for 10-meter frangible wind masts. Typically, concrete foun- a CAT II to $285,000 for a CAT III. dations for wind masts and standards and structures are excluded, as these usually are provided by the aerodrome TABLE 3.4.10  Single Runway Costing for Each Category Type itself and built to specifications provided by the AWOS Item CAT I CAT II CAT III supplier. It can be useful to break down costs related to System $170,000 $230,000 $270,000 each field station for the TDZ, MID position, STOP-END, Spares $15,000 $15,000 $15,000 and met garden. Total $185,000 $245,000 $285,000 ■ Computer equipment. This centers on a redundant serv- er installed in a 19-inch rack in a hot-standby configura- Source: Estimates are from various AWOS suppliers, confidential information tion. Some network equipment is included, along with a per supplier, but allowed to be used as general overview numbers. UPS, KVM switch, and CPUs, monitors, keyboards, point- ers, and other necessary components for the servers and 3.4.11.2  Recommended Spares workstations. Recommended spares should be acquired with the initial ■ Software licenses. AWOS systems are generally available purchase of the AWOS system and are essential for the sus- either on the basis of a one-time licensing fee or as an tainable operation of the system and to support the calibra- annually recurring licensing fee. Costs can differ consid- tion strategy. In table 3.4.10, for example, these come to erably per supplier. Besides the AWOS software, licenses $15,000, regardless of the type of runway. They should be might be needed for other software packages (such as MS purchased in quantities suggested by the manufacturer or Windows or MS SQL). supplier. Commonly, obtaining one spare of each item is suf- ■ Factory services. These services include project man- ficient. An additional set of spares should be acquired at the agement, complete system integration of equipment and 7- to 8-year mark, as this is when the original sensors are software, testing, and factory training for the owners. It is nearing end-of-life and might require replacement. Typically, common for owners to show up for several days of train- a spare is swapped for a failed or worn sensor, or one nearing ing ahead of the Factory Acceptance Test. Packing and 212    Automated Weather Observing Systems calibration expiry. Sensors with calibrations that have ex- 3.4.11.4  Owner Project Costs pired or are nearing expiry can be returned to the manufac- The owner of the AWOS is typically either the NMS or the air- turer or to a suitable facility for recalibration. A complete set port authority for the aerodrome. In either case, the owner of spares should be maintained at all times to ensure sustain- will incur the following additional costs to complete the able and uninterrupted operations. project: In addition, there will be a need for present weather sensors, Cost for civil works and site preparation. New AWOS systems ceilometers, and servers (these are not included in table are typically procured to upgrade an aerodrome from manu- 3.4.10). al weather observations to fully automated AWOS operation, or to replace an aging system. This means that field stations ■ Present weather sensors. These are calibrated in the field with cables for AC mains power and data (or other types of with a scatter plate, which is purchased together with the data communication) are already in place. This infrastructure instrument. When maintained properly, present weather should be reused as much as possible, though existing copper sensors can remain operational for up to 20 years. cables should be checked for condition before use with the ■ Ceilometers. These require some maintenance, which can new AWOS and replaced as necessary. Costs for civil works be done in the field. Some components (such as the laser) activities that require digging (such as the installation of be- may need replacement after several years of operational low-ground cabling) can range from just a few thousand dol- use, in which case the ceilometer can be returned to the lars to several tens of thousands of dollars. The total distances manufacturer or a local representative for repair. Since the between the field station (or stations when more than one is ceilometer is the most expensive instrument in an AWOS, needed) and equipment rooms can be several kilometers. it is the owner’s choice whether to purchase a spare or op- erate for a short period without one. Radio communications can provide a good option, provided ■ Servers. The AWOS system is configured with dual re- there are free lines of sight between the field stations and dundant systems. Typically, standard off-the-shelf servers the location where the receiving radio antenna is installed— should be used so that if one needs to be replaced, it can including during airport operations, as airplanes’ tails can be purchased locally and integrated into the system. interfere with the radio signals if they pass through the line of sight. When considering the use of radios, national and in- 3.4.11.3  Optional Items ternational regulations regarding frequency bands should be There are two categories of optional parts listed for all three considered, and permits should be obtained as necessary. categories of runway: In addition to communications and mains power infrastruc- ■ Lightning detector. This is a sensor that can be added to ture, the owner is also responsible for concrete foundations one of the field stations in the aerodrome. Most AWOS for towers and sensor bases and lightning rod earth connec- software can display and integrate data from a lightning tion at field stations. And for AWOS equipment that will be detector into the present weather codes for METAR, SPECI, installed in the equipment room and at user desks, the owner and other reports. A lightning detector is recommended for will need to budget staff and third-party contractor costs for aerodromes with frequent lightning events for airport staff ensuring (1) AC mains power connections, (2) 19-inch rack safety reasons. space in the equipment room, (3) data communication con- ■ Runway light monitoring field station. This measures the nections with field stations and server, and (4) a UPS. flow of current through the runway light’s power cable and is needed if there is no other way to assess runway lighting While “Civil Works” appears as a line item in the costing exer- intensity. The preferred method of obtaining runway light cise, the owner will need to work with the supplier and a local settings is via a serial or network connection to the runway contractor to obtain accurate pricing. light system. Automated Weather Observing Systems    213 Cost for installation efforts. The AWOS system supplier will final training. For larger airports and when large numbers of install the system on site. The owner should assign its own technicians and users are to be trained, multiple training ses- staff to work with the supplier, which should be budgeted for: sions (taking up to five days) may be required. ■ Two days per field station, which are required for installa- 3.4.11.5  Operational Costs tion (depending on the amount of equipment). Installation Costing of the operational budget is an important component of the field stations generally requires support from one or for understanding the financial commitment of keeping the two local technicians. system operational for a period of (at least) 10 years. Each of ■ Two to four days, which are required for installing the the following activities has associated expenses, which con- 19-inch rack with servers and other equipment including tribute either annually or randomly within the 10-year opera- workstations. tional period of the system. These costs include (1) preventive maintenance and calibrations, (2) corrective maintenance, Cost for project management. The owner should appoint a (3) adaptive maintenance, and (4) life-cycle management. project manager to coordinate the installation, testing, and commissioning activities with the supplier, which will involve Preventive maintenance and calibrations. This consists of requirement specifications, a design review and approval regular checks of the equipment, routine basic cleaning ac- process, coordination of local civil works, and preparation of tivities, and checking the conditions at the field stations and the installation location. For a typical AWOS project for one equipment room. Performing preventive maintenance will runway system, it is expected that the owner’s project manag- extend the operational life of the equipment, as it will detect er will spend 20–30 days of work on the project. and prevent problems from becoming more severe. Preventive maintenance routines are equipment- and site-specific and Cost for factory training and Factory Acceptance Test (FAT). should be carried out in accordance with manufacturer’s When factory services for setting up the AWOS system are recommendations. For calibrations, sufficient spares should ready, the owner usually sends a group of people to the sup- be available to swap freshly calibrated sensors for sensors plier’s factory for factory training (for a duration of five days), with expiring calibrations. Sensors with out-of-date calibra- travel logistics, and the FAT. This group typically includes: tions should be sent to the manufacturer or a local facility for calibration. ■ Two technicians to participate in the FAT and receive tech- nical training Corrective maintenance. This includes all unscheduled re- ■ Two users, typically observers, to participate in the FAT pairs to equipment. While it is not possible to know with and receive user training certainty when a piece of equipment will fail, the possibility ■ A project manager to participate and approve the FAT and of failure should be accounted for in the budgeting process. participate in the training. Corrective maintenance can be estimated as a fraction, often 15 percent, of the preventive maintenance budgeted amount. Cost for on-site training and Site Acceptance Test (SAT). The SAT is a shorter version of the FAT and focuses on testing the Adaptive maintenance. This maintenance is required to im- AWOS in its operational environment. It will confirm that data plement changes in the AWOS system. Suppliers usually work are communicated properly between the sensors, AWOS serv- with an Engineering Change Proposal (ECP)—a document ers, and workstations—and that observers and the ATC have that allows a user to request a change to the AWOS system. access to the required data and reports. A typical SAT can be The supplier will analyze the request and prepare a quota- completed in one to two days for a single runway aerodrome; tion for executing the work, to be agreed on by both the user larger and multi-runway aerodromes typically take longer. and the supplier. Examples include implementing changes One or two NMS representatives of the owner must witness introduced by ICAO Annex 3, or if the AWOS operator wishes the test and sign off when it is confirmed that the AWOS is to add a new instrument to the system (such as a lightning working correctly. Typically, two days are required for this detection sensor). Adaptive maintenance is not a regular 214    Automated Weather Observing Systems activity but commonly involves a significant amount of work. system and budgeted for annually or in the required year of Often, changes in the AWOS software require additional de- the expenditure. sign, software, testing, and an additional SAT when installed at the aerodrome. These costs through ECPs can typically 3.4.12 Recommendations range from 10 percent to 25 percent of the AWOS system’s initial purchase price, and one to four ECPs might be required In sum, when developing a strategy for implementing an during the AWOS’s operational life. The owner should allocate AWOS network, an NMS should consider the following 20 percent of the AWOS’s original purchasing value over the recommendations: lifetime of the system for adaptive maintenance. 1. Appoint a project manager. Project management is Life-cycle management. During the operational life of the critical to the successful implementation of an AWOS AWOS system, several parts will have to be replaced as they system. The project manager is the key focal point of age and become unreliable. Fortunately, the most-expensive communications between the supplier and the aero- sensors in the AWOS system (ceilometers and forward scatter drome and must have decision-making authority over sensors) are also typically the most reliable and are known the implementation phase. to be able to continue operating for periods of 20 years. For 2. Investigate the costs of implementing service-level some other equipment (such as the Field Data Collection Unit agreements. When budgeting the TCO of an AWOS net- and communications devices), 20 years may also be achiev- work, it is critical to consider the value proposition pro- able if maintained properly. AWOS servers should be preven- vided by service-level agreements, given the improved tively replaced after 6 to 7 years of operation. As for optical functionality and features provided in annual software instruments, which have an anticipated life span exceeding upgrades (including the incorporation of changes in the 10-year cost of operation, the owner may wish to replace ICAO procedures that will need to be incorporated to one or both sensors as part of the 10-year life cycle, depend- maintain an international aerodrome status). ing on reliability and condition. 3.4.11.6  Upgrade Strategy AWOS systems are built according to ICAO and WMO regula- tions, which do not often change. ICAO Annex 3 is updated about every two years. These updates might include changes to METAR or other reports, which may require modification of the AWOS software. In general, ICAO allows two years to implement and activate the changes. An operational AWOS system is typically designed and built with additional requirement specifications from the owner. If the owner wishes to implement changes in the system, this can be done with adaptive maintenance. Most AWOS suppli- ers can offer service-level agreements, which describe the services the supplier can offer to support the AWOS system following certain well-defined processes and procedures. A service-level agreement may be based on an annual fixed fee for the ongoing support effort, with additional fees for ECPs. Costs for service-level agreements and ECPs should be inves- tigated with the supplier at the time of acquiring an AWOS Automated Weather Observing Systems    215 3.4.13 References WMO (World Meteorological Organization). 2014. Guide to Meteorological Observing and Information Distribution Hodgson, Steven. 2022 in process. Paper on meteorological Systems for Aviation Weather Services. WMO-No. 731, 2014 law and the global weather enterprise in process. edition. Geneva: WMO. https://library.wmo.int/doc_num. php?explnum_id=8627. ICAO (International Civil Aviation Organization). 2017. Manual of Aeronautical Meteorological Practice, Document WMO (World Meteorological Organization). 2018a. Guide to 8896 AN/893, 11th edition. Montreal: ICAO. http://icscc. Instruments and Methods of Observation, Volume I: Measure- org.cn/upload/file/20190102/Doc.8896-EN%20Manual%20 ment of Meteorological Variables. WMO-No. 8. Geneva: WMO. of%20Aeronautical%20Meteorological%20Practice.pdf. https://library.wmo.int/doc_num.php?explnum_id=10616. ICAO (International Civil Aviation Organization). 2018a. WMO (World Meteorological Organization). 2018b. Guide to Annex 3 to the Convention on International Civil Aviation, Instruments and Methods of Observation, Volume III, Observing Meteorological Service for International Air Navigation, 20th Systems. WMO-No. 8. Geneva: WMO. https://library.wmo.int/ edition, July 2018. Montreal: ICAO. file:///C:/Users/Owner/ doc_num.php?explnum_id=9872. Downloads/icao_annex_3_meteorologicalserviceforinterna- tionalairnavigation.pdf. WMO (World Meteorological Organization). 2018c. Technical Regulations: Basic Documents No. 2, Volume II: Meteorological ICAO (International Civil Aviation Organization). 2018b. Service for International Air Navigation. WMO-No. 49, 2018 Annex 6 to the Convention on International Civil Aviation, edition, updated 2021. Geneva: WMO. https://library.wmo. Operation of Aircraft, Part I – International Commercial Air int/doc_num.php?explnum_id=10733. Transport – Aeroplanes, 11th edition. Montreal: ICAO https:// ffac.ch/wp-content/uploads/2020/09/ICAO-Annex-6- Operation-of-Aircraft-Part-I-International-commercial-air- transport.pdf. ICAO (International Civil Aviation Organization). 2018c. Annex 14 to the Convention of International Civil Aviation, Aerodromes, Volume I, Aerodrome Design and Operations, 8th edition, July 2018. https:// www.worldcat.org/title/manual-of-barometry-wban/ oclc/11814353&referer=brief_results. 216    3.5 Automatic Weather Stations 3.5.1 The Purpose of Automatic Weather Stations Data from weather stations are used for a wide range of applications, includ- ing real-time weather observations; forecasting general weather and extreme weather events; providing information for sectors such as agriculture, hydrol- ogy, and climatology; and ensuring road, marine, and aviation transporta- tion safety. In low-income countries, weather observations might be made through a combination of manual weather stations (MWS) and automatic weather stations (AWS), as determined by cost, capacity, capability, and gov- ernment policy. Whether a given station is manually operated or automated can also depend on its application. For example, a country might use an au- tomated weather observing system (AWOS) to support international airport obligations and World Meteorological Organization (WMO) data exchange re- quirements but employ human observers to gather weather observations for agricultural operations or long-term records aimed at understanding climate Photo: © Gerald Corsi | istock.com variability and change. Whether meteorological data are obtained using observers or automated sta- tions, many of the same principles apply with respect to system design, spa- tial distribution, and measurement of meteorological parameters. Similarly, even though the use of an AWS eliminates the need for a weather observer, regular maintenance of the AWS is essential for ensuring the uninterrupted flow of quality data. AWS provide reliable, accurate, and continuous observations of surface weather conditions without the need for human observers. To accomplish this, an AWS employs sensors that convert meteorological conditions to elec- tronic signals, which are then measured and recorded by a data logger. In most cases, the AWS then transmits the recorded data to a central informa- tion processing system (IPS) via telephone, cellular, satellite, internet, or radio communications. Data received by the IPS from the various AWS instal- “AWS provide reliable, lations in the network are recorded, checked for quality and completeness, analyzed, compiled into forecasts and other products that serve the National accurate, and continuous Meteorological Service (NMS) mandate of public safety, and transmitted to end users. The end goal is always to design a system where the data recorded observations of surface are fit for purpose in support of its intended use. weather conditions without the need for human observers.” Automatic Weather Stations   217 FIGURE 3.5.1  Typical Configuration of an AWS Installation operational costs) over the approximate 10-year life cycle of the network. Lightning rod Wind speed and It is important to consider each of the above principles as in- direction sensor terdependent components of a complete weather monitoring system. For example, even the highest-specification instru- ment can fail to deliver useful measurements if it is improp- erly installed; similarly, a properly specified and installed AWS will fail to provide data to the IPS if the power supply or communications fail. Perhaps most critically, no AWS net- work will deliver quality measurements over its expected life Solar panel cycle if it is not carefully and regularly maintained. This chapter will guide the NMS through some of the key points to consider when building a sustainable weather ob- Air temperature and servation network, with a primary focus on AWS rather than relative humidity sensor MWS. If the goal of sustainability is to be achieved, there are with radiation shield Solar radiation sensor two key commitments that the NMS and sponsoring agencies Enclosure houses barometric pressure sensor, datalogger, must make: power supply, and modem Raingage ■ The commitment to solve a specific weather-related prob- lem (such as improving forecast accuracy or developing Grounding rod an extreme weather event warning system) through a Source: Campbell Scientific Inc. dedicated observation network. Whether the network is composed of AWS, MWS, or a combination of the two, the goal of reliably gathering high-quality data fit for purpose The general principles that should be applied when designing must be viewed as critical. It is only through this commit- an AWS network that supports the data requirement over a ment that long-term operational requirements (such as in- 10-year life cycle include: strument recalibration and maintenance and proper data management) will be fulfilled. ■ Choosing which environmental parameters to monitor ■ The commitment to maintain the long-term financial vi- ■ Selecting instruments to measure these parameters to re- ability and sustainability of the network. Typically, the quirement specifications cost of acquiring equipment for an AWS network rep- ■ Configuring the AWS, including installation site, mounting resents just a fraction of the TCO over a 10-year period. structure, sensor mounting, and power supply (see figure The remaining costs—which are associated with station 3.5.1) maintenance and life-cycle management communications, ■ Determining where to place individual stations in the net- data management, and personnel—must be underpinned work across the landscape by a stable financial budget. ■ Ensuring that data are reliably transmitted from each AWS to the IPS Because of the large number of decisions that must be made ■ Planning for proper maintenance and upgrades for the when designing, specifying, installing, operating, and main- AWS network and the IPS taining a weather monitoring network, it is strongly recom- ■ Estimating the total cost of ownership (TCO) of the AWS mended that if the NMS does not have the required in-house network (including capital and installation costs plus experience to make the necessary decisions, it should consult with knowledge experts from an experienced NMS currently 218    Automatic Weather Stations operating a sustainable AWS network, or with a consultant or for general weather forecasting can be found in WMO-No. industry expert. 8 (WMO 2018) and in the Manual on the WMO Integrated Global Observing System (WIGOS) (WMO 2021); for other 3.5.2 Environmental Parameters Measured by applications, requirement specifications might differ. WIGOS AWS specifies three different levels of performance: Different monitoring applications require the observation of ■ Threshold is the minimum requirement to be met to en- different environmental parameters in order to answer the sure that a given observation is useful. specific questions being asked of nature. For example, an- ■ Breakthrough is an intermediate performance level be- swering the question “What will the weather be like here on tween threshold and goal and represents a significant im- Tuesday?” or “How severe will the storm be?” requires a dif- provement over threshold specifications. ferent set of observations than would be needed for “On what ■ Goal is an ideal requirement above which further improve- day should I plant my crops?” or “What species of crop should ments are unnecessary. I plant?” As a result, different applications also have differing sets of requirement specifications for meteorological observa- It is important to stress that these levels of uncertainty per- tions, which are designed to ensure that the measurements tain to the entire measurement system, including the sensor are of sufficient quality to answer the particular questions itself, how and where it is mounted, and how it is measured being asked—that is, that the measurements are fit for pur- or sampled (for example, by a data logger). For each mea- pose. These specifications generally include requirements for sured parameter, requirements typically include the follow- (1) which parameters are measured; (2) the accuracy of the ing specifications: measurements (typically expressed in terms of uncertainty, a mathematical expression of accuracy; see WMO 2018, Chapter ■ Range. This is the total range of the parameter being mea- 1); (3) the range of conditions over which the specified uncer- sured over which the sensor is required to report values tainties must be met; and (4) how and where sensors are to within the required uncertainty. be installed as well as their geographical spatial distribution ■ Required uncertainty. This is the maximum permitted un- (which determines the density of measurements). certainty (accuracy) over the total measurement range of the measured parameter. This section examines the set of environmental parameters ■ Time constant. This is the time it takes after a step change most commonly measured by AWSs for general weather in the value of the parameter for the instrument to report forecasting: air temperature, relative humidity, barometric a change equal to 63 percent of the change. For wind sen- pressure, surface winds, precipitation, and global horizontal sors, this is typically expressed as a distance constant, irradiance. For each of these, it highlights the importance of which is the distance air must travel following a step the parameter for understanding the environment, the types change in wind speed or direction for the sensors to report of sensors used with the AWS, and the instruments used to a change equal to 63 percent of the change. measure the parameters in MWS. Determination of the spe- ■ Sampling frequency. This is the suggested rate at which cific set of environmental parameters to be monitored for a the data logger should sample values from the sensor; al- particular application should be made in consultation with ternatively, this frequency can be set to 25 percent of the knowledge experts. (For information on measuring environ- sensor’s time constant. mental parameters not discussed here, see WMO 2018 and ■ Operating range. This is the range of environmental con- references contained therein.) ditions (typically, temperature and relative humidity) over which the sensor is required to report values within the Once the set of environmental parameters to be measured is required uncertainty. determined, it is important to select appropriate requirement specifications for the particular application for which the Requirement specifications for a particular application should AWS network is needed. Requirement specifications suitable be determined in consultation with a qualified consultant or Automatic Weather Stations   219 an industry expert. The following information regarding mea- FIGURE 3.5.2  Radiation Shields surements of air temperature, relative humidity, atmospheric pressure, surface winds, precipitation, and global horizontal a. Passive b. Aspirated irradiance is summarized from WMO 2018. 3.5.2.1 Air Temperature Air temperature is perhaps the most fundamental of all mete- orological observations. The temperature of a body of air im- pacts its density, which gives rise to pressure variations, and, for a given concentration of water vapor, the temperature de- termines the relative humidity. Air temperature determines whether precipitation is liquid or solid and also determines the rate of evaporation, and it strongly influences the growth rate of crops and other plants. Air temperature measurements at an AWS are typically made Source: RM Young Company. using platinum resistance thermometers or thermistors. Note: Panel a shows a passive radiation shield; panel b shows a forced ventilation radiation shield. Because direct exposure to solar radiation can cause signif- icant measurement errors, the sensor should be mounted in a well-ventilated screen (see figure 3.5.2); this screen also re- 3.5.2.2 Relative Humidity duces measurement errors caused by precipitation. Airflow Relative humidity is a measure of the amount of moisture through radiation screens can be passive or fan forced. While present in the air, expressed as a percentage of the amount forced ventilation can help minimize uncertainties when needed for saturation under the same temperature and air wind speeds are low (< 1 m s-1), it can also cause measurement pressure conditions. The relative humidity is closely related errors if fog or rain droplets are drawn onto the sensor by to (and can be used to calculate) the dew point temperature, the fan. Note that the type of ventilation (passive or forced) which is defined as the temperature at which air with a given can influence whether a given temperature/relative humidity water vapor pressure would reach saturation if cooled. (See sensor will meet requirement specifications. WMO 2018, Chapter 4, for more information on the measure- ment of the humidity of air.) Air temperature measurements at an MWS can be made using an electronic thermometer, typically with a digital readout, Relative humidity measurements at an AWS are most com- or with a liquid-in-glass thermometer; in the latter case, mer- monly made using electrical capacitance hydrometers (WMO cury-in-glass thermometers are no longer acceptable given 2018, Chapter 4), which use the variation in the dielectric the Minamata Convention and, if currently in use, should be properties of a hygroscopic material to measure atmospheric replaced by modern alternatives and disposed of properly. As water vapor. Electrical capacitance sensors typically include with electronic thermometers, liquid-in-glass thermometers an electrical interface built into the probe, which simplifies should be shaded from direct solar heating by a suitable ra- data logger connection and programming. Because relative diation shield. When reading a liquid-in-glass thermometer, humidity is a function of both water vapor pressure and tem- care should be taken to avoid parallax error by ensuring that perature, relative humidity sensors are commonly built into the line of sight is at a right angle from the thermometer stem probes along with air temperature sensors. This places the rel- and at the same level as the liquid’s meniscus. ative humidity sensor inside a radiation shield, which protects the hydrometer from damage due to direct contact with water. Relative humidity measurements at an MWS can be made using an electronic relative humidity sensor or psychrometer. 220    Automatic Weather Stations If a psychrometer is used, it is important to follow manufac- much weaker than those in the horizontal direction, only the turer instructions for maintaining the water reservoir, wick, horizontal wind speed and direction are generally measured. and wet-bulb sleeve. Measurement of the dry and wet ther- Wind speed is typically measured in units of meters per sec- mometers should be made as close to simultaneously as ond (m/s) or kilometers per hour (km/h), whereas wind di- possible. rection is defined as the direction from which the wind blows and is generally reported in polar coordinates with numbers 3.5.2.3 Atmospheric Pressure increasing clockwise from geographical north. Both quanti- The atmospheric pressure—also called barometric pressure— ties are typically measured at a standard height of 10 meters at a given location is a measure of the weight of the atmo- above the ground surface, and it is usual to report time-aver- sphere above that point. Atmospheric pressures change with aged values for speed and a vector for direction. variations in air density, which are in turn driven by an un- even heating of the Earth’s surface due to changes in solar Rapid fluctuations in winds are referred to as gustiness, with heating, surface type and cover, and weather. Atmospheric single fluctuations called gusts. For applications that require pressures generally decrease with increasing altitude (as information on the variability or gustiness of the wind, three the farther up one goes, the less atmospheric mass remains quantities are reported: peak gust, which refers to the max- above one’s location), and so atmospheric pressures are typ- imum observed wind speed over a specified time interval; ically converted to mean sea level (MSL) pressures using a standard deviation of wind speed; and standard deviation of standard model of the atmosphere’s pressure variations with wind direction. height (see WMO 2018, Chapter 3, for more information on atmospheric pressure and its measurement). Wind speeds are typically measured using a mechanical cup anemometer or propeller anemometer (see figure 3.5.3), Most electronic barometers sense changes in atmospher- while wind direction measurements are made using a me- ic pressure by measuring the displacement of a diaphragm chanical vane, either as a separate device (typically in combi- using strain gauges or potentiometric, piezometric, or ca- nation with a cup anemometer) or built into a common body pacitive displacement sensors. Other electronic barometers with a propeller anemometer. Ultrasonic instruments, which sense deflection of the walls of a thin-walled cylinder with a carry the significant advantage of having no moving parts, pickup coil mounted inside the measurement cylinder. In ei- are becoming increasingly common, but they have their own ther case, atmospheric pressure changes are measured with disadvantages. For example, precipitation can introduce sig- respect to the reference pressure (typically a vacuum) in an nificant noise to the signal, and their geometry means that adjacent sealed chamber. Pressure variations are converted one measurement probe can cast a wind shadow onto anoth- to a change in some analog electrical quantity (such as volt- er probe when the wind is blowing from certain directions, age or frequency), or to a digital signal that can be read di- increasing measurement error. In addition, both cup and rectly by the data logger. vane instruments and ultrasonic instruments are prone to problems caused by birds, and the movement inherent to me- Barometric pressure at an MWS is measured using an elec- chanical wind sensors can draw the attention of inquisitive tronic barometer or aneroid barometer, or barograph. If an animals, such as bears. aneroid barometer is used, it should be mounted in a location that has a relatively uniform temperature throughout the day, Surface winds at an MWS are typically measured by electron- and in the same orientation (vertical or horizontal) as during ic cup-and-vane wind gauges or simpler alternatives, while calibration. wind speed can be measured using a hand-held anemometer, and wind direction can be determined from a vane or banner 3.5.2.4 Surface Wind mounted on a pole that has indicators for the principal com- pass directions. Alternatively, wind speed and direction can The speed and direction of winds vary in both time and be estimated from a calibrated windsock of the type frequent- space. Because winds in the vertical direction are typically ly used in undeveloped airports. Automatic Weather Stations   221 FIGURE 3.5.3  Types of Wind Sensors a. Cup Anemometer and Vane b. Propeller Anemometer and Vane c. Ultrasonic Wind Sensor Source: Met One, RM Young Company, Gill Instruments. 3.5.2.5 Precipitation FIGURE 3.5.4  Precipitation Gauges Precipitation is defined as the liquid or solid products of the a. Tipping Bucket b. Weighing Gauge condensation of water vapor, whether falling from clouds or deposited from air onto the ground, and includes rain, hail, snow, dew, rime, hoar frost, and fog precipitation. Precipitation amounts are expressed in terms of the vertical depth of water to which it would cover a horizontal surface, usually expressed in units of mm. Precipitation amounts < 0.1 mm (< 0.2 mm for the United States) are generally referred to as trace (see WMO 2018, Chapter 6 for more information on precipitation). Precipitation at an AWS is often measured using electronic precipitation gauges; if only liquid precipitation can be accu- rately measured (as is the case with non-heated gauges), they are called rain gauges. These gauges measure the volume (as is the case for most rain gauges, which use a funnel to guide Source: Environmental Measurements Limited (EML) and OTT HydroMet. water to a tipping bucket (see figure 3.5.4a) for rain gauges) or weight (as is the case for most solid or combined liquid and Precipitation at an MWS is commonly measured using a grad- solid precipitation gauges—see figure 3.5.4b). Precipitation uated rain gauge, with rainfall amounts being measured at gauges are particularly sensitive to undercatch due to wind, regular intervals. As with liquid-in-glass thermometers, care and thus the best locations for mounting these sensors are should be taken to avoid parallax error when taking measure- commonly poorly suited for mounting wind sensors. The use ments by ensuring that the line of sight is at a right angle of a wind shield can often improve the accuracy of precipi- from the rain gauge stem and at the same level as the water’s tation measurements. Because high-intensity rainfall events meniscus. can also cause inaccurate measurements, it is recommended that expert advice be sought before selecting a precipitation gauge for use in areas where heavy rains are expected. 222    Automatic Weather Stations 3.5.2.6 Global Horizontal Irradiance typically used. If a Campbell-Stokes recorder is used, the card Electromagnetic radiation emitted by the Sun provides most needs to be replaced daily, with the card inserted into the of the energy that warms the Earth’s surface and atmosphere, appropriate set of grooves for the time of year. evaporates water, drives weather, and fuels photosynthe- sis of plants. The Sun emits radiation largely in the spectral FIGURE 3.5.6  Campbell-Stokes Sunshine Recorder range of 290–3,000 nanometers (nm), commonly referred to as solar or short-wave radiation. Portions of this energy are absorbed by the Earth’s surface and the gases, aerosols, and clouds of the atmosphere, which then re-radiate the energy in longer wavelengths given their lower temperatures. This re-radiated energy is almost entirely longer than 3,000 nm and is commonly referred to as terrestrial or long-wave radi- ation. Additional portions of radiation are reflected from the Earth’s surface and atmosphere without being absorbed or re-radiated. Measurements of radiation include direct solar radiation, global and diffuse sky radiation, and sunlight duration. For some applications, it is desirable to make separate measure- ments of short-wave and long-wave components of radiation Source: Photo courtesy of Fairmont Instruments. or incoming and outgoing radiation. Other applications bene- fit from measurements of net (incoming minus outgoing) total 3.5.2.7 Data Loggers (short-wave plus long-wave) radiation. Of these measure- Typically, meteorological sensors in an AWS are controlled by ments, the most commonly made by an AWS is global hori- an industrial data logger, which can be programmed to per- zontal irradiance (typically referred to as incoming shortwave form numerous additional tasks. These tasks can include: radiation—see figure 3.5.5), which is measured with a pyra- nometer (for more information regarding radiation measure- ■ Monitoring a real-time clock to execute specific tasks ac- ments, see WMO 2018, Chapter 7). cording to a schedule ■ Sampling sensors at desired intervals FIGURE 3.5.5  Solar Radiation Sensor ■ Converting sensor outputs to meteorological data by ap- plying calibration factors ■ Calculating minimum, maximum, average, and other math- ematical values (such as standard deviations) ■ Performing quality control and data reduction ■ Storing, transmitting, and displaying data ■ Performing system checks and any additional tasks (such as changing configuration settings) requested by the com- munication platform. Source: Hukseflux Thermal Sensors. Data loggers typically use an operating system and program- ming language that is specific to the manufacturer. They can Solar radiation at an MWS is typically measured using a pyra- be easily configured to operate a wide range of instruments nometer or, alternatively, if hours of bright sunshine are re- in accordance with the particular needs of the user. They typ- quired, a Campbell-Stokes sunshine recorder (figure 3.5.6) is ically operate on very little power and over a wide range of Automatic Weather Stations   223 environmental conditions, making them ideal for remote AWS ■ Security. Is the area safe enough that theft and vandalism installations. would not be significant issues? Is the site in a location safe for NMS personnel servicing the station? 3.5.3 Determining Geographical Location of ■ Land availability. Does the government own the land Individual AWSs across the Landscape where the station is to be sited? Will the government need to purchase the land (resulting a capital cost) or lease it If an NMS does not have experience in siting a network of sta- (resulting in an annual budgetary cost)? tions, it should employ an experienced consultant or consider ■ Communications. It is vital that the site have a reliable seeking the advice of a more mature NMS on working through means of communication; otherwise, the site is not ideal the process. There are several factors to consider when lo- for operational meteorology. cating individual stations. Each site selected will need to p If cellular is the preferred communication method, be balanced against all of these factors. The simplest way is there reliable coverage in the area? What is the to identify possible site locations is to review a map of the monthly cost of the communication plan based on the country or area in which the desired network will be installed. volume of data to be transmitted? Costly plans in an First, the desired spacing of stations should be determined. area may suggest that alternative methods of trans- Ideally, stations would be installed in an array or grid with mitting the data should be considered. Cellular sig- uniform north-south and east-west spacing; the chosen spac- nal strength at the site should be measured, either by ing would be selected to provide meteorological data that checking the bars on the phone or by using a signal meet the requirements for their intended end use (or uses). power meter. Stronger signals lessen the impact of Using this spacing, potential locations for the individual sta- adverse weather and provide more reliable communi- tions can be identified on the map. cation between the AWS and the IPS. p Is there landline availability for telephone or internet Each of these locations should be evaluated based on instru- communications? If not, what is the cost of bringing ment siting requirements (see WMO 2018, Annex 1.D), along the service to the site? with the sites’ logistics, security, land availability, commu- p Is there a satellite communication system that can be nications, power, and environmental conditions. Based on a used to support data transmission? Satellite systems cost analysis of each of these factors, the final site location such as INSAT, METEOSAT, and GOES have one-way might differ by several kilometers from the ideal: limited data transmission and high capital costs but low or no usage fees. Other satellite systems, such as ■ Logistics. The remoteness of a site can add significantly to Iridium and Globalstar, allow two-way data transmis- the operational costs of a network. sion and have lower capital costs, but typically incur p Is the site easily accessible? Sites on the tops of moun- high usage fees. tains that require a helicopter or that are in extremely p If using geostationary satellite communications, is remote locations will increase the travel costs signifi- there a clear view of the region of the sky in which cantly for equipment repair and station maintenance. the satellite is located? Does the antenna mount allow p Is there an alternative location that is less expensive accurate pointing of the antenna? to travel to that would provide the same or similar p If using radio communications, is there a line of site data that supports the end use, reducing logistical that ensures the data can be transmitted to a site challenges and costs? The more time spent traveling where the data can then be forwarded to the IPS? If to sites increases the human resources required by not, how many repeater stations will be required to the NMS to support the network. get the data to a location where it can be forwarded p What are the additional costs for the civil works for back to a central site? supporting the station given the remoteness of the ■ Power. Similar to communications, power will have its site? own set of concerns. 224    Automatic Weather Stations p When using solar panels does the site have a clear 3.5.4 Configuring the AWS view of the Sun? p When using solar panels, site security will be an issue. This section considers the different elements that go into con- Is there local human presence on or near the site to figuring the AWS. deter theft and vandalism? p Is there mains power at the site? Can mains power 3.5.4.1 Installation Site be brought to the site, and if so at what cost? In ei- Careful selection of the specific installation site is critical ther scenario, there will be an ongoing cost for power for obtaining the highest-quality data from a weather sta- usage. tion. This is true whether observations are automated or ■ Environmental conditions. Environmental conditions at performed by a human observer. When planning for a new the site play a large role in how often the site will require installation, it is important to consider the spatial scale over maintenance and repair. Is the ideal location for the site which a given measurement needs to be representative—that subject to: is, the degree to which it accurately describes conditions p Extreme wind, dust, or debris that can affect measure- over a larger area. Different applications can have different ment accuracy, requiring more frequent maintenance preferred spatial and temporal scales of representativeness. visits for cleaning rain gauges or radiation sensors For example, observations taken for agricultural meteorology and shields? need to be representative on relatively small scales, whereas p Nearby bodies of fresh or (in particular) salt water, which can increase corrosion rates and require more observations for long-range weather forecasting should be frequent maintenance or replacement? representative of much larger areas. Each of the above factors will need to be considered and bal- How representative the meteorological observations made anced against the operational costs and the data’s represen- at a station are for the surrounding area is strongly influ- tativeness requirements. In the end, the ideal site may still be enced by the physical characteristics of a station’s setting. the best location to support the data needs. By weighing the Meteorological observations made at stations located on above factors, an NMS should begin to understand the cost steeply sloping ground or near open water, paved areas, tall implications of their decisions. buildings, or trees will be influenced by these features, and thus be less representative of conditions over larger scales. The AWS network should be constructed from high-quality, For this reason, potential installation sites should be evaluat- research-grade instruments and data loggers for the collec- ed with respect to characteristics that might influence mea- tion of synoptic (forecast) data. Although less expensive sys- sured values. If possible, stations should be located in a flat, tems are available (such as all-in-one stations), they do not open area, away from buildings, trees, open water, or other meet the accuracy needs of synoptic-grade stations, are less features, as those situated in hilly areas, coastal settings, or flexible in operation, and suffer from design compromises near obstacles are unlikely to be representative of areas larg- that sacrifice measurement quality for installation ease. For er than a few kilometers. For stations that are used to satisfy example, all-in-one stations often include both surface wind several purposes—for example, those used for both weather and precipitation sensors in the same compact system, even forecasting and agricultural meteorology—the most stringent though siting requirements for these two environmental pa- requirements should be followed when selecting installation rameters are incompatible. That said, all-in-one stations are sites and instrumentation. Once these choices have been well suited for use as secondary or infill stations, where re- made, other considerations, such as power and communi- quirement specifications are less stringent. Unlike modular cations, can be determined (for more information about site systems where a sensor can be replaced easily, all-in one sta- selection, see WMO 2018, Section 1.3 and Annex 1.D; Ehinger tions require full replacement when a single sensor fails. 1993). Automatic Weather Stations   225 A nearly ideal AWS site, in Canada, is pictured in photo 3.5.1. In the case of a typical AWS used for synoptic or climatolog- It is located in an open field, providing very good exposure ical station in a regional or national network, WMO gives the in all directions for all of the sensors, and it is grass covered following specifications for site selection and instrument lay- and well maintained. Environmental parameters measured by out (WMO-No. 8, WMO 2018, Chapter 1): the AWS include all parameters discussed in this chapter (air temperature, relative humidity, barometric pressure, wind ■ Outdoor instruments should be installed on a level piece speed and direction precipitation, and global horizontal ir- of ground, preferably no less than 25 m × 25 m where there radiance) plus snow depth and soil temperatures at depths are many sensors installed, but the area may be much of 5, 10, 20, 50, 100, and 300 centimeters (cm) below the smaller if relatively few sensors are installed. The ground ground surface. should be covered with short grass or a surface represen- tative of the locality and surrounded by open fencing to exclude unauthorized persons. PHOTO 3.5.1  View of Environment Canada AWS Station near Lacombe, Alberta, Canada (WMO ID: 71242) a. View from the North, Looking South b. View from the East, Looking West c. View from the South, Looking North d. View from the West, Looking East Source: Photos © 2021 by J. Sowiak, Government of Canada. Note: WMO ID: 71242 is a specific number attributed by WMO to this weather station identifying site location, parameters measured as well as other metadata. AWS = automatic weather station; WMO = World Meteorological Organization. 226    Automatic Weather Stations ■ If soil temperatures are to be measured within the enclo- 3.5.4.2 Mounting Structures sure, a bare patch of ground of about 2 m × 2 m should be Although MWSs are often simpler than the typical AWS, reserved for observations of the state of the ground and of structures similar to those for an AWS might be necessary if soil temperature at depths of equal to or less than 20 cm; an MWS includes electronic instruments that are read by a soil temperatures at depths greater than 20 cm can be human operator. Mounting tower/mast and sensor placement measured outside this bare patch of ground. are the same for both AWS and MWS. ■ There should be no steeply sloping ground in the vicinity, and the site should not be in a hollow. If these conditions Meteorological sensors are commonly installed on 10-meter are not met, the observations may show peculiarities of masts; in some installations, additional components (such as entirely local significance. the data logger enclosure, solar panels, and communications ■ The site should be well away from trees, buildings, walls, antenna) might also be mounted to the mast. A hinged mast or other obstructions. The distance of any such obstacle or one that can be tipped from the base will facilitate access (including fencing) from the precipitation gauge should to the sensors for maintenance and replacement. In regions not be less than twice the height of the object above the where strong winds are prevalent, guyed towers are preferred rim of the gauge, and preferably four times the height. to hinged masts. ■ Radiation, precipitation, and surface wind sensors must be exposed according to their requirements, preferably on the To install the mast, a concrete foundation of about 1 m × same site as the other instruments. 1 m × 1 m is required and should be prepared before AWS ■ Note that the fenced enclosure might not be the best place system installation. The mast manufacturer will provide de- from which to estimate the wind speed and direction; an- tailed drawings for the mast foundation and installation in- other observing point, more exposed to the wind, may be structions. Proper grounding of the site, including the use of desirable. lightning rods and transient voltage suppression—such as a ■ Very open sites that are satisfactory for most instruments spark gap or transorb, which shunt transient voltages away are unsuitable for precipitation gauges. For such sites, the from the expensive electronics of the data logger and com- precipitation catch is reduced in conditions other than munications devices to an earth ground—is critical. Note that light winds. For these locations, some form of wind shield transient voltage suppression does not protect a tower when can improve measurement accuracy. a direct lightning strike occurs, but it does lessen the dam- ■ If objects such as trees or buildings significantly obstruct age to key components, thereby reducing repair costs. The the horizon around the enclosure, alternative viewpoints AWS installation site should be enclosed with a secure fence. should be selected for observations of radiation. Chain link fencing topped with security wire works well for ■ At coastal stations, it is desirable that the station com- this purpose, given its open structure, which minimizes air mands a view of the open sea. However, the station should flow disturbance (see photo 3.5.2). not be too near the edge of a cliff, because wind eddies created by the cliff will affect the wind and precipitation The data logger is typically mounted in a waterproof enclosure measurements. along with the communications interface, battery, solar/alter- nating current (AC) charge controller, surge protector, and ba- These criteria should be considered along with those relat- rometer. This enclosure should be large enough to allow easy ed to logistics, security, land availability, communications, access to each component; vented so that solar heating of the power, and environmental conditions. Similar criteria should enclosure does not introduce errors into the barometric pres- be used for selecting the site of a manual weather station. sure readings; and carefully sealed against water, insects, and Note that several of these criteria are contradictory and will pests. In cold or high humidity environments, desiccant pack- require some compromise. ages should be used to prevent condensation of water vapor onto the electronics at low temperatures. Automatic Weather Stations   227 or NEMA-4X (weather proofness and corrosion resistance) PHOTO 3.5.2  AWS with 10-Meter Mast and Security Fence standards. To prevent water infiltration, enclosures should be installed with cable entry ports (which allow easier expansion with ad- ditional sensors) or bulkhead connectors (which offer lower maintenance and installation costs) that face downward. This ensures that all wires run below the bottom of the enclosure in a “J” loop (or dew loop), causing water to drip off the cable at the bottom of the loop. When designing the station, suit- able instrument cable jacket materials should be specified. For example, UV-resistant coatings prevent deterioration due to exposure, armored cables protect against gnawing by ro- dents, and Teflon coatings remain flexible in cold climates. Source: Photo courtesy of OTT HydroMet. Communications interfaces can be installed inside the data Note: AWS = automatic weather station. logger enclosure for ease of wiring. An antenna (if necessary) can be mounted according to manufacturer’s specifications, commonly on the tower (along with key sensors—see box 3.5.1) or inside the data logger enclosure. The enclosure can be mounted on the instrument mast or on a separate standard and should be mounted at a level that gives convenient access. It should meet, at minimum, IP-64 or NEMA-4 (weather proofness) or, alternatively, IP-66 BOX 3.5.1  Installation of Key Automatic Weather Station Sensors Air temperature and relative humidity sensors. These should be mounted to the instrument mast at a height of 1.5–2.0 meters above ground level in an appropriate radiation shield. Alternatively, the sensor can be housed inside a Stevenson screen at a separate location near the mast. Wind sensors. These should be mounted at a height of 10 meters above ground level. If possible, they should be installed at the top of the instrument mast; if not, they should be mounted on booms that are at least 2 meters long, and preferably long enough to place the sensors at a distance equal to or greater than 10 diameters from the mast. Masts with open structures are recommended, as these have less influence on wind measurements than do solid structures. Precipitation sensors. Because the accuracy of precipitation gauges is strongly affected by winds, they should be mounted near ground level where wind has less effect, but away from structures (such as the mast) to prevent water from dripping off of the structure into the gauge. The ground surface around the precipitation gauge can be covered in short grass or gravel, but hard surfaces (such as concrete) should be avoided to reduce errors due to splashing. The effect of wind on precipitation measure- ments can be further reduced by using a wind shield around the gauge. Global horizontal irradiance sensors. These should be mounted to the instrument mast on the side facing toward the Equator (that is, on the south side of the mast in the Northern Hemisphere and the north side of the mast in the Southern Hemisphere) to prevent the mast’s shadow from passing over the sensor. The sensor should be mounted on an arm that places it away from the mast, oriented so that the wire or connector is facing toward the mast to minimize solar heating of the sensor’s electronics. The sensor should be carefully leveled after mounting using the leveling screws and provided spirit level, then secured by tightening the retaining screws. 228    Automatic Weather Stations 3.5.4.3 Power Supply the charging process. This issue can be alleviated by prop- A reliable and well-designed power supply is a critical com- er ventilation, allowing gases to escape from the enclosure. ponent of any AWS installation. Automated stations are com- Otherwise, the battery (or batteries, wired in parallel to main- monly powered by one or more 12-volt batteries, which are tain a 12-volt supply) can be housed in a separate enclosure recharged by solar panels. Such supplies are safe, reliable, on the ground, placed as near to the data logger enclosure as and relatively low in cost. The total required battery capacity practical to minimize the wire run distance. In areas where (in DC ampere hours, or Ah) and solar panel wattage should rodents are problematic, protecting exposed power wires by be determined from a careful analysis of the total average using an armored shielding or running the cables in conduit current draw of the system—including all components (such can prevent chewed wires. In many cases, solar panels can be as all instruments, data loggers, and communications) to be mounted to the instrument mast, provided that care is taken powered by the supply. It is important when sizing the battery to avoid influencing sensor readings. Alternatively, panels capacity, which determines the number of days a station can can be mounted on a separate standard, potentially shared operate without charging, to consider environmental condi- with the data logger enclosure. Solar panels should generally tions at the site during periods when no or minimal charging be aimed toward the equator and angled upward (at an angle will occur (such as during monsoon season and periods of that depends on the local latitude) to maximize charging. In excessive cloud cover, fog, or snow). Depending on location, areas subject to snow, solar panels should be mounted ver- these environmental conditions can last from a few days to tically so that accumulation does not significantly reduce several months. Battery capacity must be sufficient to power charging efficiencies. the station over the longest expected period during which no charging occurs. Sizing of the solar panel(s) should also be Note that solar panels are a desired commodity in the de- calculated from the total average current draw of the system, veloping world and subject to theft. Thus, an NMS may em- and if charging is expected to be frequently interrupted be- ploy some security strategies to reduce the cost due to loss cause of environmental conditions, consideration should be through theft, including by: given to how quickly the batteries will need to be recharged when favorable conditions return. Given the importance of ■ Installing solar panels 6 meters or higher on the mast with the power supply and the complexity of the power calcula- anti-climb panels and either welding them to their mounts tions, consultation with an expert is advised. Alternatively, or attaching them with tamper-resistant bolts; or most meteorological equipment manufacturers understand ■ Placing the solar panels at a secure location away from the power requirements and can assist in designing the power AWS site, and then swapping out the batteries on a regu- supply. lar basis (maybe every one to three months based on AWS power requirements). While this strategy would require Instruments with significant power demands—such as sen- purchasing two complete sets of batteries, it would reduce sors that require heaters (such as some precipitation gauges the opportunity for theft and the associated costs of solar and radiation sensors) or aspirated radiation shields—can panel replacement. be difficult to operate from a standalone battery supply. Stations using these components will need AC (mains) power. 3.5.5 Planning for Operation, Maintenance, and AC power can also be used to maintain the state of battery Upgrades of the AWS Network charge in stations with modest power requirements; in these cases, a battery power supply acts as both power conditioner When considering sustainability of an AWS network, serious and backup power supply in the case of AC power outages. consideration and planning should be given to its day-to-day operations and its maintenance, as well as future upgrades. If a sufficiently small battery powers the AWS, the battery can be housed in the data logger enclosure. But caution should 3.5.5.1 Operational Requirements be used when charging batteries inside the data logger enclo- The most important factor in determining the long-term oper- sure, given the potential for buildup of explosive gases during ation and sustainability of a weather observation network is Automatic Weather Stations   229 routine and timely maintenance. Despite this, ongoing main- processing, and weather forecasting to be performed at a cen- tenance is frequently overlooked or underfunded, resulting in tral IPS. But if only lower-bandwidth communications options AWSs commonly failing years before they should. MWS sites are available, summary reports might need to be generated provide an interesting contrast. Because a human observer at the AWS and sent to the IPS in lieu of the observational visits them daily, they are generally well maintained. With data. The choice of communication technology can also im- every measurement, the observer checks the sensor; if there pact a station’s power consumption, capital cost, and operat- is an issue (such as debris in a rain gauge orifice or peeling ing costs, and, in some instances, location (for more on AWS paint on a Stevenson screen), it can be immediately fixed, communications, see WMO 2013, Chapter 1). thus ensuring high data quality. When feasible, communication technologies that allow two- While an AWS can provide data at a higher temporal reso- way communication between the AWS and the IPS should be lution than an MWS (and in all weather conditions), it will employed. This facilitates, for example, acknowledgment of generally not have the benefit of daily inspections. To ensure correct receipt of data transmissions, sending remote com- that any issues at an AWS are discovered and rectified in a mands and updates, and reconfiguration and/or reprogram- timely manner, each AWS should be visited on a weekly basis, ming of the AWS. Two-way communication also allows for with daily inspections if possible. But given the education or trouble shooting of the site to identify failures so that field training required of NMS personnel to operate these complex personnel understand how to repair the fault, which can save systems, it is frequently too costly (and logistically imprac- multiple trips. If only one-way communication technologies tical) to do so. If that is the case, the NMS should strive to are feasible, it is advisable to format data messages with con- visit each site two to four times per year, or once per year trol codes or checksums to allow the data center to confirm if a station is in a remote location, although this lower visi- that the message has been accurately received. In all cases, tation frequency still leaves a considerable gap between the it is prudent to consider whether communication technolo- care and attention paid a typical AWS site and an MWS site. gies might change during the 10-year life cycle of the AWS One way to significantly close this gap is by employing local network and to plan for migration to newer platforms as they people to carry out the less technical aspects of maintenance, become available. leaving the more technical work (such as recalibration of sen- sors and repairing stations that are down) to NMS personnel In the case of MWS networks, or individual manual weather during their less frequent visits. The bottom line is that there stations embedded in an AWS network, communication re- needs to be a detailed plan created and executed by the NMS quirements are generally simpler and can frequently be met on carrying out regular inspections and maintenance. by providing the operator with a cellphone for calling in ob- servations or, alternatively, a computer and email account. 3.5.5.2 Communication Requirements For a manual climate observation station, observations can Besides ensuring that each AWS in a network is properly simply be mailed in or emailed at regular intervals. configured, constructed, and located, it is critical that the data gathered by each station are reliably communicated to 3.5.5.3 Maintenance of Automatic and Manual Weather a central IPS. This allows the data to be archived; checked Stations for completeness and quality; and used for weather forecast- AWS and MWS sites should be regularly maintained to op- ing, meteorological reports, and other end uses. Because the timize operation and measurement of meteorological pa- choice of communication technology has a significant impact rameters. Annual inspections should be made of all physical on the function and operation of the AWS and network, it is structures (including fencing, mounting masts and standards, important to carefully consider regionally available commu- and enclosures). Mounting hardware and guy lines should be nications technologies and their associated costs. For exam- inspected for corrosion and tightness, and concrete founda- ple, if a high-bandwidth communication technology is used, tions should be inspected for cracks, crumbling, and corro- the AWS can be configured to send raw observational data sion of mounting platforms. Cable ties that are used to affix to the data center, allowing data quality management, data sensor wires to the mounting structures should be inspected 230    Automatic Weather Stations for deterioration and replaced as necessary. Ground cover as replacements. (For more information on maintenance and should be maintained (including mowing and watering grass, calibration of temperature and relative humidity sensors, see trimming vegetation, and picking up debris). WMO 2018, Chapters 2 and 4, respectively.) 3.5.5.4 Maintenance of Sensors Barometers. These are usually installed inside the data log- Meteorological sensors are continuously exposed to the en- ger enclosure and connected to a static pressure head. The vironment, and thus must be maintained according to man- static head must be inspected regularly to keep it free of pol- ufacturer’s specifications to ensure accurate measurements lution (such as grass, spiderwebs, bugs), and the tube from and reliable performance. Here the focus is on maintenance the static head to the sensor should be inspected for blockag- for the sensors commonly used in AWS (for more information, es and damage. A cartridge with desiccant should be inserted see WMO 2018). The manufacturers or suppliers should be between the static head and the sensor, with filters to avoid consulted at the time of purchasing for recommendations on desiccant particulate matter from entering the barometer. maintenance and calibration frequencies and procedures. NMSs that do not have in-house or in-country calibration fa- The instruments must be calibrated or controlled regularly cilities may consider purchasing this service from the suppli- against a standard barometer using approved procedures. er as part of the capital acquisition of the network for a period The interval between calibrations must be short enough to of years. Ultimately, the NMS should develop a maintenance ensure that the total absolute measurement error will meet plan that includes the number of site visits and activities by accuracy requirements. Because instruments have to be cal- field technicians (usually two to four times per year), as well ibrated in a suitable laboratory away from their operational as other general maintenance activities (usually weekly) that location, care must be taken to avoid factors that can impact are performed by local, less-technical individuals. the stability or accuracy of the barometer. These factors in- clude mechanical shocks and vibrations, displacements from Temperature and relative humidity sensors. Regular main- the vertical, and large variations in pressure (such as may be tenance includes visual inspection and cleaning of the sen- encountered during transportation by air). (For more infor- sor’s Teflon cap and radiation shield, removal of insects and mation on maintenance and calibration of barometers, see spiders, and checking cables for damage. Field verification WMO 2018, Chapter 3.) with a handheld electronic temperature/relative humidity sensor or sling psychrometer can be done, but because such Surface wind sensors. Routine maintenance for both types sensors are very sensitive to the exposure conditions (tem- of wind sensors includes regular visual inspection and re- perature, wind cooling, shading, and exposure to direct solar moval of debris, spiderwebs, and so on. Mechanical wind radiation), large differences between the handheld and the sensors use ball bearings and potentiometers, which should operational measurements can be observed. be checked regularly and replaced according to manufac- turer recommendations (or more frequently if a location is Calibration of temperature and relative humidity sensors particularly windy or dusty, or if excessive wear is noticed). should either be performed to manufacturer’s specifications Ultrasonic wind sensors do not have mechanical parts that or the sensors should be returned to the manufacturer or sup- are subject to wear and tear, but they are sensitive to dam- plier for calibration. Alternatively, some temperature/relative age and fouling by birds: the transmitter/receiver units usu- humidity probes are designed with detachable heads that can ally have soft caps, which birds can damage; birds can also be sent back to the manufacturer for calibration—and pur- perch or build nests on the sensors if there is no protection or chasing spare heads that can be replaced at regular intervals screening. Bird spikes can reduce the occurrences of this type in the field will ensure that the measurements remain within of damage. (For more information on maintenance and cali- specification. Heads that are returned from the field can be bration of surface wind sensors, see WMO 2018, Chapter 5.) sent back to the manufacturer for calibration, and newly cal- ibrated sensor heads can be returned to the pool to be used Precipitation gauges. Routine maintenance for both precip- itation gauge types should include cleaning dust and debris Automatic Weather Stations   231 from the funnel and catchment container, inspecting for loose installing a new AWS network and the ongoing costs associ- or damaged components and for frayed or otherwise damaged ated with regular maintenance. Germany operates the largest wiring, and ensuring that the gauge is level. For tipping buck- network of the developed countries considered here with 997 et gauges, it is recommended that the unit be replaced annu- AWS installations (161 are Global Basic Observing Network, ally with a newly calibrated unit; weighing gauges should be or GBON, compliant), followed by Australia (701), Canada recharged with oil (for rain-only measurement) or antifreeze (585), the United Kingdom (300), and Austria (270). All five and oil (for rain and snow measurement). countries exhibit excellent performance with uptime of 95 percent or more, indicating high reliability in the flow of qual- Calibration of a tipping bucket precipitation gauge should ity data. Because uptimes less than 90 percent prevent the be performed to manufacturer’s specifications or returned true value and benefit of an AWS network from being realized, to the manufacturer or supplier for calibration. Weighing it is recommended that developing nations set an initial per- precipitation gauges have few moving parts, and as a result, formance target no less than an uptime of 90 percent. Once they require infrequent calibration. (For more information the NMS has gained experience operating and maintaining on maintenance and calibration of precipitation gauges, see AWS stations, the target should be increased to 95 percent. WMO 2018, Chapter 6.) With regular maintenance and repair, proper management of a spare parts inventory, life-cycle management of component Radiation sensors. Pyranometers and other radiation sensors parts and well-trained maintenance personnel, it is reason- should be cleaned as frequently as practical (ideally, daily) to able to target an expected lifetime of 10 years as experienced reduce the effect of dust, bird droppings, and so on. If possi- by the developed countries. An NMS may consider extending ble, cleaning should be performed at times when the sensor is the life of an AWS to 15 years when the network is still re- not actively taking measurements. Snow, frost, and rime can liable and technologically viable, deferring full capital cost be removed using de-icing fluid. The sensors should be regu- replacement by 5 years. larly inspected for condensation inside the dome, and desic- cant should be replaced as necessary. Any forced air systems The bottom line is that annual operations and maintenance should be inspected and cleaned regularly. The manufacturer (O&M) costs per installation benchmark in the data from or supplier should perform calibration. (For more information the middle- to upper-income developing countries ($2,800) on maintenance and calibration of surface radiation sensors, is less than half that of the developed country benchmark see WMO 2018, Chapter 7.) ($6,200). This gap can be attributed to potentially less net- work complexity and possibly underspending on the part of 3.5.6 Estimating the Total Cost of Ownership of the lower-income countries. While the information is inter- the AWS Network esting, based on the small sample size of available data from these developing countries, no further analysis is offered, al- How do NMS decisions impact the TCO over a 10-year period? though more investigation is warranted. The reality is that money used to purchase a system that is too large or too complex will be wasted if the annual bud- For developed countries, although O&M costs are moderate get is too small to support the network. To aid in the cost- when each station is viewed in isolation—with O&M averag- ing process, the NMS is directed to complete the Total Cost of ing about 11 percent per year of the replacement cost with Ownership Exercise. (To calculate the TCO exercise, see chap- installation (RCI) (benchmark of $56,000)—for a large net- ter 3.8, using the additional information provided below that work, they represent a significant annual investment. Smaller is specific to the AWS network as an aid.) networks should consider a value of 15 percent to 20 per- cent for O&M (without labor) or more of the RCI benchmark 3.5.6.1 AWS Network Operating, Maintenance, and of $56,000. However, the RCI costs for an NMS upgrading an Life-Cycle Cost Examples existing network of stations will be significantly lower if the An AWS network represents a significant long-term in- current infrastructure has been well maintained and does not vestment for an NMS, given the high costs associated with require replacement. 232    Automatic Weather Stations As for labor costs, there is considerable variance among the Table 3.5.1 shows the costs related to the initial purchase, developed countries, among the developing countries, and operation, and maintenance for AWS networks in Australia, between the developed and developing countries. Major fac- Austria, Canada, Germany, and the United Kingdom—all of tors are economic circumstances, country size, and the tech- which operate successful and well-maintained networks with nical level of the AWS. For the United Kingdom, O&M plus consistently high levels of performance. Based on costing and labor costs are higher, owing to the significant cost for a num- life-cycle information from these countries, benchmark tar- ber of AWS systems that are more than a simple technological gets have been set as a guide when determining the true cost solution. of ownership of an AWS network. TABLE 3.5.1  AWS Annual Network Cost of Operations and Maintenance Data for Example Countries Lifetime Annual Annual Annual (10 years) O&M cost labor O&M plus O&M plus TCO over Expected RCI (without cost labor cost labor cost 10 years AWS sites Performance lifetime per site labor) per per site per site per site per site Country (number) (uptime, %) (years) (US$) site (US$) (US$) (US$) (US$) (US$) Developed countries Australia 701 95% 10 $56,000 $5,900 $4,200 $10,100 $101,000 $157,000 Austria 270 96% 10 $58,000 $6,200 $3,800 $10,000 $100,000 $158,000 Canada 585 96% 10 $53,000 $6,000 $7,200 $13,200 $132,000 $185,000 Germany 997 98% 10 — $5,300 $7,000 $12,300 $123,000 — United 300 96% 10 — $7,400 $7,300 $14,700 $147,000 — Kingdom Benchmark 1 96% 10 $56,000 $6,200 $6,000 $12,200 $122,000 $178,000 Eight developing upper-middle-income countries Average — 80% 10 — $2,800 $1,400 $4,200 $42,000 — Range — 50%–95% — — $900– $350– — — — $6,500 $4,700 Source: Data provided by NMS to the GFDRR of the World Bank. Note: All data are from 2020 except Australia (2021). Data for the eight upper-middle-income developing countries are from the Dominican Republic, Guyana, Jamaica, North Macedonia, Panama, Paraguay, South Africa, and Surinam, which were considered to be the most reliable. Benchmark data are based on five developed countries, of which Australia and Canada provided the fullest information. Annual O&M costs include those related to site lease or rent costs, utilities (electricity, telecommunications), security, spare parts and components, instrument and site maintenance, and life-cycle maintenance per station. AWS = automatic weather station; O&M = operations and maintenance; RCI = replacement cost with installation; — = not available. There are two critical considerations for a successful AWS component costs typically range between 10 percent to 20 network: the availability of spare equipment for maintenance percent of the capital equipment cost, depending on network and life-cycle management. The number of spare compo- size and configuration. Because temperature/relative humid- nents—including instruments, data loggers and other periph- ity and solar radiation sensors require frequent maintenance erals (such as communication interfaces), structural elements and recalibration and may have relatively high failure rates (towers and mounting brackets), and power supply compo- based on environmental conditions, the number of spares nents (batteries, charge regulators, and solar panels)—should needed to maintain the reliable operation of a given network be determined in consultation with equipment manufactur- will be higher than, for example, the number of data loggers, ers or vendors, NMSs with significant experience in AWS net- which require less frequent calibration and have lower failure work operation, or entities with comparable expertise. Spare rates. Automatic Weather Stations   233 It is important to note that the spare parts inventory serves the network. Caution should be used as this estimate may un- two purposes: first, it allows for the rapid replacement of sen- dervalue the cost of FTEs depending on staff size and makeup sors or other components in the case of malfunction, failure, or of the NMS. For example, a high number of FTEs are required damage; second, it allows instruments and other components to staff a network of manual stations and the pay scale for this that are nearing the end of their calibration certification to work may be much smaller than the cost of more technically be replaced with new, serviced, or recalibrated components. trained staff to operate AWSs, skewing the average signifi- Whenever an AWS component is removed from a station for cantly lower. any reason, it should be serviced by qualified NMS personnel, a qualified service center, or the component manufacturer. If Personnel required to operate and maintain an AWS net- the component cannot be brought back into specification, the work can be broadly divided into three categories: (1) field component should be replaced. Components that have short- technicians, (2) site maintenance personnel, and (3) support er expected lifetimes than 10 years should be scheduled for personnel. Field technicians (system operators/observers) replacement as part of the overall network life-cycle manage- perform the more technical tasks of maintenance (scheduled ment plan. In such cases, it is good practice to replace these and unscheduled), repair, and upgrades to the AWS systems. components during scheduled maintenance, on a rolling In some countries, they also perform regular upkeep and basis over several years, to reduce needed capital investment minor maintenance of the sites (such as mowing grass; trim- in any single fiscal year. ming vegetation; and cleaning dust and debris from precipi- tation, wind, and solar radiation sensors). In other countries, The total cost of a new AWS site is strongly dependent on its these duties are performed by separate site maintenance configuration and location. The procurement cost of a new personnel. AWS system that includes the suite of instruments already discussed (air temperature, relative humidity, atmospheric Field technicians. Ideally, these technicians should possess pressure, surface wind speed and direction, precipitation, a degree in a related technical discipline to perform the re- and global horizontal irradiance), plus the instrument tower quired tasks. In Canada, a technical degree is required for and mounting system, communication interface(s), and employment in this position; in some other countries, such power supply, is about $25,000 to $35,000. This amount as Norway and Estonia, engineers perform these functions. does not include the cost of land acquisition, site prepara- Field technicians typically receive additional training from tion, construction of perimeter fencing, installation, or con- the manufacturer of the AWS systems or NMS, typically at the nections to utilities. site(s) that they will be supporting. 3.5.6.2 Personnel Site maintenance personnel. Given that their tasks generally The cost of staffing will vary between countries as a result of require much less skill or training, site maintenance person- differences in AWS network designs, individual circumstanc- nel positions are often filled by members of the local commu- es, and salary scales. Countries should use their pay scales nity. They typically earn lower salaries than field technicians, to calculate the costs of staff salaries. Countries that do not and they often live near the AWS(s) they service, which helps have this information readily available and need to estimate to reduce overall NMS costs, increase overall station and net- costs associated with personnel necessary to operate an AWS work performance, and enhance security. network can use an average cost for personnel, expressed as a full-time equivalent (FTE). This can be roughly estimated NMS support personnel. The support personnel are needed by dividing the total salary budget for the NMS by the total for data management and analysis, maintenance planning number of staff it employs. This value may provide a reason- and scheduling, maintaining an inventory of spare parts, and able weighted average of the cost of an individual NMS staff life-cycle management. Their roles include but are not limited member. Multiplying this value by the number of personnel to: (FTEs) required to operate the AWS network will provide an estimate of the annual personnel costs required to support 234    Automatic Weather Stations ■ Specialists/scientists to analyze, interpret data, and create of stations. As a result, the relationship between the number data products of stations and number of field technicians is nonlinear. ■ Information technology (IT) specialists to support data ingestion, quality assurance/quality control (QA/QC) func- For example, Canada occupies a large geographical area (table tions, and storage and flow of data and products to fore- 3.5.2) and employs 0.09 FTEs per AWS station. But the United cast models Kingdom, where the country and network size are 97 percent ■ Network management and planning—including service, and 50 percent smaller respectively than Canada, maintains a incident, change and process improvement management, similar FTE count at 0.08 FTEs per AWS station, due to tech- and life-cycle support nical complexity of its AWSs. Australia operates the second ■ 24/7 operational support from the IT service desk to log largest number of AWS installations at 701 sites but employs field site, communications, server failures, and so on, to 0.05 FTE per AWS. Germany with the largest and most dense ensure repairs are affected in a timely manner. network of 997 AWS station is third in geographical area but has the highest complement of FTEs at 0.11. Their high FTE The number of required support personnel positions will de- count per site was due to absorbing a high number of staff pend on the size and complexity of the network, but this is into the meteorological service during reunification. Over time generally smaller in terms of a per-station ratio than that re- Germany plans to reduce its FTE complement through retire- quired for field technicians. ments and attrition. Austria, which is much smaller than the other countries, employs only 0.04 FTE per AWS. The total number of staff required to ensure reliable op- eration of an AWS network will depend on several factors, For budgeting purposes for large networks (100 or more), an including the number of installations in the network, the com- NMS should use a costing ratio of one FTE for every 10 to 20 plexity of equipment, the number of site visits per year, and stations in the network, depending on the geographical size the geographical distribution of stations. A country with a of the country for a basic network configuration. For a net- large network covering a large surface area will require more work size of less than 100 stations, the NMS will need to plan field technicians and site maintenance personnel to operate the resources for all three categories based on the size of the its network than will a smaller country with the same number country and station density. TABLE 3.5.2  Full-Time Equivalent Staff Numbers for Example Developed and Developing Countries Support personnel per Station density (km2) site (total number in Country Area (km2) AWS sites (number) per station FTEs) Developed countries Austria 84,000 270 311 0.04 United Kingdom 242,500 300 808 0.08 Germany 357,400 997 358 0.11 Australia 7,692,000 701 10,973 0.05 Canada 9,093,500 585 15,544 0.09 Benchmark 0.08 Eight developing upper-middle-income countries Average — — — 0.19 Range — — — 0.03–0.75 Source: Data provided by NMS to the GFDRR of the World Bank. Note: All data are from 2020. Data for the eight upper-middle-income developing countries are from the Dominican Republic, Guyana, Jamaica, North Macedonia, Panama, Paraguay, South Africa, and Surinam, which were considered to be the most reliable. AWS = automatic weather station; FTE = full-time equivalent. — = not available. Automatic Weather Stations   235 How do the FTE averages from developed and upper-middle- 3.5.7 Recommendations income developing countries compare? As table 3.5.2 shows, the benchmark for developed countries (0.08) is consider- In sum, when developing a strategy for implementing ably lower than that of the average upper-middle-income an AWS network, an NMS should consider the following developing ones (0.19), indicating that the sample develop- recommendations: ing countries have, on average, a larger number of FTEs to support a network of AWSs. The range of FTEs for upper-mid- 1. Evaluation is the first place to start. The NMS should dle-income developing countries (0.03–0.75) is also much evaluate and articulate their data needs against the sta- wider than the narrow range of FTEs in sample developed tus of existing resources—including in-house staff tech- countries (0.04–0.11); the latter is possibly explained by sim- nical capabilities, the current state of existing networks, ilar business models. The broader range of FTEs exhibited by and the annual O&M budget—before embarking on ac- the developing countries may be due to differences in busi- quiring and operating a new system. A gap analysis of ness models, with some countries choosing to employ larger current capabilities versus what is required to support numbers of staff in the NMS. a new system is imperative to developing a meaningful implementation plan and TCO. As for pay, an NMS looking to acquire a system of AWS sta- 2. Be sure to include O&M costs. When budgeting the TCO tions will need to adjust the salary base to attract and retain of an AWS network, it is critical to consider the signif- the technical personnel required to operate the more com- icant annual O&M costs over the approximate 10-year plex network. Key factors to keep in mind are: expected life cycle, which often exceed the costs associ- ated with initially establishing the network. ■ The salary scale required to attract and retain the techni- 3. Use FTEs to develop personnel costs. The number of cal personnel needed to operate the more complex AWS field technicians, site maintenance personnel, and sup- network is significantly higher than it is for the MWS. port staff required for reliable network operation should ■ While the staffing required to operate the AWS network is be determined through consultations with other NMSs less, the talent pool with the technical knowledge in-coun- with significant experience in AWS network operation, try is typically smaller and may present recruitment or with entities with comparable expertise. challenges. 4. Timing of installations. If installing a large network, the ■ Staff training costs to support an AWS network are high timing of installations should be staggered. This allows and should be carried out annually or at regular intervals the NMS not only to develop experience in all aspects to ensure that skill sets are maintained and grow as system of network and data management as the network comes enhancements are added. online but also to stagger the replacement of the systems ■ FTEs are required in three areas: (1) field technicians with when they reach end-of-life. This lessens the significant the necessary technical background; (2) site maintenance capital cost burden of replacing the entire network. personnel, who are less technical but important for main- 5. Do long-term upgrade planning. Once the network is taining data quality; and (3) NMS support personnel, who installed, careful planning is essential to ensure that require technical training in data sciences and analysis as the supplier updates and enhancements are installed well as management skills for project, network, and staff. and incorporated. Key to the ongoing success of an AWS network is long-term planning for upgrades to computer Staffing for the talent required is a key element in developing equipment, software, and incoming data analysis. a sustainable AWS network that provides quality data fit for 6. Understanding the TCO is important. Given that AWS its intended purpose. networks are expensive to purchase, install, and main- tain—and can have substantial annual operating costs if the network comprises a large number of individual stations—it is vital to understand the TCO, along with 236    Automatic Weather Stations ensuring that the year-over-year funding is available WMO (World Meteorological Organization). 2013. Manual prior to purchase. on the Global Observing System: Volume I – Global Aspects. 7. Share data openly and freely. Broader sharing of me- WMO-No. 544. Geneva: WMO. https://community.wmo.int/ teorological data to GBON requirements and WMO data wmo-no-544-manual-global-observing-system. sharing policies will provide better forecasting ability and a greater return on investment for the entire global WMO (World Meteorological Organization). 2018. Guide to weather community. Instruments and Methods of Observation, Volume III: Observing Systems. WMO-No. 8. Geneva: WMO. https://library.wmo.int/ 3.5.8 References doc_num.php?explnum_id=9872. Ehinger, J. 1993. Siting and Exposure of Meteorological WMO (World Meteorological Organization). 2021. Manual Instruments (SEMI). Instruments and Observing Methods on the WMO Integrated Global Observing System. WMO-No. Report No. 55 (WMO/TD-No. 589). https://library.wmo.int/ 1160. Geneva: WMO. https://library.wmo.int/?lvl=notice_dis- doc_num.php?explnum_id=9608. play&id=19223#.Ynk624zMKUk.    237 Upper-Air Systems 3.6.1 The Purpose of Upper-Air Systems 3.6 Upper-air observation—also known as a sounding or upper-air sounding—is considered a primary measurement of atmospheric conditions aloft. These soundings provide a three-dimensional representation of the atmosphere for forecasters to use in evaluating the vertical distribution of temperature, humidity, and wind—all critical information for forecasting severe thunder- storms and tornadoes in the summer and winter storms in the winter. Over land, vertical profiles of atmospheric conditions are typically obtained using radiosondes launched from dedicated ground stations and carried aloft by gas-filled balloons. Radiosondes contain sensors for measuring upper-air environmental parameters, electronics for converting the sensor output into digital signals, and a radio transmitter. Radio signals are received by an an- tenna and radio receiver at the ground station, decoded, and converted to Uganda. Photo courtesy of V. Tsirkunov, WBG. meteorological observations, which are compiled into sounding reports that are transmitted to the National Meteorological Service (NMS). Upper-air data are part of the World Meteorological Organization (WMO) Global Basic Observing Network (GBON) and are shared with the international community following WMO protocols and standards. An NMS will commonly use information from upper-air soundings of the at- mosphere supporting (1) localized forecasting for assessing the potential for thunderstorms and heavy rain; (2) precipitation typing (liquid, freezing, or frozen); (3) the location of high-, mid-, and low-level jet streams, turbulence, and wind shear zones; and (4) the potential for the atmosphere to disperse smoke or pollution (stable layers and inversions). In regions prone to strong inversions, the NMS can provide smoke dispersion forecasts or alerts to min- “Upper-air data are part of imize industrial activities that would increase harmful pollutants that could collect over a city. the World Meteorological Organization (WMO) Global The NMS should consider acquiring an upper-air system or systems with the three basic measurements (temperature, relative humidity, and wind) that Basic Observing Network are used in the initialization of the analysis of numerical weather prediction (GBON) and are shared (NWP) models for operational weather forecasting (see WMO 2018a, Chapter 12). with the international community following WMO As with all meteorological data, upper-air data increase in value when sup- porting additional meteorological applications, such as: protocols and standards.” 238    Upper-Air Systems ■ Initialization of NWP models, especially regional and local has been set to achieve 95 percent data availability from the forecasting and nowcasting upper-air stations with two daily launches and 85 percent of ■ Environmental pollution studies BUFR reports to heights of 100 hectopascals (hPa) within 50 ■ Atmospheric stability monitoring minutes of launch time. ■ Civil aviation ■ Artillery and space vehicle launch calculations Accuracy requirements for the different applications (global ■ Upper-air climate change studies. NWP, climate monitoring, and nowcasting) are given in WMO 2018a, Chapter 12; WMO 2019; and WMO OSCAR. It is nec- Unlike other in situ meteorological observation systems, essary to review the requirements for all the expected ap- upper-air measurements are highly dependent on the use of plications when determining the accuracy requirements for one-time consumables (radiosondes). At two launches per a particular site or network. Upper-air measurements repre- day, a single upper-air station will require 730 radiosondes sent several distinct challenges. First, the rapid rate of verti- to support the measurement of the upper atmosphere. In de- cal ascent (typically 5–6 meters per second, or m/s) requires veloping countries, NMSs typically do not have the budget or sensors with very short response times. Second, both air do not set up the necessary processes required to maintain a temperature and water vapor content can vary significantly constant supply of radiosondes. This shortfall has historical- over short vertical distances, and thus short response times ly represented a significant point of failure in most upper-air are necessary to reduce errors associated with lagged sensor measurement programs. This chapter will outline some of the responses; this is exacerbated at high altitudes by the ex- key elements of planning, implementation, and costing for tremely low atmospheric pressures encountered there, which both manual and automated upper-air systems. increase sensor response times. Third, the meteorological parameters of interest vary significantly over the elevations 3.6.2 Environmental Parameters Measured measured by radiosondes, which requires systems to operate during Upper-Air Soundings over an extremely wide range of environmental conditions. Accurate measurements of the vertical structure of tempera- 3.6.2.1 Air Temperature ture and water vapor fields in the troposphere are extreme- Upper-air temperature measurements are typically made ly important for all types of forecasting, especially regional using electrical resistance thermometers (such as thermis- and local forecasting and nowcasting. Radiosonde observa- tors, thermocapacitors, or thermocouples). Temperatures tions are used regularly for measurements up to heights of taken during a sounding can vary over extremely wide ranges, Photo 25–35 courtesy of Vladimir kilometers. Tsirkunov If observations are to be made for climate and thus sensors should be able to operate between −95°C monitoring purposes, flights should reach 30 kilometers on a and +50°C. They should have a short response time to re- regular basis. Requirement specifications can differ from ap- duce errors resulting from thermal lag during ascent and be plication to application, so if radiosonde observations are to designed to be as free as possible from errors due to solar be used for more than one purpose (such as regional weath- heating. Temperature sensors should be mounted in a posi- er forecasting, aviation safety, and climate monitoring), the tion above the main body of the radiosonde, on an arm or most stringent of requirements should be followed. outrigger that holds the sensor in a fixed position away from any air that might have been heated or cooled by contact with Launches should be performed twice per day as per GBON the balloon or radiosonde body. All radiosonde temperature requirements; WMO No. 1160 (WMO 2019) recommends sensors will be impacted by heating during daytime sound- that observations should be recorded at 0000 and 1200 ings as a result of direct solar radiation and radiation back- Coordinated Universal Time (UTC) (although this recommen- scattered from the Earth’s surface and from clouds. In modern dation is currently under review). Upper-air applications will radiosonde systems, the resulting temperature errors are cor- benefit from data only when they are distributed internation- rected during data processing. ally, with high availability, and in near-real time. For example, in WMO Region VI (Europe and the Middle East), the target Upper-Air Systems   239 3.6.2.2 Relative Humidity cases barometric pressure; a GPS unit for location track- Upper-air relative humidity measurements are most common- ing and wind speed and direction calculation; electronics ly made in modern systems with a thin-film capacitor, which that convert measured values to a digital signal that can be is mounted externally without a cover on an arm or outrigger transmitted to the ground station; a radio transmitter with that holds the sensor in a fixed position away from any air that antenna; and batteries. might have been heated or cooled by contact with the balloon or radiosonde body. Accurate measurements require that the Depending on station configuration, these components are sensor be well ventilated but protected from the deposition assembled either by the human operator or by the launch sys- of water or ice from clouds. Modern radiosonde relative hu- tem automatically. Modern radiosondes use GPS navigation midity sensors perform accurately over temperatures ranging to provide location information, allowing direct measure- from −70°C to +40°C; however, because water saturation ment of geometric height (height above mean sea level), from cannot be accurately measured below about −50°C, relative which geopotential height (which accounts for variations in humidity should be calculated using an appropriate mathe- gravity due to latitude and altitude) is calculated. matical expression. Batteries used in radiosondes should be of sufficient capacity 3.6.2.3 Upper Wind to provide power for the required flight time, including pos- sible delays; generally, three hours is sufficient for flights up Upper-wind measurement is accomplished by accurately to 5 hPa. Batteries should have a long storage life and be as tracking the three-dimensional position of a balloon. Most light as possible, and not pose an environmental hazard once commonly, tracking is performed by a Global Positioning the radiosonde falls to the ground. Batteries used in modern System (GPS) receiver onboard the radiosonde, allowing si- radiosondes are generally dry-cell (alkaline) or lithium. multaneous measurement of upper-air temperature, relative humidity, and wind. It is assumed that the balloon is freely Radio transmitters used in radiosondes today usually trans- carried along by the wind, and that its motion accurately re- mit at 400 megahertz (MHz) (maximum power: 250 milli- flects winds at the elevation of the balloon. Wind speeds are watts, or mW) or 1,680 MHz (maximum power: 330 mW). reported in units of (m/s). Wind direction is reported with 0° The NMS should ensure the radiosonde radio transmissions (or 360°) representing wind blowing from the north, and 90°, comply with local regulations. 180°, and 270° representing winds from the east, south, and west, respectively. 3.6.3 Determining Where to Place Individual Stations in the Network across the 3.6.2.4 Radiosonde Launch Components Landscape Typically, a radiosonde launch comprises the following components: Typical spacing between upper-air stations is 100–500 ki- lometers. If an NMS does not have the experience in siting ■ A balloon to contain the lifting gas upper-air stations, it should employ an experienced con- ■ Gas for lift, typically helium or hydrogen sultant, work with WMO personnel, or consider seeking the ■ A parachute, if required to reduce hazards associated with advice of a more experienced NMS on working through the uncontrolled radiosonde descent (especially in populated process of site selection. areas) ■ A tether to suspend the radiosonde far enough below the When selecting installation sites for each station, there are balloon that air temperature and relative humidity mea- several factors to consider. Each site selected will need to be surements are not affected by air warmed or cooled by balanced against all these factors. And each location should passing over the balloon be evaluated based on siting requirements specified by the ■ A radiosonde unit with the following components: sensors launch system supplier(s), advice from experts, and the con- to measure air temperature, relative humidity, and in some siderations discussed below regarding the sites’ logistics, 240    Upper-Air Systems security, land availability, communications, power, and envi- ■ Power. Typically, upper-air stations require mains power, ronmental conditions. Based on a cost analysis of these fac- as solar panels cannot produce enough energy for up- tors, the final site location might differ by several kilometers per-air systems. when considering the following: p Is there mains power at the site? p Can mains power be brought to the site, and if so at ■ Logistics. The remoteness of a site can add significantly to what cost? the operational costs of a network. ■ Environmental conditions. Environmental conditions at p Is the site easily accessible? Sites that are in extreme- the site play a large role in how often the site will require ly remote locations will increase the travel costs sig- maintenance and repair. Is the ideal location for the site nificantly for twice daily launches, equipment repair subject to: and station maintenance. p Extreme wind, dust, or debris that can affect the abili- p Is there an alternative location that is less expensive ty to safely assemble, handle, and launch balloons, or to travel to that would provide the same or similar affect the measurement accuracy of the launch site’s data that supports the end use, reducing logistical automatic weather station (AWS)? challenges and costs? Time spent traveling to sites p Nearby bodies of fresh or (in particular) saline water, increases the human resources required by the NMS which can increase corrosion rates and require more to support the network. frequent maintenance or replacement? p What are the additional costs for the civil works for ■ Location of other in-situ observation systems. Collocating supporting the station given the remoteness of the upper-air stations with other in-situ instrumentation in- site? stallations should also be considered to facilitate infra- ■ Security. Is the area safe enough that theft and vandalism structure and operational cost efficiencies. would not be significant issues? Is the site location safe for the NMS personnel who will operate and service the sta- Each of the above factors will need to be considered and bal- tion? Note that manual radiosonde operations commonly anced against operational costs and spatial density require- require personnel to work at night, which should be con- ments of the data. In the end, the ideal site may still be the sidered in the context of security. best location to support the data needs. By considering the ■ Land availability. Does the government own the land above factors, an NMS should begin to understand the cost where the station is to be sited? Will the government need implications of their decisions. to purchase the land (resulting a capital cost) or lease it (resulting in an annual budgetary cost)? 3.6.4 Configuring and Installing the Upper-Air ■ Communications. It is extremely important that the site Station have a reliable means of communication; otherwise, the site is not ideal for operational meteorology. Upper-air stations can be configured for human-assisted or p If cellular is the preferred communication method, automated operation. In automated operations, an operator is there reliable coverage in the area? What is the is required to load the radiosondes into an automated sys- monthly cost of the communication plan based on the tem only as required, typically every two to four weeks. The volume of data to be transmitted? Costly plans in an choice of system largely depends on the availability of the area may suggest that alternative methods of trans- operators for the planned stations and the cost of sustaining mitting the data should be considered. Cellular sig- operations. The ground station typically consists of the fol- nal strength at the site should be measured, either by lowing components: checking the bars on the phone or by using a signal power meter. ■ A radio antenna and receiver, computer with specialized p Is there landline availability for telephone or internet software, and communications interface communications? If not, what is the cost of bringing ■ A building for mounting the radio receiving antenna and the service to the site? housing the radio and computer equipment Upper-Air Systems   241 ■ If tracking is not performed by GPS onboard the radio- A human-operated launch station (also known as a manual sondes, balloon tracking equipment launch station) requires (1) a secure building for housing the ■ An AWS with air temperature, relative humidity, and baro- radio receiver and computer equipment, with storage space metric pressure sensors for making surface observations for sounding equipment (balloons, radiosondes); (2) a cov- to be included in the meteorological messages ered preparation area for filling the balloons and preparing ■ A system for checking radiosonde measurements prior to the radiosondes for launch; and (3) a platform from which the launch radiosondes can be safely released figure 3.6.1). Photo 3.6.1 ■ Safe storage of the gas cylinders or, if hydrogen is to be shows a manual launch of a radiosonde just prior to release produced on-site, a hydrogen generator and the equipment at Vienna Hohe Warte by the Austrian Meteorological Service and materials to safely produce the gas (ZAMG). ■ Storage space for radiosondes and balloons ■ A system for filling the balloons FIGURE 3.6.1  Manual Launch Site ■ A launching platform with shelter for balloon preparation ■ A secure perimeter fence. It is suggested that an industry expert or an experienced NMS be consulted when developing an upper-air network to en- sure that the observational needs of the NMS are met. The installation site should be large enough to accommodate the buildings or automated launching system and should not have any high obstacles near the launch area. Once the gen- eral area for siting a station has been chosen, it is necessary to select a specific site for the facility. The following criteria should be considered when selecting the final site for both Source: Figure courtesy of Vaisala. automated and manual launch sites: Note: (1) = radiosondes preparation area; (2) = covered area for filling the balloons; (3) = radiosonde launching area. ■ The site must be accessible by an all-weather road to facil- itate delivery of supplies and proper station maintenance. There should be a storage facility for the gas cylinders, or a ■ The site should have good drainage and should not be in a hydrogen generator and associated equipment and materials location prone to flooding. if hydrogen is to be produced on-site (not shown). If the lift- ■ The site should be free from natural or manmade obstruc- ing gas to be used is hydrogen, additional safety precautions tions that would interfere with the flight path of balloons and measures must be taken given the extremely flammable or the tracking of the radiosonde signal. nature of the gas (see WMO 2018b, Chapter 8 and references ■ Utilities such as electric power, water, sewage, and com- therein). An AWS (not shown) or environmental chamber is munications must be available. required for checking radiosonde measurements immediately ■ The site must be surveyed to ensure that the radio frequen- prior to launch. Prevailing wind conditions should be consid- cy band, electronic equipment, and telecommunications ered when planning the layout of the station area, and local are free from any interference. regulations and requirements must be followed when estab- ■ Local weather conditions should be suitable for safe oper- lishing the site and the buildings. ation, as high winds and extreme temperatures can impact the capability to launch balloons. For general guidance on establishing an upper-air site, see WMO 2019, Chapter 3.3. 242    Upper-Air Systems PHOTO 3.6.1  Human-Operated Launch Site Source: Photo courtesy of M. Staudinger, Austria Meteorological Service. An automated launch system requires (1) a gas storage area FIGURE 3.6.2  Automated Launch Site located close to the automated system; (2) a level installa- tion site, such as a concrete or steel platform; (3) an AWS for performing surface observations prior to each sounding; and (4) a security fence (figure 3.6.2). As with the manual station, the layout of the station area should consider prevailing wind conditions, and the platform’s elevation above ground should consider snow or water amounts. In addition, local regula- tions and requirements must be followed when establishing the site and its components. Photo 3.6.2 shows an automated launch facility operated by Vaisala on the coast of the Gulf of Finland. Source: Figure courtesy of Vaisala. Note: (1) = automated launch system; (2) = gas storage; (3) = level installation site; (4) = AWS; (5) = security fence. Upper-Air Systems   243 PHOTO 3.6.2  Automatic Sounding System ■ Parachutes to reduce the hazard posed by falling radio- sondes, especially in densely populated areas. ■ Gas, either hydrogen or helium, to fill the balloons. Both hydrogen and helium can be supplied in pressurized bot- tles; alternatively, hydrogen can be generated at the site. The determining factors for the choice of gas are safety and local availability. The use of hydrogen requires ad- ditional measures (including building design, protective clothing, and operator training) to ensure safe operation and storage. (Guidelines for safe handling are provided in WMO 2018b, Chapter 8, but local regulations and safety procedures must always be followed.) At manual stations, tasks associated with each sounding include: Source: Photo courtesy of M. Lehmuskero, Vaisala. ■ Operating the sounding system and corresponding work- station and software for radiosonde preparation, data re- 3.6.5 Planning for Operation, Maintenance, and ception, and data dissemination Upgrades for the Upper-Air Network ■ Preparing the radiosonde, including unpacking the radio- sonde and performing ground checks according to the ven- The three categories of requirements for the operation, main- dor instructions tenance, and future upgrades of an upper-air network are ad- ■ Filling a balloon with gas and attaching the radiosonde to dressed in this section. the balloon ■ Checking radiosonde air temperature and relative humid- 3.6.5.1 Operational Requirements ity readings against values recorded by the station AWS ■ Releasing the balloon Radiosonde observations require the following consumables: ■ Entering surface observations into the sounding software (if not done automatically) ■ Radiosondes to measure the upper-air temperature, hu- ■ Ending meteorological data (if not done automatically). midity, pressure, height, and wind speed and direction. They should be supplied with an unwinder for deploying a Each launch will generally take two to three hours of operator tether that holds them at a suitable distance from the bal- time. Depending on the number of daily launches, sustainable loon and eases the launch, especially in windy conditions. operations will need two to four persons capable of handling ■ Meteorological balloons to carry the radiosondes; their the tasks (including day and nighttime launches, weekends, size depends on the desired measurement altitude range. and holidays). For automated launches, a human operator Larger balloons reach higher altitudes but are more expen- is needed to load the consumables to the system every few sive and require more lifting gas. To prevent degradation weeks, depending on the overall capacity of the automatic of the balloons during storage, they should be stored in a sounding system. Each visit takes a maximum of one day, not location away from heat and ozone sources, and the mini- including travel time to and from the station. mum number of balloons required to maintain reliable op- erations should be ordered at regular intervals. (See WMO 3.6.5.2 IT Requirements 2018b, Chapter 8, for more information on balloon materi- als and properties, ascent behavior and performance, stor- Upper-air systems are generally operated with a computer age, and handling.) workstation that can be supplied by the vendor or sourced locally. A complete sounding archive takes 10–20 megabytes 244    Upper-Air Systems (MB) of storage space, so one complete year of upper-air data techniques to be followed in determining and correcting mal- consumes about 7–15 gigabytes (GB) of storage. Computer fail- functioning equipment. ure is one of the main causes of upper-air operation downtime. Given the relatively low cost of computer equipment and spe- 3.6.6 Estimating the Total Cost of Ownership of cialized software (provided by the radiosonde manufacturer), a Upper-Air Stations spare computer, loaded and tested with the required software, should be held on-site or at a central NMS facility. How do NMS decisions impact the total cost of ownership (TCO) over a 10-year period? The reality is that money used Operational software is typically configured as part of the to purchase a system that is too large or too complex will be commissioning of the upper-air system by the vendor. This wasted if the annual budget is too small to support the net- software package usually (1) includes a visual data display; work. To aid in the costing process, the NMS should complete (2) performs automatic correction for errors due to changes the Total Cost of Ownership Exercise. (See chapter 3.8, using in sensor response time with elevation, radiative heating of the additional information provided below that is specific to temperature sensors, extremely low temperatures, and other upper-air stations as an aid.) sources of error; and (3) interpolates values when data are missed due to poor reception of the radiosonde radio signal. The network cost is the sum of the cost of individual stations. (For additional information about software requirements and Note that individual stations might have different operational radiosonde profile corrections, see WMO 2018a, Chapter 12.) requirements, as will be the case if the network includes both human-operated and automated stations. 3.6.5.3 Maintenance of Upper-Air Stations A maintenance program is required to keep equipment in 3.6.6.1 Procurement Costs operating condition and to ensure reliable data quality. Procurement costs for an upper-air site include: Maintenance should include preventive maintenance, re- quired equipment checks and recalibration, periodic cleaning ■ Land costs and lubrication, corrective and adaptive maintenance, and ■ Investments on buildings, balloon filling facility (manual equipment replacement and upgrades, as necessary. system) or platform (automated launching system), and associated civil works Preventive maintenance is important and must be conducted ■ Procurement cost of the sounding system regularly on all equipment. It is more efficient and effective to ■ Gas storage or generation facilities and hydrogen genera- keep equipment running than to repair it following a break- tor, if applicable down. Scheduling preventive maintenance is a necessity for ■ Project management costs for establishing a station, with the sustainable operation of an upper-air system. Through installation, training, and any acceptance costs testing and evaluation programs, upper-air equipment man- ■ Power, other utilities, and communication infrastructure ufacturers have determined appropriate preventive main- costs. tenance schedules and procedures, which must be closely followed through the life cycle of the equipment to ensure Once the site civil works are accomplished, an upper-air sys- correct functioning. Systems should require only one preven- tem can generally be commissioned within one to two weeks. tive maintenance visit per year, but more frequent mainte- nance might be required if local conditions increase wear 3.6.6.2 Consumables compared with typical installations. The annual consumable costs can be estimated once the An effective corrective maintenance program involves en- launching schedule (for example, two launches per day) and suring the availability of adequate supplies, spares, and the measurement altitude (for example, 30 kilometers) range trained electronic and other maintenance personnel. Original are determined. Total annual costs for the consumables are equipment manufacturers usually prescribe procedures and calculated as: Upper-Air Systems   245 ■ [365 days] × [Number of daily launches] × [Cost of radio- the system, the maintenance needs and information on the sonde, balloon, tether, and parachute]. spares should be requested as a part of the initial procure- ■ [365 days] × [Number of daily launches] × [Volume of bal- ment. In addition, buildings at a manual upper-air site re- loon in m³] × [gas cost per m³]. quire maintenance, both manual and automated launching sites require electricity and communications, and computers This is the nominal cost of the consumables with the fixed and other information technology (IT) infrastructure need re- schedule. The quality costs due to missing or failed sound- placement typically every four to five years. The annual main- ings should be considered as well. Furthermore, additional tenance costs, including spare parts and maintenance visit (adaptive) soundings should be considered when calculating costs, should be tracked to ensure that the costs remain at the cost. These can be performed, for example, in the event of the expected level. approaching severe weather and can mean 10 percent more flights in addition to the fixed synoptic schedule (see table 3.6.6.5 Administrative Costs 3.6.2). Administrative costs include operative network management costs, procurement cost, operational quality costs, costs for 3.6.6.3 Operator Costs recruitment and training, and the cost of leasing the site if The method of launch (manual or automated) determines the not government owned. They are largely shared with all the costs of the operators: stations in the upper-air network and not multiplied by the number of stations. One of the unique features of radiosonde ■ With manual launches, operator costs depend on wheth- measurements is the complete dependance on consumables. er the operators are employed full time or part time, and Because of this, restocking and procurement contracts are how many operators are needed. It may also be possible to critical, and should be carefully maintained throughout the have contracted operators to do the launches. life cycle of the upper-air system or network. This requires ■ With automated launches, operator costs depend on the that operational costs are budgeted and provided on a recur- number of annual site visits. Depending on the capacity of ring basis for a similar duration; radiosonde, balloon, and lift- the launching system, this can range from weekly to month- ing gas inventories are carefully monitored; and new stock is ly visits. It may also be possible for contracted operators to ordered well before inventories are depleted. do the loading of sondes into the carousel. There are extra costs associated with monitoring and regular checks of the Quality and the related costs of the upper-air network should systems, along with responses due to automated alerts or be monitored to ensure that the network reaches the targeted non-reporting. data availability. Reasons for missed or failed soundings, re- sulting in loss of data and increased quality costs, should be Especially with manual launches, there may be a need for resolved within the network operations or with the vendor. more recruitment or training to sustain the operations. For a new station, training can be provided by the senior operators 3.6.6.6 Upper-Air Network Operating, Maintenance, or the vendor. Typically, an operator training session can be and Life-Cycle Cost Examples completed within one week. An upper-air network represents a significant long-term in- vestment for an NMS, given the high costs associated with 3.6.6.4 Maintenance Costs installing a new upper-air system and the ongoing costs Automated systems typically require one annual preventive associated with regular launches. Costs related to the op- maintenance visit at each site as well as possible corrective erations and maintenance (O&M) for upper-air networks in maintenance visits, along with spare parts, which are usual- Australia, Austria, Canada, Germany, and the United Kingdom ly recommended by the vendor. Maintenance can be carried are shown in table 3.6.1. All five countries operate successful out by local engineers at the NMS or by the vendor, depend- and well-maintained networks, with consistently high levels ing on the region and the system’s location. When procuring of performance. 246    Upper-Air Systems ■ Australia operates the largest network of this group with has purchased 2 automated systems now under test. There 38 upper-air installations. These are automated sites, with are two balloon launches per day. the exception of 6 that are manual. There are two balloon ■ Germany operates 10 automated sites. There are 2 balloon launches per day. launches per day. ■ Austria operates 1 upper-air installation, which is manu- ■ The United Kingdom operates 7 sites—with a mix of 2 man- ally operated, but it is investigating an automated system. ual stations operating two flights a day; 4 automated sta- There are two balloon launches per day. tions operating one flight a day; and 1 station on a military ■ Canada operates the second largest network with 30 in- base operating as required. stallations. These are all manually operated, but Canada TABLE 3.6.1  Upper-Air Network Cost of Operations and Maintenance Data for Example Countries Annual O&M Annual O&M Upper- Expected cost (without Annual labor plus labor cost Lifetime O&M air sites Performance lifetime RCI per site labor) per site cost per site per site (US$) + labor cost per TCO per site Country (number) (uptime, %) (years) (US$) (US$) (US$) site (US$) (US$) Developed countries: Manual upper-air sites Austria 1 100% 20 $980,000 $118,000 $309,000 $427,000 $8,540,000 $9,520,000 Canada 30 90% 20 $1,580,000 $269,000 $90,000 $359,000 $7,180,000 $8,760,000 United 2 96% 20 $860,000 $128,000 $85,000 $213,000 $4,260,000 $5,120,000 Kingdom Manual 1 95% 20 $1,000,000 $200,000 $223,000 $423,000 $8,460,000 $9,460,000 benchmark Developed countries: Automated upper-air sites Australia 32 95% 15 $750,000 $209,000 $82,000 $291,000 $4,365,000 $5,115,000 Canada 2 —  15 $680,000 $128,000 $85,000 $213,000 $3,195,000 $3,875,000 Germany 10 99% 15  — $221,000 $72,000 $293,000 $4,395,000 $4,395,000 Automated 1 97% 15 $715,000 $200,000 $82,000 $282,000 $4,230,000 $4,945,000 benchmark Developing country sample average: Manual upper-air sites Upper- 1 86% 20 — $80,000 $21,000 $101,000 $1,515,000 — middle income Source: Data provided by NMS to the GFDRR of the World Bank. Note: All data are from 2020. A 15-year period was used for automated systems and a 20-year period for manual systems when calculating the lifetime O&M plus labor cost per site to normalize the data between countries. Manual station benchmark calculations are based on sites from Austria (1), Canada (30), and the United Kingdom (2) and include equipment, computer, and balloon-filling shed. Automated station benchmark calculations are based on sites from Australia (32), Canada (2), and Germany (10) and cover the automated systems and computers but not civil works and installation, which are minimal as the system is containerized. Austria and the United Kingdom are not included, as their O&M costs are significantly below the others for country-specific reasons. Upper- middle-income country data are based on averages for Colombia, the Dominican Republic, Panama, Paraguay, and South Africa, and the collective performance (uptime) of 86 percent is arguably lower than shown (see box 3.6.1). O&M = operations and maintenance; RCI = replacement cost with installation; TCO = total cost of ownership; — = not available. Upper-Air Systems   247 Typical uptime performance among developed nations ranges The biggest chunk of the annual O&M costs (without labor) from 90 percent to 100 percent. Because uptimes less than for developed countries—about 65 percent—is associated 90 percent prevent the true value and benefit of upper-air with the cost of consumables related to balloon launches. systems from being realized, developing nations should set Australia and Canada have the largest networks and by far the an initial performance target of 90 percent. Once the NMS most launches per year (table 3.6.2). This suggests that these has gained experience operating and maintaining its upper-air two countries might have much lower per flight costs than systems, the target can be increased to the performance Austria and the United Kingdom. But this is not the case; all of benchmark of more than 95 percent. With regular main- these countries stand at about $170 per launch. One reason tenance and repair, proper management of a spare parts may be that a significant component of the consumable cost inventory, and well-trained maintenance personnel, it is rea- in Australia and Canada, which are large geographically, is sonable to target an expected life cycle of 20 years for a man- related to shipping launch materials to remote sites. Another ual upper-air site and 15 years for an automated site. reason may be the variable cost of helium from country to country. That said, Germany’s per flight cost of consumables For developed countries as a group, annual O&M costs (with- is higher than that of the other countries, indicating that out labor) per site (assuming two balloon launches daily) an NMS needs to consider its own individual circumstances have a benchmark of $200,000 for both manual and auto- when budgeting for all launch materials. mated systems and include costs for consumables; building and site leases; utilities (electricity, water, and telecommuni- BOX 3.6.1  Performance in Developing Countries cations); security; spare parts and components; and mainte- Restated nance of the launch system, infrastructure, and site. However, labor costs vary, ranging from $223,000 for manual systems to $82,000 for automated ones. As a result, annual O&M plus As for developing countries, annual operations and labor cost per site reaches $423,000 for manual systems, far maintenance (O&M) costs (without labor) stand at above the $282,000 for automated ones. $80,000, which is 40 percent of the developed coun- tries’ benchmark (table 3.6.1). With a per site bench- Among the five developed countries shown in table 3.6.1, mark for consumables in developed countries of there is a wide range for the annual O&M costs (without $125,000 annually, it is evident that the majority of labor), reflecting different approaches to O&M outlays. For developing countries are not achieving the standard example, Austria’s O&M number is considerably less than of two balloon launches per site per day in accordance the others, as it is not investing in its current site while it with the Global Basic Observing Network (GBON). A explores acquiring an automated system. For the United revised estimate of developing country uptime would Kingdom, current O&M costs are also significantly lower than be closer to 50–60 percent instead of the 86 percent. they are for the others, as there was no substantial, non- capital investment in its upper-air network in terms of life- cycle improvements. However, the United Kingdom does make life-cycle improvements to its network every few years, which enables it to repair (rather than replace) some compo- nents, lowering hardware costs and deferring (and thereby reducing) annual O&M costs. In addition, the United Kingdom incurs significant radio spectrum access costs, but these were not included. Not all NMSs pay for frequency licenses, but in countries where applicable, the NMS must include this cost as part of their annual O&M budget. 248    Upper-Air Systems TABLE 3.6.2  Costs Associated with Balloon Launches Total cost of Consumable cost Upper-air Launches per consumables per site per year installations year by country Cost per flight (2 launches per day) Country (number) (number) per year (US$) (US$) (US$) Australia 38 27,740 $4,526,000 $163 $119,105 Austria 1 730 $118,000 $162 $118,000 Canada 30 21,900 $3,600,000 $164 $120,000 Germany 10 7,300 $1,324,000 $181 $132,400 United Kingdom 7 4,015 $660,000 $164 $120,000 Benchmark  n.a. n.a. n.a.  $170 $125,000 Source: Data provided by NMS to the GFDRR of the World Bank. Note: This table assumes launches of two balloons a day in accordance with the Global Basic Observing Network (GBON). Total cost of consumables includes lifting gas (helium) for the five countries. n.a. = not applicable. The cost of a new upper-air site is strongly dependent on its can rectify the problem immediately with a manual system configuration, features, and location. The procurement cost of by launching another balloon. And it should carry out due dil- a new manual upper-air system is benchmarked at $1 million igence when purchasing an autosonde system, weighing the (about $1.5 million with a hydrogen generator for the lifting cost of missing data due to failed launches. gas) for a developed country—including the radio receiver and antennas, computer and peripherals, system installation, Automated launch stations typically use helium as the lift- and the calibration system or AWS necessary to confirm ra- ing gas, which can pose significant challenges if helium is diosonde calibration before launch. The cost of land acquisi- not readily and reliably available. In contrast, hydrogen can tion, site preparation, construction of a secure building and be generated onsite, but is expensive to purchase, at about perimeter fencing, and connections to utilities may be on top $500,000, exclusive of the building necessary to house the of these costs, which vary considerably between countries system. The use of hydrogen also requires a clean water and sites. And developing countries may experience a lower source, and operating personnel require specialized training cost, given that land procurement and civil works costs play a and personal protective equipment (PPE), given the explosive smaller role than they do in developed countries. potential of the gas. These factors can significantly increase the total cost of an automated upper-air system. Autosondes are typically recommended for remote sites—and where the NMS lacks personnel experienced with upper-air 3.6.6.7 Personnel launches. Automated launch systems are usually provided as Countries seeking to estimate costs associated with person- a containerized system with a developed country benchmark nel necessary to operate an upper-air network can use an av- cost of $715,000 (about $1.2 million with a hydrogen gen- erage cost for personnel, expressed as a full-time equivalent erator for the lifting gas). This cost will vary depending on (FTE). This can be roughly estimated by dividing the total features (like carousel size, which determines the length of salary budget for the NMS by the total number of staff it em- time the system can perform autonomously, typically 1 to 4 ploys. This value will provide a reasonable weighted average weeks). Although the capital and annual labor costs are lower of the cost of an individual NMS staff member. Multiplying than the cost of a manual launching system, their actual life- this value by the number of personnel required to operate time is also five years shorter indicating the need for more the upper-air network will provide an estimate of the annu- frequent replacement. The NMS should also carefully con- al personnel costs required to support an upper-air network. sider the cost of lost data when an automated launch fails, Caution should be used as this estimate may undervalue the as a launch cannot be redone; in contrast, onsite personnel cost of FTEs, depending on staff size and makeup. Upper-Air Systems   249 Staffing costs vary significantly (from $87,000 to $309,000) For manual stations, each balloon launch takes about two among the developed countries, with Austria appearing to three hours to complete, with launches being performed four times higher—explained by the fact that its FTE count at 00:00 and 12:00 GMT each day. Typically, upper-air field is roughly four times higher than that of Australia and the technicians are trained to perform maintenance and other United Kingdom, which operate larger networks with an FTE tasks on other components of an NMS’s observational system count of about 1 per site (table 3.6.3). Austria’s average labor (such as weather radar or AWS), which reduces the overall cost per FTE is about $84,000. NMSs operating a low number number of trained and qualified personnel that the NMS is re- of upper air sites need to be aware of the minimum comple- quired to keep on staff. Furthermore, by providing additional ment of FTEs required to support operations. This variability hours, this practice changes field technician positions from is even broader when comparing within the developing coun- part-time to full-time, making it easier to recruit and retain tries (average labor cost of $21,000). qualified personnel. Note that for accounting purposes, the FTE value assigned to a full-time field technician might be TABLE 3.6.3  Full-Time Equivalent Staff Numbers for Example less than 1.0 if that technician spends only a portion of their Countries time operating and maintaining upper-air systems. Support personnel Upper-air per site How much dedicated staffing is required to support launch- Country sites (number) (total number, in FTEs) es (without manual surface weather observations)? As table Developed country 3.6.3 shows, a site needs about 3.34 FTEs to cover a 24-hour Australia 38 1.04 period; if manual weather observations are also performed Austria 1 3.66 at the location, 6 to 7 FTEs may be required. If a network of Canada 30 1.11 15 manual sites needs to be staffed, the FTE complement to operate it would require 62 FTEs. But for a country with au- Germany 10 1.17 tomated sites, the dedicated staffing needs are lower. Using United Kingdom 7 0.97 Australia’s 32 automated sites as a benchmark, a single site Benchmark requires only 1.04 FTEs (see box 3.6.2 for the approach used Manual —  2.84 by Meteorological Services Canada, or MSC). Automated  —  1.59 Developing country The total number of field technicians required to ensure Upper-middle — 2.17 reliable operation of the upper-air network will depend on income several factors, including the number of installations in the Source: Data provided by NMS to the GFDRR of the World Bank. network, type of systems installed, and the geographical dis- Note: All data are from 2020. Upper-middle-income country data are based tribution of stations. A country with a large network covering on averages for Columbia, the Dominican Republic, Panama, Paraguay, and South Africa. FTE = full-time equivalent; — = not available. a large surface area will require more field technicians to op- erate its network than will a smaller country with the same number of stations. As a result, the relationship between the The personnel required to operate and maintain an upper- number of stations and number of field technicians is nonlin- air network can be broadly divided into two categories: field ear. For a single manual station, one to two FTE field techni- technicians and support personnel. For manual upper-air cians are recommended to allow deployment of balloons at stations, field technicians (system operators/observers) per- the prescribed 12-hour intervals and to ensure reliable sys- form radiosonde calibrations, fill balloons with lifting gas, tem operation should one employee be absent (such as sick assemble the launch vehicles, and launch the balloons twice leave, vacation leave, or departure from employment). each day. For automated stations, field technicians resup- ply the systems at regular intervals, often monthly. In both cases, field technicians also perform maintenance (sched- uled and unscheduled), repair, and upgrades to the systems. 250    Upper-Air Systems ■ 24/7 operational support from the IT service desk to log BOX 3.6.2  Meteorological Services Canada Contracts field site, communications, server failures, and so on to Out O&M ensure repairs are affected in a timely manner. In Canada, Meteorological Services Canada (MSC) con- The number of required support staff positions is often over- tracts out the operations and maintenance (O&M) of 25 looked in the TCO and will depend on the size and complex- of its 30 upper-air sites to private sector partners. Given ity of the network, but it is generally smaller in terms of a the country’s large size, it is more cost-effective to use per-station ratio than that required for field technicians. community or private sector organizations to cover the 24-hour period of operation of the sites. Because launch- ing and maintenance tasks take only a small portion of 3.6.7 Recommendations the day, this also allows the private sector partners to perform other work for other organizations. For exam- In sum, when developing a strategy for implementing an ple, contracted staff can work as electricians in the area upper-air station or network, an NMS should consider the fol- around the upper-air site they operate, while being avail- able for scheduled balloon launches. lowing recommendations: In community organizations, system operators/observ- 1. Evaluation is the first place to start. The NMS should ers often support other municipal functions. MSC also operates a school for training both contractors and MSC evaluate and articulate their data needs against the staff in upper-air O&M. The small number of stations that status of existing resources—including in-house staff MSC directly operates and maintains are in remote loca- technical capabilities, current state of existing net- tions. For these stations, the staff complement (employed works, and annual O&M budget—before embarking on by MSC) is five to seven; MSC also provides housing for acquiring and operating a new system. A gap analysis of these personnel. Sites operated by MSC include three in the High Canadian Arctic and one on an island off the current capabilities versus what is required to support Atlantic coast, plus the training school. Together, these a new system is imperative to develop a meaningful im- have a combined complement of about 30 FTEs. plementation plan and TCO. 2. Be sure to include O&M costs. When budgeting the TCO of an upper-air station or network, it is critical to con- sider the significant annual O&M costs over the approx- Regardless of whether the upper-air installations are oper- imate 15- to 20-year expected life cycle of an upper-air ated and maintained by NMS personnel, a private-sector system or network, which exceed the costs associated partner, or in combination, the NMS will require dedicated with initially establishing the system(s)—especially support personnel for data management and analysis, main- tenance planning and scheduling, maintaining an inventory when considering the significant investment in launch of spare parts, and life-cycle management. These personnel consumables. are in addition to the field technicians responsible for sup- 3. Carefully weigh auto launch versus manual launch porting O&M of the weather radar network. Roles for support systems. When planning for an upper-air network, the personnel include but are not limited to: NMS should consider whether manual launching sys- tems, automated launching systems, or a combination ■ Specialists/scientists to analyze, interpret, and apply radar of the two technologies will best suit its needs and bud- information get in light of labor, staff retention, and ongoing training ■ IT specialists to support data ingestion, quality assurance/ costs. The NMS should also consider security of the site, quality control (QA/QC) functions, and storage and flow of given that manual stations are staffed twice daily ver- data and products to forecast models sus an automated station that is visited once every two ■ Network management and planning—including service, to four weeks. The NMS should discuss the benefits and incident, change and process improvement management, drawbacks of each system type with upper-air system and life-cycle support manufacturers, an NMSs with significant experience in Upper-Air Systems   251 upper-air network operation, or entities with compara- 3.6.8 References ble expertise. 4. Stay on top of replacement needs. Given the depen- WMO (World Meteorological Organization). 2018a. Guide to dance of upper-air observations on consumables, re- Instruments and Methods of Observation, Volume I: Measure- stocking and procurement contracts are critical and ment of Meteorological Variables. WMO-No. 8. Geneva: WMO. should be carefully maintained through the life cycle https://library.wmo.int/doc_num.php?explnum_id=10616. of the upper-air system or network. This requires long- term budgeting for related costs, careful monitoring of WMO (World Meteorological Organization). 2018b. Guide to consumable inventories, and timely ordering of stock to Instruments and Methods of Observation, Volume III: Observing ensure uninterrupted operations. Systems. WMO-No. 8. Geneva: WMO. https://library.wmo.int/ 5. Use FTEs for personnel costs. The number of field tech- doc_num.php?explnum_id=9872. nicians and support staff required for reliable network operation should be determined through consultations WMO (World Meteorological Organization). 2019. Manual on with NMSs with significant experience operating a the WMO Integrated Global Observing System. WMO-No. 1160. similar upper-air network, or entities with comparable Geneva: WMO. expertise. 6. Consider outsourcing O&M. NMSs that do not have expe- WMO OSCAR (World Meteorological Organization Observing rience with O&M for upper-air systems or networks, are Systems Capability Analysis and Review Tool. https://www. planning for smaller networks, or anticipate challenges wmo-sat.info/oscar/. in recruiting or retaining qualified personnel should consider contracting the O&M functions of the network to an experienced private sector partner. This approach might also result in higher uptime performance in cases where the NMS does not have experience operating Photo courtesy of Vladimir Tsirkunov upper-air systems, providing greater return on investment. 7. Do long-term upgrade planning. Key to the ongoing success of an upper-air network is long-term planning for upgrades to computer equipment, software, and analysis of the incoming data. Upper-air systems are expensive to purchase, install, and maintain, and they have substantial annual operating costs given the con- sumable costs of radiosondes, balloons, parachutes, and lifting gas. Careful planning is required to ensure that manufacturer updates and enhancements are installed and incorporated. 252    3.7 Weather Radar Systems The Purpose of Weather Radar Systems 3.7.1 Over the past 30 years, the need for weather radars has grown significantly for a variety of reasons. National Meteorological Services (NMSs) have adopted the use of weather radars primarily to support precipitation forecasting, pub- lic warnings of severe weather events, ground truthing of satellite information, and more. While radar information is now beginning to be incorporated into numerical weather prediction (NWP) models to improve weather predictions, NMSs primarily use radars for nowcasting applications and early warnings for rainfall intensity, flash flooding events, and dangerous thunderstorms. In colder climates, radars are also important for distinguishing between rain and snow events. Radars are also capable of assessing, monitoring, and predicting char- acteristics of wind and shear zones. This is particularly useful for determining Photo: © Anthro | istock.com the intensity and severity of thunderstorms and supercells, and for predicting maximum wind speeds, gusts, turbulence, and shear. Radar information is not routinely shared regionally or globally; in fact, that is done only under national or regional agreements (for example, the OPERA program of EUMETNET, a grouping of 31 European National Meteorological Services). However, the June 2021 World Meteorological Organization (WMO) “Over the past 30 years, Unified Data Policy (Resolution 1, Cg18.5) (WMO 2021b) has broadened the application of WMO policy for the international exchange of Earth system data the need for weather radars and created a more open environment for the free and unrestricted exchange has grown significantly of data. So far, there are no increased obligations regarding which data shall be shared, but that may change as technical regulations are successively re- for a variety of reasons. viewed to better reflect global modeling requirements and better serve WMO Members and the associated global community. National Meteorological Services (NMSs) have Due to its unique capabilities, weather radar—as illustrated in photo 3.7.1— can form a significant component of a modern weather observation infra- adopted the use of weather structure. WMO’s Guide to Instruments and Methods of Observation (WMO-No. radars primarily to support 8), Section III, Chapter 7 (WMO 2018) describes weather radars in the follow- ing manner: precipitation forecasting, public warnings of severe Radars provide localized, highly detailed, timely and three-dimensional sensing and observing capability that no other meteorological monitor- weather events, ground ing system can provide. They can measure variations in precipitation rates at a resolution of a few square kilometers or better and at time truthing of satellite cycles of the order of a few minutes. They provide the capability to mon- information, and more.” itor rapidly evolving weather events, which is critical for the provision Weather Radar Systems   253 PHOTO 3.7.1 NEXRAD Located at the National Weather Center (NWC) in Norman, Oklahoma Source: Wikipedia. of early warnings of severe and hazardous weather. This weather radars that an NMS should take into consideration includes heavy rain, hail, strong winds (for example, tor- are: nadoes and tropical cyclones) and wind shear; the con- ditions that have the highest impact on society of all the ■ Surveillance of synoptic and mesoscale weather systems weather elements. ■ Severe weather detection, tracking, and warning (includ- ing severe wind hazard detection) The initial capital purchase, site preparation, and ongoing ■ Nowcasting operations and maintenance (O&M) make radar systems a ■ Estimation of precipitation intensity, echo classification costly investment for any NMS. The data they provide, while ■ Wind profiling and wind mapping. highly valuable, are complex and require significant educa- tion and training to interpret. An NMS should consider how to Radar systems are costly to acquire, operate, and main- best use this asset to its fullest capabilities prior to acquiring tain. While their life cycle can span 15 years and beyond for to ensure proper implementation, along with operational and well-maintained systems, most NMSs in lower-income coun- resource planning. The main meteorological applications of tries lack the security of having adequate budget and technical 254    Weather Radar Systems resources to support them over this lifetime. Understanding The most common approach is to specify the technical pa- the full financial and human resource implications through rameters of the system, which is often combined with spec- detailed planning from acquisition through to operation is ifying key features. However, it is important to keep in mind key to capitalize on the advantages that a radar system(s) can that the overall performance of the system is quite difficult provide. This chapter will guide the NMS through some of the to determine from a set of technical parameters. System set- key points to consider when building a sustainable weather tings depend on the particular application of the radar and radar network. can have a significant effect on performance. Emphasizing performance for one application may decrease performance 3.7.2 Weather Radar Observations in another, although some tradeoffs are often inevitable. Observational specifications for a radar network should be Perhaps the best approach to defining radar specifications is developed from considerations of local weather hazards to use specifications based on performance in the selected (such as intensity of precipitation) and the locations to be key applications and use cases, appended with specifications monitored (such as city, airport, or river basin). It is then crit- for significant features, although this requires a thorough un- ical to determine the applications and services that the radar derstanding of radar technology and applications. network is expected to serve (such as thunderstorm warnings, wind shear alerts, hydrology, climate), given that a radar op- It is usually not advisable to specify a radar system that dif- timized for one application might not perform well in other fers from standard industry products, especially in hardware applications. features. Custom radars can be manufactured, but the cost is likely to be higher than the cost of standard products, and The WMO Guide (WMO 0218) contains a wealth of informa- support for such radars may be more difficult to obtain during tion on radar technology and performance. Additionally, a the long life cycle of a radar system. The risk for technical common ISO/WMO standard ISO 19926 (ISO 2019) pro- problems is also greater, and custom radar systems might vides clear definitions for radar parameters as well as highly lack approvals for electrical safety and/or electromagnetic useful guidance on radar calibration and maintenance. The compatibility, for example. standard is also reproduced as Annex 7.A in the WMO Guide (WMO 2018). At the time of writing, WMO is developing an For defining requirement specifications, it is highly recom- Operational Weather Radar Best Practices Guidance for weath- mended to study the WMO Guide (WMO 2018) and the draft er radars (WMO 2021a), which aims to provide practical guid- Best Practices Guidance (WMO 2021a), and—most critically— ance on network design, radar procurement, radar technology to employ an expert in weather radar systems. It is also advis- and data exchange. A partial draft is publicly available. able that after-sales support and maintainance be included as a mandatory component of the radar system specifications, Weather radar specifications are based on the definition given its critical role in sustainablility. of three components, each of which provides a distinctive approach: WMO Guide 8 (WMO 2018) provides an example of radar specifications in Attachment B Sample Radar Specifications. ■ Features are defined as the functionality of the system. This example, shown in table 3.7.1, is a limited set of spec- ■ Performance is defined as the measurement of performance. ifications that defines a number of technical parameters but ■ Technical considerations are defined as the technical pa- no features. rameters of the system. Weather Radar Systems   255 TABLE 3.7.1  Common Specifications of Weather Radar System performance requirements for weather radar Radar operating frequency Radar 1 Radar 2 Radar 3 Fundamental parameters Criteria (S-band) (C-band) (X-band) Sensitivitya Minimum detectable reflectivity must be A dBz or less at < 18; 240 < 13; 120 < 8; 60 a distance up to B km, where max unambiguous velocity of more than ±48m/s is attained with 2-stagger fPRF of either 2:3 or 3:4 or 4:5 Spatial resolution Beam resolution must be ƟH and ƟV (in degrees) or less <1 <1 <1   Range resolution must be ∆R (in m) or less ≤ 150 ≤ 150 ≤ 150   Antenna side lobe must be ∆Vpa (in dB) or less < –25 < –25 < –25   Range side lobe must be ∆Vpr (in dB) or less for < –50 < –50 < –50 pulse-compression radar Phase stability Phase stability should be Ɵps (in degrees) or less 0.3 0.6 1.2 Accuracy of dual- Cross polarization radio must be XPDsys (in dB) or less < –30 < –30 < –30 polarization measurement Radar 1 Radar 2 Radar 3 Other key parameters Criteria (S–band) (C–band) (X–band) Maximum rotation speed Antenna maximum rotation speed must be Rmax (in rpm) ≥6 ≥6 ≥6 or more Acceleration As EL antenna acceleration, EL drive time from 0 to 90 < 20 < 20 < 20 (in degrees), and 90 to 0 (in degrees) must be less than taEL (in sec)   As AZ antenna acceleration, time from maximum speed <5 <5 <5 to complete stop must be less than taAZ (in sec) Antenna pointing Antenna pointing accuracy must be ƟAZ (in degrees) or < 0.1 < 0.1 < 0.1 accuracy less and ƟEL (in degrees) or less Dynamic range Dynamic range must be LVd (in dB) or more > 100 > 100 > 100 Unwanted emissions b The level of unwanted emissions must be A dB or less at < –60; 5 < –60; 5 < –60; 5 B MHz away from the central frequency f0 (in MHz) Source: WMO 2018. Note: a. For typical pulse-compression radar the minimum detectable reflectivity at short ranges cannot be calculated from the given values. b. National requirement might request different values. MHz = megahertz. The choice of radar frequency band will have a significant used radars in the world, providing a long measurement range impact on procurement and life-cycle costs. The higher at a more affordable cost than the S-band. the operating frequency, the smaller the radar can be, and the lower the cost. However, high-frequency signals in the Typical output data for a modern dual-polarization weather X-band can experience significant attenuation in heavy pre- radar are as follows: cipitation, which limits the measurement range of X-band radars and, consequently, their applications. S-band radars ■ Reflectivity factor (Z, or Zh for horizontal polarization and were traditionally used in the tropics, as signal attenuation Zv for vertical polarization) due to heavy precipitation could not be corrected in single- ■ Differential reflectivity (Zdr ) polarization radars. With the introduction of dual-polariza- ■ Doppler velocity (V) tion technology, C-band radars can now effectively correct ■ Spectrum width (W) for signal attenuation. C-band radars are the most commonly 256    Weather Radar Systems ■ Differential phase (Фdp) and specific differential phase typically available at the IPS or in multiple workstations re- (Kdp) motely connected to the IPS. ■ Correlation coefficient between Zh and Zv (ρhv). Weather radars are highly specialized systems and can be These primary variables (also called data moments) are then configured to provide insight into a wide range of weather processed into a number of derived radar products, which phenomena. Because of this, a detailed discussion of the ca- are disseminated and displayed to users. Examples of typical pabilities and limitations of the various radar bands is be- radar products include precipitation intensity, precipitation yond the scope of this document (see WMO 2018, Section III, accumulation, hydrometeor classification, wind field, wind Chapter 7 for more information). It should be noted that a shear, and echo tops. (See WMO 2018, Chapter 7, for a more radar system configured for optimal performance in one ap- detailed list of meteorological products.) plication (such as measuring precipitation intensity) might not perform well in another application (such as wind map- The chain of radar data processing can be divided into four ping); similarly, each radar must be set up in accordance with levels: local conditions. ■ Level 1. Data in sensor units are also known as time se- As a result, the NMS needs to have a clear understanding of the monitoring purpose(s) that the weather radar will fulfill if ries or I/Q (in-phase and quadrature) data. Produced and the system is to be properly specified and configured. Given processed by the instrument’s signal processor, such data the complexity of weather radar systems, the level of exper- are generally not recorded except for limited durations on tise required to install, configure, and maintain a weather operational radars and used for research purposes. radar is very high. Thus, if an NMS is contemplating the pur- ■ Level 2. Derived radar variables or moments (for example, chase of a weather radar system but does not have the nec- reflectivity, radial velocity, differential reflectivity) are at essary in-house expertise to configure the system, it is highly full resolution after aggregation and filtering. They are or- recommended that it contract with a consultant. ganized in polar coordinates by rays, range bins, and quan- tities. They are also known as sweep and volume scan data. When specifying a weather radar system or network, the NMS ■ Level 3. These are radar products derived primarily from should be prepared to answer the following questions: Level 2 data. They may be in the Level 2 polar coordinates or in other coordinate systems, such as vertical profiles ■ What meteorological problem or problems will the weath- or Cartesian grids. Examples of Level 3 products: con- er radar system or network address? stant altitude plan position indicator, or CAPPI; rain rate p General weather surveillance? estimates. p Specific severe weather detection and warning (such ■ Level 4. Higher-order products may include data from mul- as flood warning)? tiple radars or variables from other observation systems p Nowcasting (like for an airport)? (such as lightning location systems, rain gauges). p Precipitation measurement? p Wind mapping or profiling? In a typical system setup, Level 1 and 2 processing is done ■ What high-priority areas will be served by the weather at the radar station, while product generation (Levels 3 and radar system or network? 4) is done at a central information processing system (IPS) p Densely populated areas? location. Product generation can also be implemented at the p Airports? radar station (for redundancy or single radar installations, p Major agricultural areas? for example). Data visualization and analysis software are p Flood-prone areas? Weather Radar Systems   257 3.7.3 Determining Where to Place Individual FIGURE 3.7.1  Beam Height as a Function of Range Stations in the Network across the Landscape 20 Weather radars can be configured and used collectively as a 15 network. Such networks can, in turn, be used in cooperation Convective Elevation and collaboration with the networks of partner organizations 2.5º Height (km) and agencies, countries, and regions (that is, multiple coun- 10 tries), provided that a sufficient standardization of operation 1.5º is agreed upon and deployed. In general, operational weather radars are sparsely distributed because of the high cost of 5 Stratiform their purchase, installation, and maintenance. In some cases, 0.5º an NMS will install several weather radar systems but fail to network them; this lowers the value of the data produced by 0 Low level snow 0 50 100 150 200 250 the individual systems. It is highly desirable to develop a Range (km) network that provides uniform and optimal coverage, partic- ularly over those areas for which data provision and weather Source: WMO 2021a. monitoring applications are required. Note: Each color arc is marked with an angle that relates to three beam elevation angles (0.5°, 1.5°, and 2.5°) for a radar beam of width 1°. The radar Due to the curvature of the Earth, a radar is unable to de- beam elevation angle is the angle at which the antenna is pointed up from horizon. The horizontal color bands marked with “Convective,” “Stratiform,” tect weather near the surface, where accurate information and “Low level snow” relate to different types of weather that can produce is needed. Depending on the application, data aloft (several precipitation (rain, snow, and other). Convective weather (thunderstorms) may reach high altitudes, while stratiform weather stays at lower altitudes, kilometers above the ground) can be extrapolated down to and some types of weather such as low-level snow exist at low altitudes. the surface. However, extrapolation might not be adequate Convective weather can be detected at long distances while low stratiform for applications where data accuracy requirements are more weather will be challenging and low-level snow impossible to detect over long distances. This is related to the concept of deciding the distance stringent. Hence, the network design is critical to the effica- between radars based on local climate: what types of weather are important cy of the warning service. (WMO 2021a, the Best Practices in your region / country, what do you need to detect reliably? Guidance Part A, Annex D, provides a brief overview of the key considerations in designing radar networks.) elevation angle of 0.5 degrees is normally the lowest usable For national radar networks, uniform radar coverage of the angle for radars with a beam width of 1.0 degrees. country is often the ultimate goal. But for practical purposes, priority is typically given to densely populated or other criti- This issue of radar station spacing is applicable to radars op- cal areas (such as major agricultural or flood-prone areas, or erating in the C- or S-band with a 1° beam width. These radars significant airports). Radar density in the network is strongly are typically built with capabilities to measure long ranges, dependent on the radar applications and the local climate. up to 300 to 400 kilometers. Due to the high attenuation in If shallow precipitation needs to be detected reliably, or the the X-band frequencies, radars operating in the X-band are quantity of precipitation measured accurately, radars should restricted to shorter ranges. High performance X-band ra- be located less than 200 kilometers apart. On the other hand, dars can be expected to provide measurement ranges reach- detection of convective storms is possible with radar spacing ing about 60 to 100 kilometers, depending on local climate of up to 400 kilometers. Many national radar networks find (intensity of typical precipitation). Here radar attenuation a compromise between these two parameters and are built is simply the absorption or reflection of radar signals as the with spacing between 200 kilometers and 400 kilometers. radar pulse penetrates an area of precipitation—preventing Figure 3.7.1 gives a good visual indication of the limitations that radar from detecting any additional cells that lie behind of measurement caused by the Earth’s curvature. A beam the first storm. These radars are typically used as gap fillers in 258    Weather Radar Systems national networks, or as dedicated radars for a specific appli- conditions, and cost. Based on this analysis, the final site lo- cations and users (such as at airports). Gap filling with X-band cation might differ by several kilometers from the ideal when radars can be a good strategy—for example, in the case where the following factors are considered: mountains would block the observations of longer-range S- or C-band radars. ■ Logistics. The remoteness of a site can add significantly to the operational costs of a network. The WMO Guide 8 (WMO 2018) gives the following guidance p Is the site easily accessible by road? on radar locations, siting a study by Leone et al. (1989): p Is there an alternative location that is less expensive to travel to that would provide the same or similar Optimum site selection for installing a weather radar data that supports the end use, reducing logistical depends on the intended use. When there is a definite challenges and costs? Time spent traveling to sites zone that requires storm warnings, the best compromise increases the human resources required by the NMS. is usually to locate the equipment at a distance of be- p What are the additional costs for the civil works for tween 20 and 50 km from the area of interest, and gen- supporting the station given the location of the site? erally upwind of it according to the main storm track. It ■ Security. Is the area safe enough that theft and vandalism is recommended that the radar be installed slightly away would not be significant issues? Is the site in a location from the main storm track in order to avoid measurement safe for NMS personnel servicing the station? problems when the storms pass over the radar. At the ■ Land availability. Does the government own the land same time, this should lead to good resolution over the where the station is to be sited? Will the government need area of interest and permit better advance warning of the to purchase (resulting in a capital cost) or lease the land coming storms (Leone et al. 1989). (resulting in an annual budgetary cost)? ■ Communications. Given the volume of data that weather Suitable radar locations are strongly constrained by the re- radar systems produce, the most reliable and cost-effective quirement to provide an unobstructed view to the radar. form of data communication is a direct connection to the Radar measurement can be blocked by hills or mountains, or internet. by nearby tall buildings. Especially in mountainous areas, it p Is there infrastructure in place for internet communi- is often not possible to find a location providing a fully un- cations? If not, what is the cost of bringing the service obstructed view at the lowest antenna pointing angle, which to the site? would allow the longest usable measurement range. Radar p Modern cellular data communication may be an op- locations in such areas should be selected by considering the coverage of the whole network; if an area is blocked to one tion, but this uses high frequencies that suffer from radar, adding another radar into the network might be able attenuation in heavy precipitation events and data to provide coverage to that area. Coverage of critical areas communication may deteriorate at a time when the should be prioritized. Radar beam blocking by terrain can be data would be most valuable. As a result, cellular com- simulated with suitable tools, using digital elevation data. munication is not recommended. However, if cellular Such simulations are highly useful in selecting the best loca- is the preferred communications method, is there reli- tions for the radar stations. able coverage in the area? What is the monthly cost of the communications plan based on the volume of data When selecting installation sites, there are several factors to be transmitted? Costly plans in an area may sug- to consider when locating individual stations. Each site se- gest the need for alternative methods of transmitting lected will need to be balanced against all of these factors. the data. The cellular signal strength at the site should Each of these locations should be evaluated based on siting be measured either by checking the bars on the phone requirements specified by the supplier(s); advice from ex- or by using a signal power meter. perts; and considerations related to the site’s logistics, secu- p Is the communication connection reliable, is data rity, land availability, communications, power, environmental throughput stable or are there breaks in the connection? Weather Radar Systems   259 ■ Power. Weather radar stations require reliable and sta- ■ Safety. People should not be directly exposed to the radar ble mains power, as solar panels cannot produce enough beam within a safe distance specified for the radar. energy to power the systems. When power is not stable, ■ Location of other observational systems or networks. the radar should be supported with an onsite generator to This should be considered to maximize the benefit of ensure operation remains continuous maximizing uptime. the weather radar system(s). Although it can complicate p Does the site have mains power? If not, can it be planning, it can bring significant benefits and efficiencies brought to the site, and if so, at what cost? with respect to operational costs (including infrastructure, p Is mains power reliable, how long breaks in power are telecommunications, personnel). (See the network design expected? principle recommendation in WMO 2018a.) ■ Environmental conditions. Environmental conditions at the site play a large role in how often the site will require All of the above factors will need to be weighed and balanced maintenance and repair. against the cost of operations and spatial density require- p Is the location free of electromagnetic noise that ments of the data. In the end, the ideal site may still be the will interfere with operation of the weather radar? best location to support the data needs. By considering the Because some types of interference are intermittent, above factors, an NMS should begin to understand the cost an electromagnetic survey of the site should be un- implications of their decisions. dertaken over a period of time (at least one week) to ensure that electromagnetic noise is not an issue. 3.7.3.1 Configuring the Weather Radar System p Does the radar have an unobstructed view of the sky in all directions? Are there trees, buildings, or topo- A typical weather radar station contains a radar tower and a graphical features that will block the radar signal? room for electronic equipment, as shown in figure 3.7.2. The p Are there nearby bodies of fresh or (in particular) salt radar antenna system is installed in the radar tower and pro- water that can increase corrosion rates and require tected with a radome. The local climate should be weighed to the site to undergo more frequent maintenance or determine the maximum wind speed tolerance of the radome replacement? and tower, as well as requirements for lightning protection. FIGURE 3.7.2  Layout and Setup of a Typical Weather Radar Installation Source: P. Utela, Vaisala. Note: (1) = radome; (2) = radome tower; (3) = equipment room; (4) = security perimeter fence; (5) = access road. 260    Weather Radar Systems Radar systems include the following components: ■ Transmission Control Protocol/Internet Protocol (TCP/ IP) communication. This is the standard communication ■ Radar antenna systems. These are typically rated to op- method used by modern weather radars. The ISO standard erate in a wide range of temperature and humidity condi- (ISO 2019) recommends 8 megabits/second (Mbit/s) for tions. However, the radome may need to be equipped with dual-polarization radars. Lower communication speeds forced ventilation, heaters, air conditioning, or a dehumid- can be used, but full capabilities of a modern dual- ifier, depending on local weather extremes. polarization radar may be restricted by the communica- ■ Radar electronics and other auxiliary systems—such as tion speed. an uninterrupted power supply (UPS) and telecommuni- cation devices. These are installed in the equipment room, 3.7.4 Configuring the Weather Radar Network which should be weatherproof and provide controlled tem- perature conditions. Depending on the local climate, it will Configuring the weather radar network needs to take into require heating, cooling, or humidity control. account six categories of requirements: operational require- ■ Security protection. This is needed to protect the station ments, information technology (IT), operational personnel, against unauthorized access. It is provided either by a maintenance, upgrades, and spare parts. These are consid- perimeter fence or locked access doors to the equipment ered in detail in this section. room and tower. Other security measures (such as camera surveillance) might be necessary. 3.7.4.1 Operational Requirements ■ Accessibility. The radar station needs to be accessible by Weather radar represents arguably the most complex instru- road that is suitable for a mobile crane large enough to lift ment that a meteorological service or service provider must the radar and radome on the top of the tower. Note that a service and maintain (WMO Guide 8, 2018). It requires a very radar system can also be installed by helicopter if the sta- high level of training and skill development. Maintenance is tion is built at a location where road construction would be critical to keeping the radar operating, and calibration is crit- prohibitively expensive. Nevertheless, some means to ac- ical to the quality of the data. According to the ISO standard cess the radar will be needed for future maintenance visits. (ISO 2019), competent and regular maintenance should re- ■ A flat surface. This is needed close to the tower to facili- sult in radar availability greater than 90 percent of the time tate the assembly of the radar antenna system and radome, on a yearly basis, with standard weather radar equipment. with enough space for the crane. (General guidance on the requirements for operating a radar ■ A radar tower. This is typically the tallest feature in the sur- network is given in the draft Best Practices Guide and the ISO rounding area and thus prone to attract lightning strikes. standard of WMO-No. 8.) The radome should be equipped with proper lightning pro- tection, following the latest guidance (such as FAA-STD- The NMS will need to answer the following key questions 019E, or later—see EverySpec, no date). when considering the operation of the weather radar network ■ Sufficient electrical power. The radar should be equipped and to work through the total cost of ownership (TCO) (civil with a UPS to ensure continuous observations in the case works and annual costs) exercise. of short power outages. If reliable power cannot be provid- ed, the station might need to be equipped with an auto- ■ Weather radar systems have different levels of routine main- matic backup generator. tenance requirements, frequency band, type of radome, and ■ Reliable communications. It is critical to have a reliable station infrastructure. The NMS will need to assess its own link from the radar station to the data center. An electrical in-house expertise, and decide: What is the recommended or optical communication cable is preferred over wireless level of maintenance it can perform? Will it be performed communication methods, as radio frequency signals suf- by NMS personnel or covered by maintenance contracts, fer from attenuation in heavy precipitation. In the worst or both? The NMS should investigate the various types of case, real-time radar data could be lost at a time when they service-level agreements and costs and select the one that would be most valuable. matches the NMS’s level of in-house expertise. Weather Radar Systems   261 ■ What should the NMS’s spare parts strategy be? Purchase archiving should follow established protocols. As a general rec- and own a comprehensive stock of spare parts, or purchase ommendation for dedicated radar-related personnel, the ISO a limited stock of spare parts and hold a contract with the standard states that the ratio of full-time radar specialists to radar supplier? radar stations should be greater than 1:1 (ISO 2019). ■ How will station safety be ensured? Will there be security fencing, locked buildings, surveillance, or security guards? Ideally, the organization operating a radar network should ■ How will the network be monitored, through general ob- have one or more radar specialists (staff with expertise in servation network monitoring or specific radar network radar measurement). The role of a radar specialist is to man- monitoring? age the settings of the radars (such as adjusting settings re- ■ How will operations of the IPS be organized? lated to data filtering or radar scan configurations) and to p On-premises or cloud-based? provide second-line support to more complex issues with the p Are there recurring software licensing fees? operations and measurement performance. Suitable educa- p How often are upgrades issued and is there a cost? tional backgrounds for the radar specialist role include me- p What personnel does the NMS have to support the teorology or radar technology. System-specific training from IPS? the radar manufacturer is also necessary and is usually pro- ■ How will radar performance be optimized to local condi- vided as a component of radar project delivery. tions after radar delivery? Will optimization be part of the contracted services from the vendor? Radar data are mostly viewed and used by forecasters, who ■ How often will NMS personnel require refresher training? will need training to interpret radar observations. Unless the forecasters have received this training as a part of their 3.7.4.2 IT Requirements basic meteorological education, they should receive dedicat- ed training in radar meteorology. Forecaster competence in Data from the radar network is collected at a central IPS, typ- radar measurements should also be kept up to date with suit- ically operated by the NMS. A dedicated computer or comput- able refresher training courses. ers should have licensed software that is configured to ingest data from the radar (or radars), archive data, generate radar 3.7.4.4 Maintenance products (levels 3 and 4) and to provide tools for analyzing Station maintenance is often organized in two levels: and visualizing the products. These computers and related telecommunication equipment should be connected to a UPS. First level: Basic service capability. This includes station The IPS is a fundamental component of the radar system, and maintenance (air conditioning, backup generator, infrastruc- as such should have appropriate service-level agreements, ture) and possibly visual inspection of radar and radome or technical support (which might require different expertise to simple repairs. First-level maintenance requires little radar-spe- that of the weather radar), and redundancy to ensure 24/7 op- cific training and can usually be carried out by personnel with erations. Additionally, the radar network can be equipped with a wide range of general maintenance duties (such as with other a number of workstations for data analysis and display. These weather observation systems). It might also be possible to out- computers are typically installed in the offices of data users. source these duties to a suitable, usually local company. 3.7.4.3 Operational Personnel Second level: Radar technician/engineer. This includes an- Radar network operation should be monitored, just as are other nual or bi-annual calibration and radar troubleshooting and observation systems. Often, the same personnel can monitor repair. Second-level maintenance typically requires more in- several different observation systems from a central location, depth understanding of radar operation and training in radio possibly including all national meteorological observation in- frequency electronics; these activities can only be outsourced frastructure. Radar system–specific training should be provid- to a company with some expertise in radar technology, poten- ed to the personnel responsible for monitoring the network. tially to the radar manufacturer. The maintenance personnel Personnel trained in IT are needed for data center O&M. Data can also manage an inventory of local spare parts. For very 262    Weather Radar Systems large radar networks, it may be feasible to maintain a dedi- acquiring a comprehensive set of spares is to acquire only a cated national depot with a comprehensive set of spare parts limited set of critical parts and to enter into a service contract and highly trained personnel capable of more complex repair with the radar manufacturer, so that the non-spared parts can operations. More often, the radar manufacturer acts as the be quickly obtained. service depot. It is crucial to maintain the competence of all maintenance personnel; additional training by the radar man- 3.7.5 Estimating the TCO of Weather Radar ufacturer may occasionally be necessary (possibly at a local Networks radar station, service center, or data center). How do NMS decisions impact the TCO over a 15-year period? A suggested schedule for preventive station maintenance vis- The reality is that money used to purchase a system that is its would be: (1) two to six first-level visits per year; and (2) too large or too complex will be wasted if the annual budget two second-level visits per year, in between first-level visits. is too small to support the network. To aid in the costing pro- This schedule can be modified based on operational expe- cess, the NMS is directed to carry out the TCO Exercise. (To rience—for example, depending on the general condition of calculate the TCO exercise, see chapter 3.8, using the addi- station infrastructure and age of radar equipment. Annual (or tional information provided below that is specific to weath- more frequent) visits are recommended to ensure proper cali- er radar as an aid.) This exercise should be completed even bration, but less frequent visits by second-level experts might when the radar is donated to a developing country. Both the be acceptable in cases where the radar system is stable and NMS and the international financial institution (IFI) or donat- offers advanced remote monitoring and control. ing country should understand the lifetime operational costs. Based on the TCO, the donor country or IFI are encouraged to Unscheduled maintenance visits might be needed to correct consider longer-term support if both the donor and recipient faults at the station and to replace any worn-out parts. Because NMS are to realize the full value of weather radar. of the complexity of a radar and station infrastructure, such visits may be necessary once or more often every year, and 3.7.5.1 Procurement Costs travel costs should be part of estimated maintenance costs. Successful procurement and implementation of a radar proj- ect will require resources from the organization acquiring the 3.7.4.5 Upgrades radar. (The Best Practices Guidance in WMO 2021a contains Weather radar systems, if well maintained, typically have more detailed guidance for implementing a radar program.) long life cycles of more than 15 years. It is advisable to plan One key resource is a program or project manager, who should for a more extensive service operation or upgrade 5–10 years act as the main contact point for the radar supplier’s project after installation of a radar system. Such upgrades often in- manager. Additional subproject manager or managers may be clude selected mechanical parts and computer hardware and necessary, for example, to oversee land acquisition and con- software. Selected electronics can also be replaced, based on struction work when a completely new radar station is built. analysis of maintenance history. The purchasing organization should have sufficient expertise in weather radars to oversee installation and monitor accep- 3.7.4.6 Spare Parts tance tests. Moreover, future maintenance personnel should Spares should be acquired for the most sensitive and difficult- be available for training by the radar supplier. to-source components (such as power tubes or solid-state components, boards, chassis, motors, gears, and power sup- Weather radars are typically purchased as turn-key projects. plies). The number of operational spares should be driven The common scope of such a project includes the following by the mean time between failure of the unit and the time items, which should form a basis of tender procurement to acquire replacement items from the manufacturer; it also documents: radar, radome, data center with software, test depends on the number of radars in the national network. equipment, selected spare parts, relevant services, Factory The spare part requirement or policy should be based on the Acceptance Test, installation and commissioning, site ac- manufacturer’s recommendations. An alternative strategy to ceptance test, and user and maintenance training. The radar Weather Radar Systems   263 tower and a suitable equipment room might also be included ■ How will spares be budgeted? Given both the importance when there is a completely new radar site. As a general rule, and high cost of spares, it is worth considering including all civil works related to building a new radar station or refur- a long-term service-level agreement with the manufacturer bishing an old station should be contracted locally, preferably as part of the procurement cost. This transfers the costs of under a separate contract. spares from the ongoing operational budget (which can be difficult to maintain) to part of the capital investment cost. The total cost of a turn-key project is largely dependent on the ■ What personnel are required for the operations and main- contract’s scope (such as the type of radar, extent of services, tenance of the weather radar network? spare parts). However, table 3.7.2 gives typical expected ■ How much training of personnel will be needed, and who prices for radar turn-key projects, including the items listed will participate in training? above but excluding land, equipment room, fence (or other p Maintenance personnel? security measures), radar tower, spares, telecommunication p Radar specialists? and electricity lines, potential backup generator, all related p Scientists? civil works, and the data center. It shows that these costs can p Forecasters? range from $850,000 for a magnetron X-band to $3.2 million- for a klystron or solid-state S-band—not including the cost of 3.7.5.2 Operational Costs land installation and or spare components. The main operational costs, excluding personnel, are relat- ed to electricity, telecommunications, maintenance, updates, TABLE 3.7.2  Estimated Procurement Costs for Single X-Band, and spares. The total annual O&M cost for a radar network C-Band, and S-Band Weather Radar Systems with 1.0° Beam can be estimated at about 6–8 percent of the total procure- Widths ment cost of all stations in the network, not including person- X-band 1.0° C-band 1.0° S-band 1.0° nel costs. Operational costs include: System type (US$) (US$) (US$) Turn-key $850,000 $1,900,000 $2,900,000 ■ Electrical supply to the radar station, including: (magnetron) p Radar system Turn-key n.a. $2,100,000 $3,200,000 p Auxiliary systems (security, telecommunications) (klystron) p Ventilation, heating or cooling Turn-key (solid $900,000 $2,100,000 $3,200,000 p Radio frequency licenses state) ■ Cost of telecommunication link to the data center Source: Based on observations from Vaisala from pre-2021 projects. ■ Cost of maintenance of radar station infrastructure, equip- Note: These indicative estimates are based on competitive prices in years 2020–21 and do not include spare components. n.a. = not applicable. ment room, tower, radome, fence, and so on ■ Upgrades ■ Spares (if not included in a service-level agreement). The data center cost will depend on the amount of computer hardware and number of software licenses. The price for a Operational personnel. Once the personnel required for typical basic setup can be estimated at $120,000–$150,000; sustained operation of the weather radar network have been this is on top of costs related to the individual weather radar identified, associated costs (such as salaries, benefits, travel, systems. training) can be estimated. It is advisable to request budgetary quotations from potential Maintenance. Station maintenance is often organized into two radar suppliers to get up-to-date cost information for plan- levels: the first level (performed by personnel with basic ser- ning the budget of a radar project. Key questions to answer vice capability) and second level (performed by a radar tech- when considering procurement costs include: nician/engineer). It is recommended that between two and six first-level visits and two second-level visits be performed each year. Routine maintenance costs include consumables. 264    Weather Radar Systems The only consumable with a significant cost is the transmit- S-band at the time of writing, resulting in a low cost of an- ter tube, used in magnetron or klystron-based radars. Modern nual O&M cost per radar, as equipment is still under man- magnetrons can be expected to operate for more than 3 years, ufacturer’s warranty. generally up to 5–7 years, while klystron tubes typically last ■ Germany operates 18 sites. Its radars are C-Band. longer, from 8 to 12 years. ■ The United Kingdom operates 15 sites. Its network is C-band. Upgrades. Weather radar systems typically have long oper- ational lifespans, which often means that significant techno- Among developed nations, a performance benchmark of 95 logical advancements are made during a system’s life cycle. percent uptime is typical, with uptimes of 95 percent or high- Budgeting for system upgrades at a point 8–12 years after er commonly achieved. At the time of writing, one of Austria’s installation should be included in the initial life-cycle man- five radar systems was inoperable in 2020; hence its 80 per- agement plan. cent uptime performance rating. Upon completion of repairs, it is expected that Austria’s radar network performance will Spare parts. The ISO standard states that, based on experi- return to about the performance benchmark. ence, it is desirable to acquire spare parts for up to 30 percent of the initial cost of the radar (ISO 2019). This percentage is Developing nations should set an initial performance target of the cost of the radar equipment, rather than of the whole of 90 percent, because uptimes less than 90 percent prevent station. If there are several similar radars in the network, this the true value and benefit of weather radar systems from cost can be lower, as spare parts can be stored at a central being realized. Once the NMS has gained experience oper- location and shared between radars. Another strategy, more ating and maintaining its radar systems, the target should be common in the case of small networks, is to acquire only a increased to 95 percent or better. At this point, as table 3.7.3 limited set of critical spare parts and to sign a service-level shows, the benchmark performance for upper-middle-income agreement with the radar manufacturer for swift delivery of countries is 87 percent where of 4 out of 6 countries report- spare parts from the factory. ed 90 percent or more uptime, 2 showed a value of 80 per- cent or lower. Moreover, as table 3.7.4 shows, 47 percent of 3.7.5.3 Weather Radar Operating, Maintenance, and low- and upper-middle-income countries (noted earlier) and Life-Cycle Cost Examples high-income developing countries are not achieving the min- Weather radar systems are costly to operate, especially when imum 90 percent performance uptime. Of the 14 countries the TCO over the lifetime of the equipment is factored in. sampled, only 7 (operating 25 radars) reported performance Thus, while weather radar systems can represent an import- of 90 percent, with 7 countries (operating 22 radars) reported ant component of a modern national weather observation at less than 90 percent. While the data are limited, it is clear network, it is strongly recommended that the NMS ensure that a significant percentage of developing countries are not that it has sufficient funds and technical staff to support long- receiving the full value from the investment in weather radar term O&M before committing to the purchase of a weather systems and further investigation is warranted. radar system or systems. Costs related to the O&M for radar installations in Australia, Austria, Canada, Germany, and the United Kingdom are shown in table 3.7.3. ■ Australia operates the largest network of this group, with 67 radar sites in a mix of C-band (51) and S-band (16). ■ Austria operates 5 sites. Its radars are C-band. ■ Canada operates the second largest network, with 30 sites. Its radar network is undergoing full replacement to Weather Radar Systems   265 TABLE 3.7.3  Weather Radar Cost of O&M Data for Benchmarking and for Developed and Developing Countries Annual Lifetime (15 Annual O&M Annual O&M plus years) O&M Radar Expected cost (without labor cost labor cost plus labor Country or sites Performance lifetime RCI per site labor) per site per site per site cost per site radar type (number) (uptime) (years) (US$) (US$) (US$) (US$) (US$) TCO Developed countries Australia 67 95% 15 $4,500,000 $124,000 $78,000 $202,000 $3,030,000 $7,530,000 Austria 5 80% 15 — $572,000 — $572,000 $8,580,000 $8,580,000 Canada 30 95% 15 $3,400,000 $106,000 $92,000 $198,000 $2,970,000 $6,370,000 Germany 18 97% 15 $2,800,000 $65,000 $107,000 $172,000 $2,580,000 $5,380,000 United Kingdom 15 98% 15 $3,200,000 $71,000 $66,000 $136,000 $2,040,000 $5,240,000 Developed countries benchmarks S-band 1 95% 15 $3,900,000 $105,000 $85,000 $190,000 $2,850,000 $6,750,000 C-band 1 95% 15 $3,000,000 $87,000 $86,000 $173,000 $2,595,000 $5,595,000 Developing countries summary information Upper-middle — 87% 15 — $47,000 $22,000 $69,000 $1,035,000 — income Source: Data provided by NMS to the GFDRR of the World Bank. Note: All data are from 2020. Developing country data are based on averages from six upper-middle-income countries (Guyana, North Macedonia, Panama, Paraguay, South Africa, and Surinam). RCI per radar includes all equipment and installation costs. Annual O&M per radar column includes costs related to site lease or rent costs, utilities (electricity, telecommunications), security, spare parts and components, maintenance of radar and site, as well as life-cycle maintenance over the lifetime. United Kingdom annual O&M cost leaves out radio frequency spectrum access charge, which is significant. O&M = operations and maintenance; RCI = replacement cost including installation; TCO = total cost of ownership; — = not available. TABLE 3.7.4  Developing Country Weather Radar Uptime Performance Uptime Percentage C-, and S- band), and the geographical size of the country. performance Radar sites of station These costs include the purchase and replacement of spare Metric (%) (number) count parts, electrical power to operate the site (including backup Average 82% 47 100% generator costs), communication to facilitate data transfer, Greater than 90% 25 53% training NMS staff, and system upgrades, but do not include Less than 90% 22 47% salaries of staff required to maintain or support the radar sys- tem. Table 3.7.3 indicates that lifetime O&M is about a third Total   47   of capital cost of a C-band, with S-band being slightly higher Source: Data provided by NMS to the GFDRR of the World Bank. at about 40 percent. Note: All data are from 2020. Developing country performance data are from a combination of high-income, upper-middle-income, and low-income countries (Argentina, Bahamas, Botswana, Colombia, Costa Rica, Curacao, It should be noted that both Australia and Canada operate Ecuador, Estonia, Guyana, Hungary, Jamaica, Panama, Romania, and South similar sized networks over large geographical areas, al- Africa). though Canada’s O&M cost is $18,000 less per site (or 15 percent less) per year—reflecting the fact that its radar sta- Annual O&M costs. Annual O&M costs (without labor) per tions are considerably newer than Australia’s. For developing site are benchmarked at about $87,000 for C-band and countries, annual O&M spending (without labor) is $47,000; $105,000 for S-band per year. Factors affecting this cost will this cannot be compared with that of developed countries as be the number of sites, the variety of radars deployed (X-, the age or type of the radars are unknown. 266    Weather Radar Systems Business model. In Australia, Canada, Germany, and the network. Caution should be used as this estimate may under- United Kingdom, weather radar systems are both owned and value the cost of FTEs, depending on staff size and makeup. operated by the NMS. But in Austria, the weather radar sys- tems are owned, operated, and maintained by a third-party Table 3.7.5 shows a significant difference in staffing levels government partner, from which the NMS has contracted to between the developed and developing countries, where the purchase data fit for its observational and forecasting require- latter uses a higher staff complement to support the radar net- ments. Although the amount budgeted for Austria’s O&M cost work—possibly reflecting the technical expertise of personnel appears to be significantly greater, this discrepancy is largely and different government employment strategy. Germany’s due to the fact that it includes costs for personnel related to number is higher in comparison to the other developed coun- the O&M of the radar network, as well as depreciation and tries due to reunification, which resulted in more staff being replacement, which appear in different budget categories for absorbed. This number will be reduced over time through at- the other four countries. Taking this into account, Austria’s trition and retirements, bringing its FTE count more in line overall cost for weather radar data is comparable to that of with the other developed countries. Australia over the lifetime (15 years) and the TCO is only mar- ginally more, at less than $70,000 per year. This arrangement TABLE 3.7.5  Full-Time Equivalent Staff Numbers for Example means that Austria’s NMS does not need to recruit, train, or Countries retain personnel for these roles. The Austria business model Support personnel for the operations of their weather radar may be difficult to Radar sites per site implement fully as a result of logistic issues, safety concerns Country or radar type (number) (total, in FTEs) of personnel, and available technical staff in developing Developed countries countries. A developing country NMS may wish to consider Australia 64 0.75 a modified version of the outsourcing business model where Austria 5 —  the NMS is responsible for first-level maintenance and out- Canada 30 1.14 sources the more technical second-level maintenance, where Germany 18 1.73 finding or retaining qualified personnel is expected to be United Kingdom 15 0.75 a challenge. Australia is evaluating a third model with one radar where its Bureau of Meteorology purchased the sta- Developed countries benchmark tion but contracts out the maintenance and provision of data S-band — 1.09 under a service-level agreement. Even though the evaluation C-band — 1.09 is still in its early stages, the bureau sees several advantages Developing countries summary information of cost reductions in field staff and training, stocking spare Upper-middle income  — 2.73 parts, life-cycle management, and administrative expenses. Source: Data provided by NMS to the GFDRR of the World Bank. Note: All data are from 2020. Developing country data are based on averages 3.7.5.4 Personnel Costs from six upper-middle-income countries (Guyana, North Macedonia, Panama, Paraguay, South Africa, Surinam). Australia operates 67 radars across Countries seeking to estimate costs associated with the per- the country of C- and S-band. Austria’s data were not used in benchmark sonnel necessary to operate a weather radar network can calculation; it outsources the operations and supply of radar data to a different government department and has no maintenance responsibilities use an average cost for personnel, expressed as a full-time for the radar network. An estimate of the total number of support personnel equivalent (or FTE). This can be roughly estimated by divid- required to operate a large network of radars can be calculated by multiplying ing the total salary budget for the NMS by the total number the Support personnel per site (total, in FTEs) by the number of radar installations for that country. FTE for S-band is calculated from Australia and of staff employed by the NMS. This value will provide a rea- Canada. FTE for C-band is calculated from Australia, Germany and the United sonable weighted average of the cost of an individual NMS Kingdom. FTE = full-time equivalent; — = not available. staff member. Multiplying this value by the number of per- sonnel required to operate the weather radar network (field technicians and support staff) will provide an estimate of the Personnel required to operate and maintain a weather radar annual personnel costs required to support a weather radar network can be broadly divided into two categories: field Weather Radar Systems   267 technicians and support personnel. Field technicians main- weather radar network. Roles for support personnel include tain operation of the radar installations through maintenance but are not limited to: (scheduled and unscheduled), repair, and upgrades. The num- ber of field technicians required to ensure reliable operation ■ Specialists/scientists to analyze, interpret and apply radar of the radar network depends on several factors, including information the number and type of radar installations in the network and ■ IT specialists to support data ingestion, quality assurance/ the geographical distribution of stations. A country with a quality control (QA/QC) functions, storage and flow of data large network covering a large surface area will require more and products to forecast model field technicians to operate its network than will a smaller ■ Network management and planning, including service, in- country with the same number of stations; as a result, the cident, change and process improvement management and relationship between the number of stations and number of life-cycle support field technicians is nonlinear. For a single station, three FTE ■ 24/7 operational support from the IT service desk to log field technicians are recommended to allow reliable opera- field site, communications, server failures, and so on to tion in the case of the absence of one employee (such as sick ensure repairs are effected in a timely manner. leave, vacation leave, or departure from employment). The number of field technicians per station can be reduced below The number of required support staff positions will depend this 3:1 ratio as the network grows, provided technicians are on the size and complexity of the network but is generally able to reliably travel to more than one site within a reason- smaller in terms of a per-station ratio than that required for able timeframe and at reasonable cost. field technicians. For reference, the ratio of support staff to radar installations in both Canada and the United Kingdom Countries that operate and maintain their own weather radar is roughly 0.5:1. systems will require qualified and trained field technicians. These individuals should possess a degree in a related techni- 3.7.6 Recommendations cal discipline to perform the tasks of the position. In Canada, a technical degree is required for employment in this posi- In sum, when developing a strategy for implementing a tion; in some other countries, an engineering degree is re- weather radar system or network, an MHS should consider quired. Field technicians typically receive additional training the following recommendations: from the manufacturer of the weather radar systems, either at the manufacturer’s headquarters or at the site(s) they will 1. Evaluation is the first place to start. The NMS should be supporting. Alternatively, third-party partner can provide evaluate and articulate their data needs against the O&M of the radar systems, or a hybrid model can be employed status of existing resources—including in-house staff where the NMS provides first-level support and the third-par- technical capabilities, current state of existing net- ty partner is responsible for the more technical aspects of works, and annual O&M budget—before embarking on radar maintenance and repair. Under this model, staffing of acquiring and operating a new system. A gap analysis of field technician positions can be included as part of the con- current capabilities versus what is required to support tracted services, as is the case with Austria. a new system is imperative for developing a meaningful implementation plan and TCO. Regardless of whether the radar installations are operated 2. Be sure to include O&M costs. When budgeting the TCO and maintained by NMS personnel, a third-party partner, or of a weather radar station or network, it is critical to in combination, the NMS will require dedicated support per- consider annual O&M costs, on top of the initial procure- sonnel for data management and analysis, maintenance plan- ment cost of the radar system(s), given that these costs ning and scheduling, maintaining an inventory of spare parts, typically result in a rough doubling of costs (not includ- and life-cycle management. These personnel are in addition ing staffing costs) over the approximate 15-year expect- to the field technicians responsible for supporting O&M of the ed life cycle of a radar system or network. 268    Weather Radar Systems 3. Consider O&M financial support. IFIs or donor countries they require careful planning to ensure that the manu- should consider longer-term support beyond the initial facturer’s updates and enhancements are installed and capital investment for annual operational maintenance incorporated. of the radar(s), along with ongoing training of technical and scientific staff. By doing so, the donor and recipient country can realize the full value of the capital cost. 3.7.7 References 4. Provide for spares. Given the high cost of spares, it is recommended that a provision for spares be included in EverySpec. No date. FAA-STD-019D, FAA STANDARD, a long-term service-level agreement, rather than consid- LIGHTNING AND SURGE PROTECTION, GROUNDING, ered as an ongoing operational cost. BONDING, AND SHIELDING REQUIREMENTS FOR FACILITIES 5. Use FTEs for personnel costs. The number of field tech- AND ELECTRONICS EQUIPMENT 9. http://everyspec.com/ nicians and support staff required for reliable network FAA/FAA-STD/FAA_STD_019d_2298/. operation should be determined through consultations with the radar system manufacturer, NMSs with signifi- Leone, D.A., R.M. Endlich, J. Petriceks, R.T.H. Collis, and cant experience in radar weather network operation, or J.R. Porter. 1989. “Meteorological Considerations Used in entities with comparable expertise. Planning the NEXRAD Network.” Bulletin of the American 6. Consider outsourcing second-level maintenance func- Meteorological Society 70: 4–13. tions. NMSs that do not have experience operating and maintaining weather radar networks, are planning for ISO (International Organization for Standardization). 2019. smaller networks, or anticipate challenges in recruit- ISO 19926-1:2019 Meteorology – Weather Radar – Part 1: ing or retaining qualified personnel should consider a System Performance and Operation. ISO. https://www.iso.org/ modified version of the Austrian model by outsourcing obp/ui/#iso:std:iso:19926:-1:ed-1:v1:en. the highly technical second-level maintenance of the network to an experienced third-party partner. This ap- WMO (World Meteorological Organization). 2018. Guide to proach might also result in higher uptime performance Instruments and Methods of Observation, Volume I: Measure- in cases where the NMS does not have experience oper- ment of Meteorological Variables. WMO-No. 8. Geneva: WMO. ating complex weather radar systems, providing a great- https://library.wmo.int/doc_num.php?explnum_id=10616. er return on investment. 7. Fit the network into the location of other systems. WMO (World Meteorological Organization). 2021a. When designing the weather radar network, the location Draft Operational Weather Best Practices Guidance. of other stations or systems in the NMS’s observation May 2021. https://community.wmo.int/activity-areas/ network should be considered to maximize the benefit weather-radar-observations/best-practices-guidance. of the weather radar system or systems. 8. Timing of installations. If installing a large network, the WMO (World Meteorological Organization). 2021b. WMO timing of installations should be staggered. This allows Unified Data Policy Resolution (Resolution 1, Cg18.5). https:// the NMS to not only develop experience in all aspects public.wmo.int/en/our-mandate/what-we-do/observations/ of network and data management as the network comes Unified-WMO-Data-Policy-Resolution. online but also to stagger the replacement of the sys- tems when they reach end-of-life. This lessens the signif- icant capital cost burden of replacement of the network. 9. Do long-term upgrade planning. Key to the ongoing success of a weather radar network is long-term plan- ning for upgrades to computer equipment, software, and analysis of the incoming data. Weather radar systems are expensive to purchase, install, and maintain, and    269 Total Cost of Ownership Exercise 3.8.1 Introduction 3.8 When a National Meteorological Service (NMS) is interested in purchasing a weather observing system or network that is designed to support the NMS’s data requirements, a key tool is the total cost of ownership (TCO), which looks at the long-term affordability and financial requirements of implementation and operation of the system (see chapter 3.3). This tool helps the NMS gain clarity on the purpose of the monitoring system, the data required, and the design of the system’s main components (including individual stations, the communications network, and how the data will be ingested). This chapter examines how the TCO is calculated—an exercise that is intend- ed as a first step in ensuring that the NMS has the financial resources to oper- Photo: © Dreamstimepoint | Dreamstime.com ate the observation system prior to obtaining or receiving a system. It should not be seen as a substitute for a proper financial budgeting process. However, upon completion, it should be possible to: ■ Compare the capital cost of the project (equipment, computers, installa- tion) to the capital budget available. If the capital costs exceed the avail- able budget, examine the network to see where its size can be reduced. ■ Compare the projected cost of annual maintenance to the current or pro- jected operations budget. If the projected cost exceeds the current or fu- ture operations budget, look at where costs can be reduced that will not negatively impact the network design goals. Completing the exercise with as few assumptions as possible should provide an NMS, its respective government, and the World Bank with an indication as to whether the observation system under consideration is affordable and sustainable over its lifetime. After reviewing the TCO Exercise, an example of how to complete the calculation is provided in chapter 3.9. “This chapter examines how 3.8.2 Total Cost of Ownership Calculation the TCO is calculated—an We start with a look at the form that the NMS will need to fill out—the Total Cost exercise that is intended as of Ownership Summary Form (TCOSF) (figure 3.8.1)—which is divided into a first step in ensuring that three sections: (1) Total Initial Investment Cost (Lines 1–5), (2) Total Annual Operations and Maintenance Cost (Lines 6–16), and (3) Total Life-Cycle Costs. the NMS has the financial Each section of the TCOSF follows with instructions on completing the table resources to operate the to understand the TCO over the meteorological observation system’s lifetime. The blank version of the TCOSF can be printed and completed using the line- observation system prior by-line entry instructions that take up the rest of this chapter. The form uses to obtaining or receiving a red boxes for values that will go into Line X of the form, and blue boxes for working amounts. system.” 270    Total Cost of Ownership Exercise FIGURE 3.8.1  Total Cost of Ownership Summary Form Total Cost of Ownership Summary Form Project:   Date:   Capital Budget:   Currency:   Operation Budget:         Initial Investment (Capital Costs) Line   Description Cost   Total 1 Cost of Capital Equipment       2 Cost of Civil Works       3 Total Cost of Supplier Services       4 NMS Costs to Install Operational System       5 Total Initial Investment (Capital Costs) enter the total of lines 1, 2, 3, 4 into yellow box Cost of Annual Operations 6 Business Costs       7 Administrative Costs       8 Other Operating Costs       9 Total Cost of Annual Operations enter the total of lines 6, 7 and 8 into blue box © Dreamstimepoint | Dreamstime.com Cost of Annual Maintenance     10 Preventive Maintenance       11 Corrective Maintenance       12 Adaptive Maintenance       13 Total Cost of Annual Maintenance enter the total of lines 10, 11 and 12 into blue box 14 Total Cost of Annual Operations and Maintenance enter the total of lines 9 and 13 into green box 15 Lifetime of System (Years) enter the expected lifetime of the system into white box 16 Total Cost of Annual Operations and Maintenance over Lifetime multiply line 14 (green box) by line 15 (white box) and enter result into yellow box Life-Cycle Costs over Lifetime Total Life-Cycle Costs 17 enter number into yellow box Total Cost of Ownership over Lifetime 18 enter total from all yellow boxes (lines 5, 16 and 17) into gray box Total Cost of Ownership Exercise    271 3.8.2.1 Header The header provides data entry for the Project Name/Number and short description, date, budget numbers (if known) and the currency of the values used to complete the rest of the table. 3.8.2.2 Initial Investment (Capital Costs) The Total Initial Investment Cost includes the Cost of Capital Equipment, the Cost of Civil Works, Total Cost of Supplier Services, and Internal NMS Project Costs to install an operational system. When added together, this number represents the total cost of a fully installed operational system—that is, a system that is providing real-time data from field sites for NMS use. Keep in mind that when developing a TCO for a replacement project, significant costs will be reduced in the civil works area if the existing infrastructure has been well maintained and does not need major refurbishment. A gap analysis between the existing and new systems may reveal savings in other areas as well. Line 1 Cost of Capital Equipment Enter the costs associated with the procurement of hardware including cost of freight, insurance, and other carriage charges for all sites next to the descriptions below. It is suggested that these costs be obtained from one or more ven- dors through a request for information (RFI) process. ■ Total cost of all field equipment to be installed, including supporting hardware. This value can be obtained from one or more vendors. ■ Total cost of spare parts, sensors, or equipment to facilitate the repair of field equipment to maintain the operational status of the site. This value can be obtained from one or more vendors but is typically in the range of 10 percent to 20 percent. Vendors can provide recommendations on the level of spare equipment required. ■ Cost of mobile field equipment, including transportation equipment (truck, boat, trailer) and tools in support of the installation and repair of field equipment. Vendors can provide recommen- dations on the tools required. ■ Total cost of information processing system (IPS) hardware, including computer servers, soft- ware (if applicable), power supplies and batteries, backup systems, communication interfaces, and so on necessary to support retrieval of data from the field, quality assurance/quality control (QA/QC), data storage, archive, dissemination, and so on. This value can be obtained from one or more vendors. Vendors can provide recommendations on the equipment required to support the IPS. Enter the total of the above four numbers here and transfer the value to Line 1 of the TCOSF. 272    Total Cost of Ownership Exercise Line 2 Cost of Civil Works Enter the costs associated with the civil works development for all sites using the checklist below as a guide to gener- ate a total: ■ Total cost to procure land and buildings for all applicable sites. ■ Infrastructure costs for all sites. p Total cost of site preparation for all sites, which might include leveling and/or clearing the land and construction of roads to facilitate the installation and repair of field equipment. This value can be obtained from one or more vendors or contractors at the discretion of the NMS. p Installation of utilities, including mains power, water, sewer, and communications as required. p Construction of major supporting structures (buildings, tower bases, and towers) for observ- ing systems such as weather radar and upper-air systems. p Cost of security, including perimiter fencing, cameras, and alarms for security. p Construction of offices and server room for computer server and supporting hardware, includ- ing air conditioning, server racks, mains power, backup power, communications, and so on. Enter the total of the above six numbers here and transfer the value to Line 2 of the TCOSF. Line 3 Total Cost of Supplier Services Enter the costs associated with services provided by the vendor(s) for all sites next to the descriptions below, where applicable: Caution: Some or all the following numbers may be included in the Cost of Capital Equipment (Line 1). Please ensure that the numbers below have not been previously recorded. ■ Installation of equipment is to be carried out by the vendor(s). If the NMS is performing the installation of equipment at field sites, enter zero (0) for a value. These values where applicable will need to be provided by the vendor during the RFI process. ■ Factory Acceptance Testing is the testing of the equipment to be purchased; this is typically carried out at the vendor's facility. ■ Site Acceptence Testing should be included in the costs when the NMS contracts the installation of the field equipment to ensure all systems are operating to specification. ■ Training of staff on the new equipment and systems to ensure NMS can operate and maintain the equipment and systems. ■ Prepaid calibration of sensors. Some manufacturers may provide prepayment of calibration ser- vices for key components or sensors over a period of years at the time of the acquisition of the system. Enter the total of the above five numbers here and transfer the value to Line 3 of the TCOSF. Total Cost of Ownership Exercise    273 Line 4 NMS Costs to Install Operational Systems Enter the staff costs associated with the management and installation of site equipment for all sites next to the de- scriptions below where applicable. If the staff costs are not known, calculate a weighted average cost of a full-time equivalent (FTE) as described in chapters 3.5, 3.6, and 3.7. Many of the costs in this section will be based on estimates depending on the information available to the NMS. Care should be used in providing the best estimation of costs to reflect an accurate TCO. ■ Project management should be carried out by an NMS staff member (if experienced) or by a consultant hired to act on behalf of the NMS. The project manager is responsible for development and analysis of the RFI process; developing the resulting tender documentation; and coordinating and scheduling with the vendor(s) equipment delivery, installations, training, civil works, and testing to ensure the project runs smoothly and is completed on time. Time required for project management can be as little as six months or, depending on scope and complexity, as long as one year or more. ■ Cost of logistics: Estimate the cost for staff travel expenses (food, lodging, etc.) ¨ To travel to the site(s) to complete the installation (staff costs for travel and other travel costs such as fuel, air fare, and more), ¨ Time on site to complete the installation including costs for accommodations and food, ¨ Time of staff at a vendor facility for Factory Acceptance Test, training, travel time and cost of travel including public transportation (air, bus, rail), vehicle rental, fuel, food, and accommodations. Optional Costs ■ Installation time costs for NMS personnel may be calculated here if the agency needs to trap these costs specific to the project whether the NMS will carry out the full installation of the equipment or support the vendor. However, the costs associated with NMS personnel will be included as part of the annual costs associated with Operations and Maintenance further down. If the NMS wishes to include a total cost of time against the project, use the following calculation as a guide. p FTE cost (per field technician) × number of FTEs related to the field installation of equip- ment × the total number of FTE days to carry out the installation, including travel time. p For example: if the cost of a single FTE is $150.00 per day and 2 FTEs are required to install a network of 10 AWSs at 2 days per site, plus 12 days of travel, the calculation is: p $150.00/day × 2 FTEs × [(10 sites × 2 days/site) + 12 days] = $9,600.00 for the installation. ■ Cost of staff training on the new equipment and systems to ensure the NMS is able to operate and maintain the equipment and systems. These costs are for onsite training of NMS person- nel and may be calculated here if the agency needs to trap these costs specific to the project. Calculate the FTE cost (per field technician) × number of FTEs to be trained × the total number of training days. p For example: if the cost of a single FTE is $150.00 per day and 5 FTEs are to be trained, the calculation is: p $150.00/day × 5 FTEs × 10 training days = $7,500.00 for the training. Enter the total of the above two numbers and the optional numbers (if completed) here and trans- fer the value to Line 4 of the TCOSF. 274    Total Cost of Ownership Exercise Line 5 Total Initial Investment Cost On the TCOSF, total Lines 1 to 4 and enter value into the Yellow Box, Line 5, which represents the Initial Investment Cost. 3.8.2.3 Total Cost of Annual Operations The Total Cost of Annual Operations includes the yearly costs to ensure that system performance (uptime) is met and is pro- viding quality data fit for purpose. The NMS can choose to do this work in-house or have the maintenance work completed by a third-party vendor. Regardless, these costs are significant when totaled over the lifetime of the equipment. Line 6 Business Costs Business operating costs include operating expenses such as rent (office, furniture, and so on), equipment (servers, computers, vehicles), inventory costs, payroll, insurance, computational software costs. Enter the total of Business Costs here and transfer the value to Line 6 of the TCOSF. Line 7 Administrative Costs Administrative costs include the costs of support staff (FTEs) salaries required to carry out the work of the day-to-day financial and human resource management, scheduling site maintenance, warehousing and maintaining inventory con- trol of spares, ordering replacements of failed equipment, reviewing network health, developing processes, and life-cy- cle planning and scheduling. This number should also include the costs of benefit packages or other remuneration. For example: if the average cost of a single FTE is $38,000 per year and there are 9 FTEs required support and operate the observation system, the calculation is: $38,000/year × 9 FTEs = $342,000 in annual staff costs. Enter the total of Administrative Costs here and transfer the value to Line 7 of the TCOSF. Line 8 Other Operating Costs Enter the costs associated with other annual costs for all sites, using the checklist below as a guide to generate a total for the following: ■ Calibrations are required to ensure data quality. Calibrations are typically carried out by re- turning the equipment to the manufacturer or a local calibration facility. Follow manufacturer’s recommendation for calibration intervals for equipment. Enter an estimate of calibration costs, including shipping for all equipment to be calibrated. Some manufacturers offer a prepayment of calibration including transportation costs as a service at the time of purchasing the system. If a prepaid calibration option has been included in procurement cost, enter a value of 0. ■ Consumables might or might not be a significant cost depending on the observation system under consideration. This cost is very significant for upper-air systems, which require consumables for each launch (twice per day; see table 3.6.2). Another example of consumables is fuel for backup generators for weather radar systems; that should also be entered here. ■ Software licenses are an annual cost that provide various levels of upgrades and support. Total Cost of Ownership Exercise    275 ■ Service-level agreements for software are common. An NMS should also consider implementing a service-level agreement to service complex equipment such as radar or computer room servers to reduce the expense of training and salaries of highly qualified technical staff. ■ Land lease costs for use and access to land not owned by the government. ■ Utilities, including mains power, water, sewer, and communications, as required for all sites combined. Enter the total of the above six numbers here and transfer to Line 8 of the TCOSF. Line 9 Total Cost of Annual Operations On the TCOSF, total Lines 6, 7, and 8 and enter value into Yellow Box, Line 9, which represents the Total Cost of Annual Operations. 3.8.2.4 Total Cost of Annual Maintenance As discussed in chapters 3.4, 3.5, 3.6, and 3.7, there are three types of maintenance: preventive, corrective, and adaptive. A maintenance program is required to keep equipment in operating condition and to ensure reliable data quality over the antici- pated lifetime of the system. Each type of maintenance needs to be accounted for below: Line 10 Preventive Maintenance Preventive maintenance is carried out on a regular basis following vendor guidelines, and should include required equipment checks and recalibration, periodic cleaning and lubrication, and equipment replacement and upgrades, as necessary. Routine preventive maintenance should follow vendor guidelines on frequency. The cost of FTEs and the cost of logistics was calculated above as a one-time cost to install the system and annual costs for these categories will need to be included below. The largest costs typically associated with maintenance activities are those of field staff and travel activities. Cost of field staff includes the necessary staff to carry out the maintaince work (field technicians) at all the sites. This number should also include the costs of benefit packages or other remuneration. ■ Field technician staff carry out the more technical duties of annual maintenance and repair of the observation system. Calculate the staff salaries similar to the example provided above in Administrative Costs (Line 7). ■ Site maintenance personnel carry out the less technical duties (for example, cutting of grass, cleaning of sensors, security). ■ Cost of logistics estimates the cost for staff to travel to the site(s) to carry out maintenance ac- tivities. This number should include public transportation (air, bus, rail), vehicle rental, vehicle maintenance (if owned) fuel, food, and accommodations. ■ Cost of mobile field equipment estimates the cost of maintenance for vehicles and other modes of transportation owned by the NMS. ■ Cost of replacement parts estimates the cost of replacement parts that have been damaged by environmental conditions, vandalism, theft, and so on. 276    Total Cost of Ownership Exercise ■ Cost of IPS computer equipment estimates the annual cost to maintain computer, servers, and other hardware. Enter the total of the above six numbers here and transfer to Line 10 of the TCOSF. Line 11 Corrective Maintenance Corrective maintenance is the repair of unanticipated equipment failures damaged by the environment, vandalism, theft. It is suggested that a value of 10 percent to 15 percent be used from the Benchmark Preventive Maintenance values presented in chapters 3.5, 3.6, and 3.7 per site until the NMS has gained enough experience to develop its own cost understanding through the operation of the system. This number is an estimate and includes the cost of replacing equipment that is no longer serviceable. Enter the total of Corrective Maintenance here and transfer to Line 11 on the TCOSF. Line 12 Adaptive Maintenance Annual Maintenance entails annual planned support and upgrades of the entire value chain, from measurement to data processing and training. Enter the total of Adaptive Maintenance here and transfer to Line 12 on the TCOSF. Line 13 Total Cost of Annual Maintenance Calculate the Total Cost of Annual Maintenance by adding Lines 10, 11, and 12 and enter the result on Line 13 on the TCOSF. Line 14 Total Cost of Annual Operations and Maintenance Calculate the Total Cost of Annual Operations and Maintenance by adding Lines 9 and 13 and enter the result on Line 14 on the TCOSF. Line 15 Lifetime of System in years Enter the lifetime of the observation system under consideration, in years—based on the bench- marks calculated from chapters 3.5, 3.6, and 3.7—on Line 15. Line 16 Total Cost of Annual Operations and Maintenance over Lifetime To calculate the Total Cost of Annual Operations and Maintenance over the lifetime of the observation system, multiply the Total Cost of Annual Operations and Maintenance (Line 14) by Lifetime of System (Line 15) and enter the result into Yellow Box on Line 16. 3.8.2.5 Replacement Costs While the expected lifetime of the system can be as high as 20 years or more (upper air), parts and sensors will typically have shorter lifetimes. As a result, these components of the system will require more frequent replacement. Manufacturer’s recom- mendations on the frequency of replacement should be followed. Line 17 Total Life-Cycle Costs Life-cycle costs typically fall into one of three categories: the cost of field equipment (parts and sensors), the cost of mobile field equipment (vehicles and so on), and the cost of IPS computer and software (server hardware). Each follow the same process for calculating the value for cost of replacements over the lifetime of the system, and the NMS will need to separately calculate a value for each category, where applicable, as well as for each specific part. Total Cost of Ownership Exercise    277 Lifetime of System Cost of Replacement × × Number of Components Component Useful Life in Years ■ Cost of Field Equipment at end-of-life. For example, the lifetime of an AWS system is bench- marked at 10 years, but certain components (such as a relative humidity sensor, batteries) may have a useful life of 5 to 7 years if properly maintained and calibrated. Data loggers and com- munications devices may have lifetimes as high as 20 years and will not factor into a life-cycle replacement cost. 10 Years Cost of Replacement × × Number of Components 7 Years = 1.4 (since the component parts will be replaced only once in the lifetime, use the value “1”). Use the equation above to calculate this cost. ■ Cost of other field equipment, such as vehicles used in supporting the observation system for installation, maintenance, and so on. ■ Cost of IPS computer and software. The IPS computer hardware typically has a lifetime of 5 to 7 years. Costs associated with relevant system components (such as servers, uninterruptable power supply or UPS, communications) need to be calculated. The software may require period- ic updating for enhancements throughout the life of the system. For example, the International Civil Aviation Organization (ICAO) will make changes to the automated weather observing sys- tem (AWOS) (every 2 to 3 years), which need to be updated in order to maintain an aerodromes compliance. Enter the total of the above three numbers here and transfer the value to Line 17 of the TCOSF. 3.8.2.6 Calculate the Total Cost of Ownership Over Lifetime The Total Cost of Ownership Over Lifetime can now be calculated by adding the three yellow boxes (Lines 5, 16, and 17) on TCOSF and enter the result in Line 18. This value represents the TCO. If the budget numbers for capital and operation of the observation system are not known, the NMS can use the TCO to support the development of the budget. If the NMS does have capital and operation budgets allocated, they will need to evaluate these numbers against the following questions: 1. Does the capital cost (Line 5) of the project (equipment, computers, installation) exceed the capital budget available to the NMS? 2. Does the projected cost of annual operations and maintenance (Line 14) exceed the current operations budget available to the NMS? The NMS may use the information on the TCOSF to apply for both a capital and annual operation budget. If either the projected capital cost (Line 5) or the projected operations and maintenance cost (Line 14) exceeds the budget available, the NMS will need to make adjustments to its funding or costs through one or more of the following measures: 1. Obtaining more financial resources to meet the projected needs. 2. Looking for ways to decrease these projected costs by reducing the network size to meet available budget. 3. Looking at other business models that might be more cost-effective. 278    3.9 Total Cost of Ownership Example Calculation Note to Reader The purpose of this exercise is to provide an example of how to cal- culate the total cost of ownership (TCO) and complete the Summary Table. The values represented are not transferable as they are specific to this example. Each National Meteorological Service (NMS) must do its own research to develop a proper TCO for the weather observing system it wishes to implement. The guidance for this example comes from the TCO concepts in chapter 3.3, the automatic weather station (AWS) discussion in chapter 3.5, and the TCO exercise instructions in chapter 3.8. 3.9.1 Context for TCO Example A developing world NMS wants to add 21 AWS sites to augment its national manual surface weather observation network of 28 stations. These AWSs will Photo: Petrovich9 be new stations and used to infill data holes not covered by their manual weather stations. This will be the NMS’s first project to incorporate automa- tion and it has set a target performance of 95 percent. The new stations will be installed by a third-party contractor and will include a new server room at the NMS headquarters to ingest, process, and store the data automatically. Quality assurance/quality control (QA/QC) of the data will be done prior to forwarding to the forecast modeling center. Considerations for the development of the TCO include a site breakdown as follows: ■ The NMS has a capital budget of $1.1 million to purchase and install all equipment and an annual operations and maintenance budget of $150,000 per year. ■ As all the sites are to provide synoptic data to a forecast model, they will be configured as discussed in chapter 3.5 with: p A sensor complement of air temperature and relative humidity, wind speed and direction (measured at a height of 10 meters above the ground surface), atmospheric (barometric) pressure, precipitation (via a tipping bucket rain gauge suitable for measurement of tropical rain events), and solar radiation. Total Cost of Ownership Example Calculation    279 p Data acquisition with redundant satellite and cellular remove debris from the rain gauge three times per week communications, housed in a weatherproof enclosure as well as cut grass as needed. To maintain and repair sta- with batteries and power supply. tions, the NMS field staff will be dispatched from a central p A 10-meter tower with mounting hardware to support location. It will take 42 days in travel time for the technical all equipment. field staff to visit each site once from the NMS central loca- ■ Five sites will require the purchase of land, 4 sites will be tion (table 3.9.1). situated on leased land, and 12 sites will be installed on ■ Development of a server room at the NMS headquarters land owned by the government. with air conditioning and server hardware, uninterrupt- ■ Eighteen sites will be powered by solar panels and batter- able power supply (UPS), and communications links. ies, and 3 will have mains power with battery backup. ■ All sites will require varying degrees of civil works site TABLE 3.9.1  Travel Time Needed for Site Visits preparation that include: p The clearing of land of vegetation, some degree of lev- Travel time (days) Number of sites Total travel (days) eling, and planting of grass. 1 9 9 p The pouring of concrete for tower bases, precipitation 2 4 8 gauge and installation of a 2-meter tall security fence. 3 7 21 p The addition of mains power for three sites. 4 1 4 ■ To facilitate the maintenance of the sites, the NMS will Total 21 42 employ local personnel to clean the radiation sensor and 280    Total Cost of Ownership Example Calculation 3.9.2 Total Cost of Ownership Example Calculation Using the concepts from chapters 3.3 and 3.5, the instructions in chapter 3.8, and the context above, the following provides guidance on completing the Total Cost of Ownership Exercise and the TCOSF. Line 1 Cost of Capital Equipment Total cost of all field equipment. The average station cost from the request for information (RFI) pro- $525,000 cess was $25,000/station (21 stations), including sensors and mounting hardware, data acquisition and communications hardware, 10-meter tower, and power supplies (batteries and solar panels). Total cost of spare parts, sensors, and equipment. A budgetary figure of 15 percent of the capital $78,750 cost of the equipment (above) was used to calculate this value. Cost of mobile field equipment. To facilitate site visits, the NMS will purchase one truck as well as four complete sets of tools necessary to support the installation and repair of field equipment. $30,000 Total cost of information processing system (IPS) hardware. This cost includes computer serv- ers, applicable software, power supplies and batteries, backup systems, communication inter- $35,000 faces, and so on necessary to support retrieval of data from the field, quality assurance/quality control (QA/QC) data storage and archival, dissemination, and so on. $668,750 Enter the total of the above four numbers here and transfer the value to Line 1 of the TCOSF. Line 2 Cost of Civil Works Enter the costs associated with the civil works development for all sites using the checklist below as a guide to gener- ate a total: ■ Total cost to procure land for five sites at an average cost of $2,500 per site. $12,500 ■ Infrastructure costs for all sites. p Total cost of site preparation for all sites includes leveling of the land, construction of roads, and constructing tower bases to facilitate the repair of field equipment to maintain $178,500 the operational status of the site. This value was obtained from three local contractors and was estimated to be, on average, $8,500 per site. p Installation of utilities of mains power at three sites at an average cost of $750 per site $2,250 from estimates from three local contractors. p Construction of major supporting structures. None of the 21 sites require buildings as the $0 AWS was designed as a standalone system and a value of $0 was budgeted. p Cost of security perimiter fencing is $1,000 per site for 21 sites from estimates from three $21,000 local contractors. p Construction of offices and server room for server hardware including air conditioning, $21,000 server racks, mains power, backup power, communications, and so on. This value is an es- timate obtained from three local contractors. No additional office space was required and not included in the costing. Enter the total of the above six numbers here and transfer the value to Line 2 of the TCOSF. $226,250 Total Cost of Ownership Example Calculation    281 Line 3 Total Cost of Supplier Services Enter the costs associated with services provided by the vendor(s) for all sites next to the descriptions below, where applicable: Caution: Some or all the following numbers may be included in the Cost of Capital Equipment (Line 1). Please ensure that the numbers below have not been previously recorded. ■ Installation of equipment is to be carried out by the equipment supplier with support from NMS personnel. The estimated cost of $750 per day per person including supplier logistics $133,500 (travel to country and so on) for two FTEs was used, based on information obtained from the RFI process. Each site will take 2 days to install and travel time to the various sites is estimated at 42 days for installation and 42 days for logistics of travel for a total of 84 days. Additionally, the supplier is responsible for the installation of the server room equipment adding 5 addition- al days per person in costs. Calculation: 2 Personnel X [84 Days (travel and installation) + 5 Days (server room equipment installation)] X $750 = $133,500 ■ Factory Acceptance Testing (FAT) is to be carried out at the vendor’s facility over 3 days. This $5,000 cost represents a flat fee to perform the FAT. ■ Site Acceptence Testing was estimated from the RFI process at $200 per site to ensure all $4,200 systems are operating to specification. ■ Training of NMS staff by the vendor on the new equipment and systems to be carried out over $12,000 10 days, including supplier logistics costs. ■ Prepaid calibration of sensors. The NMS chose to include calibration of the relative humidity sensor, barometric pressure sensor and pyranometer including shipping to and from the NMS’s $50,000 facility at a cost of $50,000 for 5 years. Enter the total of the above five numbers here and transfer the value to Line 3 of the TCOSF. $204,700 282    Total Cost of Ownership Example Calculation Line 4 NMS Costs to Install Operational Systems Although the NMS has chosen to contract the installation of the sites, the building of the server room, and civil works, it will still have significant responsibilities and costs associated with managing the project and ensuring that the finished sites meet its requirements. The NMS should use the actual costs for field technicians, a senior manager, and other internal staff, if available. In this example, it is assumed the costs for the FTEs is an average value of $10,400 per FTE per year. Based on 220 working days per year, this represents a daily rate of approximately $47.00. This value was used to develop NMS daily costs per FTE. ■ Project management. As the NMS has no experience in the management of a large automation $164,000 project, the NMS hired a consultant to manage the project. The contracted project manager is responsible for assisting in site selection, review of site communication availability, devel- opment and release of the RFI, analyzing the results, developing the Request for Tender doc- uments, coordinating and scheduling with the vendor, installations, training, civil works, and testing to ensure the project runs smoothly and is completed on time. Time required for the management of this project is 180 days, at a cost of $800 per day, plus logistics is estimated at $20,000. ■ Cost of logistics estimates the cost of: p Time for staff to travel to the site(s) to complete the installation $15,000 p Time for staff on site to complete the installation p Time of staff at a vendor facility for Factory Acceptance Test, training, travel time, and cost of travel including public transportation (air, bus, rail), vehicle rental, fuel, food, and accommodations. Optional Costs Assumption: It is assumed that NMS personnel are full-time staff and the costs associated with installation time and cost of staff training below are therefore captured in the next section, Line 6, Total of Cost of Operations and Maintenance and are not part of the total in Box 4, on Line 4 of the TCOSF. As a result, the calculations for these two sections are not provided in this example. ■ Installation time $0 ■ Cost of staff training $0 Enter the total of the above two numbers and the optional numbers (if completed) here and $179,000 transfer the value to Line 4 of the TCOSF. Total Cost of Ownership Example Calculation    283 Line 5 Total Initial Investment Cost On the TCOSF, total Lines 1, 2, 3, and 4 and enter value into the Yellow Box, Line 5, which rep- $1,278,700 resents the Total Initial Investment Cost. 3.9.3 Total Cost of Annual Operations The Total Cost of Annual Operations includes the yearly costs to ensure that system performance (uptime) is met and provid- ing quality data fit for purpose. The NMS can choose to do this work in-house or have the maintenance work completed by a third-party vendor. Regardless, these costs are significant when totaled over the lifetime of the equipment. Line 6 Business Costs Business operating costs are part of the government’s facility management and there are no additional costs to the NMS. Enter the total of Business Costs here and transfer the value to Line 6 of the TCOSF. $0 Line 7 Administrative Costs The NMS has 2.5 FTEs carrying out support staff duties for financial and human resource management, scheduling site maintenance, warehousing and maintaining inventory control of spares, ordering replacements of failed equipment, reviewing network health, development of processes and life-cycle planning and scheduling. This number includes the costs of benefit packages or other remuneration. Calculation: 2.5 FTEs at $10,400 per year equals an administrative cost of $26,000 Enter the total of Administrative Costs here and transfer the value to Line 7 of the TCOSF. $26,000 Line 8 Other Operating Costs Enter the costs associated with other annual costs for all sites, using the checklist below as a guide to generate a total for the following. ■ Calibrations are required to ensure data quality. Calibrations for relative humidity sensor, barometric pressure sensor, and pyranometer were costed under prepaid calibrations. The NMS $0 still must calibrate the network of precipitation gauges, but as this is completed by field tech- nicians, there is no additional cost to the NMS. ■ Consumables for an AWS are typically not significant and are usually limited to desiccant for the barometric pressure sensor electronics enclosure, and so on. A value of $25 per site per $525 year is budgeted. ■ Software licenses are an annual cost that provide various levels of upgrades and support. The $5,000 cost of the software license running on the NMS servers is $5,000. ■ Service-level agreements have been allotted for the maintenance of servers annually that sup- ports the 24/7 operation of the NMS minimizing downtime of the IPS. This has been contracted $12,000 to an outside vendor at a value of $12,000 per year. ■ Land lease costs for the four sites on leased land have a nominal cost of $300 per site per year $1,200 or a total value of $1,200. 284    Total Cost of Ownership Example Calculation ■ Utilities total annual cost for this scenario includes: p Mains power at 3 field sites at $15 per month, or $180 per site annually (total $540) $17,580 p Cellular service at 21 sites at $20 per month, or $240 per site annually (total $5,040) p Internet connection for the server room plus mains power and air conditioning at $12,000 annually. Total cost of utilities is $17,580. Enter the total of the above six numbers here and transfer to Line 8 of the TCOSF. $36,305 Line 9 Total Cost of Annual Operations On the TCOSF, total Lines 6, 7, and 8 and enter value into the Yellow Box, Line 9, which rep- $62,305 resents the Total Cost of Annual Operations. 3.9.4 Total Cost of Annual Maintenance As discussed in chapters 3.4, 3.5, 3.6, and 3.7, there are three types of maintenance: preventive, corrective, and adaptive. A maintenance program is required to keep equipment in operating condition and to ensure reliable data quality over the antici- pated lifetime of the system. Each type of maintenance needs to be accounted for below: Line 10 Preventive Maintenance is carried out on a regular basis, following vendor guidelines. ■ Cost of field staff includes the necessary staff to carry out the maintaince work (field techni- cians) at all the sites. This number should also include the costs of benefit packages or other remuneration. p Field technician. The NMS has 2.5 staff to carry out the maintenance work at a cost of $26,000 $10,400 per FTE. p Site maintenance personnel. The NMS will contract local individuals to ensure the site is secure, grass cut, and minor site maintenance done, including cleaning of the radiation $5,250 sensor dome and the radiation shield and precipitation gauge of debris daily at a cost of $250 per site per year. ■ Cost of logistics. It is estimated that the NMS will spend about $2,000 for public transporta- $8,000 tion, food, and accommodations to visit all sites once. Given each site will be visited four times per year, the cost of logistics is estimated at $8,000. ■ Cost of mobile field equipment. The NMS will spend about $2,200 to visit all sites once, using $8,800 the purchased vehicle (maintenance, insurance, fuel). Given each site will be visited four times per year, the cost of logistics is estimated at $8,800. ■ Cost of replacement parts. The NMS estimates the cost of replacement parts that have been damaged by environmental conditions, vandalism, theft, and so on to be 10 percent of the cap- $52,500 ital cost of the field equipment. ■ Cost of IPS computer equipment. The NMS has a service-level agreement that includes the cost $0 of maintenance. Enter the total of the above six numbers here and transfer to Line 10 of the TCOSF. $100,550 Total Cost of Ownership Example Calculation    285 Line 11 Corrective Maintenance is used for unanticipated failures of network equipment. This value includes the cost of replacement parts to be returned to the spare parts inventory and the cost of $15,100 logistics. As the NMS has allocated for replacement parts for the network, parts are not included in this value. The NMS will monitor its corrective maintenance costs over the next three years to develop a cost based on in-country experience and has budgeted an approximate value of 15 percent of the Preventive Maintenance Cost. Enter the total of Corrective Maintenance here and transfer to Line 11 on the TCOSF. Line 12 Adaptive Maintenance is planned support and upgrades of the entire value chain, from measurement to data pro- cessing. In this example, the NMS has signed a service level agreement with a vendor, who will provide updates and upgrades as part of the service-level agreement. In cases where the NMS will perform the adaptive maintenance, an appropriate value (determined in consultation with vendors) should be entered here. A value of $0 is recorded in this example as the cost will be accounted for in Other Annual Costs under service-level agreements. Enter the total of Adaptive Maintenance here and transfer to Line 12 on the TCOSF. $0 Line 13 Total Cost of Annual Maintenance Calculate the Total Cost of Annual Maintenance by adding Lines 10, 11, and 12 and enter the $115,650 result on Line 13 on the TCOSF. Line 14 Total Cost of Annual Operations and Maintenance Calculate the Total Cost of Annual Operations and Maintenance by adding Lines 9 and 13 and $177,955 enter the result on Line 14 on the TCOSF. Line 15 Lifetime of System(s) in Years Enter the lifetime of the observation system under consideration, in years—based on the bench- marks calculated from chapters 3.5, 3.6, and 3.7—on Line 15. 10 years Line 16 Total Cost of Annual Operations and Maintenance over Lifetime To calculate the Total Cost of Annual Operations and Maintenance over Lifetime of the observation $1,779,550 system, multiply the Total Cost of Annual Operations (Line 14) and Maintenance by Lifetime of System (Line 15) and enter the result into Yellow Box on Line 16. 286    Total Cost of Ownership Example Calculation 3.9.5 Replacement Costs While the expected lifetime of the system can be as high as 20 or 25 years (weather radar), parts and sensors will typically have shorter lifetimes. As a result, these components of the system will require more frequent replacement. Manufacturer’s recom- mendations on the frequency of replacement should be followed. Line 17 Total Life-Cycle Costs The expected lifetime of 10 years was used for the AWS, but the air temperature and relative humidity sensors, batter- ies, and server equipment have shorter lifespans—5 years for each of these components. As a result, these components will require replacement at least once prior to the end-of-life of the AWS if the network is to remain operational. ■ Cost of parts and sensors in this example include replacement of the air temperature/relative humidity sensors and station batteries for each of the 21 stations at the end of these compo- $78,750 nents’ useful life of 5 years. This cost is estimated at $1,875 per site. Given that two potential replacements are required during the 10-year lifespan of the AWS, the total life-cycle cost to ensure operations of the AWS is $78,750. ■ Cost of mobile field equipment NMS will replace the field maintenance vehicle once in the 10- $25,000 year period at an approximate cost of $25,000. ■ Cost of server hardware for the IPS typically has a lifespan of 5 to 7 years. Costs are associated $35,000 with servers, UPS, communications, and so on. Given this is shorter than the life of the network, these components will require replacement at a budgeted cost of $35,000. ■ To account for worst case cost of inflation and exchange rate risk, it is prudent to multiply the $30,525 above sum by a factor greater than unity. In this example, annual inflation and exchange rate risk over the 5-year period before replacement of the components is assumed to be 4 percent, or a factor of 1.04. Over 5 years, this factor is (1.04)5 = 1.22 or 22 percent of additional cost. Multiplying the calculated total life-cycle cost of ($78,750 + $25,000 + $35,000) or $138,750 by 22 percent yields a cost of inflation and exchange risk of $30,525. The Total Life-Cycle Costs of $169,275 is entered in Box 17, on Line 17 of the TCOSF. Enter the total of the above four numbers here and transfer the value to Line 17 of the TCOSF. $169,275 3.9.6 Calculate the Total Cost of Ownership Over Lifetime Line 18 Total Cost of Ownership Over Lifetime The Total Cost of Ownership Over Lifetime can now be calculated by adding the three yellow boxes on the TCOSF to- gether and entering the result into the white Line 18, which equals $3,227,525 (figure 3.9.1). Total Cost of Ownership Example Calculation    287 FIGURE 3.9.1  TCOSF Example 1 Calculation Total Cost of Ownership Summary Form Project:  Example 1 Calculation Date:  1/20/22 Capital Budget:  $1,100,000 Currency: US dollars  Operation Budget:  $150,000 per year       Initial Investment (Capital Costs) Line   Description Cost   Total 1 Cost of Capital Equipment $668,750     2 Cost of Civil Works $226,250     3 Total Cost of Supplier Services $204,700     4 NMS Costs to Install Operational System $179,000     5 Total Initial Investment (Capital Costs) $1,278,700 enter the total of lines 1, 2, 3, 4 into yellow box Cost of Annual Operations 6 Business Costs $0     7 Administrative Costs $26,000     8 Other Operating Costs $36,305     9 Total Cost of Annual Operations $ 62,305 enter the total of lines 6, 7 and 8 into blue box Cost of Annual Maintenance     10 Preventive Maintenance $100,550     11 Corrective Maintenance $15,100     12 Adaptive Maintenance $0     13 Total Cost of Annual Maintenance $115,650 enter the total of lines 10, 11 and 12 into blue box 14 Total Cost of Annual Operations and Maintenance $177,955 enter the total of lines 9 and 13 into green box 15 Lifetime of System (Years) 10 enter the expected lifetime of the system into white box 16 Total Cost of Annual Operations and Maintenance over Lifetime $1,779,550 multiply line 14 (green box) by line 15 (white box) and enter result into yellow box Life-Cycle Costs over Lifetime Total Life-Cycle Costs 17 $169,275 enter number into yellow box Total Cost of Ownership over Lifetime 18 $3,227,525 enter total from all yellow boxes (lines 5, 16 and 17) into gray box 288    Total Cost of Ownership Example Calculation At this point, the NMS needs to return to the key cost ques- 8 days of logistic costs of the 4 stations, adding an tions to see if there is a need to reduce the TCOSF, and if so, additional $24,800 in savings. by how much can this be done. p The NMS looked at other potential areas of savings, realizing a learning opportunity by cutting the proj- 1. Does the Total Initial Investment Cost (Line 5) of ect management consultant’s time from 180 days to the project (equipment, computers, installation) 75 days. An internal manager within the NMS would exceed the capital budget available to the NMS? carry out most of the project management work, using the consultant as an adviser to ensure the proj- The NMS originally had a budget limit of $1.1 million as a ect does not suffer delays in the schedule and imple- Total Initial Investment Cost for the AWS Network. Line mentation of the network. This created an additional 5 on the TCOSF above exceeds the budget by $178,700. savings of $84,000. It is clear from this example that the NMS will need to modify its expectations and review the network design. These cost saving measures balanced the Total Initial In this case study, the NMS chose to make two changes Investment Cost for the AWS Network project under budget, in the initial investment cost: at $979,900, leaving contingency of 11 percent ($120,100) for minor cost overruns. Additionally, there was a significant p The first change was to reduce the network size to 17 savings in the life-cycle management due to the smaller num- stations, removing 4 sites that were considered the ber of field stations as well as the IPS computer hardware least significant contributors to their design goals, at totaling a savings of $61,000. The amended numbers are $25,000 per site. It was felt the removal of 4 sta- summarized in figure 3.9.2 for the TCOSF for Example 1 (21 tions would have the least impact and still allow the stations) and Example 2 (17 stations). NMS to provide better spatial coverage of the country for the purposes of increasing public forecasts and Figure 3.9.2 shows the reduction in all four areas of the Total severe weather alerts. This created an immediate Cost to Develop an Operational System (Lines 1 to 4). savings of $100,000 plus a reduction in the cost of spare parts. 2. Does the projected Total Cost of Annual Operations p The second change was through the RFI process, and Maintenance (Line 14) exceed the current opera- with some of the respondents proposing a cloud ser- tions budget available to the NMS? vices solution. This saved the NMS from purchasing IPS servers and associated hardware ($35,000), as The NMS had a budgeted annual increase in addition to well as the construction cost of a new server room their original budget for Operations and Maintenance ($12,000). The NMS did take the opportunity to add of $150,000 in support of the new network of stations. $10,000 of computers to assist the field technicians Line 14 on the TCOSF (figure 3.9.1) exceeds the budget and internal support staff to facilitate management by nearly $28,000. While a seemingly small amount, of the network. The net savings to the capital equip- over the 10-year lifetime of the network represents ment cost was $37,000. $280,000. Again, the NMS modified its expectations and the annual Operations and Maintenance Costs. The reduction in the network size generated flow- through reductions in other areas: Fortunately, the reduction of 4 stations also generated some cascading savings in the annual operations and p With a reduced station count of 17, there was a maintenance, including the following reductions: $40,000 savings in the Cost of Civil Works. p The reduction in stations reduced the supplier costs p Reductions in the costs of logistics, consumables, where the NMS saved 8 days of installation costs and calibrations, and cell phone data charges Total Cost of Ownership Example Calculation    289 p Reductions in the number of local staff required to These cost saving measures balanced the total annual carry out basic maintenance cost of operations and maintenance to $143,848, slight- p One field technician who would carry out mainte- ly under the $150,000 budget. nance half-time was due to retire. The NMS decided not to hire a replacement, for an additional cost sav- In this example, the NMS went through the process of devel- ings of $5,200. It was felt that the remaining techni- oping the Total Cost of Ownership twice. The first iteration cians could handle the maintenance of the reduced showed that the NMS did not have either the capital budget network. The NMS also trained an internal support to meet their original design needs or the operating budget to staff member on field maintenance as a back-up. sustain AWS network operations over its lifetime of 10 years. p By going to a cloud software–hosted solution, the Determining where the NMS could modify the size of the net- NMS did add $20,000 annually, but it saw an addi- work and looking at alternative ways to deliver the data in tional savings of $12,000 as there was no longer a the second iteration provided a solution that met their de- need for service-level agreement to support the IPS. sign goals and was sustainable financially. The process of the There was a savings of $5,000 as an annual software TCOSF exercise aided the NMS in planning the network from license was included in the cloud data hosted pack- equipment acquisition to human resources requirements and age. Additionally, the cost for utilities was reduced operational plans, allowing it to ensure consistent data flow by $7,000 as mains power was no longer required to supporting their timely forecast needs over the lifetime of 10 support a server room for network operations as well years. as air conditioning. 290    Total Cost of Ownership Example Calculation FIGURE 3.9.2  TCOSF Example Calculation (1 and 2 Combined) Total Cost of Ownership Summary Form Project: Example 1 Calculation Date: 1/20/22 Capital Budget: $1,100,000 Currency: US dollars Operation Budget: $150,000 per year                   Initial Investment (Capital Costs)     Line Description Example 1 Example 2 Difference 1 Cost of Capital Equipment $668,750 $518,750 –$150,000 2 Cost of Civil Works $226,250 $186,250 –$40,000 3 Total Cost of Supplier Services $204,700 $179,900 –$24,800 4 NMS Costs to Install Operational System $179,000 $95,000 –$84,000 5 Total Initial Investment (Capital Costs) $1,278,700 $979,900 –$298,800 Cost of Annual Operations     6 Business Costs $0 $0 $0 7 Administrative Costs $26,000 $15,600 –$10,400 8 Other Operating Costs $36,305 $31,245 –$5,060 9 Total Cost of Annual Operations $62,305 $46,845 –$15,460 Cost of Annual Maintenance     10 Preventive Maintenance $100,550 $84,350 –$16,200 11 Corrective Maintenance $15,100 $12,653 –$2,448 12 Adaptive Maintenance $0 $0 $0 13 Total Cost of Annual Maintenance $115,650 $97,003 –$18,648 Total Cost of Annual Operations 14 $177,955 $143,848 –$34,108 and Maintenance 15 Lifetime of System (Years) 10 10   Total Cost of Annual Operations and 16 $1,779,550 $1,438,475 –$341,075 Maintenance over Lifetime Life-Cycle Costs over Lifetime       17 Total Life-Cycle Costs $169,275 $108,275 –$61,000 18 Total Cost of Ownership over Lifetime $3,227,525 $2,526,650 –$700,875