Water Security and Drought Resilience in the South of Angola Aleix Serrat-Capdevila Natalia Limones Javier Marzo-Artigas Marcus Wijnen Bruno Petrucci About the Water Global Practice Launched in 2014, the World Bank Group’s Water Global Practice brings together financing, knowledge, and implementation in one platform. By combining the Bank’s global knowledge with country investments, this model generates more firepower for transformational solutions to help countries grow sustainably. Please visit us at www.worldbank.org/water or follow us on Twitter at @WorldBankWater. About the Korea Green Growth Trust Fund In partnership with the World Bank, the Korea Green Growth Trust Fund (KGGTF) was established in 2011 to strengthen and expand the World Bank’s global green growth portfolio by tapping expertise from Korea’s successful green growth experience and investment through public and private resources. Its central approach is to support the World Bank and its clients to operationalize inclusive green growth initiatives, strategies, and investments. Please visit us at www.kgreengrowthpartnership.org About GWSP This publication received the support of the Global Water Security & Sanitation Partnership (GWSP). GWSP is a multidonor trust fund administered by the World Bank’s Water Global Practice and supported by Australia’s Department of Foreign Affairs and Trade, Austria’s Federal Ministry of Finance, the Bill & Melinda Gates Foundation, Denmark’s Ministry of Foreign Affairs, the Netherlands’ Ministry of Foreign Affairs, the Swedish International Development Cooperation Agency, Switzerland’s State Secretariat for Economic Affairs, the Swiss Agency for Development and Cooperation, and the U.S. Agency for International Development. Please visit us at www.worldbank.org/gwsp or follow us on Twitter at @TheGwsp. WATER SECURITY AND DROUGHT RESILIENCE IN THE SOUTH OF ANGOLA Aleix Serrat-Capdevila, Natalia Limones, Javier Marzo-­Artigas, Marcus Wijnen, and Bruno Petrucci © 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 This work is a product of the staff of The World Bank with external contributions. The findings, interpre- tations, 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 judg- ment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Please cite the work as follows: Serrat-Capdevila, Aleix, Natalia Limones, Javier Marzo-Artigas, Marcus Wijnen, and Bruno Petrucci. 2022. “Water Security and Drought Resilience in the South of Angola.” World Bank, Washington, DC. Any queries on rights and licenses, including subsidiary rights, 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 Front and back cover photos: Aleix Serrat-Capdevila/World Bank. All interior photos were provided by the authoring team. Cover design: Will Kemp. Contents Preface: Drought and Resilience ix Acknowledgments xi Abbreviations xiii Executive Summary 1 How Can Angola Build Drought Resilience and Water Security? 2 The Value of Data and Continued Analysis 4 Chapter 1  Introduction 5 Drought in Angola 5 Objectives, Audience, and Approach 5 Country Background 6 The Study Area 7 Notes 9 Chapter 2  Characterizing Drought Risk in the South of Angola 11 Vulnerability and Exposure to Drought in the Region 11 Drought Characterization and Direct Hydrologic Impacts 19 Expected Impacts: Overlaying Vulnerability Conditions and the Severity of the Drought Event 23 Notes 24 Chapter 3  Understanding Impacts Observed on the Ground 25 Gathering Feedback From Local Communities 25 Main Issues Observed in the Field 30 Notes 35 Chapter 4  Building Drought Resilience in the South of Angola: Assessing Options for Rural Water Infrastructure Investments 37 Selection of Most Suitable Infrastructure Options 37 Case Study 1: Options to Increase Water Security in the Cuvelai Basin 39 Case Study 2: Identification of River Sections with Potential for the Construction of Small-Scale Managed Aquifer Recharge In Namibe Province 73 Infrastructure Costing 92 Identification of Risks Associated with Building Small-Scale Water Supply Infrastructure 102 Notes 103 Water Security and Drought Resilience in the South of Angola iii Chapter 5  Conclusions and Recommendations: Building Drought Resilience in the South of Angola 105 Recommendation 1: Invest in Information and Knowledge 106 Recommendation 2: Invest in Rural Infrastructure 108 Recommendation 3: Invest in Institutions and People 110 Bibliography 115 Glossary 119 Appendix A  Main Aquifers in the Cuvelai-Etosha Basin 123 Appendix B  Water Balances for Below-Average Rainfall Years 125 Appendix C  Water Harvesting Infrastructure Designs 127 Appendix D  Topographic and Geognostic Surveys 129 Appendix E  Mission Schedule 133 Figures 1.1. Some Facts about Angola’s Country Background 8 2.1. Commune Scores and Percentiles for Indicator 1: Water Source Unreliability 14 2.2. Commune Scores and Percentiles for Indicator 2: Water Source Unsafeness 14 2.3. Commune Scores and Percentiles for Indicator 3: Water Financial Dependence 15 2.4. Temporal Graphs Depicting the Evolution of the Drought Exceedance Probability Index in 21 Points of the Region, and the Provincial Averages 21 2.5. Average Monthly Precipitation Minus Actual Evapotranspiration in the Region for 1994–2019 22 2.6. Difference Between the Average Monthly Values of Precipitation Minus Actual Evapotranspiration in the Period 2012–19, and Average Monthly Values Calculated for the Entire Period 1994–2019 23 4.1. Infrastructure “Decision-Tree” for Finding the most Adequate Small-Scale Rural Water Supply Infrastructure to be Explored Further in Each Priority Area, Based on its Natural Conditions 38 4.2. North-South Cross-Section of the Western Part of the Cuvelai Basin, from the Headwaters till Lake Oponono 43 4.3. Monthly Rainfall in the Upper Catchment, Angola Lowland, and Namibia Lowland Subbasins, 2003–19 44 4.4. SPEI Drought Index at Different Time Scales for Ombala Yo Mungu 45 4.5. Precipitation in Upper Catchment 54 iv Water Security and Drought Resilience in the South of Angola 4.6. Mean Monthly Precipitation (P), Actual Evapotranspiration (ETa), and P—ETa Values for the Three Cuvelai Subbasins 57 4.7. Average Annual Precipitation (P) for the Three Combined Cuvelai Subbasins, 2003–18 58 4.8. Correlation between Annual P and Annual ETa for the Angola Lowland Subbasin, 2003–18 58 4.9. Annual Exceedance Probability of Annual Precipitation Values, Lowland Angola 61 4.10. Annual Exceedance Probability of “Wet Season Excess P-ETa” Values for Lowland Angola 62 4.11. Annual Exceedance Probability “Wet Season P-ETa Excess” Values for Lowland Angola 63 4.12. Correlation Between “Wet-Season P-ETa Excess” and “Wet-Season P” 67 4.13. Wet Season Rainfall P (October–March) and Calculated Wet Season Net Rainfall Pnet at Omabala Yo Mungu, for the Rainy Seasons 68 4.14. Daily Rainfall and Estimated Net Daily Rainfall in Ombala Yo Mungu During the Rainy Seasons of 1998–2019 68 4.15. Cumulative Rainfall and Net Rainfall in Ombala Yo Mungu for an Average Rainy Season 69 4.16. Filling Curve of Standard Chimpacas and Cisterns for Different Impluvium Sizes During a Rainy Season with Average Rainfall (left) and Below-Average Rainfall (AEP = 60 percent) (right) 69 4.17. Proposed Dam Profile and Rough Bedrock Pattern 85 4.18. Dam Cross-Sections 85 4.19. Proposed Dam Profile and Rough Bedrock Pattern 87 4.20. Dam Cross-Sections 87 4.21. Dam Profile and Rough Bedrock Pattern 88 4.22. Dam Cross-Sections Show Two Different Patterns Due to the Absence/Presence of a Sandy Cover of the Granite Basement 90 5.1. Decision Process for the Prioritization and Selection of Beneficiary Communes for Water Resources Investments 111 C.1. Schematic Cross-Section of a Chimpaca 127 C.2. Schemtic Layout of a Chimpaca 127 C.3. Schematic Cross-Section of a Cistern 128 C.4. Schematic Layout of a Cistern 128 Maps 1.1. The Five Provinces of this Study and their Average Monthly Rainfall 7 2.1. Population Distribution Maps Showing the Cunene, Huíla, and Namibe Communes with Lower Population Densities 18 3.1. Communes Where Detailed Field Work was Undertaken (Including Interviews and Discussions in Municipality Government Offices 26 4.1. Maps of Cunene Province (left) and the Cuvelai Transboundary River Basin (right) 40 4.2. Subbasins of the Cuvelai Basin used in this Study 43 4.3. Location Map of Sites Across the Cuvelai Basin Visited during the Field Survey 49 4.4. Satellite-Derived Meteorological Grid 53 4.5. Detail from the Hydrogeological Map of Africa at 1/10 M Scale Showing Modeled Groundwater Recharge Values 55 Water Security and Drought Resilience in the South of Angola v 4.6. Map of Bibala and Camucuio Municipalities in Namibe Province 74 4.7. Zoom to the Giraul Basin Area, with Potential Sites for Managed Aquifer Recharge Solutions Labelled “GB” 74 4.8. Zoom to the Chingo area, with potential sites for managed aquifer recharge solutions labelled “CH” 75 Photos 3.1. Volanta Pump (Left); A Solar-Powered Water Point (Right) 30 3.2. A Cacimba Excavated In A Riverbed 31 3.3. Truck Used in Onkokwa, Provided by the National Service of Civil Protection and Firemen 32 3.4. Silted Chimpaca with Very Limited Remaining Live Storage (Left); A Cacimba (Open Well) Dug Inside a Silted Chimpaca to Find a Small Quantity of Water (Right) 33 3.5. A Drip Irrigation Plot in Bentiaba 35 4.1. Google Earth View of the Study Area Landscape Units in the Angolan Part of the Cuvelai Basin 40 4.2. Ground View of the Cuvelai Landscape 41 4.3. Geological Map of the Cuvelai-Etosha Basin (left); Quarry Near Ombala Yo Mungu (right) 42 4.4. Aerial and Side Views of a Cuvelai Chimpaca 46 4.5. Typical Cacimba in a Kimbo Near Ohenghali, Mongua Commune (left); Open Well at Chiulo, Mukope Commune (right) 47 4.6. Borehole with Volante Handpump Near Onjiva (left); Borehole Equipped with Solar Pump and Storage Reservoir at Omambodi, Mongua Commune (right) 47 4.7. Soil and Subsurface Geology Observations Near Ombala Yo Mungu Sede (left); Groundwater Level Measurement Near Mongua Sede (middle); and Water Quality Measurement in Naulila Commune (right) 51 4.8. Main Components of the Cuvelai Subbasins’ Water Balance Considered in this Study 54 4.9. Borehole Equipped with Solar Pump and 5 m3 Storage Reservoir at Ondjiva Sede 64 4.10. Borehole Equipped with Volante Pump for Community and Husbandry Water Supply at Omulova, Namakunde Commune 65 4.11. Active Channel of the Cuvelai Drainage System in the Cuvelai Lowlands (left); Kimbo Near Ohenghali, Mongua Commune, with Field Prepared for Growing Rainfed Crops During the Upcoming Rainy Season (right) 66 4.12. Sediment trap protecting cistern (“berkat”) in Somaliland (left two images) and Rock Filter Sand Sediment Trap for two Different Cisterns in Djibouti (right two images) 71 4.13. Artesian Deep Borehole at Okashana (left) and Stable Isotope Concentrations in Groundwater in the Namibian Part of the Cuvelai-Etosha Basin (right) 72 4.14. View of the Possible Dam Layout at Site GB2 from the Right Bank 78 4.15. Possible Dam Layout, Throwback, and Geological Features at Dam Site GB2 79 4.16. View of the Possible Dam Layout at Site GB3 from the Right Bank 80 4.17. Possible Dam Layout, Throwback, and Geological Features at Dam Site GB3 80 4.18. Site CHB1 82 4.19. New Proposed Dam (blue line) and Throwback Area 83 4.20. Details of the GB3 Dam Site Area 88 vi Water Security and Drought Resilience in the South of Angola 4.21. Details of the Dam Area 90 D.1. Site GB2 Survey Area 130 D.2. Site GB3 Survey Area 131 D.3. Site CHB1 Survey Area 132 Tables 2.1. Indicators Used to Characterize Sensitivity and Vulnerability to Drought In the Study Region from a Domestic Water Supply Point of View 12 2.2. Ranking of the Communes of the Five Studied Provinces Based on their Vulnerability to Drought 15 3.1. List Of Impacted Municipalities Based on the Information Provided by Key Informants and Degree of Adjustment to the Results of this Study 28 3.2. Impacts Reported for Each of the Sampled Priority Communes 29 4.1. Range of Electrical Conductivity Values for Different Types of Water Points 50 4.2. Tentative Water Balances for the Three Cuvelai Subbasins, 2003–18 59 4.3. Annual Exceedance Probabilities for Different Climate Parameters (Upper Catchment) 62 4.4. Surface Water Harvesting Potential for the Angola Lowland Subbasin, at Different AEP Values 63 4.5. Required Impluvium Sizes for Standard and Large Chimpacas and Cisterns, Depending on the Amount of Net Seasonal Rainfall Received (October to March) 70 4.6. Locations of the Selected Points and Hydrological Basin Parameters 75 4.7. Stream Flow and Dam Parameters 84 4.8. Stream Flow and Dam Parameters 86 4.9. Stream Flow and Dam Parameters 89 4.10. Cost Estimates for a Standard Chimpaca and a Strategic Large Chimpaca 95 4.11. Cost Estimates for Standard and Large Cisterns 96 4.12. Costs Analysis for the GB2 Site Sand Dam 97 4.13. Costs Analysis for the GB3 Site Sand Dam 99 4.14. Costs Analysis for the CH1 Site Sand Dam 100 4.15. Overview of Cost Estimates for Different Types of Rural Water Supply Infrastructure 101 A.1. Classification and Main Characteristics of Aquifers of the Kalahari Sequence in the Namibian Part of the Cuvelai-Etosha Basin 123 B.1. Tentative Water Balances for the Three Cuvelai Subbasins Drawn with P, ETa, “Wet Season P-ETa Excess,” and “Dry Season P-ETa Déficit” Values at 60 Percent, 70 Percent, 80 Percent, and 90 Percent Annual Exceedance Probabilities 125 Water Security and Drought Resilience in the South of Angola vii © Marcus Wijnen/World Bank Preface: Drought and Resilience The effects of drought are modulated by the presence—or the absence—of appropriate policies and insti- tutional efforts, adequate infrastructure and the degree of its functionality, among other factors. As the major droughts of the past years have proven in many countries, institutions—formal or informal—are a key determinant of drought resilience. Given a meteorological anomaly, vulnerability to drought will depend on a range of structural conditions, the management context, the state of routine maintenance and corresponding budget allocations, the presence of information systems and flows, and so forth. These factors in turn condition communities’ preparedness and capacity to efficiently respond to drought when it hits. Resilience depends on the interplay of the natural environment, built infrastructure, institutions, and human behavior. Resilience is about having options, and buffers. This study responds to a request from the Government of Angola in late 2018 to help it transition from a reactive response mode to a proactive and resilience-focused model of dealing with drought in the South of Angola. Toward this end, the study presents a practical approach and actionable proposals to support the Government of Angola in its shift towards fostering climate resilience. The work presented here aims at providing a better understanding of southern Angola’s vulnerability to drought, focusing on the structural conditions of access to water as well as the governance context. Based on this improved understanding, the study informs the design of key solutions, including strengthening governance to promote resilience, investing in the sustainable development of water resources, and a strategy for prioritizing related interventions across the region. This study focuses on rural communities across the South, and specifically on helping to find ade- quate water-access solutions for them. Its approach and findings complement parallel, ongoing efforts by the Government of Angola to increase water security through the construction of dams and water transfers in the Cuvelai, Cunene, and Cubango basins, as well as dams in the Namibe coastal basins. This study targets those rural communities that will be beyond the reach of these government efforts. Water Security and Drought Resilience in the South of Angola ix © Marcus Wijnen/World Bank Acknowledgments The “Drought Resilience in the South of Angola” study of the World Bank was launched in February 2019 under the supervision of task team leader Aleix Serrat-Capdevila, Senior Water Resources Management Specialist. The work presented in this report was carried out with a team of four specialists: Natalia Limones Rodríguez, Drought Resilience and Climate Data Specialist; Javier Marzo Artigas, Statistics and Geographic Information Systems (GIS) Data Specialist; Marcus Wijnen, Senior Groundwater Specialist; and Bruno Petrucci, Senior Water Harvesting Specialist. This study would not have been possible without the broad support and collaboration of government institutions, provincial and municipal administrations, and a range of development partners. The team received valuable support and information from a number of partners acknowledged below. The field sup- port of Evanilton Pires (Instituto de Tundavala, IT) and Marco Paulo Carlos (Universidade Agostinho Neto, UAN), thanks to the facilitation of professors Carlos Ribeiro (IT), Gabriela J.P.T. Pires (UAN), and Gabriel Luis Miguel (UAN), is gratefully acknowledged, as well as that of Cecile Wijnen and Melissa Aguirre. From the Government of Angola, the team gratefully acknowledges the institutional support of the Instituto Nacional de Recursos Hídricos (INRH) and the Gabinete para Administração das Bacias Hidrográficas do Cunene, Cubango e Cuvelai (GABHIC), through their respective teams: Manuel Quintino, Francisco Quipuco, Emmanuel Ferreira, and Narciso Ambrosio (INRH); and Carolino Mendes, Carlos Andrade, Falco Kiowa, and Jose Kapu (GABHIC), who also provided valuable assistance in the field, as well as the institutional support from the Direção Nacional das Aguas through Elsa Ramos and Antonio Quaresma, the valuable comments from Aleixandrina Pires and the support from Guillermo Távara (Executive Coordinator of the PDISA2 Project). We would also like to acknowledge the essential support received from the Fundo de Apoio Social (FAS) during field missions, through Santinho Figueira, Pedro Bell, Manuel Esteves, Vicente Joao, and their extended teams. We are thankful for valuable insights and assistance in the field from Ditutala Lucas Simao (Coordinator, Programa Nacional de Bovinocultura), as well as for valuable feedback from Segundo Comandante José Horacio da Silva, the Director Adjunto Comandante Edson Fernando of the Servico Nacional de Proteção Civil e Bombeiros; and from Hermenegildo Keane dos Santos, National Director of Hidráulica Agrícola e Engenharia Rural. From development partners, the team wants to gratefully acknowledge Tomas Lopez de Bufala and Edson Monteiro from the United Nations Children’s Fund for their valuable water point database contri- butions, information, and valuable insights on the water sector in Angola; John Mendelsohn of RAISON for his digital population database of Southwest Angola; as well as the following colleagues, who pro- vided valuable context, information, and support: Allan Cain (Development Workshop); Keita Sugimoto, Goeth Schroth, and Henrik Larsen (United Nations Development Programme); Mateo Tonini and Vitor Serrano (Food and Nutrition Security and Resilience Strengthening [FRESAN] project); and Yuzo Kitamoto and Yoshihiro Miyamoto (Japan International Cooperation Agency, JICA). From the World Bank, we would like to thank peer reviewers Erwin De Nys (Program Leader), Eileen Burke (Lead Water Resource Management Specialist), Tesfaye Bekalu (Senior Water Supply and Sanitation Specialist) and Tomas Lopez de Bufala (Head of WASH, UNICEF Angola) for their valuable Water Security and Drought Resilience in the South of Angola xi insights, as well as acknowledge the overall guidance of Maria Angelica Sotomayor Araujo (Practice Manager) and the support of Olivier Lambert (Country Manager), Ana Maria Carvalho (Senior Operations Specialist), Benjamim Mutti, Jesus Lino and the entire Angola Country Management Unit (CMU). Fayre Makeig and Erin Ann Barrett provided great editorial support. Camilo Lombana Cordoba, Marco Antonio Aguero, and Luis Alberto Andres were valuable sources of insight and support in Angola. This work was funded by the Global Water Security and Sanitation Partnership, with additional funding support from the Water Expert Facility and the Korean Green Growth Trust Fund. Finally, we would like to thank Lucrécio A.M. da Costa, State Secretary for Water (Ministério da Energia e Águas, MINEA), for his constant support on this front, and His Excellency João Baptista Borges, Minister of Energy and Water, for his initial request to pursue this study and for his efforts to shift the government’s mindset to drought in Angola from reactive to proactive. xii Water Security and Drought Resilience in the South of Angola Abbreviations ASA Advisory Services and Analytics AEP Annual exceedance probability CEB Cuvelai Etosha Basin CH/CHB Chingo area (basin) DNA Direccao Nacional das Aguas DEM Digital Elevation Model DEPI Drought Exceedance Probability Index EPAS Empresas Provinciais de Agua e Saneamento EC electrical conductivity ET evapotranspiration ETa actual evapotranspiration FAO Food and Agriculture Organization FAS Fundo de Apoio Social FG fresh granite FGD Focus Group Discussion FRESAN Fortalecimento da Resiliência e da Segurança Alimentar e Nutricional em Angola GABHIC Gabinete para Administração das Bacias Hidrográficas do Cunene, Cubango e Cuvelai GAS Grupos de Agua e Saneamento GB Giraul Basin GIS Geographic Information System GLEAM Global Land Evaporation Amsterdam Model GPS Global Positioning System IGRAC UNESCO International Groundwater Resources Assessment Centre IND indicator INRH Instituto Nacional de Recursos Hídricos de Angola MAR managed aquifer recharge O&M operation and maintenance P precipitation, rainfall PDNA Post-Disaster Needs Assessment Pnet net rainfall SbSD subsurface dam SD sand dam SISAS Sistema de Informacao do Sector de Agua e Saneamento SPEI Standardized Precipitation-Evapotranspiration Index SPI Standardized Precipitation Index TDS total dissolved solids TMPA TRMM Multisatellite Precipitation Analysis TRMM Tropical Rainfall Measuring Mission Water Security and Drought Resilience in the South of Angola xiii UNDP United Nations Development Programme UNICEF United Nations Children’s Fund WASH water supply, sanitation and hygiene WHO World Health Organization xiv Water Security and Drought Resilience in the South of Angola Executive Summary This study responds to a request from His Excellency, the Minister of Energy and Water of Angola, to help the Government of Angola transition from a reactive response mode to a proactive and resilient approach in the face of drought and water insecurity in the South of Angola. Thus, this first World Bank report on drought in Angola aims to provide a practical approach and actionable measures to support the Government of Angola in its paradigm shift toward drought and climate resilience. To provide a clear picture of the problem and its solutions, the report focuses on understanding the structural causes of water access vul- nerability across the region, informing the prioritization of investments and the selection of water supply solutions at the community level, and providing a strategy for the water sector to build resilience to drou- ghts and climate variability, which is also likely to intensify with global warming. The effects of drought are modulated by the actions of institutions and their policies (or lack thereof), adequate infrastructure, and the degree of its functionality among other factors. The center, and espe- cially the South, of Angola, suffered a devastatingly severe drought from the 2012/2013 rainy season until the 2019/2020 rainy season. The drought affected mainly the provinces of Cunene, Namibe, and Huíla, and, to a lesser extent, Benguela and Cuando Cubango. The impacts of the long and severe drou- ght in the South were exacerbated by the widespread disrepair of water points and a lack of drought-re- silient infrastructure. This study finds four main causes of drought vulnerability in the South. These are (1) a lack of informa- tion about the state of water points and deficient knowledge about the potential of water resources; (2) insufficient investments in water resources at the community level (resulting in a lack of water supply options; (3) a lack of efficient mechanisms to repair, maintain, and guarantee the functionality of infras- tructure at the community level; and (4) limited institutional capacity at various levels to prepare for, mitigate, and respond to droughts. A detailed regional view of priorities is needed. The Government of Angola is mobilizing funding for drought interventions (including new dams and water transfers in areas of the South) with the support of several donors, but a comprehensive spatial view of specific needs is lacking. Why are some rural areas particularly affected by drought, and where are they located? Which investments might reduce their vulnerability to drought? How can communities and investments be prioritized? A data-based approach is presented and implemented in this study to help prioritize locations for com- munity-level interventions across the southern Angola region. This study uses satellite data to characte- rize the meteorological drought, and census data (INE 2014) for the 130 communes (comunas) of five provinces—Huíla, Benguela, Namibe, Cunene, and Cuando Cubango—to characterize access to water at the commune level. Three indicators (measuring the degree to which water is unreliable, unsafe, and requires a financial transaction for access), in combination with population density data, allowed for the identification of vulnerable and exposed communes. The resulting map was then overlapped with the spatial distribution of the area’s drought hazard (based on a study of drought intensity and duration Water Security and Drought Resilience in the South of Angola 1 from 2012), to identify priority communes. Feedback from local partners and data collected during field surveys were used to perfect and validate the approach and its findings, and organize a prioritized list of communes requiring intervention across the vast region. A framework to select appropriate water resources investments for each community is presented, based on hydrologic, geomorphology, and hydrogeology information and field missions. Where available, groundwater development is given priority since it is relatively unlikely to be affected by drought, though in many areas, other alternatives may be the only option. While an exhaustive hydrogeologic study is needed to define the groundwater potential across the region, this report presents an assess- ment of two types of water resource investments: (1) surface water harvesting through improved chimpacas and cisterns, and (2) sand dams and managed small-scale aquifer recharge. A hydrologic and hydrogeologic assessment of the Cuvelai basin in Cunene province confirmed that sur- face water represents the largest available resource, and that its harvesting can be expanded. The study recommends strengthening the planning and design of water harvesting infrastructure, building cisterns for water for human consumption (smaller and deeper than chimpacas, to reduce evaporation losses and contamination), and separate chimpacas for animal water supply. The costs of standard cisterns (500 cubic meters, m3) and chimpacas (12,000 m3) are estimated at US$32,000 and US$160,000 respectively. The potential for sand dams and managed aquifer recharge in Namibe province was evaluated and a number of sites were identified and characterized in the Giraul basin and Chingo and Camucuio areas. Sand dams offer an additional water supply that cover, for a minimum of 3 months to more than 6 months, a population of at least 500 persons with 500 cows and 1,000 sheep and goats, and the poten- tial to support small vegetable gardens. Preliminary cost estimates range from US$50,000 to US$70,000 for each dam. A detailed topographic and geognostic survey is needed to finalize dam designs, and is planned for the next phase of the study. To provide increased water security, especially during extended droughts, the groundwater potential should be further explored. This requires the systematic collection of information on boreholes, aquifer characteristics, groundwater levels, and groundwater quality. Medium to deep groundwater (at depths between 5 m and 200 m) is expected to have limited potential for development and use, and the presence of deep groundwater resources needs to be confirmed. Considering that current groundwater recharge rates are very low, deep groundwater resources may have limited long-term potential. How Can Angola Build Drought Resilience and Water Security? Drought impacts are modulated by the systems in place. Drought impacts are an expression of the degree of risk, which depends on vulnerabilities modulated by institutions, existing systems, and investments. The study found that drought impacts were particularly severe because of a lack of (1) information and communication regarding the state of water points and water resources; (2) reliable access to water due to insufficient investments at the community level; and (3) capacity to systematica- lly repair water points and prepare for droughts, from the community level to provincial and national levels. Thus, this study makes the following recommendations: 2 Water Security and Drought Resilience in the South of Angola •• Invest in information and knowledge. Address the lack of information on water points by strengthe- ning the quality and frequency of their monitoring. The functionality of water points in the South of Angola is severely hampered by a lack of adequate information on the status of water infrastruc- ture and water resources. Establish mechanisms to update information, and to manage it, share it, and use it. Operationalize the Sistema de Informacao do Sector de Agua e Saneamento (SISAS) by connecting it to systematic reporting from the field administrations, as well as to repair and main- tenance actions. Improve infrastructure planning and design by investing in water resources knowledge—groundwater studies, surface water and water balance studies—and continuous moni- toring mechanisms. Involve universities and share data. •• Invest in community-level infrastructure. Implement an integrated program of physical investments to develop and mobilize water resources reliably and sustainably for and with rural communities. These investments can include large diameter wells, boreholes with solar pumps, storage tanks, drinking taps and basins for cattle, boreholes with a small desal unit in key locations, deep boreho- les, runoff harvesting sand dams, chimpacas and cisterns, bulk water through piped supply, old weirs and small dam rehabilitation, managed aquifer recharge, and watershed storage. Make deci- sions based on data and studies. Continuously update the prioritization of vulnerable communes based on the nature of ongoing investments and interventions from partners. •• Resilience means having more than one option. It is essential to plan for redundancies and buffers in the form of storage. Plan with a regional strategic view, focusing on communities but also on trans- humance corridors. •• Invest in people and institutions. Investments and resources are needed to guarantee minimal levels of institutional capacity at all levels, to ensure water security in the region and systematically maintain water supply systems. It is essential to build institutional preparedness and capacity to reduce drought impacts in the form of devastated livelihoods, lives lost, and stunted children. Strengthen and scale up community-level organizations (i.e., Modelo de Gestao Comunitaria del Agua, MoGeCA) and regional water point governance (e.g., foster an enabling environment for the MoGeCA). Strengthen the Empresas Provinciais de Agua e Saneamento, as well as the capacity of basin organizations such as the Gabinete para a Administração das Bacias Hidrográficas do Cunene, Cubango e Cuvelai, and national agencies such as Direção Nacional das Aguas and Instituto Nacional de Recursos Hídricos. Invest in training and in the professionalization of the private sector (i.e., drilling best practices, geophysics studies). •• Strengthen government institutions. Specifically, it is important to build capacity to (1) gather and manage relevant information on climate, hydrology, and hydrogeology; (2) strengthen information-to-action mechanisms for drought preparedness and drought response programs and ­ associated training; (3) implement and operationalize a continuous monitoring approach for water points and other infrastructure; (4) operationalize water resources and water infrastructure databases and dissemination mechanisms; (5) develop financial resilience and efficient budgetary ­ Water Security and Drought Resilience in the South of Angola 3 management at all levels to guarantee basic functionality of water points and other essential servi- ces; (6) develop a strategic vision for storage investments, combining the use of dams and reser- voirs, aquifers, and watershed storage; (7) integrate no-regrets flood planning and flood control infrastructure investments, aiming at synergies between flood mitigation and storage options; and (8) develop basin master plans for the Namibe coastal basins, as well as support the updating of the Cunene River Basin Plan, to ensure a knowledge-based foundation for all actions going forward. While we could estimate that with US$200 million the story of drought in the South of Angola could be rewritten, we have to understand that building things is only half of the solution. The other half, which poses the biggest obstacle to water security and climate resilience in Angola, is maintaining and opera- ting them: strengthening systems and institutions that are fit for purpose to fulfill their operational mandates, and growing staff capacity to serve the population and be held accountable. The Value of Data and Continued Analysis This report is a stepping stone and sets the stage for the next steps. This study builds on past govern- ment investments in data collection, such as the National Census (INE 2014), highlighting the impor- tance of having good data to make well informed analysis and evidence-based decisions. Specific to drought, it also builds on the Post Disaster Needs Assessment (UNDP 2016), deepening the analysis with many additional sources of data and field observations, and increasing the resolution of analysis from the provincial to the commune level. Continuing policy dialogue and analytical work with the Government of Angola are essential going forward and will include the following: (1) an analysis of options for the systematic monitoring of water points and information mechanisms within government hierarchies (i.e., to feed the SISAS with periodic information that may be used to trigger repairs and interventions); (2) a deeper analysis of the linkages between community-based management and municipal and provincial governance, budgeting for maintenance and repairs, and the role of the private sector; (3) efforts to increase the availability and accessibility of hydrogeologic information across the region and finding mechanisms for systematic monitoring; (4) a scale up of the characterization of investments at the community level across the region; (5) a mapping of agriculture and deeper understanding of transhumance dynamics; and (6) ongoing contributions to the training and capacity building of government agencies, universities, and the private sector. 4 Water Security and Drought Resilience in the South of Angola Chapter 1 Introduction Drought in Angola The South of Angola suffered a severe drought from the time of the weak rainy season in 2012/2013 until the arrival of the 2019 rainy season, affecting the provinces of Cunene, Namibe, and Huíla, and signifi- cant parts of Benguela and Cuando Cubango. Sporadic rains during this period brought some relief but were not enough to initiate a recovery. Some areas in southern Angola, as well as other parts of southern Africa, recorded the driest season in 35 years in 2015/2016, a peak of severity linked to the El Niño effect. From 2013 to 2016, between 76 percent and 94 percent of the populations of the provinces of Namibe, Cunene, and Cuando Cubango were affected by the drought. According to the Post-Disaster Needs Assessment (PDNA) performed under the supervision of the United Nations Development Programme (UNDP), by 2016 the drought had affected 1,139,064 people in Cunene, Huíla, and Namibe. Of this total, half were in Cunene. The economic impacts across all sectors are estimated at over US$749 million in the three most affected provinces, with the agriculture, livestock, and fisheries sector being by far the most damaged. Apart from the directly monetizable losses, the PDNA acknowledges a rising trend in malnutrition, family abandonment, domestic violence, charcoal production, and deforestation. The PDNA reports that about 80 percent of existing boreholes were nonfunctional in 2016 due to water scarcity and disrepair (approxi- mately 2,400 boreholes were damaged) in the three provinces. Drought in the South of Angola is an old problem, with severe droughts in the 1990s and long before. While droughts will continue to occur periodically in the region, it is likely that the effects of global warming may increase both the frequency and the magnitude of droughts and floods in the future. ­ Thus, building climate resilience by strengthening institutions and designing investments that can manage both floods and droughts is essential. For this, the consolidation of integrated water resource management frameworks, integrated planning, water resource information and knowledge, and their multisectoral use and allocation are much needed in the path to navigate water insecurity. Objectives, Audience, and Approach The objectives of this study are to map and characterize the drought vulnerabilities of communities in the South of Angola, focusing on their conditions of access to water, and to explore the feasibility of sustainable rural water supply options that would enhance these communities’ drought resilience. By addressing these topics, the study is supporting the Government of Angola in its efforts to develop a program of interventions to enhance water security in the South of Angola. The information presented here will also help the government and development partners to better coordinate the actions of various donors in the South of the country, maximizing the efficiency of development aid. Water Security and Drought Resilience in the South of Angola 5 This report is intended for practitioners from government agencies and development organizations that are working to combat vulnerabilities and increase resilience to drought in Angola and beyond. The work was organized in three stages, reflected in the sequence of chapters in this report. •• In February 2019 to April 2019, desk-based research was conducted to characterize the drought risk, and analyze the exposure and vulnerability of human populations in southern Angola to drought. Based on this assessment, priority areas for follow-up work were identified. •• In April 2019, the findings of the desk-based research were tested and verified on the ground, through field trips and consultations with local experts and stakeholders in the identified priority areas. •• In July 2019 to early 2020, detailed diagnosis and prefeasibility assessment of infrastructure options for water security were conducted in a set of priority areas experiencing high drought risk. Country Background Angola is a beautiful country of varied landscapes, climates, and ecozones covering 1.25 million square kilometers (km2) and with a population of about 30 million people, of which 8 million reside in the capi- tal, Luanda. Following its independence from Portugal in 1975 and a long civil war (1975–2002), efforts have been made to repair and improve infrastructure and strengthen public institutions. With an oil boom that peaked in 2010–12, and whose benefits generated extreme inequality, followed by an oil bust in 2015, Angola’s economy is still highly dependent on oil exports. In the midst of significant debt and liquidity problems, Angola is making efforts to diversify the economy, a task made even more difficult in the current context of global economic recession due to the COVID-19 pandemic. In terms of access to water and sanitation, it is estimated that 62 percent of Angola’s population have access to an improved drinking water source, and 70 percent to an improved sanitation facility, but vast differences exist between urban and rural contexts and between the rich and the poor.1 While basic access2 to drinking water in urban areas reaches 70 percent (90 percent for sanitation), in rural areas it averages under 37 percent (27 percent for sanitation). For the population below the median wealth level and for the poorest, water access falls to 30 percent and 15 percent, respectively (these figures are 33 percent and 9 percent for sanitation). Drinking water coverage has seen little improvement in the past decade (partly due to population growth and rural-to-urban migration), and 6 million continue to practice open defecation. The leading cause of death among children under five in Angola continues to be diarrheal disease. The rate of stunting in this age group is 37.6 percent, with some provinces peaking well above 40 percent.3 The rural southern provinces of Benguela, Namibe, Huíla, Cunene, and Cuando Cubango illustrate very well the disparities and inequalities of access to services when it comes to rural areas and the poor. The region is host to a range of landscapes from the high elevation plateau of Huíla to its piedmont and the plains of the Cuvelai, the forests of Cuando Cubango, the rugged terrain in northern Namibe, and 6 Water Security and Drought Resilience in the South of Angola the deserts in the South. Improved roads exist mainly to connect provincial capitals. Rural livelihoods in the region depend on pastoralism, with seasonal transhumance routes; agriculture for domestic con- sumption and limited sales (mostly dryland agriculture, and some of it irrigated when near rivers); as well as fishing along the coast. The region also enjoys the cultural diversity of numerous ethnolinguistic groups including the Kuvale, Kwisi, Muhakaona, Himba, Mwila, Kuanhama, and many others, which could be grouped under five main communities—Herero, Nyanyeka-Humbe, Ganguela, Mbundu, and Ambo—in addition to the San people and people of European descent. An excellent and expansive por- trait of land and life in Namibe, Huíla, and Cunene is available in Mendelsohn and Mendelsohn (2019). The Study Area This Advisory Services and Analytics (ASA) encompasses the south-central part of Angola, namely, the provinces of Benguela, Cunene, Huíla, Namibe, and Cuando Cubango. It is a large and elongated region from 11.5S to 17.5S and from 11.5E to 23.5E (see map 1.1). MAP 1.1. The Five Provinces of this Study and their Average Monthly Rainfall Benguela Huila Namibe Cuando Cubango Cunene Monthly average rainfall (mm) 250 200 150 100 50 0 br y M ry Ap h ril ay ne em t De em r r ce ber Se ug ly No to r be O be pt us v be Fe uar c Ju ua M ar Ju m n c Ja A Namibe Cunene Benguela Cuando K Huila Source: Average Monthly Rainfall calculated using data from the Tropical Rainfall Measuring Mission (TRMM) Multisatellite Precipitation Analysis (TMPA). Water Security and Drought Resilience in the South of Angola 7 This area shows less socioeconomic development than the remaining parts of Angola; it is sparsely populated with many marginalized rural communities. Its provinces have a combined population of nearly 6.5 million, most of which is concentrated in Huíla and Benguela. The rural population is about 65 percent of that figure. In terms of climate and landscape, there is also a difference between the South and the rest of the coun- try. These provinces fall within the arid and semiarid agroecological zone in Angola that is characterized by desert, savannah grass, and woodlands. Temperatures are milder on average than in the rest of Angola (Huntley 2019). It rains less than 800 millimeters (mm) per year on average in most of the South, but there are parts of the southern coast where it rains less than 200 mm. Although total rainfall is significant across most of the region, there are substantial fluctuations across seasons and years. As shown in figure 1.1, some rain falls at the beginning of October and later at the end of the season in March/April. In between, the highest rainfall intensities can occur at any time but are most frequent from January to March, especially in parts more to the south. Year-by-year variance is FIGURE 1.1. Some Facts about Angola’s Country Background Water On average, only 36 percent of ... and current levels of spending are less than the water and sanitation sector’s annual 25 percent of the amount needed budget is executed... (as a percentage of GDP) to achieve the sustainable development goals. Progress in the water and sanitation sector ... while nascent regulations and a poorly is significantly constrained by a centralized implemented tariff-setting framework add governance structure and weak to the challenges. institutional capacity... 62 percent of the population has access ... yet the rate of access rose only 4 percent to an improved drinking water source... between 2000 and 2017. Sanitation 70 percent of the population has access ... still, 6 million people continue to practice to an improved sanitation facility... open defecation. Hygiene Only 42 percent of Angolans have access ... well below the to a handwashing facility... average of 59 percent for Sub-Saharan Africa. Childhood health Children with limited access to water and ... and more than one-third of all children under sanitation service are more susceptible to five in Angola are stunted, 5 percent are wasted, malnutrition... and 19 percent are underweight. Poverty The poor are disproportionately ... 47 percent of the poorest households rely disadvantaged across all indicators... on surface water and 82 percent practice open defecation. Source: Lombana et al. 2021. 8 Water Security and Drought Resilience in the South of Angola moderate across the entire region (variation coefficients are frequently beyond 50 percent), which con- tributes to the unpredictability of annual rainfall and makes it difficult to use averages as a reliable reference. Notes 1. According to the standards set by the United Nations’ Millennium Development Goals for 2015, “improved” drinking water sources are those that, by nature of their design and construction, have the potential to deliver safe water, while “improved” sanitation facilities are those designed to hygienically separate excreta from human contact. 2. That is, not “improved.” 3. A detailed diagnostic of the water sector in Angola can be found in Lombana et al. (2021), whose main recommendations include to invest more, and more wisely, in the WASH sector; to strengthen institutional capacity and participatory approaches; and to establish systems to monitor and collect data to make good decisions and build resilience. Water Security and Drought Resilience in the South of Angola 9 © Aleix Serrat-Capdevila/World Bank Chapter 2 Characterizing Drought Risk in the South of Angola The region affected by the recent drought in south-central Angola is very large, spanning five provinces. This chapter describes the mixed-methods approach taken by this study to understand the structural causes of drought vulnerability across the region, the physical hazards, the distribution of impacts, and ultimately where the greatest needs and vulnerabilities are, and how to prioritize them. It starts with an analysis of data from the 2014 National Census, focusing on data at the commune (comuna) level related to structural variables that explain drought vulnerability. The chapter then con- trasts the results of this analysis (of vulnerability to a potential hazard) with the characterization of the meteorological drought, and the observed impacts on the ground (from the occurrence of a hazard event). The outcome is a list of communes at risk, by order of priority, that can guide decision-makers with a regional view of where investments may be most needed to promote drought resilience. This prioritized list of all the communes in the five provinces can be used as an evidence-based road map for targeting response interventions in this very large region. Vulnerability and Exposure to Drought in the Region The goal of the efforts described in this chapter is to map and characterize the drought vulnerabilities of human populations in the region, including their conditions of access to water supply. Understanding the structural characteristics of the region’s population within its environment is key to understand its susceptibility against climate threats like droughts. Faced with the same hazard, one community could be resilient while another one could suffer important impacts. A high degree of vulnerability to drought is a factor that constitutes a threat to livelihoods and to meeting the most basic needs. On the other hand, drought exposure can be derived from the distribution of population (and assets or goods) that can be affected by drought events. Data and Methodology Step 1: Finding the most vulnerable Vulnerability is the state that exists within a system before it encounters a hazard (Brooks 2003). It is defined by the structural conditions of the system. According to the Intergovernmental Panel on Climate Change (IPCC) definition, vulnerability is the propensity or predisposition to be adversely affected. Within this paradigm, vulnerability can be defined, in turn, as a function of sensitivity and coping capa- city (Shahid and Behrawan 2008; Luetkemeier and Liehr 2018). A set of variables have been used to characterize vulnerability to drought at the commune1 level (see table 2.1). These were selected based on sensitivity indices related to access to water supply defined by Luetkemeier and Liehr (2018) and using the exhaustive commune-level data of the National Census Water Security and Drought Resilience in the South of Angola 11 TABLE 2.1. Indicators Used to Characterize Sensitivity and Vulnerability to Drought In the Study Region from a Domestic Water Supply Point of View Indicators of IND1: Water Main source of drinking water (%). In line with Luetkemeier and Liehr (2018), seasonal a commune’s sources’ changes (wet/dry) of the sources of drinking water indicate reliability levels. sensitivity unreliability Indicator building: Households that strongly depend on unreliable sources are highly to drought, sensitive to drought events and suffer second-order effects if they are not able to switch reflecting local to more reliable sources (Luetkemeier and Liehr 2018). Starting from this premise, water supply Luetkemeier and Liehr (2018) establish reliability levels for different water source types, conditions by analyzing seasonal changes (wet-dry). The categories in the National Census of Angola (Luetkemeier and are very similar, and easy to adapt to this methodology. Then, each reliability level score Liehr 2018) is multiplied by the percentage of the commune relying on that type of water source and added up per commune. IND2: Water Main source of drinking water (%). Data from the National Census of Angola offer an sources’ opportunity to differentiate between safe and unsafe (i.e., poor-quality) water source unsafeness types. Indicator building: Communities that are more exposed to poor-quality water are more likely to suffer impacts such as diseases or even increased mortality rates. In this sense, the census provides a differentiation between safe and unsafe water source types. Based on this classification, each water source has been scored according to whether it is unsafe or not, with a dichotomic 1–0 score. Then, each score is multiplied by the percentage of the commune relying on that type of water source and added up per commune. IND3: Water Main source of drinking water (%). Source types that require financial inputs: boreholes and financial tanker trucks. dependence Indicator building: There are two water source types that can be related with financial (the degree to dependence: boreholes and tanker trucks. which water Boreholes functionality is extremely low in some areas because of financial or technical access depends incapacity to perform repairs. Obviously, this situation has an impact on water availability on financial for inhabitants and transhumance. transactions) Considered a reliable source once drought occurs, tanker trucks can imply a perverse effect, as they normally depend on a direct economical transaction. Communities normally resort to water vendors due to the lack of other free or safer water sources. Some inhabitants might not have the financial capacity to buy water, so eventually they might be forced to acquire it from unhealthy sources. Based on this, a dichotomic 0–1 score has been assigned to the type of sources, boreholes, and tanker trucks rated with 1 because they require economic inputs to be useful. Then, each score is multiplied by the percentage of the commune relying on that type of water source and added up per commune. Source: Original compilation, using the methodology of Luetkemeier and Liehr (2018) and data from IND (2014). of Angola (2014). These indicators measure the degree to which water sources are unreliable (IND1), unsafe (IND2), and require a financial transaction for access (IND3). The calculation details and the ratio- nale behind their use is further explained in table 2.1. A broader vulnerability assessment was also performed, using all the variables and indicators described by Luetkemeier and Liehr (2018), but its results are not as water-focused to prioritize water access inter- ventions as those obtained with just the combination of the three indicators used for this report. The results of this broader analysis using 11 indicators can be found in a parallel paper (Limones et al. 2020). 12 Water Security and Drought Resilience in the South of Angola Once calculated, the normalized scores of the three indicators have been added up to rank the relative vulnerability of all communes in the five provinces under study, in terms of their conditions of access to water. Step 2: Consideration of the exposed vulnerable population Apart from targeting the most vulnerable areas according to those three indicators, the interventions to be designed as a result of this study must benefit the maximum amount of population. Identifying den- sely populated areas within the communes recognized as most vulnerable adds another layer of useful information and can be utilized as a proxy for drought exposure. Therefore, it is valid to consider a dou- ble benefit from this population screening: the implemented solution would impact a significant amount of people and, at the same time, the most exposed areas are being targeted. For this exercise, a radius of 5 km is used to consider people benefiting from a given infrastructure point (note that for DNA and OMS the criteria is 2 km), therefore assuming an area of impact of around 78 km2 per infrastructure investment. If the population density was considered as homogeneous, which is rarely the case, a minimum of 6.5 inhabitants/km2 is needed to reach at least 500 inhabitants within the radius of the constructed infrastructure. Consequently, in the list of selected communes per province, chosen because they show high scores in the sum of normalized indicators (IND1, IND2, and IND3), all those communes that do not reach the 6.5 inhabitants/km2 of population density were highlighted to be further analyzed. For each one of the highlighted communes inside Namibe, Cunene, and Huíla, the high-resolution population grid from Mendelsohn and Mendelsohn (2019) has been utilized to search for population clusters that could merit an intervention because they are exposure hot spots, even if the average population density in the commune is lower than our established threshold. Unfortunately, such spatial information is not yet ­ available for Benguela and Cuando Cubango. Main Results Figures 2.1, 2.2, and 2.3 show the results of the application of the three indicators across the region, before averaging their normalized scores to get a final ranked list of communes based on their condi- tions of access to water. Table 2.2 ranks the communes in order of vulnerability, using the methodology described above and the combination of the three indicators. Based on the table, we can generate an initial list of priority com- munes per province. The communes that do not reach the population density threshold explained in the previous paragraph are highlighted in red, both in the list and in the table: •• Benguela priorities: Kapupa, Wyiagombe, Lambala, Passe, Cayavi. •• Namibe priorities: Lucira, Camacuio, Caitou, Chinquite, Cahinde, Chingo. •• Huíla priorities: Tchipungo, Chimbemba, Chiange, Kapunda Kavilongo, Jau, Kusse, Cutenda, Uaba, Kalukembe, Kalepi. Water Security and Drought Resilience in the South of Angola 13 FIGURE 2.1. Commune Scores and Percentiles for Indicator 1: Water Source Unreliability Water source unreliability indicator (percentiles) % % % shallow Order Commune Score chlmpacas Caclmbas sources 1 Ombala-yo-Mungu 69,37 79,8 18,6 98,4 2 Cafima 61,40 47,2 47,1 94,3 13 3 Môngua 59,42 31,7 64,5 96,2 12 4 Chiede 58,41 19,7 79,2 98,9 5 Evale 57,63 45,9 36,0 81,9 11 10 6 Mucope 55,97 49,4 26,3 75,7 7 Humbe 55,26 55,8 13,9 69,7 N 6 9 8 Naulila 52,23 52,5 13,9 66,4 5 2 7 3 9 Tchimporo-Yonde 51,08 5,4 79,8 85,2 8 1 4 10 Quipungo 48,95 0,6 81,6 82,2 0 100 200 300 400 11 Matala 48,06 0,4 82,8 83,2 Km 12 Cusse 47,59 0,3 78,8 79,1 13 Capupa 46,62 0,2 74,9 75,1 0–10 10–20 20–30 30–40 40–50 50–60 60–70 70–80 80–90 90–100 Note: Commune colors indicate score percentiles, with the 13 communes listed in the top percentile of unreliability. FIGURE 2.2. Commune Scores and Percentiles for Indicator 2: Water Source Unsafeness Water source unsafeness indicator (percentiles) % % % % % water Order Commune Score water chim chim tanker springs bodies pacas bas trucks 1 Rito 100 100 0 0 0 0 2 Neriquinha 100 100 0 0 0 0 9 3 Lupire 100 100 0 0 0 0 4 Bondo 100 100 0 0 0 0 13 11 8 3 5 Baixo Longa 100 100 0 0 0 0 N 6 Cueio 99,9 99,3 0 0,4 0,1 0 5 6 7 Chamavera 99,8 99,7 0,1 0 0 0 2 8 Longa 99,3 98,1 0,4 0,8 0 0,1 10 4 1 9 Bolonguera 98,7 89,5 1,2 8,7 0,2 0,1 12 10 Evale 98 13,5 45,9 35,0 3,4 0,1 0 100 200 300 400 7 Km 11 Dinde 97,7 74,1 0,3 19,7 3,5 0 12 Maue 97,5 87,7 0 9,9 0 0 0–10 10–20 20–30 30–40 40–50 50–60 60–70 70–80 80–90 90–100 13 Lola 97,4 79,3 0,3 16,0 1,9 0,1 Note: Commune colors indicate score percentiles, with the 13 communes listed in the top percentile of unsafeness. •• Cunene priorities: Mongua, Evale, Kafima, Ombala yo Mungu, Humbe, Mukope, Oximolo, Shiede, Naulila, Onkokwa, Otchinjau. •• Cuando Cubango priorities: Kutuile, Neriquinha, Luengue, Rivungo, Kuito Kuanavale. Map 2.1 shows the lower-density communes that were highlighted in red in the priority areas’ list for Namibe, Cunene and Huila provinces. Cells in dark blue are locations with more than 30 people/km2, 14 Water Security and Drought Resilience in the South of Angola FIGURE 2.3. Commune Scores and Percentiles for Indicator 3: Water Financial Dependence Water financial dependence indicator (percentiles) % Tanker Order Commune Score % Boreholes trucks 1 Lucira 96,2 51,3 44,9 2 Bentiaba 63,4 7,8 55,5 10 3 Cutuile 51,0 0 51,0 11 1 4 Moçâmedes 40,9 37,6 3,2 2 5 Cahama 35,2 3,2 31,9 4 6 Chimbemba 29,4 0,9 28,4 12 8 3 N 7 Chiange 23,1 0,1 23,0 6 7 5 8 Cunjamba 22,0 0 22,0 13 9 9 Luengue 19,0 0 19,0 10 Dombe Grande 18,5 0,4 18,2 0 100 200 300 400 11 Ekimina 17,2 8,9 8,3 Km 12 Cuito Cuanavale 14,2 4,1 10,0 13 Otchinjau 14,0 0 14,0 0–10 10–20 20–30 30–40 40–50 50–60 60–70 70–80 80–90 90–100 Note: Commune colors indicate score percentiles, with the 13 communes listed in the top percentile of financial dependence. TABLE  2.2. Ranking of the Communes of the Five Studied Provinces Based on their Vulnerability to Drought Ranking of Ranking of vulnerability vulnerability to drought Population to drought Population scores  Commune  Province  Population  density  scores  Commune  Province  Population  density  1  Lucira  Namibe  10,986  7.70 66  Kakonda  Huíla  59,685  50.90 2  Ombala yo Cunene  59,331  30.56 67  Kuvati  Cunene  3,774  0.82 Mungu  3  Kafima  Cunene  50,229  14.43 68  Kuvango  Huíla  36,317  8.71 4  Evale  Cunene  55,386  26.95 69  Kunjamba  Cuando Cub.  1,064  0.16 5  Shiede  Cunene  55,747  5.74 70  Chila  Benguela  17,666  12.69 6  Humbe  Cunene  35,985  12.57 71  Chikuma  Benguela  44,762  47.98 7  Mongua  Cunene  75,811  30.47 72  Bondo  Cuando Cub.  1,868  0.29 8  Mukope  Cunene  78,225  26.12 73  Kaiundo  Cuando Cub. 14,665  2.41 9  Kutuile  Cuando Cub.  2,334  0.35 74  Kanjala  Benguela  13,684  9.73 10  Oximolo  Cunene  5,189  0.49 75  Kalonga  Cunene  8,619  4.16 11  Tchipungo  Huíla  56,657  5.83 76  Longa  Cuando Cub.  15,104  1.07 12  Naulila  Cunene  60,854  29.07 77  Kueio  Cuando Cub.  4,247  0.66 13  Otchinjau  Cunene  29,296  5.62 78  Lupire  Cuando Cub.  1,600  0.26 14  Onkokwa  Cunene  17,923  4.93 79  Baixo Longa  Cuando Cub.  1,352  0.25 15  Kapupa  Benguela  63,440  54.60 80  Rito  Cuando Cub.  906  0.34 16  Mupa  Cunene  20,681  9.42 81  Xa-mavera  Cuando Cub.  4,975  0.88 table continues next page Water Security and Drought Resilience in the South of Angola 15 TABLE 2.2. Continued Ranking of Ranking of vulnerability vulnerability to drought Population to drought Population scores Commune Province Population density  scores Commune Province Population density  17  Wyiagombe  Benguela  16,983  36.88 82  Cavimbe  Benguela  30,788  26.59 18  Chitado  Cunene  23,163  5.52 83  Matala  Huíla  141,159  99.19 19  Chimbemba  Huíla  38,839  8.89 84  Munhino  Namibe  14,316  3.68 20  Kuvelai  Cunene  24,324  3.68 85  Dongo  Huíla  43,756  10.02 21  Lambala  Benguela  51,147  64.56 86  Impulo  Huíla  15,721  13.54 22  Chiange  Huíla  40,623  10.70 87  Kihita  Huíla  24,817  20.88 23  Neriquinha  Cuando Cub.  1,331  0.14 88  Chibia  Huíla  74,331  53.72 24  Caitou  Namibe  10,486  6.92 89  Babaera  Benguela  27,044  46.32 25  Kapunda Huíla  66,337  44.14 90  kalai  Cuando Cub.  20,091  6.13 Kavilongo  26  Jau  Huíla  25,186  21.70 91  Cacula  Huíla  34,429  45.77 27  Kusse  Huíla  46,725  53.69 92  Chinguanja  Cuando Cub.  3,612  0.90 28  Camacuio  Namibe  41,449  17.83 93  Kanhamela  Benguela  19,754  32.61 29  Chinquite  Namibe  5,407  1.96 94  Baia dos Namibe  3,170  4.19 Tigres  30  Cutenda  Huíla  40,847  16.19 95  Kilengue  Huíla  39,511  10.93 31  Cahinde  Namibe  17,348  3.19 96  Kuchi  Cuando Cub.  20,361  3.67 32  Luengue  Cuando Cub.  3,547  0.18 97  Cubal  Benguela  166,783  128.66 33  Uaba  Huíla  37,953  45.87 98  Bocoio  Benguela  63,438  54.71 34  Kalukembe  Huíla  96,099  83.41 99  Ebanga  Benguela  23,984  24.52 35  Namakunde  Cunene  86,299  70.63 100  Chingo  Namibe  3,493  1.42 36  Rivungo  Cuando Cub.  15,115  1.96 101  Chingongo  Benguela  20,619  37.26 37  Bentiaba  Namibe  8,952  23.65 102  Dirico  Cuando Cub.  12,620  1.64 38  Passe  Benguela  14,846  18.89 103  Jamba  Huíla  35,023  10.82 39  Kalepi  Huíla  30,700  54.78 104  Chindumbo  Benguela  26,374  35.38 40  Cayavi  Benguela  3,101  7.86 105  Ganda  Benguela  93,609  116.34 41  Dinde  Huíla  20,102  16.35 106  Egito-Praia  Benguela  13,198  13.64 42  Chicomba  Huíla  28,939  17.99 107  Vikungo  Huíla  7,989  6.25 43  Virei  Namibe  15,097  1.49 108  Humpata  Huíla  35,688  32.34 44  Hoque  Huíla  64,475  38.23 109  Namibe  Namibe  272,598  39.94 45  Onjiva  Cunene  18,7914  89.60 110  Xangongo  Cunene  70,568  27.32 46  Gungue  Huíla  23,457  12.21 111  Chongoroi  Benguela  67,539  24.17 47  Kahama  Cunene  40,766  8.65 112  Menongue  Cuando Cub.  292,533  26.83 48  Katengue  Benguela  15,347  11.94 113  Kaimbambo  Benguela  35,654  53.21 table continues next page 16 Water Security and Drought Resilience in the South of Angola TABLE 2.2. Continued Ranking of Ranking of vulnerability vulnerability to drought Population to drought Population scores Commune Province Population density  scores Commune Province Population density  49  Kuito Cuando Cub.  22,774  2.22 114  Mavinga  Cuando Cub.  20,251  1.63 Kuanavale  50  Lola  Namibe  14,286  8.80 115  Kutato  Cuando Cub.  15,031  6.59 51  Galangue  Huíla  34,238  9.60 116  Monte Belo  Benguela  37,947  33.26 52  Quendo  Benguela  24,262  23.34 117  Ekimina  Benguela  1,752  1.12 53  Kapelongo  Huíla  49,447  21.34 118  Cuangar  Cuando Cub.  13,300  3.11 54  Savate  Cuando Cub.  13,291  1.56 119  Dombe Benguela  41,434  14.24 Grande  55  Kassinga  Huíla  26,310  6.21 120  Lubango  Huíla  581,180  752.43 56  Bambi  Huíla  14,760  11.28 121  Mucusso  Cuando Cub.  3,602  0.68 57  Mulondo  Huíla  24,822  4.70 122  Bibala  Namibe  25,416  20.78 58  Huíla  Huíla  60,278  102.16 123  Balombo  Benguela  47,806  34.96 59  Bolonguela Benguela  9,726  3.03 124  Nancova  Cuando Cub.  1,890  0.25 60  Kaseke  Benguela  46,087  33.18 125  Biopio  Benguela  5,488  21.42 61  Chipindo  Huíla  33,153  13.21 126  Baia Farta  Benguela  61,572  47.15 62  Ngola  Huíla  53,132  23.74 127  Lobito  Benguela  366,198  598.36 63  Luiana  Cuando Cub.  8,010  0.62 128  Tombua  Namibe  52,324  3.02 64  Kalahanga  Benguela  3,082  2.84 129  Benguela  Benguela  561,775  236.98 65  Maue  Cuando Cub.  831  0.17 130  Katumbela  Benguela  95,034  289.38 Note:  Communes of low population total and density are highlighted in red. The ranking is according to the sums of the normalized values of three vulnerability indicators – IND1, IND2, and IND3  – for each commune as defined by Luetkemeier and Liehr (2018).  while the baby blue cells depict areas with more than 200 people/km2. The part of each commune with a density lower than 30 people/km2 is shown in red. According to the visual population check, only Oximolo in Cunene could be questioned and taken out of the preliminary priority list, since in any case there are other parts of this province with high levels of vulnerability. As mentioned above, these population grids are not available for Benguela and Cuando Cubango. For the first case, this turned out not to be an issue because it is a densely populated province and none of the selected communes fell below the established threshold. For Cuando Cubango, with a low population density of less than 6.5 people/km2, the five communes in the list can be tentatively retai- ned, pending more exhaustive follow-up work using satellite imagery to map agriculture in the con- text of follow-up work. Water Security and Drought Resilience in the South of Angola 17 MAP 2.1. Population Distribution Maps Showing the Cunene, Huíla, and Namibe Communes with Lower Population Densities Chinquite Cahinde Population is scarce, but in the northwest of the A few thousand people in the areas in the northeast. commune there are a few clusters. Any infrastructure built in this area could serve hundreds of people. Tchipungo Oximolo Around 3,000 people in the light blue areas, so there are There are not enough densely populated areas, so it might be a important settlements; also meets conditions in the candidate to be taken out of the list. Another option is to plan north. the infrastructure around its southwest border, as close to the border there are a few more populated communes. Shiede Onkokwa Plenty of opportunities in the western part of the There are two areas in which we find hundreds of commune. clustered people (see baby blue points). Otchinjau Otchinjau The central part of the commune contains certain areas The population is extremely concentrated in the head of the with clusters of population, so it must be kept. commune and a bit in the northwest edge, but there are hundreds of people in these clusters. Source: Original compilation, using population data from Mendelsohn and Mendelsohn (2019). 18 Water Security and Drought Resilience in the South of Angola Drought Characterization and Direct Hydrologic Impacts This part of the study aims to provide an overall understanding of the dynamics of water availability in the region and to describe the evolution of those dynamics from 2012 to the time of the completion of this ASA. The progression of the meteorological drought hazard and the monthly yields are analyzed for that period. Data and Methodology To characterize the meteorological drought over the past years and understand its influence across the region, a drought index was calculated using rainfall data estimated by satellite. It is essential to deter- mine where drought has recently hit with significant intensity and duration and where the phenome- non is getting increasingly more threatening. Quantifying the meteorological drought also provides insights into the relative roles of the rainfall anomaly (nature dependent, beyond our reach) and the structural vulnerabilities of access to water (dependent on infrastructure, within our reach), in modula- ting the severity of observed impacts on the ground. The Drought Exceedance Probability Index (DEPI) is used for this analysis. This index is a modification of the Standardized Index of Rainfall Drought (indice standardisé de sécheresse pluviométrique, ISSP; see Pita [2000] to learn more about the index formulation). The DEPI is an index based on the calculation of cumulative monthly rainfall anomalies, in a similar way to the Standardized Precipitation Index (SPI) of McKee, Doesken, and Kleist (1995) or the Standardised Precipitation-Evapotranspiration Index (SPEI) (Vicente-Serrano, Beguería, and López-Moreno 2010). However, in this case each monthly DEPI score represents the empirical probability of exceeding the level of drought experienced in that month, so low DEPI values describe more severe drought (a value of 0.5 is representative of the normal cumulative anomaly; below that level is considered drought). Since rainfall stations are currently scarce in the region and their records are not continuous, satellite estimates of rainfall provide a good alternative, being periodic and spatially continuous. Consequently, satellite rainfall estimates freely available online from the Tropical Rainfall Measuring Mission (TRMM) Multisatellite Precipitation Analysis (TMPA) have been used as input data. The TMPA has been selected because it performs accurately in the South of Angola (Pombo, Oliveira, and Mendes 2015). The DEPI has been applied to the TMPA monthly values from January 1998 till January 2019 for the 21 points distribu- ted across the region that appear in map 1.1. The evolution of the DEPI from 2012 to 2019 has been exa- mined in detail in all the points.2 To understand the evolution of water availability during this period, the monthly precipitation minus the actual evapotranspiration (P-ETa) has been computed at a medium spatial resolution (0.25°, ~25 km) across the entire region. Different studies consider the result of P-ETa as a simple and reliable proxy to the yield or the produced runoff per unit watershed area (Guerschman et al. 2008; Weiss 2009). Actual evapotranspiration data has also been obtained from satellite estimates, namely the monthly actual evapotranspiration (ETa) data sets from Global Land Evaporation Amsterdam Model (GLEAM; https://www.gleam.eu), available from 2013 to present. Water Security and Drought Resilience in the South of Angola 19 The full monthly series P-ETa from 2003 to January 2019 was obtained for all the pixels of the region, and some statistics were calculated to understand the normal (non-drought) regimes and what is diffe- rent in the most recent regimes from 2012 to 2019. These results are aimed at confirming the areas expe- riencing significant deficits in water availability. Main Results The application of the DEPI drought index to the TMPA monthly series of precipitation corrobora- tes that the five provinces suffered moderate to severe meteorological drought from 2012 to 2019, but with differences in the temporal evolution and intensity of the phenomenon across the terri- tory (see figure 2.4): •• The southern part of Benguela had been suffering drought conditions since 2012, while the nor- theast of the province experienced it for the last 2 years of the analyzed series. In general, the phe- nomenon got progressively more intense in this province. •• The entire provinces of Huíla and Cunene experienced a progressive increase in drought conditions that started out slow in 2013 and grew to be extremely intense from 2016—a season from which all the consecutive rainy seasons were anomalously dry until the end of year 2019. •• The highlands of Namibe—which correspond to the eastern part of the province—experienced a gradually intensifying drought from 2012, relatively parallel to the evolution of the drought in Huíla or Cunene, while in the coastal areas—the western half—drought conditions occurred intermit- tently throughout that period. Moreover, the phenomenon became more and more intense in the eastern highland, while the coast seems to be recovering. •• Last, Cuando Cubango has faced more intermittency than the rest of the region, but is registering short deficits from 2013. In 2015, a more continuous drought started in the entire province, but the eastern part has been steadily recovering from 2017, while the western part has sustained a rainfall deficit similar to that of its neighboring provinces. All this proves that the central part of the region was hit more intensely and continuously by drought and, in addition, across most of it the deficit continued to accumulate until the end of the study period. Hence, the intensity of the event grew gradually over the study period. This conclusion justifies focu- sing mainly on the central core of the region to evaluate the drought impacts from 2012 and propose investments: the provinces of Huíla and Cunene, the western part of Cuando Cubango, the south of Benguela, and the eastern part of Namibe. On the other hand, the account of the monthly P-ETa points to the same problematic geographic area: the central core of the southern region. Not only is this area most affected by the meteorological drought from 2012 to 2019 but it is among those with the lowest values of produced runoff per unit of watershed area under normal conditions (see figure 2.5); also, it experienced relatively more significant reductions in hydrologic yields in the study period (see figure 2.6). 20 Water Security and Drought Resilience in the South of Angola FIGURE 2.4. Temporal Graphs Depicting the Evolution of the Drought Exceedance Probability Index in 21 Points of the Region, and the Provincial Averages Evolution of the drought in the five provinces of the study area 1.00 0.75 0.50 0.25 N 0 98 Ja 9 00 Ja 1 Ja 2 Ja 3 Ja 4 Ja 5 Ja 6 Ja 7 Ja 8 09 10 11 12 3 14 15 16 17 18 19 0 0 0 0 n- 0 0 9 0 0 n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Benguela NE Benguela NW Benguela SE Benguela SW Average 1.00 0.75 0.50 0.25 N 0 98 99 00 01 02 03 04 05 06 07 08 09 10 11 12 3 14 15 16 17 18 19 n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Huila NE Huila NW Huila SE Huila SW Average 1.00 0.75 0.50 0.25 N 0 98 99 00 01 02 03 04 05 06 07 08 09 10 11 12 3 14 15 16 17 18 19 n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Namibe NE Namibe NW Namibe SE Namibe SW Average 1.00 0.75 0.50 0.25 0 N 8 Ja 9 00 Ja 1 Ja 2 Ja 3 Ja 4 Ja 5 Ja 6 Ja 7 Ja 8 09 10 11 12 3 14 15 16 17 18 19 0 0 0 0 n- 0 0 -9 0 -9 0 n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- Ja n n Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Cunene NE Cunene NW Cunene SE Cunene SW Average 1.00 0.75 0.50 0.25 N 0 8 Ja 9 00 Ja 1 Ja 2 Ja 3 Ja 4 Ja 5 Ja 6 Ja 7 Ja 8 09 10 11 12 3 14 15 16 17 18 19 0 0 0 n- 0 0 0 9 0 -9 0 n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- n- Ja n Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Cuando C NE Cuando C NW Cuando C SE Cuando C SW Cuando C center Average Note: Values between 0.5 and 0 indicate drought, more intense as the DEPI values decrease. Water Security and Drought Resilience in the South of Angola 21 The largest negative anomalies in yields during the years 2012–19 are concentrated in the wet mon- ths, which are the only ones in which the production of runoff is possible in most of the region (according to figure 2.5), especially from December to March. Again, these deficits are particularly substantial and continuous in the central core of the region: the entire provinces of Huíla and Cunene and the adjacent halves of the other provinces. These areas registered average deficits well beyond 20 mm of yield per month in the rainy seasons of 2012–19 (figure 2.6), which significantly impacted water availability. This analysis highlights the area that experienced the worst hydrometeorological conditions during the last drought. It represents the physical magnitude of the hazard, and provides a sense of the exposure experienced by the affected communes. FIGURE 2.5. Average Monthly Precipitation Minus Actual Evapotranspiration in the Region for 1994–2019 January February March April May June July August September October November December Precipitation minus actual evapotranspiration <–200mm –200––50 –50––20 –20–0 0–20 20–100 100–500 > 500 mm 22 Water Security and Drought Resilience in the South of Angola FIGURE 2.6. Difference Between the Average Monthly Values of Precipitation Minus Actual Evapotranspiration in the Period 2012–19, and Average Monthly Values Calculated for the Entire Period 1994–2019 January February March April May June July August September October November December Difference between the mothly mean 2012–19 and the total monthly mean from 1998 <–200mm –200––50 –50––20 –20–0 0–20 20–100 100–500 > 500 mm Expected Impacts: Overlaying Vulnerability Conditions and the Severity of the Drought Event Overlaying the drought hazard, exposure, and commune vulnerability analyses performed in the prece- ding sections, it is possible to identify the vulnerable areas with the highest levels of drought hazard intensity for the period 2012–19. Doing this, a number of communes emerge as having the highest prio- rity in terms of drought risk. These communes constitute the geographical core of the ­drought-­impacted region. From the list of vulnerable communes in the previous section (based on conditions of access to water), those in Namibe and Benguela were in parts of their provinces severely affected by the physical drought. Water Security and Drought Resilience in the South of Angola 23 Neriquinha, Rivungo, and Kutuile, meanwhile, are located in the eastern part of the Cuando Cubango province, in areas that did not experience dry periods of more than 2 years’ duration throughout the years 1998–2019. After further analyses, these were replaced with Savate, Bondo, and Maue (the next ones in the priority list for Cuando Cubango, located toward the province’s western part), pending the confirmation of population clusters in them, since they seem to also be sparsely populated. Having said this, it is important to note that, in general, the communes of this province have lower vulnerability scores (and poorer data) than the rest of the region. Last, since the entire two provinces of Huíla and Cunene were hard hit by the drought, and with increa- sing severity until 2019, all the priority communes listed before are retained. Therefore, the final list of priority areas per province based on the desk study of drought vulnerability, exposure, and drought severity, remains very similar to the previous vulnerability list based on the con- ditions of access to water in each commune: •• Benguela priorities: Kapupa, Wyiagombe, Lambala, Passe, Cayavi. •• Namibe priorities: Lucira, Camacuio, Caitou, Chinquite, Cahinde, Chingo. •• Huíla priorities: Tchipungo, Chimbemba, Chiange, Kapunda Kavilongo, Jau, Kusse, Cutenda, Uaba, Kalukembe, Kalepi. •• Cunene priorities: Mongua, Evale, Kafima, Ombala yo Mungu, Humbe, Mukope, Shiede, Naulila, Onkokwa, Otchinjau, •• Cuando Cubango priorities: Luengue, Kuito Kuanavale, Savate, Maue, Bondo. Notes 1. In Angola, provinces are divided into municipalities, which in turn are subdivided into communes (comunas). The study area is made up of 130 communes, which have been the basic administrative unit for the calculation of the indicators in this step, being the smallest spatial entity with publicly available disaggregated census data. There are drawbacks in the use of communes, as opposed to the use of communi- ties (aldeias/bairos), as different communities within a given commune may have different levels of vulnerability. 2. Further in-depth analyses are possible to construct comprehensive composite indices that combine the magnitude and frequency of occu- rrence of droughts throughout a time series, in order to evaluate the overall hazard posed by drought in a particular point (Dabanli 2018b; Shahid and Behrawan 2008; Rajsekhar, Singh, and Mishra 2015). This will be done in the framework of follow-up work. 24 Water Security and Drought Resilience in the South of Angola Chapter 3 Understanding Impacts Observed on the Ground While the previous chapters describe a desk analysis based on valuable census data and satellite infor- mation on drought conditions, this chapter provides the essential insights gained during visits to the South of Angola and discussions with impacted communities, partners, and government authorities. These efforts enabled a better understanding of the actual impacts suffered on the ground, the challen- ges, and the potential solutions to build drought resilience. All drought indicators need to be validated with observations on the ground—that is, compared with information representative of local drought conditions and impacts, which are “observable losses or chan- ges that occurred because of drought” (National Drought Mitigation Center 2019), although that compari- son rarely happens (Bachmair et al. 2016; Van Loon et al. 2016). Despite the valuable findings presented in the Post-Disaster Needs Assessment (PDNA), the document has limited value for ground validation because it presents information aggregated at the provincial level and there is a general lack of compre- hensive databases or formal records of drought impacts in the country. Thus, the work at this second stage focused on complementing the PDNA findings and providing understanding of the situation at a higher spatial resolution, through field work and with feedback and help from a network of local partners. The mission meetings and field work took place in several locations in Huíla, Cunene, and Namibe, as these areas were confirmed by the desk work as the most affected by the drought from 2012 and the most vulnerable ones too. This approach was endorsed by the staff of several development agencies consulted in Luanda before field missions were planned in the South. Appendix D contains a log of the mission schedule, field work sessions, and meetings, whose main outcomes are summarized in the following subsections. Gathering Feedback From Local Communities Methodology The first round of local experts and stakeholder consultations consisted of successive key informant interviews of a total of 21 representatives of different entities working on the ground in water resources management, livestock and agriculture, support to emergencies, relief, and related topics. These inclu- ded staff of the Fundo de Apoio Social (FAS), United Nations Children’s Fund (UNICEF), Proteção Civil e Bombeiros, Instituto Nacional de Recursos Hidricos (INRH), and the Food and Agriculture Organization (FAO) in the South of Angola. Some of these meetings took place in Luanda. The start of the exercises consisted in asking a series of questions about the impacts of the 2012–19 drou- ght, revolving systematically around the following topics: 1. Start of the drought, intensity and duration, geographical differences in the phenomenon; 2. Main affected sectors and significant impacts in the different areas of the region and throughout time; and 3. Drivers of resilience in the areas that coped relatively well. Water Security and Drought Resilience in the South of Angola 25 It continued with a presentation of the outcomes of the two screening analyses, aimed at finding the areas with the highest hazard, exposure, and vulnerability levels (stage 1 results), in order to discuss the relevance of the measured parameters and confirm the accuracy of the results. It was essential to follow this order so that the results of the researchers’ analyses did not influence the discourse of the consulted individuals. Building from the feedback received, a second round of consultations was organized in the municipali- ties that contained a higher number of identified priority communes, facilitated by FAS (see the visited priority communes in map 3.1). The meetings developed as informal focus group discussions (FGDs) and  took place mainly in the headquarters of the government of the selected municipalities and communes. In all cases, a group of 10–15 members from the communities participated in each of the sessions. A communal or municipal representative (“soba,” or “regidor”), several farmers and herders’ represen- tatives from the different communes, a member of the municipal government in charge of water supply, and another one involved in livestock and agriculture management were always present. MAP 3.1. Communes Where Detailed Field Work was Undertaken (Including Interviews and Discussions in Municipality Government Offices 26 Water Security and Drought Resilience in the South of Angola A detailed line of questioning was used to guide the interviews/discussions with participants, aimed at gaining understanding of: 1. The socioeconomic and environmental characteristics of the municipality and the communes that condition resources management and water supply, beyond what can be concluded from the census data. 2. The nature, degree, and distribution of impacts suffered in 2012–19 in each of the communes within the municipality. 3. The interventions to mitigate them. In this case, an informal FGD was facilitated in every place visi- ted. Specific questions were posed to the entire group, but efforts were made to ensure that all parti- cipants offered their individual views independently. Main Results These two types of consultations were excellent platforms, through facilitated open dialogues, to collect information on circumstances that determined vulnerability and impacts of past droughts and that will govern risk in the future. These discussions were essential to test our methods. In particular, they were very useful in identifying and evaluating potential actions for drought preparedness and mitigation, considering the drought hazard, impacts, vulnerabilities and development goals (Andreu et al. 2015). All the experts consulted in the first round confirmed the need to give priority to Cunene, Huíla, and Namibe, in that order. Seventeen of the key informants, mainly those coming from entities coordinating or collaborating with drought relief campaigns, were able to provide a list of municipalities suffering the most severe impacts and receiving continuous government and donor support (see table 3.1). It is important to highlight that these municipalities, especially the ones in Cunene and Huíla, were consistently mentioned by the key informants. Three respondents provided specific feedback regarding the administrative level of the communes. Interestingly, Chiange (in Huíla), Chingo (in Namibe), Onkokwa and Ombala Yo Mungu (in Cunene) were repeated as most affected in the three cases, and they well match the selection in stage 1. Therefore, the team could corroborate the validity of the list obtained based on the analyses of vulnerabi- lity and drought hazard up to the level of the municipality, since more than two-thirds of the listed priority communes were contained in the enumeration of municipalities delivered in this first round of consulta- tions. From the municipalities identified as problematic by the consulted local experts, only the Namibe coastal municipality was not fully underlined as a high priority by the performed analyses (see table 3.1). On the other hand, the visit to a sample of the communes (see map 3.1) and the informal FGDs in them allowed the team to get familiar with the water supply and resilience deficiencies at local scales in the South. Table 3.2 shows a summary of the main problems and drought impacts experienced from 2012 in the visited communes, according to the responses and the issues raised in the FGDs. Water Security and Drought Resilience in the South of Angola 27 TABLE 3.1. List of Impacted Municipalities Based on the Information Provided by Key Informants and Degree of Adjustment to the Results of this Study Municipalities most impacted (according to Were they recognized as areas at risk in the vulnerability and hazard key informants) analysis? Province they belong to Gambos Yes. It is the municipality of Huíla characterized by the worst vulnerability and hazard indicators. Gambos comprises 2 communes: Gambos (also referred to as Chiange) and Chimbemba, in the stage 1 list. Huíla Caconda Yes. Caconda comprises 4 communes. Of them, Uaba and Kusse are in the stage 1 list. Bibala Yes. This is an area with generally poor vulnerability and hazard indicators. Caitou is one of its 4 communes and it is in the stage 1 list Camucuio Yes. This is another area with generally poor indicators. Camucuio comprises 3 communes: Camucuio, Chinquite (or Mamue), and Chingo, which are all in the stage 1 list. Namibe Namibe (municip.) Yes, although most of this coastal area had less hydrometeorological issues from 2012 than some other parts of the province. Namibe comprises communes with poor vulnerability indicators. Of these, Namibe and Lucira are in the stage 1 list, mainly because of the degree to which water access depends on financial transactions Curoca Yes. Curoca comprises 2 communes, Chitado and Onkokwa. Of these, Onkokwa is included in the stage 1 list. Ombadja Yes. Ombadja comprises 5 communes: Xangongo, Mukope, Humbe, Naulila, and Ombala Yo Mungo. The last 4 of these are in the stage 1 list. Cuanhama Yes. Cuanhama also appears as one of the most vulnerable municipalities. Cunene It comprises 4 communes: Ondjiva, Mongua, Kafima, and Evale, of which the last 3 are in the stage 1 list. Namacunde Yes. Namacunde comprises 2 communes: Namacunde and Shiede, the second of which is included in the stage 1 list. Self-reported drought measures are considered more valid if the self-assessments of people living in proximity are similar (Hunter, Gray, and Edwards 2013). In table 3.2 we can see that the sample respon- ses are very consistent even though they span three different provinces, but the issues are especially similar among communes that are close to each other, like Chimbemba and Chiange (Gambos). Apart from the information summarized in the table, a key result of these consultations was the confir- mation that the total of the cases and respondents (100 percent) experienced what they perceived as an exceptionally long and intense drought. All of them—with no exception—reported that most of the impacts registered started immediately after the DPEI meteorological drought index initiated its decline in each of the areas (see figure 2.4). The only exception was that the failure of water points tapping groundwater happened later, as drought conditions persisted beyond about 6 months.1 Interviews of key informants and the informal FGDs also confirmed the indicators’ ability to reveal vul- nerability to drought and potential impacts in the South of Angola. In particular, high scores for the degree to which water sources were unreliable and unsafe matched fully with the areas reported as the 28 Water Security and Drought Resilience in the South of Angola TABLE 3.2. Impacts Reported for Each of the Sampled Priority Communes (aka Chiange) Chimbemba Mamue (aka Ombala Yo Chinquite) Chingo Sector Issues/critical circumstances/ commune Onkokua Gambos drought impacts Mungu Water points not georeferenced x x x x x (But x (But making making progress) progress) Lack of knowledge of water x Issues linked to water supply and WRM points functionality Water points failure x x x x x x Points not repaired due to a mix x x x x x of lack of technical capacity, spare parts, and financial capacity Long trips to fetch water x x x x x x High dependence on cacimbas x x x x High dependence on chimpacas x High dependence on water trucks x Hydrological drought (sources x x x x x x experienced water deficit) Poor water quality and diseases x x x Transhumance receiving area x x Transhumance exodus area x x x Issues linked to primary activities Significant livestock mortality x x x reported Decrease of crop production and x x x x Not much x crop failure in rainfed farms is planted in this commune Decrease of crop production and x x x They normally rely Not much x crop failure in irrigated small on rainfed agriculture is planted gardens since they had very in this favorable conditions commune (until the drought) Source: Information compiled in FGDs. Note: WRM = water resources management. most broadly and deeply affected, showing that the issues of reliability and safety are among the most urgent to tackle in order to increase resilience in the region. Interestingly, municipalities and communes where access to water relied on financial transactions to a high degree—specifically those along the Namibe coast requiring tanker trucks—were also pinpointed, even though were not the worst hit by the physical phenomenon and therefore were not considered as Water Security and Drought Resilience in the South of Angola 29 urgent by the desk analysis. This points to structural water access issues that need to be addressed, beyond just drought risk. The coast of Namibe is an arid area with a reduced buffer in terms of water availability and few currently developed options beyond truck deliveries. Water tanker truck service can become scarce, expensive, and unreliable in many situations, like a mild drought one when they need to provide for more people and with less reliable sources than is normal.2 Thus, the financial dependence of access to water constitutes another valid indicator to highlight zones at risk. Main Issues Observed in the Field This section provides a summary of the most important findings from the field work (see map 3.1 to check the communes visited at this stage). Regarding rural drinking water supply, the conditions and the governance of sources vary with the geographical setting, but are in general quite challenging: Many water points are not working, and many new ones should be set up across large areas that are populated but do not count on reliable water sources. Estimates of the percentages of failed water points are difficult to make and are contradictory, but for sure more than half of them are not available. The time required to fetch water is the greatest that everybody in the visited communities remembers. This also takes a toll on livestock. Local administrators keep lists of all existing water points, the new ones that should be set up, and the ones that need rehabilitation. The lists include information about the village in which the points are located, well characteristics (depth, static water level, dynamic water level, yield, type of pump; see photo 3.1), and reasons for non-operability (in inventories of rehabilitation needs). The information is never georeferenced, and it is uncertain if and when these lists are sent to higher authorities, and if they are accounted for to mobilize resources for repairs and maintenance. In all the locations visited, stake- holders pointed to a lack of resources for carrying out these tasks. PHOTO 3.1. Volanta Pump (Left); a Solar-Powered Water Point (Right) Source: World Bank. 30 Water Security and Drought Resilience in the South of Angola Manual pumps are frequently destroyed by misuse, and the tanks and reservoirs of the water points are of bad quality. There is a need to modernize infrastructure; at the same time, however, there are a significant number of failing water points using new technologies and based on solar energy (the team ­ visited a few in the three provinces). Since the reasons for their breakdown frequently relate to aban- donment, vandalism, and thievery, there is an urgent need for resources and training in awareness raising and operation and maintenance (O&M). Solar water points, due to lack of maintenance and even ­ acts of vandalism, are very susceptible to inoperability. In the context of ongoing studies, dual pumps are foreseen, that is, in each system supplied by solar pumps, a system with volanta pumps in parallel. Failure to wash the solar panels is one of the main causes of failure. The most common way to obtain drinking water in the region is through shallow cacimbas (hand-dug wells that are generally less than 4–5 m deep), excavated mainly in river beds or other loose sediments near the surface (see photo 3.2). In average years, these cacimbas run dry by July, but in the past few PHOTO 3.2. A Cacimba Excavated in a Riverbed Source: World Bank. Water Security and Drought Resilience in the South of Angola 31 years most of them have dried up by early May. Thousands of people rely on these hand-dug wells; more than half of them cover long trips of more than 10 km to get to some of these wells, even competing with animals for the water. It is rare to find cacimbas that never dry out and have a considerable yield. In the vulnerable areas of Cunene belonging to the Cuvelai catchment, cacimbas are a common drinking water source; however, the most characteristic elements of this environment are the chimpacas, large shallow ponds that collect rainwater and runoff coming from small catchments (photo 3.4). Obviously, the water level in chimpacas depends on the previous season’s rains, and also on their design and main- tenance, especially desilting. Once chimpacas dry up, the users dig holes (cacimbas) inside them to access water contained in the chimpacas’ sediments. During the April 2019 field mission, only 4 chimpacas out of 39 had water in the commune of Ombala Yo Mungo. Such conditions, indicating an emergency situa- tion, are to be found throughout most of the Cuvelai catchment. In coastal villages and towns, the majority of the population gets their regular supply of drinking water through tanker trucks,3 as confirmed by the local authorities (photo 3.3). Tanker trucks of various private companies offer water supply services for around US$0.01 per liter, although this price varies. Those that cannot afford it in these areas—which is always a significant percentage of the PHOTO 3.3. Truck used in Onkokwa, Provided by the National Service of Civil Protection and Firemen Source: World Bank. 32 Water Security and Drought Resilience in the South of Angola PHOTO 3.4. Silted Chimpaca with Very Limited Remaining Live Storage (Left); A Cacimba (Open Well) Dug Inside a Silted Chimpaca to Find a Small Quantity of Water (Right) Source: World Bank. Note: Such water points are unsanitary, as the bottom of the chimpaca is covered by excreta from animals and humans, and by all kinds of waste. This chimpaca has 57 cacimbas excavated in its dry bed. Each hole is “protected” and belongs to one family. Of the 57 families interviewed, most travel more than 10 km to fetch water. population—end up searching for water in contaminated streambed cacimbas (such as in Lucira) or simi- lar sources. The need for truck deliveries originates from the poor quality of groundwater access in most of these zones, and the fact that where there are rudimentary connections to boreholes that are further from the coast, the maintenance of the pipelines and the rest of the system is lacking, leading to fre- quent service interruptions and failures. Some washes in the arid lands of Namibe seemed to have the potential for sand dam construction, so several waypoints were registered in order to facilitate an expert follow-up analysis. For now, the use of sand dams is not widespread in Angola, except for dams built in colonial times. There are several damaged weirs and small dams in the southwest of Angola. The municipalities keep records of all them. These were constructed in colonial times and were destroyed by flash floods or filled up with sediment. The design of all types of supply systems is generally very poor. Often the team recognized problems that included components’ breakdown, problems caused by the pressure in the pipelines, underestimated sizes of reservoirs at the water point, inadequate pumps, inadequate installation depths, no separation between animal and human drinking points, etc. All of these challenges should be, in principle, easy to resolve, but they are the result of a lack of technical knowledge, supply chains for spare parts, and financial capacity. Related to all of the above, the study also identified the impacts and disruptions that the drought was causing (and the main vulnerabilities) with regard to access to water for economic and domestic activi- ties among rural livelihoods. These included the following. Water Security and Drought Resilience in the South of Angola 33 There are parts of the south that are under increasing pressure from transhumance, and that pressure affects their water points, especially in Huíla province. Some communes receive livestock from other areas, normally from Cunene and the more southern parts of the region, and sometimes even from Namibia. This is causing fodder- and water-related conflicts with migrants. Because of the drought, transhumance begins around April instead of June, which has additional conse- quences like a lack of vaccination for livestock or children’s absences from school. In most of the south, agriculture is not developed and, wherever it exists, it depends almost totally on rainfall, which has been very irregular and sparse in the past few years. Irrigation techniques—in the rare case where they exist—and seeds and the varieties used may not be the most adaptable, and there is little know-how about these issues. Livestock is suffering the worst impacts, though it is undoubtedly the most important economic resource of the rural population. In Cunene, the communes most affected by the drought are those in the Cuvelai basin mentioned above. Massive loss of livestock—which is the base of their livelihoods—has occurred mainly in the last 5 years, even after adjusting transhumance patterns over time, while trying to find fodder and water. People are suffering from a lack of drinking water and water for the most basic needs. To cope with this situation, local governments are supplying water using tanker trucks—sometimes at no cost; in other cases, at high cost—and, in the most vulnerable settlements, some portable reservoirs have been installed for trucks to replenish them with water however periodically. In some of the isolated and more humid parts of the visited provinces (such as the communes of Chinquite or Mamue), the practice of rainfed agriculture is very important for the economy and at the same level as cattle breeding. This is in contrast with the rest of the region (such as in the adjacent comune of Chingo), where cattle breeding is predominant. In years with normal rainfall people do not need to move far to let their cattle graze. In fact, they receive transhumance from other areas, so invest- ment in these areas is especially relevant for regional resilience. Many of the deep boreholes in these areas do not work, and people rely on alternative sources of shallow groundwater and surface water. Before the intense drought, these villages were prosperous and their water resources seemed historically less impacted by droughts. However, their economic activities were severely disrupted by this drought, perhaps even more than other drier parts of the region which might have had better drought knowledge and more coping capacity. Another exception to the predominant pastoralism can be found in the lower sections of the largest wadis in Namibe, close to the mouths, where many livelihoods are based on a very productive irrigated horticulture that has evolved quickly to include new technologies and methods (photo 3.5). In these areas, the population is much more concentrated than in the rest of the region. They receive migrants from other provinces like Huíla and Benguela, and even from other countries like the Democratic Republic of Congo and Zambia. They experienced a noticeable decrease in productivity due to the drou- ght, but were certainly less affected than other parts of the south. 34 Water Security and Drought Resilience in the South of Angola PHOTO 3.5. A Drip Irrigation Plot In Bentiaba Source: World Bank. Minority communities of hunter-gatherers, who do not have cattle or practice agriculture, are likely to be among the first and most affected by the drought, as stated consistently by residents of various areas of the municipality of Camucuio. Local representatives stated4 that these communities had been relying on water from working pumps at the headquarters of local municipalities and communes for the entire duration of the drought event. In sum, the dispersed nature of settlements in most of the South requires the implementation of distri- buted solutions focusing on small communities. Notes 1. This proves the value of meteorological indices as predictors of hydrological issues and to determine the onset of drought-related problems in communities relying mainly on livestock and rainfed agriculture (also showed in other studies connecting the phenomenon and its impacts like Erfurt, Glaser, and Blauhut 2019. 2. According to oral communications by FAS and successive confirmations in the field, these trucks tend to use water bodies and aquifers that are far inland, in areas that were registering a more severe deficit than the Namibe coastal ones during the study period. 3. The rest of the region relies on trucks to supply drinking water only in emergencies. 4. April 19 communication. Water Security and Drought Resilience in the South of Angola 35 © Authoring team/World Bank Chapter 4 Building Drought Resilience in the South of Angola: Assessing Options for Rural Water Infrastructure Investments Selection of Most Suitable Infrastructure Options Thanks to the results of the stages described hitherto, the team was able to identify priority areas for which to plan follow-up knowledge activities in this third and last stage of the Advisory Services and Analytics (ASA). Stage 3 is essential for designing interventions to address both the current emergency situation and—even more important—the structural causes of drought vulnerability: pump repairs, small infrastructure of harvesting and storage, new wells, surveys, and institutional strengthening, among other key factors. Depending on resource availability, topography, soils and geology, different options for water supply and storage infrastructure were considered for further analysis across the visited priority areas. To guide future work at the community level, the team developed a conceptual framework, or decision-tree, for determining the infrastructure options to explore in the priority areas (see figure 4.1). The intention is to expand and develop this approach for detailed analysis across all of the priority areas. This process should also be informed by ongoing discussions with local residents about the infrastruc- ture best suited to the needs of each communities, possible alternatives, information required for ­ further planning, and the governance arrangements needed to make key investments and sustain their functionality. The framework for decision making on water supply investments, to be implemented in priority com- munes across the region, can be summarized as follows: 1. In areas with confirmed groundwater resources, priority should be given to wells for drinking water supply. If wells or boreholes with sufficient capacity already exist, these wells should be maintained (redeveloped/rehabilitated), and the equipment and pumps repaired/renewed. In case no wells exist, new wells should be planned and drilled. Relatively shallow groundwater bodies are cheaper to mobi- lize (possibly through open wells or cacimbas) and should be assessed first. If groundwater is present at intermediate depths, wells equipped with solar pumps should be installed. In the case of brackish or saline groundwater, the technical and financial feasibility of a small groundwater desalination plant could also be explored. 2. In areas with confirmed deep groundwater resources, deep wells with a depth up to 300 m can be drilled at strategic locations to supply larger communities or groups of small communities. These wells are not optimal for small communities because of their complexity and cost, but can supply larger communities or be strategic water points to provide water security for a larger area. All new Water Security and Drought Resilience in the South of Angola 37 FIGURE 4.1. Infrastructure “Decision-Tree” for Finding the Most Adequate Small-Scale Rural Water Supply Infrastructure to be Explored Further in Each Priority Area, Based on its Natural Conditions Confirmed GW resources? Yes No Fresh GW @ shallow depth? Conditions for underground storage? (< 10 m) Open well GW @ intermediate depth? Can sand dams meet demand? Can SW be harvested directly? (10 < 200 m) Is GW fresh? Can other communities benefit? Bore hole Sand dams Can other MAR options Only human WS? Identify options be implemented? to club WS with other Bore hole w/ desal unit communities, use tankers, Other MAR connect to structures Deep bore hole piped WS Cistern Chimpaca and cistern Note: GW = groundwater; MAR = managed aquifer recharge; SW= surface water; WS = water supply. boreholes should be equipped with associated supply infrastructure as needed. Many parts of the Cuvelai basin may fall in this category but hydrogeological studies need to be conducted to confirm deep groundwater potential. 3. In case shallow/intermediate groundwater is not available in sufficient quantities, the implementa- tion of managed aquifer recharge (MAR) techniques should be considered. In different parts of Namibe province, for example, where bedrock is close to the surface and limited amounts of loose sediment are available, the potential for constructing sand dams has been confirmed in the Giraul watershed as well as in the Chingo and Camucuio. Depending on the geology/morphology and flow regime, other MAR techniques like check dams, subsurface dams (SbSD), and diversion weirs can also be considered. 4. In case sufficient groundwater resources are not available to meet water demand, surface water har- vesting options should be evaluated.1 Considering the hydroclimatic conditions in the South of Angola, it would be difficult to build water security under drought conditions through the harvesting of surface water alone. Therefore, surface water harvesting may be complemented by other water 38 Water Security and Drought Resilience in the South of Angola supply options to face emergency situations. Nevertheless, in parts of the South a large potential for surface water harvesting exists (for example, in Cunene province), and this abundant resource should be mobilized to its maximum extent to meet the bulk of the water demand. The identification of the most adequate surface water harvesting structure will depend on the hydrological, topographical, and hydrogeological conditions. 5. Small weirs/dams (2–5 m maximum) or old dams from colonial times should be rehabilitated, and new small dams can be built in sites with propitious conditions. Most of this activity may consist of old dam rehabilitation. 6. If communities are close to piped bulk water supply and they can be reached by building an offshoot pipe, options should be explored to implement this solution. Using this decision-tree, two water supply options analyses were immediately pursued in two of the priority areas visited during stage 2: •• The Cuvelai basin (Cunene province), including Ombala Yo Mungo and other priority communes, where case study 1 did a deep analysis on how to improve the use of chimpacas and increase water security by embedding surface water harvesting in a broader rural water supply strategy; and •• The communities in the Giraul middle basin (Namibe province) and other adjacent streams around priority communes like Chingo, where case study 2 developed a prefeasibility study of the cons- truction of sand dams in the province. Both case studies, presented following, aimed to explore low-cost options to enhance shallow ground- water recharge and water harvesting, but each fell in different arms of the decision-tree/framework— due to their different geologic contexts—and therefore different types of structures and ideas are discussed and tested for both of them. Case Study 1: Options to Increase Water Security in the Cuvelai Basin Background Landscape Units The eastern part of the Cunene province includes part of the Cuvelai river basin, an endorheic trans- boundary river basin shared between Angola and Namibia (see map 4.1). Once the Cuvelai reaches the flat plains in the central part of the basin it transforms into a braided river, splitting into a large number of ephemeral channels and pans (or chanas) that converge and diverge and that flood seasonally, sepa- rated by slightly elevated islands (or mufilos). The channels form an inland delta that widens toward the border with Namibia. South of the border the channels converge toward Lake Oponono that occasiona- lly overflows, along with some smaller channel systems, into Etosha Pan. Photo 4.1 shows the typical landforms encountered in the lowland part of the Cuvelai basin and photo 4.2 shows a ground view of this landscape. Water Security and Drought Resilience in the South of Angola 39 MAP 4.1. Maps of Cunene Province (left) and the Cuvelai Transboundary River Basin (right) PHOTO 4.1. Google Earth View of the Study Area Landscape Units in the Angolan Part of the Cuvelai Basin Note: There is a clear distinction between active flood plains and channels, with isolated trees and filled with water-borne clays (gray soils), intermittently flooded areas with a denser tree cover and washed clayey to sandy soils (yellow soils) and islands composed of wind-blown sands with a dense natural tree and shrub vegetation (dark brown). 40 Water Security and Drought Resilience in the South of Angola PHOTO 4.2. Ground View of the Cuvelai Landscape Source: World Bank. Note: This shows the ochre to red sandy soils of the oasis or mufilo (foreground) and, at a slightly lower elevation, the flat and barren chana or channel with clayey soils and salt deposits (background). Geology A large part of the Cuvelai basin is underlain by Tertiary and Quaternary sediments that belong to the Kalahari Sequence (see photo 4.3). In the low-lying part of the basin in Cunene province, compact white marls from the Kalahari Limestone Group are uniformly covered by a thin layer of Quaternary sediments. Below the islands the sediment cover is composed of reddish iron-rich aeolian sands (see figure 4.3) with a thickness of 2–3 m. In intermittently flooded areas the sands have been bleached and are yellowish. Below channels and pans the sand cover has been eroded and the limestone is covered by 1–2 m of green to gray alluvial silt and clay, with white salt crusts resulting from the evaporation of flood waters. An overview of the aquifer layers within the Kalahari Sequence is included in Appendix A. Topography The flat lowland part of the Cuvelai located within Angola receives runoff from (part of) the upper Cuvelai catchment and drains toward the lowland part of the Cuvelai located in Namibia. This part of the Cuvelai basin drains toward Lake Oponono in Namibia, which overflows into Etosha Pan only during exceptionally wet years. The subbasin covers an area of nearly 36,000 square kilometers (km2) and effec- tively represents an endorheic basin. The Cuvelai lowlands, situated at an elevation between 1,080 and 1,125 m above sea level, are extremely flat, with islands that are just elevated a few meters above the numerous channels. The terrain has a regional gradient of about 0.3 o/oo. To the north they are bordered by a pediplain that is less intersected, with a regional gradient of about 0.6 0/00. The upper part of the catchment is hilly between 1,160 m and 1,475 m above sea level. Water Security and Drought Resilience in the South of Angola 41 PHOTO 4.3. Geological Map of the Cuvelai-Etosha Basin (left); Quarry Near Ombala Yo Mungu (right) Source: Dill et al. 2012. Note: The map (left) shows the Tertiary (Neogene) and Quaternary sediments of the Kalahari Group that are uniformly present in the central part of the Cuvelai-Etosha basin. In the photo of the quarry (right), red iron-rich aeolian sand cover can be seen on top of white marls of the Kalahari Group. A topographical cross-section along the profile line shown in map 4.2 is shown in figure 4.2. For the purpose of assessing the water resources the subbasin has been divided in three subbasins that corres- pond to the hilly Upper Catchment (including the pediplain), the lowland part of the basin situated within the territory of Angola, and the lowland part of the basin in Namibia. Climate Daily precipitation (P) totals derived from 3-hour Tropical Rainfall Measuring Mission (TRMM) Multi- satellite Precipitation Analysis (TMPA) data are available at a 0.25° spatial resolution for the period 2003 to the present (https://gpm.nasa.gov, 3B42 Research Version). TRMM pixels measure approxima- tely 26.5 x 27.5 km at the latitude of South Angola. An overlay of the satellite-derived meteorological data grid and the three subbasins used in this study is shown in map 4.2. Based on the pixel data sets, average monthly P values for each of the three subbasins have been calculated, and the time series are shown in figure 4.3. 42 Water Security and Drought Resilience in the South of Angola MAP 4.2. Subbasins of the Cuvelai Basin used in this Study FIGURE 4.2. North-South Cross-Section of the Western Part of the Cuvelai Basin, from the Headwaters till Lake Oponono Min, Avg, Max Elevation: 1,079, 1,188, 1,472 m South North Elevation above mean sea level (m) Range totals: Distance: 378 km Elev gain/loss: 1,300 m, –922 m Max slope: 2.9%. –2.0% Avg slope: 0.3%, –0.3% 1,472 m 1,425 m Lowland with chanas and islands Pediptain Dissecleted upper catchment 1,350 m 1,275 m Lake oponomo Namibia Angola 1,200 m 1,125 m 1,125 m 1,079 m 0.0% 172 km 50 km 100 km 150 km 200 km 250 km 300 km 378 km Longitudinal distance (km) Note: For a map of this area, see figure 4.2. Water Security and Drought Resilience in the South of Angola 43 FIGURE 4.3. Monthly Rainfall in the Upper Catchment, Angola Lowland, and Namibia Lowland Subbasins, 2003–19 Monthlu rainfall for upper, middle and lower Cuvelai basin 500 450 400 350 Monthly rainfall (mm) 300 250 200 150 100 50 0 03 04 04 06 07 08 09 10 11 12 13 14 15 16 17 18 2/ 2/ 2/ 2/ 2/ 2/ 3/ 3/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 2/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ /3 1/ 1/ 1/ 12 P upper catchment P Lowland angola P Lowland namibia Rainfall decreases from North to South across the basin, following a similar pattern during dry and wet years (see map 4.2). Mean annual rainfall for the three subbasins for the period 2003–18 amounts to 614 millimeters per year (mm/yr), but interannual variability is high. Annual rainfall received by the basin was as low as 360 mm/yr in 2015 and as high as 1,008 mm/yr in 2009. Rainfall is concentrated over a 6-month rainy period, generally from early October till late March/early April. The months of May to September account for less than 1 percent of the annual rainfall. Monthly actual evapotranspiration (ETa) data sets from the Global Land Evaporation Amsterdam Model (GLEAM) (https://www.gleam.eu) are available for the period 2003–18. Actual ETa in the Cuvelai Basin is high and, during a year of average rainfall, ETa amounts to 96 percent of the rainfall received, confirming the enclosed character of the basin, with almost no in- or outflow. During dry years, the ETa can represent up to 130 percent of the annual rainfall received by the basin, depleting any remaining stocks of water in the basin. The climate of the Cuvelai basin is characterized by recurrent drought episodes. The Standardised Precipitation Evapotranspiration Index (SPEI) is a drought index based on climatic data. It can be used for determining the onset, duration, and magnitude of drought conditions with respect to normal conditions. Figure 4.4 shows the SPEI index for different reference period durations (3, 6, 12, 24, and ­ 48 months) for the Ombala Yo Mungo commune in the central part of the Cuvelai basin (see figure 4.1). Starting from 2013 this part of the Cuvelai basin has entered a period characterized by nearly continuous moderate-to-severe drought conditions. 44 Water Security and Drought Resilience in the South of Angola FIGURE 4.4. SPEI Drought Index at Different Time Scales for Ombala Yo Mungu 3 2 1 SPEI Index 0 –1 –2 –3 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ Date SPEI_48 SPEI_24 SPEI_12 SPEI_6 SPEI_3 Note: See figure 4.1. SPEI = Standardised Precipitation Evapotranspiration Index. Water supply challenges in the Cuvelai Use of Surface Water The population of the Cuvelai basin is vulnerable to climatic variations due to the lack of water security caused by the absence of secure water points. Parts of the Cuvelai rely almost entirely on the harvesting of surface water for human and animal water consumption—relying on the seasonal flooding during the rainy season that occurs between October and March—in chimpacas, or tanks (see photo 4.4). Notwithstanding the seasonal abundance of surface water, harvesting of flood waters is challenging due to the extremely flat topography of the Cuvelai basin, allowing only for shallow reservoirs that have high siltation rates due to the abundance of fine-grained sediments in the Cuvelai inner delta and that are exposed to high evaporation losses (the Cuvelai lowlands are situated between 16° and 18°S), reducing their live storage. Because of the limited active storage capacity, most chimpacas provide limited water security, often sufficient to bridge the dry season in a year with average rainfall, but falling dry during periods of below-average rainfall or in case rains are delayed. Besides the limited water security provided by chimpacas, they also pose a high health risk due to fre- quent contamination. Most chimpacas provide water to human populations and cattle, especially during drought periods when the number of functioning water points dwindles. Cattle will enter the reservoir to drink, resulting in microbial and pathogen contamination. Efforts to keep animals out, by fencing off the water reservoir and its immediate surroundings, often fail as fences are vandalized to let animals drink. Chimpacas are typically positioned in the downstream parts of chanas or pans. Usually they collect the runoff from the adjacent part of the impervious floodplain, rather than from active channels that cross Water Security and Drought Resilience in the South of Angola 45 PHOTO 4.4. Aerial and side views of a Cuvelai chimpaca Source: World Bank. the Cuvelai inner delta. Floodplains are created by the deposition of clay and silt in the low-lying chan- nel or depression and may have salt crusts at the surface as a result of the repeated evaporation of stag- nant flood waters. Chimpacas are constructed by excavating the clayey infill of the floodplain or pan, over a depth of a few meters, often till the top of the limestone substratum. As both the clayey sediments and the underlying limestone are impervious, infiltration losses are limited. Typical dimensions for modern chimpacas are 80 m by 60 m, with a depth of around 4 m, representing a maximum storage capacity of about 20,000 m3. Based on estimates of open water evaporation, a full chimpaca loses about half of its storage over a 12-month period. The life storage of chimpacas is further reduced by the deposition of sediments, carried by each flood. No systematic desilting practices exist in the Cuvelai, and chimpacas are often abandoned when entirely filled up with sediment. When no other water points are available in the vicinity, villagers resort to dig- ging cacimbas in the sediment infill from which very small quantities of water are then harvested. Use of Groundwater Across the Cuvelai, shallow groundwater (encountered at less than 10 m depth) is accessed through open wells or cacimbas (unlined shallow wells; see photo 4.5) tapping thin perched aquifers or semi-per- vious layers, typically with a very low yield and vulnerable to drought episodes. Cacimbas typically serve one or several families. Often the water stored in the cacimbas is used during the day and the well slowly fills up during the night. The use of groundwater through deeper boreholes is limited because of the abundance of brackish groundwater at intermediate depths (10 < 200 m depth) and the generally low yield from the marly limestone Kalahari Formation uniformly present below the Cuvelai lowlands. Where yield is sufficient, boreholes have traditionally been equipped with Volante handpumps (see photo 4.6). Due to their robustness, a large share of the Volante pumps is still in use. 46 Water Security and Drought Resilience in the South of Angola PHOTO 4.5. Typical cacimba in a Kimbo near Ohenghali, Mongua Commune (left); Open Well at Chiulo, Mukope Commune (right) Cacimba near Mongua (Cunene) Cacimba near Chiulo (Cunene) Source: World Bank. PHOTO 4.6. Borehole with Volante Handpump Near Onjiva (left); Borehole Equipped with Solar Pump and Storage Reservoir at Omambodi, Mongua Commune (right) Volante near Ondjiva (Cunene) Borehole with solar pump near Mougua Source: World Bank. Water Security and Drought Resilience in the South of Angola 47 Most recently, drilled wells are equipped with solar pumps and overhead tanks of 5–10 cubic meters (m3). A significant number of wells with solar pumps are out of order due to pump or solar panel failure. No well-established mechanism for maintenance and repair of pumps exists, and wells often remain out of order for months or even years before they get fixed as part of a well rehabilitation. The knowledge about the quality and dynamics of groundwater in the Cuvelai is extremely limited and drilling data on existing boreholes is not available, hampering more systematic planning and develop- ment. The failure rate of boreholes is very high because of the lack of reference information and the practice of drilling without preliminary geophysical investigations. The presence, potential, and quality of deep groundwater resources (> 200 m depth) need to be confir- med as no systematic study of the groundwater resources below the Angolan part of the Cuvelai basin has been performed. Field Survey Design The series of field visits conducted in April 2019 as part of Stage 2 of this study was aimed at verifying drought impacts across Namibe, Huíla, and Cunene provinces through stakeholder consultations. Following the selection of priority areas, a field survey was conducted in the central part of Cunene pro- vince in July 2019 as part of Stage 3 of the work plan, with the specific objectives to: •• Check chimpaca designs and functionality to assess opportunities to optimize storage, strengthen hygienic conditions, and improve reliability in face of drought conditions; •• Verify surface geological conditions across the region, in view of enhancing surface water harves- ting potential and to evaluate groundwater recharge conditions; •• Survey cacimbas and boreholes to evaluate current groundwater use and potential; •• Measure basic surface- and groundwater quality parameters. During the 3-week field survey a technical team supported by Gabinete para Administração das Bacias Hidrográficas do Cunene, Cubango e Cuvelai (GABHIC), Universidad Agostinho Neto, Instituto Superior Politécnico Tundavala, and the World Bank collected data on water points and surface geology, covering the Angolan part of the Cuvelai basin. A total of 54 sites was visited (see map 4.3), at which discussions were held with representatives of the commune and other stakeholders. At each location a series of observations and measurements were made: •• Site description: Type of infrastructure, longitude, latitude, elevation (meters above sea level); •• Water point description (including chimpacas, cacimbas, open wells, and boreholes): Year built, depth, diameter/dimensions, volume stored, energy source, type of use, daily consumption, operational status, months of operation per year; ­ 48 Water Security and Drought Resilience in the South of Angola MAP 4.3. Location Map of Sites Across the Cuvelai Basin Visited During the Field Survey •• Water resources’ condition: Static water level, dynamic water level, temperature, salinity (electrical conductivity); and •• Soil and subsurface geology: Soil type, soil thickness, nature of subsurface geological formation(s), observed thickness of formation. Observations on the Subsurface Geology The entire region is uniformly underlain by white marly limestone with a low permeability. The limes- tone poses a barrier to any water that infiltrates the top soil and contains limited amounts of groundwa- ter that is often brackish. The islands and channels forming the Cuvelai lowlands represent a uniformly thin cover of limestone, composed respectively of windblown (red to yellow) sands and impervious water-born clays, rarely exceeding 3 m in thickness. Water Security and Drought Resilience in the South of Angola 49 Natural groundwater recharge occurring on the sandy islands (covering about 80 percent of the area) will largely remain in the sandy top layer and function as a reservoir for the abundant vegetation of the region. Most of the recharge water will return as evapotranspiration to the atmosphere during the dry season. Observations on the Hydrogeology Chimpacas are typically excavated in the clayey channel infill, till the top of the limestone substratum, and therefore have low infiltration losses. The limited thickness of the top sand layer (and its position at the same elevation as the chimpaca) make the underground storage of chimpaca water impractical. Cacimbas typically tap small perched aquifers in the sandy topsoil or sandy horizons in semiactive floo- dplains but their potential is generally very limited. Boreholes are typically drilled into the underlying limestone layers and have a generally modest yield and variable salinity. No records are kept of failed or successful boreholes and no groundwater monito- ring data are available in the Angolan part of the Cuvelai-Etosha basin. According to groundwater studies in the Namibian part of the Cuvelai-Etosha basin (BIWAC 2006), the deep-seated multilayered Kalahari Aquifer is recharged in Angola, and groundwater flows in a southern direction toward the Etosha Pan and the Okavango River. Several shallow aquifers of the Kalahari sequence superimpose the deeper Kalahari aquifer. The partia- lly saline groundwater originates from regular floods in the Cuvelai basin. Observations on Water Quality Salinity measurements (expressed in electrical conductivity, or EC; see table 4.1) of 28 water points, taken in July of 2019 in the Cuvelai basin (photo 4.7) during the elaboration of this report, reveal that: •• The best water quality is found in springs (EC = 205 microsiemens per centimeter, μS/cm) and cacimbas (average EC = 503 μS/cm). •• Chimpacas have a low salinity level when they fill up but this salinity gradually increases during use as a result of evaporation (average EC = 670 μS/cm). The main quality-related concern regarding TABLE 4.1. Range of Electrical Conductivity Values for Different Types of Water Points Type of Minimum Ec Maximum Ec Standard Deviation Waterpoint No of Samples (μs/Cm) (μs/Cm) Mean Ec (μs/Cm) (μs/Cm) Borehole 9 426 1,481 932 402 Chimpaca 8 234 1,489 671 357 Cacimba 10 91 999 503 299 Spring 1 205 205 205 0 Total 28 91 1,489 400 678 Note: EC = electrical conductivity. 50 Water Security and Drought Resilience in the South of Angola PHOTO 4.7. Soil and Subsurface Geology Observations Near Ombala Yo Mungu Sede (left); Groundwater Level Measurement Near Mongua Sede (middle); and Water Quality Measurement in Naulila Commune (right) Source: World Bank. chimpacas is bacteriological and pathogen contamination due to the frequent presence of animals. •• Higher salinity values (and higher variability) are encountered in wells across the region (average EC = 930 μS/cm). •• The highest salinity values encountered are just below 1,500 μS/cm, equivalent to a total dissolved solids (TDS) value of about 960 milligrams per liter (mg/L). Although there is no World Health Organization (WHO) drinking water guideline for TDS, all samples are below the 1,000 mg/l refe- rence value adopted by a number of countries. Establishing the Water Balance Approach and Assumptions To study the surface water harvesting potential, a preliminary water balance was established for the basin. In order to gain better control of some water balance components, the water balance was compi- led for three subbasins (see map 4.2): the upper catchment, the lowland part of the basin within Angola, and the lowland part of the basin located in Namibia that drains toward Lake Oponono. Considering that Lake Oponono overflows into Etosha Pan only during exceptionally wet years, the three subbasins combined effectively represent an endorheic basin: surface water inflow and outflow are considered to be negligible. Groundwater flow in or out of the basin is difficult to assess in the absence of more detailed knowledge about the geometry and hydraulic characteristics of aquifers and a groundwater contour map. Pending Water Security and Drought Resilience in the South of Angola 51 a more detailed assessment of the groundwater conditions, groundwater flow out of the basin is consi- dered equal to the amount of groundwater recharge within the subbasins. ETa during the dry season (April to September) is fed by water stored in the basin during the preceding wet season (October to March). Water storage in the basin occurs in the form of open water in channels and depressions, as soil moisture and by vegetation. Considering the scarcity of monitoring data, various components of the water cycle can only be estima- ted. Notwithstanding this limitation, the water balance based on monthly monitoring data allow us to understand the seasonal dynamics of the subbasins. For the two main components of the water balance—that is, rainfall (P) and ETa—monthly pixeled data series are available from remotely sensed data. Excess rainfall (P-ETa) during the rainy season is the reservoir for dry season ETa. Calibration of the water balance includes matching the wet season P-ETa excess and dry season P-ETa deficit. During the wet season excess rainfall (P-ETa) will: 1. Run off within the basin and be stored as flood water in the channels and ponds (ΔSlake); 2. Infiltrate in the sandy soils of the “islands” (ΔSsoil), from where part of the infiltrated soil moisture may percolate further down as groundwater recharge (Igw); 3. Be absorbed by the vegetation in the basin (ΔSveg). During the following dry season the water stored in the basin will gradually dry up, through evapotrans- piration of water stored in the channels and ponds, soil moisture, and from vegetation. Validation of Meteorological Data A water balance has been prepared for each of the three subbasins, using satellite-derived precipitation (P) and ETa data. Daily P totals derived from 3-hour TMPA data are available at a 0.25° spatial resolution for the period 1998 to the present (https://gpm.nasa.gov, 3B42 Research Version). TRMM pixels measure approxima- tely 26.5 x 27.5 km at the latitude of South Angola. An overlay of the satellite-derived meteorological data grid and the three subbasins used in this study is shown in map 4.4. Daily P data for the period 1998–2018 have been used for infrastructure design analysis (next chapter). Monthly ETa data sets from the GLEAM model (https://www.gleam.eu) are available for the period 2003–18 and have been used, along with TMPA monthly precipitation data, for water balance calculations. ETa data sets using the SSEBop and GLEAM models have been used. Based on the water balance calcu- lations for the endorheic basin, the SSEBop data appear to overestimate the dry-season ETa (see figure 4.5). Therefore, ETa data from GLEAM have been used. 52 Water Security and Drought Resilience in the South of Angola MAP 4.4. Satellite-Derived Meteorological Grid Based on the pixel data sets, average monthly ETa and P values for each of the three subbasins have been calculated. In the Cuvelai basin rainfall occurs from October till April, but only during the period between December and March does monthly rainfall exceed the monthly ETa. As the three subbasins together form an endhoreic basin, the only water balance component potentially leaving the basin, besides ETa, is lateral groundwater flow, induced by groundwater recharge ­ (Igw in photo 4.8). Groundwater Recharge To assess groundwater recharge (and groundwater flow) groundwater conditions in the basin need to be mapped, as currently no groundwater level and groundwater quality data are being collected systema- tically. In the absence of groundwater monitoring data, recharge estimates have been obtained from global scale modeling data (Döll and Flörke 2005). According to the Hydrogeological Map of Africa (see Water Security and Drought Resilience in the South of Angola 53 FIGURE 4.5. Precipitation in Upper Catchment 200 Mean monthly P and ET data for upper catchment (2003–18) 180 160 Monthly P and ET (mm) 140 120 100 80 60 40 20 0 y y ch ril ay ne ly st r er r r ar ar be be be Ap Ju gu ar ob M Ju nu ru m m em M Au ct b ve ce Ja Fe pt O De No Se Month of the year Rainfall ETa (SSE Bop) ETa (GLEAM) Note: P-ETa = precipitation minus the actual evapotranspiration. PHOTO 4.8. Main Components of the Cuvelai Subbasins’ Water Balance Considered in this Study P ETa Rin Sveg R Isoil Rout Slake Ssoil Igw map 4.5), groundwater recharge in the Angolan part of the Cuvelai basin ranges from 20 mm/yr in the southeast to 70 mm/yr in the north of the basin. Mapping of subsurface geological conditions during the field survey revealed the consistent presence of a thick impervious marl layer below the thin top sediment layer. As a result, conditions for recharging groundwater below the Cuvelai lowlands are unfavorable. 54 Water Security and Drought Resilience in the South of Angola MAP 4.5. Detail from the Hydrogeological Map of Africa at 1/10 M Scale Showing Modeled Groundwater Recharge Values Source: Seguin 2008. Water Security and Drought Resilience in the South of Angola 55 In the Cuvelai basin the shallow topsoil absorbs excess rainwater during the wet season (Isoil) that is sub- sequently released during the dry season, through evaporation from soils and transpiration from plants. It is anticipated that only a very small share of the soil water (a few mm/yr) manages to percolate down to the aquifer (Igw). Extensive groundwater studies conducted in the Namibian part of the Cuvelai basin suggest that the Oshana aquifer is mainly recharged by regular flooding of the Cuvelai drainage system (BIWAC 2006). The groundwater gradient is reportedly very flat, indicating low recharge dynamics. Seasonal Dynamics and Climate Variability Excess rainfall during the wet season either infiltrates the top soil layer, or runs off into channels and pans. Average monthly P-ETa maps for Southern Angola (figure 4.3) show the excess rainfall during the December–March period. For the low-lying part of the Cuvelai basin inside Angola (covering approxima- tely 10,200 km2), the average cumulative excess P for the rainy season from December to March amounts to ~224 mm/yr (period 2003–18). As the graphs in figure 4.6 show, ETa is maintained during the start of the dry season, fed by transpira- tion from green vegetation, evaporation of soil moisture, and open water evaporation from flooded channels and pans. Starting from July the ETa declines when vegetation dries and channels and ponds dry up. Water balance calculations show that, over the 16-year observation period, 89 percent of the wet season excess rainfall in lowland Angola is retaken by ET during the dry season. As the graphs of mean monthly P and ET for the three sub-basins show, there is a regional trend of both P and ETa decreasing from North to South. Besides this spatial variability there is also a strong interan- nual variability in P and ETa (see figure 4.7). Annual P for the three catchments combined fluctuated over the 2003–18 period between 360 mm/yr and 1,008 mm/yr, the mean being 614 mm/yr. The last 6 years of the record show below average rainfall. During dry years ETa can be higher than P, suggesting that some water stored in the basin (in channels, soil, or vegetation) remains at the end of the dry season. At the same time, there is a fairly good correla- tion between annual P and annual ETa (see figure 4.8) indicating that actual ET depends to a significant degree on the rainfall received by the (sub) basin during the same period and cannot rely on significant storage in the basin and is not influenced by permanent water sources like rivers, lakes, and irrigated fields. Constrained by 16 years of spatially differentiated monthly P and ETa data, together with estimates of seasonal soil water storage, runoff to downstream subbasins, and magnitude of groundwater recharge, a tentative water balance has been drawn for the three subbasins. Within the error margin of the P and ET data, the share of annual rainfall that is not taken back by ET during the dry season represents 4 percent of the annual rainfall received by the three subbasins. This volume of water leaves the basin, either as groundwater flow or occasional outflow from Lake Oponono during excep- tionally wet years. 56 Water Security and Drought Resilience in the South of Angola FIGURE 4.6. Mean Monthly Precipitation (P), Actual Evapotranspiration (ETa), and P—ETa Values for the Three Cuvelai Subbasins a. Upper catchment (2003–18) b. Lowland angola 200 200 P, P-ET, and ETa (mm) 150 150 P, P-ET, and ETa (mm) 100 100 50 50 0 0 –50 –50 –100 –100 ch ne pt st ch ne pt st ve r ve r Fe ry ry A y O er r Fe ry ry Au y O er ce r r ce r ril ay ril ay e No be be De be be De be l l Se ugu Se gu Ju Ju ob ua ua a a b b ar ar Ju Ju Ap Ap M M nu nu o m m m m em em M M br br ct ct Ja Ja No Month of the year Month of the year c. Lowland namibia 200 150 P, P-ET, and ETa (mm) 100 50 0 –50 –100 ch ne pt st ve r Fe ary ry A y ce r r O er ril ay e De be be l Se ugu Ju ob ua b ar Ju Ap M nu m m em M br ct Ja No Month of the year Rainfall P-ET (GLEAM) Eta GLEAM The indicative water balance is shown in table 4.2 and allows the evaluation of the part of surface runoff that remains in the subcatchments, flooding channels and pans, and that can potentially be harvested during a year with average rainfall. Water Harvesting Potential Study of the tentative water balance of the Cuvelai basin shows us that in a year with average rainfall the amount of flood waters accumulated in the lowland part of the basin and within the Angola territory amounts to some 360 Mcm, equivalent to a ± 20 cm layer of water covering all the channels and pans. During the months following the rainy season this volume of water is available for certain uses but if not harvested this resource will evaporate and will not provide water security year-round. Moreover, due to quality concerns this resource is not to be used for human consumption, unless har- vested under suitable conditions shortly after rainfall events. Water Security and Drought Resilience in the South of Angola 57 FIGURE 4.7. Average Annual precipitation (P) for the Three Combined Cuvelai Subbasins, 2003–18 Annual P and ETa for the 3 sub-basins combined (period 2003–18) 1,000 800 Annual P and ET (mm) 600 400 200 0 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 Year Rainfall ETa (GLEAM) Note: ETa = actual evapotranspiration. FIGURE 4.8. Correlation between annual P and Annual ETa for the Angola Lowland Subbasin, 2003–18 800 750 R2 = 0.6017 700 Actual evapo-transpiration (mm/yr) 650 600 550 500 450 400 350 300 300 400 500 600 700 800 900 1,000 Rainfall (mm/yr) 58 Water Security and Drought Resilience in the South of Angola TABLE 4.2. Tentative Water Balances for the Three Cuvelai Subbasins, 2003–18 Average P and ET Sub-basin Upper Cuvelai Angola Lowland Namibia Lowland 3 Sub-Basins Basin area (km ) 2 13,122 10,206 12,393 35,721 Dryland (%) 95 80 80 86 Mean annual rainfall (mm/yr) 765.9 593.9 470.0 614.1 Eta GLEAM (mm/yr) 726.9 582.6 447.9 588.9 Excess P-ET (rainy season) (mm/yr) 289.5 224.0 187.4 235.4 Deficit P-ET (dry season) (mm/yr) −236.9 −199.4 −152.8 −197.0 Deficit = % Excess 81.8 89.0 81.5 83.8 Infiltration rate (%P) @ dryland 36 44 40 39.4 Infiltration I (mm/yr) @ dryland 276 261 188 182.4 Groundwater recharge (%P) @ dryland 2.5 1.0 5.0 2.9 Groundwater recharge (mm/yr) @ dryland 19 6 23 17.0 Soil water reservoir (mm/yr) 257 255 164 165 Channels and depressions 5 20 20 14 Net Surface Water Availability (mm/yr) 28 15 37 27.2 Net Surface Water Availability (MCM/yr) 362 152 458 973 % runoff leaving sub-basin 75 15 7 4 Runoff leaving basin (MCM/yr) 272 64 37 37 Harvesting potential (MCM/yr) 91 360 485 936 Floodable land (km ) 2 656 2,041 2,479 5,176 Inundated depth (mm) 138 177 196 169.1 Wet season storage in basin (mm/yr) @ basin 251 240 171 166 Consumptive uses (mm/yr) @ basin 0 0 0 0 Groundwater recharge (mm/yr) @ basin 18 5 19 15 Runoff out of basin (mm/yr) @ basin 21 6 3 1 Inter Basin Transfer (IBT) (MCM/yr) 0 0 0 0 P − ETa − R − G + IBT (year) 0 0 0 10 Dry season water balance (mm/yr) @ basin 14 40 18 −31 Note: See map 4.2. P-ETa = precipitation minus the actual evapotranspiration; GLEAM = Global Land Evaporation Amsterdam Model. Water Security and Drought Resilience in the South of Angola 59 Due to terrain morphology, the connectivity and dynamics of drainage channels, and land use as well as technical and financial constraints, the share of floodwaters that can actually be harvested is much sma- ller. Suitable areas need an impluvium of adequate size, slope, terrain configuration, land use, and soil/ substratum type, or have to be connected to active channels or local drainage systems that flood regularly. In the absence of a hydrological model of the Cuvelai basin, a Digital Elevation Model (DEM), and soil and land use maps, it is assumed that 10 percent of the flood waters can potentially be harvested. The resource for water harvesting, which can be counted during a year of average rainfall, is therefore estimated at 10 percent of 360 Mcm or about 36 Mcm/yr. Considering that at present chimpacas in the Angolan part of the Cuvelai harvest approximately 3 Mcm/yr, it is expected that there is scope for expan- ding surface water harvesting. Considering Climate Variability In the Cuvelai basin there is significant interannual variability of rainfall. This variability needs to be considered when designing robust water harvesting infrastructure that can be relied on most of the year. For this purpose, P and ETa frequency curves have been prepared for each of the three subbasins, using the 2003–18 datasets (example shown figure 4.9). Similar frequency curves have been prepared for the “Wet-season P-ETa excess” and “Dry-season P-ETa deficit” values. For example, figure 4.10 shows the annual exceedance probability (AEP) for the cumulative excess of P-ETa during the 6-month wet season (October to March) for the years 2003–18. Table 4.3 summarizes the AEP for different probability values, for each of the four climate parameters used in the water balance calculations, for each of the three subbasins. Plotting rainfall and ETa frequency curves (see figure 4.11) for the subbasins shows that in years with less than average rainfall, ETa quickly overtakes the amount of rainfall received in the same year (that is, AEP equals about 45 percent). This implies that in any year with below average rainfall the evapotranspiration is sustained from water available in the basin, resulting in a reduction of moisture stored in vegetation and soils and likely of groundwater recharge. Runoff will be reduced, leading to a reduction in the harvesting potential and stagnant water in channels and pans will evaporate more quickly. Tentative water balances have been drawn for the three subbasins using the P, ETa, “Wet-season P-ETa excess,” and “Dry-season P-ETa deficit” values at 60 percent (3 years out of 5), 70 percent, 80 percent (4 years out of 5), and 90 percent (9 years out of 10) probability of exceedance. Results are attached in Appendix B. Estimated surface water harvesting potential for the Angola lowland subbasin at different AEP values are summarized in table 4.4. Although with the lack of hydrological data water balances are only indicative, it is evident that the surface water harvesting potential reduces quickly during dry years. ­ 60 Water Security and Drought Resilience in the South of Angola FIGURE 4.9. Annual Exceedance Probability of Annual Precipitation Values, Lowland Angola Annual rainfall frequency curve (Lowland Angola) 1,200 1,100 1,000 900 800 Annual rainfall (mm) 700 600 500 400 300 200 100 100 90 80 70 60 50 10 Annual exceedance probability (%) Source: Based on 2003–18 data. In addition, as storage of water in the basin is limited the potential will further decrease during a second or third drought year in a row, when any residual stocks of water in the basin will have been exhausted. Planning of Infrastructure The choice of infrastructure to be built in each community will largely depend on the availability, varia- bility, and quality of the water resources that can be mobilized in the vicinity, as well as of the total water demand, for human consumption and other needs. It requires sufficient knowledge of the groundwater and surface water regimes of the region, as well as engagement with the community to assess demand and discuss governance arrangements. A Decision Tree Framework for Rural Water Supply and Harvesting infrastructure is presented in section 3.1 of this report. Guiding principle of the Decision Tree is to provide the most reliable and ­ safe water supply source, or combination of sources, in view of prevailing climate variability in South Angola. Use of Groundwater Groundwater is the preferred water supply option, because it tends to be less vulnerable to short-term climatic variations and is largely protected from anthropogenic pollution (including agriculture and Water Security and Drought Resilience in the South of Angola 61 FIGURE 4.10. Annual Exceedance Probability of “wet season excess P-ETa” Values for Lowland Angola Annual wet season excess P-ETa frequency curve (Lowland Angola) 600 500 Annual west season P-ETa excess (mm) 400 300 200 100 0 100 90 80 70 60 50 10 Annual exceedance probability (%) Source: Based on 2003–18 data. TABLE 4.3. Annual Exceedance Probabilities for Different Climate Parameters (Upper Catchment) Annual Exceedence Probability 50% 60% 70% 80% 90% Upper Catchment Rainfall (mm/yr) 700 650 610 565 530 ETa (mm/yr) 720 670 620 585 550 Wet season excess P-ETa (mm/yr) 240 205 170 140 115 Dry season deficit P-ETa (mm/yr) −245 −255 −265 −270 −275 Angola Lowland Rainfall (mm/yr) 540 490 455 420 390 ETa (mm/yr) 555 535 520 500 490 Wet season excess P-ETa (mm/yr) 185 150 120 95 70 Dry season deficit P-ETa (mm/yr) −210 −220 −228 −236 −243 Namibia Lowland Rainfall (mm/yr) 420 375 345 310 290 ETa (mm/yr) 425 405 385 370 360 Wet season excess P-ETa (mm/yr) 145 115 90 65 45 Dry season deficit P-ETa (mm/yr) −167 −180 −190 −200 −207 Source: Based on 2003–18 data. Note: P – ETa = precipitation minus the actual evapotranspiration. 62 Water Security and Drought Resilience in the South of Angola FIGURE 4.11. Annual Exceedance Probability “wet season P-ETa excess” Values for Lowland Angola 1,200 Annual rainfall and ETa frequency curve (Lowland Angola) 1,100 1,000 900 Annual rainfall (mm) 800 700 600 500 400 300 200 100 100 90 80 70 60 50 10 1 Annual exceedence probality (%) Rainfall ETa GLEAM Log. (Rainfall) Log. (ETa GLEAM) Source: Based on 2003–18 data. Note: ETa = actual evapotranspiration. TABLE 4.4. Surface Water Harvesting Potential for the Angola Lowland Subbasin, at Different AEP Values Lowland Angola Surface Water Harvesting Potential Rainfall Received Mcm/Yr % Average 360 100% 60% AEP 277 77% 70% AEP 202 56% 80% AEP 169 47% 90% AEP 103 29% Source: Based on 2003–18 data. Note: AEP = Annual exceedance probability. husbandry). However, groundwater conditions vary across the Cuvelai basin. Shallow groundwater resources, that are tapped through cacimbas and open wells, have a limited potential. Therefore, groundwater studies need to confirm where intermediate or deep aquifers are available and groundwa- ter quality is acceptable, in order to assess where and to what extent groundwater can sustainably meet the required water demand. Besides the urgent need to map the groundwater resources across the Cuvelai basin, to monitor ground- water conditions to establish their dynamics, and to assess their potential it is also required to conduct Water Security and Drought Resilience in the South of Angola 63 systematic site investigations in preparation of the drilling of new boreholes, to increase their success rate, reduce their cost, and increase their sustainability. Results of groundwater studies, groundwater monitoring data (including groundwater levels, groundwater quality, and groundwater uses), and data on the technical details and operational status of wells/boreholes need to be publicly available, in order to support the development of this strategic resource and to ensure its sustainability. Based on the current (limited) knowledge of the groundwater resources in the Cuvelai basin it appears that the overall groundwater potential at shallow to medium depths (0 < 200 m) is limited and that groundwater cannot be tapped anywhere. For this reason, groundwater can be considered as a strategic and complementary resource, providing safe drinking water for human consumption and water secu- rity during drought episodes, while as large as possible a share of the water demand can be met from surface water harvesting. Currently a large share of existing boreholes is out of order for extended periods of time because of failure of the pump, solar panel, or other parts or because of deterioration of the borehole yield or reduced resource availability. Setting up of local Water Committees, responsible for the collection of user fees and in charge of maintenance and small repairs, will reduce the break- down rate of boreholes and ensure quick repairs (see photos 4.9 and 4.10). For larger repairs, planning, funding, and technical capacity needs to be developed at the local level (commune and province). Evidence from Namibia suggests that at a larger depth (> 200 m) deep groundwater resources are available under parts of the Cuvelai basin (photo 4.13). Even if regional studies (including at the ­ PHOTO 4.9. Borehole Equipped with solar Pump and 5 m3 Storage Reservoir at Ondjiva Sede Source: World Bank. Note: The borehole is managed by a local water committee (left); water point manager, appointed by the committee, with records of fees paid by users (right). 64 Water Security and Drought Resilience in the South of Angola PHOTO 4.10. Borehole Equipped with Volante Pump for Community and Husbandry Water Supply at Omulova, Namakunde Commune Source: World Bank. Note: The pump is managed by the local water committee. transboundary level) confirm the deep groundwater potential, this does not significantly alter the priorities for rural water supply in the Cuvelai. Tapping groundwater from deep aquifers is costly in development (high cost of boreholes and pumps) and in operation and maintenance (high cost for pumping and spare parts). Deep groundwater will be more expensive than the available low- cost water supply solutions and require more specialized expertise and capacity that is most likely not available locally. Use of Surface Water At present surface water represents by far the largest confirmed water resource in the Cuvelai basin. In order to further develop the surface water harvesting potential in the region it is required to conduct studies that will: 1. Map the active channels (see photos 4.11) and extent of seasonally flooded areas for a sufficient num- ber of years to identify the dynamics and evolution of the Cuvelai drainage system in relationship to the prevailing climatic conditions. 2. Hydrologically model the Cuvelai drainage system to confirm spatial distribution and variability of flows in the various channel systems to confirm the surface water harvesting potential. 3. Based on the results of the hydrological studies and GIS analysis of DEM, and land use, soil, and other maps, map suitable locations for the construction of chimpacas and cisterns. Water Security and Drought Resilience in the South of Angola 65 PHOTO 4.11. Active Channel of the Cuvelai Drainage System in the Cuvelai Lowlands (left); Kimbo Near Ohenghali, Mongua Commune, with Field Prepared for Growing Rainfed Crops During the Upcoming Rainy Season (right) Source: World Bank. These studies should identify locations that feature, besides reliable surface water flows during most of the years (either from active channels or from flooding of sufficiently large pans and depressions), favo- rable geological, geomorphological, soil, and land use conditions that allow for the delineation of suffi- ciently large impluvia. A major drawback of chimpacas is the almost systematic contamination due to the presence of animals in or near the reservoir, creating a public health hazard for human populations using the same chimpa- cas for their drinking water supply. Efforts to keep animals out, by fencing off the water reservoir and its immediate surroundings, often fail as fences are vandalized to let animals drink. For this reason, it is recommended to construct chimpacas in pairs with cisterns that collect water exclusively for human uses. Cisterns harvest smaller volumes of water than chimpacas but ensure enhanced hygiene. They typically have a larger depth/area ratio and can be lined and covered, to reduce evaporation losses. Appendix C includes standard layouts and cross-sections for chimpacas and cisterns. Chimpaca and cistern design should include sufficiently large impluvia to provide a 80 percent reliability to fill up in a ­ given year (filling up at least 4 years out of 5). In areas with low harvesting potential a lower reliability level can be applied, provided that the water security of the populations can be ensured by other nearby sources during drought spells. Sandy islands generate very low runoff, because of the loose sandy soils and the near flat topography. In addition, many islands are occupied by kimbos that prepare the surroun- ding land for growing rainfed crops for which the topsoil is labored to enhance infiltration. Therefore, the design of chimpacas and cisterns should be based on the runoff generated by the channels, floodp- lains, and depressions with silty and clayey soils and scant vegetation. “Wet-season P-ETa excess” cumulated per rainy season (October till March) shows a strong correlation with “wet-season P” (figure 4.12). The P-ETa (October–March) is a good measure of the net rainfall Pnet received by channels and depressions with low infiltration and will be used to determine design speci- fications for chimpacas and cisterns. 66 Water Security and Drought Resilience in the South of Angola FIGURE 4.12. Correlation Between “wet-season P-ETa excess” and “wet-season P” Wet season P-ETa vs, wet season P lowland Angola (2003–18) 600 (P-ETa) = 0.93 P – 336 Seasonal excess P-ETa (Oct-Mar) 500 R2 = 0.90 400 300 200 100 0 –100 200 300 400 500 600 700 800 900 1000 Seasonal rainfall (Oct-Mar) Source: Based on monthly data accumulated for the rainy seasons from October till March in the 2003–18 period. Note: ETa = actual evapotranspiration. The correlation between P-ETa (October–March) and P (October–March) for the Angola Lowland subba- sin has been used to estimate the daily net rainfall amounts that can be harvested (figure 4.14), based on daily rainfall data at Ombala Yo Mungu (location shown in figure 4.1) for the period 1998–2019. First, for every rainy season of the 1998–2019 period the cumulative net rainfall has been calculated, based on seasonal cumulative rainfall amounts (see figure 4.13). As can be seen, seasonal Pnet is highly variable, fluctuating between 0 mm and 423 mm per season (ave- rage is 120 mm) over the 21-year period, and 5 out of the last 7 years show zero Pnet values. Second, based on the ratio of Pnet/P determined for every individual season, the daily rainfall amounts have been converted into daily Pnet. Results are shown in figure 4.14. From the daily P and Pnet values cumulative rainfall graphs can be drawn for a rainy season. Figure 4.15 shows how, based on the rainfall data at Ombala Yo Mungu, cumulative P and Pnet evolve throughout the 2003/2004 rainy season, which was characterized by an average amount of rainfall. Based on the evolution of the cumulative Pnet, the amounts of surface water harvested for different sizes of impluvium can be calculated. Figure 4.16a shows for different impluvium sizes how fast a stan- dard cistern (measuring 20 m x 5 m x 5 m) and a standard chimpaca (measuring 80 m x 60 m x 4 m) fill up during a year of average rainfall. With respectively a 0.5 hectare (ha) and a 20 ha size impluvium a standard cistern and a standard chimpacas fill up in about 160 days. Similar graphs can be drawn for years with below average rainfall, with different impluvium sizes and for different chimpacas/cistern sizes, to explore which chimpacas/cistern sizes and what impluvium sizes provide maximum harvesting potential at the highest degree of reliability. Water Security and Drought Resilience in the South of Angola 67 FIGURE 4.13. Wet Season Rainfall P (October–March) and Calculated Wet Season Net Rainfall Pnet at Omabala Yo Mungu, for the Rainy Seasons Net rainfall (Oct-Mar) for Ombala Yo Mungu (1998–19) 900 800 700 600 Rainfall (mm/season) 500 400 300 200 100 0 99 99 20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08 20 09 20 0 20 1 20 2 20 13 20 4 20 15 20 6 20 17 20 18 9 –1 –1 –1 –1 –1 –1 – – – – – – – – 19 8– – 10 – – – – 12 16 0 14 11 17 13 15 18 09 00 06 02 01 04 07 03 05 08 –2 9 19 P (Oct-Mar) Net rainfall (Oct-Mar) Source: Based on daily data accumulated for the rainy seasons from October till March in the 1998–2019 period. FIGURE 4.14. Daily Rainfall and Estimated Net Daily Rainfall in Ombala Yo Mungu During the Rainy Seasons of 1998–2019 Daily P and Pnet at Ombala Yo Mungu (1998–19) 100 90 80 70 Rainfall (mm/day) 60 50 40 30 20 10 0 1– 8 30 99 30 0 1 2 3 1– 4 5 1– 6 7 8 1– 9 0 1– 1 1– 2 1– 3 2– 4 1– 5 6 2– 7 8 –0 –1 –0 –1 –0 –0 –1 –0 –1 –1 –0 –1 –0 –0 –1 –9 –0 –1 –0 –1 – 10 10 10 10 10 10 10 10 10 –9 10 10 10 10 10 10 –9 10 10 10 –9 1– 1– 1– 1– 1– 30 1– 1– 1– P (mm/day) Pnet (mm/day) 68 Water Security and Drought Resilience in the South of Angola FIGURE 4.15. Cumulative rainfall and net rainfall in Ombala Yo Mungu for an average rainy season 600 Cumulative P and Pnet at Ombala Yo Mungu (year of average rainfall) 500 Cumulative rainfall (mm) 400 300 200 100 0 20 40 60 80 100 120 140 160 180 Days since start of rainy season FIGURE 4.16. Filling Curve of Standard Chimpacas and Cisterns for Different Impluvium Sizes During a Rainy Season with Average Rainfall (left) and Below-Average Rainfall (AEP = 60 percent) (right) a. Degree of filling of Chimpaca / Cistern for b. Degree of filling of Chimpaca / Cistern for different impluvium sizes (Year of averge rainfall) different impluvium sizes (60% AEP of rainfall) 100 100 80 80 Degree of filling Degree of filling 60 60 40 40 20 20 0 20 40 60 80 100 120 140 160 180 0 20 40 60 80 100 120 140 160 180 Days since start of rainy season Days since start of rainy season Impluvium cistern: 0.5 ha 1 ha 1.5 ha Impluvium chimpaca: 10 ha 15 ha 20 ha Figure 4.16b shows for the same impluvium sizes the degree of filling up in a year with below average rainfall (Annual exceedance probability of 60 percent). As can be seen, cisterns require a larger than 1 ha impluvium to fill up and chimpacas only fill up half with a 20 ha impluvium. Based on daily rainfall data for the period 1998–2019 at Ombala Yo Mungu, the required impluvium sizes to entirely fill up chimpacas and cisterns have been determined for different probabilities of exceedance of seasonal rainfall (see table 4.5). As Ombala Yo Mungu receives 13 percent less rainfall than the average for Lowland Angola, required impluvium sizes are somewhat smaller in other parts of the basin. Water Security and Drought Resilience in the South of Angola 69 TABLE 4.5. Required Impluvium Sizes for Standard and Large Chimpacas and Cisterns, Depending on the Amount of Net Seasonal Rainfall Received (October to March) Chimpaca Size Cistern Size Aep of P Pnet (Oct-Mar) Standard 80 X Large 120 X 80 X 8 Standard 20 X 5 X 5 Large 30 X 5 X 8 M (Oct-Mar) (%) (Mm) 60 X 4 M 19,200 M3 M 76,800 M3 M 500 M3 1,200 M3 33 140 13.7 54.7 0.4 0.9 40 123 15.6 62.4 0.4 1.0 50 80 24.0 96.0 0.6 1.5 60 46 41.3 165.2 1.1 2.6 75 20 94.0 376.1 2.4 5.9 Note: AEP = Annual exceedance probability. As required impluvium sizes increase rapidly for years with below average rainfall, standard size chimpacas and cisterns may not be able to harvest sufficient volumes of water during moderately dry ­ years and will harvest no water at all during the driest years (1 year out of five). At the same time, they will reach their maximum storage capacity long before the end of the rainy season during wet years. This limitation can be (partially) overcome by ensuring that: •• Very large chimpacas or cisterns are built at suitable, strategic locations, and fill up during avera- ge-to-wet years, storing water that can be used during the following year; •• Chimpacas and cisterns are connected via active channels with the Cuvelai drainage system to benefit from seasonal flows; and •• Backup water supply solutions, like boreholes, secure at least the human water supply during drought years. The following additional comments can be made for the design and operation of chimpacas and cisterns: •• To optimize live storage, desilting of chimpacas and cisterns needs to be done on a regular basis. Ideally, desilting is managed by the community and financially enabled by water point user fees but could alternatively be managed at the commune level. •• Chimpacas have high evaporation losses as a result of their large surface and limited depth. For the building of large chimpacas it is recommended to increase the depth/area ratio by excavating seve- ral meters of the underlying marls. •• To protect the quality of the harvested water, the impluvium of cisterns has to be demarcated/ enclosed to avoid animals grazing within. •• For cisterns it is recommended that debris and sediment traps be built and that rock filters are used to regulate flow into the cistern (see photo 4.12). 70 Water Security and Drought Resilience in the South of Angola PHOTO 4.12. Sediment Trap Protecting Cistern (“berkat”) in Somaliland (left two images) and Rock Filter Sand Sediment Trap for Two Different Cisterns in Djibouti (right two images) Source: World Bank. Water Security and Drought Resilience in the South of Angola 71 PHOTO 4.13. Artesian Deep Borehole at Okashana (left) and Stable Isotope Concentrations in Groundwater in the Namibian Part of the Cuvelai-Etosha Basin (right) 10 GMWL 0 LMWL Rain monthly Rain events –10 –20 10382HVSMOW –30 –40 Shallow wells –50 y = 4.66 x –14.65 R2 0.94 –60 deep wells y = 6.01 x –7.88 R2 0.93 –70 –10 –9 –8 –7 –6 –5 –4 –3 –2 –1 0 1 2 103 818OVSMOW Source: Hamutoko et al. 2017. •• To generate sufficient storage and reduce evaporation losses, cisterns should be partially excavated in the underlying marls. •• Water from cisterns can be supplied through taps using solar energy. Remarks, Recommendations, and Conclusions on Water Supply Options in the Cuvelai Basin No single water resource can provide water security for all uses in the Cuvelai. Therefore, a mix of solu- tions is recommended, with surface water harvesting providing the bulk of the rural water demand and groundwater providing water security to human populations and cattle in critical areas and during dry years. Surface water represents the single largest available water resource in the Cuvelai, and its harvesting can be expanded. Surface water can provide the bulk of animal demand for water (except during droughts) and, if correctly designed, also part of the human demand. The current shared use of chimpacas for human and animal consumption creates a public health hazard. It is recommended to build chimpacas in pairs with cisterns, solely dedicated to human consumption. To provide enhanced water security, for example along transhumance corridors, it is recommended that larger and, more importantly, deeper chimpacas are built at suitable locations. Such strategic chimpacas should be designed to have sufficient storage to bridge short drought spells. To optimize live storage, desilting of chimpacas and cisterns should be done on a regular basis. Ideally, desilting is managed by the community and financially enabled by water point user fees. 72 Water Security and Drought Resilience in the South of Angola Shallow groundwater, tapped through open wells and cacimbas, provides good-quality water from shallow perched aquifers but their potential is very limited. Where present it can be used to supply ­ ­individual families. Medium to deep groundwater (up to 200 m depth) appears to have moderate potential in the Angolan part of the Cuvelai. The medium-deep groundwater potential should be further mapped and explored, requiring systematic collection of information on boreholes, aquifer characteristics, groundwater levels, and groundwater quality. A regional study assessing the medium-deep groundwater potential is a priority. Currently, a large share of existing boreholes are out of order for extended periods of time because of failure of pump, solar panel or borehole. Setting up local Water Committees, responsible for the collec- tion of user fees and in charge of maintenance and small repairs, can greatly improve borehole functionality. Evidence from Namibia suggests the presence of a deep regional aquifer below part of the Cuvelai- Etosha basin. It is recommended that, in collaboration with Namibia, investigations be conducted of this potentially strategic transboundary water resource. However, confirmation of a deep aquifer would not significantly alter the priorities for rural water supply in the Cuvelai, as tapping groundwa- ter from deep aquifers is costly and will be more expensive than the available low-cost water supply solutions. Case Study 2: Identification of River Sections with Potential for the Construction of Small-Scale Managed Aquifer Recharge In Namibe Province Introduction This section presents the results of efforts to identify river sections with the potential for the construc- tion of small-scale MAR solutions at the community level, approximating the design, the potential sto- rage capacity for each site under consideration, and the expected benefits. The analysis focuses on two areas: the municipalities of Bibala and Camucuio in Namibe province (see maps 4.6–4.8 and table 4.6). The communes in these two areas are in the list of priority communes developed in chapters 2 and 3. They were selected for this initial evaluation during previous field missions, when the potential for these types of structures in the area was identified. Physical Geography of the Study Area The geology of the area consists of extensive granite and other igneous rock outcrops, forming plains and hills and it is, in principle, not suitable to the accumulation of water in the subsoil, beyond what is held in the alluvial deposits filling the stream courses. One of the main opportunities to increase the availability of water in the dry season consists in the accumulation of surface water in structures such as dams, including sand dams. With regard to climatology, what has been explained in the previous chapters is logically applicable to Namibe. It is important to note that recent years, which were drier for the entire region, made this part Water Security and Drought Resilience in the South of Angola 73 MAP 4.6. Map of Bibala and Camucuio Municipalities in Namibe Province Chingo area in Camucuio municipality Middle part of the Giraul basin area in the Bibala municipality MAP 4.7. Zoom to the Giraul Basin Area, With Potential Sites for Managed Aquifer Recharge Solutions Labelled “GB” 74 Water Security and Drought Resilience in the South of Angola MAP 4.8. Zoom to the Chingo Area, with Potential Sites for Managed Aquifer Recharge Solutions Labelled “CH” TABLE 4.6. Locations of the Selected Points and Hydrological Basin Parameters Location Basin Max Runoff Latitude Longitude Elevation Area Height Length Slope% Rainfall Volume % Volume Site XX.xxxxx° YY.yyyyy° m asl km2 m km % mm m3 m3 CHB1 −14.19159 12.78743 545 1,184 950 74 0.55 300 3.55E+08 20% 7.10E+07 CHB2 −14.12294 12.94516 650 – – – – – – – – CHB3 −14.22442 12.82309 578 34 780 13.5 1.50 300 1.01E+07 30% 3.02E+06 CHB4 −14.22065 12.76221 560 163 690 25.5 0.51 300 4.89E+07 20% 9.78E+06 GB1 −15.03947 12.98671 388 26.7 1,710 11 12.0 400 1.07E+07 30% 3.20E+06 GB2 −15.07770 13.02150 413 169 2,300 30 6.3 500 8.45E+07 30% 2.54E+07 GB3 −15.10450 13.02660 407 341 2,290 35 5.4 500 1.71E+08 30% 5.12E+07 GB4 −15.11051 13.02050 399 61.6 1,730 21 6.3 400 2.46E+07 20% 4.93E+06 Note: GB = Giraul Basin; CHB = Chingo area (basin). of the province—which in normal years registers little rainfall—experience consistent aridity. For the Giraul area, the mean rainfall dropped in the last 6 years from 740 mm/yr to 500 mm/yr; in the Chingo area, it diminished from 500 mm/yr to 300 mm/yr. Both study areas are characterized by water courses in which the flow is not fed by significant springs. The hydrological pattern of this type of basin is conditioned by geomorphological and climatic features, mostly by the rainfall, the slopes, and the geology. Water Security and Drought Resilience in the South of Angola 75 To a certain extent, Namibe province is geomorphologically homogeneous. It is mostly characterized by the escarpment that divides the Huíla Plateau from the plain that descends seaward. Therefore, most of the wadis in the area are parallel, flowing from the escarpment to the sea, although with differences in their slopes (for example, in the Giraul zone, basin slopes are around 4 percent, while in Chingo they rarely go beyond 0.5 percent). The height of the plateau, that reaches elevations up to 2,300 m above sea level, and the Leba escarp- ment ensure a high volume of rainfall that supplies running water in the headwaters at the peak of the rainy season. The flow at the major rivers follows the pattern of the rainfall distribution, with peaks in the months of March and April and 5 months of nil flow from May to October. There is little information on whether the patterns have shifted in the past years (precipitation patterns have not; only the quantities have been reduced). A review of past Google Earth images of Namibe shows important flows and some flooding only in the month of April 2008 and a minor one in January 2018. Suitable Sites: Characterization and Challenges From maps 4.7 and 4.8, it can be seen that numerous potential points were preselected in previous mis- sions or by preliminary desk work, but in August 2019, after further evaluations and field visits, only three sites were suitable for piloting this infrastructure successfully. In this summary report, only the analyses of the adequate locations are included. Two types of small dams are suitable for environments in which evaporation is an issue: SbSD and sand dams. The typical sand dam is built where a stream has excavated its course in an impervious basement, such as a crystalline basement or a clayey formation or marls where the rock is exposed. A sand dam can also be built on different types of rock, such as limestone, but close attention must be given to the presence of fractures or cavities. The dam is expected to be filled by sediment (normally sand) carried by the stream flows. In the study area, the transport of sediment has not yet been studied and the filling time is unknown. The dam action may prompt an initial free water harvesting, then a gradual filling with sediments. In this period, the water is subjected to evaporation. Once the dam is filled with sediments, these are saturated by water at each flow and can store it for a long time. Normally a dam of this type is 2–5 m high and less than 200 m wide. The geomorphological features to successfully build a SD are: (1) an impervious basement; (2) a hydrau- lic basin wide enough to supply a certain number of flows with suitable yields; (3) a height of the banks of not less than 2 m (otherwise, water flows might flood the surrounding area or the volume of sand/ water accumulated might be insignificant; (4) a significant transport of sediments with scarce matrix and sandy composition to fill the throwback area in a reasonable time; (5) a stream bed slope not excee- ding 1–2 percent to form a long throwback (as an example, a 2 m high dam with a bed slope of 0.5 percent 76 Water Security and Drought Resilience in the South of Angola shall extend its influence up to 400 m back, while with a slope of 2 percent the area of interest shall be 100 m); and (6) the selected site should be located in a narrow course. The volume of water stored that can be extracted depends on the sand porosity (effective porosity), which normally varies from 8 percent to 20 percent. On the other hand, SbSD are built where a stream has excavated an impervious basement and filled the depression with plenty of sandy deposits. Normally in such a case a seasonal or permanent aqui- fer already exists. The base of a SbSD shall reach the basement and the top shall be normally buried or slightly higher than the riverbed. Its function is to stop (or simply slow) the underground flow to raise the water level and increase the volume of water available. The secondary effect, most evident in the seasonal aquifers, is to extend in time the presence of water or transform the aquifer from sea- sonal to stable. The depth of the impervious basement should be no more than 5–6 m because beyond that the excavation of the alluvial deposits is more difficult and may require equipment and metho- dologies not easily available in developing countries. The only conditions for the construction are: (1) an impervious basement below the alluvial deposits at a shallow depth; (2) a concave pattern of the basement perpendicular to the stream axis; and (3) the presence of a narrow course to reduce the costs of construction. SbSDs are built after the excavation of a trench up to the basement; the dam can be constituted by a concrete wall or by large plastic sheets folded and filled with the material excavated. From an analysis of Google Earth images and observations made during field visits, no appropriate sites were identified for SbSDs, and the three chosen sites—GB2, GB3, and CH1—are adequate for sand dam construction with a reasonable certainty of creating significant water storage benefits for local residents. Site GB2 This site is located at the confluence of two branches surrounding an island in the NW affluent of the Giraul course (Rio das Mangueras) at the foot of a hill (NW-ward) about 200 m higher than the stream. At the site selected for a possible sand dam, fresh gray granite outcrops form the NW bank (FG1, see photos 4.14 and 4.15), the top of the SE bank (FG2), and the stream bottom (FG3) some 50 m downstream. The NW outcrop extends upstream and is constituted of weathered granite (WG), forming a wall of about 3 m, while adjacent to the SE there is a fresh granite outcrop (FG1) 2–2.5 m higher than the stream bed. At the foot of the bank the stream bed is about 0.5 m higher than at the center of the course. The FG2 outcrop extends on top of the SE bank for a length of about 30–40 m and its top is 3.0–3.5 m higher than the stream bed. In the lower section of the bank a terrace of coarse alluvial deposits (sand, gravel, and some boulders) descends gradually to the course bed. The FG3 outcrop occupies a large part of the riverbed downstream. Another thin outcrop crosses the course upstream almost completely. At the center and in the lower part of the SE bank, alluvial deposits (medium-coarse sand and gravel) having an expected thickness of up to 1–2 m below ground level, are present. According to observations and information from the community there is no allu- vial aquifer. Water Security and Drought Resilience in the South of Angola 77 PHOTO 4.14. View of the Possible Dam Layout at Site GB2 from the Right Bank Source: World Bank. Note: The granite outcrops are in violet. The slope of the wadi bed from the FG3 outcrop upstream, along the SE branch, is estimated to be less than 1 percent, while the in the NW branch the slope seems to be higher (1.0–1.5 percent) and abruptly increasing with the presence of a massive outcrop. The basin extension, calculated at site GB2, is 169 km2. It extends from 413 m asl to 2,300 m asl with a medium slope of 6.3 percent. The slope in the lower section, from the foot of the escarpment to the site, is 1.5 percent. Annual rainfall is estimated at 500 mm. Assuming this value is valid for the whole basin, the rainfall volume is approximately 85 million of m3/year. In the upper basin (about one-third of the catch- ment), bare rock and steep slopes prevail, while the lower section is composed of a plain where granite ­ oefficient is outcrops are separated by sandy deposits. Given the conditions outlined above, the runoff c estimated at almost 30 percent and the runoff volume is about 25 million m3/year. Considering only the rainfall volume of the peak month (March, 220 mm), 30 days of running water and a higher runoff coeffi- cient caused by the ground saturation, a mean yield of 8.6 m3/s is expected. According to the information supplied by local community representatives the water flows without interruption for about 2 months in the period of the rainfall peak (from February to March) and stops a few days after the end of the rainy season. According to the information collected by the community, no tardive flows occur. 78 Water Security and Drought Resilience in the South of Angola PHOTO 4.15. Possible Dam Layout, Throwback, and Geological Features at Dam Site GB2 According to geomorphological features, the possible sand dam should have a length of 75 m and a height of 2 m above the stream bed. The throwback should extend no less than 300 m in the SE branch ­ and about 150 m along the NW branch. The throwback area has an extension of about 13,500 m2. Considering a mean height of 2.0 m (sum of the thickness of the fresh deposits and of the existing ones), the estimated sand volume is 27,000 m3. The sediments filling the course are mostly made by medium- coarse sand and gravel, mostly siliceous with scarce or absent loamy matrix. Considering a mean value of 12 percent, a volume of water of 3,240 m3 can be stored by each important stream flow. Some more water could be stored if natural sandy deposits extend also beyond the throwback area. About the possible utilization of the water, in the Google Earth image dated August 2018, a dozen of huts with circular animal fences can be seen in a radius of 1.5 km, hence the area is considered suitable for pastoral temporary settlements. The flat areas closer to the site can also be considered suitable for small farms activity. The dam site is accessible by a dirt road made for the installation of a power line, accessible up to 500 m of distance from the site. The community representatives agreed on this site selection. Site GB3 This site is located along the main branch of the Giraul stream (Rio Kapangombe), as seen in photo 4.17. It is in the center of a plain mostly made by alluvial deposits with scattered granite outcrops forming small mounds slightly higher (normally 1–3 m) than the plain. In the area selected a wide granite outcrop (G1) has been cut by the stream course and forms riverbed and banks (see photo 4.16). Two more out- crops are present on the right bank (G2) and on the left one (G3). It is probable that granite is present in continuity at shallow depth in the whole area. Upstream and downstream sandy deposits with boulders Water Security and Drought Resilience in the South of Angola 79 PHOTO 4.16. View of the Possible Dam Layout at Site GB3 from the Right Bank Source: World Bank. PHOTO 4.17. Possible Dam Layout, Throwback, and Geological Features at Dam Site GB3 fill the course. Their thickness may range from 1 m to 3 m. A hole excavated some 50 m upstream of the granite outcrop G1 found the basement at 1.4 m; some 300 m upstream there is a hand-dug well (hdg 2) about 3 m deep. At the dam site the course is divided in two branches by a rocky ridge of about 40 m in length; the lower branch runs to the foot of the left bank (eastward) about 3.0 m below the outcrop top. The right (western) bank is topped by a flat plain, about 5 m higher than the stream bed (2–3 above the 80 Water Security and Drought Resilience in the South of Angola possible dam top); the left (eastern) flank is topped by a plain about 2–3 m higher than the stream bed (no more than 0.5 above dam top). The uncertainty regarding the elevations is caused by the oscillations affecting the Global Positioning System (GPS) measurements. The mean slope of the course upstream is about 2 percent in the rocky section of the course bed (near the dam site), while it is less than 1 percent in the sandy section upstream. According to observations on the ground and to information collected from the community there is no alluvial aquifer, but the presence of the hand-dug well with water is evidence of a limited flow that feeds the depressions in the basement and ensures a small amount of water in the months November-December. The presence of the water point and the gradual seasonal descent of the water level is very important because it is evidence of a seasonal aquifer. A future sand dam shall stop the already existing flow and the water accumulated shall extend upstream beyond the throwback limit. The hydrological basin for the selected point extends for about 355 km2. The basin starts at an elevation of about 2,300 m above sea level and the site is at an elevation of almost 400 m above sea level. The upper section of the basin extends for almost 200 km2 at elevations above 1,000 m. The stream is 45 km long with an average slope of 4.2 percent, but considering only the section at the foot of the escarpment (about 22 km) the slope drops to 1.3 km and in the last 5 km to almost 0.8 percent. Assuming the value of 500 mm/year as rainfall for the whole basin, the rainfall volume is approxi- mately 180 million m3/year. The morphology and geology of the basin are similar to those of the site GB2. Since measurements of the stream flow in the area are not available, given the conditions outlined above, the runoff coefficient is estimated at almost 30 percent and the runoff volume at 53 million of m3/year. Considering only the rainfall volume of the peak month (March, 220 mm), 30 days of running water and a higher runoff coefficient of 50 percent, a mean yield of 15 m3/s is expected in the month. According to the information collected by local community representati- ves, the water flows without interruption in the months of the rainfall peak (from February to April) and stops a few days after the end of the rainy season. In total a period of 3 months of continuous flow is expected. Since the peak rainfall occurs toward the end of the rainy season, normally no further flows occur. The length of the proposed dam is about 85 m; the proposed height is 2.0 m. The volume of sand accu- mulated is obtained multiplying the throwback area (estimated 12,000 m2) times an average thickness of 2.5 m that includes new and existing deposits; the total volume of sand is 30,000 m3. The effective poro- sity is estimated 12 percent and the volume of the stored water in the fresh sand is 3,600 m3. An additio- nal volume of water is supplied by the damming of the underground flow that shall accumulate also beyond the throwback probably up to 400–500 m from the dam. Site GB3 is at the center of the plain south of the Kapangombe village where about 500 people live in scattered settlements, with some 1,000 animals. A settlement with few huts is located about 100 m far from the dam on the western bank. The community representatives confirmed the importance of increa- sing the water availability in terms of quantity and duration mostly for the sites GB2 and GB3. Water Security and Drought Resilience in the South of Angola 81 Site CH1 Site CHB1 is located in the stream called Rio Manoca, the most important of the study area, about 12 km downstream and westward of the Chingo village (photo 4.18). In the 1970s a dam was built with squared blocks and some clay matrix, 3.7 m high on the present course bed. According to the information collec- ted from the community, and to the observed dam geometry, the wall was thought to stop the water flow, which at the time ran where there is now the left section of the wall and the granite basement is outcropping. Presently, as it can be seen in photo 4.18, the course runs where the left section was and the bed is made by sandy deposits, 1.4–1.7 m thick, datum verified by the excavation of two pits. From the satellite image the deviation of the old course after the dam break and the fresh erosion in the right bank is clearly visi- ble. Very likely the dam broke up at a weak point where the wall did not reach the granite at depth and connected to the earth wing built on the right bank. In that area, the stream course was once filled with sand. From the observations on the ground it is estimated that the course bed at the time of the dam’s construction was about 1–1.5 m higher than it is now. A water pool is present about 100 m upstream of the dam and supplies livestock and human needs. The dam, the courses, and the granite outcrops are visible in photo 4.19. The hydrological basin of the stream at the selected point extends for about 1,190 km2. The basin head starts at an elevation of about 940 m above sea level while the dam site is at an elevation of 546 m above sea level. In the upper section there are scattered reliefs up to a maximum elevation of 1,600 m asl. PHOTO 4.18. Site CHB1 Note: Site CHB1—red line: dam wall—red dashed line: earth right wing—dashed green line: previous course—blue line: new course—yellow dashed line: fresh erosion—blue filled circle: water pool—violet polygons: basement outcrops. 82 Water Security and Drought Resilience in the South of Angola PHOTO 4.19. New Proposed Dam (blue line) and Throwback Area The stream is 74 km long with an average slope of 0.53 percent, that in the last 10 km drops to almost 0.3 percent. The average annual rainfall dropped in the last 7 years from 490 mm to 300 mm. Assuming the lower value for the whole basin, the rainfall volume is approximately 350 million of m3/year. Most of the catchment is made of bare rock but is characterized by gentle slopes with crystalline out- crops separated by sandy deposits. The runoff coefficient is estimated at almost 20 percent and the runoff volume at 71 million m3/year. Considering only the rainfall volume of the peak month (April) and 15 days2 of running water, an average yield of about 40 m3/s is expected in the month. The water flow stops a few days after the end of the rainy season. Since the peak rainfall occurs toward the end of the rainy season, no further flows are expected. Actually, given the basin extension, it is possible that there are some minor flows in the months from November to January, when minor rains occur. Keeping in mind the changes that occurred to the course after the dam break, the best solution is to build a new dam in a different layout, with different geometry. Based on the information currently available, the dam could be founded on the rocky outcrops downstream of the dam about 5–10 m from the remains of the old one. The old dam should be destroyed and the material utilized for the construc- tion of the new one. Its center should be located in the middle of the new course occupied by sandy deposits of 1.5–2.0 m in thickness. The height should be 2.0–2.5 m on the present stream bed, so that in the deepest point the dam height shall reach 3.5–4.0 m on the crystalline basement. The throwback, given the very gentle slope (0.3 percent), should extend up to 500 m upstream. In such a basin the solid transport could be very poor and the dam could take a very long time before filling up with sand. The stored water volume before sand filling is estimated at about 18,000 m3 (18,000 m2 x 1.0 m). The volume Water Security and Drought Resilience in the South of Angola 83 of the sands already deposited in the throwback area before damming is estimated at about 27,000 m3 (18,000 m2 x 1.5 m). Once completely filled, the total possible volume of sand saturated would be about 45,000 m3, that considering a porosity of 12 percent should supply 5,400 m3 of groundwater. Considering the low slopes (0.3 percent), abundant sandy deposits, and the presence of a water pool also in the dry season, the section of saturated sand should extend beyond the throwback area. Site 1 is the priority for the administration of the Chingo village. The resident population can be estima- ted at about 200 persons and the livestock in some thousands (cows, goats, and sheep). During the site visits there were animals drinking in the pool and people coming to fill jerry cans with water all the time. The situation from the point of view of health is dramatic because the only water available is that of the pool and it is very dirty with animal feces scattered all around. The humans excavated a side hole about 1–2 m away but contamination is unavoidable. This site should be given maximum priority. The site is accessible by dirt road. Suitable Sites: Dam Designs Site GB2 The characteristics of the proposed dam are synthesized the table 4.7. The possible dam profile is illus- trated in figure 4.17, and its cross-sections in figure 4.18. The basement pattern in the figure is only a possible profile since on the top of the right bank, on the stream bottom and on the left flank, the granite is covered by sandy deposits. However, given the pre- sence of scattered outcrops down from upstream a maximum thickness of 1.5 m of alluvial deposits (sand gravel and blocks) has been considered. Assuming a mean section of about 3.3 m2, and a length of 75 m, the expected volume of the dam body is 282 m3. A slope of 50 percent has been assumed for the downstream flank of the dam (see figure 4.18) and a foundation with a section of 1 m2 in the basement. With this configuration a water volume of 3,600 m3/flow can be stored. The volume could increase if upstream there is a significant thickness of alluvial deposits. The material suggested for the dam construction is masonry stone (cement and blocks). To the downs- tream foot of the wall a bed of blocks (rip-rap) of about 50 m of length and 0.8 m of thickness with a total volume 144 m3 must be lay down to avoid the erosion from the waterfall. Blocks and sand (for the cement) are available 100–200 m from the site. The dam site is accessible by a dirt road made for the installation of a power line, accessible up to 500 m ESE of the site. Full accessibility can be easily secured by a bulldozer in a day’s work. TABLE 4.7. Stream flow and dam parameters Mean Dam Yield Stored Height on Max Runoff Peak Peak Water Dam Stream Alluvial Dam Wall Rip-Rap Volume Month Yield Volume Length Bed Thickness Volume Volume Runoff 3 3 3 3 m /year m /sec Days m /sec m /flow m m m m 3 m3 2.54E+07 7.2 60 68 3,240 75 2.0 2.0 282 144 84 Water Security and Drought Resilience in the South of Angola FIGURE 4.17. Proposed Dam Profile And Rough Bedrock Pattern Site GB2- Dam profile 417 Right bank (NW) Left bank (SE) Cross-section S1 416 Cross-section S2 415 Elevation (m) 414 Spill out 40 m 413 Granite Alluvial deposits 412 411 410 –20 0 20 40 60 80 100 Cross-section distance (m) Granite Dam profile Bedrock FIGURE 4.18. Dam Cross-Sections a. Cross-section S1 b. Cross-section S2 2.0 3.5 Spill-out level 3.0 1.5 Masonry stone 2.5 Elevation above rock basement (m) Elevation above rock basement (m) 1.0 2.0 Rip-rap protection Geotextile 1.5 0.5 1.0 0 0.5 Stream sand –0.5 0 Basement –0.5 –1.0 Basement –1.0 –1.5 –1.5 –0.5 0 0.5 1.0 1.5 2.0 2.5 –0.5 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 Longitudinal distance (m) Longitudinal distance (m) Site GB3 The characteristics of the proposed dam are synthesized in table 4.8. The possible dam profile is shown in figure 4.19, the cross-sections in figure 4.20, and the planimetry in photo 4.20. In the proposed layout 1, a wall should be added to the main dam body, to block the small channel between the two outcrops passing behind the outcrop G2a. This wall would connect the left (eastern) Water Security and Drought Resilience in the South of Angola 85 TABLE 4.8. Stream Flow and Dam Parameters Mean Stored Dam Height Max Dam Runoff Yield Peak Peak Water Dam on Stream Alluvial Wall Upstream Volume Month Yield Volume Lenght Bed Thickness Volume Wall Runoff 3 3 3 3 3 m /year m /sec Days m /sec m /flow m m m m m3 5.33E+07 15.1 60 144 3,240 75 2.0 2.0 340 30 abutment leaning against the outcrop G2a, with the outcrop G2b about 15 m upstream. The wall should be about 24 m long and 0.5 m thick. The channel behind the wall should be filled with large stones. The spill out section shall be 60 m long, and 0.7 m lower than the crest; if layout 2 is selected, which accounts for a possible increase in height and length, it is suggested to increase the spill out length to 90 m and the height to 0.8 m. At this site a rip-rap is not necessary because the stream bed is already rocky, but to reduce the erosion of the basement and of the dam base the wall has been designed with a slope of 40° for the downs- tream side. In case layout 2 is chosen a rip-rap could be necessary in the left section of the course. The mean height is about 3.0 m considering also the depth of the foundation (see figure 4.21). The volume of the dam body is estimated 340 m3. The volume of the lateral wall is 30 m3. In case layout 2 is ­chosen, the dam volume increases to about 470 m3. The material to be used is cement and blocks (masonry stone). Blocks, stones, and sand (for the cement) are available 100–200 m from the site. The site is ­ eachable from reachable by dirt road (7.8 km long) from the EN 280. The top of the right bank is easily r the road but a section of about 100 should be cleaned to access the site; given the rocky nature of the site, further works can be required to insure the movement on the site of mechanics’ machines and trucks. In photo 4.20 it is clearly shown that the basement is outcropping along the whole section to be dam- med. The proposed dam height (layout 1) ranges between 2 m and 2.5 m above the lowest points of the section, which in the figure are at elevation 410 m asl. At the lowest points the alluvial deposits are expected to be less than 1.0 m thick; for this reason the maximum height expected is 3.0 m. The dam length (along the arc) is about 85 m. With this configuration a stored water volume of 3,600 m3 is expected. Presently it is difficult to estimate the sand volume using normal GPS. Actually, the main doubts regard the slope of the stream bed, upstream of the main granite outcrops. The GPS data con- firm that the slope is very small between 0.5 percent and 1 percent: with such values it is expected that the throwback area extends up to a minimum 300 m, but could be even longer. The second doubt regards the left shoulder: the elevations of the terrace behind the left abutment seem almost flat and could require a change in the dam layout (layout 2) to reach the outcrop G3. In such a case the dam length could increase up to 130 m, the height could increase by 0.5–1.0 m, and the throwback area could increase to 30,000 m2 against 12,000 m2 of the proposed project. The stored water volume 86 Water Security and Drought Resilience in the South of Angola FIGURE 4.19. Proposed Dam Profile and Rough Bedrock Pattern Site GB3 - Dam and ground profile 415 414 Right bank (W) 413 Section S2 Section S1 Left bank (E) 412 Elevation (m) 411 Spill out 60 m 410 409 408 407 406 405 0 20 40 60 80 100 120 140 160 Cross-section distance (m) Ground surface Dam top 2 Foundation FIGURE 4.20. Dam Cross-Sections a. Cross-section S1 b. Cross-section S2 3.0 2.0 Masonry stone 2.5 Spill-out level 1.5 2.0 Masonry stone Elevation above rock basement (m) Elevation above rock basement (m) 1.0 1.5 40° 1.0 0.5 40° 0.5 0 0 –0.5 –0.5 –1.0 –1.0 –1.5 –1.5 –0.5 0 0.5 1.0 1.5 2.0 2.5 3.0 –0.5 0 0.5 1.0 1.5 2.0 2.5 Longitudinal distance (m) Longitudinal distance (m) Water Security and Drought Resilience in the South of Angola 87 PHOTO 4.20. Details of the GB3 Dam Site Area Note: Detail of the dam area—violet polygons: granite outcrops—cyan line: proposed dam (layout 1) green line: wall on left bank to seal erosion channel—yellow polygon: channel area to be filled—red line: layout 2—blue arrows: main water flow direction—black arrow: possible water flow during peak flows. FIGURE 4.21. Dam Profile and Rough Bedrock Pattern Site CHB1 (Rio Chingo) - proposed dam profile and morphology 550 Cross-section S1 Cross-section S2 549.5 549 0.7 m Elevations in m asl 548.5 548 Spill out 50 m 547.5 547 546.5 Granite Sand 546 545.5 545 0 10 20 30 40 50 60 70 80 90 100 Distance in m Ground surface Dal top Bedrock Excavation 88 Water Security and Drought Resilience in the South of Angola should reach 7,200 m3. Unfortunately, the alternative was not clear during the mission and the alter- native layout was not surveyed by GPS. A topographic survey with proper equipment can clarify the doubts. In the proposed layout 1, a wall should be added to the main dam body, to block the small channel between the two outcrops passing behind the outcrop G2a. This wall would connect the left (eastern) abutment leaning against the outcrop G2a, with the outcrop G2b about 15 m upstream. The wall should be about 24 m long and 0.5 m thick. The channel behind the wall should be filled with large stones. The spill out section shall be 60 m long, and 0.7 m lower than the crest; if layout 2 is selected, which accounts for a possible increase in height and length, it is suggested to increase the spill out length to 90 m and the height to 0.8 m. At this site a rip-rap is not necessary because the stream bed is already rocky, but to reduce the ero- sion of the basement and of the dam base the wall has been designed with a slope of 40° for the downstream side. In case layout 2 is chosen a rip-rap could be necessary in the left section of the course. The mean height is about 3.0 m considering also the depth of the foundation (see figure 4.21). The volume of the dam body is estimated 340 m3. The volume of the lateral wall is 30 m3. In case layout 2 is chosen, the dam volume increases to about 470 m3. The material to be used is cement and blocks (masonry stone). Blocks, stones, and sand (for the cement) are available 100–200 m from the site. The site is reachable by dirt road (7.8 km long) from the EN 280. The top of the right bank is easily reachable from the road but a section of about 100 should be cleaned to access the site; given the rocky nature of the site, further works can be required to insure the movement on the site of mechanics’ machines and trucks. ­ Site CH1 The characteristics of the proposed dam are synthesized in table 4.9. The possible dam profile is shown in figure 4.22, the sections in figure 4.23, and the planimetry in photo 4.21. The site CHB1 is the only one visited that fits social and geomorphological requirements in the Chingo area. As described above, in the site there is an old dam, partially destroyed. After the dam break, the stream took a different course in the new gap and probably the erosion dropped its bed of almost 1 m. The proposed layout, yet to be verified, is located very close to the existing dam, with the central section TABLE 4.9. Stream Flow and Dam Parameters Mean Dam Yield Stored Height on Max Runoff Peak Peak Water Dam Stream Alluvial Dam Wall Rip-Rap Volume Month Yield Volume Length Bed Thickness Volume Volume Runoff m /year 3 m /sec 3 Days m /sec 3 m /flow 3 m m m m 3 m3 7.10E+07 20.6 60 197 5,400 95 2.0 2.0 344 100 Water Security and Drought Resilience in the South of Angola 89 FIGURE 4.22. Dam Cross-Sections Show Two Different Patterns Due to the Absence/Presence of a Sandy Cover of the Granite Basement a. Cross-section S1 b. Cross-section S2 2.5 4.0 3.5 Spill-out level 2.0 3.0 Masonry stone 1.5 Elevation above rock basement (m) Elevation above rock basement (m) 2.5 Rip-rap protection Geotextile 1.0 2.0 1.5 0.5 1.0 0 0.5 Stream sand 0 –0.5 –0.5 Granite Granite –1.0 –1.0 –1.5 –1.5 –0.5 0 0.5 1.0 1.5 2.0 –0.5 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 Longitudinal distance (m) Longitudinal distance (m) PHOTO 4.21. Details of the Dam Area Note: Detail of the dam area—violet polygons: granite outcrops—blue line: proposed dam—blue arrows: water flow direction—green area: rip rap. 90 Water Security and Drought Resilience in the South of Angola based in the new course. The outcrop on which the left abutment can be based extends below the section of the truncated dam, while the right abutment can be based on an outcrop near the edge of the ­ old right wing. In two pits excavated in the middle of the new course the basement was found at 1.4 and 1.7 m of depth. The dam should rise about 2 m on the stream bed with a spillout of 70 cm of height and 50 m of length. Since the maximum depth of the basement is estimated at 2.0 m, the maximum height of the dam can reach 4 m, the length 90 m. With this configuration a throwback of almost 500 m is expected, since the slope is estimated to be less than 0.5 percent and the total volume of sand that can be saturated is about 45,000 m3. The stored water volume should reach 5,400 m3 per flow. The old dam should be dismantled, at least in the section near the course, and the blocks reused for the new dam. On the basis of the results of a future topographic survey, some minor works could be required to connect the new dam with leftovers of the old one. The dam body volume is about 330 m3. A rip-rap made of large blocks must be set at the dam foot in the central section where alluvial deposits are present; a geotextile sheet is set at the base. A volume of 108 m3 (140 m2 x 0.8 m) of granite blocks is estimated. The material to be used for the cons- truction is cement and blocks (masonry stones). The dam should be anchored in the basement in a section of 1 m2. Blocks, stones, and sand are available 100–200 m from the site; the blocks, apart from ­ those supplied by the old dam leftovers, should be obtained by breaking the outcrops in the surroun- dings, since the deposits are mostly made of sand and gravel. The site is accessible by dirt road, starting a few kilometers after the town of Bentiaba. Remarks, Recommendations, and Conclusions on Managed Aquifer Recharge Options •• It is clear that the above investments would provide significant water storage benefits. At present, water needs are only partially satisfied by seasonal events and hand-dug wells (cacimbas), deepe- ned until the dry basement is reached, in unacceptable hygienic conditions. •• The proposed solutions offer an additional water supply that for a minimum of 3 months to more than 6 months would cover a population of a minimum of 500 persons with 500 cows and 1,000 sheep and goats and the possibility of being used for small gardens. The local communities would need to balance the use the new resource to cover domestic needs, livestock use, and potentially vegetable gardens. •• Environmental co-benefits are expected from the development of these sites. For example, in the calculations of possible water storage, a secondary positive effect that, at least in the sites GB3 and CH1, is revealed by the presence of a limited groundwater flow that gives birth to seasonal water points is not being accounted for. In this case the damming effect shall supply an additional recharge for limited periods even after the end of the rainy season. •• The three sites selected present suitable conditions for the construction of sand dams. Nevertheless, the height of the banks and the low morphology of the surrounding areas make the sites subject to Water Security and Drought Resilience in the South of Angola 91 risk of flood during the peak of the rainy season. In this situation, it is necessary to perform a proper topographic and geognostic campaign.3 Infrastructure Costing There are two types of costs that need to be considered: the capital expenses to build infrastructure, and the operational expenses to operate, maintain, and repair the infrastructure during its lifetime of use. In Angola, lots of resources have been spent building infrastructure such as wells and boreholes, but much of this infrastructure has fallen into disrepair and is not functional due to the lack of resour- ces allocated for routine maintenance and repairs. It is estimated that 30 percent of new water points fail within the first year of use, and another 30 percent fail within the following 2 years due to lack of proper maintenance and repairs. Once a water point fails, years can go by before it is repaired, leaving the community to struggle without water, due to the lack of capacity to perform repairs that can often be of very low cost. In brief, building things is only half of the solution. The other half, which is the biggest obstacle to water security and climate resilience in Angola, is about strengthening systems and institutions that are fit-for-purpose, and growing staff capacity to deliver a service and be accountable. While a detailed analysis of an efficient maintenance system and its operational costs for the South of Angola will be the object of the next report, this section provides a first estimation of the costs of ­ ­ evelopment investments, based on prices consulted locally, for the various types of water resources d infrastructure presented at the beginning of this chapter. Cost Estimates of Boreholes Borehole Cost Components The cost of a borehole includes a number of components: •• Siting of suitable drilling location, including a desktop study of information on geology, hydrogeology (including data from successful and failed boreholes in the region), etc., and site ­ investigations comprising field survey, geophysical investigations, etc.; •• Mobilization of rig and drilling crew; •• Drilling, with the most adequate technology depending on targeted depth and type of geological formations encountered; •• Well development and testing; •• Installation of well equipment (casings, screens, well head protection, etc.); •• Cost of water supply infrastructure (pump, tank, taps, platform, fence, etc.); •• Cost to organize community management and training 92 Water Security and Drought Resilience in the South of Angola Because each of these cost components can vary widely from site to site, the cost of a borehole is highly variable. The main factors influencing local borehole costs are: •• Hydrogeological setting and complexity. Depending on the geology and climate of the region, groundwater can be encountered at different depths and in different types of geological forma- tions. Depending on the size, nature, and depth of an aquifer, the chances of encountering useful quantities of fresh groundwater vary widely and will require more or less advanced and elaborate field investigations and analysis, to enhance the chance of a successful borehole. •• Availability of drilling- and other (hydro)geological data. The knowledge of the prevailing groundwa- ter conditions in a region directly relies on the availability of information from existing successful and failed boreholes. Where this information is not available, more in-depth investigations for a new borehole will be required. •• Locally available drilling technology, expertise, and materials. The breadth and depth of expertise and quality and variability of drilling equipment available in a region will increase the options for choosing the most efficient, advanced, and cost-efficient technology and will result in higher suc- cess rates and lower unit costs of new boreholes. •• Remoteness of location. Mobilization costs for crews and equipment for complex field surveys and drilling depend directly on the remoteness and accessibility of the location. •• Unit costs of well- and water supply equipment and parts. Depending on the presence and impor- tance of local markets for drilling services, the cost of living, and the prevailing taxes and govern- ment regulations, the cost of well and pump equipment and services vary widely from country to country and from region to region. Borehole Cost in South Angola Based on evidence from the field surveys the cost of a borehole in South Angola tends to be high com- pared to other countries in the region. Besides the sum of the cost components that make up the price of a single well, the actual cost of a func- tioning well depends directly on the drilling success rate. Poor siting and low levels of expertise result in high drilling failure rates, requiring to drill several dry boreholes for every successful one, multiplying the unit cost of a functioning borehole. In addition, poor drilling practices and lack of maintenance arrangements reduce the lifetime of boreho- les, increasing further the unit cost of a functioning borehole. In sum, the absence of information, adequate siting practices, and infrastructure maintenance arrange- ments contributes to the high cost of boreholes in South Angola. Investments in new water supply infrastructure, aimed at strengthening drought resilience, should be accompanied by efforts to increase the regional knowledge base and to increase institutional and stakeholder capacity in the development Water Security and Drought Resilience in the South of Angola 93 and maintenance of groundwater infrastructure, in order to increase success rates and lower the unit cost of wells. The numbers below indicate cost ranges for different categories of boreholes: •• The cost of a borehole tapping a shallow aquifer (depth range 20–150 m) can vary from US$20,000 to US$100,000 (including taps, pump, solar panel and reservoir). •• The cost of a borehole increases sharply with increasing depth, due to the necessity to use heavier drilling equipment and to increase the drilling diameter, and as a result of the increased complexity and correlated risk of failure. The cost range for a borehole tapping a deep aquifer (depth range 150–300 m) can vary from US$75,000 to US$500,000. •• For deep boreholes, abandoning the well because of high salinity levels proves to be very costly. The installation of small solar desal units to lower salinity levels to drinking water standards can be an economic alternative. The indicative additional capital investment for such a unit is about US$140,000. Preliminary Costing of Surface Water Harvesting Infrastructure Chimpaca and Cistern Cost Components In the geographical setting of the Cuvelai lowlands, the cost of chimpacas and cisterns depends to a large extent on the excavated volume of sediments and, even more, on the underlying limestone. Deep ­ chimpacas and cisterns increase water security but are significantly more costly due to the greater depth of limestone to be excavated. Other cost elements of chimpacas are the construction of the boundary talus, water intake, fencing, and water distribution infrastructure (solar panels and pump, piping, storage reservoir, and troughs). A schematic layout and cross-section for a standard chimpaca are included in Appendix C. Besides excavation costs, other cost elements for cisterns are the construction of concrete walls on top of the limestone substratum, the stormwater intake, and sediment trap to reduce the rate of silting of the cisterns, the roof and water distribution infrastructure (solar panels and pump, piping, storage reservoir, and taps). Adding a roof to cisterns reduces contamination by bird droppings and organic matter and, if designed properly, evaporation losses. A schematic layout and cross-section for a cistern are included in Appendix C. Chimpaca and Cistern Cost Estimates Detailed chimpaca cost estimates are shown in table 4.10. The cost for a standard 80 x 60 m chimpaca (maximum storage ~12,000 m3) is ~ US$160,000. The cost for a “strategic” large and deep 120 x 60 m chimpaca (maximum storage ~38,000 m3) is ~ US$800,000. 94 Water Security and Drought Resilience in the South of Angola TABLE 4.10. Cost Estimates for a Standard Chimpaca and a Strategic Large Chimpaca Unit Cost Description Unit L (M) W (M) H (M) Quantity (US$) Total (US$) Cost-Estimate of Chimpaca So It 60 X 4 M (Exclusive of Land Acquisition) 1 Mobilization 1 4,000 4,000 2 Removal of trees/shrubs m2 104 84 8,736 0.5 4,368 3 Excavation of top sediments (slope 3/12) m 3 80 60 3 9,792 S 48,960 4 Excavation of limestone substratum with m 2 56 36 1 2,016 40 80,640 excavator and hydraulic hammer S Construction of talus m 92 72 328 20 6,560 6 Rock filter storm water intake m 2 18 10 180 15 2,700 7 Fencing m 104 84 376 10 3,760 S Throughs, piping, solar pump, solar 1 10,000 10,000 panels, 5,000 L reservoir Maximum storage (m3/season) 11,808 Total cost: 160,988 Cost-Estimate of Chimpaca 120 X 80 X £ M (Exclusive of Land Acquisition) 1 Mobilization 1 4,000 4,000 2 Removal of trees/shrubs m2 144 104 14,976 0.5 7,488 3 Excavation of top sediments (slope 3/12) m3 120 80 3 22,032 5 110,160 4 Excavation of limestone substratum with m3 96 56 3 16,128 40 645,120 excavator and hydraulic hammer S Construction of talus m 132 92 448 20 8,960 6 Rock filter stormwater intake m 2 18 10 180 15 2,700 7 Fencing m 144 104 496 10 4,960 8 Throughs, piping, solar pump, solar 1 10,000 10,000 panels, 5,000 L reservoir Maximum storage (m3/season) 38,160 Total cost: 793,388 The unit cost per cubic meter of stored water is higher for a large chimpaca, due to the increased depth. Reducing the area/depth ratio reduces daily evaporation losses and increases the capacity of the chimpaca to bridge a year of below-average rainfall. These cost estimates do not include the cost of ­ ­ preliminary studies and surveys or of the acquisition of land. Detailed cistern cost estimates are included in table 4.11. The cost for a standard 20 x 5 m cistern is about US$32,000. The cost for a large and deep 30 x 5 m cistern is about US$62,000. These cost estimates do not include the cost of preliminary studies and surveys, the acquisition of land and the cost for demarcation or fencing of the impluvium. For public health reasons the impluvia of cisterns need to be protected from pollution, including animal droppings, household waste, and fuel and lubricants from vehicles or engines. Water Security and Drought Resilience in the South of Angola 95 TABLE 4.11. Cost Estimates for Standard and Large Cisterns L W H Unit Cost Total Description Unit (M) (M) (M) Quantity (US$) (US$) Cost-Estimate of Cistern 20 X 5 X 5 M (Exclusive of Land Acquisition) 1 Mobilization 1 2,000 2,000 2 Removal of trees/shrubs m 2 30 15 450 0.5 225 3 Excavation of top sediments (slope 3/12) m3 20 5 3 378 5 1,890 4 Excavation of limestone substratum with m3 20 S 2 200 40 8,000 excavator and hydraulic hammer 5 Reinforced concrete wall m3 20 5 4 60 120 7,200 6 Concrete sediment trap & rock filter m2 8 4 1 1,500 1,500 7 Corrugated metal roof m2 20 5 100 65 6,500 8 Fencing m 30 15 90 10 900 9 Taps, piping, solar pump, solar panels, 1,000 L 1 6,000 6,000 reservoir Maximum storage (m3/season) 500 Total cost: 32,215 Cost-Estimate of Cistern 30 X 5 X 3 M (Exclusive of Land Acquisition) 1 Mobilization 1 2,000 2,000 2 Removal of trees/shrubs m2 40 15 600 0.5 300 3 Excavation of top sediments (slope 3/12) m3 30 5 3 558 5 2,790 4 Excavation of limestone substratum with m 3 30 750 40 30,000 excavator and hydraulic hammer 5 Reinforced concrete wall m3 30 5 4 84 120 10,080 6 Concrete sediment trap & rock filter m 2 8 4 1 1,500 1,500 7 Corrugated metal roof m 2 30 5 150 65 9,750 8 Fencing m 40 15 110 10 1,100 9 Taps, piping, solar pump, solar panels, 1,000 L 1 6,000 6,000 reservoir Maximum storage (m3/season) 1,200 Total cost: 61,520 Preliminary Costing of Sand Dams Evaluation of Water Demand For a preliminary cost-benefit analysis, the capital cost of building the proposed sand dams was ­estimated.4 The estimate is based on the designs defined in the previous section. With regard to bene- fits, there are problems with the estimate of the resident population, since the administrations were not able to supply data for an area subject to nomadism. For this reason, a rough analysis of the presence of single huts in a radius of 2.5 km has been done via Google Earth satellite images. It is important to note that this radius makes the study extremely conservative, since the communications with the adminis- tration revealed that people and cattle travel daily more than 5 km, and in some cases around 10 km. 96 Water Security and Drought Resilience in the South of Angola In the Giraul area the sites selected (GB2 and GB3) are at 3 km of distance each other so the same popu- lation can be considered sharing the same sources. It should be considered also that, according to the declaration of the community leaders, the huts can be easily shifted if a permanent or semi-permanent source is created. About 50 huts have been detected; assuming an average family composition of 10 persons, this makes a total number of 500 persons. This is also the number supplied by the community leaders. It is more difficult to estimate the number of animals; a tentative number of 500 cows and 1,000 among sheep and goats is assumed for the benefits analysis. It should be remembered that three more water points in the Giraul area exist, one of which is at about 1 km from GB3 and 3 km from GB2. In this location there are two to three hand dug wells with brackish water that is used mainly for livestock. Moreover, a perma- nent captured spring (and some hand-dug wells) are at 4.5–5.0 km from the two sites. For this reason, the water from a possible new aquifer from the future dam shall supply mainly water for humans and for some rural activity. The Chingo area is less populated and in a radius of 3 km, 25 huts have been counted. The demand of water is assumed as 25 l/day/person, 5 l/day per sheep or goat, 20 l/day/cow, and 5 m3/day per irrigated hectare. Site GB2 An approximate estimate of the capital investment is provided in table 4.12 along with other key parameters: In the target area, in a radius of 2.0 km from the source, a number of 25 huts has been detected that, assuming an average family composition of 10 persons, makes a total number of 250 persons. A number of 250 cows and 500 sheep and goats is assumed. Of course, in case a dam is built in an alternative to site GB3 the numbers of population and livestock can easily double. For this reason, we shall consider this option for the cost-benefits analysis. TABLE 4.12. Costs Analysis for the GB2 Site Sand Dam Description Unit Quantity Unit Cost (US$) Total (US$) 1 Mobilization 2,000 2 Excavation in soft riverine alluvium m 3 500 6 3,000 3 Excavation in crystaline bedrock (granodiorite) with use of m 3 90 80 7,200 excavator and hydraulic jack hammer 4 Rip-rap downstream protection (stones on site) m3 144 40 5,760 5 Polyester geotextile to prevent the sinking of embankment m 3 200 3 600 6 Masonry stone wall m 3 282 100 28,200 Total Cost (USD) 46,760 Water Security and Drought Resilience in the South of Angola 97 Based on the above, the sand dam with a storage capacity of 3,240 m3 and a capital cost of US$46,760, serving a population of 500 people, has been used. According to the estimates, the target beneficiaries own, altogether, an estimated 1,000 sheep and goats and 500 beef cattle. In addition, it is considered that 5 acres of irrigated farms could develop. The daily water demand for the area would be approxima- tely 38 m3/day. This means that the water stored by the dam can supply the population for about 3 mon- ths after the end of the rainy season. Considering only consumption for domestic use the period extends to 9 months that covers the whole dry season. The capital cost of the earth dam is estimated at (US$46,760 / 500) = US$93.52 per person. The sand dam is assumed to have a lifespan of 30 years, considering the total volume stored (30 x 3,240 = 97,200 m3) a cost of US$0.48 is obtained. The dam presence in the area surrounding the site GB2 does not present evident hydrological risks because the banks morphology insures a natural protection against flooding events. Actually, the left bank (SE) is topped by a plain that is almost 1.5 m higher of the dam top, while the right bank (NW) raises with a slope of 20–30 percent toward the hilltop. However, in the site a peak flow of 68 m3/s is expected, only in case of exceptional rainfall, and the designed spill out can face a flow up to 60–70 m3/s. There is a slight risk of small landslides on this flank, but it is minimized by the presence of trees and by the probable presence of the basement at shallow depth. Site GB3 An approximate estimate of the capital investment for the two possible dam layouts5 is provided in table 4.13, along with other key parameters: In the target area, in a radius of 2.0–2.5 km from the source, a number of 50 huts has been detected that, assuming an average family composition of 10 person, makes a total number of 500 persons. This is also the number supplied by the community leaders. It is more difficult to estimate the number of animals; a tentative number of 500 cows and 1,000 sheep and goats is assumed for the benefits analysis. Based on the above, the sand dam with a storage capacity of 3,600 m3 (layout 1) and a capital cost of US$39,840 has been used. According to the estimates, the target beneficiaries own, altogether, an estimated 1,000 sheep and goats and 500 beef cattle. In addition, it is assumed that 5 acres of familial ­ agriculture could be developed. The daily water demand for the area would be approximately 38 m3/day. This means that the water stored by the dam can supply the population for about 3 months and a week after the end of the rainy season. Considering only the consumption for domestic use the period extends to more than 9 months that covers the whole dry season. The capital cost of the earth dam is estimated (US$39,840 / 500) US$79.7. The sand dam is assumed to have a lifespan of 30 years, considering the total volume stored (30 x 3,600 = 108,000 m3) a cost of US$0.37/ m3 is obtained. Considering the dam layout 2, the storage capacity increases to 7,200 m3 (layout 1) and the capital cost to US$60,400. The capital cost per person of the earth dam is estimated (US$60,400 / 500) US$120.8. Since the daily water demand is the same (38 m3/day), the storage can supply water for 6 months and 2 weeks. In this case the unit water cost drops to US$0.28/m3. 98 Water Security and Drought Resilience in the South of Angola TABLE 4.13. Costs Analysis for the GB3 Site Sand Dam Description (Lay-Out 1) Unit Quantity Unit Cost (US$) Total (US$) 1 Mobilization m3 2,000 2 Excavation in sandy riverine deposits m3 40 6 240 3 Excavation in crystaline bedrock (granite) with use of m3 85 80 6,800 excavator and hydraulic jack hammer 4 Masonry stone dam m3 340 80 27,200 5 Reinforced concrete wall m 3 30 120 3,600 Total Cost (USD) 39,840 1 Mobilization m 3 2,000 2 Excavation in sandy riverine deposits m 3 400 2,400 3 Excavation in crystaline bedrock (granite) with use of m 3 130 80 10,400 excavator and hydraulic jack hammer 4 Rip-rap downstream protection (stones on site) m3 100 40 4,000 5 Masonry stone dam m 3 520 80 41,600 Total Cost (USD) 60,400 Actually, in this site, given the presence of a semi-perennial water source an underground recharge can be expected at least for 2–3 months after the rains end. In this case the dam presence contributes to an extension upstream, beyond the estimated throwback, of the alluvium saturated section. In other terms the water accumulated in the throwback area is substituted, at least partially from the upstream flow and the period of water availability increases of 2–3 months. This hypothesis can be verified only after the dam construction. The Rio Kapangombe is the main branch of the Giraul stream and has a wide basin and a peak flow of about 144 m3/s. The spill out in the design has a section of about 40 m2 and even considering a velocity of 2 m/s can contain only a flow of 80 m/s. In other terms it is possible that in case of exceptional events the water not only can surmount the whole crest, but it could flood the left bank and lap on the top of the right bank. Actually, community leaders confirmed the latter event and the absence of huts on the left bank could confirm the former. The same community leaders declared their availability to shift the nearest hut settlement (3–4 huts) to higher areas, in case of the dam construction. Site CH1 An approximate estimate of the capital investment is provided in table 4.14, along with other key parameters: It has to be considered that further works can take place in the target area. In a radius of 4 km from the site, a number of about 25 huts has been detected that, assuming an average family composition of 10 persons, makes a total number of 250 persons. Considering that after the dam construction some Water Security and Drought Resilience in the South of Angola 99 TABLE 4.14. Costs Analysis for the CH1 Site Sand Dam Description Unit Quantity Unit Cost (US$) Total (US$) 1 Mobilization 4,000 2 Excavation in sandy riverine deposits m3 400 6 2,400 3 Excavation in crystaline bedrock (granite) with use of m3 95 80 7,600 excavator and hydraulic jack hammer 4 Polyester geotextile to prevent the sinking of m3 150 3 450 embankment 5 Rip-rap downstream protection (stones on site) m3 112 40 4,480 6 Masonry stone wall m 3 344 100 34,400 Total Cost (USD) 49,330 more pastoralists might use the source since the area westward seems drier than the Giraul area, it has been considered that the resident population could increase to 500 persons. Based on the above, the sand dam with a storage capacity of 5,400 m3 and a capital cost of US$49,330 has been used. According to the estimates, the target beneficiaries own, altogether, an estimated 1,000 sheep and goats and 500 beef cattle. In addition, it is considered that 5 acres of irrigated farms could develop. The daily water demand for the area would be approximately 38 m3/day. This means that the water stored by the dam can supply the population for about 5 months after the end of the rainy season. Considering only the consumption for domestic use the period extends to more than 1 year. The capital cost of the sand dam is estimated at (US$49,330 / 500) US$98.7 per person. The sand dam is assumed to have a lifespan of 30 years; considering the total volume stored (30 x 5,400 = 162,000 m3) a cost of US$0.30 per cubic meter of water is obtained. Actually, in this site, given the presence of a semi-perennial water source, and the absence of rocky outcrops in the stream bed for a large section, an underground recharge can be expected at least for ­ 2–3 months after the rains end. The main risk of the area is the high peak flow that can be expected at the top of the rainy season that is about 200 m3/s. Actually, the destruction of the old dam is evidence of strong flows.6 Since the slope of the stream has been estimated as no more than 0.3 percent, the velocity of the flow should not exceed 1.5 m/s. For this reason, the designed spill out can channel no more than 60 m3/s. Therefore, after the topographic-geognostic survey is done, the chance to increase the length and the height of the dam, possible only if the basement extends and rises under the alluvial cover at the edges, should be verified. In this way also the spill out can be enlarged and deepened and allowed a larger flow. In a negative case great attention should be given to the abutments and to the dam crest to protect them from erosion. Furthermore, it is suggested that stakeholders evaluate possible design solutions to connect the new dam to the earth wing on the right bank and to the remains of the dam wall on the left bank. 100 Water Security and Drought Resilience in the South of Angola In the coast analysis, the construction of lined wells linked to the dams has not been included. Given the shallow depth of the basement also after the sand accumulation in the throwback area (3–4 m), the costs of such works should not be significant. Conclusion The capital costs of the proposed sand dams range from US$40,000 to US$60,000. The difference mainly depends on the length of the dam wall that ranges between 70 m and 100 m and on the height that varies between 2 m and 3 m. To this should be added the optional cost of a shallow well per dam, never beyond a few thousand US dollars. The cost of one cubic meter of water obtained by the dam construction ranges between US$0.3 and US$0.5, the investment per person, calculated based on 500 persons per dam, varies from US$80 to US$120. The assumption at the base of these calculation is a dam lifespan of 30 years, but it should be considered that if well-built the dam life cycle can extend to 50 years or more, with a significant drop of the last two parameters. In any case, this preliminary cost analysis is highly conservative in the benefits estimates and probably the development of the infrastructure is more profitable than expressed. Summary of Infrastructure Cost Estimates Cost ranges for various types of water supply and water harvesting infrastructure have been compiled from local sources and are shown below (table 4.15). As few cost examples were available from Angola, infrastructure costs from other countries in the region and from African countries with similar hydro- geological and climatological settings have been included in the comparison. Costs of infrastructure tends to be higher in Angola, partly because of capacity and governance issues mentioned above. TABLE 4.15. Overview of Cost Estimates for Different Types of Rural Water Supply Infrastructure Harvested/Produced Type of Infrastructure Typical Dimensions Volume of Water Cost Range (U$) Groundwater Open well Dia 1.5 x D20 to Dia 4 x D10 1 to 10 m3/day 6,000–12,000 Borehole w/solar pump Depth 20 to 150 m 5 to 20 m /d ay 3 20,000–100,000 Borehole w/desal unit Depth 20 to 150 m 5 to 20 m /day 3 160,000–240,000 Deep borehole Depth 150 to 300 m 5 to 100 m /hr 3 75,000–500,000 Water Harvesting Infrastructure Sand dam L50 x H1.5 to L100 x H2.5 720 to 5400 m3/season 50,000–150,000 Chimpaca L80 x W60 x D4 to L120 x W80 x D6 19,200 to 57,600 m3/season 120,000–800,000 Cistern L2Q x W5 x D5 to L30 x W5 x D8 500 to 1,200 m3/season 30,000–75,000 Water Security and Drought Resilience in the South of Angola 101 A more detailed economic analysis, using local equipment, material and expertise costs, and based on actual infrastructure designs reflecting local hydrological and hydrogeological conditions will be made in a follow-up phase. Identification of Risks Associated with Building Small-Scale Water Supply Infrastructure The South of Angola represents a complex operational environment with risks of different natures that could hamper the achievement of the objective of building drought resilience through the construction of proposed small-scale infrastructure solutions. Some of the most significant ones, and the measures that will be implemented to mitigate them, are: •• The political and institutional environment at the higher level (national, provincial, river basin agencies, etc.) can have a direct impact on the success of development projects. As explained throughout chapter 4, the intended pilots designs are small-scale and community based and as such are more disconnected from the high-level political economy and therefore protected from the potential higher-level political issues. •• Inadequate budgetary and institutional capacity for project implementation and supervision could hamper success and sustainability of the infrastructure pilots, too. The pilots will be designed with additional levels of monitoring and implementation support to assist in the creation of the required structures. •• Coordination at the different administrative levels is weak in the region and the strategies are often not coherent within and among relevant sectors and agencies. The proposed rural infrastructure pilots will be an opportunity not to be missed to enable dialogue and create platforms for policy dialogue at local levels. •• Communities can abandon the pilot projects after exit of the donors. Stakeholder engagement and community capacity building will be stressed in order to reduce this risk. Community engagement needs to start from the first planning phases and will need to be prolonged and intensive enough, so community ownership is built and operation and maintenance is planned and in place. Local priorities, livelihood dynamics, and resource conditions will inform and drive the activities to come. •• Poor community management of rehabilitated or newly created water points can degrade far- mlands and water quality. Communities’ capacity to manage the infrastructures and to develop sustainable and adequate water-use plans must be emphasized. Capacity-building components to understand better the significance of these risks in the context of southern Angola and to work collectively to avoid them are planned for the follow-up phase of this work. •• Localized conflicts around the expansion of water access and resource availability can arise in this semiarid region, especially as droughts unfold. Availability of water and fodder controls movement 102 Water Security and Drought Resilience in the South of Angola and migration of herders in this region, which can lead to increased pressures on pilot sites. The pilots have the potential to disrupt current community dynamics and interrelationships among groups. To minimize these threats, the follow-up work will include a full component to understand agropastoral habits and transhumance migration routes, in preparation of posterior piloting of the infrastructures. Also, the pilots will be small- scale and focused on the limited community demands, assuming that these will not create a prominent contrast between settlements. •• Technical design and planning of the infrastructure options is risky. Poor planning and design can lead to infrastructure failure or insufficient storage, leading to frustration in the communities. To mitigate this, part of the follow-up work focuses on conducting targeted water resources studies, working extensively with geospatial tools and field surveys to support infrastructure planning, and recognize priority catchments for water harvesting or adequate locations for groundwater develo- pment, depending on the selected solutions. •• Climate risks can impact the reliability of proposed rural water infrastructure investments. Alleviating potential climate risks and strengthening the robustness of infrastructure in the face of drought are precisely at the heart of the technical design of the proposed options for rural water supply investments. However, floods are a real threat to some of the envisioned water infrastruc- ture, so especial emphasis will be put on designing them to withstand extreme flows. •• Interventions based on water harvesting, such as chimpacas or sand dams, can entail a modifica- tion of downstream water availability. In addition, some of the identified priority areas are within transboundary basins or aquifers. These risks will be carefully studied and evaluated from the next phase of work phase onward. The implementation of the infrastructure options suggested in chapter 4 through an investment project financing, would need to observe a range of Environmental And Social Standards (ESS) from the World Bank Environmental and Social Framework, potentially including the following depending on design and location: ESS1-Assesment and Management of Environmental and Social Risks and Impacts; ESS4- Community Health and Safety; ESS5-Land Acquisition, Restrictions on Land Use and Involuntary Resettlement; ESS6-Biodiversity Conservation and Sustainable Management of Living Natural Resources; ESS7-Indigenous People’s/Sub-Saharan African Historically Underserved Traditional Local Communities; ESS8-Cultural Heritage; and ESS10-Stakeholder Engagement and Information Disclosure. Notes 1. If the topography, geology, and surface hydrology are suitable for the construction of surface water runoff harvesting and storage, cisterns and chimpacas should be built in pairs, separating human and animal water consumption. Cisterns for human consumption are smaller and deeper than chimpacas to reduce evaporation losses and contamination, and are typically covered and equipped with a sediment filter/trap. Water can be supplied through taps using solar energy. The traditional chimpacas can provide the bulk of water demand for animals (except during prolonged droughts). 2. Information supplied by administrators and confirmed by community members. 3. Details for the topographic and geognostic campaign are given in Appendix D. Water Security and Drought Resilience in the South of Angola 103 4. The analysis is based on African countries where the prices may be higher (10–20 percent) than in Angola. This research will be fine-tuned in a next phase. 5. Two possible layouts have been taken in consideration: the first (layout 1) considers a dam length of 75 m and a mean height of 2.0–2.5 m, the second (layout 2) is 130 m long with a height of 2.5–3.0 m. The decision about the final layout shall be taken after the topographic-geognostic survey. 6. It is to be noted that the old dam was not founded in the basement and at the point where the water flow broke it, the dam base did not reach the basement. 104 Water Security and Drought Resilience in the South of Angola Chapter 5 Conclusions and Recommendations: Building Drought Resilience in the South of Angola Drought impacts are far from being an “Act of God”: they are modulated by institutions, management systems, and investments. They are also heavily modulated by the functionality of existing systems and infrastructure during non-drought times, and having fit-for-purpose institutions and resources. In any context, building resilience means to have robust and functional systems, mechanisms for fixing them, having many options in case some fail, and being prepared and flexible to adapt as shocks arrive. In the course of the most recent drought (2012–19), the Government of Angola and its partners progres- sively mobilized numerous emergency resources to mitigate drought impacts when these were already occurring and the suffering in the South was widespread. The government is well aware of the need to shift from a reactive mode toward a preparedness mode, and this awareness led to the ministerial request for this study. The resources spent in emergency responses usually achieve limited results and come at very high costs. On the other hand, these same resources yield much more long-term benefits when they are spent in preparedness and in building resilience before a drought. The ultimate goal of this work is to understand the spatial distribution of water-related factors contribu- ting to drought vulnerability, in order to prioritize needs and inform the design of solutions to increase drought resilience across the South of Angola. This study has found that the main factors that determine vulnerability to drought in the South and Center of Angola, in regard to water resources access, are the following: 1. The general lack of efficient monitoring and information communication on the state of water points, as well as a very partial knowledge base of water resources. The lack of systematic informa- tion collection on the state of water points is the first missing link in the chain to maintaining water points’ functionality. The lack of knowledge on water resources is at the origin of numerous misgui- ded investments (i.e., the large number of dry wells drilled in response to the drought). 2. The lack of reliable access to water resources due to the insufficiency of resilient investments at the community level. Many areas have a high dependency on unprotected cacimbas and chimpacas for access to drinking water, which are the first affected during times of drought, and where quality deteriorates rapidly. 3. The lack of national and regional capacity to systematically repair water points, prepare for drou- ghts, and respond to droughts, from the community level to provincial and ministerial levels. This is partly due to the lack of allocated budgets and adequate resources for agency staff, which results in a lack of capacity to plan and manage, and a lack of information-to-action processes. Water Security and Drought Resilience in the South of Angola 105 This report provides a comprehensive approach to assess drought vulnerability across five provinces in the South of Angola. The approach combines a desk analysis of census data (INE 2014) at the commune level, feedback from partners and government agencies, and field work observations made during various missions. A regional vulnerability mapping and prioritization across the 130 communes of the 5 provinces is presen- ted here based on the structural water access conditions. The list of 130 communes ranked in order of decreasing vulnerability, based on their conditions of access to water, is presented in table 2.2. This list is a clear guide for helping prioritize water resource investments across the region. The vulnerability mapping has then been overlapped with the spatial distribution of drought intensity and severity as observed by satellites, and impacts reported and observed on the ground. This overlap of vulnerability and drought intensity and ground information has yielded the following picture: the communes hit hardest by the drought have been Mongua, Evale, Kafima, Ombala yo Mungu, Humbe, Mukope, Shiede, Naulila, Onkokwa, and Otchinjau in Cunene province; Tchipungo, Chimbemba, Chiange, Kapunda Kavilongo, and Jau in Huíla province; and Lucira, Camacuio, Caitou, Chinquite, Cahinde, and Chingo in Namibe province. In addition, the study also identified very poor water access conditions in locations where the drought has not had a strong physical signal. Sometimes, these conditions were found to be not too expensive to fix and would mostly require some local governance leadership (this is the case of Lucira, just to cite an example visited in stage 2). This study identified a range of water resource investments that could help build water security for resi- lience at the community level. It also proposes a decision framework to guide what types of investments will be best in each community or location, across the varied contexts of southern Angola. Within that context, the team explored the feasibility of some adapted solutions in more detail, such as coupled chimpaca and cistern systems in Cunene province, and the construction and rehabilitation of sand dams in Namibe, identifying and characterizing a number of proposed sites. Based on the findings of this study and looking forward, the following recommendations are made to build resilience in water access conditions in the South of Angola. Recommendation 1: Invest in Information And Knowledge 1. Address the lack of information and monitoring mechanisms of the state of water points. The func- tionality of water points is severely hampered in the South of Angola by the lack of adequate infor- mation on the status of water infrastructure and water resources. This can be addressed by taking the following steps: a. Update current information. Conduct a comprehensive inventory of existing data on water points, boreholes, and other water supply infrastructure in the region. This should focus on water points intended to supply the local population with drinking water. The following parameters should be collected: conductivity, pH, GPS coordinates, static water level and depth, color/smell of water, type of water point, functionality of water points, method of 106 Water Security and Drought Resilience in the South of Angola lift, pump type, need for rehabilitation, duration if the pump is out of order, and dewatering events of the point during the year. b. Establish mechanisms to update information. Information systems need to be alive, with new information updating the system regularly. These mechanisms in contexts like Angola should focus on people and information flows, as this does not require large investments, only intentionality, the accountability of institutions to fulfill their responsibilities, and training to georeference information and use simple databases. c. Manage information, share it, and use it. Operationalize information systems like the Water and Sanitation Sector Information System (Sistema de Informação do Sector de Água e Saneamento, SISAS) and associated platforms to periodically update and integrate information. A platform that only includes old information such as the current SISAS, is outdated and of limited use. Link this information system to an actions system: link it to maintenance and repairs programs. Enable sharing data within and across institutions. Train people in the provinces and municipalities to understand the system and use it. Revitalize the Groundwater Group (“Núcleo de Aguas Subterrâneas”), or a similar program or agency. 2. Invest in water resources knowledge. The limited number of available geologic hydrologic and hydrogeologic studies, as well as lack of data pose obstacles to designing good water resources investments. Good hydrogeologic studies and basin plans are needed to design basin-wide sustai- nable investments for resilient water resources management and access to water. Monitoring sys- tems (in situ, and remote sensing) are needed to monitor water resources—that is, groundwater and surface water, both in terms of quality and quantity. a. Conduct groundwater studies. Integrating existing and new information for strategic aquifer use in the region. For aquifer and groundwater systems, the following variables are critical: the geometry (extent and thickness) of the aquifer or aquifer systems; the boundary conditions, recharge areas, and outflows; the aquifer type; the hydraulic parameters; quantification of water balances and current uses and simulation models of main bodies; and characterization of groundwater quality, hazards, and use risks/threats. b. Conduct surface water and water balance studies. Understand water partitioning and dynamics with land use and land cover, through water accounting studies in the region. Use available data from both ground observations and satellite imagery. c. Establish Continuous monitoring of water resources. Continue and expand ongoing surface water monitoring efforts through hydrometric stations, integrate remote sensing estimates, and expand efforts to include groundwater monitoring. Monitoring groundwater dynamics is extremely important to understand aquifers, calibrate models, and close water balances. In addition, use digital platforms to integrate, store, and visualize water resource information. Examples, for the case of groundwater, include the Groundwater Information System (IGRAC) or the US Groundwater Watch. Water Security and Drought Resilience in the South of Angola 107 d. Involve Angolan universities. All the activities above are ideal to invest in local knowledge and make use of local talent and expertise, even when combined with international level assistance. University involvement in data and analysis tasks has many advantages, such as the cost-effective use of students to develop a body of knowledge, investing in future generations, etc. Share data. It is unfortunate that significant investments have been made in geology and e.  groundwater explorations but sharing the water-related data (e.g., the PLANAGEO study, implemented by the Ministry of Oil and Mineral Resources) seems to be an impossible task, even in the context of a national catastrophe such as the last drought in the South. Recommendation 2: Invest in Rural Infrastructure While the government of Angola is investing in new dams and water transfers in the Cunene-Cuvelai basins, and plans six new dams in the Namibe coastal basins (Programa de Medidas Estruturantes 2019), rural communities in large areas of the South of Angola will not benefit from these investments. A com- plementary program of investments at the community level will increase access to water resources and more spatially distributed small-scale storage. It is important to note that this ASA has concluded that the priority communities often cannot rely on a single solution to guarantee their domestic water supply and a conjunctive use scheme is essential for them. A range of infrastructure solutions can be used in combination to target and address the needs of these rural communities across the region. A few recom- mendations in this regard are listed below. 1. Implement an integrated program of complementary physical investments to develop and mobi- lize water resources reliably and sustainably for rural communities: The commune prioritization approach of chapter 2 and the decision framework to select feasible investments in a given location described in chapter 4 provide a good roadmap for physical investments in rural communities across the South. The overall approach for planning water resources investments at the community level is summarized in figure 5.1. These investments include the following. a. Large diameter wells to tap groundwater in shallow aquifers: Often located along (dry) riverbeds or depressions, the diameter varies between 1 m and 3 m and their depth is generally less than 20 m. Concrete or brick lining is used to stabilize the well. In low permeability formations larger wells are built that create sufficient storage that fills up during the night. Considering their generally low yield, cacimbas are suitable for small, isolated communities. b. Boreholes with a solar pump, storage tank, drinking taps, and basin for cattle: Where groundwater potential has been confirmed, boreholes equipped with a solar pump, storage tank, and taps can be installed. Groundwater often provides a clean and reliable source of drinking water, providing water security in times of drought. Wells typically range in depth between 20 m and 200 m, depending on the local geology. Solar pumps will be used to lift the groundwater to storage tanks from where it will be distributed to taps. Separate drinking 108 Water Security and Drought Resilience in the South of Angola basins can be installed for cattle. Groundwater presence across the South of Angola is variable, and in many parts current knowledge on groundwater dynamics is inadequate. The yield of small-diameter wells with PVC casing and a solar pump typically fluctuates between 5 m3/day and 20 m3/day. c. Boreholes with a small desal unit where brackish groundwater is present and alternative resources are scarce: For dispersed and remote communities far from the main water pipes and not easily accessible by road, treatment of mineralized groundwater could be a solution for drinking water supply. The use of compact solar-powered small desalination units of brackish groundwater can be a reliable and cost-effective alternative in certain settings. It is important to accompany the construction of desal units with adequate financial and technical management arrangements, considering that operation and maintenance (O&M) is what normally threatens desalination units. Operators conducting O&M shall be identified and the spare parts supply chain needs to be planned. d. Deep boreholes (more than 200 m depth): From the Namibian side of the border the presence of regional aquifers at greater depth has been confirmed. The presence and potential of deep aquifers in South Angola has yet to be confirmed but if confirmed, deep wells could be a strategic resource in view of the long-term water security of South Angola, providing secure water points for the region around them. Well depths will depend on the geometry and dynamics of the aquifers but would vary between 200 m and 400 m. Drilling and O&M costs are higher for large diameter deep wells, and adequate governance arrangements need to be put in place. Yields of some of the deep wells in the Namibian part of the Cuvelai basin exceed 100 m3/hr. e. Sand dams in areas with ephemeral streams but very small shallow aquifers: The typical sand dam is constructed where a stream has excavated its course in an impervious basement, such as a crystalline basement or a clayey formation or marls, and the rock is exposed. The dam is expected to be filled by sediments (normally sand) carried by the stream flow. As soon as a partial filling occurs, the sediments are saturated by water at each flow and can be tapped through shallow wells. Normally a dam of this type is 2–5 m high and less than 200 m wide. Depending on the flow regime, the geology/morphology and the water needs, other types of water harvesting and managed aquifer recharge (MAR) structures can be built, including check dams or spreading sills. f. “Improved” chimpacas: Direct harvesting of surface water in areas with limited or no groundwater potential. Chimpacas, excavated rectangular large pits in the very flat landscapes of Cunene, trap runoff water during the rainy season. Typically, mechanically excavated chimpacas store about 20,000 m3 of water per season but their design will be adjusted to water demand, surface water potential, and purpose (human water security or also animal water supply). The potential for harvesting surface water in the Cunene basin is substantial and should be developed even though surface water storage alone will not be sufficient to provide Water Security and Drought Resilience in the South of Angola 109 water security in times of prolonged droughts, and thus needs to be complemented by other methods. g. Runoff harvesting cisterns in small watersheds protected from pollution: These can store up to 1,000 m3 of fresh water and may be lined, covered, and equipped with solar pumps and taps. Traditionally chimpacas are used for both human and animal water supply, often leading to unhygienic conditions. For human water supply smaller cisterns can harvest water from comparatively small impluvia that need to be protected from grazing and other potentially polluting activities. Cisterns are partly excavated in the substratum and typically equipped with rock-and-gravel filters and screens, designed to reduce the sediment intake and provide clean drinking water storage during the dry season. Typically, cisterns can store up to 1,000 m3 of fresh water. Depending on the size of a community, several cisterns can be built. h. Bulk water supply through piped supply (piped connections to existing water pipelines). i. Old weirs and small dam rehabilitation (the Portuguese built many small weirs/dams in small rivers, which could be rehabilitated). j. Watershed storage, soil, and water conservation measures (traditional, local land management approaches to slowing down runoff water so that it does not erode and promotes infiltration). 2. Resilience means having more than one option: it is essential to plan for redundancies and storage buffers. The choice of infrastructure that can be built depends on the availability of groundwater or surface water resources in the region. Considering the overall demand of a community and the seasonality of the resources, a combination of infrastructure options may be necessary. 3. Plan with a regional strategic view: As described in chapter 4, planning must consider two levels: (1) the decision best suited to the community; and (2) regional resilience. Owing to the fact that different types of infrastructure might be planned in groups (a “chimpaca plan” based on a surface water resources study, a “strategic deep borehole plan” based on transhumance corridors, etc.), a regional vision for resilience also needs to be overlaid. 4. Continuously update the prioritization of vulnerable communes based on the nature of ongoing investments and interventions from partners. The approach presented in chapter 2 also streng- thens government institutions with a strategic regional vision, demonstrating practically how to integrate data from different sources at different levels (census data, field data, satellite data), with community involvement, for good evidence-based decision-making. Recommendation 3: Invest in Institutions and People Investments need maintenance. Resources are needed to guarantee minimal levels of institutional capacity to ensure water security in the region and systematically maintain water supply systems. This includes the systematic monitoring of water points and water resources, making decisions on 110 Water Security and Drought Resilience in the South of Angola FIGURE 5.1. Decision Process for the Prioritization and Selection of Beneficiary Communes for Water Resources Investments Government and project Community level confirmation inputs from field studies and involvement Water access: Regional • Complement, update and validate vulnerability assessment • Confirmation of need the prioritization of communes, • Census 2014 data at comuna integrating new information • Involvement in the selection of level covering all the region. (project studies, surveys). best solution • Water acces indicators used to • Institutional strengthening with • Organization and creation of calculate vulnerability ranking. regional strategic vision and an committee for managing use and evidence-based data approach maintenance to decision making • Entry point for Water, Sanitation and Hygiene (WASH) to reduce stunting New information Water point surveys (functionality, location, type/state of water point) Water resources investment Hydrogeological and hydrologic studies with community involvement and organization Other partner and government info interventions, planning and investments, and implementing them. It also includes the capacity to anticipate and plan ahead of droughts, floods, and other shocks. It is essential to build institutional preparedness and capacity to react before paying the high costs of devastated livelihoods, and lives lost and stunted. 1. Strengthen community-level organization through Water Point Committees or Water and Sanitation Groups (Grupos de Agua e Saneamento, GAS) and training in the protection and management of water points, operation and maintenance of water points, as well as water supply, sanitation, and hygiene (WASH), to protect water sources, water use, and contribute in the fight against stunting. The Modelo de Gestao Comunitaria da Agua (MoGeCA) works well at the community level and needs to be scaled up, but past pilots have failed due to the lack of an enabling environment and support at the municipal and provincial levels. 2. Strengthen regional water point governance and maintenance, focusing on the connectivity and capacity between the community and municipal, provincial government levels (Provincial water utilities, directorates of technical services and infrastructure, and municipal administrations) and national levels (National Water Directorate [DNA]), as well as other sector institutions. This will include mechanisms to ensure the availability of spare parts and repair technicians, considering local entrepreneurs, small-scale public-private partnerships, and nontraditional actors. Activities will include a study to check the required institutional arrangements, staffing and gaps, and deter- mine the most suitable interventions/approaches to support them, in parallel with technical assis- tance to Provincial Departments/EPAS. This provincial and municipal capacity is essential to enable the conditions for the MoGeCA model to be successful at the community level. Water Security and Drought Resilience in the South of Angola 111 3. Strengthen the provincial projection of Empresas Provinciais de Agua e Saneamento (EPASs) to be able to support rural areas and help mitigate drought pressures throughout the province. The EPASs are key resilience actors in the provinces and will help strengthen the capacity of provincial governments and municipalities to climate extremes (floods and droughts). This activity will be in the form of rural and drought resilience technical assistance to EPASs. 4. Strengthen the capacity of GABHIC (Gabinete para Administração das Bacias Hidrográficas do Cunene, Cubango e Cuvelai), DNA (Direccao Nacional das Aguas), and INRH (Instituto Nacional de Recursos Hidricos) to monitor, prepare, and respond to climate events (i.e., drought monitoring, preparedness, and flood and drought emergency and disaster response programs), in coordination with provincial and local actors. Technical assistance in looking at required institutional and inter-institutional arrangements, building on recent successful international experiences, would have several key objectives: a. Analyze staffing and gaps, determine the most suitable processes, reorganizations, and decision tools to strengthen preparedness, monitoring, and response capacity. b. Improve their capacity to gather and manage relevant climate, hydrology, hydrogeology, and water-related information, share it, make decisions based on it, and operationalize such decisions in order to maintain water security in the region. c. Review and strengthen information-to-action mechanisms for drought preparedness and drought response programs within relevant actors. This will include drought monitoring and drought preparedness systems and programs, as well as flood and drought emergency and disaster response programs. d. Develop drought preparedness and drought emergency water supply programs and associated training (including for provincial governments and municipalities to climate extremes). 5. Implement and operationalize a continuous monitoring approach and database system for wells and other water infrastructure performance using low-cost, easy, appropriate technologies. Consider the capital cost of the installation of sensors and tools, sampling costs, personnel and logistics, maintenance, analytical costs for laboratories, and data storage. Coordination is needed between key agencies to agree on data collection needs and procedures. How can data feed institu- tional systems of groundwater and water point information? Dedicated training of agencies is required. 6. Operationalize databases collecting data from the monitoring system and other information sour- ces: systematic entry of groundwater use, groundwater levels, measured in boreholes or wells: water levels, pumping rates, aquifer properties, estimated sustainable yields for boreholes or wells; measured groundwater quality, chemistry and microbiological parameters; and borehole drilling logs with geological information. Sharing and storage protocols need to be agreed between all involved agencies. 112 Water Security and Drought Resilience in the South of Angola 7. Develop financial resilience and efficient budgetary management at all levels to guarantee basic functionality of water points and other essential services, and cope with shocks. Focus on efficient budgeting of operational expenses to maintain functionality, and establish robust mechanisms to maximize results and capacity to operate during climate and other shocks. 8. Develop a strategic vision for future storage investments in the South but also nationwide, following a strong analytical approach. Support an integrative view for long-term strategic storage planning: coordinated use of dams and reservoirs, aquifers, and watershed storage. This vision also integra- tes no-regrets flood planning and flood control infrastructure investments, aiming at synergies between flood mitigation and storage options. The INRH can lead this work in the Namibe cos- tal basins and nationwide, and GABHIC can lead the efforts in the Cunene, Cuvelai and Cubango basins, in coordination with INRH. These strategic storage plans can also be part of Integrated River Basin Plans. 9. Develop integrated master plans for the Namibe Coastal basins and others, as well as supporting the updating of the Cunene River Basin Plan, to ensure a knowledge foundation for all actions going forward. Also, in the Benguela Southern basins. This study builds on past government investments such as the National Census of Angola (INE 2014), essential for the vulnerability analysis presented in chapter 2, highlighting the importance of having good data for well-informed analysis and evidence-based decisions. Specific to drought, it also builds on the Post-Disaster Needs Assessment (UNDP 2016), deepening the analysis with many additional sources of data and field observations, and increasing the resolution of analysis from the provincial to the com- mune level. This report is a stepping-stone and sets the stage for the next steps. Continuing policy dialogue and analytical work with the Government of Angola are essential going forward and will include (1) an options analysis for the systematic monitoring of water points and infor- mation mechanisms within government hierarchies (i.e., to feed the SISAS with periodic information and to use it to trigger repairs and interventions); (2) a deeper analysis of the linkages between commu- nity-based management and municipal and provincial governance, budgeting for maintenance and repairs, and the role of the private sector; (3) efforts to increase the availability and accessibility of hydrogeologic information across the region and finding mechanisms for systematic monitoring; (4) a scale up of the characterization of investments at the community level across the region; (5) a mapping of agriculture and deeper understanding of transhumance dynamics; and (6) ongoing training and capa- city building with government agencies, universities, and the private sector. Water Security and Drought Resilience in the South of Angola 113 © Aleix Serrat-Capdevila/World Bank Bibliography Andreu, J., A. Solera, P.-A. Javier, D. Haro-Monteagudo, and H. Van Lanen. 2015. Drought: Research and Science-Policy Interfacing. 1st ed. Leiden, Netherlands: CRC Press. Bachmair, S., K. Stahl, K. Collins, J. Hannaford, M. Acreman, M. Svoboda, C. Knutson, K. H. Smith, N. Wall, B. Fuchs, N. D. Crossman, and I. C. Overton. 2016. “Drought Indicators Revisited: The Need for a Wider Consideration of Environment and Society.” Wires Water 3 (4): 516–36. Bates, R. L., and J. A. Jackson. 1987. Glossary of Geology. 3rd ed. Alexandria, VA: American Geological Institute. BIWAC (Bittner Water Consult CC). 2006. Desk Study Report: Cuvelai-Etosha Groundwater Investigation. Windhoek: BIWAC. Bonaccorso, B., I. Bordi, A. Cancelliere, G. Rossi, and A. Sutera. 2003. “Spatial Variability of Drought: An Analysis of the SPI in Sicily.” Water Resources Management 17 (4): 273–96. Brant, S. 2007. “Assessing the Vulnerability to Drought in Ceará, Northeast Brazil.” Master’s Thesis, University of Michigan, Ann Arbor, MI. Brooks, N. 2003. “Vulnerability, Risk and Adaptation: A Conceptual Framework.” Tyndall Centre for Climate Change Research Working Paper 38, School of Environmental Sciences, University of East Anglia, Norwich. Cardona, O. D., M. K. Van Aalst, J. Birkmann, M. Fordham, G. McGregor, R. Perez, R. S. Pulwarty, E. L. F. Schipper, and B. T. Sinh. 2012. “Determinants of Risk: Exposure and Vulnerability.” In Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation, edited by C. B. Field et al., 65–108. A Special Report of Working Groups I and II of the Intergovernmental Panel on Climate Change (IPCC). Cambridge, UK: Cambridge University Press. Cohen-Shacham, E., G. Walters, C. Janzen, and S. Maginnis. 2016. Nature-Based Solutions to Address Global Societal Challenges. p. 97, Gland, Switzerland: IUCN. Dabanli, I. 2018a. “Drought Hazard, Vulnerability, and Risk Assessment in Turkey.” Arabian Journal of Geosciences 11 (538): 1. Dabanli, I. 2018b. “Drought Risk Assessment by Using Drought Hazard and Vulnerability Indexes.” Natural Hazards and Earth System Sciences. https://www.nat-hazards-earth-syst-sci-discuss.net/nhess-2018-129/. Dill, H. G., R. Dohrmann, J. Fenner, U. Gersdorf, S. Kaufhold, R. Kringel, F. Lindenmaier, R.-R. Ludwig, R. McG. Miller, A. Nick, U. Noell, and A. Walzer. 2012. Development and Dating of Duricrusts as Base for a Paleogeographic and Hydrogeological Model of the Cuvelai-Etosha Basin in North Namibia. Technical report prepared for the Kalahari Research Project. Hanover, Germany: Federal Institute for Geosciences and Natural Resources (BGR). Döll, P., and M. Flörke. 2005. “Global Scale Estimation of Diffuse Groundwater Recharge.” Frankfurt Hydrology Paper 03, Institute of Physical Geography, Frankfurt University, Frankfurt am Main. Dulal, H. B., G. Brodnig, C. G. Onoriose, and H. K. Thakur. 2010. “Capitalizing on Assets: Vulnerability and Adaptation to Climate Change in Nepal.” Social Development Working Papers 121, World Bank, Washington, DC. Erfurt, M., R. Glaser, and V. Blauhut. 2019. “Changing Impacts and Societal Responses to Drought in Southwestern Germany since 1800.” Regional Environmental Change 19: 2311–23. doi: 10.1007/s10113-019-01522-7 FAO (Food and Agriculture Organization). N.d. “FAO Term Portal,” (accessed December 2017), http://www.fao.org/faoterm/en. Guerschman, J.-P., A. I. J. M. Van Dijk, T. R. McVicar, T. G. Van Niel, L. Li, Y. Liu, and J. Peña-Arancibia. 2008. Water Balance Estimates from Satellite Observations over the Murray-Darling Basin. A report to the Australian Government from the CSIRO Murray-Darling Basin Sustainable Yields Project. Australia: CSIRO. GWP (Global Water Partnership). N.d. “Glossary,” (accessed December 2017), http://www.gwp.org/en/Website-Information/Glossary. GWP CEE (Global Water Partnership Central and Eastern Europe). 2015. Guidelines for Preparation of the Drought Management Plans: Development and Implementation in the Context of the EU Water Framework Directive. Bratislava: Global Water Partnership Central and Eastern Europe. Hahn, C. 2019. “Seasonal Effects of Drought on the Productivity and Fodder Quality of Temperate Grassland Species.” Doctoral Thesis, University of Basel, Faculty of Science. Water Security and Drought Resilience in the South of Angola 115 Hamutoko, J. T., H. Wanke, P. Königer, M. Beyer, and M. Gaj. 2017. “Hydrogeochemical and Isotope Study of Perched Aquifers in the Cuvelai- Etosha Basin, Namibia.” Isotopes in Environmental and Health Studies 53 (4): 382–99. Hunter, B., M. Gray, and B. Edwards. 2013. “The Use of Social Surveys to Measure Drought and the Impact of Drought.” Social Indicators Research 113 (1): 419–32. Huntley, B. J. 2019. “Angola in Outline: Physiography, Climate and Patterns of Biodiversity.” In Biodiversity of Angola, edited by B. Huntley, V. Russo, F. Lages, and N. Ferrand, 15–42. Cham, Switzerland: Springer. INE (Instituto Nacional de Estadistica). 2016. Recenseamento Geral da População e da Habitaçao de Angola. Resultados Definitivos do Censo 2014. Angola. Lasage, R., J. Aerts, G.-C. Mutiso, and A. de Vries. 2008. “Potential for Community Based Adaptation to Droughts: Sand Dams in Kitui, Kenya.” Physics and Chemistry of the Earth, Parts A/B/C 33 (1–2): 67–73. Limones, N., J. Marzo, M. Wijnen, and A. Serrat-Capdevila. 2020. “Evaluating Drought Risk in Data-Scarce Contexts: The Case of Southern Angola” Journal of Water and Climate Change 1 December 2020; 11 (S1): 44–67. https://doi.org/10.2166/wcc.2020.101. Lombana Cordoba, Camilo; Andres, Luis A.; Da Costa, Lucrecio A.M.; Fenwick, Crystal. 2021. “Diagnosing Angola’s WASH Sector: An Urgent Call to Action.” World Bank, Washington, DC. https://openknowledge.worldbank.org/handle/10986/35591. Luetkemeier, R., and S. Liehr. 2018. “Household Drought Risk Index (HDRI): Social-Ecological Assessment of Drought Risk in the Cuvelai- Basin.” Journal of Natural Resources and Development 8: 46–68. McKee, T. B., N. J. Doesken, and J. Kleist. 1995. “Drought Monitoring with Multiple Time Scales.” In Preprints, Ninth Conference on Applied Climatology. Dallas, TX: American Meteorological Society. pp. 233–6. Mendelsohn, J., and S. Mendelsohn. 2019. Sudoeste de Angola: um Retrato da Terra e da Vida/South West Angola: A Portrait of Land and Life. Namibia: Arte e Ciência. Miralles, D. G., T. R. H. Holmes, R. A. M. de Jeu, J. H. Gash, A. G. C. A. Meesters, and A. J. Dolman. 2011. “Global Land-Surface Evaporation Estimated from Satellite-Based Observations.” Hydrology and Earth System Sciences 15: 453–69. National Drought Mitigation Center. 2019. “Drought Impact Reporter Help.” http://droughtreporter.unl.edu/help/. NOAA NWS (National Oceanic and Atmospheric Administration’s National Weather Service). n.d. “National Weather Service Glossary,” (acces- sed December 2017), http://w1.weather.gov/glossary. Pita, M. F. 2000. “Un nouvel indice de sécheresse pour les domaines méditerranéens. Application au bassin du Gaudalquivir (sudouest de l’Espagne).” Publications de l’Association Internationale de Climatologie 13: 225–34. Pombo, S., R. P. de Oliveira, and A. Mendes. 2015. “Validation of Remote Sensing Precipitation Products for Angola.” Meteorological Applications 22 (3): 395–409. Rajsekhar, D., V. P. Singh, and A. K. Mishra. 2015. “Integrated Drought Causality, Hazard, and Vulnerability Assessment for Future Socioeconomic Scenarios: An Information Theory Perspective.” Journal of Geophysical Research: Atmospheres 120 (13): 6346–78. Scoones, I. 1998. “Sustainable Rural Livelihoods: A Framework for Analysis.” IDS Working Paper 72, Institute of Development Studies, Brighton. Seguin, J.-J. 2008. “Hydrogeological Map of Africa at 1/10 M Scale.” Bureau de Recherches Géologiques et Minières (French Geological Survey) and UNESCO, Orléans, France. www.brgm.fr. Senay, G. B., S. Bohms, R. K. Singh, P. H. Gowda, N. M. Velpuri, H. Alemu, and J. P. Verdin. 2013. “Operational Evapotranspiration Mapping Using Remote Sensing and Weather Datasets: A New Parameterization for the SSEB Approach.” Journal of the American Water Resources Association 49 (3): 1–15. Shahid, S., and H. Behrawan. 2008. “Drought Risk Assessment in the Western Part of Bangladesh.” Natural Hazards 46 (3): 391–413. UNDP (United Nations Development Programme). 2016. Droughts in Angola 2012–2016: Post Disaster Needs Assessment Report. Luanda: UNDP. UNDRR (United Nations Office for Disaster Risk Reduction). N.d. “Terminology,” (accessed December 2017), https://www.unisdr.org/we​ /­inform/terminology. UNFCCC (United Nations Framework Convention on Climate Change). N.d. “Glossary of Climate Change Acronyms and Terms,” (accessed December 2018), https://unfccc.int/process-and-meetings/the-convention/glossary-of-climate-change-acronyms-and-terms. 116 Water Security and Drought Resilience in the South of Angola Urquijo, J., I. Gonzalez Tánago, M. Ballesteros, and L. De Stefano. 2015. “Assessing Vulnerability to Drought: Identifying Underlying Factors across Europe.” Proceedings of the EGU General Assembly 2015, held April 12–17, 2015, European Geosciences Union, Vienna, Austria. Van Loon, A., T. Gleeson, J. Clark, A. van Dijk, K. Stahl, J. Hannaford, G. Di Baldassarre, A. Teuling, L. M. Tallaksen, R. Uijlenhoet, D. Hannah, J. Sheffield, M. Svoboda, B. Verbeiren, T. Wagener, S. Rangecroft, N. Wanders, and H. Van Lanen. 2016. “Drought in a Human-Modified World: Reframing Drought Definitions, Understanding and Analysis Approaches.” Hydrology and Earth System Sciences Discussions 20 (9): 1–34. doi: 10.5194/hess-2016-251. Vicente-Serrano, S. M., S. Santiago Beguería, and J. López-Moreno. 2010. “A Multi-Scalar Drought Index Sensitive to Global Warming: The Standardized Precipitation Evapotranspiration Index.” Journal of Climate 23: 1696–718. Villholth, K. G., C. Tottrup, M. Stendel, and A. Maherry. 2013. “Integrated Mapping of Groundwater Drought Risk in the Southern African Development Community (SADC) Region.” Hydrogeology Journal 21 (4): 863–85. Weiss, M. 2009. Modelling of Global Change Impacts on Hydrology with Focus on Europe and Africa. Kassel: Kassel University Press GmbH. WHO (World Health Organization). 2017. Guidelines for Drinking-Water Quality. 4th ed. Geneva: WHO. WMO (World Meteorological Organization). 1974. International Glossary of Hydrology. Geneva: WMO. WMO (World Meteorological Organization). 1992. International Meteorological Vocabulary (WMO-No. 182). Geneva: WMO. WMO and GWP (Global Water Partnership). 2016. Handbook of Drought Indicators and Indices (WMO-No. 1173). Integrated Drought Management Tools and Guidelines Series 2, Integrated Drought Management Programme (IDMP). Geneva: WMO and GWP. WMO and UNESCO (United Nations Educational, Scientific, and Cultural Organization). 2012. International Glossary of Hydrology (WMO–No. 385). Geneva: WMO. Water Security and Drought Resilience in the South of Angola 117 © Aleix Serrat-Capdevila/World Bank Glossary Aquifer: Underground layers of permeable rock, sediment, or soil filled with water and interconnected, so the water stays within or flows through them (GWP n.d). Basin (or catchment): An area having a common outlet for its surface runoff (WMO and UNESCO 2012). Borehole: A well that is excavated by means of a hand or power auger (WMO 1974). Cacimba (well): Shaft or hole that is sunk, dug, or drilled into the earth to extract water (WMO 1974). Chimpaca: A shallow dam excavated in an ephemeral water course (Mendelsohn et al. 2019). Coping capacity: The ability of people, organizations, and systems, using available skills and resources, to manage adverse conditions, risk, or disasters. The capacity to cope requires ongoing awareness, resources, and good management, both in normal times as well as during disasters or adverse condi- tions. Coping capacities contribute to the reduction of disaster risks (United Nations General Assembly). Desalination: Removal of salt from sea or brackish water. It is achieved by various methods, for example distillation, reverse osmosis, hyperfiltration, electrodialysis, ion exchange, and solar evaporation followed by condensation of water vapor (FAO n.d.) Drought: Period of abnormally dry weather sufficiently prolonged for the lack of precipitation to cause a serious hydrological imbalance (WMO 1992). Drought impact: A specific effect of drought on the economy, society, and/or environment, which is a symptom of vulnerability (GWP CEE 2015). Drought index: Computed numerical representations of drought severity, assessed using climatic or hydrometeorological inputs, including precipitation, temperature, streamflow, groundwater and reser- voir levels, soil moisture, and snowpack. They aim to measure the qualitative status of drought on the landscape for a given time period. Indices are technically indicators as well (WMO and GWP 2016). Drought vulnerability assessment: It is a drought vulnerability quantification and description that consist in identifying the relevant factors influencing it, from the point of view of exposure, sensitivity, and adaptive capacity. The final aim of a drought vulnerability assessment is to identify the underlying ­ sources of drought impact (Urquijo et al. 2015). Evapotranspiration: The combined processes by which water is transferred from the Earth’s surface to the atmosphere by evaporation from the land and ocean surfaces and by transpiration from vegetation (WMO 1992). Exposure: The situation of people, infrastructure, housing, production capacities, and other tangible human assets located in hazard-prone areas (UNDRR n.d.). Water Security and Drought Resilience in the South of Angola 119 Green or natural infrastructure: That which intentionally and strategically preserves, enhances, or resto- res elements of a natural system and can be combined with gray infrastructure to produce more resilient and lower-cost services. For example, watershed conservation and managed aquifer recharge serve flood control and drought management purposes (Cohen-Shacham et al. 2016). Groundwater: Water within the earth that supplies wells and springs; water in the zone of saturation where all openings in rocks and soil are filled, the upper surface of which forms the water table (NOAA NWS n.d.). Hazard: A process, phenomenon, or human activity that may cause loss of life, injury, or other health impacts, property damage, social and economic disruption, or environmental degradation. Hazards may be natural, anthropogenic, or socio-natural in origin. Natural hazards are predominantly associated with natural processes and phenomena. Anthropogenic hazards, or human-induced hazards, are indu- ced entirely or predominantly by human activities and choices. Hazards may be single, sequential, or combined in their origin and effects. Each hazard is characterized by its location, intensity or magni- tude, frequency and probability (UNDRR n.d). Impluvium: Drainage basin. Nature-based solutions: Actions to protect, sustainably manage, and restore natural or modified ecosys- tems that address societal challenges effectively and adaptively, simultaneously providing human well-being and biodiversity benefits (Cohen-Shacham et al. 2016). No-regrets investments: In this report, we refer to those investments that will be beneficial regardless of what scenario or set of events may become a reality in the future, and that do not commit future deve- lopment to a single path. For example, green infrastructure for watershed conservation and managed aquifer recharge is beneficial both in times of drought and flooding. Resilience: The ability of a system, community, or society exposed to hazards to resist, absorb, accom- modate, adapt to, transform, and recover from the effects of a hazard in a timely and efficient manner, including through the preservation and restoration of its essential basic structures and functions through risk management (UNDRR n.d.). Risk: The potential loss of life, injury, or destroyed or damaged assets that could occur to a system, society, or a community in a specific period of time, determined probabilistically as a function of hazard, exposure, vulnerability, and capacity (UNDRR n.d.). Runoff: That part of the precipitation that flows toward a river on the ground surface (surface runoff) or within the soil (subsurface runoff or interflow) (WMO and UNESCO 2012). Sand dam: A small dam built on and into the riverbed of a seasonal sand river. Its purpose is to capture and store water beneath sand: sand is transported during periods of high flow and accumulates ups- tream of a sand dam, resulting in additional groundwater storage capacity of riverbed and banks. This reservoir fills during the wet season, providing water to cover dry periods (Lasage et al. 2008). 120 Water Security and Drought Resilience in the South of Angola Vulnerability: The degree to which a system is susceptible to, or unable to cope with, adverse effects of climate change, including climate variability and extremes. Vulnerability is a function of the character, magnitude, and rate of climate variation to which a system is exposed, its sensitivity, and its adaptive capacity (UNFCCC n.d.). Wadi: A term used in the desert regions of Southwestern Asia and Northern Africa for a stream bed or channel, or a steep-sided and bouldery ravine, gully, or valley, or a dry wash, that is usually dry except during the rainy season, and that often forms an oasis (Bates and Jackson 1987). Water governance: The political, administrative, economic, and social systems that exist to manage water resources and services and are essential in order to manage water resources sustainably and pro- vide access to water services for domestic or productive purposes (GWP n.d.). Water security: The availability of an acceptable quantity and quality of water for health, livelihoods, ecosystems, and production, coupled with an acceptable level of water-related risks to people, environ- ments, and economies (GWP n.d). Waypoint: A point of reference that can be used for location. Waypoints can be the specific latitude and longitude of a location, a well-known building or a natural feature. Water Security and Drought Resilience in the South of Angola 121 © Aleix Serrat-Capdevila/World Bank Appendix A Main Aquifers in the Cuvelai-Etosha Basin TABLE A .1. Classification and Main Characteristics of Aquifers of the Kalahari Sequence in the Namibian Part of the Cuvelai-Etosha Basin Former Group Name of Aquifer/Aquitard New Abbreviation Abbreviation Sequence (Subgroup) Formation Kalahari Sequence Aquifer K n.a. Ombalantu., Beiseb, (undifferentiated) Olukonda, Andoni, Etosha Limestone M., Recent Discontinuous Perched KDP DPA Recent Aquifer Etosha Limestone Aquifer KEL UKAEL Andoni (Etosha Limestone Member) Oshivelo multi-layered KOV n.a. Ombalantu, Beiseb, Aquifer (Undifferentiated) Olukonda, Andoni • Aquifer 1 KOV1 UKAAN Andoni Kalahari • Aquifer 2 KOV2 OAAAN Andoni, olukonda Oshana Multi-layered Aquifer KOS N/A Ombalantu, Beiseb, (undifferentiated) Olukonda, Andoni • Aquifer 1 KOS1 MSAAN Andoni Ohangwena Multi-layered KOH n.a. Andoni, Olukonda Aquifer (undifferentiated) • Aquifer 1 (Andoni Fm) KOH1 MDAAN Andoni, Olukonda • Aquifer 2 (Olukonda Fm) KOH2 VDAOL Olukonda Omusati Multi-zoned Aquifer KOM n.a. Ombalantu, Beiseb, (undifferentiated) Olukonda, Andoni Water Security and Drought Resilience in the South of Angola 123 Water Borehole Aquifer Aquifer Strike Yield Transmissivity Aquifer Occurence Characteristics Litholoqy [m b.g.l.] Water Quality (m3/H) [T m2/Day] KEL Foreland of Mainly fractured, Dolocrete, I0–100 Fresh, locally 3–100 ? dolomite mountain often karstified, calcrete/ high nitrate land, Tsumeb and locally porous, limestone, concentrations Olushandja sub unconfined. sand basins KOV2 Between Mainly porous, Conglomerate. 30–150 Fresh to 25-200? 100–10,000 Omiramba locally fractured Sand, sand- brackish Owambo and and artesian, stone, Omuthiya, Tsumeb confined dolocrete, sub-basins calcrete KOS Cuvelal drainage Mainly porous, Sand, calcrete/ 6–80 Saline to 1–30 n.a. locally fractured; limestone hypersaline unconfined KOH1 Eastern Mainly porous, Sand, 60–180 Fresh to 3–50 30–760 Ohangwena and locally fractured, sandstone brackish Oshikoto regions, confined Nipele Sub-basin Source: BIWAC 2006. Note: n.a. = not applicable. 124 Water Security and Drought Resilience in the South of Angola Appendix B Water Balances for Below-Average Rainfall Years TABLE B.1. Tentative Water Balances for the Three Cuvelai Subbasins Drawn with P, ETa, “Wet Season P-ETa Excess,” and “Dry Season P-ETa Déficit” Values at 60 Percent, 70 Percent, 80 Percent, and 90 Percent Annual Exceedance Probabilities 60% Annual Exceedence Probability 70% Annual Exceedence Probability % of Namibia % of Upper Namibia avg. Upper Cuvelai Angola lowland lowland avg. Cuvelai  Angola lowland lowland 13,122 10,206 12,393 13,122 10,206 12,393 95% 80% 80% 95% 80% 80% 650.0 490.0 375.0 610.0 455.0 345.0 670.0 535.0 405.0 620.0 520.0 385.0 205.0 150.0 115.0 170.0 120.0 90.0 −255.0 –220.0 –180.0 −265.0 −228.0 −190.0 124.4% 146.7% 156.5% 155.9% 190.0% 211.1% 80% 28.8% 35.2% 32.0% 70% 25.2% 30.8% 28.0% 187 r 172 W 120 154 140 97 70% 1.8% 0.7% 3.5% 60% 1.5% 0.6% 3.0% 11.4 3.4 13.1 9.2 2.7 10.4 176 169 107 145 137 86 5% 20% 20% 5% 20% 20% 27 12 19 24 8 13 356 123 235 314 81 158 70% 53% 11% 5% 60% 45% 9% 4% 187 33 13 142 20 7 169 277 255 173 202 170 656 2,041 2,479 656 2,041 2,479 258 136 103 264 99 69 180 162 106 151 130 83 0 0 0 0 0 0 11 3 11 9 2 8 14 3 1 11 2 1 0 0 0 0 0 0 −45 −51 −42 −29 −69 −49 −75 −58 −74 −114 −98 −107 Water Security and Drought Resilience in the South of Angola 125 80% Annual Exceedence Probability 90% Annual Exceedence Probability % of Namibia % of Angola Namibia avg. Upper Cuvelai Angola lowland lowland avg. Upper Cuvelai lowland lowland 13,122 10,206 12,393 13,122 10,206 12,393 95% 80% 80% 95% 80% 80% 565.0 420.0 310.0 530.0 390.0 290.0 585.0 500.0 370.0 550.0 490.0 360.0 140.0 95.0 65.0 115.0 70.0 45.0 −270.0 −236.0 −200.0 −275.0 −243.0 −207.0 192.9% 248.4% 307.7% 239.1% 347.1% 460.0% 60% 21.6% 26.4% 24.0% 50% 18.0% 22.0% 20.0% 122 111 74 95 86 58 50% 1.3% 0.5% 2.5% 40% 1.0% 0.4% 2.0% 7.1 2.1 7.8 5.3 1.6 5.8 115 109 67 90 84 52 5% 20% 20% 5% 20% 20% 24 6 5 24 1 −1 316 64 68 320 14 −17 38% 8% 4% 40% 30% 6% 3% 118 14 3 96 7 0 197 169 79 224 103 −10 656 2,041 2,479 656 2,041 2,479 301 83 32 341 51 −4 124 104 60 103 78 41 0 0 0 0 0 0 7 2 6 5 1 5 9 1 0 7 1 0 0 0 0 0 0 0 −36 83 −66 −32 102 −75 −146 −132 −140 172 −165 −166 126 Water Security and Drought Resilience in the South of Angola Appendix C Water Harvesting Infrastructure Designs FIGURE C .1. Schematic Cross-Section of a Chimpaca Chimpaca Chimpaca embankment embankment with inlet Stormwater inflow H max Channel infill Island composed Storage of sandy clay of aeolian sand Limestone substratum FIGURE C .2. Schemtic layout of a chimpaca Clay embankment H=2m Stormwater with inlet inflow Slope 1:5 Channel infill of sandy clay Island composed of aeolian sand Excavated storage: 80 × 60 × 4m Clay embankment H=2m Water Security and Drought Resilience in the South of Angola 127 FIGURE C .3. Schematic cross-section of a cistern Tap with solar Stormwater inflow Cistern with pump and deviated from concrete slab reservoir main channel Rock and gravel filter H max Channel infill Island composed of sandy clay of aeolian sand Limestone substratum FIGURE C .4. Schematic layout of a cistern Dug deviation channel Rock and grave Island composed filter with sediment of aeolian sand trap Flow direction Excavated cistern with concrete slab: 20 × 6 × 8 m Main channel Channel infill of sandy clay 128 Water Security and Drought Resilience in the South of Angola Appendix D Topographic and Geognostic Surveys Dimensions and Methodology To fine-tune the sketched dam dimensions, the riverbed slopes, and the banks morphology a topogra- phic survey shall be carried out in a stream section of 400 m upstream and 200 m downstream of the selected sites. The surveys can be done by drone or precision GPS or by theodolite. In the last two cases the following elements are required: (1) the execution of three continuous profile 600 m long, along the stream axis and the edges of the riverbed; (2) eight profiles perpendicular to the axis profiles, each 200–300 m long,1 that shall be located 25, 50, and 100 m downstream, on the dam axis and 25, 50, 100, and 200 m upstream; and (3) a final map with contour lines 0.5–1.0 m spaced shall be drawn, on which each significant morphological feature (such as terraces edge, drainages, rocky outcrops) and the sites where test pits or other tests are positioned shall be indicated. If the survey is done by drone the map and the 10 sections obtained by the grid, they shall be presented as described above. Measurements every 5 m along the dam axis, each 10 m in the range of 100 up and downstream and 20–30 m in the remaining area, are required. The profiles perpendicular to the stream axis must extend not less than 50 m from the course’s edge. In case the survey is done by theodolite at least 10 points scattered in the whole area shall be georeferenced. The geognostic survey comprises the following steps: •• Geological survey: All the rocky outcrops in the area described above shall be mapped and described with regard to the shape (spheroids, flat or inclined plains, ridges channels), the presence of fractu- res, and the degree of compactness. It shall be indicated if the granite is fresh, slightly weathered, or deeply weathered. The presence of blocks lenses on the riverbed or on the flanks shall be signa- led too, because this factor is important for the materials necessary for the dam construction. •• Observation pits: The pits shall be excavated up to the basement top and the material found at various levels described with regard to the composition (coarse, medium of fine sand, gravel, peb- bles, blocks, or boulder) and the possible water presence. For the coarse sediments (such as pebbles or boulders) a range of dimensions shall be indicated. A profile along the dam axis with pits 5 m spaced, a longitudinal profile (along the stream axis) with pits at 10 m downstream, and 10, 20, 50, 100, and 200 m upstream is required. In case there is an outcrop in the position suggested, the pit is shifted up or downstream from the location 10 m far from the outcrop. •• Infiltration tests2: Two infiltration test are done in one of the excavated pits, possibly in a position near the stream axis. One test shall be done on the surface, before the pit excavation, and another at 0.5 m of depth. Water Security and Drought Resilience in the South of Angola 129 •• Porosity essays3: Four porosity tests are carried out on samples taken in the riverbed in two different sites: for each side a sample shall be taken on the stream surface and another at 1 m of depth (exis- ting pits can be utilized). It is suggested to take a sample near the dam axis and another some 50–100 m upstream. •• Report: The final report shall describe all the activities carried out, a geological report with maps, at two cross-sections (one at the dam axis and another perpendicular) with pits logs, the description of the infiltration tests with the related graphs, and the results of the porosity essays. Site Adjustments In the site GB2 the topographic survey and the pits excavation shall be extended to both courses’ bran- ches and to the island dividing them (as shown in photo D.1). PHOTO D.1. Site GB2 Survey Area GB2 30 0m m 400 130 Water Security and Drought Resilience in the South of Angola In the site GB3 additional topographic profiles, perpendicular to the stream axis, shall be generated inside the stream bed each 5 m up to 20 m downstream and 40 m upstream (photo D.2). The reason is the particularly rough morphology, due to the rocky outcrop, of the stream bed. The pits cannot be excavated along the dam layout, but six locations shall be chosen, three on each bank along the dam alignment. In the site CHB1 Photo D.3 particular attention is required in the area downstream of the old dam, to select the best-fitting location for the new dam. The area 10 m upstream and 40 m downstream of the old dam must be surveyed with profiles 5 m spaced. The limits of the outcrop must be mapped with great precision. The excavation of the pits shall be concentrated in this area to find the basement between and beyond the two outcrops on the opposite flanks. The pits shall be aligned along a profile 120 m long, 10 m shifted downstream of the old dam, with pits 5 m spaced where loose deposits are present. PHOTO D.2. Site GB3 Survey Area 0m GB3 50 30 0m Note: Site GB3 survey area—green area: survey area—with dashed rectangle: area of detailed topographic survey. Water Security and Drought Resilience in the South of Angola 131 PHOTO D.3. Site CHB1 Survey Area CHB1 m 0 25 60 0 m Notes 1. See photo D.1 and figure D.2 for the maximum length of the profiles. 2. This type of essay is finalized to ascertain the possible recharge of the aquifer. 3. The porosity tests are important for a better estimate of the future volume of the stored water. 132 Water Security and Drought Resilience in the South of Angola Appendix E Mission Schedule Mission 1 (Huila, Namibe, Cunene, Stage 2, April 2019): Between April 15 and 23, 2019, different meetings with several partners took place in Luanda, the capital of Angola, and Lubango, the capital of the pro- vince of Huíla. These meetings provided key information both in terms of the drought context and in regards to the preparation of the field trip, which covered the period April 24 to May 2. The weeks of work were distributed as follows: •• April 15: Internal World Bank Team Meeting. •• April 16: Instituto De Recursos Hídricos (INRH) with director Manuel Quintino and hydrologist Narciso Ambrósio. •• April 17: United Nations Development Programme (UNDP) with Mr. Goetz Schroth. •• April 18: Instituto Nacional De Estadística (INE). •• April 19: Civil Protection Meeting with Segundo Comandante José Horacio da Silva and Director Segundo Comandante Edson Fernando. •• April 20: United Nations Children’s Fund (UNICEF) with Tomás López de Bufala, Chief of Water, Sanitation and Hygiene of UNICEF in Angola. •• April 21: Fundo De Apoio Social (FAS) with Santinho Figueira, General Director of Fundo de Apoio Social (FAS). •• April 21: Development Workshop with Mr. Allan Cain. •• April 23: Fortalecimiento De La Resistencia Y De La Seguridad Alimenticia Y Nutricional (FRESAN) with Dr. Mateo Tonini, in Lubango. •• April 24: Field visit to Chiange (Huíla). •• April 25: Field visit to Giraul Basin (Namibe). •• April 26: Field visit to Camacuio: Chingo and Mamúe (Namibe). •• April 27: Field visit to Lucira (Namibe). •• April 29: Field visit to Ombala-Yo-Mungu (Cunene). •• April 30 and May 1: Field visit to Oncocua (Cunene). Water Security and Drought Resilience in the South of Angola 133 The information gathered during this mission is the main basis for Chapter 3 and informed the overall structure of this work and the design of follow-up missions. Mission 2 (Cunene, Stage 3, July 2019): Between July 3 and 20, 2019, starting with a workshop in the World Bank offices with development partners and government agencies, to present work to date (from Stages 1 and 2), and then a field mission in Cunene province took place from the 8th to the 19th focusing on hydrogeology and chimpaca hydrology in the Cuvelai basin. Meetings with several partners and local administrators, including GABHIC and the Development Workshop, took place in Ondjiva (the capital of Cunene), as well as in many municipalities. The information gathered during this mission is the main basis for Case Study 1 in chapter 4. Mission 3 (Namibe, Stage 3, August 2019): Between August 26 to September 6, 2019, a field mission in Namibe province took place focusing on identifying and characterizing sand dam sites in the central Girault and the Chingo areas of Namibe. Support from FAS-Namibe and meetings with provincial and municipal administrators were key to the success of the mission. The information gathered during this mission is the main basis for Case Study 2 in chapter 4. 134 Water Security and Drought Resilience in the South of Angola Water Security and Drought Resilience in the South of Angola SKU W20101