1 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS 2021 The 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 volume is a product of the staff of The World Bank. The findings, interpretations, and conclusions expressed in this volume do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development / The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete infor- mation to the Copyright Clearance Center Inc., 222 Rosewood Drive, Danvers, MA 01923, USA: telephone: 978-750-8400; fax: 978-750-4470; Internet: www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA: fax: 202-522-2422; e-mail: pubrights@worldbank.org. To order additional copies of this publication, please send an e-mail to the Transport Help Desk transport@worldbank.org Transport publications are available on-line at http://www.worldbank.org/transport/ Contents Foreword....................................................................................................................7 Team & Acknowledgements.............................................................................8 Abbreviations and Acronyms...........................................................................9 Glossary of Technical Terms...........................................................................10 Structure of the Report.................................................................................... 12 Objectives of the Report..................................................................................13 Executive Summary............................................................................................15 1.  Introduction................................................................................................ 22 1.1 The Importance of Rural Roads............................................ 23 1.2 The current situation in Africa.............................................. 27 2.  The Traditional Approach to the Paving Decision....................... 34 2.2 The Traditional Cost-Benefit Analysis Approach................ 35 2.1 Reasons for Paving.................................................................. 35 3.  Challenges with the Traditional Approach ................................... 43 3.1 Main Critiques of the Traditional Approach...................... 44 3.2 Key Factors for a Systematic Paving Decision.................. 44 4.  Other Approaches in the Literature....................................................57 4.1 Cost Effectiveness Analysis (CEA)......................................... 58 4.2 Multicriteria Analysis Approach (MCA)................................ 58 4.3 Summary of Approaches....................................................... 64 5.  Decision Framework: To Pave or Not To Pave?............................... 65 5.1 Critical Questions affecting the Paving Decisions...........66 5.2 Presenting the Spade Model................................................ 67 5.2 The Spade-Plus Approach..................................................... 71 5.2 Spade Model Case Study Results......................................... 74 6.  Choice of Road Surfacing Options...................................................... 90 6.1 General Considerations.......................................................... 91 6.1 Menu of Road-Surfacing Options........................................ 92 7.  Conclusions and Recommendations................................................ 115 Bibliography......................................................................................................... 119 Annexes..................................................................................................................125 Annex A: Why Invest in Rural Road Infrastructure?.................126 Annex B: Additional Case Studies................................................127 Annex C: Road Safety Guidance...................................................132 Photo credits: Scott Wallace 4 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE FIGURES Figure 1.1: Theory of Change in Road Investment Options............................................................................................ 23 Figure 1.2: Rural Accessibility Index and Poverty Headcount: National Levels........................................................... 25 Figure 1.3: Rural Accessibility in Africa.................................................................................................................................. 27 Figure 1.4: Rural Accessibility Index, 2006...........................................................................................................................28 Figure 1.5: Rural Accessibility Index (RAI) and Poverty Headcount: Subnational Levels.......................................... 29 Figure 1.6: Percentage of Paved Roads in Africa (by country)........................................................................................ 29 Figure 1.7: Percentage of Paved Roads by Region............................................................................................................ 30 Figure 1.8: Construction/Upgrading of Paved Road Projects in Sub-Saharan Africa................................................ 31 Figure 1.9: Madagascar: Pupil Repetition Rates by Road Connectivity......................................................................... 32 Figure 1.10: Madagascar: Medicine Stock Rates by Road Connectivity........................................................................ 32 Figure 2.1: Impact of Paving on Road Roughness Over Time......................................................................................... 38 Figure 2.2: Total Project Costs vs. Initial AADT.................................................................................................................... 40 Figure 2.3: Threshold of Traffic for Paving and Investment Costs................................................................................. 41 Figure 2.4: Threshold of Traffic for Paving and Investment Costs with 24-Month Grading.................................... 42 Figure 3.1: Seven Key Considerations for a Systematic Paving Approach................................................................... 45 Figure 3.2: Fuel Levy on Petrol (in US$/liter)........................................................................................................................ 46 Figure 3.3: Road Fund Resources as a Share of GDP (%)................................................................................................. 46 Figure 3.4: Road Fund Resources per Length of Road..................................................................................................... 47 Figure 3.5: Links between Climate Co-Benefits (CCBs) and Risk Screening................................................................ 52 Figure 3.6: Resilience Attributes from the World Bank Sectoral Checklist................................................................... 53 Figure 3.7: Criticality and NPV of Rural Road Investment in Mozambique.................................................................. 54 Figure 3.8: Relationship of Vehicle Speed and Dust Formation..................................................................................... 56 Figure 4.1: Three-Step Paving Framework...........................................................................................................................62 Figure 4.2: Screenshot of Kansas Paving Program............................................................................................................ 63 Figure 4.3: Three Approaches to the Economic Evaluation Problem of Paving......................................................... 64 Figure 5.1: Factors Considered in the SPADE..................................................................................................................... 69 Figure 5.2: SPADE-PLUS Approach Framework.................................................................................................................. 73 Figure 5.3: Distribution of SPADE Model Recommendations from 54 Case Studies................................................ 77 Figure 5.4: Comparison of SPADE and RED Model Paving Positivity Results*............................................................ 77 Figure 5.5: FR1 in Karongi District Before and After Gravel Road Improvement Works.......................................... 83 Figure 5.6: Landslide-Affected Sections of FR1 in Karongi District................................................................................ 83 Figure 5.7: FR-14 in Karongi District: Four Different Pavement Trials........................................................................... 84 Figure 5.8a. Steep Section of Poor Earth Road Surface (Laos) ...................................................................................... 86 Figure 5.8b. Typical Small Wooden Bridge .........................................................................................................................86 Figure 6.1: Typical Road Asset Management Considerations and Processes............................................................ 91 Figure 6.2: Natural Earth Roads.............................................................................................................................................. 94 Figure 6.3: Shaping and Forming Track to Engineered Natural Surface (ENS).......................................................... 94 Figure 6.4: Poor Performance Due to the Quality of Materials in Wet or Very Dry Environments........................ 95 Figure 6.5: Well-Maintained ENS Roads in Wet and Moderate Climates..................................................................... 95 Figure 6.6: Engineered Gravel Surface (EGS) Road........................................................................................................... 96 Figure 6.7: Roughness Deterioration at 5 Bladings Per Annum (AADT = 324 vpd)................................................... 96 Figure 6.8: Roughness Deterioration at 2 Bladings Per Annum (AADT = 66 vpd)..................................................... 97 Figure 6.9: Blading Frequency and Steady-State Roughness......................................................................................... 97 5 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Figure 6.10: Performance of Various Additives................................................................................................................100 Figure 6.11: Typical Salt Roads in Namibia........................................................................................................................100 Figure 6.12: Koekenaap DR2225 - 75 x 7 Axles HV/Day.................................................................................................101 Figure 6.13: Effect of Regular Sea Water Spray vs. Normal Grader Maintenance...................................................101 Figure 6.14: Respondent Feedback Regarding the Use of Chemical Additives.......................................................102 Figure 6.15: Sealed Surfacing Alternatives........................................................................................................................102 Figure 6.16: Different Types of Bituminous Seals............................................................................................................103 Figure 6.17: Rejuvenation Sprays [Sabita Manual 40].....................................................................................................107 Figure 6.18: Concrete Blocks.................................................................................................................................................109 Figure 6.19: Clay Bricks...........................................................................................................................................................110 Figure 6.20: Dressed Stone...................................................................................................................................................111 Figure 6.21: Cobblestone.......................................................................................................................................................112 Figure 6.22: Nonreinforced or Reinforced Concrete.......................................................................................................113 Figure B.1: Photos from Ghana Case Study......................................................................................................................128 Figure B.2: Tangible Benefits of Dust and Mud Suppression.......................................................................................128 Figure B.3: Geocell and Concrete Application in Kiribati................................................................................................129 Figure B.4: Photos from the Nicaragua Adoquines Road Surfacing Experience.....................................................130 TABLES Table 1.1: Sub-Saharan Africa Road Density Compared to Other Low-Income Countries...................................... 30 Table 2.1: Comparison of Gravel vs. Paved Road Options by Discount Rate.............................................................. 39 Table 4.1: Criteria and Weights for Feeder Roads, Rwanda............................................................................................ 59 Table 4.2: MCA and RED Model Compared (Rwanda Case Study)................................................................................. 60 Table 4.3: Score Sheet Summary (Henning et al)...............................................................................................................61 Table 5.1: Reconciliation of Key Paving Factors against the SPADE Bucket Parameters......................................... 70 Table 5.2: Summary of SPADE-RED Test Results for 54 Roads....................................................................................... 74 Table 5.3: Summary of SPADE-RED Model Results for Roads in Nyagatare District, Rwanda................................ 80 Table 5.4: Summary of SPADE-RED Model Results for Roads in Nyabihu District, Rwanda.................................... 81 Table 5.5: Summary of SPADE-RED Model Results for Roads in Karongi District, Rwanda..................................... 82 Table 5.6: Summary of SPADE-RED Test Results for Three Roads in Tanzania........................................................... 85 Table 5.7: Summary of SPADE-RED Test Results for a Road in Laos............................................................................. 86 Table 5.8: Summary of SPADE and RED Test Results for 7 Roads in Nicaragua........................................................ 87 Table 5.9: Summary of CEA Results for Case Study Roads with AADT under 200..................................................... 88 Table 6.1: Menu of Pavement Surfacing Options............................................................................................................... 92 Table 6.2: Surface Type Decision Framework (STDF).......................................................................................................114 Table A.1: List of Positive Effects and Impacts of Rural Roads......................................................................................126 Table C.1: Increase in Risk with Change in Speed............................................................................................................133 Table C.2: Throughout Routes..............................................................................................................................................134 Table C.3. Protection for Vulnerable Road Users.............................................................................................................135 Table C.4. Safety Measures at Curves in the Road...........................................................................................................136 Table C.5. Safety Techniques at Intersections...................................................................................................................138 Photo credits: Wenxin Qiao 7 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Foreword Whether a road or roads should be paved is maintenance practices. This is critical a crucial decision for planners and officials in because many of the reasons that justify the many developing countries, with far-reaching paving of a road cannot easily be quantified social, environmental and fiscal ramifications. in monetary terms, yet have a disproportion- We are pleased to present this technical ate effect on the beneficiaries’ quality of life, report, that provides an innovative approach and contribute to wider economic benefits in to dealing with this complex longstanding a way that traditional appraisal approaches challenge. cannot capture. Examples include: more reli- able access during rainy seasons, improved Pablo Fajnzylber In the face of increased financial constraints security and safety, business and tourist Director, Strategy and Operations Acting Global Director, Transport and with the real threats posed by climate attraction, improved land values, elimina- change, development practitioners have tion of dust nuisance and associated health to be ever more mindful about how scarce impacts, better market prices for farmers, resources are allocated among competing reduced prices of inputs and consumer priorities. From prioritization of road invest- goods, increased opportunities for public ments, to ensuring that the provided assets transport and mobility, and a move for some are sustained, tough development circum- towards non-farm employment. stances call for innovative development solu- tions. World Bank teams and client countries The approach presented in this report working on rural road projects have strug- (SPADE-PLUS approach) fills important gled with the dilemma of whether to pave a knowledge gaps to overcome limitations of Maria Marcela Silva road or not, and how to prioritize a given set traditional economic evaluation approaches, of roads for investment. The new approach allowing practitioners to effectively incorpo- Practice Manager, Transport Southern Africa, Ethiopia, Sudan, presented in this report provides the tools rate other critical strategic, social, environ- South Sudan to prioritize and justify pavement decisions, mental, and technical justification parameters. using a multi-criteria analysis backed up by By internalizing difficult-to-quantify benefits economic reasoning, and in a systematic, that are normally not considered in conven- robust and defensible way. This will have a tional methods, this work proposes a para- significant impact on the management of digm shift that will move us from designing rural road infrastructure and most important- projects that focus on the priorities of vehicle ly on how benefits for all road users, not just users to a more people-centered approach vehicle users are captured and materialized. that caters to the needs of all users. Equally as important, the approach incorporates the The proposed framework improves the deci- new climate change realities of more extreme sion-making process for paving low volume and more frequent weather events and will Binyam Reja roads by bringing into consideration external provide more defensible paving decisions. Global Practice Manager, variables that are normally not included Transport in the traditional economic analysis, which We believe that this new approach will be hinge the decision to pave or not a road on the next frontier in rural road prioritization cost benefit analysis driven by traffic volumes. and evaluation and we encourage practi- As a result, many rural roads, especially in tioners and policy makers to incorporate it in developing countries in Africa and elsewhere the design of governments and internation- are not justified for paving, and are increas- al financial institutions financed projects to ingly subject to rapid deterioration due to move towards a more robust and justified use the effects of climate and lack of adequate of scarce resources for rural road investment. 8 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE Team & Acknowledgements This report, and the associated SPADE model deliverables (the SPADE model in Excel Format, and the SPADE User Manual in pdf format) were prepared by a team led by Stephen Muzira (Senior Transport Specialist, and Task Team Leader); Atsushi Iimi (Senior Economist, and Co-Task Team Leader); and Wenxin Qiao (Transport Specialist, and Co-Task Team Leader). Other contributors from the World Bank team (to the writing of this report, to the development of the accompanying SPADE model deliverables, and to general task delivery) include: Rodrigo Archondo Callao (Senior Highway Engineer), Ben Gericke (Lead Transport Specialist), Radoslaw Czapski (Senior Transport Specialist), Karla Dominguez Gonzalez (Gender Specialist), Jing Xiong (Senior Transport Specialist), Bezawit Fantta (Transport Specialist), Kanta Kumari Rigaud (Lead Environmental Specialist), Maria Cordeiro (Climate Change Consultant), Robbie Mutyaba (Climate Change Specialist), Ivan Mwondha (Senior Transport Specialist), Oceane Keonu (Transport Specialist), Sudeshna Mitra (Transport Specialist), Matthew Blair Turner (Extended Term Consultant), Sam William Johnson (Extended Term Consultant), Leah Mbugua (Road Safety Specialist), William Majani Wambulwa (Road Safety Specialist), Kazuyuki Neki (Junior Professional Officer), Desta Woldeargey (Program Assistant), and Almaz Bedada (Program Assistant). The team also extends its heartfelt thanks to its government col- leagues, especially from the road agencies in Ethiopia, Mozambique, Rwanda, and Tanzania, for their active help in providing some of the data and case studies used in this work, and for test-running the SPADE model. The following expert consultants were extremely valuable in the task delivery, including their contributions to the report and model development. The team highly commends them for their work: Charles Schwartz (Consultant, SPADE model development), Antonio Ciampa (Consultant, Economist), Jasper Cook (Consultant, Engineer), Gerrie Van Zyl (Consultant, Engineer), and Cecilia Escalante Hernandez (Consultant, Gender Specialist). The team is deeply grateful to Hafez Ghanem (Vice President, Africa) for availing funds for this priority task, and to his front-office team led by Thomas O’Brien (Senior Adviser) for the excellent leadership and guidance to the task team in delivering this work. The team also extends its gratitude to the Infrastructure Practice Management: Makhtar Diop (Vice President), Riccardo Puliti (Regional Director), Maria Marcela Silva (Practice Manager), and the other Africa Transport practice managers: Aurelio Menendez, Ben Eijbergen, Ibou Diouf, and Nicholas Peltier (out- going), for their invaluable guidance and support in the development and delivery of this work. The team thanks the following peer reviewers for their excellent critiques and contributions to the quality enhancement of both the report and the proposed model approach: Cecilia Briceño Garmendia (Lead Economist), Kulwinder Rao (Lead Transport Specialist), Satoshi Ogita (Senior Transport Specialist), Eric Lancelot (Program Leader), Julie Rozenberg (Senior Economist), Sevara Melibaeva (Senior Transport Specialist), Ashok Kumar (Senior Highway Engineer), and James Markland (Senior Transport Specialist). Last but not least, the team extends its grat- itude to Ramon Munoz-Raskin (Senior Transport Specialist), Rakesh Tripathi (Senior Transport Specialist), Cecilia Escalante (Consultant), Emmanuel Taban (Highway Engineer), Maria Jose Pelufo (Transport Consultant), Carlos Bellas (Young Professional), and Antonio Ciampa (Consultant) for their contributions to the case studies. 9 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Abbreviations and Acronyms AADT Average Annual Daily GNI Gross National Income PPS Paving Priority Score Traffic GRSF Global Road Safety Facility PSD Particle Size Distribution AICD Africa Infrastructure Coun- try Diagnostics GWC Gravel Wearing Course RAI Rural Accessibility Index CBA Cost Benefit Analysis HDM-4 Highway Design and Man- RED Road Economic Decision agement Model Model CBR California Bearing Ratio HICS High Income Countries RF Road Fund CEA Cost Effectiveness Approach HMA Hot Mix Asphalt ROCKS Road Costs Knowledge System (World Bank) CEI Cost Effective Indicator IHDS India Human Development Survey RUC Road User Costs DBST Double Bituminous Sur- face Treatment IRI International Roughness SAR South Asia Region Index DFID UK Department for Inter- SATCC Southern Africa Trans- national Development KRRP Kiribati Road Rehabilitation port and Communications Project Commission DID-PSM Difference in Difference with Propensity Score LAC Latin America and the SDG Sustainable Development Matching Caribbean Goals EAP East Asia and the Pacific LCC Life-Cycle Cost SEACAP South East Asia Communi- ty Access Program EEA European Environmental LCCA Life-Cycle Cost Analysis Agency SPADE Systematic Paving Decision LMICS Low and Middle-Income Model EGS Engineered Gravel Surface Countries SPADE- 2-Stage Combination of EIRR Economic Internal Rate of LVR Low Volume Road the SPADE model and the PLUS Return RED or Cost-Effectiveness MCA Multi-Criteria Analysis Approach ENS Engineered Natural Surface MDD Maximum Dry Density Sub-Saharan Africa Trans- SSATP port Policy Program EPA Environmental Protection MNA Middle East and North Agency Africa Surface Type Decision STDF Framework ESAL Equivalent Standard Axle NMT Non-Motorized Transport Load Transport Research TRL NPV Net Present Value Laboratory FR Feeder Road PEA Pure Economic Analysis VOC Vehicle Operating Costs GDP Gross Domestic Product PI Plasticity Index VPD Vehicles Per Day GHG Greenhouse Gas Emissions 10 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE Glossary of Technical Terms ASPHALT Hot or cold mix consisting of graded aggregates, filler, and bitumen, laid, and rolled/ compacted to form final road-wearing course surface BITUMINOUS MACADAM Graded crushed stone (mixed with coarse, medium, fine aggregates and filler material), and infused with bitumen to form a bituminous base-like road surface CALIFORNIA BEARING Penetration test used to evaluate the strength of subgrades and pavements RATIO (CBR) CAPE SEAL Bituminous seal with prime application, then bitumen binder and stone aggregates; and finished off with slurry seal CHEMICALLY Natural gravel that has had chemical additives to improve properties and performance. STABILIZED GRAVEL Could be chemical additives like lime or cement, or bituminous additives, with water added and layer compacted CHIP SEAL Also known as surface dressing; see surface dressing ENGINEERED NATURAL Earth soil that has been graded, shaped, and compacted SURFACE (ENS) GRADED CRUSHED STONE An aggregate mix with suitable grading and other desirable material properties to improve pavement performance GRAVEL A mix of stone, sand, and fine-sized particles used as a sub-base, base, or surfacing on a road. In some regions, it may be defined as aggregate. Gravel may come from natural, or quarry sources. LIQUID LIMIT Moisture content at which a cohesive soil will just start to flow when subjected to a small shearing force under the Casagrande test MECHANICALLY A gravel-wearing course that is improved by mixing two or more different grading soil STABILIZED GRAVEL or aggregate materials, normally gravel, with crushed aggregates to improve properties and performance OTTA SEAL Bituminous seal with graded aggregate added to the bitumen binder, (No prime) and rolled/compacted to form final wearing course. Can be a single or double application, to form a single otta seal or a double otta seal PARTICLE SIZE A list of values or mathematical functions that defines the relative amount, typically by DISTRIBUTION mass, of particles present according to size in a granular material PLAIN CONCRETE Either manually or mechanically mixed concrete made of cement, sand, aggregates, and water 11 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS PLASTIC LIMIT Moisture content at which a cohesive soil will change from a plastic state to a semisolid state that will begin to crumble when rolled under the Casagrande test PLASTICITY INDEX The plasticity index (PI) is a measure of the plasticity of a soil. It measures the size of the range of water contents where the soil exhibits plastic properties. The PI is the difference between the liquid limit (LL) (the state in which the material liquefies) and the plastic limit (PL) (the state in which the material becomes plastic or solidifies (PI = LL-PL). RECYCLED ASPHALT Existing asphalt surface that is scarified, and new asphalt or bitumen injected, mixed, rolled/compacted to form a new, improved wearing course surface REINFORCED CONCRETE Concrete with steel reinforcement incorporated into the final wearing course to add to tensile strength of the pavement RURAL ROAD A publicly-owned road whose primary purpose is to provide access for rural villages and communities to economic and social services. In extended definitions, a rural road can be considered as any public road outside the urban conurbation; hence interurban roads can also be considered as rural roads. SAND SEAL Bituminous seal with sand added to the bitumen film, and rolled/compacted to form the final wearing surface SELF-HEALING ASPHALT This is a relatively novel concept: it has been little tested, and even less used, but it is a highly promising option in which chemical additives are added to the asphalt pavement to give it self-healing properties and eliminate the need for continuous overlays. SLURRY SEAL Bituminous seal with fine aggregates, water, bitumen emulsion, cement or other addi- tives mixed, laid and compacted to form a wearing layer STABILIZED RECYCLED Existing pavement that is stabilized chemically with recycled or other waste materials to PAVEMENT improve properties and performance STONE PAVING BLOCKS Using stones, either in their naturally occurring state, or in processed (cut) form as the final wearing course surface on top of sand bed; normally hand-packed SUBGRADE Native/pre-existing material under a constructed road, pavement or railway track SURFACE DRESSING Bituminous seal with chip aggregates added to the bitumen binder film and rolled/com- pacted to form the final wearing course. Normally this can be a single application (sin- gle-surface dressing), or a double application (double-surface dressing). For double-chip seal, different gradings are used, normally with 14/20- mm aggregates in the first layer, and 6/10-mm aggregates in the second and final layer. 12 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE Structure of the Report This report is organized into seven chapters, with three supporting appendixes. Chapter 1 presents an introduction to the topic. Chapter 2 discusses the traditional approach used in the paving decision-making process based on an economic (cost-benefit) analysis, and provides some examples. Chapter 3 draws on the literature to critique the traditional economic approach, and outlines the key factors that ought to affect the paving decision, but that are rarely taken into consideration. Chapter 4 presents other approaches that are used in the literature to arrive at a paving or non-paving decision, including multicriteria, and cost-effectiveness analyses. Chapter 5 presents the proposed two-stage SPADE-PLUS approach, which is showcased as an improvement over conventional pure-economic approaches with the combination of a multicrite- ria analysis (using the novel SPADE model); and an economic analysis, using either a cost-benefit analysis (CBA) with the Road Economic Decision (RED) Model, or a cost-effectiveness analysis (CEA), depending on current traffic levels. This approach includes innovative permutations to circumvent the traditional critiques of existing methodologies. Chapter 6 presents an engineering summary of the possible surfacing options available to the decision maker, including their pros and cons, and the conditions that align with their adoption on a technical basis. Chapter 7 provides concluding remarks and considerations. Appendix A draws on the literature to provide a summary overview of the reasons that it is import- ant to invest in rural roads. Appendix B delves into some interesting case study applications. Appen- dix C provides road safety guidance teams, to ensure that the paving decision does not translate into a road safety disaster. This report has a dual audience that includes all those who are usually tasked with the preparation and evaluation of rural roads projects, but who face the difficult paving conundrum without many analytical tools at their disposal. That dual audience is (i) the World Bank and other developmental agency task teams; and (ii) government counterparts in the technical agencies. It is hoped that both of these audiences will benefit from this study, and will apply the proposed SPADE-PLUS approach to tackling the longstanding challenge of paving decisions. 13 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Objectives of the Report Investments in road infrastructure as a means to Unpaved roads have a surfacing that consists of enhance access and mobility have been an import- earth or gravel; they therefore demand high levels ant part of the World Bank’s strategy of fighting of maintenance when a certain traffic threshold is poverty and increasing shared prosperity since its reached, or where adverse climatic conditions exist. inception. Rural roads are necessary in order to pro- There are other inconveniences related to unpaved vide people with mobility and access to opportunities roads, including dust in the dry season, mud and pass- in rural areas. Demand for these roads is strong, since ability in the wet season, and the need for the contin- around 45 percent of the world’s 7.8 billion people live ual depletion of borrow pits for regravelling the roads. in rural areas (UN 2020). Just like their urban counter- Given these factors, the desire of many beneficiaries parts, rural households aspire to improve their liveli- is normally to pave the roads. However, with scarce hoods and quality of life, with access and mobility at the resources, not all roads can be paved, and a systematic top of their agendas. Studies suggest that transpor- way to make the paving decision is needed. tation infrastructure triggers economic development through reductions in transport and trade costs, which The difficult task for policy makers, road plan- in turn leads to upgraded access to markets and social ners and engineers is to provide roads that meet services (health, education, administrative, leisure). It their functional and structural requirements in a also fosters agricultural production, alters production cost-effective way. To do this, the spectrum of options decisions, stimulates off-farm diversification, and cata- ranges from providing an engineered natural surface, lyzes other income-earning opportunities. gravel, or low-cost surface treatment, to full paving with asphalt or a concrete pavement surface. Other While the benefits of rural road investment are tra- critical improvements may also be needed for drainage, ditionally determined by economic considerations climate resilience, crossing structures, or road safety that hinge on vehicular travel time and vehicle enhancements. operating cost savings, social and technical con- siderations also play a large part over a project’s The standard practice in rural road project evalu- lifecycle; and also ought to be seriously considered. ation is based on a traditional economic analysis. As a variate means to different ends, farmers use rural However, this analysis usually yields regravelling as roads to take their produce to markets; workers to the solution in most low-volume rural road contexts, travel to their places of employment; tourists to head to since the calculated benefits of travel-time savings and their destinations; the pregnant and sick to seek urgent reduced operating costs are not enough to justify the medical attention; children to get to school; trans- higher costs associated with bringing the road to some porters to move their goods; and families and friends form of paved standard. The assumption under the to visit their loved ones. Speed, safety, and efficiency traditional economic approach is that once regravelled, are all at stake. On the other side of the equation, the if properly maintained, these low-volume traffic roads actual costs of the physical road are defined by the type would be fit for purpose. This is based on a flawed of materials used, the topography of the land, and the assumption that the maintenance will indeed be under- maintenance cycles. taken. Compounding matters, the incidence of high 14 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE levels of rainfall, floods, and landslides, especially in the context of climate change, causes many gravel roads to be washed away, leading to a waste of resources. Many of the benefits of road intervention are hard to quantify, and are not usually captured in a tra- ditional economic analysis. This challenge provides the motivation for the current study. Many relevant attributes—such as climate, the availability of materials, land-value improvements, the attraction of new busi- nesses, the generation of farm and nonfarm employ- ment, the nuisance of dust, travel comfort, the benefits to nonmotorized transport users, reductions in costs of inputs and goods received, increases in prices of crops and goods produced, and the reliability of the road in the rainy season—are difficult to quantify in monetary terms. Road improvement benefits notwithstanding, the negative externalities of community and worker health and safety, road safety, and the environment related to unpaved roads cannot be overlooked, and must also be considered. As Covid-19 has demonstrat- ed, roads can also be a vector that aids the transmis- sion of a deadly virus if transport movements are not properly controlled and monitored during a pandemic. This study proposes an alternative approach that considers the various variables that affect decisions about paving in a holistic way, while maintaining an economic justification. It presents a systematic deci- sion framework that road authorities and development practitioners can use to help them make the difficult decision on whether to pave or not to pave, and any other alternative options that may exist. Photo credits: Curt Carnemark 15 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Executive Summary Investments in road infrastructure as a means for Given scarce public resources, the difficult task granting access and mobility have been an import- for policy makers, road planners, economists, and ant part of the World Bank’s strategy of fighting engineers is to provide roads that meet their func- poverty and increasing shared prosperity since its tional and structural requirements in a cost-effec- inception. Studies suggest that road infrastructure tive way. To do this, the spectrum of available options triggers economic development through reductions ranges from providing an engineered natural surface, in transport and trade costs, which in turn leads to gravel, or low-cost surface treatment, to full paving upgraded access to markets and social services (health, with an asphalt or concrete pavement surface. Other education, administrative, leisure); fosters agricultural critical improvements may also be needed for drainage, production; alters production decisions; stimulates off- climate resilience, crossing structures, or road safety farm diversification; and catalyzes other income-earn- enhancements. The progressive step up along the lad- ing opportunities. As a variate means to different ends, der of surfacing options results in improved capacity for farmers use rural roads to take their produce to mar- handling traffic loads, as well as resisting the damaging kets; workers to travel to their places of employment; effects of climate (rainfall, heat). This does involve high- tourists to head to their destinations; the pregnant and er construction costs, albeit with a reduced need for sick to seek urgent medical attention; children to get to frequent regravelling costs. The challenge, therefore, is school; transporters to make their deliveries; and fami- to find the “sweet spot” where the increase in construc- lies and friends to visit their loved ones. tion costs is balanced by the savings in maintenance, as well as the benefits that will accrue to the target ben- Bridging Africa’s infrastructure gap is key to over- eficiaries over the lifetime of the project. This calls for coming the continent’s development challenges. greater reliance on a proper road asset management Road infrastructure is a key component of this effort. system (AMS) that captures the life-cycle costs of var- Inadequate road infrastructure retards economic ious road surfacing and intervention solutions, rather growth potential by undermining the export compet- than basing decisions solely on the initial upfront costs. itiveness of agricultural produce and other manufac- tured goods; curtails the opportunity for employment In the face of projected massive infrastructure and business development; and impedes human devel- investments, climate change risks2 also pose an opment efforts in health and education. World Bank indomitable challenge. The provision of road infra- estimates indicate that Africa needs $93 billion a year structure lends itself to lock-in effects. Once a road for its infrastructure sectors, with about two-thirds of it is built, with proper maintenance, it is expected to required for new investment in physical infrastructure, last a long time (more than 20 years). The effects of and the other third for maintenance and operations. Of increasing climate variability, and the more frequent this amount, road infrastructure is expected to take up occurrence of more extreme climatic events cannot be about $18 billion.1 ignored. Roads must be designed and built to with- stand the changing climate elements to which they will be exposed over the course of their life cycle. Upstream 1 Africa Infrastructure summary diagnostic available online at https://www.icafrica.org/fileadmin/documents/Knowledge/ICA_publications/ Overview_AICD_Africas_Infrastructure_ENGLISH.pdf 2 Climate change considerations include floods, landslides, extreme heat, sea-level rise, storm surges, hurricanes, and cyclones. 16 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE considerations need to integrate resilience elements seven categories: (i) fiscal and distributional constraints; into the road asset provision exercise to avoid costly (ii) economic and strategic imperatives; (iii) institutional, repair bills later. It makes little sense to regravel a road, network, and system dynamics; (iv) the differentiated only to watch it be washed away in the next rainy sea- needs of various road users and target beneficiaries; son. Continual regravelling also leads to environmental (v) engineering concerns; (vi) climate change consider- degradation as borrow pit materials are depleted. In ations; and (vii) other externalities. However, more fre- Africa in particular, another challenge is that the pre- quently than not, the decision of simply gravelling the sumed routine, periodic maintenance of gravel roads road is made on the pure consideration of traffic vol- is almost never carried out, making a mockery of the umes without factoring in other elements. This simpli- assumptions made in pure economic analyses. In short, fication can turn out to be problematic when ignoring paving a road is one measure that can make a road the not-easy-to-quantify factors, such as climate impact more resilient to the effects of climate change, given (such as increasing rainfall); the type of terrain; a large proper drainage, maintenance, and slope stabilization. number of nonmotorized users among the population being served, but few vehicles; lack of quality gravel How does a policy maker make the decision to materials; and dust pollution issues, among others. pave a road, given such complexities? There are Ignoring these factors and concentrating solely, or even primarily three approaches to choose from: a pure primarily, on an economic evaluation driven by vehic- economic analysis (PEA) using a cost-benefit anal- ular traffic levels that focuses the design on the needs ysis (CBA); a cost-effectiveness analysis (CEA); or a of vehicles rather than the needs of all users can lead multicriteria analysis (MCA). Given the preponder- to a road being left unpaved, with a variety of negative ance of factors that impinge on the paving decision, the consequences for its intended beneficiaries. MCA approach offers the best hope of a meaningful and workable solution to the paving decision problem. To address the conundrum of whether to pave However, since this approach doesn’t yield an econom- or not to pave a road, this work proposes a new ic parameter—that is, a net present value (NPV), or an approach that considers critical influencing internal rate of return (IRR)—it needs to be supple- factors and makes them explicit for the paving mented by either a CBA or a CEA, for a more compre- decision-making process. The approach is simple hensive solution. yet robust, and is designed to be defensible and amenable for application over a wide swathe of Under the traditional approach, using a pure eco- countries and local contexts. This novel, systemat- nomic analysis (a CBA), gravel surfacing is recom- ic paving decision approach is based on a hybrid of mended for many rural roads because the traffic MCA, CBA, and CEA, and is known as the SPADE-PLUS levels are not high enough to justify positive NPVs. approach. It builds especially on the work of Benchmark This approach ignores critical considerations that Engineers (2006), Henning et al (2006), and Dissan- would make for a more defensible paving decision. The ayake and Patel (2016). World Bank task teams, road multitude of other factors can be roughly grouped into authorities, and development practitioners planning, 17 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS prioritizing, and preparing rural road projects can use Bringing the explicit recognition of the multitude the proposed SPADE-PLUS approach to make the dif- of factors that impinge on the paving decision into ficult decision of whether or not to pave a road, or to the first stage of analysis makes the process more decide which roads should be prioritized in a network holistic and more well aligned with development planning exercise, as well as to explore any other alter- realities. This is critical, because many of the reasons native options that may exist. that justify the paving of a road do not have a direct equivalent in monetary values, yet have a dispropor- SPADE-PLUS is a two-stage, sequential approach tionate effect on the quality of life of the intended ben- that combines the Systematic PAving Decision eficiaries, and contribute to the economic development (SPADE) model,3 and either the Road Economic Deci- of an area in a way that traditional approaches cannot sion (RED) model4 or a CEA. It allows for handling capture. (For example, attracting businesses and tour- very complex variables in a simple manner. (See ists, improving land values, getting better market prices Box 1 for details). A sequential approach is preferred for farmers, and moving some toward nonfarm employ- to a one-stage approach because it enables decision ment. There are 41 factors in the SPADE model that can makers to weigh the merits of paving a road from be grouped into five categories: the country context, an MCA approach. This is important given that some the regional context, the operational environment, the factors are not easily quantifiable in monetary terms. road context, and the engineering context. Using the SPADE can also provide economic justification using the SPADE approach, most of the information and data traditionally accepted CBA or CEA approaches. If a pure- used has been selected in a very pragmatic and oppor- ly economic approach is used, however, the justification tunistic manner to facilitate the analysis, and to tap into for paving many roads will be either nonexistent or readily available road agency reports, and road project very unconvincing. Undertaking the traditional eco- feasibility studies. The factors selected as being the nomic approach without making use of the SPADE most pertinent are presented in Chapter 5. model in Stage 1 runs into the challenges that are fully described in this report; and undertaking the SPADE model analysis on its own, without the sec- ond step of economic analysis (with either the RED model or a CEA) results in an incomplete economic evaluation exercise. 3 The SPADE model is an innovative Excel-based model that groups MCA factors to generate a paving priority score that determines whether a road is worth paving or not. 4 The RED model is traditionally used by World Bank transport teams and their partners in many rural road project economic evaluations that are using a cost-benefit analysis. 18 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE Box 1. The Two Stages of the SPADE-PLUS Approach In Stage 1, a multicriteria analysis (MCA) is undertaken that brings together the country, region, operational environment, road, and engineering contexts. Scores are assigned to each of the factors in the SPADE model using the scales provided, with some of the factors noted as being either normal (sin- gle weighting); critical (triple weighting); or super-critical (quintuple weighting). The model then combines the scores from all of these contexts for a total weighted score with scaled percentages of (10 percent, 10 percent, 10 percent, 40 percent and 30 percent) being assigned to each of the buckets respectively. A total weighted score is then obtained, to generate what is termed a Paving Priority Score (PPS). Next a check is made on the total weighted scores (the PPS), to determine the way forward on the paving decision. For roads with low scores (an overall PPS at or below 50 percent), there is no compel- ling justification to pave the road under evaluation. The recommendation in such a case is to do nothing, or to explore other simple alternatives like spot regravelling, or drainage improvements. For roads with a medium score (a PPS between 51 and 70 percent), low-cost paving options like double bituminous surface treatment or otta seals should be evaluated. For roads with a high score (a PPS over 70 percent), there is a compelling justification to pave, and the full menu of paving options should be considered. In Stage 2, an economic analysis is carried out for roads that have passed the SPADE prioritiza- tion for paving with a medium or high priority rating, using either a RED model analysis or a Cost Effectiveness Analysis (CEA) to complete the economic justification exercise. The objective here is not to undermine the comprehensive work done at the MCA stage but rather to complement it. The pro- posed bifurcation is to use the RED model for higher-traffic roads (>200 AADT ) and the CEA for lower-traf- fic roads (<200 AADT) in the economic evaluation of these roads. This average annual daily traffic (AADT) threshold has been adopted based on an extensive review of the literature that suggests that below this threshold, economic analysis models that heavily depend on vehicle operating cost and travel time-saving benefits tend to yield negative results. It is worth noting that 200 AADT is not the threshold put forward for the paving decision, but is only used to define which economic justification tool is to be used in the decision-making process. It is designed to steer the economic evaluation of roads below the threshold to a CEA approach, and those above the threshold to a more traditional CBA approach. It is highly unlike- ly that the economic evaluation will fail for any road that has successfully passed Stage 1. In the remote event that a road fails under Stage 2, an exception could conceivably be provided by the preparation team, using the results of the MCA assigned in Stage 1 for justification to move forward. Therefore, the economic justification stage under the SPADE-PLUS approach is useful primarily for validation purposes. For this reason, it is not recommended that road ranking and prioritization decisions be made based on the Stage 2 results, but rather the Stage 1 results. 19 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS The SPADE-PLUS approach has been calibrated and approach, all of these roads would be recommended validated with case studies from six countries in for paving (100 percent). In a subset of three complet- three continents. The selected case studies are from ed gravel roads totaling 50 km that were damaged by Ethiopia, Mozambique, Rwanda, and Tanzania in Africa; floods and landslides, SPADE-PLUS would have recom- Laos in Asia; and Nicaragua in South America. A total mended that these roads be paved in the first place. of 54 sample rural roads from different programs were This is not surprising, as a paved surface is better able selected for testing, and evaluated using input data to withstand the erosive and abrasive forces of water, drawn from available feasibility studies. Roads that wind, and traffic than a gravel road surface. deserved to be paved were paved given the multitude of positive benefits that they offered, or the contexts in Here is another example of how SPADE-PLUS picks which they were situated (abundant rainfall, large pop- up socioeconomic aspects in a way that the tra- ulations, dust nuisance issues, low-lying flood plains, ditional economic analysis does not. A hypotheti- etc). Further details are presented in the case studies cal test was undertaken to depict a typical rural road section in Chapter 5. setting and paving decision conundrum. A profile of a 10 km road serving 5,000 people, with a high incidence An assessment of the case studies suggests that of poverty, in a high-rainfall area, with all other factors the traditional decision-making approach tends to typical of a rural setting in a low-income country, but overlook critical social, environmental, and eco- with AADT of 100 vehicles per day was generated. The nomic variables that are relevant at the time of pav- RED model analysis yielded a negative NPV (-$2.8 mil- ing, and that more than justify the potential addi- lion) for a Double Bituminous Surface Treatment (DBST) tional upfront cost of paving. The SPADE-PLUS model sealing solution. However, the SPADE-PLUS approach yields a higher percentage of roads recommended to captured the importance of paving this road, which has be paved than the traditional approach: for example, it a PPS of 66, and provides a medium-priority recom- recommended that 53 out of 54 the case study roads mendation for a low-cost paving solution like DBST, otta be paved (98 percent). Under the traditional economic seals, cobblestone, or similar. This recommendation approach, 16 percent of the roads considered in these is further justified with a CEA, which shows a result of case studies would not have been seen as justified $500 per beneficiary, far below the threshold of $1,000 for paving. Notably, the SPADE-PLUS model captures per beneficiary.5 climate resilience as a defining factor for paving some roads. There is enormous value to be gained from using the SPADE-PLUS approach in supporting a more The Rwanda case study offers one of the best les- transparent allocation of resources for paving deci- sons learned. In the Rwanda case study, 163 kilome- sions. The SPADE-PLUS approach equips road prac- ters (km) out of a total of 433 km of roads tested (38 titioners and transport policy makers with qualitative percent) would be recommended for gravelling under and policy arguments in support of attracting resourc- the traditional approach; but under the SPADE-PLUS es for rural road challenges. It also equips transport 5 As of July 1, 2020, low-income economies are defined as those with a Gross National Income (GNI) per capita, calculated using the World Bank Atlas method, of $1,036 or less. A $1,000 threshold is adopted based on a rounding of this number. 20 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE practitioners with the elements they need to commu- used to estimate climate co-benefits (CCBs). Accord- nicate to non-transport specialists—who are in most ing to the World Bank Group’s Reference Guide on cases the ones allocating the scare public resources Adaptation Co-Benefits, the three principles that should needed to pay for the paving of certain rural roads. be used to estimate climate co-benefits are: (i) a context For example, if a road is in a very high-rainfall area with and location-specific focus, meaning that the project poor quality gravel materials, and serves a significant document needs to describe local vulnerability, make population (say over 5,000 inhabitants), running a a statement of intent, and describe how the relevant traditional economic evaluation approach that hinges components will enhance climate resilience; (ii) granu- on vehicular traffic volumes would likely recommend a larity of the information provided that enables estima- gravel road, while taking all of the factors in the SPADE tion of the incremental cost of adaptation measures, model into consideration is likely to result in a PPS that or of the proportion of project components that are justifies paving. There may be some unusual situations addressing climate change vulnerability; (iii) conserva- that may justify paving at all costs, but the SPADE model tiveness--meaning that if the information provided in does a good job of balancing so many different critical the project document is not enough to estimate based paving consideration factors that these cases are the on the second principle, the climate change team will exception rather than the norm. make a conservative estimate of the percentage of total financing that would be needed for the project compo- The SPADE-PLUS approach is presented and justi- nent or activity. fied in this work as the next frontier in rural road prioritization and evaluation; this report showcases The considerations for paving have an impact on the improvements it offers over current approach- the question of climate co-benefits both incremen- es. The SPADE-PLUS approach fills an important knowl- tally and particularly, because gravel is a wasting edge gap in terms of how to overcome the limitations material. When exposed to the actions of traffic, wind, of the traditional CBA approach used in rural road oper- and the effects of rainfall, the material is continuous- ations in Africa, and elsewhere. It effectively addresses ly lost from the pavement surface, and runs into the this challenge, since it can prioritize roads that have drains, or into adjacent fields and properties. If a road low-traffic volume but are critical for local communities is unpaved in a high-rainfall area, the surface is eas- by taking into account the difficult-to-quantify social, ily washed away, leading to faster deterioration and economic, equity, engineering, environmental, and cli- damage of the pavement structure. The models for mate change benefits that are normally not considered the prediction of gravel losses estimate a loss of 20-25 in conventional paving decisions. millimeters (mm) of the wearing course every 5-6 years for gravel roads that receive 800 mm of rain and have In its selection of factors, the SPADE model under- a traffic level of 200 AADT (Van Wijk et al 2019). Jones et scores the need to address climate change consid- al (1984) also isolated the effect of rainfall from all other erations not only in terms of the paving decision parameters, and found that the surface roughness of variables,6 but also in following the three principles an unpaved road increases annually between 970 and 6 Factors related to climate change: A6: Country disaster risk exposure; D11: Redundancy (If major link is disrupted, is road one of viable alternatives?); E1: Drainage: (Is proper drainage provided, or will it be easily integrated into the design?); E4: Climate/Rainfall; and E5: Flooding risk due to low lying road plain/embankment. 21 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS 1100 mm/km per meter of rainfall. Rainfall is therefore Given the holistic nature of the overall situation a critical constraint in determining the performance of regarding the paving of rural roads, it is our recom- an unpaved road. Paving would help to mitigate these mendation that World Bank task teams, and devel- effects. Further analytical work is needed to fully quanti- opment agency and road agency personnel who are fy the paving CCB, but a default of 30 percent is consid- preparing rural road projects, use the SPADE-PLUS ered reasonable. approach to arrive at more robust justifications for paving. The SPADE-PLUS approach can help solve the The SPADE model also offers flexibility because it long-standing dilemma that has plagued practitioners can be easily complemented by other factors, and for so long–not knowing whether to pave a road or the weights can be adjusted accordingly. Practi- not–by applying the SPADE model for the initial screen- tioners that seek to use other factors in their prioritiza- ing and paving recommendation, and the RED model tion or justifications need to draw upon other arguments and CEA for the economic justification. This approach is or sources beyond the present work. Finally, because in line with World Bank’s economic evaluation guidance, of a need to standardize the model for application on a but builds even further with specific recommendations broader scale, the parameters that have been chosen, for practitioners in the rural road projects space, irre- and the weightings applied to them individually, as well spective of practice affiliation. Users are encouraged as to the broader grouping buckets have been pre-set. to standardize the collection of data for the indicated This does not prevent a government or road authority SPADE model variables, in order to achieve the most that has very specific policy objectives from using the robust results. default SPADE model as a template, and customizing it according to their needs. This report does not recom- mend such an approach, but cannot preclude it. 22 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE 1. Introduction   Photo credits: Trevor Samson 23 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS 1.1 The Importance of Rural Roads In research work done for the World Bank, Berg disposal. They can fund infrastructure investments such et al (2015) have noted that roads are the arteries as building a new road; they can use price instruments through which the economy pulses. By linking pro- such as taxes on gasoline, or subsidies for public tran- ducers to markets, workers to jobs, students to schools, sit; or they can issue regulations such as fuel efficiency and the sick to hospitals, roads are acknowledged as or safety standards. With these tools, policy makers being vital to any development agenda. can affect both the supply and demand for transport, which in turn lead to changes in the costs of transport Berg et al have also developed a theoretical frame- services, accessibility, and many other externalities. work for the translation of possible interventions These changes stimulate economic responses in terms in the transport infrastructure space into econom- of trade, location choices, or transport use, and thus ic outcomes, as shown in Figure 1.1. They note that shape the ultimate development outcomes that policy policy makers have three main instruments at their makers seek (Berg et al 2015). Figure 1.1: Theory of Change in Road Investment Options Policy Focus of Outputs Responses Economic instruments Intervention Outcomes Investments Physical ∆ Transport costs Trade Growth infrastructure (new, (incl. time costs) (∆ Production and Price upgrading, Location productivity) instruments Operations & ∆ Access to Maintenance) transport Transport use Inclusion Regulations services / (∆ Opportunity) Transport connectivity service Sustainability ∆ Environmental (∆ Environment & Technology externalities quality of life) Source: Berg et al 2015 24 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE For rural roads in particular, the direct benefits Beyond its effects on agriculture, the provision of of travel time-savings and vehicle operating cost rural roads also contributes to an increase in the reductions from the provision of infrastructure nonfarm employment rolls. In research for the World decrease the costs of transport access to markets, Bank on the $40 billion India Rural Roads Program, jobs, schools, hospitals, and other social, economic, using regression discontinuity Asher & Novosad (2018) and administrative services.7 Enhancements in pro- found that the main effect of the provision of rural ductivity, competitiveness, and profitability for farmers roads was to enable workers to obtain nonfarm work. and other business owners are also associated with rural road improvements, especially in cases where Despite endogenous issue caveats, the clear con- distance leads to isolation. Key outcomes for human sensus is that improved rural roads contribute to capital development, economic development, and pov- poverty reduction and livelihood improvement. erty reduction have been evaluated and discussed in For instance, the lack of accessibility to rural roads has wide-ranging literature to support these claims, using been identified as one of the main causes of poverty various statistical approaches.8 among rural people (Lebo and Schelling 2001). In Ethi- opia, Dercon et al (2008) showed that all-weather rural The impacts of improved rural roads are particularly roads reduced poverty by 6.9 percentage points, and significant in the agricultural sector, which remains increased household consumption by 16.3 percent.11 In among the most important economic sectors in more recent work, an impact evaluation study in Ethi- Africa. In Sub-Saharan Africa, there are still about opia has demonstrated the importance of rural roads, 370 million people, 54 percent of the total labor force, particularly for building resilience to climate change. An who engage in agriculture, though its share has been econometric analysis using geospatial and household declining over the long term.9 The majority of them data suggests that rural road development in Ethiopia are poor and live in rural and remote areas. Africa has has increased household welfare by 22 percent between great potential for agriculture;10 however, poor connec- 2012 and 2016 for roads built between 2010 and 2012. tivity hampers rural farmers’ timely access to advanced Poorer households and those in more remote areas inputs such as fertilizer, pesticides, and improved seeds, were found to have particularly benefited. Improved thereby reducing their productivity. Poor accessibility rural accessibility also appears to have increased house- to output markets also reduces their bargaining power holds’ resilience to drought. Households in places affect- and keeps farm-gate prices low, disincentivizing farmers’ ed by the 2015/16 El Nino drought coped better when adoption of new crops and technologies, and a more rural connectivity had improved: without rural roads, commercialized production system. (Humphreys et al household consumption levels would have been 30 per- 2008; Brenneman & Kerf 2002). cent lower in the drought-hit areas. 7 Rural roads are defined here simply as secondary and tertiary roads, excluding urban roads. National roads could also be classified as rural roads (interurban) but since this paper examines only the lower-level traffic paving decision, these are excluded from the definition used in this paper. 8 See Dercon et al 2008; Humphreys et al 2008; Fan & Chan-Kang 2005; Brenneman & Kerf 2002; Escobal & Ponce 2002; Corral & Reardon 2001; de Janvry & Sadoulet 2001; Lanjouw et al 2001; Lebo & Schelling 2001; and Liu 2000 among others. 9 According to World Development Indicators. 10 Together with agribusiness, it is estimated that agriculture currently generates $31 billion, or nearly half of the GDP of the region. This is projected to continue growing to $1 trillion by 2030 (World Bank 2013). 11 The study covers 15 villages in Ethiopia, and uses the instrument’s variables model to control for household-specific fixed effects. 25 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Figure 1.2: Rural Accessibility Index and Poverty Headcount: National Levels 80 Poverty headcount at national 70 MDG BDI 60 LSO SLE poverty line (%) ZMB LIB 50 MWI MOZ NGA RWA 40 MLI KEN 30 TZA ARM ETH NPL BGD 20 PER UGA 10 0 0 20 40 60 80 100 RAI (%) Source: World Bank 2019 In general, there is significant correlation between institutions, and cultural norms, but not enough atten- a lack of rural accessibility and poverty, as shown in tion was paid to the labor market opportunity structure Figure 1.2.12 Almost half of the multidimensional poor13 that constrains women’s labor market activities. Using in Sub-Saharan Africa—that is, 28 percent of a total of data from the India Human Development Survey (IHDS) 64 percent—experience simultaneous deprivations in in 2004–05 and 2011–12, Lei et al examined how village consumption, education, and access to basic infrastruc- transportation infrastructure affects both women’s ture. The cumulative deprivations reinforce each other, and men’s agricultural and nonagricultural employ- and make it much harder to combat poverty. ment. Results from their fixed-effect analysis show that access by paved or unpaved roads and frequent bus The gender-differentiated impacts of road trans- service increase the odds of nonagricultural employ- port infrastructure are compelling. Lei et al (2019) ment among both men and women. The effect of road investigated the gender-differentiated effects of rural access on nonfarm employment however, is stronger road transport infrastructure improvement on both among women than among men. They also found that farm and nonfarm employment in India. The labor force an improved transportation infrastructure has a stron- participation of Indian women is extremely low, and ger positive effect on women’s nonfarm employment in women are much less likely than men to work in the communities with more egalitarian gender norms. nonfarm sector. Earlier research had explained wom- en’s labor supply by individual characteristics, social 12 The Rural Accessibility Index is defined as the proportion of rural inhabitants who live within 2 km (typically equivalent to a walk of 20 to 25 minutes) of an all-season road, which is defined as a road that is usable year-round by motorized transport. 13 Multidimensional poverty encompasses the various deprivations experienced by poor people in their daily lives – such as poor health, lack of education, inadequate living standards, disempowerment, poor quality of work, the threat of violence, and living in areas that are envi- ronmentally hazardous, among others. 26 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE In a quasi-experimental study using Difference in women ages 15–19, and 50.9 percent ages 20– 34, indi- Difference with Propensity Score Matching (DID- cated that distance was the main factor limiting their PSM), and DID-PSM with covariates that was under- access to health care facilities. Infrastructure mobility taken for a rural roads project in Nicaragua it was barriers can disproportionately affect women’s specific found that the project intervention had a positive health issues, such as pregnancy complications that impact on the welfare of its beneficiaries, but with require immediate care (for example, anemia, eclamp- significant differences by gender (Jimenez 2013). sia, or postpartum infections). This limited accessibility Both men and women’s incomes increased due to the to basic services is exacerbated when roads are not effects of improved rural roads, but the increase was resilient to natural disasters. not statistically different from the increase in men’ monthly average income in the control group, while it A largely overlooked aspect that has a significant was statistically significant for the women. On average, impact is the effect that paving roads has on land women’s income in the intervened municipalities was value improvement, and attracting business devel- estimated to increase 77 percent (according the simple opment. Gonzalez-Navarro and Quintana-Domeque DID-PSM model) or 82 percent (under the DID with (2016) found significant infrastructural benefits from covariates); the results were significant at 90 percent, paving streets in Mexico, with increased consumption and 95 percent confidence intervals respectively. seen within two years of the intervention, based on resi- dents’ capture of their increased property values. Their Paving rural roads also has important implications experimental study found that the valuation of street for access to health, especially women’s health. For asphalting as captured by property value improve- instance, in Mozambique, according to the Demograph- ments was about as large as the construction costs ic and Health Survey, 48.7 percent of the interviewed themselves. 27 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS 1.2 The current situation in Africa Africa lags behind the rest of the world in rural East Asia and the Pacific are above the 80 percent mark accessibility, with an access rate of only 34 per- (Figures 1.3 and 1.4). In the Africa region, most farmers cent (that is, the percentage of the population that has are still disconnected from local, regional, and global access to all-season roads). On the opposite end, Latin markets. Local business development has been ham- America and the Caribbean (LAC), the Middle East and pered because of high transport costs, unreliable logis- North Africa (MNA), and South Asia are close to the 60 tics, and therefore, high inventory costs, all of which percent mark, while Europe and Central Asia, as well as reduce the continent’s competitiveness. Figure 1.3: Rural Accessibility in Africa RAI (%) Pop. w/o access (millions) Population without access (million) 100 500 89.9 443 81.8 Rural Access Index (%) 80 400 59.4 59.4 57.3 289 60 300 40 33.9 200 117 20 54 100 48 51 0 0 East Asia & Pacific Europe & Central Asia Latin America & Caribbean Middle East & North America South Asia Sub-Saharan Africa Source: World Bank Rural Accessibility Indicator Calculations 28 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE Figure 1.4: Rural Accessibility Index, 2006 Source: Roberts et al 2006 Overall, rural areas remain largely underserved, inequitable access makes the flow of people, goods, yet 63 percent of Africa’s population lives in these and services to and from rural areas difficult, and costly areas. The worldwide poverty trend, with 79 percent for the people who can afford it the least. And where of the world’s poor living in rural areas, is replicated in accessibility is low, poverty remains especially high in Africa, where the vast majority of the 413 million poor the region (Figure 1.5.) live in rural areas (World Bank 2018). This unequal and 29 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Figure 1.5: Rural Accessibility Index (RAI) and Poverty Figure 1.6: Percentage of Paved Roads in Africa (by country) Headcount: Subnational Levels Percentage Paved 0 10 20 30 40 50 60 70 80 90 100 Africa Algeria Angola Benin Botswana Burkina Faso Cameroon Cape Verde Central African Republic Chad Comoros Democratic Republic of Congo Republic of Congo Côte d’Ivoire Source: World Bank (2019) Djibouti Egypt Eritrea Ethiopia Gabon At the same time, the vast majority of rural roads Gambia, The in Africa are unpaved.14 Though available statistics Ghana Guinea are sometimes outdated, only about 22 percent of the Guinea Bissau region’s total roads are paved (Figure 1.6). If rural roads Kenya Lesotho are disaggregated, this percentage is even lower: the Liberia average share of paved rural roads is estimated at 10.5 Libya percent.15 Of course, there is a wide variation across Madagascar Malawi countries, from less than 5 percent in Chad and the Mali Central African Republic to nearly 100 percent in the Mauritania Mauritius small island countries, but on average Africa compares Morocco unfavorably with the rest of the world (Figure 1.7). This Mozambique Namibia is clearly one of the important constraints to economic Niger development in Africa (World Bank 2010). Nigeria Rwanda Sao Tome & Principe Senegal Seychelles Sierra Leone Somalia South Africa Sudan 14 Unpaved roads are defined as roads without a permanent Tanzania waterproof surface, and that are made predominantly of earth Togo or gravel. Tunisia 15 Secondary and tertiary roads are traditionally considered rural Uganda roads, though wider definitions may incorporate even national Zambia interurban roads. This statistic captures the former, but the Zimbabwe overall report allows for the wider definition. Source: World Bank Development Indicators* *(Latest information, though year of country data varies) 30 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE Figure 1.7: Percentage of Paved Roads by Region To compound matters, Africa’s road networks are often poorly maintained. Although there is no com- parable global data, the share of roads in poor or very poor condition is generally significantly high, possibly more than half. In Tanzania, about 44 percent of the road network is in poor condition. (World Bank 2019a). In Liberia, more than 90 percent of paved roads are in good or fair condition, but nearly 60 percent of the unpaved roads are in poor condition (IImi and Kulwind- er 2019). In Madagascar, 70 percent of the unpaved roads are in poor condition, and the rate is similar in Ethiopia (World Bank 2019b). The region seems to have Source: World Development Indicators a build-bust-rebuild culture, which eventually turns out to be several times more costly.16 The seminal Africa Infrastructure paper commissioned To provide universal access in Africa, significant by the World Bank noted that the paved road density resources would likely be needed. The Sustainable and total road density of Sub-Saharan Africa’s low-in- Development Goals (SDGs) aim to build resilient infra- come countries is far below the levels observed in structure, promote inclusive and sustainable industrial- other low-income countries (World Bank 2010d). (See ization, and foster innovation (Goal 9), for which Target Table 1.1). 9.1 is “to develop quality, reliable, sustainable and resilient infrastructure to support economic develop- ment and human well-being, with a focus on affordable Table 1.1: Sub-Saharan Africa Road Density Compared to and equitable access for all.” Significant unmet demand Other Low-Income Countries exists in Africa. While AICD (2010) estimated that $9.6 billion, or 1.5 percent of GDP, is needed annually in Normalized Sub-Saharan Other the road sector, the World Bank estimates the aver- units Africa low- low- age annual road investment cost at 2.5 percent of GDP income income (World Bank 2019c). Different studies have assumed countries countries different targets and different levels of accessibility. But Paved-road 31 134 it is evident that achieving universal access is an ambi- density tious challenge in Africa, and one that the continent can ill afford to ignore. Total road density 137 211 Source: World Bank 2010 16 Various studies have found that it costs 3-5 times more to reconstruct a road whose maintenance has been deferred. 31 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Figure 1.8: Construction/Upgrading of Paved Road Projects in Sub-Saharan Africa a. (By Subregion) b. (By Year) Source: AfDB 2014 However, available resources for road development average daily traffic (AADT) of less than 400 vehicles per and maintenance are always limited. There have day (vpd),17 and about half of the roads have less than long been significant efforts by governments and devel- 1,000 vpd. For these low-traffic roads, investment and opment partners to increase road-sector revenues in maintenance that is too costly may not be justifiable order to secure sustainability of investment and mainte- from the economic efficiency point of view. nance in the sector. Since the 1990s, most African coun- tries have adopted road finance mechanisms based on In addition, road paving works have become more road funds. Currently, more than $110 million is collect- expensive in Africa, although there is a wide varia- ed, mainly from fuel levies, in the 25 countries where tion—from less than $100,000 to more than $1 million recent data are available. Still, on average this is merely per kilometer—depending on the technical specifica- 0.5 percent of GDP, far below the need, even with sup- tions. The cost of rehabilitating and upgrading rural port by development partners. This resource constraint roads also changes significantly over time. Although it complicates the political decision-making process. remains a challenge to draw any statistical conclusions from the limited sample of projects, in general the costs The selection of cost-effective options is particularly of both graveling and upgrading have increased over challenging in Africa, where average traffic is often time in the region. In Kenya, for instance, the nominal low, particularly in rural areas. As is well known, the costs of regraveling a road increased from KES1.9 mil- population density in Africa is low (51 persons per km2 lion ($29,000 per km) in 2007 to KES3.7m/km ($43,000 of land), much lower than in other developing countries per km) in 2013, a nominal annual average growth of (for example, there are on average 131 persons in East 17 percent. And in real terms, the costs per kilometer Asia and the Pacific, and 380 in the South Asia Region). rose by approximately 7 percent annually. The increas- In rural settings in particular, this must of necessity ing costs of paving are of course making the political translate into a low level of traffic on rural roads: in Afri- decision of whether or not to pave even more compli- ca, secondary and tertiary roads often have an annual cated. (Figure 1.8). 17 This threshold is adopted from AASHTO (2001) and Keller and Shearer (2003), though different sources quote different thresholds for this definition. 32 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE Rural roads have a critical role to play in connect- Figure 1.9: Madagascar: Pupil Repetition Rates by Road ing target rural populations to their essential Connectivity social needs, such as education and health care. In Madagascar, for instance, the pupil repetition rates are high where road connectivity is poor (Figure 1.9). In the health sector, poor road conditions hamper people’s access to health care services in rural areas, not only because of the impassability of roads, but also the poor quality of services provided. For example, with limited connectivity, it is difficult to deliver medicines to local health facilities (Figure 1.10), and increasingly more difficult for pregnant women to access prenatal care and delivery services. Thus, the pure economic analysis approach is insufficient for rural road investment deci- sion making: there are potentially sizable noneconomic benefits that cannot easily be monetized. Source: Iimi and Rajoela 2018 Figure 1.10: Madagascar: Medicine Stock Rates by Road Connectivity Source: Iimi and Rajoela 2018 33 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Making matters worse, Africa’s roads are becom- Resilience elements must be built into the road ing increasingly more vulnerable to climate events. infrastructure so that these valuable assets can Because of the lack of maintenance and the low withstand the climate forces and effects that they level of the standard, coupled with a rapidly chang- will be subjected to over their life cycle. Suggested ing climate, Africa’s rural roads are particularly resilience measures include the provision of proper vulnerable to extreme climate events. Many African drainage facilities (culverts, side drains, catchwater countries have been experiencing cyclones, floods, drains, bridges, and freely-draining stone aggregate landslides, and heat waves, which often cause signifi- sub-bases); embankment raising; slope stabilization cant damage to the road networks. Now the effects of (proper sloping, slope drainage, vegetation, afforesta- climate change may be more important than ever. In tion); the use of gabion boxes and retaining walls; lining March and April 2019, for instance, southern African side drains; and/or provision of a more permanent countries such as Comoros, Malawi, Mozambique, and impermeable road surface (paving). Zimbabwe were hit by Cyclones Idai and Kenneth, which caused massive flooding in the region. In Mozambique, Although it is hard to have a one-size-fits-all solu- a total of 3,600 kilometers of roads was severely dam- tion, there is a pressing need to standardize and aged, causing significant disruptions in the transport systematize the paving decision process. Such a tool system and the entire economy. Thus, certain roads, is expected to help policy makers and other stakehold- such as lifeline roads, may need to be more resilient. ers consider this question in a holistic way, resulting Because building resilient infrastructure is normally in significant savings of public resources spent on the more costly, it is becoming increasingly important to road sector over the long term, while achieving and bal- encourage further development of rural road networks ancing various development objectives. in an affordable and sustainable way by efficiently using local resources to provide a cost-effective transport infrastructure (Cook et al 2013). 34 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE 2. The Traditional Approach   to the Paving Decision Photo credits: Kubat Sydykov 35 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS 2.1 Reasons for Paving In theory, there are three main reasons for pav- ing:18 (i) to provide a surface that keeps water out of the underlying layers of the road, and prevents them from being weakened by moisture;19 (ii) to cushion the lower-lying, weaker layers from the disruptive and dam- aging effects of traffic by providing additional structur- al support to absorb the imposed vehicle loads, thus preventing premature failures;20 and (iii) to provide a smooth, dust-free, all-weather, fast and reliable surface for both vehicular and nonvehicular traffic that allows Image credit: Curt Carnemark for reductions in travel times, as well as in vehicle oper- ating costs.21 Given these benefits, the choice of many target beneficiaries is almost always to get a paved road. And as economies grow, the demand to pave roads increases (Henning et al 2005). 2.2 The Traditional Cost- Benefit Analysis Approach Image credit: Thomas Sennett The traditional approach is to undertake a pure economic analysis (PEA), using a cost-benefit anal- ysis (CBA) based purely on the quantifiable dis- counted costs of the project (that is, the investment and maintenance costs over the project’s lifetime) versus the discounted quantifiable benefits over the same period of time. The benefits are usually concentrated on travel time and vehicle operating cost benefits, as well as savings in maintenance. Image credit: Wenxin Qiao 18 This report refers to any unsealed road--for example, any earth or gravel road--as unpaved, and any road that has sealing—whether asphalt, chip seal, sand seal, otta seal, concrete, or cobblestone—as paved. In most instances “paved” refers to asphalt, chip seals, and concrete type surfacings as the most prevalent. 19 Though proper camber is the critical element, the imperviousness of the paved surface helps with camber as well. 20 This reason is that it is more supported by the full-pavement options of asphalt or concrete than with the lower-cost surfacing alternatives. 21 For example, Cygas & Zillioniene (2002) found that VOCs on gravel roads are higher because driving speeds are 20 percent less than on paved roads; also, the price of freight carriage is 1.35 to 1.9 times more for gravel roads versus paved roads. 36 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE Traditional economic analysis of road investment The CBA of a road intervention depends on a con- using a CBA is normally based on the consumer fluence of factors. These include the evaluation period surplus approach, where the savings of road-user and discount rate adopted; the project road terrain, costs (RUC), including vehicle operating costs and pas- condition, and traffic; the paving characteristics and senger time costs, are compared with the costs of road costs; the vehicle fleet characteristics; the expected nor- rehabilitation and maintenance. By undertaking a road mal traffic growth rate, and generated and/or induced improvement project (such as rehabilitation, recon- traffic; and the assumptions made regarding the costs struction, upgrading, and/or maintenance), road rough- for maintenance of the road—with and without the ness is typically reduced, vehicle operating costs (VOCs) project—over the evaluation period. are lowered, and users can travel faster, on a smoother surface. As a result, passengers can save time, use less The key factors that influence the CBA results for a fuel, and have less costly vehicle repair bills. On the oth- given paving technology are the following:24 er hand, roads deteriorate with traffic and other condi- tions, such as precipitation. Therefore, the net value of 1. Paving Cost. The higher the paving cost, the low- road investment is determined through the dynamics er the economic justification. The paving cost is a of all of these factors. function of the unit costs of materials, labor, and equipment, and the quantities needed for paving There are several different economic and engi- the road. neering models that carry out CBA in this process, among which some of the most common include: 2. Road Traffic. The greater the volume of traffic on the road, the higher the economic justification. The • The Highway Development and Management Model level of traffic is determined by: (HDM-4),22 which estimates over time the deteriora- • Total normal traffic, in vehicles per day tion of paved and unsealed roads, and the resulting • Normal traffic composition, in percentage agency and user costs used to perform the CBA of • Normal traffic annual growth rate, in percent per various alternatives; or year • Generated traffic, as percentage of the normal • The Roads Economic Decision Model (RED)23 is an traffic Excel model customized for the economic evaluation • Induced traffic, in vehicles per day of low-volume roads; it performs a simplified CBA for • Diverted traffic, in vehicles per day road projects, based on the HMD-4 principles. 3. Road Terrain. The worse the terrain in terms of hor- izontal and vertical profile, the lower the economic justification for paving the road.25 The terrain type 22 http://www.hdmglobal.com/ 23 https://collaboration.worldbank.org/content/sites/collaboration-for-development/en/groups/world-bank-road-software-tools.html 24 For example, paving with a 50 mm asphalt concrete wearing course over a granular base, or paving with a 25 mm double-surface, dress- ing-wearing course over a granular base. 25 With technical or engineering considerations, the opposite situation arises. When there is worse terrain on the vertical profile, there is stronger justification to pave because the unpaved material is more likely to deteriorate. 37 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS (flat, hilly, or mountainous) affects the paving costs 7. Vehicle Fleet Characteristics. The higher the as well as the road-user cost benefits. Paving costs vehicle operating costs and travel-time costs, the are typically higher in mountainous terrain due to higher the economic justification. The vehicle oper- the additional resources needed. The expected ating costs are a function mainly of the price of new road-user cost benefits are also lower in mountain- vehicles and fuel costs, while travel-time costs are a ous terrain, due to the speed constraints caused by function of the value of time. Thus, economies with the geometry of the road. higher vehicle usage costs and higher wages have a greater economic justification for paving unsealed 4. Road Climate. The more rainfall, and the greater roads. the frequency of extreme weather events, the higher the economic justification for paving. Heavy rainfall, 8. Economic Evaluation Parameters. The lower the whether on its own, or through induced floods and discount rate, the longer the evaluation period, and/ landslides affects the road surface, embankment, or the higher the salvage value, the higher the eco- and slope deterioration or destruction, and increas- nomic justification. es the maintenance needs of an unsealed road. 9. Maintenance Assumption. The more the amount 5. Road Surface Type. The worse the all-weath- of maintenance that needs to be planned on an er access, the higher the economic justification. unsealed road, the higher the economic justifica- The surface type affects all-weather access to an tion for paving.26 The planned maintenance over unsealed road. Typically, a packed-earth road does the evaluation period under the without-project not guarantee all-weather access for the whole year, scenario (that is, what it would be if keeping the particularly during periods of heavy rainfall; thus, road unsealed) greatly affects the average condition given the same amount of traffic, there is more justi- of the unsealed road over time. Thus, if a low level fication for paving an earth road than a gravel road. of maintenance will cause the unsealed road to be in worse condition, there is an increased justifica- 6. Road Condition. The worse the condition of the tion for paving. The maintenance assumption under road, the higher the economic justification to pave. the “with project” scenario (for a paved road) is less An unsealed road in very poor condition has a high- critical, because the condition of the paved road will er economic justification for paving than a road in not change very much either with or without main- fair or poor condition. The condition of an unsealed tenance, at least in the initial years, due to the low road can vary greatly during the year (for example, traffic of the road. before or after maintenance, or rainfall); thus, the average road condition over the course of a year The CBA of paving unsealed roads depends greatly or over time, not just the road condition on a given on the definition of the “without project” alterna- date, should be evaluated. tive, which should show what would happen to the road over the evaluation period if the project is not executed. On one extreme, one can consider that no 26 This is the reality of many African road contexts, but it creates the perverse incentive problem. If maintenance is going to be poor, then the model would justify paving, but this is not what any credible program should be driving toward in terms of desirable outcomes. 38 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE maintenance, or very low maintenance will be done on In this example, the EIRR is 16 percent and the NPV, at the unsealed road without the project; on the other a 6 percent discount rate, is $1.84 million. Because the extreme, a high level of maintenance might be done. NPV is positive, at the given discount rate we can say This assumption will have a significant effect on the final that the project is economically justified. economic evaluation model results. Figure 2.1: Impact of Paving on Road Roughness Over Under the typical rural road setting, the main eco- 27 Time nomic benefits from paving are generally attribut- ed to reduced roughness of roads, thereby, faster travel speed and lower vehicle operating costs. In Figure 2.1 the current roughness is around 14 IRI m/ km, which decreases to around 3 IRI m/km, once the road is paved. Without the project, the road will receive three regravellings, while with the project, no periodic maintenance is expected during the evaluation period due to the low traffic of the road. Vehicle operating costs (for fuel, lubricants, tires, maintenance, crew, depreciation, and interest costs), and vehicle travel speeds are a function of the roughness of the road, the terrain of the road, and the vehicle fleet characteristics. The HDM-4 model has equations that compute vehicle operating costs and travel time costs based on the road and vehi- cle fleet characteristics. Thus, it is necessary to define the vehicle fleet characteristics of the country (the cost of new vehicles, new tires, fuel, lubricants, maintenance crews, depreciation, interest, value of time, annual vehi- cle use in kilometers, annual vehicle use in hours, num- ber of passengers, weight of vehicles, and equivalent standard axles of the vehicle) for each vehicle observed Source: on the project road traffic composition. 27 The evaluation period is assumed to be 20 years, and the first year of the evaluation period is 2006. The discount rate is 6 percent, the economic costs (net of taxes) are 80 percent of the financial costs, and the investment cost residual value at the end of the evaluation period is 40 percent. The evaluation did not consider either road accident or CO2 emissions net benefits. The road length is 10 kilometers, the carriageway width is 7 meters, and the road is a gravel road with a roughness of 14 (IRI), meters/kilometer. The current traffic is 300 vehicles per day, of which 25 percent is trucks; the normal traffic growth rate is 3 percent per year for all vehicles during the entire eval- uation period, and no generated, induced, or diverted traffic is expected with the project. The financial paving cost of double bituminous surface dressing is assumed to be $200,000 per km, and the construction will take one year. Without the project, the unsealed road will remain a gravel road and will receive proper maintenance, which will include routine maintenance, two gradings per year and regravelling ($10,500 per km) when the gravel thickness reaches 50 millimeters (mm). With the project, once the road is paved, it will receive routine maintenance, pothole patching, and periodic maintenance in the form of 25mm reseals ($24,000 per km) when the damaged pavement area reaches 30 percent. 39 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Under a traditional economic analysis, the selec- Under the traditional approach, there is a con- tion of road interventions is highly dependent on sensus that a threshold of 100-200 AADT should current and future traffic, as well as on investment determine the decision of whether to pave a road. costs. As described above, various factors are interde- In general, gravel roads are cheaper to invest in than pendent upon one another, when it comes to determin- paved roads, but the maintenance required is more ing the economic value of a road investment. To make costly. Road maintenance costs are typically higher than the paving decision, some factors are presumably more bituminous-surfaced roads when the AADT is greater important than others. Among 21 parameters, the than 100 vehicles per day (Rukashaza-Mukome et al following five parameters, which the economic analy- 2003). Maintenance costs increase sharply, particular- sis is presumably sensitive to, are examined: (i) normal ly when the traffic exceeds 200 vehicles. Zimmerman traffic; (ii) passenger time costs; (iii) dry season rough- and Wolters (2004) show a similar result by compar- ness; (iv) wet season roughness; and (v) investment ing life-cycle project costs: a gravel roadway surface is costs. Regardless of the pavement technology used, the the most effective surface type for AADT from 0-150 volume of traffic is the most relevant determinant of the vehicles per day (vpd), but the chip-sealed and paved economic performance of road investment, followed by surface is more effective for an AADT of 150–660 vpd investment costs. and greater than 660 vpd. Table 2.1: Comparison of Gravel vs. Paved Road Options by Discount Rate Initial Net Present Value at Discount Rate of I.R.R (%) Traffic 0% 5% 10% 15% 20 -0.215 -0.217 -0.218 -0.219 -20.7 50 -0.134 -0.163 -0.179 -0.189 -8.6 75 -0.064 -0.119 -0.149 0.166 -3.3 100 0.024 -0.066 -0.114 -0.141 1.1 200 0.296 0.107 0.007 -0.051 10.5 400 1.362 0.754 0.435 0.254 32.6 700 3.527 2.081 1.322 0.891 71.4 1000 5.817 3.502 2.281 1.584 109.2 Source: Kerali et al 1991 40 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE Figure 2.2: Total Project Costs vs. Initial AADT Cost (C£m) Gravel 7 Traffic Growth - 2% 6 Discount Rate - 10% 5 Sealed 4 3 2 1 0 0 100 200 300 400 500 600 700 800 900 1000 Initial Average Daily Traffic (ADT) Source: Kerali et al 1991 The suggested threshold seems to be robust In addition to the volume of traffic, the paving against the choice of a social discount rate. With a decision is also highly dependent on paving costs. higher discount rate, the initial investment becomes In our sample case, the minimum average daily traffic relatively costly because its future benefits are more needed to justify paving is 161 vpd, when the cost of heavily discounted. At a discount rate of 5 percent, the paving is $200,000 per km (Figure 2.3). But if it costs crossover point is just above 100 vpd; at a discount rate $500,000, and the minimum traffic is doubled to 300 of 10 percent, the crossover point is around 200 vpd vehicles, gravel roads could still be economically viable. (Kerali et al 1991). This supports the traditionally used If the cost of paving is significantly high, for example $1 threshold of 200 AADT (Table 2.1 and Figure 2.2.). million, the guiding threshold could be even greater, perhaps 500 vehicles. 41 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Figure 2.3: Threshold of Traffic for Paving and Investment Costs Flat/Sub Humid/With Generated/3% Traffic Growth/3 Month Grading 1.00 0.80 Paving Cost NPV at 6% discount Rate (M US$ per km) 0.60 (US$ 000/km) 0.40 200 300 0.20 400 500 0.00 600 50 100 150 200 250 300 700 -0.20 800 -0.40 900 1000 -0.60 -0.80 -1.00 Average Annual Daily Traffic (AADT) Last but not least, the paving decision is crucially Examples of the use of the traditional pure econom- related to the sustainability of road maintenance after ic analysis include the following: road rehabilitation or upgrading. Gravel roads typically need more frequent maintenance than paved roads. • Luhr & McCullough (1983). This study analyzed The quality of gravel roads is highly sensitive: without low-volume road (LVR) surface types using a proper maintenance, they can easily deteriorate or, in a pavement management system created for the worst-case scenario, be washed away. Thus, paving can U.S. Forest Service’s LVR network. Three pave- become a more economically viable option if a gravel ment types were considered for the study: aggre- road is not well maintained, as is now the case in many gate (gravel), surface treatment (chip-sealed), and African countries. If a road is graded every 24 months, hot mix asphalt (HMA; that is, paved). Total life-cycle the threshold to justify paving at a 6 percent discount costs for each surface type were estimated based on rate is 84 vpd, assuming that paving costs $200,000 per various traffic mixes and volumes. Results showed km (Figure 2.4). that the gravel and chip-sealed roads became 42 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE Figure 2.4: Threshold of Traffic for Paving and Investment Costs with 24-Month Grading Flat/Sub Humid/With Generated/3% Traffic Growth/24 Month Grading 1.00 0.80 Paving Cost NPV at 6% discount Rate (M US$ per km) 0.60 (US$ 000/km) 0.40 200 300 0.20 400 500 0.00 600 50 100 150 200 250 300 700 -0.20 800 -0.40 900 1000 -0.60 -0.80 -1.00 Average Annual Daily Traffic (AADT) more expensive than HMA-surfaced roads as traffic break-even point. They established that in mountain- increased, due to increased maintenance and reha- ous areas with high rainfall, the break-even point for bilitation costs. upgrading from gravel to sealed road ranges from 50-150 vpd, a lower threshold than in more typical • Kerali et al (1991). This work involved determin- environments. ing the economic viability of upgrading low- vol- ume roads using the HDMIII model. They found • Rukashaza-Mukome et al (2003). This study, con- that the crossover point to pave a road depends ducted for the Minnesota Department of Transpor- on the discount rate. At a discount rate of 5 per- tation, examined when it was economically advanta- cent, the crossover point is just above 100 vpd. At geous to upgrade and pave roads. The researchers a discount rate of 10 percent, it is around 200 vpd. concluded that AADT ranges from 100-200 vpd They also found that of all the other factors beyond should initiate consideration of upgrading a gravel traffic, only rainfall and high vertical alignment road. (topography) had the most significant effect on the 43 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS 3. Challenges with the   Traditional Approach Photo credits: Curt Carnemark 44 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE 3.1 Main Critiques of the 3.2 Key Factors for a Sys- Traditional Approach tematic Paving Decision There are four main critiques of the traditional pure There are seven key considerations that a systematic cost-benefit economic analysis (CBA) approach: model needs to consider in order to develop a more (i) it fails to capture important but hard-to-quantify defensible paving decision than one derived from a social and other economic benefits such as land val- pure traditional economic analysis. These factors are ue improvements, attracting business opportunities, presented below, followed by a deeper discussion, and health benefits from reduced dust, and considerations examples of some of these aspects. for nonmotorized transport users; and focuses instead exclusively on savings in road agency costs, vehicle 1. Fiscal and distributional constraints. Compet- operating costs (VOCs), and travel-time savings for vehi- ing sector demands, competing regional demands, cle users; (ii) it biases investments toward more pros- capital versus recurrent expenditures, budget con- perous areas where vehicular traffic demand levels are straints, consistency and reliability of funding; already more established due to higher motorization, and thus discriminates against investment in poorer 2. Economic and strategic imperatives. Vehicle areas; (iii) it is more suited to dealing with high-traffic operating costs, travel time, poverty, agricultural than low-traffic roads; yet many rural roads are charac- production, other economic or strategic reasons terized by lower levels of traffic volume; and (iv) it fails (tourism, security, cultural, functional classification of to factor in important climate change considerations the road); and other externalities, for example the ever-increasing extreme weather events; and it fails to properly quantify 3. Institutional, network, and system dynamics. other externalities like road safety, and greenhouse gas Institutional setup and responsibility for rural roads, (GHG) emissions from transport. Van de Walle (2002) road asset management, life-cycle considerations, agrees with the first point in particular, and finds that connectivity, network planning, standards and for rural road investment decision making, the pure specifications; economic analysis approach remains unsatisfactory because there is a potentially sizable share of the ben- 4. The differentiated needs of various road users efits that these roads bring that cannot easily be mea- and target beneficiaries. Women, children, men, sured in monetary terms, and aggregated in a consis- the elderly, people with disabilities, pedestrians, and tent manner with the monetary values arising from the other non-motorized transport (NMT), including other costs and benefits. cyclists, animal-drawn carts, motorized (including motorcycles28), commercial transport, public trans- In short, although the traditional road project port, access to schools, hospitals, markets, and other assessment tools are well-equipped to consider administrative centers, rural versus urban beneficia- many parameters based on strong engineering ries, roadside beneficiaries, public preferences; equations and economic principles, there are other important factors that are not easily monetized, but are highly relevant, and that need to be considered, particularly in the context of rural roads in Africa. 28 This work recommends that motorcycles be counted along with motorized traffic, at a factor of 0.5 to a vehicle. 45 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS 5. Engineering concerns: - material availability, mate- cycle need to be properly taken into consideration rial quality, haulage distances, axle and other traffic during the planning, design, and implementation loading concerns, the effects of wind and rainfall, phases; and drainage, structures connectivity, slope stability, soil erosion, maintenance, subgrade quality, road 7. Other Externalities. Road safety, community and condition; occupational health and safety, GHG emissions, the health impact of dusty unpaved roads, depletion of 6. Climate change considerations. With the ever-in- borrow pits, climate change considerations, environ- creasing frequency of extreme weather events (hur- mental sensitivities. ricanes, cyclones, floods, landslides, extreme heat), and the lock-in effect of the provision of transport The seven considerations are presented graphically in infrastructure, the long-lasting effects of climate Figure 3.1. change on the infrastructure of the road over its life Figure 3.1: Seven Key Considerations for a Systematic Paving Approach A. Fiscal and Distributional Constraints B. Economic G. Other and Strategic Externalities Imperatives C. Institutional, F. Climate Network and Change System Considerations dynamics D. Differentiated E. Engineering Needs of various concerns users and target beneficiaries Source: World Bank 46 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE FISCAL AND DISTRIBUTIONAL CONSTRAINTS First and foremost, many African governments liter should be collected to ensure the sustainability are faced with severe budget constraints. Available of road maintenance (World Bank 2019a). In reality, financial resources are limited, and there are many however, most of the countries in Africa charge less competing development demands. To secure resourc- than that (Figure 3.2). As a result, the available financial es for timely road maintenance and development, since resources through RFs are mostly less than 1 per- the 2000s, many African countries have adopted road cent of GDP (Figure 3.3), while the estimated require- funds (RFs), the revenue of which is largely dependent ment for achieving reasonable road connectivity and on road user fees, such as fuel levies. A recent World ensuring the sustainability of investment is double that Bank study recommended that at least $.15 cents per amount (2 percent) in Africa (Gwilliam 2011). Figure 3.2: Fuel Levy on Petrol (in US$/liter) Figure 3.3: Road Fund Resources as a Share of GDP (%) Fuel levy on petrol ($/liter) 0 5 10 15 20 RF resources per GDP (%) 0.0 0.5 1.0 1.5 2.0 2.5 Burkina Faso Mozambique Kenya Namibia Namibia Kenya Malawi Malawi Tanzania Tanzania Rwanda Zambia Cameroon Sierra Leone Côte d’Ivoire Mali Senegal Senegal Central African Republic Togo Ghana Côte d’Ivoire Togo Burkina Faso Ethiopia Chad Sierra Leone Ghana Mali Cameroon Mozambique Benin Guinea Burundi Niger Zambia Benin Niger Republic of Congo Republic of Congo Burundi Chad Guinea-Bissau Madagascar Madagascar Guinea Gambia, The Guinea Bissau Rwanda Gambia, The Ethiopia Source: World Bank 2019a Source: World Bank 2019a 47 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS INSTITUTIONAL, NETWORK, AND SYSTEM DYNAMICS Figure 3.4: Road Fund Resources per Length of Road Institutional Capacity and Coordination. Questions RF resources ($/km) that need to be asked include the following: Do the 10000 1000 2000 3000 4000 5000 6000 7000 8000 9000 nominated agencies have clear mandates for the net- 0 Mozambique works under their control, and the capacity required Senegal to properly plan, deliver, and maintain the infrastruc- Côte d’Ivoire ture, including the consideration of integrating climate Ghana change and natural disaster risk? Are there arrange- Togo ments and mechanisms to coordinate required action Tanzania across relevant ministries or agencies in place? Are Zambia the governance processes, and the flow of information Kenya clear, and working? Cameroon Namibia Network and Systems Planning. Are road inventory and condition surveys undertaken to know which assets Burkina Faso are in place, and in what condition? Is an asset manage- Malawi ment system (AMS) in place? Has mapping been done Mali in order to understand the effects of the infrastructure Benin on the network? (For example, which roads are con- Sierra Leone nected to which areas of importance? Which roads are Guinea redundant? Which ones are critical?) The definition of Republic of Congo asset location is also important in order to reduce the Chad level of exposure to natural hazards, and to consider Niger the lock-in effect and the impacts of climate change. Burundi When critical infrastructure is in a high-risk location Ethiopia then understand and provide redundancy in the net- Gambia, The work. Also, how can paved roads contribute to enhanc- Guinea-Bissau ing the resilience of communities? For example, could Rwanda infrastructure have dual functions, such as water cap- ture and containment for use during drought periods? Source: World Bank 2019a Conversely, if paved roads are built on elevated ground, can these serve as secure locations for evacuation of the population during flooding – link with contingency programming pillar. 48 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE NEEDS OF VARIOUS ROAD USERS AND TARGET BENEFICIARIES Population. In a traditional economic analysis, the to complex travel patterns. For example, in rural areas vehicular traffic (in terms of AADT) has a disproportion- they are responsible for food harvesting, transport, ate impact on the results of the analysis. This, however, storage, and marketing activities (AfDB 2007; Porter begs the question of roads that serve communities with 2011). Also, as Peters (2002), and more recently Domin- sizeable populations, but where most travel is through guez-Gonzalez et al (2020) have noted, a dispropor- nonmotorized transport (walking, bicycles, animal carts), tionate share of the transport burden falls on women, or through public transport vehicles. Some approach- yet they also have more limited access to means of es, like cost-effectiveness analyses (CEAs), have set transport. Furthermore, because women prioritize the a threshold value of dollars spent per person in an economy of care this translates into mobility patterns attempt to overcome this problem, but the cost-benefit for women of shorter distances, and slower speeds for analyses (CBAs) leave it largely unaddressed. essential household activities like fetching water and taking children to school. Presence of schools, hospitals, markets, and oth- er administrative centers. A key aspect that goes In rural areas, poor infrastructure and limited transport largely ignored by the traditional economic analysis is services constrain both male and female mobility, but whether a road serves schools, hospitals, markets, or women face additional sociocultural constraints that other administrative centers. Roads are not an end in make the negative effects for women even stronger. themselves, but a means to an end. They are supposed Given the mobility patterns of women and girls (for to be the means that enable users to access services example, reliance on walking), they are also a vul- like schools (for pupils, teachers, parents); hospitals or nerable population in terms of road crashes and/or health centers (for patients, especially vulnerable ones infrastructure designs that lack the features that can like pregnant women, and for ease of access to vacci- ease walkability, such as sidewalks. They are also neg- nation services); markets (critical for farmers to obtain atively impacted by the lack of complementary facilities inputs like seeds and fertilizers, but also to obtain more to improve security, such as lighting. Therefore, the favorable farm-gate prices, as well as to contribute to decision of whether or not to pave roads should be improved food security); and resilience to droughts or accompanied by considering design features that natural disasters. The importance of these facilities, and will enable the mobility of people traveling by foot. their consideration in the paving decision, cannot be overlooked. Various technologies for road improvement can have additional positive impacts on women’s mobility, as Gender considerations. Women and men use and well as on human capital. Improved roads can incen- benefit from roads differently. It is widely recognized tivize women’s use of intermediate modes of transport that women and men have different mobility pat- such as bicycles, enhancing their mobility and access to terns. From the users’ side, there have been important services; and better roads can increase their sense of insights into the way social structures and transport security and safety. infrastructure and services intermingle and lead to dif- ferences in the demand for and the use of mobility by These differentiated demands should be considered in men and by women. In addition, women’s multiple roles the process of making paving decisions. Based on these usually lead to juggling a variety of tasks, which leads arguments, guidance regarding the type of paving 49 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS technology to be used should consider the travel needs reinforce the gender angle on the pedestrian counts is and purposes of different groups of the population in the share of pedestrians carrying heavy loads on their the areas of planned intervention. Better understand- heads. Studies have shown that women walk more than ing of the road users, road usage, and origins and men, and in some cases they do it carrying heavy loads destinations for different population groups will make it on their heads: by counting them, and judging roads easier to identify who will be the “winners” and who the that serve a large share of people carrying goods as pre- “losers” as a result of the decision of whether or not to mium, a strengthened gender angle to the NMT counts pave low-volume roads. can be made. Dust nuisance has been mentioned earlier, but the fact that many women and children are Given the widespread use of motorcycles in Africa, pedestrians makes this issue very pertinent for them. this study recommends that the motorized traffic count include motorcycles, if it’s not already stan- Overall, from an equity and inclusion point of view, the dard practice in the country. Because of motorcy- decision of whether or not to pave can have import- cles’ low gross and axle weights, a suitable conversion ant implications; therefore, having users’ perspectives factor for vehicle equivalent can be developed.29 In in mind is important in the process of prioritizing, some rural areas, motorcycles may be the only mode selecting, designing, constructing, and maintaining of public transport available, which would make them roads. Analytical work with a gender angle—through also important for emergency response. As an example, focus groups and semi-structured interviews, women’s in Tanzania the registration of motorcycles went from safety audits, participatory mapping and the like—can less than 2,000 in 2003 to 800,000 in 2014, and recent help identify the complementary interventions that are feasibility studies conducted in rural Tanzania show that needed to ensure that different sectors of the popula- motorcycle traffic can account for up to 90 percent of tion are being considered, in order to increase mobility the total daily traffic (from 26 to 90 percent). Starkey and equal access to opportunities. For instance, in one and Hine (2014) cite similar shares for a road study in project under preparation, the Roads to Inclusion and Cameroon, and Porter (2007; 2011) showed that men Socioeconomic Opportunities (RISE) program, women are the main users of motorcycles in rural areas. How- expressed feeling unsafe and vulnerable when using ever, the provision of safer roads and infrastructure, roads in quiet areas and taking shortcuts, especially at while also addressing other barriers, can incentivize night. In such instances, women tend to avoid travel if women to use motorcycles more, which would improve possible, and only an emergency will push them to do their mobility. the trip, in which case they would prefer to use a well- known and reliable motorcycle taxi service. The project The criterion for nonmotorized transport (NMT), design uses a people-centered approach that has has a particularly strong gender angle. NMD traffic allowed it to identify the needs of different users, and counts, which includes road users like pedestrians, to include design features that address such concerns. cyclists, and animal-drawn carts, ought to be included Other rural road projects are encouraged to use the in the project preparation. One relevant indicator to approaches being used in RISE. 29 One approach would be to merely count a motorcycle as half-a-vehicle for the model purposes, but not for pavement design purposes. This is what this study recommends. 50 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE ENGINEERING CONCERNS CLIMATE CHANGE CONSIDERATIONS Availability of materials. A critical issue in the paving With more frequent, and more extreme climate decision is whether or not quality gravel materials are events, climate change is becoming an ever-in- available—that is, materials that pass the technical creasing concern in the provision of transport infra- specifications for a good surface course that can with- structure. Extreme rainfall and extreme heat are key stand the effects of traffic, rainfall, and other climate considerations, but regarding the decision of to pave or events. In work done for the Transportation Research not to pave, the rainfall question is the more pertinent Laboratory on the deterioration of gravel pavements one. The heat question is more relevant when choosing in Africa, Jones et al (1984) found that unpaved roads among various paving options (asphalt vs. chip seals vs. annually lose 10-30 millimeters of gravel per 100 vehi- concrete). cles per day. This means that 70-210 cubic meters of gravel are removed from the road surface each year. According to the World Bank Group’s Reference Guide Particle Size Distribution (PSD) was found to be a criti- on Adaptation Co-Benefits, the climate change team cal element in slowing down the rate of deterioration. follows three principles to estimate climate co-benefits Engineering analysis of the materials to be used also (CCBs): (i) context and location-specific focus— meaning indicates that the type of gravel—not only PSD, but also that the project document needs to describe vulner- other properties, like the Plasticity Index (PI)—greatly ability, make a statement of intent, and describe how influence the performance of gravel material.30 In areas the relevant components can enhance resilience; (ii) where the materials for gravel are poor, an economic granularity of the information provided, enabling esti- evaluation that comes up with the recommendation mation of the incremental cost of adaptation measures, to keep the road unpaved but add a layer of gravel (in or of the proportion of the project component that is other words, to re-gravel the road) misses the point: addressing climate change vulnerability; (iii) conserva- this is not advisable if the only available materials will be tiveness—meaning that if the information provided in washed away or wasted in a short period of time. the project document is not enough to estimate based on the second principle, the climate change team will Connectivity of structures. In certain situations, the make a conservative estimation of the percentage of question may not be whether or not to pave, but rather total financing for the project component or activity. whether to provide connectivity across natural barriers like rivers, streams, or valleys that are preventing the There are two approaches to assigning CCBs for road transportation of goods or people due to geographical infrastructure projects: and climatic constraints, especially in the rainy season. In such cases, there may be a greater benefit in provid- 1. Incrementality. At the infrastructure level this is ing crossing structures to ensure all-year accessibility estimated based on the incremental cost of works for the target population. This factor should be serious- that are embedded in the design to make a road ly analyzed, especially in areas with abundant rainfall, asset more resilient by considering the identified or areas that are experiencing increasing effects of cli- climate change vulnerability context (for example, mate change. In such cases, clear trade-offs may have moderate to high climate risk, which may require to be made. an extra layer of pavement surface; or additional, or 30 The Plasticity Index (PI) is a measure of the plasticity of a soil. It indicates the size of the range of water contents where the soil exhibits plastic properties. The PI is the difference between the liquid limit (LL) (the state in which the material liquefies) and the plastic limit (PL) (the state in which the material becomes plastic or solidifies (PI = LL-PL). 51 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS wider, drainage, or bridges with higher clearance. Task teams have an obligation to undertake screening The costs of these incremental resilience measures for climate and disaster risks, and to estimate climate are counted as CCBs (adaptation finance). When CCBs in road transport projects following the three- detailed information is lacking, an estimate of 25-30 step process shown in Figure 3.5. percent is normally adopted. The best way to estimate the CCBs assigned to paving 2. Proportionality. At the investment, or network would be to compare the co-benefits of the “with pav- level, if the whole investment or project is undertak- ing” scenario versus the “without paving” scenario of en in order to respond to climate change impacts, the same project. Further analytical work is needed. then a higher CCB can be claimed, and the whole component or project can be counted as adaptation This work has shown that the considerations for finance. paving have an impact on the question of CCBs. If a road is unpaved in a high rainfall area, the surface will Normally, a road project will receive adaptation be easily washed away, leading to faster deterioration co-benefits as follows: and damage of the pavement structure, since gravel is a wasting material. When exposed to the effects of 1. Road rehabilitation and new road works traffic, wind, and rainfall, gravel material is continual- • Use weather-resistant paving material to with- ly lost from the pavement surface, and runs into the stand extreme weather events drains, or into adjacent fields and properties. In work • Construct wider drains and larger culverts to done for the Transportation Research Laboratory on accommodate heavy precipitation the deterioration of gravel pavements in Africa, Jones et • Widen or construct side drains, catchwater al (1984) found that unpaved roads annually lose 10-30 drains, and mitre drains to properly drain water mm of gravel per 100 vehicles per day. This means that away from the pavement structure 70 to 210 cubic meters of gravel are removed from the • Undertake slope stabilization to protect against road surface each year. Jones et al (1984) also isolated climate risks the effects of rainfall from all other parameters, and found that the surface roughness of an unpaved road 2. Road maintenance increases annually between 970-1100 mm/km per • Resurface roads to prevent the surface material meter of rainfall. Rainfall is therefore a critical constraint from being washed away in determining the performance of an unpaved road. The importance of paving the road–in addition to other 3. Bridge rehabilitation and construction sound engineering measures like providing drainage • Undertake structural retrofits/reconstruction to and slope stabilization – cannot be overlooked, because withstand climate hazards paving would help to mitigate these effects. Further • Use weather-resistant paving materials on analytical work is needed to fully quantify the paving bridges CCB, but a default estimate of 30 percent is considered • Provide bridges with higher clearance reasonable. 52 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE Figure 3.5: Links between Climate Co-Benefits (CCBs) and Risk Screening Source: Climate Change Group, World Bank (2020) Integrating Resilience. It is important for teams that The sectoral checklist for transport elucidates nine are preparing rural road projects to integrate resil- resilience attributes: robustness, learning, redundancy, ience upstream in the design process by linking project rapidity, connectedness, diversity, flexibility, inclusion activities with attributes that strengthen core resil- and self-organization. These are shown graphical- ience capacities; that is, their absorptive, adaptive, and ly in Figure 3.6. The SPADE approach doesn’t replace transformative capacities. These attributes and related the need for teams to undertake this checklist, but “markers” can also help teams track progress toward it embeds the key aspects in it in the model to help resilience-building in the context of implementation, inform the final paving decision. to strengthen the delivery of resilience outcomes, and contribute to corporate climate commitments (Rigaud et al, 2020). 53 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Figure 3.6: Resilience Attributes from the World Bank Sectoral Checklist ROBUSTNESS LEARNING REDUNDANCY RAPIDITY CONNECTEDNESS DIVERSITY FLEXIBILITY INCLUSION SELF- Resilience Attributes ORGANIZATION Ability of Ability of Availability of Speed Breadth of Ability Ability Extent to Ability to the system the system additional or at which resources and of the of the which the sys- independent- withstand to gain spare resources assets structures system to system to tem embrac- ly rearrange the impact or create and/or institu- can be that a system under- respond to es equity and functions and of shocks knowledge, tions that can assessed, can access, take uncertainty, inclusiveness, processes in the and fluctu- and build be accessed in mobi- at multiple different addressing and provides face of shocks ations and the skills, case of shocks lized and levels, in order courses challeng- fair access and stressors, to maintain its attitudes or stressors, accessed to respond or of action es and to rights, diagnose prob- character- and com- and that are by system adapt to shocks and utilising the resources lems, assess istic petencies interchangeable stake- or stressors. innovate. oppor- and oppor- priorities, and/ needed to among them holders tunities tunities to its or mobilise innovate (i.e. overlap- to achieve that may members. resources to ini- and adapt to ping functions, goals in an arise from tiate solutions. change services and/or efficient change. capacities). manner. Source: Rigaud et al, 2020 Climate vulnerability, criticality, and redundancy. is typically low, tend to be easily isolated when heavy The traditional road analysis is still crucially deficient rain occurs. From the road network point of view, some in capturing resilience, redundancy, and criticality, all roads are more critical than others because there is three of which are becoming increasingly important little redundancy. in the context of climate change. Road infrastructure is expected to last decades. Gravel roads are relatively In the case of Mozambique, for rural roads that were less expensive to invest in, but are particularly vulner- evaluated the value of criticality was considered in able to climate events such as heavy precipitation and deciding which roads should have priority.31 In this case, floods, and are easily cut. Paved roads are more costly, the criticality of each road was measured by comparing but they are also more resilient and more resilient to all the current road-user costs with hypothetical road-us- weather conditions. er costs, using the assumption that the road would be impassable.32 On the other hand, the traditional Road connectivity cannot be compromised under consumer-surplus approach, where savings in road-us- certain circumstances. For example, access to health er costs are compared to investment costs, can also care services should be assured regardless of weath- be used to calculate conventional economic assess- er conditions. Rural areas in Africa, where road density ment measurements. The net present value (NPV) at a 31 Mozambique Integrated Feeder Road Development Project (P158231). 32 For the purposes of simplicity, the criticality is measured by the potential increase in road-user costs due to a 7-day disruption, consider- ing the current level of traffic. See a more comprehensive discussion in Espinet et al 2019. 54 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE discount rate of 6 percent is used here. For simplicity, Figure 3.7: Criticality and NPV of Rural Road Investment a constant unit cost for rehabilitation and upgrading in Mozambique was assumed, and a project life of 20 years was consid- ered.33 About 20 percent of rural roads are considered economically not justifiable, but it is acknowledged that they could cause significant disruption costs if they were to be cut off. Not surprisingly, there appears to be a broad correlation between the criticality measure- ment and the conventional NPV. However, there are cases where the rehabilitation investment cannot be justified economically based on the estimated NPVs being less than zero, but with a significant criticality (Figure 3.7). In total, 514 roads in Mozambique were examined. For 92 roads of the roads, the NPVs were Data Source: Piloting the Use of Network Analysis and Decision- Making under Uncertainty in Transport Operations, X Espinet, J negative, but some additional costs would occur if they Rozenberg, KS Rao, S Ogita, World Bank, 2018 were to be disrupted. In 7 cases, the costs of disrup- tion would exceed half a million U.S. dollars: thus, some additional consideration was needed for the decision, Design standards and material specifications: Do from the socioeconomic point of view. When there is design standards and materials specifications consider no alternative route in a local area, a road may need to climate change and natural disaster impacts? What sup- be developed with a higher level of standard to ensure plementary infrastructure is being deployed to protect all-weather passability, justifying the paving regardless roads? Are there dams or enhanced drainage systems? of economic efficiency. How does the choice of materials impact resilience and connectivity year-round? The issue of redundancy is a critical consideration. If two parallel roads exist between two major points, A Operations and maintenance: What are the project and B, there is less justification for paving one of these provisions in terms of maintenance when deploying roads if the other is already paved, and the resources paved versus unpaved roads? What is being planned to could better be spent on paving other roads that have ensure continued resilience to climate change effects? no redundant alternatives. Contingency programming: Are there institutional, financing, procurement, and implementation provisions in place to facilitate quick rehabilitation and restoration of road function after a natural disaster? 33 Based on a recent feeder road project in Mozambique, the average investment cost is assumed to be $70,763 per kilometer. The traffic is assumed to grow at an annual rate of 3 percent. 55 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS OTHER EXTERNALITIES Road Safety: Every year 1.35 million people die on the the risks can be far greater than the safety benefits roads; and more than 93 percent of road crash fatali- unless mitigating safety strategies are used. Therefore, ties occur in low and middle-income countries (LMICs). while in assessing the economic benefit of paving the Road crashes and injuries disproportionately affect the benefits from the reduction in travel time are consid- young, an economically productive age group, which ered, the probable increase in crash, injury, and fatality then also creates a critical impact on human capital. risks should also be considered. A deeper discussion of They not only devastate families emotionally and finan- these issues, together with some possible low-cost miti- cially, but they also take a toll on the path to develop- gation interventions that can be adopted are presented ment for many LMICs. In Africa, a significant amount of in Appendix C. fatalities and injuries happen on rural roads, although very often this is not reflected in official statistics. Greenhouse gas emissions (GHGs). Road transport While paved roads provide better access to educa- is one of the main contributors to GHG emissions, tion, health care services, and businesses, there is little accounting for 72 percent of the total for the trans- objective and conclusive information on the safety portation sector (EEA 2018). It is also responsible for impacts of paving roads (e.g. from dirt or unsurfaced 14 percent of total GHG emissions overall (EPA 2018). roads to concrete or aggregate surfacing) that would African travel, both passenger and freight, is high- enable estimating the potential number of reductions ly road-dependent, and the emissions’ share of road or increases in fatalities and serious injuries to factor transport in the transportation matrix is likely higher into the economic analysis. As such, the assumption than the 72 percent cited above. is often made in traditional economic analyses that safety benefits will ensue from paving a road. This is The argument is that given the low speeds, frequent due to the likely reductions in certain types of risks that braking and acceleration, and long travel times, with emerge from riding on an uneven surface, especially for the consequent increases in carbon dioxide and oth- two and three-wheeled vehicles. Pedestrians, cyclists, er emissions, roads in bad condition increase fuel and animal-drawn cart users benefit from improved consumption. This contributes to global warming and shoulders and sidewalks if provided, and visibility can increases outdoor air pollution, affecting the health of also be improved with less dust. All of these factors can road users and people in adjacent properties. For this aid to the safety outcomes. However, there is strong reason, the paving of low-volume roads (LVRs) is viewed evidence to suggest that paving a road can also result as a measure that may contribute to the reduction of in negative safety outcomes, especially if specific and GHG emissions, and some consider it appropriate to adequate road safety improvements are not made at incorporate this potential reduction into the exoge- the same time. neous benefits accruing to society overall. The counter- argument is that paving the road leads to an increase The biggest safety risk comes from the increase in induced traffic, which might offset some of the in speed when there is a smoother road surface. expected climate change co-benefits discussed above. Studies indicate that for rural roads, for every 1 km/ hour increase in speed, there is around a 4.5 per- The traditional road assessment tools do not center on cent increase in the risk of a fatal crash outcome (Elvik the increase or reduction of GHG emissions, though 2009). Coupled with the possible increases in traffic, they technically compute the emissions that would be 56 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE generated by a given investment scenario. For instance, Figure 3.8: Relationship of Vehicle Speed and HDM-4 considers seven different components of Dust Formation exhaust emissions: hydrocarbon (HC), carbon monox- 25 ide (CO), nitrous oxide (NO), sulphur dioxide (SO), par- ticulates (Par), lead (Pb), and carbon dioxide (CO2). The engine-out emissions are simply estimated based on 20 Emission factors lbs/veh. mile the fuel consumption rates. late rticu Despite such a detailed calculation, the traditional 15 l pa model may be underestimating the impact of GHG Tota emissions on the planet. From the economic point of view, this is an issue of the costing of GHGs emissions. 10 The social cost of carbon measures the economic harm from those impacts, expressed as a dollar value of the s on 5 icr total damages caused by emitting one ton of carbon m dioxide into the atmosphere. The current central esti- 10 er Und micron s mate of the social cost of carbon is roughly $40 per ton Under 2 of CO2; however, this value is not reflected in the prices 0 10 20 30 40 50 of Carbon Certificates markets. Because of high price Speed, mph volatility, it is recommended that GHG emissions be Source: Azulbaidi (1999) costed at very conservative levels. Dust Pollution. The issue of dust on rural roads does On unpaved roads, the effect of friction on tires as not quite get the recognition and attention it merits. they move over the road disperses the dust particles During the dry season, the vast network of unpaved and makes them airborne. These particles are then roads poses a dust nuisance, and can lead to an inhaled by road users who have direct exposure, but increase in respiratory diseases, the spread of com- also pedestrians, occupants of vehicles without air-con- municable diseases, and hinderance of road visibility, ditioning and/or circulation control mechanisms, and thus presenting a safety issue. Khan and Strand (2018) people living and working in adjacent properties. This is did an extensive review of the literature and found that a critical health factor that is not traditionally consid- road dust has harmful effects on the human body, most ered in the paving decision. specifically on the respiratory system. Beyond respira- tory damage, other literature has identified a variety of Depletion of borrow pits. In the unpaved scenar- health effects, including those associated with PM2.5 io, keeping the road in motorable condition usually which leads to cardiovascular complications from dust requires continuous grading, and periodic regravelling. inhalation (Bell et al 2007; Franklin et al 2008). The regravelling necessitates the exploitation of materi- Azulbaidi (1999) documented the harmful effects of dust als from borrow pits. This continual depletion of gravel generation relating to unpaved roads based on Lindh resources is not good for the environment, as it leads (1981) and Roberts and Walter (1975). (See Figure 3.8.) to cavernous excavations, and the depletion of natural vegetation cover. These activities can also accelerate adverse climate events, for example landslides. 57 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS 4. Other Approaches   in the Literature Photo credits: Wenxin Qiao 58 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE 4.1 Cost Effectiveness 4.2 Multicriteria Analysis Analysis (CEA) Approach (MCA) A cost effectiveness analysis (CEA) compares the costs Multicriteria analysis (MCA) is one practical solu- of an intervention to its expected impacts on the tar- tion for balancing all of the different factors that get beneficiaries. It is more generally applicable when impinge on the paving decision, and thus helps to the focus is on poverty and other social impacts of the determine the priority roads for upgrading. Under intervention rather than a pure economic impact. It is an MCA, points are given to criteria that are thought also appropriate when benefits cannot easily be quanti- to be important to the paving decision; then weights fied in monetary terms. In short, it is an attempt to find are applied to each of the criteria in a predetermined the least costly way of delivering the stated objectives. manner. Each road link is allocated a number of points corresponding to the fulfillment of the criteria, and The cost-effectiveness indicator (CEI) is a type of the corresponding weights are applied. The sum of cost-benefit ratio, which is calculated for each road link the weighted points provides a measure of the rank based on the following formula: for that road link in the investment decision. Using this approach, it is important for policy planners and evaluators to have a clear understanding of the most Cost of Improving critical factors, and how to justify the weights that are the Road Link assigned, so as not to bias the final decision. Once the CEI = criteria, points, and weights have been determined, Population Served case studies need to be undertaken that include edge- by Road Link case scenarios, in order to enable the correct thresh- olds to be established. Care needs to be taken to ensure that factors with relatively less importance don’t overwhelm more critical factors in the MCA. Proper The CEI is then used to prioritize proposed rural road weighting helps to address this concern. improvements. The highest priority is assigned to the road links that present the best CEI cost-effectiveness; that is, the lowest ratio of per capita capital invest- MCA EXAMPLE 1: CASE STUDY FROM RWANDA ment. Fiscal constraints can also determine where the cut-off point is on the CEI. Given the existing resource In the case of Rwanda, eight criteria have been selected constraints, a suitable threshold must be established. to gauge the effects and benefits of feeder road (FR) For example, it might be decided that no road proj- improvements. Weights are based on the overriding ect should exceed $1,000 per inhabitant, considering, policy objectives, and the socioeconomic and techni- for example, a Gross National Income (GNI) per capita cal features of the area that is traversed by the roads threshold in a low-income country context.34 under evaluation. The reliance of the evaluation/priori- tization will depend on the quality of the socioeconomic data used in this exercise. The proposed criteria select- ed for Rwanda are displayed in Table 4.1, along with the respective weights. 34 As of July 1, 2020, low-income economies are defined as those with a per capita Gross National Income (GNI) of $1,036 or less, using the World Bank Atlas method. In this case, a $1,000 threshold was adopted, based on a rounding of this number. 59 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Table 4.1: Criteria and Weights for Feeder Roads, Rwanda Priority Indicators - variation range [5,1] Road 1 2 3 4 5 6 7 8 Improve- Benefits indicator & ment Access to social Road Agriculture Community VUP ranking cost per Connectivity Remoteness Traffic & economic condition potential Priority Impact beneficiary services (weighted Indicator Weight by benefits’ indicators) Indicator USD/hab. 0.10 0.09 0.05 0.11 0.20 0.35 0.07 0.03 Ranking [5,1] Source: Technical and economic analysis of sealing options for low-volume roads and prioritization of roads for sealing, Final Report, World Bank, 2018 The results from an MCA are generally different • The indicator of the economic performance from those calculated using a traditional econom- calculated by the RED model is mainly the NPV ic analysis such as an Internal Rate of Return (IRR) per kilometer. The NPV is divided by the length of or net present value (NPV), which are focused on the road under evaluation to eliminate the influence efficiency in investment. An MCA has the advantage of the length on the NPV value. (In this exercise a of considering socioeconomic benefits in addition to discount rate of 6 percent has been adopted, as per economic efficiency. In the case of Rwanda’s feeder World Bank economic evaluation guidance). roads, the ranking of the socioeconomic performance obtained with the MCA is compared with the economic The MCA approach can be a foundation for consider- performance calculated with the RED model, drawing ing the question of whether a road should be paved, on a sample of 25 feeder roads in Rwanda. The criteria because with MCA all of the various development objec- used to compare the performances and set out the tives can be evaluated. ranking are as follows: (See Table 4.2.) • The evaluation and prioritization with the MCA will result from the weighing of effects (benefits) of FR improvement, and the cost per beneficiary of the improvement. The indicator used is the cost of the programmed intervention per beneficiary. 60 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE Table 4.2: MCA and RED Model Compared (Rwanda Case Study) Source: Technical and economic analysis of sealing options for low-volume roads and prioritization of roads for sealing, Final Report, World Bank, 2018 61 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS MCA EXAMPLE 2: CASE STUDY FROM WYOMING, USA Benchmark Engineers (2006) have proposed an Table 4.3: Score Sheet Summary (Henning et al) interesting approach that combines traditional pure Physical Factors economic analysis with a quantification of other non- economic factors. Most of their work is based on Zim- TOPOGRAPHY (GRADE) SCORE merman & Wolters (2004). They undertook a detailed Flat or Undulating area (<4%) 0 study to advise Laramie County in Wyoming using data Undulating to Hilly area (4-8%) 2 from the experience of other jurisdictions in the coun- Hilly to Mountainous (8-14%) 4 try. They found that while there are agency costs to con- Mountainous (>14%) 5 structing and maintaining a paved road, benefits accrue CLIMATE & SOIL CONDITIONS SCORE to road users from a road that allows faster speeds, Soils suitable for weather/traffic 0 and lower operating costs, and to property owners and Soils suitable for weather treated 3 residents from reduced air and water pollution, access Soils predominantly unsuitable 5 to schools, and potential enhancements of property val- Socio Economic Factors ues. They propose a step-by-step procedure for making NMT DEMAND FOR SURFACING SCORE this difficult decision based on both economic and non- Animal or NMT with low volume 1 economic factors. They walk the reader through these NMT with medium volume 3 steps, using examples and providing simplified work- NMT with high volume 5 sheets. The steps are: Step 1: Identify the road section; MOTORIZED TRAFFIC VOLUME SCORE Step 2: Determine the agency costs; Step 3: Determine < 50 1 the user costs (vehicle operating costs plus crash costs), and then scale the user costs; Step 4: Summarize total 50 - 200 3 costs; and Step 5: Evaluate noneconomic factors. In > 200 5 their case, Step 5 included political issues, growth rates, IMPACT OF DUST FORMING SCORE housing concentration/dust/erosion control, mail routes Slight 1 and industry/truck traffic, and the use of weight factors Medium 3 to arrive at the combined economic and noneconomic Severe 5 total score for each road. COMMUNITY IMPACT SCORE Slight 1 Medium 3 MCA EXAMPLE 3: GENERAL CASE STUDY Severe 5 In 2006, Henning et al developed a Three-Step Surfac- TRAFFIC INCREASE AFTER SEALING SCORE ing Alternative Decision Framework. In the first step, Unlikely 1 they assess the demand for a sealed surface, using a Some 3 model score sheet. The model evaluates whether an Likely 5 unpaved road should be upgraded to a paved one, or AVAILABILITY OF QUALITY MATERIAL SCORE the road should be left unpaved. The main factors con- Available and short hauling distance 0 sidered are topography, climate/soil condition, nonmo- Available but distance > 10km 3 torized traffic demand, motorized traffic volume, poten- Material is scarce or depleted 5 tial impact of dust, community impact, traffic increase Note: NMT = Non Motorized Traffic after sealing, and availability of quality materials. The summary spreadsheet is shown in Table 4.3. Source: Henning et al 2006 62 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE The total score ranges from 5 to 30. The authors then maintenance liability); (ii) factors relating to the physical note that the minimum score for a road to be consid- and social environment (traffic capacity, gradient sever- ered for surfacing depends on the level of development ity, opportunity for the creation of local employment, funding available for the country that is being assessed: flood resistance, dust suppression, and the use of finite they suggest a score of 12-15 for developed countries resources); and (iii) factors relating to the expected with stable funding regimes; 16-20 for developing performance of the surface (corrugations, potholes, countries with uncertain funding regimes; and a high erosion, raveling, dust, structural strength, rutting, and threshold of 21-30 for countries with severely under- roughness). funded networks. In the third and final step, they undertake a financial In the second step, they identify the surfacing alterna- and economic evaluation. The surfacing solution that tives that are available and take into account (i) factors offers the most compelling case from the financial and relating to construction and maintenance (design economic perspective is then adopted. standards, required production and laying equip- ment, required imported materials, skill level required, The three-step process is summarized in Figure 4.1. Figure 4.1: Three-Step Paving Framework Source: Henning et al 2006 63 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS MCA EXAMPLE 4: KANSAS CASE STUDY Dissanayake and Patel (2016) developed gravel-road classes of vehicles; the gradient and horizontal curva- paving guidelines based on research in the state of ture of the road; and the conversion factor for the costs Kansas. The guidelines were based on the MCA meth- of paved surface versus gravel surface roads. od. They identified the key factors in the decision mak- ing process as agency cost, safety, vehicle operating The final score is calculated by multiplying the weights costs (VOC), traffic volume, purpose of road usage, and of each factor and their respective scaled values. The public preference. The MCA method involves calculat- surface type with a higher score is the preferred alter- ing the weights for the factors that are important in native for a road section under consideration. A com- decision making; obtaining respective scaled values for puter-based Gravel Road Paving Guidelines program each factor for paved and gravel surfaces; and eventu- was created as a user interface, using Visual Studio to ally calculating the final score for both paved and gravel carry out all of the complex calculations for determin- surface types. Equations were formulated to carry out ing the LCC and VOC considering local variations. The the life-cycle cost (LCC) analysis along with the present program also helped to determine final total scores for worth evaluation, which provided flexibility for calcu- paved and gravel roads by considering scaled values of lating agency costs while considering local conditions. the all-important factors considered for conversion. The VOC could be calculated for both paved and gravel roads, considering variations in the speeds of different A screenshot of the program is shown in Figure 4.2. Figure 4.2: Screenshot of Kansas Paving Program Source: Kansas DOT 2016 64 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE 4.3 Summary of Approaches In summary, it can be surmised that there are basically solution to the paving decision problem. This report three approaches to the problem of economic evalu- presents a novel systematic approach based on a ation: a pure economic analysis (PEA); a cost-effective- hybrid of MCA, PEA, and CEA, but it hinges on the MCA. ness analysis (CEA); and a multicriteria Analysis (MCA) The proposed approach builds especially on the work (Figure 4.3). Given the preponderance of factors that of Benchmark Engineers (2006), Henning et al (2006), impinge on the paving decision, the MCA approach and Dissanayake & Patel (2016). offers the best hope of a meaningful and workable Figure 4.3: Three Approaches to the Economic Evaluation Problem of Paving 65 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS 5. Decision Framework:   To Pave or Not To Pave? Photo credits: Wenxin Qiao 66 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE 5.1 Critical Questions affecting the Paving Decisions Given constrained budgets, road agencies need to the alternative that maximizes the net present value identify the best use of resources at the program (NPV) of the expected net benefits, or the least-cost level. The main rationales warranting public action for alternative among the options for producing a spe- any project include the correction of market failures; cific set of expected results, unless explicit reasons the incorporation of externalities or spillovers; and are given for choosing another alternative. redistributional, social, and political concerns. Given the scarcity of public funds and implementation capacities, From a road agency perspective, five critical ques- public investment in new or ongoing programs con- tions derive from this guidance: tributes most to development when those programs are clearly directed to addressing one or more of these First Question: Should a road agency implement a concerns (World Bank 2013). Left to their own devices, paving program in this country? National or regional almost unanimously the preference of target beneficia- development plans should identify rural road improve- ries is to receive a paved road over an unpaved option. ments as a critical component of economic develop- A systematic but fair approach is therefore warranted ment or poverty eradication efforts. Ideally, the decision to ensure the optimal use of available resources. process would be guided by a rural road strategy and implementation plans, following a robust asset man- According to the World Bank Economic Guidance 2013, agement system (AMS), and would be accompanied by three critical considerations must be met when consid- relevant policies, basic design standards (both geo- ering the impacts of a project that is being put forward metric and pavement), materials specifications, and for financing: a commitment to the funding and sustaining of road maintenance practices, to ensure that the investment is • It is expected to contribute to development if the not wasted. expected benefits justify the expected cost, includ- ing benefits that can be realistically stated in mone- Second Question: Which roads are suitable to be tary terms, or other benefits that are more difficult considered for paving? The proposed SPADE model to monetize (or sometimes even to quantify) but attempts to answer this question by adopting an inno- that can be demonstrated to be important project vative multicriteria analysis (MCA) that seeks to put a outcomes. number (a paving priority score, or PPS) to a confluence of many determinant factors. • The expected benefits and costs attributable to the project are measured by comparing the situation Third Question: Which roads are economically jus- with the project to the situation without the project tified for paving? The traditionally-used RED model (the counterfactual, or a baseline). seeks to answer this question. Road investments that result in a positive NPV are economically justified for • Where plausible alternatives exist, the selected proj- paving. Cost-effectiveness analyses (CEAs) are also ect should be shown to be the preferred design. In used to justify value for money when the RED mod- other words, assessment of the development impact el is not considered appropriate, especially where should involve a serious consideration of serious social and other concerns overwhelm pure economic alternatives. Normally, the selected project would be considerations. 67 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Fourth Question: What pavement technology should A. Country Context be used (surfacing type, mode of delivery, contracting • A1: Country GDP per capita (US$) current approach)? Chapter 6 of this report provides guidance • A2: Annual funds available for the rural road network on narrowing down the pavement technologies to • A3: Country Rural Accessibility Index consider. The SPADE model proposes an approach for choosing among these technologies for low-priority, • A4: Share of population (rural) medium-priority, and high-priority roads. • A5: Percentage of paved roads in the country • A6: Country disaster risk exposure Fifth Question: Given the financial constraints and overriding policy objectives, how should the pav- B. Region Context ing program be implemented? The results from the • B1: Agricultural Production SPADE-PLUS approach (which includes the SPADE mod- • B2: Other economic or strategic purpose (tourism, el, the RED model, and CEA) dictate the priority rural industrial, security, cultural) road investments as well as the paving technologies to be implemented. Financial constraints and overrid- • B3: Poverty Incidence ing policy objectives will determine the cut-off point for which roads should be invested in, and when such C. Operational Environment investments should be undertaken. • C1: Maintenance practices: - is there continuous routine maintenance for the road network, and is adequate and timely allocation made for both rou- 5.2 Presenting the tine and periodic maintenance?) • C2: Rural roads policy, strategy, and plan: - do these Spade Model documents exist to guide the planning and invest- ment process?) In an extensive review of the literature, several factors • C3: Asset Management System (AMS). (Is there an keep recurring as the most critical ones in the decision AMS for the area of interest that includes a road of whether or not to pave a road. This work considers inventory, condition, and prioritization framework?) these factors in five buckets, cascading down from the • C4: Design standards. (Are geometric and pavement country context to the regional context, the operational design standards for rural roads in place?) environment, and the road and engineering contexts; • C5: Materials specifications. (Are materials specifi- and develops a systematic approach that factors all of cations available to guide quality material selection these relevant dimensions into the decision-making and construction?) process. D. Road Context The matrix of these categories, and the factors under each of them, form the framework for the Systematic • D1: Traffic (non-motorized): Number of NMT users Paving Decision (SPADE) Model, as summarized below. per day on road: Low <50, Medium 50-200, High > For further details, refer to the SPADE model itself, and 200 to the SPADE Model User guide. 68 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE • D2: Public transport: Public transport vehicles per • D16: Public Preference (Has the road section been day (any public transport vehicles, not just buses): identified as among the priority sections for paving Low <5, Medium 5-20, High >20 by the local population and authorities?) • D3: Commercial transport (vehicles carrying goods or produce for commercial purposes): Low <20, E. Engineering Context High > 20 • D4: Population served: Number of people served • E1: Drainage (Is proper drainage provided, or will it by the road within a 2-kilometer (km) buffer: Low be easily integrated into the design?) <1000, Medium 1000-5000, High >5000 • E2: Structures for connectivity (Will there be a need • D5: Schools (within 5 km buffer of road) for crossings or hydrological structures?) • D6: Hospitals/Health Centers (within 5km buffer of • E3: Materials availability (Are quality gravel materi- road) als available within a reasonable hauling distance (<10km)? • D7: Markets (within 5 km buffer of road) Low 1 or less, Medium 2-5, High >5 • E4: Climate (Low rainfall <500mm/yr, Medium 5000- 1500mm/yr, High > 1500mm/yr) • D8: Administration Centers (within 5 km buffer of road) • E5: Flooding risk due to low-lying road/plain/ embankment 9: Dust nuisance to road users and roadside prop- • D erties (How severe is the dust issue in the dry sea- • E6: Traffic (motorized) Low <50 vpd, Medium 50-200 son for road users and roadside properties?) vpd, High >200 vpd • D10: Road Connectivity (To higher classification level • E7: Prevailing terrain i.e. majority classification (flat, road) rolling, mountainous). Cross-slope 0-10% Flat, 10%- 25% Rolling, >25% Mountainous. • D11: Redundancy (If a major link is disrupted, is this road one of viable alternatives?) • E8: Right of Way (RoW) Issues: Is land for road con- struction available without RoW issues? • D12: Are there women and girls carrying loads on their heads along the project road during a normal • E9: Existing Subgrade: - quality of existing subgrade: day? CBR <15 weak, CBR 15-25 Fair, CBR > 25 Good • D13: Road Safety (Have the most likely increased • E10: Condition of the road Poor IRI >10; Fair IRI 5-9; crash risks (from higher speeds) been identified? Good IRI <5 Have plans been made for how to mitigate these • E11: Estimated paving cost as a multiple of the cost risks through adequate infrastructure protection for of gravel. road users (including vulnerable ones)? • D14: Environmental Sensitivity (park/reserve, flora, These factors are presented more systematically in fauna): Does the road cross or is it within a 2 km buf- Figure 5.1. fer of sensitive ecological areas? • D15: Functional Classification of the Road (primary, secondary, tertiary) 69 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Figure 5.1: Factors Considered in the SPADE A. Country Context Total Score • A1: Country GDP per capita (US$) current • A4: Share of population (rural) A • A2: Annual funds available for the rural A5: Percentage of paved roads in the country • road network A6: Country Disaster Risk Exposure • • A3: Country Rural Accessibility Index B. Region Context Total Score • B1: Agricultural Production • B2: Other economic or strategic purpose (tourism, industrial, security, cultural, trade) B • B3: Poverty Incidence C. Operational Environment Total Score • C1: Maintenance practices • C4: Design Standards • C2: Rural roads policy, strategy, and plan • C3: Asset Management System - AMS • C5: Materials Specifications C D. Road Context Total Score • D1: Traffic (non-motorized) • D2: Public transport: Public transport vehicles per day • D3: Commercial transport (Vehicles carrying goods or produce for commercial purposes) • D4: Population served: Number of people served by road within 2km buffer D5: Schools (at least 1 in 5km buffer of road) • D6: Hospitals/Health Centers (at least 1 in 5km buffer of road) • D7: Markets (at least 1 in 5km buffer of road) • D D8: Administration Centers (at least 1 in 5km buffer of road) • D9: Dust nuisance to road users and roadside properties to road users and roadside properties?) • D10: Road Connectivity (To higher level road) • D11: Redundancy (if major link is disrupted, is road one of viable alternatives?) • D12: Number of women and girls carrying loads on their heads on a normal day • D13: Road Safety • D14: Environmental Sensitivity • D15: Functional Classification of the Road (Primary, Secondary, Tertiary) • D16: Public Preference • D. Road Context Total Score E1: Drainage • E7: Prevailing Terrain • E2: Structures for connectivity • E8: Right of Way (RoW) Issues • E E3: Materials Availability • E9: Existing Subgrade • E4: Climate (Rainfall) • E10: Condition of the road • E5: Flooding Risk • E11: Estimated paving cost (multiple of the cost • E6: Traffic (motorized) • of gravel) 70 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE It is worth reconciling the seven key factors that were noted as being critical to the paving decision, versus the parameters that the SPADE model proposes to incorporate. This reconciliation is presented in Table 5.1. Table 5.1: Reconciliation of Key Paving Factors against the SPADE Bucket Parameters Key Factor vs. SPADE Country Region Operational Road Context Engineering Bucket Context Context Environment Context Fiscal and A1, A2 Distributional Constraints Economic and Strategic A4, A5 B1, B2, B3 Imperatives Institutional, Network A3 C1, C2, C3, C4, D10, D15 and System Dynamics C5 Differentiated needs of D1, D2, D3, D4, various road users and D5, D6, D7, D8, target beneficiaries D9, D12, D16 Engineering concerns E2, E3, E6, E7, E8, E9, E10, E11 Climate change A6 D11 E1, E4, E5 considerations Other Externalities D13, D14 The country context, region context, operational For a PPS below 50, there is no compelling justifica- environment, road context, and engineering con- tion to pave the road under evaluation. The recom- text are all factored in as part of the SPADE decision mendation is “Do nothing,” or explore cost-effec- process, as illustrated in Table 5.1, and in the SPADE tive alternatives like spot regravelling, or drainage model itself. Scores are assigned for each of the fac- improvements. tors in the SPADE model using the scales provided, with each being noted as normal (single weighting), critical For a PPS between 51 and 70, low-cost paving (triple weighting), or super-critical (quintuple weighting). options like chip seals, Otta seals, cobblestones, or The model then combines the scores from the country similar surfaces should be evaluated. context, region context, the operating environment, and road and engineering context for a total weight- For a PPS between 71 and 100, there is a compelling ed score, with scaled percentages of 10 percent, 10 justification to pave, and the full menu of paving percent, 10 percent, 40 percent and 30 percent being options should be considered. assigned to the buckets respectively. Total weighted scores are thus obtained to generate what is termed a From the SPADE model runs that have been undertak- Paving Priority Score (PPS). en (both on real road sections, and using hypothetical case studies), most roads score a PPS of 40-85, with scores outside this range considered outliers. 71 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS 5.2 The Spade-Plus Approach With its combined use of a multicriteria analysis (MCA) In Stage 2, for roads that have passed the SPADE and an economic analysis, and with its innovative per- prioritization for paving with a medium or high mutations for circumventing the challenges of tradition- priority rating, an economic analysis is undertak- al approaches, the proposed two-stage SPADE-PLUS en using either a RED model analysis or a CEA. The approach offers an improvement over conventional proposed bifurcation is to use the RED model for the approaches. economic evaluation of higher-traffic roads (>200 AADT) and a CEA for lower-traffic roads (<200 AADT). In the first stage, the innovative MCA SPADE model is applied. In the second stage, the economic analysis is Undertake a RED model analysis (AADT > 200) by com- undertaken, using either a cost-benefit analysis (CBA) paring different paved surfacing alternatives in order (the RED model is recommended), or a cost effective- to determine the most economically justifiable option. ness analysis (CEA). This two-stage process offers an This report emphasizes, however, the need to use optimal solution to the paving conundrum by paying the RED model with realistic assumptions about the due attention to both easy-to-quantify, and not-so-eas- maintenance regime that will be followed in prac- ily quantifiable parameters in a defensible manner. The tice. The paving options to be evaluated in the RED SPADE PLUS approach is represented graphically in model will depend greatly on local conditions, policy Figure 5.2 below. objectives, and cost. The different surfacing options and their pros and cons, as well as their suggested use are In Stage 1, the PPS is calculated through the SPADE discussed more fully in Chapter 6. model, as described above. Or they can undertake a CEA (for roads with an AADT < Next, a check is made on the total weighted 200) by comparing the investment costs of three paving score (the PPS) to determine the way forward. As options divided by the target number of beneficiaries. described above, for roads with a low score (an overall Where the resulting cost per beneficiary is less than weighted score at or below 50 percent), there is no $1,00035 or its equivalent, the investment is justified. In compelling justification to pave the road under eval- comparing the three options, the option with the lowest uation. The recommendation is to do nothing, or to cost per beneficiary is the one that is recommended for explore cost-effective alternatives like spot regravelling, adoption. or drainage improvements. For roads with a medium score (an overall weighted score from 51-70 percent), Undertaking a SPADE model analysis before the low-cost paving options like chip seals, otta seals, cob- RED model/ CEA analysis is important for three blestones, or similar surfaces should be evaluated. For main reasons: roads with high scores (an overall weighted score over 70 percent), there is a compelling justification to pave, 1. Screening purposes: The SPADE model helps to and the full menu of paving options is available. weed out nonviable candidate roads at the very beginning and stop them from being fronted in an unfounded economic analysis. 35 As of July 1, 2020, low-income economies are defined as those with a Gross National Income (GNI) per capita, calculated using the World Bank Atlas method, of $1,036 or less. A $1,000 threshold is adopted based on a rounding of this number. 72 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE 2. Multiplicity of key considerations: The paving decision is not straightforward. As has been men- tioned previously, a multitude of factors impinge on the decision, yet many of these cannot easily be translated into monetary values. Using the SPADE model allows these critical and sometimes divergent factors to be considered before an analysis of sur- facing options is evaluated on an economic basis. 3. Narrowing surfacing alternatives: Since the RED model analysis can only compare three options at a time, the initial SPADE model analysis helps narrow down the possible choices of surfacing alternatives from do nothing to gravel to low cost paving to conventional paving options like asphalt and gravel based on the SPADE model threshold results. Undertaking the traditional economic approach without making use of the SPADE model outlined in Stage 1 will result in running into the same chal- lenges that have been fully described throughout this report. On the other hand, undertaking the SPADE mod- el analysis on its own, without the second step of economic justification using the RED model or CEA, will result in an incomplete economic evaluation exercise. Photo credits: Trevor Samson 73 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Figure 5.2: SPADE-PLUS Approach Framework Country Context (Score A) Raw Score Range 6-30 points Region Context (Score B) Raw Score Range 2-15 points Operational Environment (Score C) Raw Score Range 0-35 points Road Context (Score D) Raw Score Range 17-210 points Stage 1: SPADE Model Analysis Engineering Context (Score E) Raw Score Range 26-140 points Weight A by 10 percent, B by 10 percent, C by 10 percent, D by 40 percent and E by 30 percent to get the Paving Priority Score (PPS) Low Score Medium Score High Score (Below 50 percent) (51-70 percent) (Above 70 percent) No /Limited justifica- Medium justification Compelling justifica- tion to pave. Do to pave. Explore low tion to pave. Explore nothing or explore cost paving options different paved minimal cost-effective surfacing alternatives interventions including to determine the spot regravelling or most economically drainage improvements justified option YES Is AADT NO > 200? Use Cost-Benefit Analysis Use Cost Effectiveness Analysis e.g. Red Model 1. Identify a minimum of three 1. Adopt justified discount rate based paving solutions, and estimate on World Bank Economic Analysis the investment costs for the Stage 2: Economic Analysis Guidance and country economic road under each option conditions 2. Run a Cost Effectiveness 2. Use the maintenance regime that Analysis (Cost of the Interven- is prevalent in the region of tion divided by the number of analysis not what is theoretically target beneficiaries) expected to happen 3. If the Cost per beneficiary is less 3. Evaluate Do Nothing option against than US$1000, the paving two paving alternatives (choose solution is economically justified alternatives from feasible locally available materials solutions) 4. If all three paving options are 4. Run RED Model for each road economically justified, choose under evaluation. the option that offers the least 5. Adopt paving solution that offers cost per beneficiary. highest NPV 74 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE 5.2 Spade Model Case Study Results The SPADE-PLUS model that has been put forward in To further validate the SPADE model, a worst-case this work has undergone an iterative process of calibra- scenario was also investigated, in which a hypothetical tion and final validation with case studies and hypothet- edge-case road was tested at the lower boundary. ical testing. Weights, thresholds, and other indicators and parameters have been designed and adjusted with The results of the application of the proposed SPADE- reference to work done in past studies in the literature, PLUS approach is presented below in the discussion of and with the use of Delphi-Techniques, which involves the case studies. This evaluation aimed at pointing out running model parameters through an iterative consul- the synergies and benefits of the sequential use of the tation process with sector experts. two models (SPADE and RED), and of the CEA where annual average daily traffic (AADT) is less than 200. The calibration and validation case studies cover road sections selected from six countries in three continents: It is worth noting that some cells are highlighted in Ethiopia, Mozambique, Rwanda, and Tanzania in Africa; red, to show where the RED model analysis indicated a Laos in Asia; and Nicaragua in South America. A total negative NPV, but sealing was still recommended, with of 54 sample rural roads from different programs were the understanding that it was still better than the grav- selected for testing, and evaluated using input data el option. These are the kinds of challenges that the drawn from available feasibility studies (See Table 5.2). SPADE-PLUS approach is designed to overcome. Table 5.2: Summary of SPADE-RED Test Results for 54 Roads RED NPV@12% RED NPV@6% NPV/ NPV/ NPV km km Spade SPADE rec- 000’ 000’ Best NPV 000’ 000’ Best Country Road Name km AADT PPS ommends USD USD Alternative USD USD Alternative Rwanda KRFR01 24 152 82 Pave (H) 90 4 Sealing 1931 81 Sealing Rwanda KRFR03 2 66 67 Pave (M) -77 -35 Sealing 49 22 Sealing Rwanda KRFR14 22 105 76 Pave (H) -278 -13 Sealing 1395 65 Sealing Rwanda NBFR01 14 96 71 Pave (M) -768 -56 Gravel 157 11 Sealing Rwanda NBFR02 33 99 66 Pave (M) -2873 -88 Gravel -1509 -46 Sealing Rwanda NBFR03 6 31 68 Pave (M) -155 -26 Gravel 245 40 Sealing Rwanda NBFR04 14 57 65 Pave (M) -1390 -102 Gravel -716 -52 Sealing Rwanda NBFR05 49 154 76 Pave (H) 3279 67 Sealing 8768 180 Sealing Rwanda NBFR06 2 50 62 Pave (M) -204 -89 Gravel -169 -73 Sealing Rwanda NBFR07 5 27 62 Pave (M) -433 -82 Gravel -217 -41 Sealing Rwanda NBFR08 6 36 60 Pave (M) -1063 -166 Gravel -1028 -161 Sealing 75 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS RED NPV@12% RED NPV@6% NPV/ NPV/ NPV km km Spade SPADE rec- 000’ 000’ Best NPV 000’ 000’ Best Country Road Name km AADT PPS ommends USD USD Alternative USD USD Alternative Rwanda NBFR09 9 13 64 Pave (M) -1267 -137 Gravel -950 -103 Sealing Rwanda NBFR10 6 34 57 Pave (M) -763 -120 Gravel -534 -84 Sealing Rwanda NBFR11 6 62 72 Pave (H) -325 -54 Gravel -52 -9 Sealing Rwanda NBFR12 3 59 56 Pave (M) -135 -47 Sealing 53 18 Sealing Rwanda NBFR13 5 52 71 Pave (H) -611 -120 Gravel -432 -85 Sealing Rwanda NBFR14 11 124 70 Pave (M) -488 -43 Sealing 317 28 Sealing Rwanda NBFR15 14 41 70 Pave (M) -1612 -113 Gravel -1299 -91 Sealing Rwanda NBFR16 16 77 65 Pave (M) -904 -58 Gravel 47 3 Sealing Rwanda NBFR17 1 62 60 Pave (M) -270 -284 Gravel -252 -265 Sealing Rwanda NGFR01 36 205 65 Pave (M) 606 17 Sealing 2962 82 Sealing Rwanda NGFR02 5 201 68 Pave (M) 84 17 Sealing 409 82 Sealing Rwanda NGFR03 18 518 67 Pave (M) 1887 104 Sealing 3930 218 Sealing Rwanda NGFR04 10 410 61 Pave (M) 625 64 Sealing 1506 154 Sealing Rwanda NGFR05 19 173 58 Pave (M) 69 4 Sealing 1239 65 Sealing Rwanda NGFR06 16 191 65 Pave (M) -204 -13 Gravel 553 34 Sealing Rwanda NGFR07 16 169 67 Pave (M) 393 25 Sealing 1122 71 Sealing Rwanda NGFR08 9 125 60 Pave (M) 104 12 Sealing 642 73 Sealing Rwanda NGFR09 11 105 56 Pave (M) -127 -12 Sealing 390 36 Sealing Rwanda NGFR10 18 109 66 Pave (M) 321 17 Sealing 1535 83 Sealing Rwanda NGFR11 1 127 57 Pave (M) -54 -45 Sealing -17 -14 Sealing Rwanda NGFR12 10 379 54 Pave (M) 588 61 Sealing 1076 112 Sealing Rwanda NGFR13 9 215 63 Pave (M) -370 -42 Gravel -37 -4 Sealing Rwanda NGFR14 6 91 53 Pave (M) -187 -30 Sealing 68 11 Sealing Tanzania Iringa 34 693 82 Pave (H) 17319 509 AC Sealing 33255 978 AC Sealing Tanzania Mtili-Mkuta 14 976 71 Pave (H) 5732 409 Sealing 7038 503 Sealing Tanzania Wenda-Mgama 19 1977 71 Pave (H) 4902 258 Sealing 8407 442 Sealing Mozambique NAMPULA - R686 51 735 73 Pave (H) N/A N/A N/A N/A N/A N/A Mozambique NAMPULA - R696 24 464 82 Pave (H) N/A N/A N/A N/A N/A N/A 76 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE RED NPV@12% RED NPV@6% NPV/ NPV/ NPV km km Spade SPADE rec- 000’ 000’ Best NPV 000’ 000’ Best Country Road Name km AADT PPS ommends USD USD Alternative USD USD Alternative Mozambique NAMPULA - R1156 55 303 73 Pave (H) N/A N/A N/A N/A N/A N/A Mozambique NAMPULA - R1170 38 261 49 Don’t Pave N/A N/A N/A N/A N/A N/A (L) Mozambique NIASSA - N360 67 624 82 Pave (H) N/A N/A N/A N/A N/A N/A Mozambique NIASSA - R720 64 360 84 Pave (H) N/A N/A N/A N/A N/A N/A Ethiopia Arsi Robe – Wabe 53 89 70 Pave (M) N/A N/A N/A N/A N/A N/A River Bridge Ethiopia Goge Jore - Akobo 120 146 64 Pave (M) N/A N/A N/A N/A N/A N/A Ethiopia Indiber-Qanite-Qebul 58 200 77 Pave (H N/A N/A N/A N/A N/A N/A Nicaragua Cardenas-Colon 10 218 75 Pave (H) 740 74 Sealing N/A N/A N/A Nicaragua Emp.La Mora-La 25 256 80 Pave (H) 3590 144 Sealing N/A N/A N/A Carpa Nicaragua Emp. Mina El Limon 15 416 74 Pave (H) 4080 272 Sealing N/A N/A N/A Nicaragua Granada-Malacatoya 10 452 79 Pave (H) 4830 483 Sealing N/A N/A N/A Nicaragua Quebranda Hon- 10 256 80 Pave (H) 1920 192 Sealing N/A N/A N/A da-San Francisco Libre Nicaragua Rio 10 348 82 Pave (H) 3550 355 Sealing N/A N/A N/A Blanco-Mulukuku Nicaragua Rivas-Veracruz 8 633 78 Pave (H) 3000 375 Sealing N/A N/A N/A Laos Khammouane 14 132 67 Pave (M) 58 4 Sealing 646 46 Sealing *L-Low; M-Medium; H-High The SPADE PPS ranges from 49 to 84 over the 54 recommendation not to pave); 32 roads (59 percent) selected roads, with a mean score of 68; a median for Medium Priority (low-cost paving solution recom- score of 68; and a standard deviation of 9. mended); and 21 roads (39 percent) for High Priority (all paving solutions possible). The SPADE model recommends that 53 out of the 54 case study roads be paved (98 percent). The distri- The distribution of the SPADE model recommendations bution is: 1 road (2 percent) for Low Priority (with a is shown graphically in Figure 5.3. 77 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Figure 5.3: Distribution of SPADE Model In the RED model tests, three road improvement Recommendations from 54 Case Studies alternatives were tested: Alt.1: Rehabilitation to gravel 80% standard, and widening to 6-meter roadway; Alt.2: 59% Upgrading to 6-meter width and double bituminous 60% 39% surface treatment (DBST); Alt.3: Upgrading to 6-meter 40% width and asphalt concrete surfacing. (Except in the 20% case of Nicaragua, where the concrete block (adoquines) 2% alternative was considered instead of the DBST.) RED 0% analysis data was not available for the Ethiopia and Low Medium High Mozambique case studies. For the 45 roads that also ran a traditional RED model The RED simulations used two NPV discount rates, set analysis at a 12 percent discount rate, the SPADE model at 6 percent and 12 percent respectively: the first one recommended that all 45 roads be paved (100 percent), is more oriented to taking into consideration the long- while the traditional RED model recommended that term intergenerational equity, and the second one is only 38 out of the 45 be paved (84 percent). more apt to emphasize economic performance in the short or medium term. The Nicaragua test runs used The paving positivity comparison between the SPADE only the 12 percent discount rate. The choice of the model results and the RED model results at a 12 per- discount rate was found to have a significant influ- cent discount rate is shown in Figure 5.4. ence on the outcome of the RED evaluation. At the 6 percent discount rate, 25 roads out of the 38 examined were judged economically feasible (NPV≥0) for paving Figure 5.4: Comparison of SPADE and RED Model improvement. (This did not include Ethiopia, Nicaragua, Paving Positivity Results* and Mozambique, which didn’t have RED results for 105% either the 6 percent or the 12 percent discount rates). 100% At the 12 percent discount rate, only 15 roads were 100% judged economically feasible for paving. 95% 90% The results show that the SPADE PLUS model is 84% more aligned with the NPV settled at 6 percent. 85% Also, it appears that in the case of Rwanda, some roads 80% are indicated for sealing, yet they have a negative NPV 75% (even at the 6 percent discount rate). The SPADE-PLUS SPADE RED model approach releases the policy maker from the *From 54 case studies, at 12 percent discount rate. conundrum of justifying this decision. 78 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE In the RED model, the way bicycle and motorcycle traffic centers in selected districts. This project funded 270 is handled—that is, whether or not it is included in the kilometers of rural roads in four districts for rehabilita- AADT—has a major impact on results. In the traditional tion, upgrading, and maintenance. All of the road sec- approach, this traffic is neglected. Some RED models tions under this project were graveled roads. include this traffic in the AADT, and others do not. In calculations of benefits and costs, the variation of crash In 2017, the project got additional funding from a rates has also not been included, partly due to the lack Multi-Donor Trust Fund (MDTF), and was expanded to of road crash statistics, but also because claiming safety cover an additional six districts, for a total of 450 kilo- benefits when a road is paved is hard to justify if prop- meters for upgrading, rehabilitation, and maintenance, er mitigation measures, especially against speeding, of which 250 km were to be paved. The feeder road (FR) are not introduced, and the provision of facilities for sections selected for paving were based on a feasibility nonmotorized transport are not provided. The reduc- study with an MCA approach that considered terrain, tion in GHG emissions has not been included, since the population, climate considerations, and IRR. Roads low volume of traffic makes the incidence of emissions with a high IRR were selected for paving, with the final reduction on the overall economic performance by pro- selection also being based on availability of funds. It posed road improvement alternatives negligible. The was found that roads in mountainous terrain ought to RED model sensitivity analysis highlights the fact that be paved given the negative impacts from erosion and the main determinants of the economic performance traffic if they were not sealed. include traffic volumes, construction and maintenance costs, and the improvement of roughness (that is, the In 2018, after completion of the rehabilitation works, difference between IRI before and after paving). For several sections of a gravel road (FR1) in one of the dis- most of the roads evaluated, the alternative of surfac- tricts (Karongi) was severely damaged after flash flood- ing with DBST or other low-cost paving solutions was ing during the rainy season. To enhance the resilience judged as the best-performing. of the roads in Rwanda, sealing studies had started for the roads during the second phase of the project, In this section, the case studies are discussed in further under the MDTF, for various paving options, and sev- detail. eral paving trials were then conducted in the Karongi district (about 1.8 km in length). Selected roads from this study will be sealed with various paving options; the RWANDA CASE STUDY works were to commence in July 2020 and paving was planned to be mainly DBST, which has proved to have a In 2014, the government of Rwanda launched a pro- higher IRR than other options. gram for the improvement of rural roads known as the Feeder Roads Development Project (FRDP), which For this case study, feeder roads were selected from is financed through the World Bank, with Internation- three typical districts: Nyagatare, which has flat-to- al Development Association (IDA) credits, to enhance rolling terrain, and the Karongi and Nyabihu districts, all-season road connectivity to agricultural market which have mountainous terrain. A sample of 14 FRs in 79 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS the Nyagatare district, along with a sample of 3 FRs in Some of the other key findings were: Karongi and 17 FRs in Nyabihu, have undergone a com- bined evaluation with the SPADE and RED models in 1. SPADE gave medium priority to pave; the main order to make recommendations on whether to pave. contributors included the flat to rolling terrain, the In brief, the SPADE model gave recommendations that medium connectivity to markets/school/hospital, are strongly in accordance with the project decisions and traffic volume, as well as the population served. for the second phase of the project. SPADE provided a very high PPS score for FR1 in 2. The main drivers of economic feasibility were the Karongi (which was constructed as a gravel road volume of traffic (specifically, the large number of and was seriously damaged during flash floods). motorcycles, coupled with a significant amount of Using SPADE this road could have been justified for nonmotorized traffic (NMT); paving in the first place, and with better drainage and slope stabilization it might possibly have fared 3. The moderate cost of road works, mainly due to the better during the floods. availability of construction materials in adjoining areas; The detailed results are summarized below. 4. The NPV discount rate plays a discriminatory role Nyagatare district (flat-to-rolling terrain). The screen- for projects that have economic performance lev- ing of the 14 FRs, made with the SPADE model, has els falling in the transition range from feasibility to shown that all of the FRs have overall PPS scores from unfeasibility. 50-70, indicating a medium priority to pave (low-cost paving solutions needed). The RED model results indi- The results of the SPADE and RED models applied to cated that two of the selected roads were economically the FRs of Nyagatare district are shown in Table 5.3. unfeasible for sealing if the NPV was set at 6 percent, and five of them were economically unfeasible if the NPV was set at 12 percent. The project decision for these roads was for DBST paving, irrespective of the NPV results. 80 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE Table 5.3: Summary of SPADE-RED Model Results for Roads in Nyagatare District, Rwanda RED NPV@12% RED NPV@6% Length Spade SPADE NPV 000’ NPV/km Best NPV 000’ NPV/km Best Road ID (km) PPS recommends USD 000’ USD Alternative USD 000’ USD Alternative NGFR01 36 65 Pave (Medium) 606 17 Sealing 2962 82 Sealing NGFR02 5 68 Pave (Medium) 84 17 Sealing 409 82 Sealing NGFR03 18 67 Pave (Medium) 1887 104 Sealing 3930 218 Sealing NGFR04 10 61 Pave (Medium) 625 64 Sealing 1506 154 Sealing NGFR05 19 58 Pave (Medium) 69 4 Sealing 1239 65 Sealing NGFR06 16 65 Pave (Medium) -204 -13 Gravel 553 34 Sealing NGFR07 16 67 Pave (Medium) 393 25 Sealing 1122 71 Sealing NGFR08 9 60 Pave (Medium) 104 12 Sealing 642 73 Sealing NGFR09 11 56 Pave (Medium) -127 -12 Sealing 390 36 Sealing NGFR10 18 66 Pave (Medium) 321 17 Sealing 1535 83 Sealing NGFR11 1 57 Pave (Medium) -54 -45 Sealing -17 -14 Sealing NGFR12 10 54 Pave (Medium) 588 61 Sealing 1076 112 Sealing NGFR13 9 63 Pave (Medium) -370 -42 Gravel -37 -4 Sealing NGFR14 6 53 Pave (Medium) -187 -30 Sealing 68 11 Sealing Nyabihu district (mountainous terrain). The screening were set at 6 percent. The project decision for these made with the SPADE model of the 17 FRs shows that roads was for DBST paving despite the justification 14 of them were given medium priority to pave (low- issues presented by the NPV results. cost paving solutions), and 3 of them (those with a PPS score from 55-80) were given high priority (all paving Other findings include: solutions). In other words, all of the roads were recom- mended for paving. The RED model results indicated 1. SPADE gave medium priority to pave: the main con- that only 1 of the selected roads was economically fea- tributors were the mountainous terrain, medium sible for sealing if the NPV were set at 12 percent, and connectivity to markets/schools/hospitals, traffic vol- 6 of them were judged economically feasible if the NPV ume, and the medium level of the population served. 81 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS 2. The RED results indicated the economic unfeasibility 3. The NPV discount rate played a discriminatory role of sealing when using the 12 percent discount rate for projects with an economic performance falling in that affects the majority of the Nyabihu FRs, due to the transition range from feasibility to unfeasibility. the combined effect of the high cost of construc- tion in the mountainous environment of this district, Table 5.4 summarizes the results of the SPADE and RED coupled with the general unavailability of appropri- models applied to the FRs of the Nyabihu district. ate construction materials. The RED model predom- inantly recommended improving the roads to a gravel standard. Table 5.4: Summary of SPADE-RED Model Results for Roads in Nyabihu District, Rwanda RED NPV@12% RED NPV@6% Length Spade SPADE NPV 000’ NPV/km Best NPV 000’ NPV/km Best Road ID (km) PPS recommends USD 000’ USD Alternative USD 000’ USD Alternative NBFR01 14 71 Pave (Medium) -768 -56 Gravel 157 11 Sealing NBFR02 33 66 Pave (Medium) -2873 -88 Gravel -1509 -46 Sealing NBFR03 6 68 Pave (Medium) -155 -26 Gravel 245 40 Sealing NBFR04 14 65 Pave (Medium) -1390 -102 Gravel -716 -52 Sealing NBFR05 49 76 Pave (High) 3279 67 Sealing 8768 180 Sealing NBFR06 2 62 Pave (Medium) -204 -89 Gravel -169 -73 Sealing NBFR07 5 62 Pave (Medium) -433 -82 Gravel -217 -41 Sealing NBFR08 6 60 Pave (Medium) -1063 -166 Gravel -1028 -161 Sealing NBFR09 9 64 Pave (Medium) -1267 -137 Gravel -950 -103 Sealing NBFR10 6 57 Pave (Medium) -763 -120 Gravel -534 -84 Sealing NBFR11 6 72 Pave (High) -325 -54 Gravel -52 -9 Sealing NBFR12 3 56 Pave (Medium) -135 -47 Sealing 53 18 Sealing NBFR13 5 71 Pave (High) -611 -120 Gravel -432 -85 Sealing NBFR14 11 70 Pave (Medium) -488 -43 Sealing 317 28 Sealing NBFR15 14 70 Pave (Medium) -1612 -113 Gravel -1299 -91 Sealing NBFR16 16 65 Pave (Medium) -904 -58 Gravel 47 3 Sealing NBFR17 1 60 Pave (Medium) -270 -284 Gravel -252 -265 Sealing 82 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE Karongi district (mountainous terrain). The screen- level of connectivity to markets/schools/hospitals, ing made with the SPADE model of the 3 FRs shows traffic volume, high intensity of rain, as well as the that one of them has medium priority to pave (low-cost large population served. paving solution), and two have high priority (all paving solutions), with a PPS score from 67-82. The RED model 2. Gravel rehabilitation does not meet with the envi- results indicated that only one of the selected roads was ronment challenge in this district, and FR1 is an economically feasible for sealing if the NPV was set at example of a road that failed during the rainy 12 percent, and all of them were economically feasible if season. the discount rate were set at 6 percent. Karongi district was under the first phase of the parent project, and 3. Road conditions on the sealing trials section indicate all of the roads were gravel roads. The government of satisfactory performance to date. Rwanda has adopted a strategy of sealing feeder roads using various wearing surface options to improve sus- Table 5.5 summarizes the results of the SPADE and RED tainability. To combat climate change and improve resil- models applied to the FRs of the Karongi district. ience, sealing trials were completed on one of the roads (FR14) with four options: cape seal, concrete pad, cold FR1 completed rehabilitation civil work as a gravel road. mix, and DBST. Paving was completed in March 2019, Figure 5.5 shows the before and after scenarios. In and at present the roads are still in good condition. 2018, FR1 experienced severe damage from flooding and landslides (shown in Figure 5.6). In 2019, trial sec- The findings for the Karongi district include: tions were completed on FR14 for four different paving options: all are in good condition so far (Figure 5.7). 1. SPADE gave high priority to pave. The main contrib- utors included the mountainous terrain, the high Table 5.5: Summary of SPADE-RED Model Results for Roads in Karongi District, Rwanda RED NPV@12% RED NPV@6% Length Spade SPADE NPV 000’ NPV/km Best NPV 000’ NPV/km Best Road ID (km) PPS recommends USD 000’ USD Alternative USD 000’ USD Alternative KRFR01 24.4 82 Pave (High) 90 4 Sealing 1931 81 Sealing KRFR03 2.2 67 Pave (Medium) -77 -35 Sealing 49 22 Sealing KRFR14 21.5 76 Pave (High) -278 -13 Sealing 1395 65 Sealing 83 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Figure 5.5: FR1 in Karongi District Before and After Gravel Road Improvement Works Photo credits: Rwanda Transport Development Agency Figure 5.6: Landslide-Affected Sections of FR1 in Karongi District Photo credits: Rwanda Transport Development Agency 84 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE Figure 5.7: FR-14 in Karongi District: Four Different Pavement Trials (a) Cold mix (b) Cape seal (c) Double Bituminous Surface Treatment (DBST) (d) Concrete pad Photo credits: Rwanda Transport Development Agency 85 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS TANZANIA CASE STUDY: ROADS TO INCLUSION AND MTILI-IFWAGI-MKUTA AND WENDA-MGAMA ROADS SOCIOECONOMIC OPPORTUNITIES (RISE) PROGRAM These two roads, also under the RISE program, are part of the tertiary network of district roads of the Iringa The three roads selected for testing are under the pro- region. Mtili-Mkuta (14 km) is located in Mufindi district; posed RISE project, which is in the preparation stage. it transverses rich agricultural areas for food and cash crops, and also a large-scale tea plantation and a formal IRINGA-KILOLO ROAD timber production. Wenda-Mgama (19 km) is located This road is under the RISE project. It is a secondary in the Iringa rural district; it connects the surround- road in the Iringa region, which links the Iringa regional ing communities with the TAZAM highway and the headquarters with the Kilolo district headquarters in Ihemi-Ihimbo secondary road. Both roads are located upland rolling terrain (predominantly <4-6 percent gra- in a rich agricultural area, and also provide the com- dient alignment with localized steeper sections (>8 per- munity with access to social services. Both are unpaved cent). This road has heavy traffic and is a strategic road. and currently in poor or fair condition, and experience It passes through agricultural areas with the potential difficulties in passability during the rainy season. for producing potatoes, fruit, vegetables, beans, and maize, as well as supporting livestock. This district also The SPADE model gave high-priority paving recom- has the potential to export soft wood timber products. mendations for all three roads, and the RED model Approximately 26 kilometers of the 34-kilometer link is indicated economically feasibility for sealing at both the unsealed; the remainder is already bitumen surfaced. 12 percent and 6 percent discount rates. The detailed The unsealed surface is recorded as being in fair to results are shown in Table 5.6. poor condition Some lengths of the alignment are low-lying, and subject to flooding. Some routine and Both the SPADE and RED models recommended paving periodic maintenance was observed as being carried these roads. out on the already sealed section. Table 5.6: Summary of SPADE-RED Test Results for Three Roads in Tanzania RED NPV@12% RED NPV@6% Length Spade SPADE NPV 000’ NPV/km Best NPV 000’ NPV/km Best Road Name (km) PPS recommends USD 000’ USD Alternative USD 000’ USD Alternative Iringa 34 82 Pave (High) 17319 509 AC Sealing 33255 978 AC Sealing Mtili-Mkuta 14 71 Pave (High) 5732 409 Sealing 7038 503 Sealing Wenda-Mgama 19 71 Pave (High) 4902 258 Sealing 8407 442 Sealing 86 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE LAOS CASE STUDY of wooden bridges are in poor condition and require replacement (Figure 5.8b). Little or no effective mainte- The road selected for testing has not been constructed yet. nance has been undertaken, and there is a general lack of nearby sources for rock, aggregate, and gravel. The ROAD 27.01 KHAMMOUANE PROVINCE LAOS project is likely to propose a solution of DBST, possibly This is a tertiary road in southern Laos that is being with concrete pavement in the steepest sections. Spot upgraded under a current Rural Development Program improvement could be considered. (RDP). This unsealed tertiary road was surveyed as being in predominantly poor condition and vulnerable The SPADE model recommended this road as medium to climate impact from a high level of monsoon rainfall; priority for paving (low-cost paving solutions), and the it is also subject to tropical storms. There are some RED model indicated that sealing with DBST at both the areas of very steep terrain (>10% gradient), shown in 12 percent and 6 percent discount rates was economi- Figure 5.8a. This inhibits access in the wet season; but cally feasible. The detailed results are shown in Table 5.7. the road is predominantly 0-4% gradient. A number Figure 5.8a. Steep Section of Poor Earth Road Surface (Laos) Figure 5.8b. Typical Small Wooden Bridge Photo credits: World Bank Project Photos Table 5.7: Summary of SPADE-RED Test Results for a Road in Laos RED NPV@12% RED NPV@6% Length Spade SPADE NPV 000’ NPV/km Best NPV 000’ NPV/km Best Road Name (km) PPS recommends USD 000’ USD Alternative USD 000’ USD Alternative Khammouane 14 67 Pave (Medium) 58 4 DBST Seal 646 46 DBST Seal 87 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS NICARAGUA CASE STUDY The SPADE model gave paving recommendations of high priority for all seven roads, and the RED model indi- The seven roads selected for testing in Nicaragua are cated economically feasibility for sealing at a 12 percent under the Rural Roads Infrastructure Improvement discount rate, which is consistent with the SPADE model Project, financed by the World Bank under IDA terms. results. The detailed results are shown in Table 5.8. Table 5.8: Summary of SPADE and RED Test Results for 7 Roads in Nicaragua RED NPV@12% Length SPADE SPADE NPV NPV/km Road Name (km) AADT PPS recommends 000’ USD 000’ USD Best Alternative Cardenas-Colon 10 218 75 Pave (High) 740 74 Pave (Adoquines) Emp. La Mora-La Carpa 25 256 80 Pave (High) 3590 144 Pave (Adoquines) Emp. Mina El Limon 15 416 74 Pave (High) 4080 272 Pave (Adoquines) Granada-Malacatoya 10 452 79 Pave (High) 4830 483 Pave (Adoquines) Quebranda Honda – San Francisco Libre 10 256 80 Pave (High) 1920 192 Pave (Adoquines) Rio Blanco-Mulukuku 10 348 82 Pave (High) 3550 355 Pave (Adoquines) Rivas-Veracruz 8 633 78 Pave (High) 3000 375 Pave (Adoquines) HYPOTHETICAL CASE OF A LOWER-BOUND ROAD case study helps to demonstrate the credibility of the To further validate the SPADE model, a worst-case proposed SPADE model, and the thresholds that have scenario was also investigated, in which a hypothetical been set. road was envisaged, to test the lower boundary. This hypothetical road bears the following features: the road RESULTS OF THE COST EFFECTIVENESS ANALYSIS is in a country with a relatively high GDP, a high RAI, FOR ROADS WITH AADT < 200 a relatively low share of rural population, and most of For purposes of thoroughness, the SPADE-PLUS its paved roads having a low disaster risk. This road is approach dictates that the CEA needs to be run for not in an agriculture zone, has a low level of strategic roads with an AADT of less than 200. This exercise was purposes, and there is no rural road policy or manual run for 31 roads that fall into this low-volume traffic cat- in place. This road was envisaged as low volume, not egory, almost all of which are feeder roads in Rwanda connected to markets/schools/hospitals, and in an (two are in Ethiopia, and one is in Laos). The CEA was unfriendly engineering environment. run for the DBST paving solution, which was the most popular recommendation among the results from the The hypothetical run yielded a PPS of 25, with a rec- first tests. The results are presented in Table 5.9. All ommendation of “no paving” provided by SPADE. This roads passed the CEA with flying colors, completing the economic evaluation exercise, and validating the rec- ommendations made by the SPADE model. 88 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE Table 5.9: Summary of CEA Results for Case Study Roads with AADT under 200 CEA check Length SPADE Total construc- Number of Cost/No. of Country Road Name (km) AADT Spade PPS recommends tion cost (DBST) Beneficiaries beneficiaries Laos Khammouane 14 132 67 Pave (Medium) 1,260,000 16400 77 Rwanda KRFR01 24.4 151.6 82 Pave (High) 1,144,300 52503 22 Rwanda KRFR03 2.2 66 67 Pave (Medium) 333,956 9508 35 Rwanda KRFR14 21.5 105 76 Pave (High) 3,533,820 49623 71 Rwanda NBFR01 14 96 71 Pave (Medium) 2,868,821 12678 226 Rwanda NBFR02 33 99 66 Pave (Medium) 6,344,546 50109 127 Rwanda NBFR03 6 31 68 Pave (Medium) 1,056,323 27103 39 Rwanda NBFR04 14 57 65 Pave (Medium) 3,289,920 14338 229 Rwanda NBFR05 49 154 76 Pave (High) 7,924,387 62273 127 Rwanda NBFR06 2 50 62 Pave (Medium) 377,611 5896 64 Rwanda NBFR07 5 27 62 Pave (Medium) 1,057,080 19424 54 Rwanda NBFR08 6 36 60 Pave (Medium) 1,571,386 10162 155 Rwanda NBFR09 9 13 64 Pave (Medium) 2,352,035 7734 304 Rwanda NBFR10 6 34 57 Pave (Medium) 1,467,355 14294 103 Rwanda NBFR11 6 62 72 Pave (High) 951,434 21489 44 Rwanda NBFR12 3 59 56 Pave (Medium) 559,995 11571 48 Rwanda NBFR13 5 52 71 Pave (High) 1,224,266 8947 137 Rwanda NBFR14 11 124 70 Pave (Medium) 2,297,742 15616 147 Rwanda NBFR15 14 41 70 Pave (Medium) 2,970,857 22784 130 Rwanda NBFR16 16 77 65 Pave (Medium) 3,196,149 17049 187 Rwanda NBFR17 1 62 60 Pave (Medium) 381,667 5740 66 Rwanda NGFR05 19 173 58 Pave (Medium) 2,392,851 20530 117 Rwanda NGFR06 16 191 65 Pave (Medium) 1,608,180 29259 55 Rwanda NGFR07 16 169 67 Pave (Medium) 1,766,131 40067 44 Rwanda NGFR08 9 125 60 Pave (Medium) 1,019,405 23518 43 Rwanda NGFR09 11 105 56 Pave (Medium) 1,247,877 16392 76 Rwanda NGFR10 18 109 66 Pave (Medium) 2,193,291 32507 67 Rwanda NGFR11 1 127 57 Pave (Medium) 135,564 6325 21 Rwanda NGFR14 6 91 53 Pave (Medium) 758,716 3408 223 Ethiopia Arsi Robe-Wabe 53 89 70 Pave (Medium) 5,300,000 34066 156 Ethiopia Gog Jore -Akobo 120 146 64 Pave (Medium) 12,000,000 50876 236 89 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS CONCLUSIONS FROM THE SPADE-PLUS 3. Using the SPADE-PLUS approach enables more MODEL APPLICATION roads to be paved, which is a good outcome The test results and case studies described above considering climate change, since a paved surface soundly validated the satisfactory performance of is better able to withstand the erosive and abrasive the SPADE-PLUS approach proposed in this report. forces of water, wind, and traffic than a gravel road The SPADE model is proven to work efficiently as a surface. user-friendly screening, prioritization, and guidance tool, with data inputs that are relatively easy to obtain. 4. The SPADE-PLUS approach for the first time The RED model, in conjunction with CEA, which may enables both the policy maker and the road also be used, depending on the traffic levels, is a suit- practitioner to provide a comprehensive expla- able economic justification tool that works well with the nation for why they are making the recommen- SPADE model, for the more comprehensive SPADE- dation to pave a road or not, in a way that is trans- PLUS approach. In order to achieve robust results, parent and can be easily understood by everyone, users are encouraged to systematize the collection of including non-technical persons. data needed for the indicated SPADE model variables. Given the SPADE model’s ability to capture social, eco- 5. Users are urged not to take the SPADE model nomic, and environmental factors, the paving recom- as a black box that issues an indubitable PPS. mendations provided by SPADE offer a more com- There may well be instances where common sense, prehensive perspective than the traditional economic or extraordinary situations, dictate that a road must considerations when it comes to the question of wheth- be paved irrespective of the PPS. The SPADE model er or not to pave a road. has been built in such a way that all variables are known for the sole purpose of aiding the paving Six key lessons can be drawn from this case study work, decision-making process to proceed in a holistic and are worth highlighting here: way. Practitioners that seek to use other factors in their prioritization or justifications would need to 1. If a traditional cost-benefit analysis (CBA) draw upon other arguments or sources beyond the approach were to be used, 16 percent of the present work. roads considered in these case studies would not be justified for paving; but they would be justi- 6. Because of the need to standardize the model fied for paving under the SPADE-PLUS approach. for application on a broader scale, the parame- ters that have been chosen, and the weightings 2. In one of the most interesting case studies (in applied to them individually, as well as to the Rwanda), some of the roads were recommended for broader grouping buckets have been pre-set. paving under the traditional approach, even when This does not preclude a government agency or the NPV was negative. This is a problematic case road authority that has very specific policy objectives to argue, and the use of the SPADE-PLUS approach from using the default SPADE model as a template helped to overcome this burden. and customizing it according to their needs. This report does not recommend such an approach, but cannot preclude it. 90 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE 6. Choice of Road   Surfacing Options Photo credits: Curt Carnemark 91 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS 6.1 General Considerations Good road asset management requires that the pro- maintenance, rehabilitation, reconstruction, and cesses must be in place for determining the optimum upgrading, combined with other best-practice engi- funding level needed to preserve the existing road neering measures like the provision of drainage, network, and to provide suitable access to people in erosion control techniques, and slope stabilization, underserved areas. While good knowledge exists for as shown in Figure 6.1. how to determine and prioritize remedial or improve- ment needs on the existing road network, the prioritiza- Detailed investigations that are focused on the decision tion of potential upgrading projects requires significant of whether “to pave or not to pave” should consider the additional detailed information for the purposes of appropriateness and costs of alternatives for the specif- comparison. Typically, the solution is not a binary ic environment of each potential project. Stage 1 of the choice between paving and not paving, but runs SPADE model can provide guidance in terms of which the gauntlet from routine maintenance, periodic pavement surfaces should be considered. Figure 6.1: Typical Road Asset Management Considerations and Processes Existing road network New roads Transportation master plans and Monitoring and identification Monitoring and identification Monitoring and identification regional needs – need identification (Structures, Slopes, Furniture) (Surfaced roads) (Unsurfaced roads) (New roads) Identify need to Model alternatives for Model alternatives for provide, repair, replace, different funding different funding improve scenarios scenarios Strategic Periodic Major Periodic New roads Maintenance rehabilitation/ Maintenance Upgrading to and tactical surfaced (Repair & Reseal) improvements e.g. regraveling evaluation Detailed Additional Detailed Additional Detailed Detailed investigation investigation if investigation investigation if investigation investigation required required Optimum funding requirement and proportional allocation (including routine maintenance) Optimization of available fundings and prioritisation (including routine maintenance) - Prepare Work Programs Detailed design and procurement Execution Source: GD van Zyl 92 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE 6.1 Menu of Road-Surfacing Options Given the multitude of road-surfacing options, there is It is advisable to run the model comparing for a maxi- a need to whittle down the number of options by sub- mum of three surfacing alternatives. The menu of sur- jecting them to the RED model, or to CEA testing under facing options is shown in Table 6.1. Stage 2. Table 6.1 indicates 25 options, and they are not exhaustive. Table 6.1: Menu of Pavement Surfacing Options No. Road Surfacing Summary Description 1 Natural Earth Existing soil 2 Engineered Natural Surface (ENS) Natural soil that has been graded, shaped, and compacted 3 Gravel Layer of natural gravel-wearing course, usually compacted (typically 10-20 cm thick) 4 Chemically stabilized gravel Natural gravel that has had chemical additives to improve properties and performance (e.g. lime or cement, or bituminous additives, with water added and layer-compacted) 5 Earth with dust suppressant Natural earth, but with dust suppressants added to the surface 6 Gravel with dust suppressant Gravel, but with dust suppressants added to the surface 7 Stabilized recycled pavement Existing pavement that is stabilized chemically with recycled or other waste materials to improve properties and performance 8 Mechanically stabilized gravel Gravel-wearing course that is improved by mixing two or more different grading materials (nor- mally gravel with crushed aggregates) to improve properties and performance 9 Graded crushed stone Graded crushed stone aggregate mix with suitable grading and other desirable material proper- ties to improve pavement performance 10 Stone paving blocks Stones, either in their naturally occurring state, or in processed (cut) form, used as the final wear- ing course surface on top of a sand bed; normally hand-packed 11 Concrete blocks Concrete blocks, usually laid as the final wearing course, on a sand bed, which in turn is on top of an existing or improved gravel road base; normally hand-packed 12 Plain concrete Either manually or mechanically mixed concrete laid as the final wearing surface 13 Reinforced concrete Concrete with steel reinforcement incorporated into the final wearing course to add tensile strength to the pavement 14 Earth bricks Bricks formed with earth and water, and dried, or fired up to harden, and then laid as the final wearing surface 15 Clay bricks Bricks formed with clay and water, and dried, or fired up to harden, and then laid as the final wearing surface 93 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS No. Road Surfacing Summary Description 16 Waste bricks/surfacing Bricks or surfacing made totally of waste materials, and dried or fired up to harden, then laid as the final wearing surface 17 Sand seal Bituminous seal with sand added to the bitumen film, and rolled or compacted to form the final wearing surface 18 Slurry seal Bituminous seal with fine aggregates, water, bitumen emulsion, and cement or other additives mixed, laid, and compacted to form the wearing layer 19 Cape seal Bituminous seal with prime application, then bitumen binder, stone aggregates, and finished off with slurry seal 20 Otta seal Bituminous seal with graded aggregate added to the bitumen binder, (no prime) and rolled or compacted to form the final wearing course. Can be a single or double application (to form a single Otta seal or a double Otta seal) 21 Chip seal (also known as surface Bituminous seal with chip aggregates added to the bitumen binder film, and rolled or compact- dressing) ed to form the final wearing course. Normally can be a single application (single- surface dress- ing), or a double application (double-surface dressing). For double chip seal, normally different gradings are used, with 14/20 mm aggregates in the first layer, and 6/10 mm aggregates in the second and final layer. 22 Recycled asphalt Existing asphalt surface that is scarified, and new asphalt or bitumen injected, mixed, and rolled or compacted to form a new and improved wearing-course surface 23 Bituminous macadam Graded crushed stone (mixed with coarse, medium, and/or fine aggregates and filler material) and infused with bitumen to form a bituminous base-like surface 24 Asphalt concrete A hot or cold asphalt mix consisting of graded aggregates, filler, and bitumen, laid and rolled or compacted to form a final wearing surface 25 Self-healing asphalt This is a relatively little-tested, and even less used, but highly promising option where chemical additives are added to the asphalt pavement to give it self-healing properties and eliminate the need for continual overlays.36 These options are discussed below, followed by a summary table that compares all 25 options (Table 6.2). 36 See Tabakovic and Schlangen 2015. “Self-Healing Technology for Asphalt Pavements.” 94 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE THE NATURAL EARTH OPTION This is the bare minimum. In essence, this is the operational, and road context results have been so do-nothing option, and will only be realistic for scenari- low that no engineering-type intervention is warranted os where the SPADE model’s combined country, region, (Figure 6.2). Figure 6.2: Natural Earth Roads Photo credits: GD van Zyl THE ENGINEERED NATURAL SURFACE (ENS) OPTION The first stage of improving the level of service from a the in-situ material, the amount of rainfall, the topogra- track or a naturally developed road is to provide side phy, the amount of traffic, and whether the road is reg- drainage and shape the existing road in such a way as ularly maintained. Even with good drainage provided, to shed water from the road surface (Figure 6.3). The poor quality in-situ materials can easily result in poor performance of an ENS is dependent on the quality of riding quality, resulting in inaccessibility (Figure 6.4). Figure 6.3: Shaping and Forming Track to Engineered Natural Surface (ENS) Source: GD van Zyl 95 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Figure 6.4: Poor Performance Due to the Quality of Materials in Wet or Very Dry Environments Photo credits: GD van Zyl Shaping and maintaining a road with good-quality The cost of converting a track to an ENS road is depen- in situ material could provide an acceptable level of dent on the existing subgrade conditions and the service for very low traffic volumes in any climate for a minimum geometric standards required; the latter is a while, provided that the shape of the road can be main- function of the terrain, the climatic environment, and the tained (Figure 6.5). expected service life before further upgrading is needed. Spot improvements in the form of imported fill, cut-to-fill, cross drainage, and capping layers are often required. Figure 6.5: Well-Maintained ENS Roads in Wet and Moderate Climates Photo credits: GD van Zyl 96 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE THE ENGINEERED GRAVEL SURFACE (EGS) OPTION EGS roads are defined as unpaved roads with adequate engineered natural surface (ENS) road might require, side drainage, and sufficient strength to carry the traffic in addition to a suitable wearing course, additional load and protect the subgrade (Figure 6.6). Upgrad- imported gravel to improve drainage and/or to protect ing from a naturally developed unpaved road or an the subgrade. Figure 6.6: Engineered Gravel Surface (EGS) Road The provision of proper drainage, the selection of suit- 300 vehicles per day (vpd). Based only on the vehicle able materials, good-quality construction, and regular operating cost (VOC) benefits that this road accrued and appropriate maintenance measures could result in due to upgrading to surfaced standards, the unpaved excellent performance. Figure 6.7 shows an example maintenance strategy was still more cost-effective. of a road in a moderate climate that carries more than Figure 6.7: Roughness Deterioration at 5 Bladings Per Annum (AADT = 324 vpd) MR276: km1.8 - 2.3: Roughness Deterioration (Model vs. Actual) 18 16 Riding Quality (IRI) 14 12 10 8 Reshaping 6 4 2 0 0 1 2 3 4 5 6 7 8 Years since construction Adjusted Model Constr f=0.8 Actual QI HDM: Blading (normal compaction) Photo credits: Van Zyl et al, 2008 97 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Figure 6.8: Roughness Deterioration at 2 Bladings Per Annum (AADT = 66 vpd) MR270: km32 - 32.5: Roughness Deterioration (Model vs. Actual) 12 10 Riding Quality (IRI) 8 6 4 2 0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 Years since construction Actual QI HDM4: Blading HDM Adj Photo credits: Van Zyl et al, 2008 Figure 6.8 shows the performance of a well-constructed • Traffic annual average daily traffic (AADT) road in a moderate to dry environment with 66 vpd and (16 – 725 vpd) maintained at 2 bladings per annum. • MMP (10 – 100 mm) • Variety of different materials The results obtained on 67 road sections over a period • Grade (0.2% - 9%), Rise and Fall = 1.5 - 110 meters/ of three years, from a current HDM4 calibration study kilometer (km) by the Western Cape Province in South Africa highlight- • Curvature = 5 - 296 Deg/km ed the fact that an average steady-state roughness lev- el could be maintained at IRI= 6 with an average of 4.5 The distributions of the traffic, blading frequency, and bladings per annum. The sample included the following steady-state roughness are presented in Figure 6.9. variables: Figure 6.9: Blading Frequency and Steady-State Roughness Blading Frequency (No/year) Steady State Road Roughness (IRI) 20 12 International Roughness Index (m/km) Number of blading per year 18 10 16 14 8 12 10 6 8 4 6 4 2 2 0 0 0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100 Percentile below Percentile below Source Van Zyl et al. 2020 98 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE Given that gravel is a wasting surface, its application raffic is relatively low (<400 vpd). Gravel loss accel- • T needs to be carefully considered. A gravel wearing erates with increases in traffic volumes. It is uncom- course (GWC) should only be seriously considered where: mon that a GWC will be cost-effective at traffic flows above 400 vpd, and it may even be unjustified at • Maintenance is guaranteed (both routine and peri- much lower traffic levels, though as the examples odic maintenance will be implemented according to above from South Africa indicate, local context mat- an asset management schedule);37 ters a lot, and there can be exceptions to this rule. • Gravel quality is adequate38 (See the Ethiopia and Malawi Rural Roads Design Manuals for suggested THE MECHANICALLY MODIFIED, CHEMICALLY TREAT- materials specifications for grading, soil classifica- ED, AND STABILIZED ENS OR EGS OPTIONS tion, Atterberg limits, California Bearing Ratio (CBR), maximum dry density (MDD), and moisture content); Treatment or modification of unpaved road materials could include: • Gravel deposits are still plentiful, and are not caus- ing an environmental nuisance or worry about • Mechanical modification; depletion of existing natural resources, or of • Chemical treatment; and/or deforestation; Stabilization. • • Hauling distances are reasonable (<10 km); Mechanical Modification • There is low to moderate rainfall (<1500 mm per Mechanical modification to improve road strength and year); performance could include: • There are no steep gradients (<12 percent), (though • Processing of materials (grid rolling, or crushing, in sometimes localized solutions for this are possible); in-line or mobile crushing plants). Significant benefits in terms of performance and • There are no residences, or only a limited number reduced maintenance costs have been realized by of residences, in a 500-meter buffer of road that reducing oversized materials (Henderson & van Zyl). are exposed to dust in dry season areas, since dry weather facilitates the binding fines to be removed • Blending of materials (cohesionless materials with from the surface by traffic or wind. Dust causes neg- clay or clay-like materials with suitable grading mate- ative health and safety impacts. (Localized low-cost rials like sand or aggregates). paving solutions may need to be adopted for limited stretches of the road); and Guidelines are available for how to select suitable mix proportions of various natural material sources.39 37 Normally routine maintenance activities like culvert cleaning, drain cleaning, grass-cutting, sign maintenance, and limited carriageway clearance and repairs should be continual throughout the year. Grading of the GWC is expected at least once a year, and regravelling at least every three years. In high rainfall areas, such a schedule will be inadequate. 38 Gravel should comply with grading and plasticity requirements and should not break down under traffic, or it will be lost from the surface at a high rate. Natural gravel quality varies substantially within each pit location and with depth. Great care is essential in order to ensure that only suitable material is selected, and that the mixing of marginal or unsuitable material is avoided. 39 Draft TRH20. “Structural Design, Construction, and Maintenance of Unpaved Roads.” Committee of State Road Authorities. Pretoria, South Africa, 2012. 99 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Chemical Treatment • Products are generally expensive; • Bituminous sealed sections always outperform More than 200 products are available worldwide that chemically treated sections; could be mixed into the wearing course material, or sprayed onto it, both to suppress dust and/or to All treatments (except those using enzymes) result- • improve wearing course performance. These products ed in a benefit within one year; could be classified as: Treated sections perform better than untreated • • Wetting agents (water, ground water containing dis- sections; solved salts or wetting agents) • Performance of the untreated gravel sections is • Hygroscopic salts (calcium chloride) highly dependent on material properties, traffic, and • Lignosulphonates environmental conditions; • Modified waxes First-year benefits indicate that the chemical treat- • • Synthetic polymers ment of gravel roads could be more cost-effective if short-term solutions are required (for example, for • Petroleum resins temporary bypasses); and • Bituminous products. Additives can improve the performance of marginal • Guidelines have been compiled through experience materials. and testing and are available to assist practitioners in selecting products that can work in specific situations40. Performance of the various products are displayed in Figure 6.10. Various products have been tested over a period of more than 30 years in South Africa, and elsewhere in None of the chemically treated unsealed sections the world (Uys, van Zyl & Truter 2013). lasted more than one year; but all treated and sealed sections were still in excellent condition The outcome of most studies confirms the following: after three years. This finding suggests that chemically treating a road but leaving it unpaved may be a fool’s • There is no product that will fit all materials and errand in many locations, and is likely not the best use situations; of public resources. • Most products will result in better performance for treated sections than for untreated sections; 40 Draft TRH20. “Structural design, construction and maintenance of unpaved roads. Committee of State Road Authorities. Pretoria, South Africa, 1990.” 100 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE Figure 6.10: Performance of Various Additives Source: Uys et al, 2013 Some cases have been recorded where specific prod- Figure 6.11: Typical Salt Roads in Namibia ucts or chemical treatments performed exception- ally well. The use of brine with a high salt concentra- tion (>20%) on gravel roads close to the west coast of Namibia is well-established, and well-known for its excellent performance close to the coast (Figure 6.11). Regular application of normal sea water (with an approximate salt concentration of 2-3 percent) on several roads in the western parts of South Africa also resulted in excellent performance, as shown in Figures 6.12 and 6.13. Photo credit: GD van Zyl 101 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Figure 6.12: Koekenaap DR2225 - 75 x 7 Axles HV/Day Photo credit: GD van Zyl Figure 6.13: Effect of Regular Sea Water Spray vs. Normal Grader Maintenance There has also been interesting feedback from 164 respondents who are using chemical additives (mainly calcium chloride) in the United States, as shown in Figure 6.14. 102 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE Figure 6.14: Respondent Feedback Regarding the Use of Chemical Additives Number of respondents 0 20 40 60 80 100 120 140 160 180 Control dust Reduce road maintenance costs Extend grader maintenance intervals Preserve gravel Improve level of service for road user Improve safety by minimizing washboard on grades Aid compaction Control erosion Improve strength of in situ or imported materials Improve wet weather drivability Other Source: Skorseth & Sellim, 2005 SEALED SURFACING ALTERNATIVES The most prominent sealed surfacing alternatives are shown in Figure 6.15. Figure 6.15: Sealed Surfacing Alternatives Surfacing Bituminous Non-Bituminous Surface treatments Asphalt Stone Clay Stone Sprayed Slurry Combination Rejuvenation Hot mix Cold mix Natural Dressed Clay Blocks Slabs seals seals seals sprays (20-40mm) (20-40mm) bricks Sand seals Slurry Cape seal Reinforced Graded Micro- Slurry-bound aggregate surfacing Macadam seals Single seal Double seal Source: World Bank 2018 103 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Figure 6.16: Different Types of Bituminous Seals Variations and additional alternatives include: • Penetration macadam eclaimed bituminous surfacings • R (such as asphalt millings, or slurry from hydrocutting) Roller-compacted concrete • • Geocells arious forms of concrete blocks • V (geocells, grass blocks) rack applications, as an alterna- • T tive to full-width applications. This section highlights the most com- mon types of bituminous surfacing used.41 For bituminous seals, Ben- nett et al 2002 provide an excellent illustration of possible options (Figure 6.16). All surface treatments are sensitive to the quality of the base and the con- struction, the materials used, quality control during the laying operation, and timely maintenance. Source: Bennett et al 2002 41 The information was obtained and/or adapted from SABITA Manual 40; Cook et al 2013; and Bennett et al 2002. 104 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE Sand Seals Option General description and application Principle advantages Principle concerns Consists of a film of binder (preferably • Cheapest bituminous • Highly sensitive to turning actions cutback bitumen or emulsion) followed surface treatment • Highly sensitive to rain shortly after by a graded natural sand or fine sand, • Very suitable for labor- construction machine or hand-broken aggregate based construction, • Highly sensitive with urban drainage systems (maximum size typically 6 – 7 mm), especially where emul- (erosion between curbs and sand getting which must then be compacted. sions are used, and into subsurface drains) requires only a simple • Sand needs to be boomed back into the Single sand seals are not very durable, plant for construction. “worn” wheel tracks. There is an extended but performance can be improved with • Especially useful if good curing period (a minimum of 8 – 12 weeks) the application of a second seal within aggregate is hard to find. between application of the first and second the first year, depending on traffic. • Highly suitable for seals needed, to ensure complete loss of vol- Could then last for another 6-9 years periodic maintenance atiles and thus to prevent bleeding before another reseal is required. (resealing) on low-vol- • Road markings must be postponed until the ume roads seal has settled Otta Seals Option General description and application Principle advantages Principle concerns An Otta seal consists of a relatively • Suitable for all LVR traffic. • Requires specific types of soft bitumen, which thick layer of bitumen binder fol- • Proven sealing technique may not be readily available in some regions. lowed by a layer of aggregate that is in a number of road • Extensive rolling with heavy rubber-tired rolled into the binder using a heavy environments. compaction plant is essential, therefore not pneumatic-tired roller, or loaded suitable for small contractor/community • A wide range of natural trucks. A graded gravel or crushed implementation or processed aggregate aggregate (19 mm down) is used in may be used. • Significant 2-axle rubber-tired traffic required comparison to single-sized material after construction to bring up the bitumen. used in conventional chip seals. Its • Highly sensitive to rain shortly after success depends on the binder being construction squeezed up through the aggregate by the action of extensive rolling by • Highly sensitive to urban drainage systems pneumatic-tired rollers and by traffic. (for example, sand getting into subsurface A single Otta seal plus a sand seal or a drains) double Otta seal is recommended as • Road markings must be postponed until the initial construction seals. seal has settled. 105 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Single Chip Seals Option General description and application Principle advantages Principle concerns A single-chip seal consists of the • Cheaper solution than a • Accurate application of binder and aggre- application of a suitable binder, a double seal gate is required. single-sized aggregate, and rolling, to • Could be applied by hand • Highly sensitive to turning/braking orientate and embed the stone in the or with small equipment on actions and erosion if not blinded with bituminous binder. Application of an very low-volume roads sand. emulsion cover spray reduces the risk of early aggregate loss. An addition- al application of a coarse sand in the emulsion cover spray further reduces the risk of aggregate loss. Double Chip Seals General description and application Principle advantages Principle concerns A double-chip seal consists of two • Proven performance in all • Sensitive to erosion on steep slopes with applications of a suitable binder, and climates, suitable for rural urban drainage systems (curbs) then spreading and rolling two layers situations. • Good distribution of the binder is of single sized aggregate. The second • Double seals are durable required. layer of aggregate is normally half the (typical initial life 8 – 14 • Initially sensitive to cold weather, and to size of the first layer, which provides years) turning and braking actions if the second the seal structure with stability and • The combination of a large aggregate layer is more than half the size minimizes the risk of early aggregate aggregate, and 1/3 of the of the first layer loss. A diluted emulsion cover spray large aggregate size is • 20/10 double seal is considered the high- could be applied as a final layer, fur- considered the lowest-risk est-risk double seal for early aggregate ther reducing the risk of aggregate double seal. loss. loss. 106 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE Slurry Seals and Microsurfacing General description and application Principle advantages Principle concerns Slurry seals are a mixture of well-graded fine • Proven performance in all • Conventional slurry can take sev- aggregate, bitumen emulsion, filler (usually climates, suitable for both eral days to cure in conditions of Portland cement or lime), and additional water. rural and urban situations. high humidity They are mixed in a concrete mixer or other • Suitable for light traffic, can • Thin layers are highly sensitive to purpose-built equipment, and are spread on a be used in combination turning and braking actions pre- prepared surface using wheelbarrows and with other seals and as • A stiff layer, due to low bind- squeegees, or spreader box/drag spreaders. The maintenance treatment. er film thickness, can result in slurry seal can be spread by hand to the thick- • Construction of con- early fatigue cracking or crack ness of the large aggregate fraction. Following ventional slurry does reflection. application at ambient temperature the water not require expensive • Initial high permeability if not separates from the emulsion and evaporates, equipment. compacted with pneumatic tire leaving the residual bitumen in place to adhere • Suitable for construction rollers to the pavement surface and aggregates. by small contractors or • Requires local sources of suitable community groups. emulsion. Microsurfacing differs from conventional slurry in that chemicals are used to speed up the • Safer for operatives and • Emulsion needs to cure before breaking process (the separation of the water). local village personnel to the road can be opened to traffic. The rapid curing characteristics require applica- construct and maintain • A thin layer (<15 mm) is not rec- tion by spreader box only, but could be applied than hot bitumen ommended as an initial seal, even to a 30-mm thickness, and in conditions of high for a low volume of traffic. humidity. Cape Seals General description and application Principle advantages Principle concerns A Cape Seal is a multiple surface treatment con- • Proven performance in all • Requires local sources of suitable sisting of an application of a single bitumen chip climates, suitable for both emulsion. seal followed by a single or double application rural and urban situations. • Emulsion needs to cure in each of bitumen slurry seal. Usually a first seal of 13 • Suitable for light to heavy application, and before opening mm of chipping is combined with a single slurry traffic and turning/braking the road to traffic. application. A 19 mm chipping first seal is nor- actions. • Over-application of stone will mally combined with a double slurry application. • Construction does result in the slurry not filling all The aim is that on completion the tops of the not require expensive the voids. stone chips will be exposed just above the slurry, equipment. • Irregular application of stone will which will fill the interstices between the stones. • Suitable for construction result in poor surface quality by small contractors or • Use of solvent in the binder can The Cape Seal technique is durable (typical initial community groups. result in bleeding. life 8 – 16 years), and it enables a heavy-du- ty surfacing to be constructed with minimal equipment. 107 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Slurry-Bound Macadam Seal General description and application Principle advantages Principle concerns The slurry-bound macadam seal is a • Excellent performance in excess of • The stiff layer, due to low binder combination of a single-sized aggre- 10 years in an urban environment film thickness, could result in gate and fine slurry. that has a high frequency of turning early fatigue cracking on a soft vehicles upper base. A dry layer of 14 or 20 mm of aggre- • Resistant to water flow in urban • Expensive (double the cost of a gate is placed at a thickness of 20 – 50 environments 14/7 double seal) mm, then levelled and compacted. • Suitability for labor-enhanced con- • Slow construction A fine slurry is then vibrated into the struction (very high) Constructed stone matrix, using a pedestrian roller. using only manual labor and small After curing the slurry, the layer is equipment rolled with a static roller, and a final • Ability to smooth out irregularities 4-6 mm layer of slurry is applied. on the base, and could be used on coarse base textures Rejuvenation Sprays Figure 6.17: Rejuvenation Sprays [Sabita Manual 40] Photo credits: GD van Zyl 108 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE General description and application Principle advantages Principle concerns Diluted stable-grade, or invert cut-back • Easy application by hand or • Must be applied before wide emulsion is sprayed on an existing sur- distributor cracks occur face treatment to prolong the effective • Could extend the service life of • Could run off at steep grades service life. Typical rates of application surface treatments for more than • Invert cut-back emulsions can are 0.8 l/m2 for diluted emulsions, and three years stay tacky for several days. Only 0.45 l/m2 for invert cut-back emulsions. suitable for very low-volume Most aggregates in Africa are nega- traffic, or where traffic can be tive-charged; therefore, anionic emul- accommodated for the required sions are preferred to run into the seal period to prevent pickup structure and not adhere immediately to • Sufficient voids must be present the stone. in the seal structure, as over-ap- plication of emulsion could reduce macro texture and skid resistance. Asphalt General description and application Principle advantages Principle concerns Hot premixed asphalt consists of grad- • Suitable for rural or urban applica- • Most expensive bituminous ed crushed aggregate, bitumen, and tion in all climates. surfacing an active filler (for example, lime). It is • Local labor employment with cold • Relatively stiff surfacing has thin normally mixed in a plant, and paved mixes both in construction and in layers that are sensitive to crack- with a custom-built paver, then com- ongoing maintenance. ing if on a poor support pacted. Note: Could be mixed on site • Strongest bituminous surfac- • High permeability of cold mixes with a small plant, spread by hand, and ing, and least sensitive to poor or poorly compacted hot mixes compacted. The thickness can vary, maintenance • Low macro texture (low skid dependening on the nominal size of the • Longest expected service life on resistance) with continual graded aggregate. good pavement structures mixes Cold mixes normally consist of an admix- • Improves road roughness (elimi- ture of graded crushed aggregate and nates small base irregularities) a stable, slow-breaking emulsion which is mixed by hand or in a concrete mixer. After mixing, the material is spread on a primed roadbase and rolled. Thickness can vary, depending on the nominal size of the aggregate (20 – 40 mm). Very suitable for labor-based construction; requires only a very simple construction plant; reduces the potential hazards of working with hot bitumen. 109 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Concrete Blocks Figure 6.18: Concrete Blocks General description and application Principle advantages Principle concerns Concrete brick paving is a well-estab- • Suitable for rural or urban applica- • The un-mortared joint option lished technique used in many countries tion in all climates. may be subject to erosion in for a variety of applications, including as • Social and economic benefits to the areas of high rainfall if not an option for low-volume rural roads. communities through local manu- well-maintained. The application is based on the proven facture of blocks. • Requires consistent production ability of individual concrete bricks to of good-quality blocks of 25MPa • Centralized brick production can effectively disperse load laterally to crushing strength. facilitate good quality control. adjacent bricks through the sand joints. • Local labor employment both • Needs control of construction This option consists of rectangular in construction and in ongoing using string lines within pre-con- concrete bricks (usually around 70 mm maintenance. structed edge constraints (curbs) thick) being laid in a herringbone or • Suitable for construction by com- • Irregular thickness of bedding other pattern to camber within confining munities or small contractors. sand quickly results in poor rid- edge-curbs (cast either before or after ing quality. placement of the bricks). They are com- • Good durability, load-bearing, and pacted into place, and sand is brushed load-spreading characteristics. • Angular sand required to in at the joints. A sand cement mortar improve friction between blocks • Low-cost maintenance procedures. joint or bituminous seal may be speci- and bedding sand performance fied to waterproof the finished surface as a separate operation, although this is usually unnecessary on a well-construct- ed sub-base. As a refinement, the con- crete bricks may be cast with a top edge chamfer to assist surface drainage. 110 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE Clay Bricks Figure 6.19: Clay Bricks General description and application Principle advantages Principle concerns Fired clay bricks are made by firing • Proven performance in all climates. • The mortared joint option may be molded blocks of silty clay. They are • Suitable for urban application if the subject to erosion in areas of high commonly used in low-cost road mortar is jointed and sealed. rainfall if not well-maintained. pavement construction in areas with • Requires consistent production of • Social and economic benefits to a deficiency of natural gravel or rock good-quality engineering bricks of communities through local brick materials. This surfacing consists of >20-2 MPa crushing strength. manufacture. placing a layer of edge-on engineering • Good carbon footprint attributes if • Needs good control of construc- quality bricks within previously installed bricks are burned using agricultural tion using string lines within edge constraints. The bricks are laid in waste or sustainable fuel. preconstructed edge constraints an approved pattern on a bedding lay- (curbs) er of sand, or on a previously laid layer • Local labor employment both in construction and in ongoing • Irregular thickness of bedding of flat-laid bricks (a “soling layer”). The maintenance. sand quickly results in poor riding joints between the bricks may be either quality. in-filled with suitable sand, or they may • Good durability, load-bearing, and be mortared in. load-spreading characteristics. • Angular sand is required to improve friction between the • Low-cost maintenance procedures. blocks, and bedding sand performance. 111 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Dressed Stone Figure 6.20: Dressed Stone General description and application Principle advantages Principle concerns Dressed stone surfacing is a historical- • Can be constructed at steep • Requires local availability of suit- ly well-established technique that has gradients. able stone. been adapted successfully as a robust • Suitable for all climates, and for • Requires skill in quarrying, dress- alternative option for low-volume rural light to heavy traffic in rural or ing, and laying to achieve a smooth roads where there is a good local sup- urban situations. finished surface. ply of suitable stone. Strong isotropic • Construction does not require • Requires a minimal compaction rocks such as granite that have inher- expensive equipment. plant. ent orthogonal joint stresses are ideal. • Suitable for construction by com- • Surface is not smooth, and a Dressed stone surfaces have good munities or small contractors. medium amount of roughness is load- spreading properties. • Suitable for remote areas with normal. This technique consists of problems accessing a construction • Stones that will be polished by traf- 150-200-mm-thick dressed stones plant. fic, or that are slippery when wet, being laid to lines and levels between • Can be easily maintained, easily must not be used. previously installed edge restraints, repairable. • Cannot be used until the mortar and compacted into a bedding layer of • Can later be upgraded by covering joints have set and hardened suf- sand, followed by cement mortaring of with a sealing layer in a staged con- ficiently (usually about 5-7 days in the joints. The dressed stones should struction strategy. hot/warm climates) normally be hand cut from solid rock • Erosion resistant, durable. and trimmed (dressed) if necessary, to form a regular rectangular shape, free from flaws and discontinuities, with a reasonably smooth top surface. 112 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE Cobble Stone Figure 6.21: Cobblestone General description and application Principle advantages Principle concerns Cobblestone paving is a historically • Suitable for all climates, and for light- • Requires suitable stone to be well-established option consisting of to-heavy traffic in both rural and available locally, and cobbles must a layer of roughly cubic-shaped or urban situations. be roughly cubical in shape. selected stones of a thickness of about • Construction does not require • Requires masonry skill in laying to 100 – 150 mm, laid on a bed of sand expensive equipment. achieve a smooth finished surface. or fine aggregate within mortared • Suitable for construction by commu- • Requires some compaction plant. stone or concrete edge restraints. The nities or small contractors. • Unsuitable for moisture-sensitive individual stones should have at least • Suitable for remote areas with prob- sub-base/subgrade in areas of one face that is smooth, to be the lems accessing a construction plant. moderate to high rainfall. upper or surface face when placed. Each stone (or cobble) is adjusted with • Can be constructed on steep • Surface is not smooth, and medi- a small mason’s hammer and then gradients. um roughness is normal (although tapped into position to the level of • Low maintenance, easily repairable. this does discourage high speeds). the surrounding stones. Sand or fine • Can be upgraded later by covering • Potential safety issue with pol- aggregate is brushed into the spaces the cobblestone with a sealing layer ished stones in wet condition, par- between the stones, and the layer is in a staged construction strategy. ticularly for two-wheeled vehicles. then compacted with a roller. • Erosion resistant, durable. Note: Stone sets is a neater alternative; it consists of a layer of cubic-shaped stones of approximately 80–100 mm, laid on a thin layer (20 – 50mm) of sand bedding. The sets can be cut by hand from a suitable hard rock like granite or basalt, that can easily break into smooth-faced pieces. Sand is brushed into the joints between the laid stones, and they are compact- ed using a vibrating plate or light roller. An edge restraint or curb constructed (for example) of large or mortared stones is required for durability. Sand-cement mortar joints can be used to improve durability and prevent water penetrating the foundation layers and weakening them. 113 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Concrete Figure 6.22: Nonreinforced or Reinforced Concrete General description and application Principle advantages Principle concerns Nonreinforced cement concrete is a • Suitable for all climates and • High initial construction cost compared to well-established form of rigid pave- for rural or urban application. most other options. ment designed to spread the applied • Robust option suitable for • Usually requires expansion and contrac- load of traffic through a slab effect. regions with high rainfall, or tion joints with steel load-transfer dowels. The option, as applied to low-volume that are prone to flooding. • May be susceptible to shrinkage and (LVRRs), usually involves the casting • Generally resistant to axle cracking unless well-constructed and of 5-meter slabs between formwork, overloading if well-construct- cured. normally with load transfer dowels ed and founded. • Tendency for laborers/contractors to add between them. In some cases, where • General concreting proce- too much water to the concrete mix to continuity of traffic demands it, these dures are understood by facilitate placement, which weakens the slabs may be half a carriageway width. small local contractors. slab and risks shrinkage cracks. The concrete slabs are cast onto a previously prepared and compacted • Minimal maintenance if prop- • Concrete must not be mixed or placed in sub-base. The concrete must be cured erly constructed and cured. ambient shade temperatures above 38 by covering it with moisture-retaining • No requirement for expen- degrees centigrade, and must be protect- material that is kept moist, normally for sive construction plants. ed from direct sunlight that would raise a minimum period of 7 days. mix temperatures to the same high levels. It is most suitable for construction in • The first and last slabs of an unreinforced areas with high rainfall, steep gradient section are subject to impact loading as alignments, and/or on routes that are vehicles move on and off the edge of susceptible to seasonal flooding and the slab; these slabs must be designed other major climatic impacts. accordingly. Continuous reinforced concrete could • Requires at least 7-14 days curing time be used as an alternative; this would following initial construction. allow for a thinner layer, but would • Requires a good sub-base, and that shoul- require higher skill levels. ders be well maintained against erosion. 114 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE A summary Surface Type Decision Framework (STDF) comparing the different surface types and their suitability for application is presented in Table 6.2. Table 6.2: Surface Type Decision Framework (STDF) Factor ENS Gravel Stone Paving Concrete Concrete Bituminous Asphalt Concrete Blocks Seals Availability of High High Medium High Impor- High High High Impor- Materials importance importance Importance tance (Cement, Importance Importance tance (Bitumen, Aggregates) (Cement, (Bitumen, Aggregates) Aggregates) Aggregates) Availability of Low Importance Medium High Importance High Medium Medium Medium Labor Importance Importance Importance Importance Importance Availability of Low Importance Medium Medium Medium Medium High High Importance Equipment Importance Importance Importance Importance Importance Knowledge of Low Importance Low Importance Low Importance Medium Medium High High Importance Construction Importance Importance Importance Method Quality Medium High High Importance High Impor- High High High Importance Assurance Importance Importance for base course tance for base Importance Importance course Gradient Not recom- Not recom- Medium Medium Medium Medium Medium mended for mended for Importance Importance Importance Importance Importance steep gradients steep gradients (>12%) (>12%) Climate/ Not recom- Not recom- Not recommended Not recom- Okay for all Okay for all Okay for all cli- Rainfall mended for mended for for extremely high mended for climates climates mates, but care is high rainfall high rainfall rainfall (>2000 high rainfall needed to ensure (>1500 mm/yr) (>1500 mm/yr) mm/yr). If used, (>2000mm/yr) If proper asphalt mix good joint seals used, good joint for very hot or very needed seals needed cold climates Traffic Not recom- Not recom- Okay for rural road Okay for rural Okay for rural Okay for rural Okay for rural road mended AADT> mended AADT traffic context road traffic road traffic road traffic traffic context 50 > 400 context context context Maintenance High High High Importance High High High High Importance Importance Importance Importance Importance Importance Level of Low Importance Low Importance Medium Medium High High High Importance Service Importance Importance Importance Importance Design Speed Low Importance Low Importance Low Importance Medium High High High Importance Importance Importance Importance 115 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS 7. Conclusions and   Recommendations Photo credits: Wenxin Qiao 116 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE This report offers the following conclusions and 4. Continued relevance of economic evaluation. recommendations: Traditional economic evaluation models for deci- sions about roads have their constraints but cannot 1. Base factors for the provision of rural roads. be fully discarded, because, given competing invest- The provision of a rural road needs to be publicly ment demands, public investments need to ensure desired, politically supported, socially acceptable, the best use of scarce resources. Of the available economically justified, safely designed, environmen- models, the RED model has been traditionally tally sustainable, technologically appropriate, and used by the World Bank and many road agencies locally contextualized. to evaluate the economic justification of proposed road investments, and has been found to be useful. 2. Importance of institutional setups and fund- Cost-effectiveness analyses (CEAs) must deal with ing for maintenance. Before a rural road paving the challenge of ensuring value for money when program should even be contemplated, it is rec- economic justifications using traditional models ommended that the right institutional setup exists, become stretched on traffic grounds in the face of with clear delineation of responsibilities, and policy competing social and other non-easily quantifiable documents including, for example, a rural road benefits. policy, strategy, and plan; an asset management system with inventory and condition surveys; and 5. The proposed SPADE-PLUS approach is the new sustainable funding and arrangements for road frontier. The SPADE-PLUS (SPADE + RED + CEA) maintenance. approach is presented and justified in this work as the next frontier in rural road prioritization and 3. Weaknesses of the traditional approaches. The evaluation, and its improvements over current traditional pure cost-benefit economic analysis: (i) approaches are showcased. The proposed approach fails to capture important but hard-to-quantify is a two-stage process that introduces a novel mod- social and other economic benefits (for example, el: the “Systematic Paving Decision,” or “SPADE” mod- land value improvements, and the attraction of el, to take into account a multicriteria analysis (MCA) business opportunities), and tends to focus rather of the prevailing macro and micro contexts in the only on savings in road agency costs, vehicle oper- first step, before the RED model analysis or a cost ating costs, and travel time; (ii) biases investment effectiveness analysis (CEA) can be undertaken in the toward more prosperous areas where vehicle traffic next stages. The SPADE model considers the coun- demand levels are already more established due to try context, the regional context, the operational greater motorization, and thus discriminates against environment, and the specific road context to make investment in poorer areas; and (iii) has methodol- a preliminary decision for which roads have either ogies that are more suited to dealing with high- no justification, or limited justification for paving; er-traffic roads than lower-trafficked rural roads; which have moderate justification for paving; and yet many rural roads are dominated by lower levels which have high justification for paving. Whether to of traffic volume. use RED or CEA hinges on current traffic levels (for 117 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS AADT over 200, the RED model is recommended, including traffic signs, sidewalks, shoulders, and traf- while for AADT under 200, the CEA is recommend- fic calming strategies in populated villages, and in ed). The SPADE-PLUS approach has been developed sensitive areas, for example, in the vicinity of schools and tested as fit-for-purpose. However, where policy or hospitals. objectives differ from the set default values, or spe- 9. Importance of quality gravel materials. Spe- cific contexts so dictate, teams may further build on cial care needs to be taken in the choice of grav- the model, or run more extended case studies to el materials to be sure that they have the proper suit the given SPADE model prototype to their spe- engineering properties (soil/aggregate type, particle cific contexts. size distribution, plasticity index) that will reasonably mportance of considering all users, not just 6. I withstand both traffic and climate effects for the vehicle users. There is a need to consider all road chosen design period. users, not just motor vehicle users, in the deci- Importance of climate, drainage, and mainte- 10. sion-making process. This includes pedestrians, nance. The damaging effects of climate (especial- cyclists, motorcyclists, and animal-drawn cart users. ly rainfall) on unpaved pavement leads to wasted This work recommends that motorcycles be includ- investment resources. It is most pernicious, and ed in the AADT, at a factor of 0.5 per vehicle. It also needs to be critically analyzed and given due con- recommends that pedestrians, cyclists, and other sideration. Whatever the chosen pavement option non-motorized users be counted, and captured in (gravel, a low-cost paving solution, or a fully paved the SPADE model. Special attention should also be solution with asphalt or concrete), proper drainage given to more vulnerable groups (the elderly, chil- (both side and cross-drainage), slope stabilization, dren, and the disabled). and both routine and periodic maintenance are of 7. Gender and mobility considerations. It is wide- critical importance, especially in a world current- ly recognized that women and men have different ly challenged by the disruptive effects of climate mobility patterns. Paving decisions should consider change. these differentiated demands and needs, especial- Showstoppers. Despite the recommended 11. ly as they relate to pedestrian and non-motorized two-stage approach using a combination of the traffic, which is usually the main mode of transporta- SPADE and RED/CEA models, some departures tion for women and girls. These considerations need from the norm may be inevitable. These are called to be factored not only into the final decision, but “show-stoppers.” They include situations in which also into the complementary design features and there is no availability of suitable gravel material; measures that are adopted, irrespective of paving excessive heavy-load traffic; or areas with high levels choice. of rainfall. In such cases, the choice of paving tech- 8. Safety considerations. Road safety is a critical issue nology will depend a lot on the results of the model, in this decision process. Paving roads leads to an but also on specifics of the local context (availability increase in vehicle speeds. Where vulnerable users of materials, the cost of various options, availabili- like pedestrians and cyclists are prevalent, special ty of labor and/or equipment, technical know-how, consideration should be made for the provision of climatic conditions, and the maintenance regime). both horizontal and vertical road safety measures, Special care should be made not to impose a paving 118 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE technology that will not perform well in the local context, especially under the straining effects of traf- fic and climate. A wide menu of surfacing options. There is a wide 12. menu of available road surfacing options. Special care should be taken to select the options that are appropriate for the traffic, climate, gradient, materi- als, and labor mix availabilities, as well as to consid- er maximal use of locally available materials, labor, and equipment. When necessary, solutions can be applied over reduced road widths, wheel paths only, or even over intermittent sections of road if justi- fied. In the medium-score range under the SPADE model, low-cost paving solutions are recommended: for example, single or double- chip seals (surface dressing); cobblestones; and the use of sand and Otta seals, among other options. In the high-score range, all paving options are on the table, including the more costly, but also more durable ones such as asphalt concrete, concrete block, and concrete sur- facing. The economic justification under Stage 2 of the SPADE model can help planners determine the more cost-effective choice in each case. 119 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Bibliography African Development Bank. 2014. “Study on Road Infrastructure Costs: Analysis of Unit Costs and Cost Overruns of Road Infrastructure Projects in Africa.” Abidjan: African Development Bank. Alzubaidi, H. 1999. Operation and Maintenance of Gravel Roads. Stockholm: Swedish National Road and Transport Research Institute American Association of Highway and Transportation Officials (AASHTO). 2001. Guidelines for Geometric Design of Very Low-Volume Roads (ADT <= 400). Washington, DC: AASHTO. _____. 1977. A Manual on User Benefit Analysis of Highway and Bus-Transit Improvements. Washington, DC: AASHTO. Archondo-Callao, R. S., 1999. Roads Economic Decision Model (RED) for Economic Evaluation of Low Volume Roads. Africa Transport Technical Notes 18. Washington, DC: World Bank. Archondo-Callao, R. S., and E. James. 2009. Economic Evaluation of Paving Rural Roads with Adoquines. Washington, DC: World Bank. Aschaeur, D. 1989. “Is Public Expenditure Productive?” Journal of Monetary Economics 23: 177–200. Asher, S. and P. Novosad. 2018. “Rural Roads and Local Economic Development.” Policy Working Paper No. 8466. Washington, DC: World Bank. Athena Sustainable Materials Institute. 2006. “A Life Cycle Perspective on Concrete and Asphalt Roadways: Embodied Primary Energy and Global Warming Potential.” Ottawa: Athena Sustainable Materials Institute. Banjo, G., H. Gordon, and J. Riverson. 2012. “Rural Transport: Improving Its Contribution to Growth and Poverty Reduction in Sub-Saharan Africa.” SSATP Working Paper 93. Washington, DC: World Bank. Bardhan, P., and D. Mookherjee. 2003. “Decentralization and Accountability in Infrastructure Delivery in Developing Countries.” Bell, M.L., K. Ebisu, B. P. Leaderer, J. F. Gent, P. Koutrakis. 2014. “Associations of PM2.5 Constituents and Sources with Hospital Admissions: Analysis of Four Counties in Connecticut and Massachusetts (USA) for Persons >= 65 Years of Age.” Environmental Health Perspectives 122: 138-144. Benchmark Engineers. 2006. Paving Decisions for Laramie County Roads. Cheyenne, WY: Benchmark Engineers, P.C. Bennett et al 2002 “Innovative Initiatives in Road Design and Construction in the Province of Kwa Zulu Natal, South Africa.” Berg, C., U. Deichmann, Y. Liu, and H. Selod. 2015. “Transport Policies and Development.” World Bank Policy Research Working Paper 7366. Washington, DC: World Bank. Bishop, T., and P. Amos. 2015. “Opportunities to Improve Road Safety Through ‘Boda-Boda’ Associations in Tanzania.” AFCAP/ ASCAP Project Management Unit, Thames OX92AH. Bougheas, S., P. O. Demetriades, and T. Mamuneas. 1999. “Infrastructure, Specialization and Economic Growth.” Brenneman, A. and M. Kerf. 2002. “Infrastructure and Poverty Linkages: A Literature Review.” Washington, DC: World Bank. British Standards Institution (BSI). 1992. “Guide for Structural Design of Pavements Constructed with Clay or Concrete Block Pavers.” London: BSI. ———. 2005. “Guide for the Structural Design of Trafficked Pavements Constructed of Natural Stone Sets and Concrete Block Paving.” BS 7533 Part 10. London: BSI. Caballero, L., and N. Alcahausi. 2007. “Gender in Peru: Can Women Be Integrated into Transport Projects?” En Breve 112. Washington, DC: World Bank. 120 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE Calderon, C., and L. Serven. 2003. “The Output Cost of Latin America’s Infrastructure Gap.” The Limits of Stabilization Infrastructure, Public Deficits and Growth in Latin America. Edited by W. Easterly and L. Serven. 95–117. Washington, DC: World Bank; Stanford, CA: Stanford University Press. ———. 2004. “The Effects of Infrastructure Development on Growth and Income Distribution.” Canning, D. 1999. “Infrastructure’s Contribution to Aggregate Output.” Policy Research Working Paper 2246, Washington, DC: World Bank. Cantarero, L., and D. Mendez. 2007. “Análisis y Evaluación de las Características Geométricas de los Pavimentos de Adoquín en las Vías Rurales de Nicaragua.” Journal article? Chilean Institute of Cement and Concrete. 1991. Design and Construction Manual: 1991 Edition. Santiago: Chilean Institute of Cement and Concrete. Cook, C., T. Duncan, S. Jitsuchon, A. Sharma, and W. Guobao. 2005. Assessing the Impact of Transport and Energy Infrastructure on Poverty Reduction. Manila: Asian Development Bank. Cook, J.R., R. C. Petts, and J. Rolt. 2013. “Low Volume Rural Road Surfacing and Pavements: A Guide to Good Practice.” Corral, L., and T. Reardon. 2001. “Rural Nonfarm Incomes in Nicaragua.” World Development 29 (3): 427–442. De Janvry, A., and E. Sadoulet. 2000. “Rural Poverty in Latin America: Determinants and Exit Paths.” Food Policy 25 389–409. ———. 2001. “Income Strategies Among Rural Households in Mexico: The Role of Off-Farm Activities.” Demetriades, P. O., and T. P. Mamuneas. 2003. “Intertemporal Output and Employment Effects of Public Infrastructure Capital: Evidence from 12 OECD Countries.” Economic Journal—Royal Economic Society 110 (465): 687–712. Dercon, S., D. O. Gilligan, J. Hoddinott, and T. Woldehanna. 2008. “The Impact of Agricultural Extension and Roads on Poverty and Consumption Growth in Fifteen Ethiopian Villages.” Available at SSRN: https://ssrn.com/abstract=1259336 or http://dx.doi.org/10.2139/ssrn.1259336 Dissanayake, S. and H. S. Patel. 2016. Gravel Road Paving Guidelines. Kansas State University Transportation Center. Kansas Department of Transportation. Dominguez Gonzalez, K., A. L. Machado, B. Bianchi Alves, V. Raffo, S. Guerrero, and I. Portabales. 2020. “Why Does She Move? A Study of Women’s Mobility in Latin American Cities.” Washington, DC: World Bank. Easterly, W., and S. Rebelo. 1993. “Fiscal Policy and Economic Growth: An Empirical Investigation.” Journal of Monetary Economics 32, 417–458. Edmonds, G. n.d. Wasted Time: The Price of Poor Access. Geneva: International Labour Organization. European Environment Agency (EEA). 2018. “Transport Greenhouse Gas Emissions.” Environmental Indicator Report. Elvik, R. 2009. “The Power Model of the Relationship between Speed and Road Safety: Update and New Analyses,” TOI Report 1034. Oslo: Institute of Transport Economics. Engineers Against Poverty. “Employment-Intensive Road Construction and Poverty Alleviation: What Is the Evidence?” Environmental Protection Agency (EPA). 2018. Environmental Greenhouse Gas Emission Data. Washington, DC: EPA. Escobal, J., and C. Ponce. 2002. “The Benefits of Rural Roads: Enhancing Income Opportunities for the Rural Poor.” GRADE Working Paper 40. Lima: Group for the Analysis of Development. Espinet et al 2019. “Piloting the Use of Network Analysis and Decision Making Under Uncertainty in Transport Operations.” Policy Research Working Paper 8490. 121 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Fan, S., and C. Chan-Kang. 2005. “Road Development, Economic Growth, and Poverty Reduction in China.” Research Report 138. Washington, DC: International Food Policy Research Institute. Franklin, M., P. Koutrakis, P. Schwartz. 2008. “The Role of Particle Composition on the Association Between PM2.5 and Mortality.” Epidemiology. 19 680-689. Gannon, C., and Z. Liu. 1997. “Poverty and Transport.” INU/TWU Series Transport Publications TWU-30. Washington, DC: World Bank. Gibson, J., and S. Rozelle. 2003. “Poverty and Access to Roads in Papua New Guinea.” Economic Development and Cultural Change 52 (1): 159–185. Gonzalez-Navarro, M., and C. Quintana-Domeque. 2016. “Paving Streets for the Poor: Experimental Analysis of Infrastructure Effects.” Available at https://are.berkeley.edu/sites/are.berkeley.edu/files/RESTAT_MS_16049.pdf Gwilliam, Ken. (2011). “Africa’s Transport Infrastructure: Mainstreaming Maintenance and Management.” Washington, DC: World Bank. Henning, T., P. Kadar, and C. R. Bennett. 2006. “Surfacing Alternatives for Unsealed Rural Roads.” World Bank Transport Note TRN-33. Washington, DC: World Bank. Humphreys, M., S. Muzira, J. Bure, E. Chesheva, and A. Guxho. 2008. “Improving the Management of Secondary and Tertiary Roads in the South East European Countries.” ECA Report No. 43979-ECA. Washington, DC: World Bank. Iimi and Kulwinder. 2019. “Spatial Analysis of Liberia’s Transport Connectivity and Potential Growth.” Washington, DC: World Bank. Iimi and Rajoela. 2018. “Transport Connectivity, Medical Supplies, and People’s Health Care Access: Evidence from Madagascar.” Policy Research Working Paper No. 8487. Washington, DC: World Bank. Intech Associates, TRL, and TDSI (Transport Development Strategy Institute). 2009. Rural Transport Project 2: Rural Road Surfacing Research; SEACAP 1 Final Report, Vol. 1. Hanoi: Vietnam Ministry of Transport. Jacoby, H. 2000. “Access to Markets and the Benefits of Rural Roads.” Economic Journal 110 (7): 713–37. Jacoby, H., and B. Minten. 2009. “On Measuring the Benefits of Lower Transportation Costs.” Journal of Development Economics 89: 28–38. Jahren, C.T., D. Smith, J. Thorius, M. Rukashaza-Mukome, D. White, and G. Johnson. 2005. “Economics of Upgrading an Aggregate Road.” St. Paul, MN: Minnesota Department of Transportation. Jimenez, B. F. 2013. “Evaluating the Impacts of a Rural Roads Project in Nicaragua: Impacts on Welfare and Education.” A World Bank Study. Washington, DC: World Bank. Joshi, S. G., and A. K. Jha. 2013. “Otta Seal Experience in Nepal.” Transportation Research Board (TRB). Paper 13-1627, present- ed at the 2013 TRB Annual Meeting, Washington, DC, January 13–17. Kerali, H.R., M. S. Snaith, and R. C. Koole. 1991. “Economic Viability of Upgrading Low-Volume Roads.” Transportation Research Record 1291; 34-40. Khandker, S., Z. Bakht, and G. Koolwal, 2006. The Poverty Impact of Rural Roads: Evidence from Bangladesh. Knapton, J., and S. D. Barber. 1979. “The Behavior of a Concrete Block Pavement.” Proceedings of the Institution of Civil Engineers. 66: 277–292. 122 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE Lanjouw, J., and P. Lanjouw. 2001. “The Rural Non-Farm Sector: Issues and Evidence from Developing Countries.” Agricultural Economics 26 (1): 1-23. Lebo, J. and D. Schelling. 2001. “Design and Appraisal of Rural Transport Infrastructure: Ensuring Basic Access for Rural Communities.” Washington, DC: World Bank. Lei, L., S. Desai, and R. Vanneman. 2019. “The Impact of Transportation Infrastructure on Women’s Employment in India.” Journal of Feminist Economics 25 (4): 94-125. Liu, Z. 2000. “Economic Analysis of a Rural Basic Access Road Project: The Case of Andhra Pradesh, India.” Transport Infrastructure Note No. RT-5. Washington, DC: World Bank. Luhr, D., and B. McCullough. 1983. “Economic Evaluation of Pavement Design Alternatives for Low-Volume Roads.” Transportation Research Record. Journal of the Transportation Research Board 898: 24-29. Mason, David J., ed. 2003. Proceedings of the Ninth Regional Seminar. Vol. 1 of Labour-Based Technology: A Review of Current Practice. Maputo, Mozambique: International Labour Organization and Administração Nacional de Estradas. McCutcheon. 2008. The Generation of Productive Employment Opportunities for the Unskilled: Principles, Potential and Pitfalls of Labor-Intensive Construction. Muzira, S., D. H. De Diaz, and B. F. J. Mota. 2015. “Rethinking Rural Road Infrastructure Delivery: Case Study of a Green, Inclusive, and Cost-Effective Road Program in Nicaragua.” Transportation Research Record. Journal of the Transportation Research Board 2472 (3): 195-202. National Cooperative Highway Research Program (NCRHP). 2011. “Guide for Pavement Type Selection.” Report 703. Washington, DC: NCRHP. Nicaragua, Government of. National Human Development Plan (Plan Nacional de Desarollo Humano). Nicaragua Technical Standard NTON 12 009-12. Olivia, S., and J. Gibson. 2008. “The Effect of Infrastructure Access and Quality on Non-Farm Employment and Income in Rural Indonesia.” Porter, G. 2007. “Transport Planning in Sub-Saharan Africa.” Progress in Development Studies 7 (3): 251-257. ———. 2011. “’I Think a Woman Who Travels a Lot Is Befriending Other Men and That’s Why She Travels’: Mobility Constraints and Their Implications for Rural Women and Girls in Sub-Saharan Africa.” Gender, Place and Culture 18 (01): 65-81. Rigaud, Kanta K.; Ospina Parada, Angelica Valeria. 2020. “Africa - Enhancing Climate Resilience through Resilience Attributes: Good Practices and Guidance Note (English)”. Washington, D.C.: World Bank Group. Roberts, P., S. KC, and C. Rastogi. 2006. “Rural Access Index: A Key Development Indicator.” World Bank Transport Paper TP-10. Washington, DC: World Bank. Roughton International. 2012. Nicaragua Fourth Roads Rehabilitation and Maintenance Project Impact Evaluation Report. Hampshire, UK: Roughton International. Rukashaza-Mukome, C. T. Jahren, C. Mary, J. M. Thorius, G.D. Johnson, and D. J. White. 2003. “Cost Comparison of Treatments Used to Maintain or Upgrade Aggregate Roads.” Proceedings of the 2003 Mid-Continent Transportation Research Symposium. Ames, IA: August 2003. Southern African Transport and Communications Commission (SATCC). 2003. Guideline: Low-Volume Sealed Roads. Gaborone: South African Development Community. 123 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Schwab, K. 2013. The Global Competitiveness Report 2013–14. Full Data Edition. Geneva: World Economic Forum. Shackel (n.d.). “The Design of Interlocking Concrete Block Pavements for Road Traffic.” Skorseth, K., and Selim, A. 2005. “Gravel Roads Maintenance and Design Manual.” Brookings, SD: South Dakota Local Transportation Assistance Program. Report No LTAP-02-002. Washington, DC: Federal Highway Administration. South African Bitumen Association (SABITA). 2020. SABITA Manual 40, 2020. Design and Construction of Surface Treatments. Cape Town, South Africa: SABITA. Spencer, E. R. 2008. “RS Means Site Work and Landscape Cost Data, 2008.” Starkey, P., and J. Hine. 2014. “Poverty and Sustainable Transport: How Transport Affects Poor People with Policy Implications for Poverty Reduction. A Literature Review.” Stock, E.A., and J. de Veen. 1996. Expanding Labor-Based Methods for Road Works in Africa. A World Bank SSATP Study. Washington, DC: World Bank. Tabakovic, A., and E. Schlangen. 2015. “Self-Healing Technology for Asphalt Pavements.” Available online at https://www.researchgate.net/publication/292986082 Tuck, L., J. Schwartz, and L. Andres. 2009. Crisis in LAC: Infrastructure Investment and the Potential for Employment Generation. A World Bank Study. Washington, DC: World Bank. Uys, R.J., Van Zyl G.D., Truter, L.Y. 2013. ARRB Low-Volume Roads Symposium, Cairns, Australia Van de Walle, D., and C. Dorothyjean. 2002. “Impact Evaluation of a Rural Road Rehabilitation Project.” Washington, DC: World Bank. Van Wijk, I. 2019. “Use of Gravel Loss Prediction Models in the Design and Management of Unsealed Road Pavements.” Transportation Research Board, September 2019. Van Zyl, G.D, Van der Gryp, A, Uys, RJ. 2020. Study report. Calibration of HDM-4 Gravel Road Performance Models. Provincial Roads and Transport Management Branch, Department of Transport and Public Works. Western Cape Government. South Africa Van Zyl, G.D. Henderson M.G. and Uys J.R. 2008. Work effects of Alternative Maintenance Measures on Unsurfaced Roads in the Western Cape. 7th International Conference on Managing Pavement Assets Warr, P. 2005. “Road Development and Poverty Reduction: The Case of Lao PDR.” Research Paper 64. Manila: Asian Development Bank Institute. World Bank. Climate Change Co-Benefits Guidance. World Bank. 2018. Guidelines for the Environmentally Optimised Design of Low Volume Roads. India World Bank. Good Practices and Guidance Note for Enhancing Climate Resilience through Resilience Attributes. Sectoral Checklists World Bank. Road Costs Knowledge System (ROCKS) (database). Washington, DC: World Bank. ———. 1971. Study of the Substitution of Labor and Equipment in Road Construction: Phase I – Final Report. A World Bank Study. Washington, DC. World Bank. ———. 1974. “Study of the Substitution of Labor and Equipment in Civil Construction: Phase H – Final Report.” Staff Working Paper 172. Washington, DC: World Bank. 124 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE ———. 1988. “Road Deterioration in Developing Countries: Causes and Remedies.” A World Bank Study. Washington, DC: World Bank. ———. 1994. World Development Report 1994: Making Infrastructure Work. Washington, DC: World Bank. ———. 2004. World Development Report 2004: Making Services Work for Poor People. Washington, DC: World Bank. ———. 2006. “Project Appraisal Document, Fourth Roads Rehabilitation and Maintenance Project.” Report 35050-NI, Washington, DC: World Bank. ———. 2010a. “Mainstreaming Gender in Road Transport: Operational Guidance for World Bank Staff.” Transport Paper TP-28. Washington, DC: World Bank. ———. 2010b. Infrastructure and Growth. A World Bank Study. Washington, DC: World Bank. ———. 2010c. “Project Paper, Additional Financing Fourth Roads Rehabilitation and Maintenance Project.” Report 53694-NI. Washington, DC: World Bank. ———. 2010d. “Africa’s Infrastructure: A Time for Transformation.” Foster, V., & Briceno-Garmendia (eds). Washington, DC: World Bank. ———. 2011a. “Project Appraisal Document, Rural Roads Infrastructure Improvement Project.” Report 61418-NI. Washington, DC: World Bank. ———. 2011b. From Growth to Green Growth: A Framework. A World Bank Study. Washington, DC: World Bank. ———. 2012a. Inclusive Green Growth: The Pathway to Sustainable Development. A World Bank Study. Washington, DC: World Bank. ———. 2012b. The International Development Association and the International Finance Corporation Country Partnership Strategy (FY 2013–2017). Washington, DC: World Bank. ———. 2012c. Inclusive Green Growth in Latin America and the Caribbean. A World Bank LCSSD Study. Washington, DC: World Bank. World Bank. 2019a. Scaling Up Private Sector Participation in Road Asset Management in Sub-Saharan Africa. ———. 2019b. Spatial analysis of transport connectivity and growth potential. ———. 2019c. Beyond the Gap: How Countries Can Afford the Infrastructure They Need While Protecting the Planet. ———. 2019d. Measuring Rural Access: Update 2017/18. Washington, DC: World Bank. Zimmerman, K. A., and Wolters, A. S. 2004. Local Road Surfacing Criteria. Pierre, SD: South Dakota Department of Transportation. 125 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Annexes Photo credits: Romel Simon 126 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE Annex A: Why Invest in Rural Road Infrastructure? Table A.1 presents an extensive, but in no way exhaustive, list of some of the positive effects and impacts of rural roads.42 Table A.1: List of Positive Effects and Impacts of Rural Roads 1. Improved access to markets for agricultural produce, and the associated improvement in farm- based incomes 2. Increased access to agricultural extension services (better seeds, fertilizer, and knowledge) 3. Lower vehicle operating costs 4. Savings in travel times 5. Improved access to health care facilities and treatment (hospitals, health centers, doctors) 6. Improved access to education (schools) 7. Improved access to employment opportunities (shorter commutes to work, ability to search for more opportunities) 8. Short-term employment opportunities during construction as well as during maintenance 9. Attraction of service providers to rural or isolated settlements (professionals want to reside and work where access is not an issue) 10. Lowered costs for input and output prices (not automatic) 11. Improved mobility through access to long-distance transport services 12. Increased opportunities for tourism 13. Improved technical capacities, and the fostering of entrepreneurship in targeted local communities 14. Improved land values 15. Improved business climate and non-farm-based income opportunities 16. Improved health effects due to dust reduction 17. Improved comfort for passengers 18. Lower losses for perishable and sensitive agricultural produce 19. Lower losses of livestock during transfer to markets 20. Catalyst for other investments (electrification, water supply, sanitation) 21. Improved stability of the country overall, through connecting isolated regions and communities 22. Contribution to alleviating rural-urban migration 23. Facilitation of the delivery of emergency relief, and provision of secure escape routes during disasters 24. Encouragement of citizen participation in the delivery of public programs 42 For additional details and more detailed study result statistics, refer to Brenneman & Kerf (2002) and Humphreys et al. (2008). 127 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Annex B: Additional Case Studies The case studies in this Appendix illustrate the deci- the gauntlet from remaining unpaved to stabilizing with sions that were made regarding whether or not to pave chemicals, different types of sand seals, cape seals, chip roads, and the solutions that were adopted based on seals (including single and double-surface treatments), local conditions and costs. It is worth noting that the slurry seals, stone paving, concrete paving, asphalt choice of pavement surface type does not have to be a paving, the use of geocells with concrete surfacing, and binary choice between unpaved and paved; it can run concrete blocks (adoquines). CASE STUDY FROM GHANA The Twifo Hemang-Baakondzidzi Road Demonstration Project43 Background: As part of its support to the Department The contractor had challenges with a faulty bitumen of Feeder Roads in Ghana, the U.K.’s Department for distributor, and insufficient rolling capacity, but after International Development (DFID) selected short sec- becoming familiar with the construction technique, the tions of feeder roads to be sealed in order to reduce results after one year of laying the Otta seal were found the high maintenance costs caused by a combination of to be satisfactory. This pilot proved that using screened high rainfall (>1000 mm/year); steep gradients (>10%); gravel instead of the more traditional and costly slippery road surfaces during the rainy season due crushed aggregate, is a viable alternative to the gravel to the relatively plastic (Plasticity Index > 10) wearing road, without going for a more expensive chip seal course gravels; and dusty conditions during the dry or asphalt pavement options. The previous problems season. associated with dust, mud, and a slippery surface were ameliorated, and users—including schoolchildren, taxi Due to the high costs arising from long hauling distanc- operators, and local residents—expressed widespread es and the high cost of traditional crushed aggregates satisfaction with the work that was done. for a bituminous surface treatment (chip seal), an Otta seal solution was implemented instead. 43 Refer to Annex D of the cited report for more technical details on how the Otta seal project was delivered. 128 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE Figure B.1: Photos from Ghana Case Study Source: SSATP Workshop Report No. 10/06; 2006 Figure B.2: Tangible Benefits of Dust and Mud Suppression Source: SSATP Workshop Report No. 10/06; 2006 129 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS CASE STUDY FROM KIRIBATI (PACIFIC ISLANDS) Whalley (2016) has noted that developing countries construction and maintenance costs. This case study are often faced with a need to extend paved road net- presents a successful application of this technique in works into remote rural areas. Paved roads significantly Kiribati, a Pacific Island nation. enhance accessibility, but they also come with high- er costs than unpaved roads. One novel solution for Geocells were used on Kiribati’s Road Rehabilitation addressing the need to pave roads with low traffic vol- Project (KRRP), which was being undertaken by the gov- umes are concrete geocell pavements. Consisting of an ernment of Kiribati with the support of the World Bank. interlocking set of unreinforced concrete blocks formed Included in the project was the rehabilitation of 6.8 kilo- by a thin high-density polyethylene lattice, the resulting meters of pothole-ridden feeder roads with geocell pave- flexible concrete surface reduces the need for gran- ments. The cost was found to be 72 percent of the cost ular pavement layers, resulting in significantly lower for chip seal surfacing, and 53 percent that of asphalt. Figure B.3: Geocell and Concrete Application in Kiribati Geocell laid ready to receive concrete Finished section in Kiribati Source: Whalley (2016) 130 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE CASE STUDIES FROM NICARAGUA Pantasma – Puente La Pavona (57.5 kilometers) This collector road is in a highly productive region. The implemented solution went a long way toward However, it also lies in a zone that is particularly vulner- providing a sustainable road surface, and drastically able to natural disasters. In the rainy season, the road improved access to essential markets and services for was not passable, leading to communities being cut off. the target population. The provision of improved drain- There was high demand from the local populace for age facilities, including a hard aggregate base in certain this road to be surfaced to bring it to a more durable sections to allow for water flow beneath the pavement, standard. The Interamerican Development Bank (IDB) proved to be a crucial pavement design decision. Figure had financed the improvement of the road to a gravel A4 shows photos from this road surfacing experience, road standard with drainage facilities, but the usage of starting from the original earth road, through the stag- the road had since increased; and in 2012, the decision es of gravel surfacing, stabilization with cement, and to pave to adoquines (concrete block) standards was finally, adoquine surfacing. made with economic justification from the RED model. Figure B.4: Photos from the Nicaragua Adoquines Road Surfacing Experience Original road with difficult access conditions (earth surface) Improvement to gravel surfacing Stabilization of road base with cement Finished adoquines road section 131 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Nueva Guinea – Naciones Unidas (18.34 kilometers) Once this lesson was learned in 2013, plans for continu- ation works on the Nueva Guinea-to-Bluefields corridor Following the success in the use of concrete block were changed to concrete surfacing, and sections paving (adoquines) on many roads in Nicaragua, this where adoquines had already been laid received further solution was extended to the Nueva Guinea-Naciones sealing with a weak sand-mortar mix to prevent further Unidas Road under World Bank financing. This sem- water ingress. inal road project sought to link the Atlantic coast to the rest of the country, and this section was the first This was a major learning experience for both the to be paved. Unfortunately, in this case the transfer of government of Nicaragua and the World Bank to a pavement solution was a mistake. First, while paving NOT replicate pavement solutions just because was merited, the Nueva Guinea area has extremely they have worked elsewhere, even in the same high levels of rainfall (average annual rainfall of 2245 country; but rather to always carefully tailor solu- mm). The adoquines construction left permeable joints, tions to the local context. It also underscored the and water ingress through these joints undermined the important factor that rainfall plays in the pavement pavement, causing certain sections to fail prematurely. surface selection process. 132 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE Annex C: Road Safety Guidance Prepared by: Blair Turner, Leah Watetu Mbugua, Radoslaw Czapski, Kazuyuki Neki, and Sudeshna Mitra, Global Road Safety Facility (GRSF). INTRODUCTION However, these safety benefits are often strongly The assumption is often made that when paving outweighed by other increases in risk, the biggest dirt or unsurfaced roads with concrete or aggre- of which comes from the increase in vehicular gate surfacing, safety benefits will automatically speed with a smoother road surface. Coupled with ensue. However, the opposite can occur, especially if possible increases in traffic, the risks can be far greater other specific road safety improvements are not made than the safety benefits unless mitigating safety strate- at the same time. This Appendix provides evidence of gies are used. this issue, and is targeted at those working in low and middle-income countries (LMICs). First, information is Objective data on this increase in risk is not readily provided on the likely safety impacts from paving. Next, available for the case of paving an unpaved road, low-cost mitigation measures are presented that might but the increase in crashes has been noted by sev- help offset this risk. eral agencies. For example, the Global Road Safety Partnership45 highlighted the fact that it is frequent- LIKELY ROAD SAFETY IMPACTS FROM PAVING ly the case that road improvement and rehabilitation There is very little objective, and conclusive, infor- schemes in LMICs result in increased traffic, higher mation on the safety impacts of paving roads. speeds, and an increased number of crashes. This issue However, it is often assumed that with paving, safety has also been documented in high-income countries benefits will automatically follow.44 There are, indeed, (HICs) with federal highways in the U.S: paved roads likely to be some reductions in certain types of risks tempt drivers to travel faster.46 These studies suggest from paving. This includes added safety benefits for that to facilitate safety in the context of increased vehi- two-wheeled vehicles, including motorcycles, especially cle speed, roads must be straighter, wider, and as free when the shoulders are also paved. In addition, vehi- as possible from obstructions. cles are more able to stay in their lanes, thus avoiding oncoming traffic and roadside hazards, if they do not The evidence from HICs indicates a very strong rela- need to maneuver around road surface hazards like tionship between changes in vehicle speed and safe- potholes. Other non-motorized users (pedestrians, ty outcomes. One comprehensive analysis47 of more cyclists, and animal-drawn cart users) also benefit from than 100 prior studies identified this relationship for improved and paved shoulders, and from the provision various types of roads. For rural roads, the results indi- of sidewalks. Visibility can also be improved with less cate that for every 1 km/hour increase in speed, there is dust, which aids safety outcomes. around a 4.5 percent increase in the risk of a fatal crash. 44 World Health Organization (WHO). 2017. Powered Two and Three-Wheeler Safety: A Road Safety Manual for Decision-Makers and Practitioners. Geneva: WHO. 45 Global Road Safety Partnership. 2008. Speed Management: A Road Safety Manual for Decision-Makers and Practitioners. Geneva: Global Road Safety Partnership. 46 FHWA 2015. Gravel Roads Construction and Maintenance Guide. Washington, DC: Federal Highways. 47 Elvik, R. 2009. “The Power Model of the Relationship Between Speed and Road Safety: Update and New Analyses. TOI Report 1034. Oslo: Institute of Transport Economics. 133 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Taking the examples above, it is likely that the surface friction, reduced dust, improved geometry, chance of death or serious injury would be very etc. However, the increased risks can be substantially low on roads with speeds around 20-30 kilometers/ offset, and in the right circumstances completely offset hour, even for vulnerable road users. However, by other improvements in road design. This is of crit- increasing these speeds could result in substan- ical importance, since paving often occurs on roads tial increases in the risk of crashes, even consid- where the existing alignment is poor, and narrow cross ering the benefits provided by better quality road sections are often retained. This has the potential of surfacing. encouraging higher speeds on poor-quality roads (for example, roads with narrow lanes; with no shoulder Using the information above on the relationship or an inadequate shoulder; with sharp curves, etc.). between changes in vehicular speed and crash risks is In LMICs, there is the added risk of vulnerable road problematic in discussing the matter of paving roads in users such as pedestrians and cyclists, being present. LMICs, because the information is derived from studies High-speed motorized vehicles mixing with vulnerable in HICs only. Putting these concerns aside, however, road users bring very high risks, whether this occurs and assuming a mean speed of 40 km/hour before the between towns and villages, or within them. change in speed and one existing fatality before paving occurred, the following estimates can be made: Summary of Crash Risks When paving roads, a review of the road geometry Table C.1: Increase in Risk with Change in Speed and use by vulnerable road users should be under- Speed before Speed after Estimate of fatalities taken in order to assess any possible increases in risk. Measures should be put in place to mitigate the 40 40 1 risks, particularly those caused by any increase in speed. 40 50 2 40 60 5 Given the lack of empirical evidence, only broad guid- 40 70 12 ance can be provided on the changes in crash risk from 40 80 23 resurfacing projects. These are: 40 90 41 40 100 67 • Where speeds are likely to increase, but there is no Source: Elvik 2009 improvement in road cross section (width and align- ment), or no safety provisions for vulnerable road users, and at intersections and curves, the risk will These results are indicative only, and will depend great- increase. Therefore, the costs of crashes will also ly on a variety of other factors. However, even if they increase, often substantially. overestimate the increase in risk ten-fold, the risk at these higher speeds is still very high. • Where mitigating measures have been used, but they are only minor (such as signs and line marking) The increase in risk from a change in speed will be and speeds remain high, risks will increase. There- slightly offset by safety improvements due to improved fore, crash costs will increase, often substantially. 134 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE • Where mitigating measures are used at key loca- MITIGATION: LOW-COST INTERVENTIONS tions, including the provision of protection for This section discusses the indicative benefits for various vulnerable road users (footpaths and traffic calming low-cost interventions. Note that engineering judgment on approach to, and through areas of activity), and is required for individual applications of them. When protection is provided for road users at intersections included in the initial design, many of these interven- and curves, risks may be reduced to a level that is tions can be included at very low, or even no additional similar to those before resurfacing occurred. cost. It is also of note that some of these interventions can only be used, or are more effective on paved roads • Where there are substantial mitigating measures, than an unpaved roads (for example traffic humps and including the provision of full protection for vulner- line marking). able road users (such as footpaths and traffic calm- ing on approach, and through areas of activity), and Although more than one intervention can be used, and protection is provided for road users at intersections this typically produces greater safety benefits, these and curves, the risks are likely to be reduced com- benefits are usually not additive, since often there is a pared to the situation prior to paving. Therefore, diminishing return when using multiple treatments at crash costs are also likely to decrease. the same location. Table C.2: Throughout Routes Treatment Description CRF Image Speed Limits Involves providing a posted speed limit using 20% 48 signs so that motorized road users can better understand the required speeds. Speed limits should be aligned to road use, particular- ly where vulnerable road users are present. For this reason, localized speed limits may be required at high-risk locations. Speed limits are often more effective if applied in conjunction with physical measures to reduce speed (such as traffic calming), or as part of a package of measures. 48 Assumes a 10 km/hr reduction in the speed limit, with a moderate level of compliance. Greater benefits will occur if there is full compli- ance. 135 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Treatment Description CRF Image Shoulders A graded, or preferably a paved, shoulder can 30% be provided to assist road users in returning to their lane when they accidentally leave the road. Shoulders provide a recovery area to reduce the likelihood of vehicles losing control. They should not be so wide as to provide an extra lane, and care should be taken to ensure that roadside traders are kept clear of traffic lanes. Visual Narrowing Narrowing of roads using painted edge lines, 40% or on wider roads, using wide painted center- lines to produce the impression of a narrow lane. These work best when coupled with lower speed limits, and in very low-traffic volume environments. Table C.3. Protection for Vulnerable Road Users Treatment Description CRF Image Segregated Paths The safest option for protecting vulnerable 50% road users is to provide full separation from motorized, high-speed traffic. This can include the provision of off-road paths for pedestrians and cyclists. These paths can be separated simply by space; or a more substantive infra- structure, such as barriers or curbing, can be provided. Widened Where segregated paths cannot be provided, 30% Shoulders a graded shoulder can provide some degree of separation for vulnerable road users. Care needs to be taken that street traders do not encroach on vehicle lanes, and that safe areas are provided for setting down and picking up passengers. Traffic Calming These are minor safety works to reduce speeds 40% at key locations, ideally to 30 km/hour where vulnerable road users are present. Measures include road humps, raised platforms, road narrowing, raised pedestrian crossings, etc. 136 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE Treatment Description CRF Image Gateway Using a combination of treatments, including 40% Treatments signs, line marking, and road narrowing to slow traffic down, and to create a visual dif- ference on entering a village or other built-up area. Signs alone (on both sides of the road) do have a benefit, but threshold treatments work better when a pinch point (some form of perceived or actual road narrowing) is used. Table C.4. Safety Measures at Curves in the Road Treatment Description CRF Image Widening Curves Shoulders provide space for a recovery area 10-50% that allows drivers to leave their travel lane without losing control of their vehicle. This is especially important at curves, where vehicles use more of the travel lane than in straight sections. Centerline and This is painted line marking to help segregate 20% Edge-Line Marking vehicles travelling in different directions. Advance Signs placed on the approach to a curve 25% Warning Signs to alert drivers to a change in the horizon- tal alignment of the road. Often an advisory speed sign is also installed underneath (see below). The signs alert drivers to the presence and alignment of the curve (left curve, right curve, reverse curve, etc.), giving additional information that is helpful in safely negotiating the curve. 137 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS Treatment Description CRF Image Advance Advisory speed signs are plates, usually 40% Warning Signs attached under a warning sign, which display with Advisory the speed limit for negotiating the curve Speeds comfortably and safely. The treatment also indicates the severity of the curve, with a lower speed indicating a more severe curve. Reduc- ing speeds on the road sections preceding horizontal curves is particularly important, because excessive speed is a significant factor in crashes at curves. Chevrons Chevron alignment markers (CAMs) placed on 30% the outside of a curve help indicate both the presence and the severity of the curve. Barriers Concrete (rigid), steel (semi-rigid), and wire- 45% rope (flexible) barriers can be used to prevent road users from striking roadside hazards, or driving over embankments and cliffs. These are used at targeted locations such as high- risk curves. Route-Based Route-based treatments are a method of 40% Curve Assessment ensuring consistency in the signing of curves and Intervention along a section of road. Each curve is classi- fied based on the risk factors, such as design speed, tangent speed, sight distances, etc. Once the risk of the curve has been identified, signs and markings for that curve are installed according to each curve’s risk category. The higher the risk category, the more treatments are installed. These include advance warning signs, guideposts, chevron markers, and pro- filed road markings. 138 MOBILITY AND TRANSPORT CONNECTIVITY SERIES TO PAVE OR NOT TO PAVE Table C.5. Safety Techniques at Intersections Treatment Description CRF Image Advance Warning Warning signs are often used in advance of 30% Signs intersections to alert motorists to the possibil- ity of an increased level of risk. It is expected that such signs will raise the attention level of motorists, and it is also possible that motorists will slow to a safer speed in some circumstanc- es. Aside from the standard warning signs, a number of different sign configurations have been employed to raise awareness at particu- larly problematic locations. Improved Sight Sight distance improvements can be achieved 30% Distance by removing vegetation and other road- side objects. This allows greater visibility to help road users detect approaching vehicles and make safer decisions about entering an intersection. Increasing the Increasing the visibility of an intersection by 35% Visibility of installing signs, islands, or painting the road Intersections surface reduces the chance of road users on a side road from missing the intersection. It may also be used to increase awareness of the intersection for those on the through road, so they can either slow or be more alert when entering the intersection. Traffic Calming Minor safety works to reduce speeds 40% on approach and through intersections (to a 50 km/hr maximum) through use of raised platforms, humps, mini round- abouts, etc. 139 DEVELOPING A FRAMEWORK FOR SYSTEMATIC DECISION-MAKING IN THE CHOICE OF PAVING TECHNOLOGIES FOR RURAL ROADS