A F G H A N I S TA N E N E R G Y S T U DY A GIS APPROACH TO PLANNING ELECTRIFICATION IN AFGHANISTAN Alexandros Korkovelos, Morgan Bazilian, Dimitrios Mentis, and Mark Howells A F G H A N I S TA N E N E R G Y S T U DY A GIS APPROACH TO PLANNING ELECTRIFICATION IN AFGHANISTAN Alexandros Korkovelos, Morgan Bazilian, Dimitrios Mentis, and Mark Howells © 2017 International Bank for Reconstruction and Development/The World Bank 1818 H Street NW, Washington, DC 20433 202-473-1000 | www.worldbank.org Some rights reserved. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Nothing herein shall constitute or be considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved. Rights and Permissions The material in this work is subject to copyright. Because the World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, World Bank Group, 1818 H Street, Washington, DC 20433, USA; fax 202- 522-2625; email: pubrights@worldbank.org. Cover map: KTH dESA. ii Contents Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vi Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii Key Terminology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix 1 Background, Context, and Scope . . . . . . . . . . . . . . . . . . . . . . . . . 1 2 Geospatial Energy Planning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 2.1 GIS data collection and processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 2.2 Identifying demand . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 2.3 Electricity access targets in Afghanistan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 2.4 Identifying supply . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 2.5 Scenario formation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 2.6 Running the model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 2.7 Results and visualization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 2.8 Bridging science and policy: Interpretation of OnSSET’s results . . . . . . . 26 3 Conclusion and Final Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . 29 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 Appendixes A Detailed results of 12 representative electrification scenarios for Afghanistan (Map and tabular format) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 B Introduction to the online interface of OnSSET: An example analysis for Afghanistan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 iii Figures ES.1 Principal components and structure of OnSSET . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x 1.1 Percentage of households with access to electricity, by source and residence type . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 2.1 Principal components and structure of OnSSET . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 2.2 Total population electrified, by technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 2.3 Newly electrified population, by technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 2.4 New capacity to be added, by technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 2.5 Investment requirements, by technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 2.6 New capacity required, by system type . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 2.7 Investment required, by system type . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 2.8 Summarized results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 A.1 Scenario 1: U4–R4, LD, 0.075RE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 A.2 Scenario 2: U4–R4, HD, 0.077 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 A.3 Scenario 3: U5–R3, LD, 0.077 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 A.4 Scenario 4: U5–R3, HD, 0.077 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 A.5 Scenario 5: U5–R3, HD, 0.075RE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 A.6 Scenario 6: U5–R3, HD, 0.075IM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 A.7 Scenario 7: U3–R3, LD, 0.075RE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 A.8 Scenario 8: U3–R3, HD, 0.077 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 A.9 Scenario 9: U3–R3, HD, 0.075IM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 A.10 Scenario 10: U4–R2, LD, 0.075IM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 A.11 Scenario 11: U4–R2, HD, 0.077 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .43 A.12 Scenario 12: U4–R2, HD, 0.075IM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 Tables 2.1 Population characteristics for urban and rural settings in Afghanistan . . . . . . . . . . . . . . . . . . . . . . 8 2.2 Indicative services that might be accessible to people, by annual electricity consumption tier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 2.3 Parameters related to the extension of the national electricity grid . . . . . . . . . . . . . . . . . . . . . . . . 10 2.4 Electricity generation technology parameters used in the model . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 2.5 Characteristics of the national power grid in Afghanistan, 2015 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 2.6 Forecast of national power grid based on current government plans, 2030 . . . . . . . . . . . . . . . . . 16 2.7 Forecast of national power grid with greater renewable energy penetration, 2030 . . . . . . . . . . 16 2.8 Forecast of national power grid with increased imports, 2030 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 2.9 Grid cost ($/kWh) under 32 electrification scenarios investigated for Afghanistan, by diesel price and electricity consumption tier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 A.1 Scenario 1: Electrification results under the scenario defined by urban demand at tier 4, rural demand at tier 4, low diesel price, grid cost at 0.075 $/kWh, and higher penetration of renewable-based technologies in the grid mix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 A.2 Scenario 2: Electrification results under the scenario defined by urban demand at tier 4, rural demand at tier 4, high diesel price, and grid cost at 0.077 $/kW . . . . . . . . . . . . . . . . 34 iv A.3 Scenario 3: Electrification results under the scenario defined by urban demand at tier 5, rural demand at tier 3, low diesel price, and grid cost at 0.077 $/kWh . . . . . . . . . . . . . . . . 35 A.4 Scenario 4: Electrification results under the scenario defined by urban demand at tier 5, rural demand at tier 3, high diesel price, and grid cost at 0.077 $/kW . . . . . . . . . . . . . . . . 36 A.5 Scenario 5: Electrification results under the scenario defined by urban demand at tier 5, rural demand at tier 3, high diesel price, grid cost at 0.075 $/kWh, and higher penetration of renewable-based technologies in the grid mix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 A.6 Scenario 6: Electrification results under the scenario defined by urban demand at tier 5, rural demand at tier 3, low diesel price, grid cost at 0.075 $/kWh, and increased imports in the grid mix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 A.7 Scenario 7: Electrification results under the scenario defined by urban demand at tier 3, rural demand at tier 3, low diesel price, grid cost at 0.075 $/kWh, and higher penetration of renewable-based technologies in the grid mix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 A.8 Scenario 8: Electrification results under the scenario defined by urban demand at tier 3, rural demand at tier 3, high diesel price, and grid cost at 0.077 $/kWh . . . . . . . . . . . . . . . 40 A.9 Scenario 9: Electrification results under the scenario defined by urban demand at tier 3, rural demand at tier 3, high diesel price, grid cost at 0.075 $/kWh, and increased imports in the grid mix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 A.10 Scenario 10: Electrification results under the scenario defined by urban demand at tier 4, rural demand at tier 2, low diesel price, grid cost at 0.075 $/kWh, and increased imports in the grid mix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 A.11 Scenario 11: Electrification results under the scenario defined by urban demand at tier 4, rural demand at tier 2, high diesel price, and grid cost at 0.077 $/kW . . . . . . . . . . . . . . . . 43 A.12 Scenario 12: Electrification results under the scenario defined by urban demand at tier 4, rural demand at tier 2, high diesel price, grid cost at 0.075 $/kWh, and increased imports in the grid mix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 v Foreword Access to affordable and reliable electricity is essential to the success of any economic growth strategy. Yet the percentage of the population with access to grid electricity in Afghanistan is among the lowest in the world. Annual per capita consumption averages 186 kilowatt-hours (kWh), far below the global average of 3,126 kWh. In rural areas, where three-quarters of Afghans live, only 11 percent are connected to the grid. Using an original tool and a new database of geospatial data, this study offers an initial analysis of the technological options and investment requirements needed to expand electricity access across Afghanistan. As part of a larger effort—the Afghanistan Energy Study—it will enable the government to “ensure access to affordable, reliable, sustainable and modern energy for all” (Sustainable Development Goal 7). Afghanistan’s diversity in terms of demographic attributes, terrain types, wealth levels, access to infrastructure, resource availability, and other factors complicates the planning of electrification. To plan in a way that will be persuasive to investors, these diverse factors need to be captured quantitatively and with local specificity using a modeling platform that will allow users to view, share, and modify underlying data and assump- tions. But acquiring energy-related data at the local level can be a challenging task, particularly in a country like Afghanistan. Fortunately, new geospatial tools have greatly reduced the costs of mapping resources and compiling geospatial datasets, thereby facilitating the creation of electrification plans with “investment grade” specificity and accuracy. In 2017, the Division of Energy Systems Analysis of the KTH Royal Institute of Tech- nology in Sweden (KTH dESA), with assistance from the World Bank, moved the electrification-planning process in Afghanistan a big step forward by building a database of geospatial information and helping Afghan planners create a modifiable least-cost electrification model, complete with targets and timetables. This model, known as the Open Source Spatial Electrification Tool (OnSSET), estimates and analyzes the most cost- effective electrification options for the achievement of electricity access goals. In Afghanistan, a plan that relies solely on grid expansion can be expected to increase the rate of electrification only slowly, particularly if donor financing for large infrastructure investment dwindles. A systematic off-grid plan that is implemented concurrently with the grid-expansion plan can help ensure that affordable, basic electricity services are made available to a wider segment of the population. I am happy to say that this study has brought such a plan within reach. vi H.E. Eng. Ali Ahmad Osmani Minister of Energy and Water Acknowledgments This report was prepared by Alexandros Korkovelos (KTH), Morgan Bazilian (World Bank), Dimitrios Mentis (KTH), and Mark Howells (KTH) under the KTH/World Bank Project Agreement for the “Afghanistan Energy Study: Activity 4.” The study benefited greatly from valuable comments and suggestions from Fanny Missfeldt-Ringius (World Bank). Several Afghan GIS experts from the intergovernmental working group of the Afghanistan Energy Study helped in calibrating the model presented in this study and in validating the study’s results. Important contributions were received from Walker Bradley (World Bank) and Niki Angelou (World Bank). This study would not have been possible without their generous and insightful cooperation. None of these individuals should be held responsible for any remaining errors in the study, for which the authors are solely responsible. The financial and technical support of the Energy Sector Management Assistance Pro- gram (ESMAP) is gratefully acknowledged. ESMAP—a global knowledge and technical assistance program administered by the World Bank—assists low- and middle-income countries to increase their know-how and institutional capacity to achieve environmen- tally sustainable energy solutions for poverty reduction and economic growth. ESMAP is funded by Australia, Austria, Denmark, the European Commission, Finland, France, Germany, Iceland, Japan, Lithuania, Luxembourg, the Netherlands, Norway, the Rock- efeller Foundation, Sweden, Switzerland, the United Kingdom, and the World Bank. vii Abbreviations BoS balance of system GDP gross domestic product GIS geographic information systems GW gigawatt GWh gigawatt hour HH household HV high voltage km kilometer kW kilowatt kWh kilowatt-hour LCoE levelized cost of electricity LV low voltage MV medium voltage MW megawatt OnSSET Open Source Spatial Electrification Tool OSeMOSYS Open Source Energy Modelling System PV photovoltaic Key Terminology Centralized electricity generation: Refers to the large-scale generation of electricity at centralized facili- ties, located usually away from end-users and connected to a network of high-voltage transmission lines (US EPA 2017). Distributed electricity generation: Refers to a variety of technologies that generate and distribute electricity at or near where it will be used. It may serve selected loads in the vicinity or it may be part of a greater system (regional and/or national grid) (US EPA 2017) (Pepermans 2005). Under this perspective and for the purposes of this report we define the following: National grid (or grid): A system of centralized and distributed electricity generation facilities that are inter-connected through an extensive transmission network spreading throughout the country. Mini-grids: Isolated power generation-distribution systems that are used to provide electricity to local communities (power output ranging from kilowatts to multiple megawatts) covering domestic, commer- cial and/or industrial demand. Stand-alone systems: Small power systems that are not tied to the national grid, operate autonomously on island mode, and can satisfy on site, low electricity demand for a limited time. viii Executive Summary This study explores the technological options and investment requirements needed to boost electricity access levels in Afghanistan and presents a method for performing a spatial based electrification analysis. As part of the World Bank’s wider effort to inform investments focused on increasing Afghans’ access to affordable and sustainable energy, the study offers an initial, “quick pass” at selected data to provide a sense of scale and to inform a more detailed analysis to be performed at a later date. The recent experience of electricity utilities in many developing countries has shown that spatial diversity—of demographic attributes, terrain types, wealth levels, access to infrastructure, and resource availability, among other factors—affects the planning of electrification. These factors need to be captured quantitatively and with local specificity, using a data-modeling platform that allows users to view, share, and modify underlying data and assumptions. The widespread availability of new and openly accessible geospa- tial information and tools reduces the costs of mapping resources and greatly assists in establishing and maintaining geospatial datasets that enable the rapid creation of electri- fication plans with quantitative and spatial specificity and accuracy. Recent global experience shows that the most effective and efficient way of achieving a rapid increase in electrification is through a sector-wide approach in which on- and off- grid electrification is pursued in a complementary manner, while also taking exogenous variables like security issues and climate change into account. Such an approach suggests solutions in line with a national-level, least-cost electrification plan. In the case of Afghanistan, a plan that relies solely on grid expansion can be expected to increase the rate of electrification only slowly, especially if available donor financing for large infrastructure investment becomes increasingly scarce. A systematic off-grid plan to be implemented concurrent with grid expansion would help ensure that affordable, basic electricity services are made available to a wider segment of the population. In 2017, the Division of Energy Systems Analysis at the KTH Royal Institute of Technol- ogy in Sweden (KTH dESA) took the electrification planning process in Afghanistan one step further by building a database of geospatial information, and helping planners create a modifiable least-cost electrification model. This model, known as the Open Source Spatial Electrification Tool (OnSSET), estimates, analyzes, and visualizes the most cost-effective electrification option for the achievement of electricity access goals. The geodatabase is available at https://energydata.info. Data were gathered from different years (2011–2016). ix Figure ES.1. Principal components and structure of OnSSET WorldPop, UN, World Bank, World Bank IEA Population Electricity access tiers Electricity demand OSM, DABS, NASA, SRTM, JRC, Natural OSM, DABS, ADB GADM NASA Merra NASA Langley ADB, USGS HydroSHEDS, JRC Earth, IEA Existing and Power plants Small/mini- Wind Diesel cost of planned and Administrative Solar hydro capacity generating transmission economic areas irradiance potential factor electricity network activities Grid Mini-grid Stand-alone Technology selection Optimal electrification split Source: KTH dESA. Note: ADB = Asian Development Bank; DABS = Da Afghanistan Breshna Sherkat; GADM = Global Administrative Areas; HydroSHEDS = Hydrological data and maps based on SHuttle Elevation Derivatives at multiple Scales; IEA = International Energy Agency; JRC = Joint Research Centre; OSM = Open Street Maps; Merra = Modern-era Retrospective Analysis; NASA SRTM = National Aeronautics and Space Administration Shuttle Radar Topography Mission; UN = United Nations; USGS = United States Geological Survey. The objective of the electrification analysis is to identify the most economic electric- ity supply mix that will allow full electrification of Afghanistan by 2030. It is unlikely that the employment of only one technology can achieve this goal. Every location (or geospatial unit) in the country has different characteristics, some of which may favor one technology over another. OnSSET considers seven technological options; these are arranged into three main electrification categories: grid, mini-grid, and stand-alone systems. (Stand-alone systems are usually a good option for remote, sparsely populated areas with limited electricity consumption needs.) OnSSET uses the levelized cost of electricity (LCoE) calculated for each geospatial unit and then finds the lowest-cost option for a particular location under all scenarios con- sidered in the exercise or defined by the modeler. To date 32 scenarios have been devel- oped for Afghanistan, considering variations of diesel price, grid cost1 and electricity 1. The cost of electricity production (US$/kWhel) by the centralized national grid. x consumption level. A few of them are presented in this study as a showcase of OnSSET’s capabilities. The results are graphically represented using maps, and are also available in tabular format so as to facilitate further analysis. Figure ES1 schematically represents the model’s main methodological processes. What is the cost of ensuring access to electricity for all Afghans by 2030? The total investment required ranges between $7.82 billion and $26.04 billion over the period 2015–30. The assumed level of electricity demand per household in each settlement, is an important factor determining which technology offers the lowest cost. At the lowest consumption levels, most population settlements close to already-electrified villages and transmission lines will find that connecting to the central electricity grid is the lowest-cost option. Elsewhere, most settlements will find that stand-alone systems are the most eco- nomical option (with PV panels a better option than diesel gensets, especially when diesel prices are high). At this low level of consumption, mini-grids play only a minor role. Assuming a midrange electricity consumption per household, makes grid connection a more viable option (to the detriment of stand-alone options). Assuming high con- sumption levels furthers the viability of increasing connections to the central grid; but, interestingly, higher consumption levels imply that mini-grids become an economically attractive option and replace stand-alone technologies in many settlements. *** OnSSET’s open-source features allow energy experts to refine results resolution and explore additional scenarios. KTH dESA has been developing an online open interface to support the use of OnSSET by professionals without knowledge or experience using geospatial software. The interface allows the user to conduct an electrification analysis of a selected country based on few key input parameters (energy access targets, population characteristics, technology costs). The results can be visualized quickly in an embedded map that shows at a glance the most cost-effective electrification pathways. This interface is accessible at OnSSET.org. OnSSET is a work in progress, especially as new satellite imagery and GIS data become available. The current analysis has limitations, of course; these, along with possible solutions, are described in the report. xi 1. Background, Context, and Scope With a gross national income per capita of $580,1 Afghanistan is the lowest-income coun- try in the South Asia region. Wracked by more than three decades of conflict, it remains an extremely fragile state and faces enormous development challenges, including high levels of poverty (39.1 percent) and unemployment (22.6 percent) (CSO 2016). Despite significant institutional advances and rapid economic growth until 2012, the trend was significantly reversed in 2013, due to declining international spending, worsening conflict and growing overall uncertainty. Following annual GDP growth averaging more than 9 percent between 2003 and 2012, growth dropped to 1.7 percent between 2013 and 2016 (World Bank 2017b). With foreign aid declining and the labor force expanding by about 300,000 a year, Afghanistan urgently needs to find ways to sustainably accelerate broad-based growth in the medium term—implying, among other things, an adequate and stable electricity supply to meet growing demand. But even under reasonably optimistic scenarios, GDP growth in Afghanistan is projected to fall from a 10-year average of more than 9 percent to between 5 and 6 percent over 2011–18. Unemployment is projected to rise further, with potentially destabilizing effects. In this context, Afghanistan is seeking ways to accelerate growth through increased private and public investment, with a particular focus on addressing the country’s severe infrastructure bottlenecks. Access to affordable and reliable electricity is essential to the success of any economic growth strategy. Afghanistan has made impressive progress in improving electric- ity access. In 2013–14, 90 percent of Afghan households had access to some source of electricity, as compared with only 42 percent in 2007–8 (CSO 2016). This huge gain was mainly driven by improved access to electricity in urban areas, where 99 percent of the population had access to electricity, and, more recently, through the promotion of community-level micro-hydropower and solar systems in rural areas, where the overall electricity access rate reached 88 percent. Nonetheless, rates of access to the electric grid are somewhat different. Although grid electricity is available to 88 percent of the urban population, it remains a serious chal- lenge in rural areas, where more than 73 percent of Afghans live and only 11 percent are connected to the grid (figure 1.1). Low access to grid electricity is also reflected in consumption levels. At the national level, per capita consumption is estimated at 186 kilowatt-hours (kWh) per year,2 which is significantly lower than the South Asia average 1. 2016 data (World Bank 2017a). 2. Based on authors’ calculations. 1 2 A GIS Approach to Planning Electrification in Afghanistan Figure 1.1. Percentage of households with access to electricity, by source and residence type 99 90 88 88 Any source Solar and wind Electric grid 60 Battery Percentage Generator 48 29 12 12 14 13 10 11 4 4 National Urban Rural Source: CSO 2016.  of 707 kWh per year and far below the global average of 3,1263 kWh. Thus, access rates alone do not provide sufficient indication of the quantity and quality of the electricity supplied. Even households with access to the electricity grid suffered prolonged power outages in the past, but the situation has improved significantly in the major cities and towns along the critical northeast corridor between Mazar-e-Sharif and Kabul, following the import of power from Uzbekistan and the rehabilitation of three hydropower plants (work on Mahipar and Sarobi is complete; Naghlu is in progress). Parts of some cities, for example, Kabul, Herat, Mazar-e-Sharif, and Pul-e-Khumri, now have a 24-hour power supply for the first time in decades. In light of budget restrictions and overall insecurity, several least-cost electricity plans have been developed over the past few years. These include a 2013 Power Sector Master Plan supported by the Asian Development Bank (ADB 2013), and the recommenda- tions of a World Bank study on uncertainty in Afghanistan’s power sector (World Bank forthcoming). The present study will build on the lessons of these previous efforts and add relevant information on sector capacity, with the aim of supporting the sustainable development of Afghanistan’s power sector. On a national level, one significant barrier to Afghanistan’s electrification is the lack of a “bankable” investment plan. While a least-cost expansion plan exists, it has not yet been translated into actionable targets and timetables. On a local level, most efforts to improve rural access to electricity were driven by support from the National Solidarity Program launched in 2003. The program promoted local—community based—governance over 3. 2014 data (World Bank 2017c). Background and Context 3 infrastructure development (Yemak, Gan, and Cheng 2013). Efforts focused mainly on the provision of electricity access through off-grid solution (e.g. solar power). As a result, 88 percent of the rural population in Afghanistan had access to electricity in 2013–14, amid which 60 percent obtained access through solar power and only 11 percent through the electric grid (figure 1.1). This study aims to develop a framework that allows for a “quick” electrification analysis and provides, useful insights into the technological options and investment requirements necessary to boost electricity access levels in Afghanistan. It is part of the World Bank’s wider effort to provide an updated assessment of the energy sector to the government so as to inform investments focused on increasing access to affordable and sustainable energy. This effort—called the Afghanistan Energy Study—is a five-part series of comple- mentary assessments and surveys being conducted over a period of four years (from June 2015 to June 2019). Its five parts are as follows: 1. Transactional advice and knowledge sharing 2. Financial, economic, and community assessment 3. Collection of household and enterprise energy diaries 4. Development of a least-cost electrification plan and investment prospectus 5. Institutional assessment This assignment corresponds to the fourth of these five parts. The recent experience of electricity utilities in many developing countries has shown that spatial diversity—of demographic attributes, terrain types, wealth levels, access to infrastructure, and resource availability, to name a few factors—affects the planning of electrification. These factors need to be captured quantitatively and with local specificity, using a data-modeling platform that allows users to view, share, and modify underlying data and assumptions. The widespread availability of new and low-cost geospatial information and tools greatly reduces the costs of mapping resources and establishing and maintaining geospatial datasets. This allows the rapid creation of electrification plans with quantitative and spatial specificity and accuracy. These relatively low-cost and improved tools also make it easier for otherwise underfunded institutions to establish and maintain datasets on infrastructure, social services (education, health), and other key resources essential for development. While the quantity and quality of efforts to gather geospatial data have improved in many developing countries around the world, much of these data have been captured in “one-off” events, creating datasets that become quickly out of date as populations grow and infrastructure is built. Going forward, a key challenge is how to provide flexible, updateable systems that will allow local practitioners to correct errors as well as to add 4 A GIS Approach to Planning Electrification in Afghanistan and update data incrementally as local conditions change. Maintaining and storing such a database at a national level requires a Web-based energy access platform. Recent global experience shows that the most effective and efficient way of achieving a rapid increase in electrification is through a sector-wide approach in which both on- and off-grid-based electrification strategies are pursued in a complementary manner, while taking exogenous variables like security issues and climate change into account (World Bank 2017d). Under such an approach, implementation will be channeled toward the deployment of solutions in line with a national-level, least-cost electrification plan, as well as financial and physical resources mobilized in a predictable and structured fash- ion, while allowing for uncertainty in security conditions. This study offers an initial, “quick pass” analysis of selected data to provide a sense of scale and to inform a more detailed analysis. In the case of Afghanistan, a plan that relies solely on grid expansion alongside coordi- nated investments in generation and transmission can be expected to increase the rate of electrification only slowly, especially if available donor financing for large infrastructure investment becomes increasingly scarce. A systematic off-grid plan that is implemented concurrent with the grid expansion plan would help ensure that affordable, basic elec- tricity services are made available to a wider segment of the population. Experience has shown that effective electrification plans build local capacity to undertake requisite energy planning exercises and tasks. This study, therefore, recommends the provision of a training package to equip the staff of relevant Afghan agencies with the necessary knowledge and tools to continue the work of sector planning in the future. Such a training package would feature easy-to-understand documents that minimize jargon. The least-cost electrification plan suggested here would serve as a fundamental part of any high-level investment prospectus for Afghanistan. To set short-term, actionable milestones, more detailed, province-level analysis is needed. This will be conducted in a later stage of this project. 2. Geospatial Energy Planning To take on the challenge of developing energy infrastructure, plan its long-term develop- ment, and make it climate resilient, national governments must answer several policy and investment questions. Will available resources be enough to meet growing demands and development needs? What will the environmental and economic costs of energy transitions be? Can a trade-off between them be found? A quantitative approach is necessary because energy system planning is essential in order to match supply with growing demand, and in the most cost-effective way. In addition, moving from planned, centralized, and expensive energy carriers toward fluctuating, decentralized, and cost-effective renewable energy production necessitates considerable modifications in energy infrastructure that must be carefully planned for optimal results. These modifications are most often motivated by geospatial concerns. Therefore, ground-level geospatial data are of key importance to help identify the most effective electrification strategy. Unfortunately, the acquisition of energy-related data at the local level is a challenging task, especially in countries where universal access to electricity has not yet been achieved. This is where a Geographic Information System (GIS) can be an asset. The integration of GIS into energy planning can have several advantages; for example, spatial data can be used to analyze demand at a particular location, while making projections that consider the location’s unique characteristics (for example, position in an urban or rural area) and corresponding energy access targets. Furthermore, GIS takes into account resource avail- ability and energy potentials. Renewable energy maps, for example, are overlaid with several socioeconomic and geographic restrictions yielding technical energy potentials at the local level in areas where such data would otherwise not be available. Moreover, GIS can be used to illustrate results in interactive maps. These graphics communicate the key indicators for electrification planning “at a glance,” and can be easily understood by policy makers with time constraints. In 2017, the Division of Energy Systems Analysis within the KTH Royal Institute of Tech- nology in Sweden (KTH dESA)4 took the electrification planning process in Afghanistan one step further by building a database of geospatial information, and helping planners create a modifiable least-cost electrification model. This model is known as the Open Source, Spatial Electrification Tool (OnSSET), and it estimates, analyzes, and visualizes the most cost-effective electrification option for the achievement of electricity access goals. The tool is focused on the assessment and deployment of primarily renewable 4. https://www.kth.se/en/itm/om/organisation/institutioner/energiteknik/forskningsavdelningar/desa. 5 6 A GIS Approach to Planning Electrification in Afghanistan technologies to “ensure access to affordable, reliable, sustainable and modern energy for all” (Sustainable Development Goal 7).5 This section will outline the methodology behind OnSSET; its application to geospatial electrification analysis; and the key, discernible steps—data mining, GIS processing, model structuring and calibration, scenario building, result aggregation, and visualiza- tion—of the analysis process, with a focus on the particular case of Afghanistan. 2.1. GIS data collection and processing OnSSET is a GIS-based tool and therefore requires data in a geographical format. In the context of the power sector, necessary data include those on current and planned infrastructure (electric grid networks, road networks, power plants, public facilities), population characteristics (distribution, location), economic and industrial activity, and local renewable energy flows. Before a model can be built, one must acquire the “layers” of data outlined OnSSET, as any other above. More often than not, each layer must be acquired on its own. For example, one layer may be administrative boundaries, another the coordi- quantitative energy nates of population settlements, and another the number of people in these settlements. Other useful data include the location of existing power plants modelling, requires and transmission networks; the transportation infrastructure; solar irradiation data acquisition levels; wind speeds; hydrological potential; and other relevant geospatial information. The final outcome is a multilayer map conveying all the informa- and a constant data tion necessary to initiate an OnSSET electrification analysis. adjustment and The spatial resolution of the final map depends on the availability of input updating. data and on the targeted level of accuracy. OnSSET can handle various levels of input data, with typical resolutions ranging from 1x1 kilometers (km) to 10x10 km. The selection of inputs usually involves a trade-off between the time needed for compu- tation and the desired level of detail. The modeler has to decide which resolution best fits the purpose of the analysis. All analyses using OnSSET require that the following layers be obtained and processed: 1. Administrative boundaries 2. Population distribution and density 3. Nighttime light maps 4. Land cover 5. Digital elevation model 5. UN Sustainable Development Goals: http://www.un.org/sustainabledevelopment/energy. A preview of this work can be found in (Mentis et al. 2015) and (Fuso Nerini et al. 2016). Geospatial Energy Planning 7 6. Mini/small hydropower potential (with restrictions) 7. Solar irradiation (with restrictions) 8. Wind power capacity factor (with restrictions) 9. Travel time to nearest town 10. Road network (existing and planned) 11. Transmission network (existing and planned) 12. Power plants (existing and planned) 13. Substations (existing and planned) 14. Quarries and mines KTH dESA and the World Bank have collaboratively collected and processed the data needed for these layers for Afghanistan, with the aim of representing the status of the country’s energy sector today as accurately as possible, given data constraints. The layers populate a geodatabase available at https://energydata.info. Please note that data were gathered from different years (2011–2016) and are considered the “best available” as of December 2016. Note that the results presented hereafter are based on the developed geodatabase. The combination of datasets with varying spatial-temporal resolutions and geographic projec- tions may have led to compounding inaccuracies and imprecisions, fact that should be taken into account when interpreting the results. Despite this, the use of the geodatabase by a wider audience is highly recommended. That said, the modeler may choose to calibrate and/or reconstruct the map. Two workshops organized in February and July 2017 as part of the Afghani- Demographic data are stan electrification project, aimed to address data acquisition challenges, demonstrate representative data preparation steps, and point to readily crucial for modelling available open-source GIS layers.6 future demand for 2.2. Identifying demand electrification of An important parameter for identifying least-cost electrification technolo- households. gies is electricity demand. Future residential electricity demand is a func- tion of projected population growth and specific assumptions regarding demand (based on, for example, the appliances that households might use). In the residential sector, universal access to electricity does not imply that living and income standards will be uniform across all settlements. Forecasting a population’s size and purchasing power is central to any electrification analysis. Modeling the dynamics of electrification over time benefits from the best possible population estimates. The number of people in a given area, and their income, 6. The workshops were held in New Delhi on February 1–2, 2017 and Dubai on July 11–13, 2017 (training and educational material are available upon request). 8 A GIS Approach to Planning Electrification in Afghanistan Table 2.1. Population characteristics for urban and rural settings in Afghanistan Parameter Metric Value 2015 Value 2030 Population, total Million persons 32.527 a 44.310 (estimated based on growth rates below) Urban population Percent of total population 26.3% b 35.8% (estimated based on growth rates below) Rural population Percent of total population 73.7% b 64.2% (estimated based on growth rates below) Urban growth Percent growth per year 3.96%b 3.49% (average value used in the model as 3.65% per year)b Rural growth Percent growth per year 1.85% b 1.12% (average value used in the model as 1.35% per year)b Electricity access Percent of total population 29%,c, d (access to 100% the national grid) Electricity access, urban Percent of urban population 88%d (access to 100% the national grid) Electricity access, rural Percent of rural population 11%d (access to 100% the national grid) People per household, urban People per household 7.4d,e 7 (assuming 5% decrease over the 15-year period)f People per household, rural People per household 8.5d,e 8.1 (assuming 5% decrease over the 15-year period)f Sources: a. World Bank 2017e; b. UN DESA 2014, 2015; c. Infrastructure Development Cluster 2012; d. CSO 2016; e. The Asia Foundation 2015; f. Ellis and Roberts 2016. are key drivers of future demand for electricity and thus the pay-back period for capital investments. But estimating population growth is not a straightforward task. Changes in socioeconomic conditions make the estimation of future fertility and mortality rates—as well migratory patterns—a complex task. Ideally, one would estimate the future population by geospatial location. In OnSSET, urban population growth estimates are separated from rural, since these two groups usually follow slightly different growth profiles. Table 2.1 shows the population charac- teristics considered in the case of Afghanistan. Following population estimates, the next step entails the estimation of electricity con- sumption in urban and rural settings. OnSSET uses five tiers for household electricity consumption, starting from very low to high consumption standards. Following the “Sustainable Energy for All” Global Tracking Framework (IEA and World Bank 2015), the model groups the assumed consumption benchmarks into tiers (see table 2.2). The lowest assumed consumption allows for no more than low-consumption tasks, such as turning on a light for a few hours or charging a mobile-phone or radio battery. The highest consumption tier allows for energy services such as continuous lightning and the running of a refrigerator, air conditioner, and so on. The model assumes that at the end year (here 2030) all persons gaining access to electricity will reach the same assumed consumption benchmark. Geospatial Energy Planning 9 Table 2.2. Indicative services that might be accessible to people, by annual electricity consumption tier Access level Tier 1 Tier 2 Tier 3 Tier 4 Tier 5 Indicative appliances Task lighting + General Tier 2 + Tier 3 + Tier 4 + powered phone charging lighting + fan + medium power Medium or continuous High power or radio television appliances appliances and continuous (i.e. general (i.e. water heating, appliances food processing, ironing, water pumping, (i.e. air conditioning) refrigeration) rice cooking, microwave) Consumption per capita and 7.7 43.8 160.6 423.4 598.6 year (kWh) Consumption per urban 54 307 1,124 2,964 4,190 household and year in Afghanistan (based on average household size: 7) Consumption per rural 62 355 1,301 3,430 4,849 household and year in Afghanistan (based on average household size: 8.1) Source: Adapted from the Global Tracking Framework (SE4ALL, 2015). 2.3. Electricity access targets in Afghanistan It is estimated that approximately 29 percent of the Afghan population has access to the national electric grid (ANPDF 2016; Infrastructure Development Cluster 2012; CSO 2016). The access rate is higher for urban households (approximately 88 percent) than for rural households (approximately 11 percent) (Infrastructure Development Cluster 2012; CSO 2016). The electricity consumption of connected households varies significantly across provinces. For example, annual electricity consumption can range from as low as 178 kWh per household in Ghor, and 551 kWh/household in Laghman, to comparatively higher levels in urban centres such as Kabul (3,000 kWh per household) and Herat (2,600 kWh per household) (ADB 2013). Interestingly, households not connected to grid electricity seem to have access to some source of electricity, mostly solar and wind power, as well as batteries. About 60 percent of Afghan households have access to electricity through such sources (see figure 1.1). This electricity is primarily used for lighting or occasionally to power low-consumption household devices (a mobile phone, fan, or radio). The use of these devices does not exceed 300 kWh per year for the average household not connected to a grid (Infrastruc- ture Development Cluster 2012). 10 A GIS Approach to Planning Electrification in Afghanistan Based on the Power Sector Master Plan estimates (ADB 2013), the average electricity consumption in Afghanistan by 2030 will be approximately 1,500 kWh/household/year (Infrastructure Development Cluster 2012). Using this estimate, tier 5 (4,190 kWh/house- hold/year) and tier 3 (1,301 kWh/household/year) were selected as base electricity access targets for urban and rural households, respectively. (See table 2.2 for tier definitions.) 2.4. Identifying supply The objective of the electrification analysis is to identify the most economic electricity supply mix that will allow full electrification of Afghanistan by the end year. It is unlikely that the employment of only one technology would achieve this goal. Every location has different characteristics, some of which might favor one technology over another. OnSSET considers seven technological options; these are arranged into three main electrification categories: grid, mini-grid, and stand-alone systems. 2.4.1. Grid This option entails the extension of the national electricity grid to settlements not yet connected. Grid extension is a capital-intensive option; however, due to economies of scale in power generation, it can provide low generating costs. Previous electrification efforts have shown that extending the electricity grid is a good option where electric- ity consumption levels are relatively high (for example, in urban and highly populated areas) or populations live relatively close to Table 2.3. Parameters related to the current grid lines (for example, within 10 km). Table 2.3 provides extension of the national electricity grid indicative values of the parameters that OnSSET considers when estimating the cost of grid extension to unelectrified areas. Parameter Cost unit HV lines (>110kV) 120,000 $/km 2.4.2. Mini-grid MV lines (20 kV) 9,000 $/km This option entails small-scale, isolated grids, able to cover the LV lines (220 V) 5,000 $/km demand of a cluster of households. The model named four types MV/LV transformer 3,500 $/unit (50 kVA) of resources—namely, wind, solar photovoltaic (PV), hydropower, Transmission losses 18.3% and diesel generators (also called gensets)—for these small-scale Connection cost per HH $122 grids. Future iterations of the model might include biomass Cost of producing electricity 0.062–0.077 $/kWha options as well as hybrid solutions combining two or more avail- able resources. Mini-grids are a good option where electricity Source: ADB (2013); a. This value was estimated based on author’s calculations consumption is moderate and the renewable resources in question available at: https://energydata.info. are abundant. Note: HH = household; HV = high voltage; kV = kilovolts; LV = low voltage; MV = medium voltage. 2.4.3. Stand-alone Stand-alone systems refer to low-capacity off-grid options used to cover the demand of single households. Mini PV installations and small diesel gensets are currently included in the model. Stand-alone systems are usually a good option for remote, sparsely populated areas with limited electricity consumption needs. Table 2.4 Geospatial Energy Planning 11 Table 2.4. Electricity generation technology parameters used in the model Plant O&M costs Fuel cost capacity (% of investment $/liter Life Plant type (kW) Investment cost ($/kW) cost/year) (future) Efficiency % Capacity factora (years)b Diesel genset 100 1,200 10.0 0.69 (1.00) 37 0.7 15 Mini grid (ADB 2013) (ADB 2013) Small hydro 1,000 2,500 2.0 - - 0.5 30 Mini grid (IRENA 2015) Solar PV 100 2,600 1.8 - - Obtained for 20 Mini grid (ADB 2013; IRENA 2015) each grid point depending on solar availability Wind 100 2,300 3.5 - - Obtained for 20 turbines (ADB 2013; IRENA 2015) each grid point Mini grid depending on wind availability Diesel genset 1 2,000 10.0 0.69 (1.00) 28 0.5 10 Stand-alone (ADB 2013) (ADB 2013) Solar PV 0.4 5,500 1.8 - - Obtained for 15 Stand-alone (ADB 2013) Including BoS each grid point costs. depending on solar availability Sources: a Adopted from IRENA (2015); b Adopted from ESMAP (2017). Note: BoS = Balance of System; kW = kilowatts; PV = photovoltaic. lists the technical and economic characteristics of the technologies considered in the electrification analysis of Afghanistan. The plant capacity presented in table 2.4 is an indicative value to illustrate the capacity range for each type of technology used in this analysis. Efficiency (or thermal efficiency) is a dimensionless factor showing the amount of input energy required by an electrical generator or power plant (usually thermal) in order to produce one unit of useful output (in this case 1 kilowatt-hour of electricity). The capacity factor is the ratio of the net elec- tricity generated, for the time considered, to the energy that could have been generated at continuous full-power operation during the same period. The technical life of a power plant refers to the estimated years of operation. These factors affect the performance of a power generator directly or indirectly, which in OnSSET’s methodology is expressed through the calculation of the levelized cost of electricity, explained below. 12 A GIS Approach to Planning Electrification in Afghanistan OnSSET uses the levelized cost of electricity calculated for each of the geospatial units and identifies which technology provides access to electricity at the lowest cost. The levelized cost from a specific source represents the final cost of the electricity required for the overall system to break even over the project lifetime. It is calculated through the following formula: (2.1) Where It is the investment expenditure for a specific system in year t; O&Mt is the opera- tion and maintenance cost; Ft is the fuel expenditure; Et is the generated electricity; r is the discount rate; and n is the lifetime of the system. Geospatial units close to an existing electricity grid might find that grid expansion is the least-cost option, since distance to the grid is a very important variable determining the cost of connection. This is especially the case where household consumption levels are potentially high. But this logic does not always hold; for example, in countries where electricity prices are very high, a grid connection may not be the lowest-cost option. If the assumed electricity demand of households is very low, other, stand-alone options might incur the lowest cost: for example, a rooftop PV panel, in locations where solar radiation is strong and diesel prices are high, or a diesel generator set where solar radiation is not strong and diesel prices are low. OnSSET makes all these calculations and finds the lowest-cost option using data relevant to a particular location for all the scenarios considered in the exercise or defined by the modeler. The results can be graphically represented using interactive maps, and they are also available in tabular format so as to facilitate further analysis. Figure 2.1 schematically represents OnSSET’s main methodological processes. 2.5. Scenario formation OnSSET provides the possibility of generating various scenarios, thus investigating alternative pathways for electrification. Thirty-two scenarios have been developed for Afghanistan that consider variations of diesel price, grid cost and electricity consumption level. The following paragraphs present in brief the main input parameters considered in the construction of these scenarios. Geospatial Energy Planning 13 Figure 2.1. Principal components and structure of OnSSET WorldPop, UN, World Bank, World Bank IEA Population Electricity access tiers Electricity demand OSM, DABS, NASA, SRTM, JRC, Natural OSM, DABS, ADB GADM NASA Merra NASA Langley ADB, USGS HydroSHEDS, JRC Earth, IEA Existing and Power plants Small/mini- Wind Diesel cost of planned and Administrative Solar hydro capacity generating transmission economic areas irradiance potential factor electricity network activities Grid Mini-grid Stand-alone Technology selection Optimal electrification split Source: KTH dESA. Note: ADB = Asian Development Bank; DABS = Da Afghanistan Breshna Sherkat; GADM = Global Administrative Areas; HydroSHEDS = Hydrological data and maps based on SHuttle Elevation Derivatives at multiple Scales; IEA = International Energy Agency; JRC = Joint Research Centre; OSM = Open Street Maps; Merra = Modern-era Retrospective Analysis; NASA SRTM = National Aeronautics and Space Administration Shuttle Radar Topography Mission; UN = United Nations; USGS = United States Geological Survey. 2.5.1. First parameter—Energy access target As described in the previous paragraphs, the level of electricity to be provided can vary significantly from area to area. In this study, we consider two population groups: urban and rural. Since, in Afghanistan, these populations have different electricity needs, the level of effort needed for full electrification also varies. To illustrate this, two sets of scenarios were constructed. The first assumes different access targets for urban and rural areas (several combinations of tiers), while the second assumes the same access target for all settlements by 2030 (either tier 3 or tier 4) (see table 2.2 for tier definitions). 2.5.2. Second parameter—Diesel price Diesel generators are an established and reliable technology used for electrification especially in remote areas and also as backup alternatives throughout Afghanistan. Despite the low capital investment required up front, diesel generators can induce high operating costs, depending on fuel price. To account for price fluctuation, two additional scenarios were constructed in OnSSET, one with a low and another one with a high fuel price. The first assumes that the price of diesel will remain close to the current low levels, 14 A GIS Approach to Planning Electrification in Afghanistan Table 2.5. Characteristics of the national power grid in Afghanistan, 2015 Base year 2015 Investment Power System in Afghanistan MW GWh (%) USD/kWh (USD/kW) Imports 3500 76.2 0.049 - Turkmenistan up to 90 595.0 17.0 0.028 - Uzbekistan up to 300 1,995.0 57.0 0.060 - Tajikistan up to 300 140.0 4.0 0.035 - Iran up to 150 770.0 22.0 0.040 - Indigenous generation 1,093 23.8 0.104 Hydro 254.0 711.9 65.1 0.091 2,200 Oil 199.6 131.1 12.0 0.300 500 Gas 67.0 249.4 22.8 0.037 1,000 Coal 0.0 0.0 0.0 0.055 1,500 Solar 0.1 0.3 0.0 0.000 1,953 Wind 0.0 0.0 0.0 0.065 1,750 520.7 4,592.8 100.0 0.062 1,722.01 Source: ADB (2013), Infrastructure Development Cluster (2012), ANPDF (2016), Shift Project Data Portal (2015). Note: GWh = gigawatt-hours; kWh = kilowatt-hours; MW = megawatts. at about $0.69/liter (45.9 afghanis (Af)/liter7), while the second assumes an increase to $1/ liter (66.5 Af/liter) based on the estimated crude oil price level over the next years (IEA 2015). 2.5.3. Third parameter—Electricity price of national grid Previous electrification efforts have shown that the expansion of an electricity grid (new capacity—transmission and distribution) is a capital-intensive process. But because of economies of scale in power generation, grid extension can provide low electricity prices to the end-user. The price at which the electricity is produced is a critical consideration. OnSSET, meanwhile, does not distinguish between the different technologies in the grid’s generation mix; it rather sees the grid as a “black box.” Other grid-optimization tools can be developed and interlinked with OnSSET to distinguish elements of the generation mix (for example, OSeMOSYS); however, they are not within the scope of this study. In this case, the national grid’s cost of generated electricity has been estimated based on a review of relevant literature and the development plans elaborated by the Afghan government. 7. Based on the exchange rate $1 = Af 66.5. (at the base year—2015). Geospatial Energy Planning 15 In 2015 the installed capacity in operation in Afghanistan was 520.7 megawatts (MW), generating approximately 1,093 gigawatt-hours (GWh). Domestic electricity production relied mainly on hydropower, oil, and natural gas and accounted for 23.8 percent of the total consumption, while the rest of the demand was covered by imports. The country imports approximately 3,500 GWh from Uzbekistan (57 percent), Iran (22 percent), Turkmenistan (17 percent), and Tajikistan (4 percent) (ADB 2013; Infrastructure Develop- ment Cluster 2012; The Shift Project Data Portal 2015). Based on this specific mix, the cost of generation was estimated at approximately $0.062/kWh.8 Table 2.5 presents the basic parameters describing Afghanistan’s power sector as of 2015. Restructuring the power sector in Afghanistan would require significant investments in additional capacity and expansion of the transmission and distribution network. Accord- ing to the National Energy Supply Plan (Infrastructure Development Cluster 2012) and the Power Sector Master Plan study conducted by Fichtner and others, approximately 3,000 MW are planned to be added to the national system by 2025 (Infrastructure Devel- opment Cluster 2012; ADB 2013). This entails 2,500 MW from 13 hydropower projects, 400 MW from coal power plants in the Aynak and Hajigak mine sites, and 200 MW from Sheberghan. The end goal is a robust and flexible power system able to effectively utilize the country’s abundant natural resources while also enhancing interconnectivity with neighboring countries (TUTAP–TAPI–CASA 1000).9 Solar and wind are additional sources with high estimated potential (wind power is estimated at 158 GW) (ADB 2013). To incorporate these plans in OnSSET, we mapped out three paths to achieving the govern- ment’s goal by 2030. The first path assumes that electricity imports will remain stable while the additional capacity will come primarily from hydropower, coal, and natural-gas-fired power plants (table 2.6). Reduction in the use of diesel generators is also included. This would force the generating cost to increase to $0.077/kWh while the capital investment requirement (that is, the value used in OnSSET to assess the investment required) is estimated at $1,970/kW. The second path was developed to assess how the increased penetration of renewable- energy projects could affect the generating cost and therefore the output of the model. It was assumed that imports will remain the same, while 40 MW of solar10 and 26 MW of wind11 will replace the oil-based generators. This would result in a lower generation cost of approximately $0.075/kWh but higher capital investment requirements at $1,989/kW. Table 2.7 presents the basic parameters of this path. 8. Based on the author’s estimations. 9. TUTAP: Turkmenistan-Uzbekistan-Tajikistan-Afghanistan-Pakistan electricity project, TAPI: Turkmenistan- Afghanistan-Pakistan India gas pipeline, CASA: Central Asia South Asia Electricity Transmission and Trade Project 10. AEIC: http://aeic.af/en/gismap/60. 11. Projects: Herat Wind Park (14 MW), Herat Solar + Wind (2 MW), Mazar Wind Project (10 MW) in Balkh. 16 A GIS Approach to Planning Electrification in Afghanistan Table 2.6. Forecast of national power grid based on current government plans, 2030 End year 2030—Path to government goal as presently planned Investment Power System in Afghanistan MW GWh (%) USD/kWh (USD/kW) Imports 3,500.0 24.3 0.066 - Turkmenistan up to 90 595.0 17.0 0.038 - Uzbekistan up to 300 1,995.0 57.0 0.081 - Tajikistan up to 300 140.0 4.0 0.047 - Iran up to 150 770.0 22.0 0.054 - Indigenous generation 10,897.8 75.7 0.080 Hydro 2,767.5 7,757.8 71.2 0.091 2,230.0 Oil 66.0 43.4 0.4 0.300 500.0 Gas 267.0 994.0 9.1 0.037 1,000.0 Coal 400.0 2,102.4 19.3 0.055 1,500.0 Solar 0.1 0.3 0.0 0.060 1,130.0 Wind 0.0 0.0 0.0 0.065 1,600.0 3,500.6 14,397.8 100.0 0.077 1,970.1 Source: ADB (2013), Infrastructure Development Cluster (2012), ANPDF (2016), Shift Project Data Portal (2015). Note: GWh = gigawatt-hours; kWh = kilowatt-hours; MW = megawatts. Table 2.7. Forecast of national power grid with greater renewable energy penetration, 2030 End year 2030—Alternative path to government goal making greater use of renewables Investment Power System in Afghanistan MW GWh (%) USD/kWh (USD/kW) Imports 3,500.0 24.1 0.066 - Turkmenistan up to 90 595.0 17.0 0.038 - Uzbekistan up to 300 1,995.0 57.0 0.081 - Tajikistan up to 300 140.0 4.0 0.047 - Iran up to 150 770.0 22.0 0.054 - Indigenous generation 11,027.7 75.9 0.078 Hydro 2,767.5 7,757.8 71.2 0.091 2,230.0 Oil 0.0 0.0 0.0 0.300 500.0 Gas 267.0 994.0 9.1 0.037 1,000.0 Coal 400.0 2,102.4 19.3 0.055 1,500.0 Solar 40.0 105.1 1.0 0.060 1,130.0 Wind 26.0 68.3 0.6 0.065 1,600.0 3,500.5 14,527.7 100.0 0.075 1,989.0 Source: ADB (2013), Infrastructure Development Cluster (2012), ANPDF (2016), Shift Project Data Portal (2015). Note: GWh = gigawatt-hours; kWh = kilowatt-hours; MW = megawatts. Geospatial Energy Planning 17 Table 2.8. Forecast of national power grid with increased imports, 2030 End year 2030—Alternative path to government goal making greater use of imports Investment Power System in Afghanistan MW GWh (%) USD/kWh (USD/kW) Imports 5,680.0 39.1 0.063 - Turkmenistan up to 90 738.4 13.0 0.038 - Uzbekistan up to 300 2,556.0 45.0 0.081 - Tajikistan up to 300 1,136.0 20.0 0.047 - Iran up to 150 1,249.6 22.0 0.054 - Indigenous generation 8,847.8 60.9 0.083 Hydro 2,254.0 6,318.3 58.0 0.091 2,230.0 Oil 199.6 131.1 1.2 0.300 500.0 Gas 150.0 558.5 5.1 0.037 1,000.0 Coal 350.0 1,839.6 16.9 0.055 1,500.0 Solar 0.1 0.3 0.0 0.060 1,130.0 Wind 0.0 0.0 0.0 0.065 1,600.0 2,953.7 14,527.8 100.0 0.075 1,603.4 Source: ADB (2013), Infrastructure Development Cluster (2012), ANPDF (2016), Shift Project Data Portal (2015). Note: GWh = gigawatt-hours; kWh = kilowatt-hours; MW = megawatts. Finally, the third path was developed to illustrate how the grid electricity cost would react in case increased imports are needed to cover the expected demand. To illustrate that, the planned domestic generation capacity was kept below 3,000 MW.12 Imports were increased. The grid cost was estimated at $0.075/kWh, with lower capital invest- ment requirements than in the previous cases, at $1,603/kW. Table 2.8 presents the basic parameters of this path. 2.6. Running the model 2.6.1. Option 1: Using the online interface of OnSSET Over the past few months KTH dESA has been developing an online open interface to support the use of OnSSET by professionals without experience in the use of geospatial software. The interface allows the user to conduct an electrification analysis of a selected country, based on few key input parameters (energy access targets, population character- istics, technology costs, and so on). The results can be visualized quickly in an embedded map that show “at a glance” the most cost-effective electrification pathways. This inter- face is accessible at OnSSET.org, and the only requirement for its use is a stable Internet connection.13 12. As of spring 2017, 513 MW of hydropower projects had not come through; gas (Sheberghan) and coal projects (in Aynak and Hajigak) were delayed. 13. For access credentials refer to appendix B. 18 A GIS Approach to Planning Electrification in Afghanistan A space in OnSSET.org has been specially created to accommodate model runs that investigate various electrification pathways for Afghanistan. More information on how to navigate and properly conduct an electrification analysis using the online interface is accessible at https://energydata.info. 2.6.2. Option 2: Stand-alone software OnSSET has been developed as an open-source tool; that is, the code and all the func- tions behind the model are accessible by any user and can be customized to serve the objectives of any analysis. The requirements for a fully customizable version of OnSSET are as follows. GIS environment OnSSET is a spatial electrification tool and as such relies on the use of GIS. A GIS envi- ronment is therefore necessary to: 1. Extract trivial characteristics for the electrification analysis from GIS layers and combine them together in a format easy to read by the Python code (a comma- separated-value file with all the attributes per population point). 2. Visualize the final results in maps. At present, OnSSET relies on ArcGIS; however, any alternative GIS environment can be used (for example, Qgis and/or Grass). Python—Anaconda package OnSSET is written in Python, an open-source programming language used widely in many applications. Python14 is a necessary requirement for OnSSET to work. Programming in Python usually relies on the use of predefined functions that can be found in so-called modules. To work with OnSSET, certain modules need to be installed/ updated. The easiest way to do so is by installing Anaconda, a package that contains various useful modules. Anaconda can be downloaded for free from https://www. continuum.io/anaconda-overview. Please make sure that you download the version that is compatible with your operating system (for example, Windows 32-bit). After the installation, you can use the Anaconda command line to run Python. Anaconda includes all the modules required to run OnSSET. Python interfaces Integrated Development Environment programs are used in order to ease programming process when multiple or long scripts are required. Many such programs have been 14. Python itself can be downloaded and installed for free using the official website: https://www.python.org/ downloads/. Geospatial Energy Planning 19 developed for Python.15 KTH dESA has been using PyCharm as the standard one to run OnSSET.16 Jupyter notebook is a console-based, interactive computing approach to providing a web- based application suitable for capturing the whole computation process: developing, documenting, and executing code, as well as communicating the results. Jupyter note- book is used for the online onset interface, and is recommended for small analyses and exploring code and results. GitHub GitHub is a Web-based Git repository hosting service. It provides access control and several collaboration features such as bug tracking, feature requests, task management, and wikis for every project. The open-source code behind OnSSET is called “PyOnSSET” and is available in KTH dESA’s Github space (https://github.com/KTH-dESA/PyOnSSET). A GitHub account will allow you to propose changes, modifications, and upgrades to the existing code. 2.7. Results and visualization In total, 32 scenarios have been created for Afghanistan. A few of them (highlighted with green in table 2.9) are presented here as a showcase of OnSSET’s capabilities. OnSSET yields two comma-separated-value files as an output for each scenario. The first file contains the information acquired from the electrification analysis for every single settlement in the country according to the specified resolution (in this case 1x1 km). This file is used to retrieve location-specific information but also to illustrate the results on detailed maps through a GIS environment. The second file contains the summarized results for the scenario, providing information about the total capacity needed, by tech- nology, and the relative investment level required to achieve the electrification target.17 Figures 2.2–2.8 provide a quick overview of the results for the 12 selected scenarios. The graphs allow a quantitative approach to the comparison of the aggregated results show- ing technology share, added capacity, and investment requirements per scenario. The maps, meanwhile, are organized in such a way as to allow a more qualitative comparison between the parameters that most influence the penetration of different technologies in the generation mix. A more detailed description of the results for each scenario is avail- able in appendix A. 15. You can find a few at: http://noeticforce.com/best-python-ide-for-programmers-windows-and-mac. 16. It can be downloaded from https://www.jetbrains.com/pycharm/. 17. The files for all the scenarios developed for Afghanistan are available through the following link: https://energydata.info. 20 A GIS Approach to Planning Electrification in Afghanistan Table 2.9. Grid cost ($/kWh) under 32 electrification scenarios investigated for Afghanistan, by diesel price and electricity consumption tier Urban and rural electricity consumption tier and average weighted household consumption Diesel price U4–R4 U5–R3 U4–R2 U3–R3 ($/L) (3,247 kWh/year) (2,433 kWh/year) (1,378 kWh/year) (1,232 kWh/year) 0.69 0.062 0.062 0.062 0.062 0.69 0.077 0.077 0.077 0.077 Low 0.69 0.075 RE 0.075 RE 0.075 RE 0.075 RE 0.69 0.075 IM 0.075 IM 0.075 IM 0.075 IM 1 0.062 0.062 0.062 0.062 1 0.077 0.077 0.077 0.077 High 1 0.075 RE 0.075 RE 0.075 RE 0.075 RE 1 0.075 IM 0.075 IM 0.075 IM 0.075 IM Source: KTH dESA. Note: RE refers to increased penetration of renewable-based technologies (see table 2.7); IM refers to increased imports from neighboring countries (table 2.8). 12 selected scenarios (highlighted in blue) are presented in the report. Figure 2.2. Total population electrified, by technology Grid connected Stand-alone diesel Stand-alone PV Mini-grid wind Mini-grid diesel Mini-grid PV Mini-grid hydro 50 40 Population (million) 30 20 10 0 Scenario 1 Scenario 2 Scenario 3 Scenario 4 Scenario 5 Scenario 6 Scenario 7 Scenario 8 Scenario 9 Scenario 10 Scenario 11 Scenario 12 Consumption tier U4–R4 U4–R4 U5– R3 U5–R3 U5–R3 U5–R3 U3–R3 U3–R3 U3–R3 U4–R2 U4–R2 U4–R2 Diesel price ($/L) LD HD LD HD HD HD LD HD HD LD HD HD Grid cost ($/kWh) 0.75RE 0.077 0.077 0.077 0.75RE 0.75IM 0.75RE 0.077 0.75IM 0.75IM 0.077 0.75IM U = urban; R = rural Source: KTH dESA. Note: Ux–Rx refers to the electrification tier for urban and rural settlements, respectively (see table 2.2); RE refers to the alternate path with increased penetration of renewable-based technologies (solar, wind) (see table 2.7); IM refers to the alternative path with increased imports from neighboring countries (see table 2.8). kWh = kilowatt-hours; LD = low diesel price; HD = high diesel price. Geospatial Energy Planning 21 Figure 2.3. Newly electrified population, by technology Grid connected Stand-alone diesel Stand-alone PV Mini-grid wind Mini-grid diesel Mini-grid PV Mini-grid hydro 35 30 Population (millions) 25 20 15 10 5 0 Scenario 1 Scenario 2 Scenario 3 Scenario 4 Scenario 5 Scenario 6 Scenario 7 Scenario 8 Scenario 9 Scenario 10 Scenario 11 Scenario 12 Consumption tier U4–R4 U4–R4 U5– R3 U5–R3 U5–R3 U5–R3 U3–R3 U3–R3 U3–R3 U4–R2 U4–R2 U4–R2 Diesel price ($/L) LD HD LD HD HD HD LD HD HD LD HD HD Grid cost ($/kWh) 0.75RE 0.077 0.077 0.077 0.75RE 0.75IM 0.75RE 0.077 0.75IM 0.75IM 0.077 0.75IM U = urban; R = rural Source: KTH dESA. Note: Ux–Rx refers to the electrification tier for urban and rural settlements, respectively (see table 2.2); RE refers to the alternative path with increased penetration of renewable-based technologies (solar, wind) (see table 2.7); IM refers to the alternative path with increased imports from neighboring countries (see table 2.8). kWh = kilowatt-hours; LD = low diesel price; HD = high diesel price. Figure 2.4. New capacity to be added, by technology Grid connected Stand-alone diesel Stand-alone PV Mini-grid wind Mini-grid diesel Mini-grid PV Mini-grid hydro 6 5 4 GW 3 2 1 0 Scenario 1 Scenario 2 Scenario 3 Scenario 4 Scenario 5 Scenario 6 Scenario 7 Scenario 8 Scenario 9 Scenario 10 Scenario 11 Scenario 12 Consumption tier U4–R4 U4–R4 U5– R3 U5–R3 U5–R3 U5–R3 U3–R3 U3–R3 U3–R3 U4–R2 U4–R2 U4–R2 Diesel price ($/L) LD HD LD HD HD HD LD HD HD LD HD HD Grid cost ($/kWh) 0.75RE 0.077 0.077 0.077 0.75RE 0.75IM 0.75RE 0.077 0.75IM 0.75IM 0.077 0.75IM U = urban; R = rural Source: KTH dESA. Note: Ux–Rx refers to the electrification tier for urban and rural settlements, respectively (see table 2.2); RE refers to the alternative path with increased penetration of renewable-based technologies (solar, wind) (see table 2.7); IM refers to the alternative path with increased imports from neighboring countries (see table 2.8). kWh = kilowatt-hours; LD = low diesel price; HD = high diesel price. 22 A GIS Approach to Planning Electrification in Afghanistan Figure 2.5. Investment requirements, by technology Grid connected Stand-alone diesel Stand-alone PV Mini-grid wind Mini-grid diesel Mini-grid PV Mini-grid hydro 30 25 20 Billion US$ 15 10 5 0 Scenario 1 Scenario 2 Scenario 3 Scenario 4 Scenario 5 Scenario 6 Scenario 7 Scenario 8 Scenario 9 Scenario 10 Scenario 11 Scenario 12 Consumption tier U4–R4 U4–R4 U5– R3 U5–R3 U5–R3 U5–R3 U3–R3 U3–R3 U3–R3 U4–R2 U4–R2 U4–R2 Diesel price ($/L) LD HD LD HD HD HD LD HD HD LD HD HD Grid cost ($/kWh) 0.75RE 0.077 0.077 0.077 0.75RE 0.75IM 0.75RE 0.077 0.75IM 0.75IM 0.077 0.75IM U = urban; R = rural Source: KTH dESA. Note: Ux–Rx refers to the electrification tier for urban and rural settlements, respectively (see table 2.2); RE refers to the alternative path with increased penetration of renewable-based technologies (solar, wind) (see table 2.7); IM refers to the alternative path with increased imports from neighboring countries (see table 2.8). kWh = kilowatt-hours; LD = low diesel price; HD = high diesel price. Figure 2.6. New capacity required, by system type Grid Mini-grid Stand-alone 6 5 4 GW 3 2 1 0 Scenario 1 Scenario 2 Scenario 3 Scenario 4 Scenario 5 Scenario 6 Scenario 7 Scenario 8 Scenario 9 Scenario 10 Scenario 11 Scenario 12 Consumption tier U4–R4 U4–R4 U5– R3 U5–R3 U5–R3 U5–R3 U3–R3 U3–R3 U3–R3 U4–R2 U4–R2 U4–R2 Diesel price ($/L) LD HD LD HD HD HD LD HD HD LD HD HD Grid cost ($/kWh) 0.75RE 0.077 0.077 0.077 0.75RE 0.75IM 0.75RE 0.077 0.75IM 0.75IM 0.077 0.75IM U = urban; R = rural Source: KTH dESA. Note: Ux–Rx refers to the electrification tier for urban and rural settlements, respectively (see table 2.2); RE refers to the alternative path with increased penetration of renewable-based technologies (solar, wind) (see table 2.7); IM refers to the alternative path with increased imports from neighboring countries (see table 2.8). kWh = kilowatt-hours; LD = low diesel price; HD = high diesel price. Geospatial Energy Planning 23 Figure 2.7. Investment required, by system type Grid Mini-grid Stand-alone 30 25 20 Billion US$ 15 10 5 0 Scenario 1 Scenario 2 Scenario 3 Scenario 4 Scenario 5 Scenario 6 Scenario 7 Scenario 8 Scenario 9 Scenario 10 Scenario 11 Scenario 12 Consumption tier U4–R4 U4–R4 U5– R3 U5–R3 U5–R3 U5–R3 U3–R3 U3–R3 U3–R3 U4–R2 U4–R2 U4–R2 Diesel price ($/L) LD HD LD HD HD HD LD HD HD LD HD HD Grid cost ($/kWh) 0.75RE 0.077 0.077 0.077 0.75RE 0.75IM 0.75RE 0.077 0.75IM 0.75IM 0.077 0.75IM U = urban; R = rural Source: KTH dESA. Note: Ux–Rx refers to the electrification tier for urban and rural settlements, respectively (see table 2.2); RE refers to the alternative path with increased penetration of renewable-based technologies (solar, wind) (see table 2.7); IM refers to the alternative path with increased imports from neighboring countries (see table 2.8). kWh = kilowatt-hours; LD = low diesel price; HD = high diesel price. 24 A GIS Approach to Planning Electrification in Afghanistan Figure 2.8. Summarized results Grid Diesel mini-grid Hydro mini-grid PV mini-grid Wind mini-grid Diesel stand-alone PV stand-alone a. Scenario 1: U4–R4, LD, 0.075 RE b. Scenario 2: U4–R4, HD, 0.077 N c. Scenario 3: U5–R3, LD, 0.077 d. Scenario 4: U5–R3, HD, 0.077 e. Scenario 5: U5–R3, HD, 0.075 RE f. Scenario 6: U5–R3, HD, 0.075 IM 0 70 140 280 420 560 Kilometers Geospatial Energy Planning 25 g. Scenario 7: U3–R3, LD, 0.075 RE h. Scenario 8: U3–R3, HD, 0.077 N i. Scenario 9: U3–R3, HD, 0.075 IM j. Scenario 10: U4–R2, LD, 0.075 IM k. Scenario 11: U4–R2, HD, 0.077 l. Scenario 12: U4–R2, HD, 0.075 IM 0 70 140 280 420 560 Kilometers Source: KTH dESA. Note: Ux–Rx refers to the electrification tier for urban and rural settlements, respectively (see table 2.2); RE refers to the alternative path with increased penetration of renewable-based technologies (solar, wind) (see table 2.7); IM refers to the alternative path with increased imports from neighboring countries (see table 2.8). kWh = kilowatt-hours; LD = low diesel price; HD = high diesel price. 26 A GIS Approach to Planning Electrification in Afghanistan Bridging science and policy: Interpretation 2.8.  of OnSSET’s results OnSSET’s results clearly show that the assumed level of electricity demand per house- hold in the settlements in each of the GIS cells is an important factor determining which technology offers the lowest cost. At the lowest consumption levels, most population settlements close to already-electrified villages and transmission lines will find that connecting to the central electricity grid is the lowest-cost option. Elsewhere, most settle- ments will find that stand-alone systems are the most economical option (with PV panels a better option than diesel gensets, especially when diesel prices are high). At this low level of consumption, mini-grids play only a minor role. Assuming a midrange electricity consumption per household makes grid connection a more viable option (to the detriment of stand-alone options). Assuming high con- sumption levels furthers the viability of increasing connections to the central grid; but, interestingly, high consumption levels imply that mini-grids become an economically attractive option and replace stand-alone technologies in many settlements. Shift- ing between low and high diesel prices not only has the expected result of reducing the importance of diesel generators for mini-grids or stand-alone systems, but it also increases the overall contribution of renewable-based mini-grids to achieving universal access to electricity. The model finds that between 55 and 73 percent of the population may be receiving electricity through off-grid technologies. Where the assumed electricity consumption is relatively high (above 1,500 kWh/hh/year), off-grid technologies might be responsible for between 40 and 50 percent of households gaining access to electricity. What is the cost of ensuring access to electricity for all by 2030? The total investment required ranges between $7.82 billion and $26.04 billion over the entire period in ques- tion, 2015–30. These figures include the up-front capital investments for extending the transmission and distribution network, building the mini-grid systems, and installing the stand-alone solar and diesel technologies. The lower a household’s electricity con- sumption and the lower the diesel price, the lower the overall investment needed to reach universal access. The lowest-cost investment ticket for providing universal access corresponds to the case where consumption is assumed to be lowest (tier 4 for urban, tier 2 for rural), and the diesel price is low. The highest ticket corresponds to the highest household consumption (tier 4 for urban, tier 4 for rural) at the high diesel price (see table 2.2 for tier definitions). The model identifies the role of renewable sources of energy in off-grid technologies. The role of renewables critically depends on the price of diesel. When diesel prices are low, renewable sources will be used to provide electricity to an average of 51 percent Geospatial Energy Planning 27 of the population. However, increasing the price of diesel to $1/liter (Af 66.5/liter) also increases the average contribution of renewable sources, which rise to 57 percent of the population. In summary: n Grid extension is the least-cost solution in densely populated areas that are already close to the existing grid network. n Low diesel prices allow for a small market capture (1–2 percent)18 of diesel gensets in the electrification mix. Increased diesel prices, on the other hand, give a com- petitive advantage to solar PV systems. n Other renewables (wind, hydropower) capture a significant percentage in areas where these resources are available; their share is moderate across Afghanistan. n As electricity demand levels rise, the most favorable electrification option shifts from (i) stand-alone systems to (ii) mini-grids to (iii) a grid connection. n The share of renewable energy sources in total electricity generation can reach more than 60 percent by 2030.19 18. An exception is the U4–R2, LD, 0.075IM where stand-alone diesel generators are found to be the most economic electrification option for 13 percent of the newly electrified population. 19. Depending on the share of grid-connected renewables. 3. Conclusion and Final Remarks OnSSET provides useful insights that may be used to assess electrification options across population settlements in well-defined locations; aggregate results reveal patterns at the national or subnational levels. The tool’s prime focus is the identification of additional capacity and investment required to fulfill energy access goals. OnSSET offers policy and decision makers support in identifying least-cost electrifica- tion strategies. Most important, the model specifically addresses the needs of the energy poor and offers tangible solutions. Its open-source features allow energy experts to refine results resolution and explore additional scenarios. When linked to OSeMOSYS, it becomes an extremely powerful analysis and planning tool. Like most open source models, OnSSET is a work in progress, especially as new satellite imagery and GIS data become available. The current analysis has several limitations, which may be overcome as follows: n The electrification mix is shown only for the end year (here 2030). Thus, the electrification mix and status in the intervening years (that is, today through 2030) are not considered. To include the whole period, it would be necessary to decide which areas need to be electrified in what order. n The breakdown of the generation mix used to consider different grid electrifica- tion costs is not detailed. It would be necessary to link OnSSET with OSeMOSYS to obtain the optimal generation mix based on the country’s resources, demand, and trade with other countries. n Another critical issue is the various resolutions of the datasets used. For example, population density data are given at 1 km while the wind speed is at 5 km. The datasets need to be harmonized to ensure better accuracy. n The analysis considers only household electrification. Other productive uses of electricity (such as in health services, schools, rural enterprises, agriculture and so on) should also be considered. These would increase the demand levels and therefore the electrification mix. n A final word of caution: the model quantifies electrification targets for Afghani- stan by 2030. It considers the least-cost mix and what the required aggregate investment would be. But it does not imply the implementation of the identified strategies or the provision of necessary finance. It highlights the challenges ahead for policy and decision makers charged with implementing energy strategies to achieve access targets, allows an analysis of trade-offs between competing demands for financial resources, and thus supports the prudent prioritization of available financial resources. 29 References ADB (Asian Development Bank). 2013. Islamic Republic of Afghanistan: Power Sector Master Plan (Financed by the Japan Fund for Poverty Reduction).Technical Assistance Consultant’s Report. Project Number: 43497. Prepared by Fichtner. May 2013. AEIC (Afghan Energy Information Center). “100 MW Renewable Energy Projects.” http://aeic.af/en/ gismap/60. ANPDF (Afghanistan National Peace and Development Framework). 2016. “National Infrastructure Plan 2017–2021.” ANPDF, Afghanistan. Asia Foundation, The. 2015. “The Asia Foundation’s 2015 Survey of the Afghan People.” Asia Founda- tion, Washington, DC. CSO (Central Statistics Organization). 2016. Afghanistan Living Conditions Survey 2013–14: National Risk and Vulnerability Assessment. Kabul: CSO. Ellis, Peter, and Mark Roberts. 2016. “Leveraging Urbanization in South Asia: Managing Spatial Transfor- mation for Prosperity and Livability.” South Asia Development Matters, World Bank, Washington, DC. https://openknowledge.worldbank.org/bitstream/handle/10986/22549/9781464806629.pdf?seque nce=17&isAllowed=y. ESMAP (Energy Sector Management Assistance Program). 2017. “Model for Electricity Technology Assessment (META).” World Bank, Washington DC. http://esmap.org/meta. Fuso Nerini, Francesco, Oliver Broad, Dimitris Mentis, Manuel Welsch, Morgan Bazilian, and Mark Howells. 2016. “A Cost Comparison of Technology Approaches for Improving Access to Elec- tricity Services.” Energy 95 (January): 255–65. http://www.sciencedirect.com/science/article/pii/ S036054421501631X. IEA (International Energy Agency). 2015. World Energy Outlook 2015. Paris: IEA. IEA and World Bank. 2015. Sustainable Energy for All 2015—Progress Toward Sustainable Energy 2015: Global Tracking Framework Report. Washington, DC: World Bank. http://www.se4all.org/sites/default/files/ GTF-2105-Full-Report.pdf. Infrastructure Development Cluster. 2012. National Energy Supply Programme (NESP). Ministry of Finance. The Islamic Republic of Afghanistan. IRENA (International Renewable Energy Agency). 2015. Renewable Power Generation Costs in 2014. Abu Dhabi: IRENA. https://www.irena.org/DocumentDownloads/Publications/IRENA_RE_Power_ Costs_2014_repor t.pdf. 31 32 References KTH dESA (KTH Royal Institute of Technology-Division of Energy Systems Analysis): https://www.kth. se/en/itm/om/organisation/institutioner/energiteknik/forskningsavdelningar/desa. Mentis, Dimitrios, Manuel Welsch, Francesco Fuso Nerini, Oliver Broad, Mark Howells, Morgan Bazilian, and Holger Rogner. 2015. “A GIS-Based Approach for Electrification Planning—A Case Study on Nigeria.” Energy for Sustainable Development 29 (December): 142–50. http://www.sciencedirect.com/ science/article/pii/S0973082615000952. Shift Project Data Portal, The. 2015. “Electricity Generation Statistics.” http://www.tsp-data-portal.org/. UN DESA (United Nations, Department of Economic and Social Affairs), Population Division. 2014. https://esa.un.org/unpd/wup/CD-ROM/. ———. 2015. https://esa.un.org/unpd/wpp/Download/Probabilistic/Population/. UN SDG (United Nations Sustainable Development Goals). “Goal 7: Ensure Access to Affordable, Reli- able, Sustainable and Modern Energy for All.” http://www.un.org/sustainabledevelopment/energy. World Bank. 2017a. Gross National Income (GNI) per capita, Atlas method (current US$). Afghanistan. https://data.worldbank.org/indicator/NY.GNP.PCAP.CD. ———. 2017b. GDP growth (annual %). Afghanistan. https://data.worldbank.org/indicator/NY.GDP. MKTP.KD.ZG?locations=AF. ———. 2017c. Electric power consumption (kWh per capita). https://data.worldbank.org/indicator/ EG.USE.ELEC.KH.PC?view=chart. ———. 2017d. State of Electricity Access Report 2017. Full report. Washington, DC: World Bank Group. http://documents.worldbank.org/curated/en/364571494517675149/full-reportated/ en/364571494517675149/full-report. ———. 2017e. Population, total. Afghanistan. https://data.worldbank.org/indicator/SP.POP. TOTL?locations=AF. ———. Forthcoming. Islamic Republic of Afghanistan Energy Security Trade-Offs under High Uncer- tainty: Resolving Afghanistan’s Power Sector Development Dilemma. Yemak, Reza; Fun Gan, Delice Cheng. 2013. Celebrating Ten Years of the National Solidarity Program (NSP): A Glimpse of the Rural Development Story in Afghanistan. Washington DC: World Bank. http:// documents.worldbank.org/curated/en/155731467993744166/Celebrating-ten-years-of-the-National- Solidarity-Program-NSP-a-glimpse-of-the-rural-development-story-in-Afghanistan. Appendix A Detailed results of 12 representative electrification scenarios for Afghanistan (Map and tabular format) Figure A.1. Scenario 1: U4–R4, LD, 0.075RE—Electrification results under the scenario defined by urban demand at tier 4, rural demand at tier 4, low diesel price, grid cost at 0.075 $/kWh, and higher penetration of renewable-based technologies in the grid mix Grid Diesel mini-grid Hydro mini-grid PV mini-grid Wind mini-grid Diesel stand-alone PV stand-alone N 0 70 140 280 420 560 Kilometers Source: KTH dESA. Table A.1. Scenario 1: Electrification results under the scenario defined by urban demand at tier 4, rural demand at tier 4, low diesel price, grid cost at 0.075 $/kWh, and higher penetration of renewable-based technologies in the grid mix People to receive electricity by 2030: 31,999,487 By technology type Share (%) Capacity (MW) Investment ($ billion) Grid extension 60.7 1,773 14.3 Mini-grids 38.3 2,921 11.07 Diesel genset 1.5 65.9 0.198 PV system 35.4 2,798 10.6 Wind turbines 0.3 23.8 0.089 Mini–small hydro 1.1 33.3 0.217 Stand-alone 1.0 73 0.4 Diesel genset 0.053 2.4 0.006 PV systems 0.9 70.6 0.388 Total 100 4,767.4 25.76 Source: KTH dESA. 33 Note: kWh = kilowatt-hours; MW = megawatts; PV = photovoltaic. 34 Appendix A Figure A.2. Scenario 2: U4–R4, HD, 0.077—Electrification results under the scenario defined by urban demand at tier 4, rural demand at tier 4, high diesel price, and grid cost at 0.077 $/kWh Grid Diesel mini-grid Hydro mini-grid PV mini-grid Wind mini-grid Diesel stand-alone PV stand-alone N 0 70 140 280 420 560 Kilometers Source: KTH dESA. Table A.2. Scenario 2: Electrification results under the scenario defined by urban demand at tier 4, rural demand at tier 4, high diesel price, and grid cost at 0.077 $/kW People to receive electricity by 2030: 31,999,487 By technology type Share (%) Capacity (MW) Investment ($ billion) Grid extension 60.6 1,769 14.22 Mini-grids 37.6 2,899 11.01 Diesel genset 0.5 20.03 0.052 PV system 35.7 2,815 10.6 Wind turbines 0.3 30.2 0.11 Mini–small hydro 1.1 34.2 0.22 Stand-alone 1.9 148 0.82 Diesel genset 0.003 0.13 0.038 PV systems 1.9 148.2 0.082 Total 100 4,816.4 26.04 Source: KTH dESA. Note: kWh = kilowatt-hours; MW = megawatts; PV = photovoltaic. Appendix A 35 Figure A.3. Scenario 3: U5–R3, LD, 0.077—Electrification results under the scenario defined by urban demand at tier 5, rural demand at tier 3, low diesel price, and grid cost at 0.077 $/kWh Grid Diesel mini-grid Hydro mini-grid PV mini-grid Wind mini-grid Diesel stand-alone PV stand-alone N 0 70 140 280 420 560 Kilometers Source: KTH dESA. Table A.3. Scenario 3: Electrification results under the scenario defined by urban demand at tier 5, rural demand at tier 3, low diesel price, and grid cost at 0.077 $/kWh People to receive electricity by 2030: 31,999,487 By technology type Share (%) Capacity (MW) Investment ($ billion) Grid extension 44.9 1,225 9.84 Mini-grids 41.4 1,252 5.7 Diesel genset 0.8 12.8 0.052 PV system 40.0 1,228 5.6 Wind turbines 0.1 4.8 0.021 Mini–small hydro 0.5 6.6 0.055 Stand-alone 13.7 383 2.1 Diesel genset 1.2 20.6 0.06 PV systems 12.5 362.3 2.0 Total 100 2,860 17.58 Source: KTH dESA. Note: kWh = kilowatt-hours; MW = megawatts; PV = photovoltaic. 36 Appendix A Figure A.4. Scenario 4: U5–R3, HD, 0.077—Electrification results under the scenario defined by urban demand at tier 5, rural demand at tier 3, high diesel price, and grid cost at 0.077 $/kWh Grid Diesel mini-grid Hydro mini-grid PV mini-grid Wind mini-grid Diesel stand-alone PV stand-alone N 0 70 140 280 420 560 Kilometers Source: KTH dESA. Table A.4. Scenario 4: Electrification results under the scenario defined by urban demand at tier 5, rural demand at tier 3, high diesel price, and grid cost at 0.077 $/kW People to receive electricity by 2030: 31,999,487 By technology type Share (%) Capacity (MW) Investment ($ billion) Grid extension 45.1 1,228 9.9 Mini-grids 41.0 1,249 5.7 Diesel genset 0.1 2.2 0.01 PV system 40.2 1,233 5.6 Wind turbines 0.2 6.6 0.027 Mini–small hydro 0.5 6.6 0.056 Stand-alone 13.9 407 2.3 Diesel genset 0.03 0.486 0.014 PV systems 13.9 406 2.2 Total 100 2,883.4 17.82 Source: KTH dESA. Note: kWh = kilowatt-hours; MW = megawatts; PV = photovoltaic. Appendix A 37 Figure A.5. Scenario 5: U5–R3, HD, 0.075RE—Electrification results under the scenario defined by urban demand at tier 5, rural demand at tier 3, high diesel price, grid cost at 0.075 $kWh, and higher penetration of renewable-based technologies in the grid mix Grid Diesel mini-grid Hydro mini-grid PV mini-grid Wind mini-grid Diesel stand-alone PV stand-alone N 0 70 140 280 420 560 Kilometers Source: KTH dESA. Table A.5. Scenario 5: Electrification results under the scenario defined by urban demand at tier 5, rural demand at tier 3, high diesel price, grid cost at 0.075 $/kWh, and higher penetration of renewable-based technologies in the grid mix People to receive electricity by 2030: 31,999,487 By technology type Share (%) Capacity (MW) Investment ($ billion) Grid extension 45.5 1,233 10.02 Mini-grids 40.6 1,236 5.64 Diesel genset 0.1 2.2 0.01 PV system 39.8 1,221 5.5 Wind turbines 0.2 6.5 0.026 Mini–small hydro 0.5 6.6 0.056 Stand-alone 13.9 406 2.23 Diesel genset 0.03 0.48 0.014 PV systems 13.9 405.5 2.2 Total 100 2,875.3 17.89 Source: KTH dESA. Note: kWh = kilowatt-hours; MW = megawatts; PV = photovoltaic. 38 Appendix A Figure A.6. Scenario 6: U5–R3, HD, 0.075IM—Electrification results under the scenario defined by urban demand at tier 5, rural demand at tier 3, high diesel price, grid cost at 0.075 $/kWh, and increased imports in the grid mix Grid Diesel mini-grid Hydro mini-grid PV mini-grid Wind mini-grid Diesel stand-alone PV stand-alone N 0 70 140 280 420 560 Kilometers Source: KTH dESA. Table A.6. Scenario 6: Electrification results under the scenario defined by urban demand at tier 5, rural demand at tier 3, high diesel price, grid cost at 0.075 $/kWh, and increased imports in the grid mix People to receive electricity by 2030: 31,999,487 By technology type Share (%) Capacity (MW) Investment ($ billion) Grid extension 45.5 1,233 8.9 Mini-grids 40.6 1,236 5.64 Diesel genset 0.13 2.2 0.01 PV system 39.7 1,221 5.5 Wind turbines 0.2 6.5 0.026 Mini–small hydro 0.5 6.6 0.056 Stand-alone 13.9 406 2.23 Diesel genset 0.03 0.5 0.014 PV systems 13.8 405.5 2.2 Total 100 2,875 16.77 Source: KTH dESA. Note: kWh = kilowatt-hours; MW = megawatts; PV = photovoltaic. Appendix A 39 Figure A.7. Scenario 7: U3–R3, LD, 0.075RE—Electrification results under the scenario defined by urban demand at tier 3, rural demand at tier 3, low diesel price, grid cost at 0.075 $/kWh, and higher penetration of renewable-based technologies in the grid mix Grid Diesel mini-grid Hydro mini-grid PV mini-grid Wind mini-grid Diesel stand-alone PV stand-alone N 0 70 140 280 420 560 Kilometers Source: KTH dESA. Table A.7. Scenario 7: Electrification results under the scenario defined by urban demand at tier 3, rural demand at tier 3, low diesel price, grid cost at 0.075 $/kWh, and higher penetration of renewable-based technologies in the grid mix People to receive electricity by 2030: 31,999,487 By technology type Share (%) Capacity (MW) Investment ($ billion) Grid extension 45.2 501 5.57 Mini-grids 41.0 1,236 5.63 Diesel genset 0.8 12.8 0.051 PV system 39.6 1,212 5.5 Wind turbines 0.1 4.7 0.021 Mini–small hydro 0.5 6.4 0.056 Stand-alone 13.7 383 2.05 Diesel genset 1.2 20.5 0.061 PV systems 12.5 362.2 1.99 Total 100 2,119.2 13.25 Source: KTH dESA. Note: kWh = kilowatt-hours; MW = megawatts; PV = photovoltaic. 40 Appendix A Figure A.8. Scenario 8: U3–R3, HD, 0.077— Electrification results under the scenario defined by urban demand at tier 3, rural demand at tier 3, high diesel price, and grid cost at 0.077 $/kWh Grid Diesel mini-grid Hydro mini-grid PV mini-grid Wind mini-grid Diesel stand-alone PV stand-alone N 0 70 140 280 420 560 Kilometers Source: KTH dESA. Table A.8. Scenario 8: Electrification results under the scenario defined by urban demand at tier 3, rural demand at tier 3, high diesel price, and grid cost at 0.077 $/kWh People to receive electricity by 2030: 31,999,487 By technology type Share (%) Capacity (MW) Investment ($ million) Grid extension 45.1 499 5.53 Mini-grids 41.0 1,245 5.68 Diesel genset 0.1 2.2 0.01 PV system 40.2 1,229 5.6 Wind turbines 0.2 6.6 0.027 Mini–small hydro 0.6 6.5 0.056 Stand-alone 13.9 407 2.23 Diesel genset 0.029 0.486 0.014 PV systems 13.9 406 2.23 Total 100 2,150.9 13.45 Source: KTH dESA.. Note: kWh = kilowatt-hours; MW = megawatts; PV = photovoltaic. Appendix A 41 Figure A.9. Scenario 9: U3–R3, HD, 0.075IM—Electrification results under the scenario defined by urban demand at tier 3, rural demand at tier 3, high diesel price, grid cost at 0.075 $/kWh, and increased imports in the grid mix Grid Diesel mini-grid Hydro mini-grid PV mini-grid Wind mini-grid Diesel stand-alone PV stand-alone N 0 70 140 280 420 560 Kilometers Source: KTH dESA. Table A.9. Scenario 9: Electrification results under the scenario defined by urban demand at tier 3, rural demand at tier 3, high diesel price, grid cost at 0.075 $/kWh, and increased imports in the grid mix People to receive electricity by 2030: 31,999,487 By technology type Share (%) Capacity (MW) Investment ($ billion) Grid extension 45.5 504 5.62 Mini-grids 40.6 1,233 5.63 Diesel genset 0.1 2.2 0.010 PV system 39.8 1,217 5.5 Wind turbines 0.2 6.5 0.027 Mini–small hydro 0.5 6.5 0.056 Stand-alone 13.9 406 2.23 Diesel genset 0.03 0.486 0.0014 PV systems 13.9 405.8 2.2 Total 100 2,142.8 13.48 Source: KTH dESA. Note: kWh = kilowatt-hours; MW = megawatts; PV = photovoltaic. 42 Appendix A Figure A.10. Scenario 10: U4–R2, LD, 0.075IM—Electrification results under the scenario defined by urban demand at tier 4, rural demand at tier 2, low diesel price, grid cost at 0.075 $/kWh, and increased imports in the grid mix Grid Diesel mini-grid Hydro mini-grid PV mini-grid Wind mini-grid Diesel stand-alone PV stand-alone N 0 70 140 280 420 560 Kilometers Source: KTH dESA. Table A.10. Scenario 10: Electrification results under the scenario defined by urban demand at tier 4, rural demand at tier 2, low diesel price, grid cost at 0.075 $/kWh, and increased imports in the grid mix People to receive electricity by 2030: 31,999,487 By technology type Share (%) Capacity (MW) Investment ($ billion) Grid extension 27.1 707 4.89 Mini-grids 1.9 32 0.13 Diesel genset 0.044 0.594 0.0018 PV system 1.8 30.2 0.122 Wind turbines 0.001 0.017 0.001 Mini–small hydro 0.073 1.38 0.0057 Stand-alone 71.0 536 2.79 Diesel genset 12.9 59.4 0.173 PV systems 58.1 476.2 2.62 Total 100 1,275.1 7.82 Source: KTH dESA. Note: kWh = kilowatt-hours; MW = megawatts; PV = photovoltaic. Appendix A 43 Figure A.11. Scenario 11: U4–R2, HD, 0.077—Electrification results under the scenario defined by urban demand at tier 4, rural demand at tier 2, high diesel price, and grid cost at 0.077 $/kWh Grid Diesel mini-grid Hydro mini-grid PV mini-grid Wind mini-grid Diesel stand-alone PV stand-alone N 0 70 140 280 420 560 Kilometers Source: KTH dESA. Table A.11. Scenario 11: Electrification results under the scenario defined by urban demand at tier 4, rural demand at tier 2, high diesel price, and grid cost at 0.077 $/kW People to receive electricity by 2030: 31,999,487 By technology type Share (%) Capacity (MW) Investment ($ billion) Grid extension 27.2 708 5.56 Mini-grids 2.0 33 0.13 Diesel genset 0.05 0.631 0.002 PV system 1.9 30.6 0.124 Wind turbines 0.01 0.165 0.0007 Mini–small hydro 0.09 1.4 0.006 Stand-alone 70.8 590 3.24 Diesel genset 0.14 0.641 0.01 PV systems 70.6 889.1 3.24 Total 100 1,330.4 8.93 Source: KTH dESA. Note: kWh = kilowatt-hours; MW = megawatts; PV = photovoltaic. 44 Appendix A Figure A.12. Scenario 12: U4–R2, HD, 0.075IM—Electrification results under the scenario defined by urban demand at tier 4, rural demand at tier 2, high diesel price, grid cost at 0.075 $/kWh, and increased imports in the grid mix Grid Diesel mini-grid Hydro mini-grid PV mini-grid Wind mini-grid Diesel stand-alone PV stand-alone N 0 70 140 280 420 560 Kilometers Source: KTH dESA. Table A.12. Scenario 12: Electrification results under the scenario defined by urban demand at tier 4, rural demand at tier 2, high diesel price, grid cost at 0.075 $/kWh, and increased imports in the grid mix People to receive electricity by 2030: 31,999,487 By technology type Share (%) Capacity (MW) Investment ($ billion) Grid extension 27.3 708 4.9 Mini-grids 2.0 33 0.13 Diesel genset 0.05 0.631 0.002 PV system 1.84 30.3 0.122 Wind turbines 0.01 0.165 0.0008 Mini–small hydro 0.09 1.4 0.007 Stand-alone 70.8 590 3.24 Diesel genset 0.14 0.641 0.0018 PV systems 70.6 588.9 3.24 Total 100 1,330.1 8.28 Source: KTH dESA. Note: kWh = kilowatt-hours; MW = megawatts; PV = photovoltaic. Appendix B Introduction to the online interface of OnSSET An example analysis for Afghanistan Workshop presentation – February 1–2, 2017, New Delhi 45 46 Appendix B Appendix B 47 48 Appendix B Appendix B 49 50 Appendix B WorldPop, UN, World Bank, World Bank IEA Population Electricity access tiers Electricity demand OSM, DABS, NASA, SRTM, JRC, Natural OSM, DABS, ADB GADM NASA Merra NASA Langley ADB, USGS HydroSHEDS, JRC Earth, IEA Existing and Power plants Small/mini- Wind Diesel cost of planned and Administrative Solar hydro capacity generating transmission economic areas irradiance potential factor electricity network activities Grid Mini-grid Stand-alone Technology selection Optimal electrification split Appendix B 51