Fikremariam Teshome Tekle

Assessment of Solar Energy

Resources in Ethiopia
Modeling solar radiation and GIS-based multi-criteria
analysis

Master's thesis in Natural Resources Management

Trondheim, May 2014

Supervisor: Ola Fredin

Norwegian University of Science and Technology

Faculty of Natural Sciences and Technology
Department of Geography
ii
Abstract
The sun is an ultimate source of energy for life on earth, and hence the knowledge of solar
radiation is essential for hydrological, climatological, biological modeling and utilization of
renewable energy resources. Solar radiation can be measured from ground instruments at
meteorological stations. Installations of instruments that directly measure global solar
radiation are quite expensive, and thus spatial density of instruments is low. For instance, in
Ethiopia there is only one Pyranometer, which is currently not functional. In addition, unlike
rainfall and temperature, solar radiation cannot be interpolated for other areas with any degree
of confidence from few sample measurements. This is mainly because the solar intensity is
highly dependent on topography and surface features. Digital Elevation Models (DEMs) are
important to derive these topographic characteristics that affect the amount of incoming solar
radiation. Estimation of solar resources in a complex topography and inaccessible areas,
where measurements are not available and/ or costly to measure over large areas from DEM is
fast, cost effective and reliable. Ethiopia is located in the tropics and thus highly endowed
with an abundance of solar energy. However, exploitation of solar energy is much
underutilized and the energy sector is the least developed. The aim of this study is to calculate
available solar energy resources and mapping the most suitable sites for large-scale PV
installations in Ethiopia. The monthly and annual global solar radiations of Ethiopia were
calculated from 30 meter resolution ASTER DEM using the ArcGIS solar radiation analysis
tools. Solar radiation potential of Ethiopia ranges from 0.2 MWh/m2/year in the lowlands
(peripheral area) to 2.6 MWh/m2/year in the central highlands. Large portion of Ethiopia
receive solar radiation exceeding 1.8 MWh/m2/year. Using GIS-based multi-criteria analysis,
about 195 sites mainly in the eastern part of the country or 6000 km2 in total, were selected as
ideal locations for large-scale PV installations in Ethiopia. The selection process takes various
topographic, economic, social and environmental factors into consideration. Selected sites
could, if exploited properly, generate more than 65 GW, which is about 1/6 of electric power
of all nuclear power plants in the world assuming 10 percent PV efficiency. Finally, a
sensitivity analysis was performed to verify the selected sites. It is hoped that this study will
promote investments in solar energy and encourage researchers for further studies in Ethiopia.

Keywords: Solar energy, Digital Elevation Models, Geographic Information Systems, Multi-
Criteria Analysis

iii
Acknowledgement
First and for most, I would like to express my deepest gratitude to Ola Fredin (Phd), who
showed interest for supervision. His constructive comments and timely guidance during this
study was very helpful. Without his support this thesis would not come to an end. My special
thanks extend to my beloved family for their moral support and encouragement during my
study. My darling ‘Enat Yene’, I thank you for your love and endless support. Finally, I like to
acknowledge the Norwegian People and Government for their financial support to my MSc
study in Norway.

iv
Contents
LIST OF FIGURES ..................................................................................................................................................... VI
LIST OF TABLES ....................................................................................................................................................... VI
ABBREVIATIONS ..................................................................................................................................................... VII
MAIN INTRODUCTION ................................................................................................................................ 1

Article manuscript
1 INTRODUCTION ..................................................................................................................................... 5
2 PREVIOUS WORK ................................................................................................................................... 8
3 MATERIALS AND METHODS ................................................................................................................. 11
3.1 THE STUDY AREA ......................................................................................................................................... 11
3.2 DATA SOURCES AND PROCESSING ................................................................................................................... 12
3.3 METHODS .................................................................................................................................................. 16
3.3.1 Modeling Solar Radiation .............................................................................................................. 16
3.3.2 Comparison Between Observed And Modeled Solar Radiation..................................................... 19
3.3.3 GIS-Based Multi-Criteria Analysis .................................................................................................. 19
4 RESULTS ............................................................................................................................................... 23
4.2 COMPARISON BETWEEN OBSERVED AND MODELED SOLAR RADIATION.................................................................. 25
4.3 GIS-BASED MULTI-CRITERIA ANALYSIS ............................................................................................................ 26
4.4 ELECTRICITY GENERATION POTENTIAL .............................................................................................................. 30
5 DISCUSSIONS ....................................................................................................................................... 32
5.1 DATA AND FACTORS..................................................................................................................................... 32
5.2 SENSITIVITY ANALYSIS ................................................................................................................................... 33
5.3 COMPARISONS ............................................................................................................................................ 35
6 LIMITATIONS AND FUTURE WORK ....................................................................................................... 38
7 CONCLUSIONS ..................................................................................................................................... 39
REFERENCES ............................................................................................................................................. 40

APPENDIX A: OBSERVED DAILY GLOBAL SOLAR RADIATION OF METEOROLOGICAL STATIONS

APPENDIX B: GIS MODELS FOR SOLAR RADIATION CALCULATION
APPENDIX C: GIS MODELS FOR GIS-BASED MULTI-CRITERIA SITE SELECTION FOR LARGE-SCALE PV
INSTALLATIONS IN ETHIOPIA
APPENDIX D: LOCATIONS AND CHARACTERISTICS OF THE SELECTED SITES FOR LARGE-SCALE PV
INSTALLATIONS IN ETHIOPIA

v
List of figures
FIGURE 1: SELECTED HIGHLY SUITABLE SITES FOR LARGE-SCALE PHOTOVOLTAIC INSTALLATIONS IN ETHIOPIA .. 3
FIGURE 2: DIGITAL ELEVATION MODEL AND LOCATION MAP OF THE STUDY AREA............................................. 12
FIGURE 3: LOCATIONS OF SELECTED METEOROLOGICAL STATIONS ..................................................................... 14
FIGURE 4: MAJOR AND MINOR ZONES ................................................................................................................. 18
FIGURE 5: DISTRIBUTION MAP OF ANNUAL GLOBAL SOLAR RADIATION IN ETHIOPIA ......................................... 24
FIGURE 6: TEMPORAL VARIATION OF MONTHLY GLOBAL SOLAR RADIATION IN ETHIOPIA ................................. 25
FIGURE 7: THE CORRELATION BETWEEN THE OBSERVED AND THE MODELED SOLAR RADIATION VALUES ........ 25
FIGURE 8: COMPARISON BETWEEN OBSERVED AND MODELED SOLAR RADIATION VALUE ................................ 26
FIGURE 9: RECLASSIFIED CRITERIA DATASETS USED IN GIS-BASED MULTI-CRITERIA ANALYSIS ........................... 28
FIGURE 10: SUITABILITY MAP FOR LARGE-SCALE PV SOLAR FARMS IN ETHIOPIA ................................................ 29
FIGURE 11: SELECTED HIGHLY SUITABLE SITES FOR LARGE-SCALE PV SOLAR FARMS IN ETHIOPIA ...................... 30
FIGURE 12: LOCATIONS OF SUITABLE AREA (PIXELS) UNDER NINE SCENARIOS ................................................... 34
FIGURE 13: ADMINISTRATIVE REGIONS OF ETHIOPIA........................................................................................... 36

List of tables
TABLE 1: CHARACTERISTICS OF SELECTED METEOROLOGICAL STATIONS............................................................. 14
TABLE 2: TYPE AND SOURCE OF DATA USED IN THE STUDY ................................................................................. 16
TABLE 3: CLASSIFICATION OF VARIOUS LAND COVERS BASED ON THEIR SUITABILITY TO SOLAR PANEL
INSTALLATION ......................................................................................................................................... 21
TABLE 4: TEMPORAL VARIATION OF MONTHLY GLOBAL SOLAR RADIATION IN ETHIOPIA ................................... 24
TABLE 5: SUITABILITY LEVEL PER CRITERIA FOR LARGE-SCALE PV SOLAR FARMS SITE SELECTION....................... 26
TABLE 6: ESTIMATED ELECTRICITY GENERATION POTENTIAL OF ETHIOPIA.......................................................... 31
TABLE 7: WEIGHT VALUE (%) FOR EACH INPUT CRITERIA UNDER NINE SCENARIOS ............................................ 33
TABLE 8: SUMMARY OF SELECTED SITES PER REGION .......................................................................................... 36

vi
Abbreviations
Abbreviation Explanation
a.m.s.l above mean sea level
AHP Analytic Hierarchy Process
ASTER Advanced Space borne Thermal Emission and Reflection Radiometer
b.m.s.l below mean sea level
cal/cm2/day calories per square centimeter per day
CIA Central Intelligence Agency
CSA Central Statistical Agency
DEM Digital Elevation Models
EEPCo Ethiopian Electric Power Corporation
ERA Ethiopian Roads Authority
ERSDAC Earth Remote Sensing Data Analysis Center
ESRI Environmental System Research Institute
FAO Food and Agricultural Organization
FDRE Federal Democratic Republic of Ethiopia
GIS Geographic Information Systems
GWh Gigawatt hour
ha hectare
km kilometer
km2 square kilometer
KWh Kilowatt hour
LCCS Land Cover Code Systems
LiDAR Light Detection and Ranging
m meter
MCA Multi-Criteria Analysis
METI Ministry of Economy, Trade, and Industry
mm millimeter
MoIWE Ministry of Irrigation, Water and Energy
MWh Megawatt hour
NASA National Aeronautic and Space Administration
NMA National Meteorological Agency
NREL National Renewable Energy Laboratory

vii
NTNU Norwegian University of Science and Technology
PV Photovoltaic
SPSS Statistical Packages for Social Sciences
SRTM Shuttle Radar Topographic Mission
TWh Terawatt hour
UN United Nations
USA United States of America
Wh/m2 watt hours per square meter

viii
Main introduction
Energy is the main economic artery of countries all over the world (Toman & Jemelkova,
2003), and can indeed be considered as the primary commodity in economy. The demand for
energy is significantly increasing with increases in population and the rapid development of
world’s economy. As a result, the price of energy from conventional sources such as fossil
fuels and nuclear power, are escalating through time. Nowadays, about 80 percent of the
global primary energy demand is comes from fossil fuels, which are limited and thus become
too expensive to exploit in the future (Uyan, 2013). While a majority of the world's energy
supply is generated from fossil fuels, release of fossil carbon into the atmosphere has
tremendous impact on the global climate. The production and consumption of energy from
fossil fuels is by far the largest sources of greenhouse gases. Increasing greenhouse gas
concentrations depletes the ozone layer, this in turn increase earth’s average temperature,
influence the patterns and amount of precipitations, reduce ice and snow cover, and raise sea
level (Hardy, 2003). These changes will impact the food supply, water resources, the
ecosystem, and earth’s environment [6]. Renewable energy systems can help to mitigate
global problems, such as climate change, poverty in developing world and insufficiency of
energy supply from the conventional sources of energy (Kaygusuz, 2001; Šúri, Huld, Dunlop,
& Hofierka, 2007).

Renewable energy refers to energy that is derived from resources that are constantly replaced
in nature and usually less polluting the environment. Examples are biomass, sunlight, wind,
water, geothermal heat and tide. They are abundant and environmental friendly sources of
energy; and thus contribute to a sustainable economic development. To exploit the potential
of renewable energy sources, there is a need to assess extent of available renewable energy
sources spatially as well as temporally (Datta & Karakoti, 2010). The geographical
dependency and dispersed nature of renewable energy generation impose questions that
require specific location-dependent answers (Šúri, Huld, Dunlop, & Hofierka, 2007).
Therefore, map based assessment of availability of renewable energy resources greatly
contributes to increase investments and setting up effective renewable energy policies (Šúri,
Huld, Dunlop, & Ossenbrink, 2007). Geographic Information Systems (GIS) helps in
mapping available renewable resources on spatial and temporal scales and identifying ideal
location for renewable energy development.

Solar energy is a vast and environmentally benign source of renewable energy, which is
available everywhere on the planet however with significant variance. Solar radiation is the
1
key deriver of natural processes on the surface of the earth (Šúri, Huld, Dunlop, & Hofierka,
2007). Air and soil temperature and moisture, snow melting, photosynthesis,
evapotranspiration processes are highly dependent on solar radiation (Tovar‐Pescador et al.,
2006). It plays a fundamental role in the energy exchange process between the atmosphere
and the earth’s surface. As a result, the knowledge of solar radiation is important for
hydrological, climatological, biological modeling and applications (Martínez-Durbán et al.,
2009). It also follows, that detailed understanding of the incoming solar radiation is vital for
utilization of solar energy resources from solar panels at different scales.

Measurement of solar radiation at a wide spatial coverage is difficult because measuring

instruments are very expensive. Unlike temperature and rainfall, solar radiation cannot be
measured at sample meteorological stations and interpolated for the surrounding areas
(Kumar, 2004).This is mainly because incoming solar radiation is highly dependent on
topography and surface characteristics (Gastli & Charabi, 2010b; Li, 2013)[2]. Digital
Elevation Models (DEMs) are useful to derive topographic characteristics such as elevation,
slope, aspect and shadowing effects of neighboring terrain features, and thus helpful to
estimate solar radiation particularly in a complex topography at a wide spatial coverage (Šúri,
Cebecauer, & Skoczek, 2011).

The overall objective of this study is to assess solar energy resources of Ethiopia. In
particular, available solar radiation potential was calculated from a high resolution DEM and
ideal locations for large-scale Photovoltaic (PV) installations were selected using GIS-based
Multi-Criteria Analysis.

The monthly and annual global solar radiations of Ethiopia were calculated from 30 m
resolution ASTER Global DEM using Environmental System Research Institute (ESRI)
ArcGIS solar radiation analysis tools. According to the GIS model, large portion of Ethiopia
receives solar radiation exceeding 1.8 MWh/m2/year. In total, 195 sites located in different
regions at 0.5 percent of the total area of Ethiopia, were selected with aggregate electricity
generation potential of 65 GW per year purely from solar radiation. The site selection takes
various topographic, economic, social and environmental criteria into consideration. Finally,
locations of the selected sites were verified through sensitivity analysis by altering the weight
values of input criteria.

2
 Figure 1: Selected highly suitable sites for large-scale Photovoltaic installations in Ethiopia

It is the aim to publish this study in an international journal such as Solar Energy, Renewable
Energy, Renewable and Sustainable Energy Reviews, and thus to disseminate the results to a
wider audience. Hence this thesis is submitted as article-based. It is hoped that at least one
paper could be published titled Assessment of Solar Energy Resources in Ethiopia: modeling
solar radiation and GIS-based multi-criteria analysis.

3
 Article manuscript:

Assessment of Solar Energy Resources in Ethiopia:

Modeling solar radiation and GIS-based multi-criteria analysis

4
1 Introduction
The sun is the ultimate source of energy for life on earth. Quantitatively, about 99.8 percent of
energy at the surfaces of earth is derived from the Sun (Dickinson & Cheremisinoff, 1980). It
is a powerhouse that sustains life on the earth surface. Solar energy is vast, free, clean and
renewable power source that is available everywhere on the planet. However, the intensity of
solar radiation reaching the surface of the earth shows considerable spatial and temporal
variation at the local and on a global scale.

Solar radiation is a critical driver of evapotranspiration, surface and air temperatures, wind,
growth and activity of plants and animals and other major processes on the earth surface
(Kumar, Skidmore, & Knowles, 1997). It plays a major role in the energy exchange process
between the atmosphere and earth’s surface. As a result, the knowledge of solar radiation is
important for hydrological, climatological, biological modeling and applications as well as
utilization of renewable energy resources (Martínez-Durbán et al., 2009).

Solar power generation is a promising source of renewable energy and could alleviate global
problems such as climate change, poverty in developing world and insufficient security of
energy in the future (Šúri, Huld, Dunlop, & Hofierka, 2007). Thus, it contributes to a
sustainable development of a region. However, in developing world the exploitation of solar
energy is at a very low level.

The development of solar energy requires accurate estimation of the available solar energy
resources and suitable sites for solar photovoltaic (PV) installations. The generation and
distribution of solar energy is highly dependent on the geographical location and topography.
Accurate knowledge of the available solar energy resource is very important for the design of
any solar-based energy system. This knowledge can significantly contribute to better siting
and economic assessment of the new installations, monitoring of their performance and
forecasting of delivered energy (Šúri, Huld, Dunlop, & Hofierka, 2007).

Solar radiation can be measured from ground meteorological stations. The values measured
from geographically dispersed stations can in theory be interpolated to generate a continuous
solar map for the surrounding region (Li, 2013). However, intensity of solar radiation is
highly dependent on topography and surface features. As a result, unlike rainfall and
temperature that can be measured at a few locations and interpolated, solar radiation cannot be
interpolated to a high degree of confidence (Kumar, 2004). Interpolation provides reasonable
estimates in homogeneous terrain with uniform climatologic characteristics but in areas with

5
considerable topographic relief these estimates become unreliable (Tovar‐Pescador et al.,
2006). Installations of instruments that directly measure global solar radiation such as
Pyranometers in various meteorological stations are quite expensive. As a result, in
developing countries there are very few instruments for direct global solar radiation
measurement. For instance, in Ethiopia there is only one Pyranometer which is currently not
functional (Mekonnen, 2007). In most cases, meteorological data are heterogeneously located
and they do not represent spatial and vertical variability of solar radiation adequately (Šúri,
Huld, Dunlop, & Hofierka, 2007).

In complex topography, data interpolation is not adequate and reliable modeling of solar
radiation is important, since intensity is highly dependent on elevation, slope, aspect and
shadowing effects of neighboring terrain features (Bosch, Batlles, Zarzalejo, & López, 2010;
Kumar, Skidmore, & Knowles, 1997). At the local scale, topography is the most important
factor that determines the distribution of solar radiation on the surface (Martínez-Durbán et
al., 2009; Tovar‐Pescador et al., 2006). Digital Elevation Models (DEMs) are very helpful to
derive these important topographic characteristics that affect the amount of incoming solar
radiation. Relatively high resolution DEMs, for instance derived from Shuttle Radar
Topographic Mission (SRTM), is very important in order to improve knowledge of solar
energy resources on a regional scale (Šúri, Huld, Dunlop, & Ossenbrink, 2007). Therefore,
estimation of solar energy resources in a complex topography on a wide spatial coverage from
DEM is fast, cost effective and reliable. Moreover, the solar radiation map derived from DEM
can be integrated with other data from various sources within Geographic Information
Systems (GIS) to identify suitable sites for large-scale PV installations, and reduce risk and
cost of potentially large energy investments. However, the estimation quality depends on
DEM resolution and accuracy. Ethiopia is located in the tropics and thus highly endowed with
an abundance solar energy. In fact, ‘13 months of sunshine’ is the motto of Federal
Democratic Republic of Ethiopian (FDRE) Ministry of Culture and Tourism. But the energy
sector of Ethiopia is the least developed in the world and highly dependent on traditional
sources of energy such as wood, crop residuals and animal waste. According to CIA, the
electricity consumption per capita of Ethiopia in 2012 was 36.8 KWh per person, which is
very low even in African standards [6]. Ethiopia is a mountainous country, where more than
46 percent of the total land mass is above 1,500 meter a.m.s.l. This highland ecosystem
supports 80 percent of the total human population of the country (Tadesse & Belay, 2004).

6
Therefore, calculation of solar radiation from high resolution DEM using ArcGIS is a
promising path to assess the solar energy potential of the country.

The aim of this study is to assess solar energy resources in Ethiopia based on high resolution
DEMs using ArcGIS tools. Furthermore, optimal sites for extensive PV installations is
calculated based on solar energy potential, population density, terrain, land use and several
other factors. By doing this it is hoped that discussions can be initiated on using renewable
solar power in general and in Ethiopia in particular.

7
2 Previous work
Related studies were conducted in different parts of the world using different approaches to
calculate solar radiation potential and select ideal sites for PV installations. I have presented
the following eight studies conducted at different scales from developing and developed
countries.

Mekonnen (2007) has applied physical and empirical models with inputs of various
meteorological data to predict the monthly average global radiation in Ethiopia for 17
stations. The models used in his analysis are the Simple Model of Atmospheric Radiative
Transfer of Sunshine (SMARTS) and Vapor Pressure Radiation Model (VP-RAD). Maximum
and minimum temperature, relative humidity, precipitation and surface pressure are some of
the data used in this study. The monthly average results for Ethiopia, which ranges from 19.5
MJ m-2 day-1 to 20 MJ m-2 day-1 are observed from the SMARTS model prediction for the 17
stations. The VPRAD model is less accurate for global radiation predictions in Ethiopia. He
concluded that Ethiopia is characterized on average, by relatively high daily global irradiance
rates and high percentage of clear days.

Datta and Karakoti (2010) have mapped the solar potential in North-West regions of India
based on global, diffuse, and direct solar radiation, which provides an estimate of potential
sites for solar power generation applying GIS. They have demonstrated that importance of
GIS for mapping both spatial and temporal variability of solar energy supply and demand.

Gastli and Charabi (2010) have developed the first geographical mapping models to locate the
most appropriate sites for different Concentrated Solar Power (CSP) technologies in Oman.
They calculated the global distribution of solar radiation within the boundaries of Oman,
based on a DEM with 3 km resolution using ArcGIS Solar Radiation Analysis tools. Only
slope analysis was considered in calculating the yearly electricity generation potential for
different Concentrated Solar Power (CSP) technologies, and thus only 10 percent of the land
of Oman with a slope of less than 1 percent was considered as exploitable land. The total
calculated potential of yearly electricity generation would be about 7.6 million GWh, which is
many times higher than the current generation supply in Oman which was about 11,189 GWh
in 2007. The results obtained show very high potential of solar radiation over all the lands of
Oman.

Janke (2010) has identified areas that are suitable for wind and solar farms using multi-criteria
GIS modeling techniques in Colorado, USA. Renewable potential (NREL wind speed

8
measurements at 50 m above the ground and NREL annual insolation data), land cover,
population density, federal lands, and distance to roads, transmission lines, and cities were
considered for finding suitable locations. The GIS model indicates that ideal areas for wind
and solar farm developments are located in northeastern and northwestern part of the state and
east of Denver at 41,850 km2 and 191 km2, respectively. The GIS model is very well
equipped at detecting regional renewable energy facilities that are capable of supporting large
urban populations, he concludes.

Bhuiyan (2013) estimated empirically global, diffuse and direct solar radiation on a
horizontal surface for ten stations equally distributed all over Bangladesh as well as
predicted correlations for them. For this study, meteorological data (temperature and sunshine
hour to predict global radiation; relative humidity and atmospheric water to estimate diffuse
radiation; and sunshine duration to estimate direct radiation) for the years 1980 to 2007 was
used. The global radiation in Bangladesh is found to be highest in the months of April and
May and lowest in the months of November and December in all the districts. He concluded
that Bangladesh is endowed with sufficient solar radiation throughout the year and
correlations proposed for Bangladesh in this study can be used in future for estimation
of solar radiations if the meteorological data are available.

Charabi and Gastli (2011) have assessed solar energy resources in Oman, in particular GIS-
based spatial multi-criteria evaluation approach. In particular the FLOWA module was used
to assess land suitability for large PV farms installations. The tool used applies fuzzy
quantifiers within ArcGIS environment allowing the integration of a multi-criteria decision
analysis. The overlay results obtained from the output maps showed that 0.5 percent of the
total land area demonstrate a high suitability level. It was also found that the CPV technology
provides very high technical potential for implementing large solar plants. In fact, if all highly
suitable land is completely exploited for CPV implementation, it can produce almost 45.5
times the present total power demand in Oman.

Li (2013) has employed GIS and remote sensing techniques for determining the optimal sites
for solar panel installation at micro and macro-scales: building rooftop installations at the
University of Waterloo main campus and ground-mounted installation in the City of
Waterloo, Canada, respectively. Light Detection and Ranging (LiDAR) and DEMs were used
to derive accumulated solar radiation energy under clear-sky and overcast conditions using
ArcGIS software. Optimal ground-mounted solar panel installation sites were determined
using a multi-criteria analysis approach that considered various environmental and

9
socioeconomic factors. He identified building roofs with a southern exposure without
obstructions and four potential sites close to the periphery of the city at micro and macro
scales. Moreover, Li conducted a questionnaire survey of three Ontario solar companies to
obtain information about how solar companies in Ontario select potential sites for solar PV
installation. Finally, a feasibility assessment was performed with ground truth information to
verify selected sites.

Uyan (2013) has determined suitable site selection for solar farms by using GIS and
Analytical Hierarchy Process (AHP) in Karapinar region, Turkey. He considered the quality
of terrain, local weathering factors, proximity to high transmission capacity lines, agricultural
facilities and environmental conservation issues in the selection process. He found that 13.92
percent of the study area (840.07 km2) as highly suitable for solar farms.

Based on these pioneering studies it is concluded that GIS analysis, where potential incoming
solar radiation is calculated using a relatively high resolution DEM, together with
considerations of physiographic and socioeconomic factors would be the most suitable to map
solar power potential in a relatively large country such as Ethiopia.

10
3 Materials and methods

3.1 The study area

Ethiopia is located between latitude 3° and 15° North and longitude 33° and 48° East in the
north eastern Africa, sometimes also known as Horn of Africa (Fig.1). It shares about 5328
km boundary line with Eritrea in the north, Djibouti in the east, Somalia in the south east,
Kenya in the south, South Sudan and Sudan in the west. With over 80 million human
inhabitants, Ethiopia is the second most populous country in Africa [3]. The total area of the
country is 1.1 million km2 [4].

Ethiopia is a country of great geographical diversity. Located within the tropics, its physical
conditions and variations in altitude have resulted in a great range of terrain, climate, soil,
flora and fauna. The geography also determines the population settlement patterns of the
country. Ethiopia has high and rugged mountains, flat-topped plateaus, deep gorges, incised
river valleys and vast rolling plains. Its altitude ranges from the highest peak at Ras Dashen in
the north (4620 meter a.m.s.l) down to the Dallol (also known as the Danakil) depression in
the east, approximately 148 meter b.m.s.l [4]. The Great Rift Valley separates the western and
the eastern highlands; and these highlands gradually descend to the lowland areas in the east,
west, and south of the country [4]. Much of the country thus consists of high plateau and
mountain ranges, which are dissected by numerous streams and rivers. Among the biggest
river basins are Blue Nile, Awash, Baro, Omo, Tekkeze, Wabe Shebelle and Genale.

Ethiopian climate varies according to the different topographical regions. The central plateau
has a favorable climate with minimal seasonal temperature variation. In contrary, temperature
variations in the lowlands are much greater. There are two distinct seasons in Ethiopia, the dry
season which is for the most part from October to May and rainy season from June to
September (Mekonnen, 2007). The mean annual temperature ranges from 20oC in the central
plateau to 30oC in the lowlands. Heavy rainfall occurs in most of the country during June,
July and August and smaller rains in February and March. The mean annual rainfall varies
between 200 mm in the east and north to 2000 mm on the central highlands.

Ethiopia is an ecologically diverse country, ranging from the deserts along the eastern border
to the tropical forests in the south to extensive Afro-mountain in the northern and
southwestern parts.

11
 Figure 2: Digital Elevation Model and location map of the study area
3.2 Data sources and processing

Various spatial and non-spatial datasets were obtained from different organizations and
processed using multiple GIS tools for mapping and analysis purposes. A relatively high
resolution, accurate and geo-referenced DEM dataset is the most important input to calculate
solar radiation using ArcGIS solar radiation analysis tools. DEMs were needed to model
important topographic characteristics such as elevation above sea level, slope inclination,
terrain orientation and shadowing effects of the neighboring terrain features that affect the
amount of incoming solar radiation. For this study DEMs were derived from the Advanced
Space-borne Thermal Emission and Reflection Radiometer (ASTER) global DEMs version 2
[9]. The Ministry of Economy, Trade, and Industry (METI) Earth Remote Sensing Data
Analysis Center (ERSDAC) in Japan and the National Aeronautic and Space Administration
(NASA) Earth Observing System (EOS) Data Information System (DIS) Land Processes (LP)
Distributed Active Archive Center (DAAC) in the United States jointly released a second
version of the ASTER Global DEM without charge as a contribution to the Global Earth
Observing System in mid-October, 2011 (P. Li et al., 2012). ASTER DEM is distributed as 1°
x 1° tiles, and therefore 175 tiles that cover about 2.1 million km2, which contained 1.1

12
million km2 study area, were downloaded from NASA thus allowing the horizon effects on
the estimated solar radiation to be taken into account properly. The ASTER Global DEM
version 2 has additional scenes to reduce artifacts, higher horizontal resolution using smaller
correlation kernel and an improved water mask compared with the previous version. The
ASTER Global DEMs obtained elevation data on a near-global scale to generate the most
complete high-resolution digital topographic database of earth. Global DEMs data were
released at a 1 arc-second (30 meter) resolution (Tachikawa, Hato, Kaku, & Iwasaki, 2011).

In order to evaluate the reliability of solar radiation values modeled from DEMs using the
solar radiation analysis tools in ArcGIS, solar intensity data of ten stations in the study area
were collected from National Meteorological Agency (NMA) of Ethiopia. The meteorological
stations are Addis Ababa, Arba Minch, Bahir Dar, Kombolicha, Debre Tabor, Dedessa, Dire
Dawa, Metehara, Nazreth and Welayita Sodo. The locations of these stations range from 930
meter a.m.s.l at Metehara to 2690 meter a.m.s.l at Debre Tabor. These stations have relatively
complete and recent daily solar intensity recorded in cal/cm2/day. Due to absence of
Pyranometers in different parts of the country and recent solar intensity values measured
using Actinographs, the daily solar intensities of these stations were thus recorded in 1990s
from Actinographs by NMA. In Ethiopia the NMA is responsible for installing meteorological
stations, collecting all meteorological data, archiving of meteorological and climatological
data as well as providing meteorological services to all stakeholders [5]. It is also responsible
for publishing and disseminates analyzed and interpreted meteorological data and
meteorological forecasts. Figure 3 and table 1 show the location and the characteristics of the
selected meteorological stations, respectively.

13
 Figure 3: Locations of selected meteorological stations
Table 1: Characteristics of selected meteorological stations

Station Station Longitude Latitude Elevation Data Recording

Number Name (East) (North) (meter a.m.s.l) (year)

1 Addis Ababa 38.45 9.02 2408 1997

2 Arba Minch 37.08 6.05 1290 1995
3 Bahir Dar 37.25 11.36 1770 1996
4 Debre Tabor 38.02 11.53 2690 1995
5 Dedessa 36.06 9.23 1200 1990
6 Dire Dawa 41.51 9.36 1260 1997
7 Kombolcha 39.44 11.07 1903 1995
8 Metehara 39.54 8.52 930 1996
9 Nazreth 39.17 8.33 1622 1997
10 Welayita Sodo 37.45 6.51 1800 1997

14
In addition to solar radiation potential, selection of ideal sites for large-scale solar panel
installations require other datasets such as land cover, existing road network, existing power
transmission network, population density, and environmental protected areas. Land cover is
important for natural resource assessment and management, environmental modeling and
decision-making. A land cover map of the study area was obtained from Food and
Agricultural Organization (FAO) of the United Nations [11]. This land cover database was
provided as ESRI shape file (vector format) and is sourced from reprocessing the raster based
Globcover database (regional version) [12]. It has been post-processed to generate a vector
version at national extent with the Land Cover Code System (LCCS) regional legend (46
classes). Globcover is currently the most recent (2005) and finest resolution (300 meter)
datasets on land cover globally. It was published in 2008 and intended for free public access
[11].

Population density refers to number of persons per km2. Population density map of the study
area was mapped using an analogue map of Ethiopian Somali region and digital map of other
regions and number of population per Woreda1 obtained from the FDRE Central Statistical
Agency (CSA) of Ethiopia. Population density mapping was thus performed through: a)
scanning, geo-referencing and digitization of Ethiopian Somali region Woreda’s boundary
from the analogue map; b) merging with other region’s Woreda map that are initially obtained
in digital format; c) joining attribute data; and d) classification based on number of people per
km2 in the Woreda boundary.

Road and power transmission network maps of Ethiopia were also important for site selection,
since existing infrastructure determine cost of the construction, transmission and maintenance
(Janke, 2010). Digital road network map and analogue power transmission network maps
were obtained from the FDRE Ethiopian Roads Authority (ERA) and the FDRE Ethiopian
Electric Power Corporation (EEPCo) respectively. To perform GIS analysis the analogue map
was scanned, geo-referenced, and finally digitized to get the digital copy of it. Maps of parks
and environmental protected areas were obtained from the FDRE Environmental Protection
Authority (EPA). In addition, lakes and river network map of Ethiopia were obtained from the
FDRE Ministry of Irrigation, Water and Energy (MIWE) in digital format. All these datasets
were used as criteria for siting large-scale solar panel installations. Type and source of data
used in this study are summarized in the following table.

1
Woreda or district: the second-level administrative divisions of Ethiopia. They are composed of Kebeles which
are the smallest unit of local government.
15
 Table 2: Type and source of data used in the study
Data Final Final
Data Source Original Data
Year Data Resolution
ASTER Global Digital Elevation The US National Aeronautics and Space
Raster 2011 Raster 30 m
Models version 2 Administration (https://earthdata.nasa.gov/)
Observed solar intensity The FDRE National Meteorological Agency Charts 1997 NA NA
Food and Agricultural Organazation of the
Land cover Digital, Polygon 2005 Raster 30 m
UN (http://www.fao.org)
Population density (Woreda Analogue& Digital,
The FDRE Central Statistical Agency 2007 Raster 30 m
boundary & No. of population) Polygon & Table
Road network The FDRE Ethiopian Roads Authority Digital, Polyline 2007 Raster 30 m
The FDRE Ethiopian Electric Power
Power transmission network Analogue, Polyline 2010 Raster 30 m
Corporation
The FDRE Ministry of Water, Irrigation and Digital, Polyline and
Rivers and Lakes 2007 Raster 30 m
Energy Polygon
The FDRE Environmental Protection
Environmentally Protected Areas Digital, Polygon 2007 Raster 30 m
Authority
Boundary of Ethiopia The FDRE Central Statistical Agency Analogue, Polygon 2007 Vector NA

3.3 Methods

Monthly and annual global solar radiation maps of Ethiopia were calculated from 30 meter
resolution DEM, which subsequently used as criteria for suitable site selection. A GIS-based
Multi-Criteria Analysis (MCA) approach was adopted to select sites for large-scale PV solar
farms in Ethiopia. This section discusses the methodology of the study in detail.

3.3.1 Modeling solar radiation

Solar radiation reaching the surface of the earth shows considerable spatial and temporal
variation at the local and on a global scale. On a global scale the major controlling factors are
the latitude, distance from the sun, and time of year [2]. Whereas at the local scale major
sources of spatial variation are elevation above sea level, surface inclination, surface
orientation, and shadowing effects of neighboring terrain features. These topographic
parameters can be modeled and derived from high resolution DEMs. As a result, the most
important input to calculate solar radiation was an accurate, geo-referenced and high
resolution DEM dataset. As mentioned earlier, for this study 30 meter resolution DEMs were
derived from ASTER Global DEM version 2.

In the last decade different individuals and institutions have developed several GIS-based
radiation models for calculating incoming radiation on global areas represented by DEMs
(Tabik, Villegas, Zapata, & Romero, 2012). Solar Analyst and r.sun are some of the tools
developed under ArcGIS and GRASS GIS respectively used to calculate direct and diffuse

16
radiation maps over smaller areas. As this study deals with calculation of solar radiation
covering comparably much larger areas using high resolution DEMs, the solar radiation tools
found in ArcGIS and developed by ESRI, was found to be a suitable tool for this study. The
solar radiation analysis tools allows mapping and analysis the effects of sun over a geographic
area for specific time periods (Charabi & Gastli, 2010; Gastli & Charabi, 2010b)[1]. It
calculates insolation based on methods from the hemispherical viewshed algorithm developed
by (Rich, Dubayah, Hetrick, & Saving, 1994). The solar radiation tool accounts for
atmospheric effects, site latitude and elevation, steepness (slope) and compass direction
(aspect), daily and seasonal shifts of the sun angle, and effects of shadows cast by surrounding
topography (Li, 2013) [2]. The output radiation raster is floating-point type, have units of watt
hours per square meter (Wh/m2) and could be easily integrated with other GIS data for further
applications.

Due to earth’s atmosphere, all the radiation, which comes from the sun will not directly reach
earth’s surface. It usually is attenuated by absorption, scattering and reflection before reaching
the land surface. The total (global) radiation at the surface of the earth is the sum of direct
solar radiation, scattered (diffuse) solar radiation, as well as reflected solar radiation (Kumar
et al., 1997). Direct (beam) radiation is the largest and the most important component of
global radiation and diffuse radiation is the second largest component. Reflected radiation
generally constitutes only a small proportion of total radiation, except for locations
surrounded by highly reflective surfaces such as areas with extensive snow cover [2].
Calculations of radiation using ArcGIS solar radiation analysis tool do not include reflected
radiation in the calculation of total radiation. As a result, the total radiation obtained from
ArcGIS is the sum of the direct and diffuse radiation. Since snow cover and other reflective
surface are low in the study area, it is assumed that absence of reflected radiation does not
affect the calculated output to a significant degree. Moreover, ArcGIS solar radiation analysis
tool do not consider the effects of clouds in the solar radiation calculation.

With large DEM dataset, the insolation would differ significantly at different latitudes (greater
than 1 degree). To analyze broader geographic regions like in this study, it is necessary to
divide the study area into zones with different latitudes. Since the study area is situated
between latitude 3° and 15° north, it is divided into 13 zones with different latitudes. In
addition to these major zones, minor zones (latitude 0.4° DEM tile) that share common area
from two consecutive major zones were created to properly take in to account the effect of
neighboring topography in the calculation (figure 4).

17
 Boundary

Major
Zone 04

Major Zone 03
Minor zone

Figure 4: Major and minor zones

Since solar radiation can be greatly affected by topography and surface features, a key
component of the calculation algorithm requires the generation of an upward-looking
hemispherical viewshed for every location in the DEM [2]. The viewshed is a raster
representation of the entire sky that is visible or obstructed when viewed from a particular
location. In addition, it needs the sun's position in the sky across a period of time (sunmap)
and the sectors of the sky that influence the amount of incoming solar radiation (skymap) [1].
The solar radiation tools calculations essentially involves four steps:

1. Calculation of an upward-looking hemispherical viewshed based on topography.

2. Overlay of the viewshed on a direct sunmap to estimate direct radiation.

3. Overlay of the viewshed on a diffuse skymap to estimate diffuse radiation.

4. Repeating the process for every location of interest to produce an insolation map.

In addition to DEM resolution, the accuracy and processing time of insolation calculations
depend on sky size and time interval. Sky size is the resolution of the viewshed, sky map, and
sun map rasters which are used in the radiation calculations (units: cells per side). Bigger sky
size and smaller time interval increases accuracy but also increases calculation time
considerably (Gastli & Charabi, 2010b; Li, 2013)[2]. As a result, in this study sky size of
2,800, which is recommended to calculate insolation for a location at the equator and the
default 14 days and 0.5 hour interval were used. To increase calculation accuracy the number
of calculation direction was also set to 32, which is adequate for complex topography [2].
Appropriate z factor was assigned to adjust the units of measure for z units as they are
different from the ground x,y units of the input DEM raster.

Solar radiation of whole year with monthly interval was calculated for each major and minor
zone. Insolation calculation from large and relatively higher resolution DEMs are time
consuming and require high speed computers. For this study the calculation of annual and
monthly global solar radiation for the whole Ethiopia takes about three months from October,
2013 to January, 2014 using multiple high speed computers running in parallel, processing

18
one zone each at the department of Geography GIS lab, NTNU. Finally, using the ArcGIS
mosaic tool annual and monthly solar radiation for each month in a year was calculated for
the whole Ethiopia. This approach has the advantage of providing fast, cost effective and
reliable results on a wide spatial coverage.

3.3.2 Comparison between observed and modeled solar radiation

In order to compare ground recorded global solar radiation with GIS-modeled solar radiation,
monthly solar radiations of the ten sample stations were also calculated using the ArcGIS
point solar radiation tool. Point solar radiation tool calculate incoming solar radiation for
specific locations in a point feature or location table from DEM surface. To compute the
correlation, mean monthly radiation value have been used for every month of the year for all
meteorological stations, providing 120 values. 26 values were excluded from the final
computation due to incompleteness. Finally, simple regression analysis in SPSS was
performed to show the correlation between 94 observational data and GIS-modeled solar
radiation.

3.3.3 GIS-based Multi-Criteria Analysis

Site selection process for large-scale solar panel installation involves various social,
technological, environmental and economic aspects. GIS-based Multi-Criteria Analysis (GIS-
MCA) provide a flexible tool that is able to handle and bring together wide range of variables
assessed in different ways and thus offer useful assistance to the decision maker in pointing at
suitable locations (Taha & Daim, 2013). It is a process that transforms and combines spatial
data and value judgments to obtain information for decision making (Malczewski, 1999).

In order to select best sites for large-scale PV solar farms in Ethiopia, I have used the
following six steps. The initial step is set the goal or problem definition. In this study the aim
of GIS-based MCA is to produce a map showing ideal sites for large-scale PV installations in
Ethiopia. The objective was specific, measurable and attainable. The second step was
determining factors/ criteria that are important for large-scale solar panel installations. The
factors used for site selection process for a particular land use should be selected by the
experts and researchers of that particular land use. Based on literatures and related studies
conducted in different areas (Charabi & Gastli, 2011; Gastli & Charabi, 2010a; Janke, 2010;
Li, 2013; Turney & Fthenakis, 2011; Uyan, 2013) solar radiation potential, land cover,
population density, proximity to existing road networks, proximity to existing power
transmission networks, and distance to water sources as well as slope and aspect were selected

19
as factors that are important for the large-scale PV solar farms site selection process in
Ethiopia. These datasets were collected from various sources and processed using ArcGIS
software for MCA. These are not the only factors that are important to select sites for large-
scale PV installations. Other datasets such as wind speed map, map of cultural heritage sites,
and future land use map at the study area, which needs to be considered were unavailable
during this study.

The third step was standardizing each factors/ criterion scores. It is important to set the
suitability values of the factors to a common scale to make comparisons possible. Therefore,
all input datasets were changed into a raster at 30 meter resolution and to a common
measurement scale using the Conversion and Reclassify tools found in ArcGIS respectively.
They were reclassified from 1 to 5; where 5 represent the most suitable conditions for large-
scale PV installations. For instance, the solar radiation raster was reclassified from 1 to 5;
where 5 represent the highest solar radiation value and thus the most favorable condition for
solar panel installation.

Land cover data set consisted of 20 land cover classifications identified within study area. The
polygons representing different land cover types were converted into a raster and finally
categorized into five classes according to their suitability for large-scale PV solar farms.
Suitable land cover includes bare areas, grassland, and savannas. Permanently unsuitable land
cover contained forest, urban areas, water bodies, flooded land. Crop lands, closed grass land,
sparse vegetation, shrub lands were not as highly suitable and categorized as marginally
suitable land cover. Table 3 below show classification of different land cover types identified
in the study area based on their suitability to solar panel installation.

Population density map was converted into a raster with a 30 m resolution and standardized
from 1 to 5; where 5 indicate the lowest population density and hence the most ideal condition
for large-scale PV installations. A Euclidian distance raster was calculated at a 30 m
resolution for the existing road and power transmission networks as well as distance to water
source. Then, the Euclidean distance raster was reclassified from 1 to 5; where 1 represent
faraway locations and thus the least desirable areas while 5 represent locations very closer to
existing roads and transmission lines and water sources, and thus the most ideal location for
solar energy development.

Slope and aspect were other factors that determine suitability of a given land for PV solar
farms. These data were derived from 30 meter resolution ASTER DEM using the ArcGIS
spatial analyst tools and reclassified from 1 to 5 based on their suitability to solar panels
20
installations. Gentle slope areas facing south, south east and south west directions are the
most desirable location for large-scale solar panel installations. Because southerly facing land
surfaces have the higher exposure to sun the northern hemisphere (Gastli & Charabi, 2010a;
Li, 2013) [15].

Table 3: Classification of various land covers based on their suitability to solar panel
installation
Highly suitable Moderately suitable Marginally suitable Not suitable Non suitable
LCCS Description LCCS Description LCCS Description LCCS Description LCCS Description

Closed to open (>15%)

Mosaic grassland (50- Mosaic forest or
broadleaved evergreen or
151 Sparse (<15%) grassland 120 70%) / forest or 14 Rainfed croplands 110 shrubland (50-70%) / 40
semi-deciduous forest
shrubland (20-50%) grassland (20-50%)
(>5m)
Mosaic cropland (50- Closed to open (>15%)
Closed to open (>15%) Closed (>40%)
70%) / vegetation (broadleaved or
herbaceous vegetation broadleaved evergreen
200 Bare areas 140 20 (grassland/ 130 needleleaved, evergreen 41
(grassland, savannas or and/or semi-deciduous
shrubland/forest) (20- or deciduous) shrubland
lichens/mosses) forest (>5m)
50%) (<5m)
Mosaic vegetation Closed to open
Consolidated bare areas Open (15-40%)
Sparse (<15%) (grassland/ shrubland/ broadleaved forest on
201 (hardpans, gravels, bare 150 30 162 60 broadleaved deciduous
vegetation forest) (50-70%) / temporarily flooded land -
rock, stones, boulders) forest/woodland (>5m)
cropland (20-50%) Fresh water
Closed to open (>15%)
grassland or woody
Artificial surfaces and
Non-consolidated bare vegetation on regularly
202 141 Closed (>40%) grassland 180 190 associated areas (Urban
areas (sandy desert) flooded or waterlogged
areas >50%)
soil - Fresh, brackish or
saline water

210 Water bodies

The fourth step was defining weights for each criterion based on its importance for the large-
scale PV installations. Several methods are available to determine the weight and in this
study the Analytical Hierarchy Process (AHP) method was adopted to assign weight for each
factor (Saaty, 1980). A pairwise comparison matrix was constructed, where each criterion was
compared with the other criteria, relative to its importance, on a scale from 1 to 9 as shown in
table 4. An Eigen-vector or weight was calculated for each criterion and used to derive a
consistency ratio of the pairwise comparisons. 0.25, 0.23, 0.05, 0.11, 0.11, 0.04, 0.02, and
0.19 was obtained as a weight value for the solar radiation potential, land cover, distance to
power transmission network, aspect, slope, distance to water source and population density
respectively with 6.4 percent consistency ratio. According to Saati, who developed AHP,
consistency ratio of less than 10 percent shows consistent comparison between the criteria and
it was deemed as acceptable (Saaty, 1980).

Step five was to aggregate the criteria using weighted linear combination and apply it in the
ArcGIS raster calculator. The output raster was converted to polygons and small polygons
with less than 2.5km2 were filtered out, since this is the smallest acreage required to produce

21
100 MW electricity from PV solar farms (Charabi & Gastli, 2010). Verification of results
was the last step and usually assessed by ground truth verification and sensitivity analysis.

Table 4: Pairwise Comparison Matrix

Solar Distance Distance Distance Population Weight

Criterion Land Cover Aspect Slope
Radiation to Grid to Road to Water Density (%)
Solar Radiation 1 1 7 2 3 7 9 2 25.13
Land Cover 1 1 5 3 3 7 9 1 23.17
Distance to Grid 0.14 0.20 1 0.20 0.20 2 6 0.33 5.04
Aspect 0.5 0.33 5 1 1 2 9 0.5 11.44
Slope 0.33 0.33 5 1 1 2 9 0.33 10.50
Distance to Road 0.14 0.14 0.5 0.5 0.5 1 3 0.14 3.77
Distance to Water 0.11 0.11 0.17 0.11 0.11 0.33 1 0.11 1.68
Population Density 0.5 1 3 2 3 7 9 1 19.28

Due to size of the study area, shortage of time, and unpracticality of field surveys only
sensitivity analysis was performed by weight value of the factors to assess the reliability of
the output.

22
4 Results
To select suitable sites for large-scale PV solar farms in Ethiopia, solar radiation potential of
the country should be assessed first. A solar radiation map was derived from high resolution
DEM employing ArcGIS solar radiation analysis tool, showing the monthly and annual solar
radiation potential of the country. GIS-based Multi-Criteria Analysis (GIS-MCA) was applied
to map suitable sites for large-scale PV solar farms in Ethiopia. In addition to solar radiation
potential land cover, population density, infrastructure and topographic characteristics of the
study area were taken into account properly in the site selection process. This section shows
the spatial and temporal variation of solar radiation and optimal locations for large scale solar
energy development in Ethiopia.

4.1 The spatial distribution and temporal variation of solar radiation

The daily, monthly and annual global solar radiation of Ethiopia was calculated from 30 m
resolution ASTER DEM using the ArcGIS solar radiation analysis tools. Figure 5 below
shows the annual distribution of global solar radiation of Ethiopia per square meter.
According to this analysis, solar radiation potential of Ethiopia ranges from 0.2 MWh/m2/year
in the lowlands (peripheral area) to 2.6 MWh/m2/year in the central highlands. The radiation
increases with increasing altitude above sea level. This is mainly because of a pronounced
increase of direct irradiance (Blumthaler, Ambach, & Ellinger, 1997). Due to their relative
location, size, and elevation regions of Oromia, Amhara, west Tigray and north Southern
Nations Nationalities and Peoples received the higher global solar radiation compared to
regions located in the lowlands and periphery parts of the country. Larger portion of the
country received solar radiation exceeding 1.8 MWh/m2/year. The geographical location of
Ethiopia, close to the equator, clearly plays an important role.

23
 Figure 5: Distribution map of annual global solar radiation in Ethiopia
Furthermore, global solar radiation of the country showed variation through months of the
year. Table 5 and figure 6 shows the minimum, the maximum and the mean global solar
radiation of each month in a year. November, December, January and February received the
lowest mean solar radiation, while May, July, August and September received the highest
mean global solar radiation in a year. Since Ethiopia is situated in the Northern hemisphere
the winter and the summer solstice is in December and June, respectively. In summer, due to
longer duration of sunlight and perpendicular angle of the sun rays reaching the surface in the
Northern hemisphere, solar radiation received in this hemisphere reaches its maximum. In
contrary, during winter the Northern hemispheres receive lower radiation as a result of shorter
duration of sunlight and oblique angle of the sun rays reaching the ground.

Table 4: Temporal variation of monthly global solar radiation in Ethiopia

Global solar radiation

Month January February March April May June July August September October November December
(Wh/m /month)

Min 1,930 2,148 3,646 15,348 6,961 3,286 4,026 11,533 6,624 2,376 1,912 1,858
Max
2

221,940 207,232 236,328 238,981 255,147 250,390 257,094 249,714 239,322 224,323 216,207 219,516
Mean 136,288 138,439 170,906 172,893 176,442 167,735 175,074 178,741 174,338 153,691 135,489 130,485
Std dev. 16,161 13,212 12,501 11,899 14,233 14,906 14,788 12,792 12,143 13,549 15,351 16,523

24
 Figure 6: Temporal variation of monthly global solar radiation in Ethiopia
4.2 Comparison between observed and modeled solar radiation

Correlation between recorded and modelled global monthly solar radiation values of ten
meteorological stations obtained from Actinograph and ArcGIS respectively were calculated
using Statistical Packages for Social Sciences (SPSS). Measurement unit for observed data is
cal/cm2 and for ArcGIS data it is Wh/m2. The R2 value of 0.65 shows that there is a
correlation between observed and modeled values.

Figure 7: The correlation between the observed and the modeled solar radiation values
June, July, August and September are rainy and hence cloudy months in Ethiopia. As
mentioned earlier, ArcGIS solar radiation tool do not consider the effects of clouds in the
solar radiation calculation. Therefore, as figure 8 shows in summer the global solar radiation
values calculated using ArcGIS solar radiation analysis tool was higher than solar radiation
values recorded in the observational data.

25
 Figure 8: Comparison between observed and modeled solar radiation value
4.3 GIS-based Multi-Criteria Analysis

Solar radiation potential, slope, distance to road, distance to power transmission grid,
population density, aspect, land cover and distance to water source were the selected criteria
for suitability analysis for large scale PV solar farms in Ethiopia. As mentioned earlier, in
order to create a suitability map that identifies potential sites, the input datasets needed to be
changed into a raster with a common resolution. Therefore, these criteria datasets were
converted into raster with 30 m resolution and reclassified from 1 to 5. 1, 2, 3, 4 and 5
indicates non-suitable, not suitable, marginally suitable, moderately suitable and highly
suitable, respectively (table 6). Based on related literature and previous studies the selected
criteria were defined as follows (Charabi & Gastli, 2011; Datta & Karakoti, 2010; Gastli &
Charabi, 2010a; Janke, 2010; Li, 2013; Turney & Fthenakis, 2011; Uyan, 2013).

Table 5: Suitability level per criteria for large-scale PV solar farms site selection
Suitability level
Measurement Weight
Criteria 5 4 3 2 1
unit (%)
Highly suitable Moderately suitable Marginally suitable Not suitable Non suitable
Solar radiation MWh/m2 /year 2.2 - 2.7 2.0 - 2.2 1.9 - 2.0 1.7 - 1.9 0.2 - 1.7 25.13
Slope % ≤5 5 - 10 10 - 15 15 - 50 > 50 10.50
2
Population density Persons/km ≤ 10 10 - 30 30 - 100 101 - 300 > 300 19.28
151 120 14 110 40
200 140 20 130 41
Land cover LCCS 201 150 30 162 60 23.17
202 141 180 190
210
South East North East North Water surfaces
Aspect Direction South East West North West 11.44
South West
Distance to road km ≤5 5 - 25 25 - 50 50 - 100 > 100 3.77
Distance to power km ≤ 10 10 - 50 50 - 100 100 - 250 > 250 5.04
transmission line
Distance to water km ≤5 5 - 15 15 - 50 50 - 100 > 100 1.68

26
As shown in table 6, the annual solar radiation map of Ethiopia obtained from DEM was
categorized into five classes and assigned discrete integer value for each class. Calculated
mean solar radiation of the study area was 1.9 MWh/m2/year. As a result, areas that receive
greater than or equal to 2 MWh/m2/year were considered as suitable site for large scale PV
solar farms. Solar radiation was the most important criterion for large scale PV solar farms.
Population density was the second most important criterion and calculated as number of
persons per km2. Since solar farms should avoid densely populated regions, sparsely
populated areas with less than 30 persons per km2 were taken as suitable sites for solar panel
installations. The third important factor was existing land cover. According to land cover
classification of FAO [11], twenty land cover types were identified in the study area and
categorized into five classes. Bare areas, grassland, savannas and shrub lands were
categorized as ideal land cover for large scale PV solar farms which is about one fourth of the
total area. While populated, reserved and sensitive land covers such as forest, urban areas,
water bodies and flooded areas were categorized as permanently unsuitable land cover. Large
scale PV solar farms also require flat or gentle slope surfaces facing south and hence areas
with less than 10 percent slope were considered as suitable. Moreover, distance to the
existing infrastructure and water sources were also taken into account. These factors are
important to minimize economic costs and increase the panel efficiency. Based on Euclidean
distance, five classes were created and areas closer to existing road networks, power
transmission lines and water sources considered as suitable for large-scale solar panel
installations. To avoid dust and flood risk, areas within 1 km buffer zone of road and river
networks were excluded from the final analysis. The results of the reclassification process of
the eight input datasets are shown in the following figures.

27
Figure 9: Reclassified criteria datasets used in GIS-based multi-criteria analysis

28
Using AHP a pairwise comparison matrix was created and criteria weights were calculated for
each factor by comparing two factors at a time using a scale with values from 9 to 1/9
developed by (Saaty, 1988). The reclassified input datasets were assigned a weight value
(table 4) to express the importance of each criterion to the other criteria for large-scale PV
solar farms. In order to select suitable sites for large-scale PV installations all the reclassified
input datasets were overlaid using the Weighted Overlay tool in ArcGIS. The result was a
raster layer that indicates suitability score for solar farms between 0 and 1 as shown in figure
10. Pixel values close to 0 indicate permanently unsuitable areas and contrary pixel values
near to 1 represents highly suitable site for large-scale solar panel installations.

Figure 10: Suitability map for large-scale PV solar farms in Ethiopia

Therefore, for this study, pixels having suitability score of 0.75 or above were chosen as
suitable and converted to polygons. Small polygons with areas less than 2.5 km2 were filtered
out and 195 large but compact shaped polygons with highest suitability score were mapped as
potential sites for large scale PV solar farms in Ethiopia as shown in figure 11.

According to GIS-based MCA, the eastern part of the country is relatively suitable for large-
scale solar panel installations. It has 1.9 MWh/m2/year of solar radiation on average with ideal

29
land cover for large-scale solar farms. Moreover, it is the least densely populated region
which also is a very important criterion.

Figure 11: Selected highly suitable sites for large-scale PV solar farms in Ethiopia
4.4 Electricity Generation Potential

One aim of this study was to calculate average electricity generation potential of Ethiopia
from PV systems. To estimate the annual electricity generation potential of the country from
PV solar panels the following equation, which was previously employed by (Gastli &
Charabi, 2010a) was used. The equation calculates the yearly potential based on mean solar
radiation, total suitable area, and efficiency of PV system.

EGP = SR × CA × AF × 𝜂

where:

EGP Electric power generation potential per year (GWh/year)

SR Annual solar radiation received per unit horizontal area (GWh/km2/year)

CA Calculated total area of suitable land (km2)

30
 AF The area factor, indicates what fraction of the calculated areas can be covered by
solar panels

𝜂 PV system efficiency

The estimated mean annual solar radiation for the suitable site was 1.9 TWh/km2. The total
suitable area for PV solar panel installation calculated using GIS-based MCA was close to
6,000 km2 which accounts to 0.5 percent of the total area of Ethiopia. The electricity
generation potential was computed by taking PV efficiency of five different solar cell
materials. I have obtained the efficiency (𝜂) values from the National Renewable Energy
Laboratory of USA (Gaur & Tiwari, 2013; Green, Emery, Hishikawa, Warta, & Dunlop,
2012) [10]. The calculations also assume that at least 50 percent of the total area can be
covered by solar panels. Annual potential of electricity generation from suitable sites for large
scale PV solar farms using different types of Solar Cell materials computed as follows.

Table 6: Estimated Electricity Generation Potential of Ethiopia

Efficiency Electricity Generation Potential

Solar Cell materials
(%) (GWh/year) (GW)

Mono Crystalline Silicon (Mono C-Si) 25 1,423,752 162.5

Poly Crystalline Silicon (Poly C-Si) 20.4 1,161,782 132.6

Amorphous Silicon (a-Si) 10.1 575,196 65.7

Cadmium Telluride (CdTe) 18.3 1,042,187 119

Copper Indium Gallium Diselenide (CIGS) 20.3 1,156,087 132

As it clearly shown in table 7 selected highly suitable sites could, if exploited properly,
generate more than 500 TWh electricity per annum using the lowest PV efficiency. According
to EEPCo currently Ethiopia is producing only 7.5 TWh electricity, mainly from hydropower
[14]. Therefore, it can be conclude that Ethiopia has very high potential to generate electricity
from PV solar farms to meet the local and neighboring countries demand.

31
5 Discussions
In this section constraint regarding the DEM and other datasets used in this study to calculate
solar radiations as well as to identify suitable sites for large-scale PV installations are
discussed. Moreover, the sensitivity analysis, which was carried out by altering the weight
values of each input criteria and comparisons of the results are presented as follows.

5.1 Data and factors

Monthly and annual global solar radiation of Ethiopia was calculated using Area solar
radiation tool of the ArcGIS based on a Digital Elevation Model (DEM). To this end, 175 1°
by 1° DEM tiles at 30 meter resolution cover the whole Ethiopia were downloaded from
NASA. It is the most recent and highest resolution DEM with global coverage (Aster, 2010;
Tachikawa, Hato, Kaku, & Iwasaki, 2011). Area Solar Radiation tool considers atmospheric
effects, site latitude and elevation, slope, aspect, daily and seasonal shifts of the sun angle, the
effects of shadows cast by the surrounding topography into calculation. These characteristics
can be modelled from the DEM. Therefore, the DEM is the primary input for solar radiation
calculation. The quality of calculated solar radiation output is dependent on DEM resolution
as well as parameters such as time interval and sky size resolution. Higher resolution DEM
with smaller time interval and bigger sky size results more accurate output but also increases
considerable calculation time. Therefore, to achieve reasonable result at a short processing
time the solar radiation map was calculated at 1,000 sky size with 0.5 hour interval.

GIS-based MCA was applied to select ideal locations for large-scale solar farms in Ethiopia.
In this study large-scale refers to sites for ground-mounted PV installations with a minimum
area of 2.5 km2 on which 100 MW electricity can be generated from solar energy. In addition
to solar radiation potential, various topographic, economic and environmental factors were
taken into account in the site selection process. The most relevant factors for finding suitable
sites for large-scale PV installations like solar radiation potential, land cover, land surface
steepness, slope direction, distance to road network, and distance to existing power
transmission lines, population density and distance from water features were considered.
Some of these data were initially collected by concerned government organizations in 2007 or
before. For instance, land cover and population density map of the study area represents data
for the year 2005 and 2007, respectively. Since these data potentially are out of date, current
data about population density, land cover are important to improve the results of this study.

32
5.2 Sensitivity analysis

To verify the results of the study, sensitivity analysis was carried out by altering the weight
values of each input criteria under nine scenarios as shown in table 8. The weight value for
each criterion was calculated using pair-wise comparison matrix of AHP with a consistency
ratio of less than 10 percent.

Table 7: Weight value (%) for each input criteria under nine scenarios

Scenario
Criteria A B C D E F G H I
Solar radiation 25.13 31.06 22.34 22.06 24.75 22.35 23.54 24.55 25.47
Land cover 23.17 20.31 28.38 19.86 21.51 21.39 21.78 22.28 23.41
Distance to power grid 5.04 4.99 4.97 5.16 4.95 8.84 4.05 4.91 5.10

Aspect 11.44 10.30 11.16 10.70 15.66 12.30 11.60 10.29 9.65
Slope 10.50 9.70 9.84 9.72 9.29 11.41 10.70 14.35 8.72
Distance to road 3.77 3.64 3.68 3.62 3.28 3.53 8.66 3.28 3.47

Distance to water source 1.68 1.69 1.66 1.64 1.66 1.59 1.55 1.64 4.54

Population density 19.28 18.31 17.96 27.23 18.89 18.59 18.11 18.70 19.65

Weight sum (%) 100 100 100 100 100 100 100 100 100
Consistency Ratio (%) 6.40 7.73 7.87 9.04 7.35 9.91 7.38 9.62 6.61
Total suitable area (%) 0.88 0.47 1.13 1.44 1.02 0.87 1.12 1.01 0.56

The suitability maps under each scenario were created using Weighted Sum tool of the
ArcGIS and are shown in the following figures.

33
Figure 12: Locations of suitable area (pixels) under nine scenarios

34
In order to compare the effect of changes in weight value to the selection process, pixel with
suitability score of greater than 0.75 were chosen as suitable pixel for large-scale PV
installations. Scenario A is the base scenario where suitable sites for large-scale PV solar are
selected in this study. According to the above figures, as the weight of solar radiation
(scenario B) increases, total suitable size diminishes from 0.88 percent to 0.47 percent. In
contrary, as the weight of land cover (scenario C) and population density (scenario D)
increases, total suitable size grows by 128 percent and 164 percent, respectively. In scenario E
a higher weight value were assigned to aspect than slope to test this effect. As a result,
percentage of total suitable area was slightly increased from 0.88 to 1.02. Under scenario F
and G the importance of distance to existing power grid and distance to road network were
compared to each other. As distance to existing power grid increases from 5.04% to 8.84%,
the sizes of suitable areas remain the same, whereas when distance to road network increases
from 3.77 percent to 8.66 percent, percentage of suitable area grows with an increasing rate of
0.24 percent. Under scenario I, where distance to water sources increases from 1.68 percent to
4.84 percent, size of total suitable area decreases from 0.88 percent to 0.56 percent. As it is
clearly shown from table 8 and figure 12, the sensitivity analysis of weight assignments under
nine scenarios show the geographic locations of suitable sites with highest suitability scores
remains the same with slight changes in size. In short, location of ideal sites under scenario A
is robust and largely consistent regardless of moderate changes in the different parameters.

5.3 Comparisons

In total 195 sites located in different regions of Ethiopia were selected with aggregate
electricity generation potential of 575 TWh/year from solar radiation. Geographic and
absolute locations and characteristics of all the selected sites are presented under appendix D.
Table 9 below shows the quantity and total areas of selected sites for large-scale PV
installations as well as electricity generation potential per region. Based on the criteria
presented in this study, the Somali, Oromia and Afar regions have the highest potential to
generate electricity from ground-mounted PV solar panels.

35
 Table 8: Summary of selected sites per region

Number of Total area Electricity Generation Potential

Region
Selected Sites (km2) (Twh/year)
Afar 30 1,228 118.5
Amhara 3 33 3.2
Baneshangul Gumuz 7 183 17.7
Gambella 1 24 2.3
Oromia 72 1,936 186.7
SNNP 7 102 9.8
Somali 73 2,429 234.4
Tigray 2 27 2.6
Ethiopia 195 5961.7 575.2

Figure 13: Administrative regions of Ethiopia

For comparison it is informative to compare this amount of power generation to some global
numeric. According to the Wolfram Alpha computational knowledge engine, 575 GWh/year
is equivalent to 65 GW, which is comparable to 1/35 of average global electric power
production or 1/6 of electric power of all nuclear power plants on the globe [13]. This shows
that Ethiopia has a very high potential to produce multiple times of the current annual

36
generation capacity. The total electric supply in 2012 was about 7,000 GWh [14]. Gastili and
Charabi (2010) estimated yearly electricity generation potential of Oman from 1,500 km2
highly suitable land area for PV installations as 360 GWh.

37
6 Limitations and future work
The accuracy of any GIS mapping and analysis output mainly dependent on the quality of
input data and the tool. Improved data sources and incorporating additional criteria in the
GIS-based MCA could improve the results of this study. Shortage of detailed, recent and
digital spatial and attribute data in the study area was the main constraint of the study. Due to
absence of higher resolution DEM or LiDAR imagery which cover the study area, I have used
30 m resolution DEM obtained from NASA to produce solar radiation potential map of
Ethiopia. Since the quality depends on DEM resolution, estimation from higher resolution
DEM such as 1 meter could produce highly reliable results. But this would also increase
calculation time considerably and would be unfeasible without access to a super computer
facility.

To test the accuracy of estimated solar radiation and to improve the tool by detecting sources
of errors, latest ground measurement from several sample points representing different
topography and climate recorded by an automatic solar radiation instrument such as
Pyranometers are required. Absence of recent ground measurement from various
meteorological stations forced me to use old and incomplete radiation values recorded using
Actinographs from a few meteorological stations. Hence these data were not ideal to assess
the reliability of the ArcGIS solar radiation tool.

Only solar radiation potential, population density, land cover, slope, aspect, distance to road
network, distance to power transmission network and distance to water sources were taken
into account in the GIS-based MCA for large scale PV solar farms in Ethiopia. Incorporating
additional criteria such as future land use, wind speed, dust risk and, cultural heritage sites
would improve the result obtained from the analysis and reduce risk of potential projects.
Since there is no national standard and guidelines for solar PV installation in the country, I
have defined the suitability levels of each criteria based on related studies conducted in other
parts of the world and personal knowledge based criteria. Further field surveys and feasibility
studies are necessary to verify the selected sites for large-scale PV installations.

Due to these limitations the suitability map provides a general overview of where large-scale
PV farms could be located in Ethiopia. However, the study provides a good starting point for
decision makers, governmental agencies and investors to evaluate Ethiopia as a promising
candidate for large-scale PV solar power generation.

38
7 Conclusions
This study assesses solar energy potential of Ethiopia based on DEM using GIS. The obtained
results show that there is a very high potential of solar power generation in extensive areas of
Ethiopia throughout the year. The geographical location of Ethiopia, close to the equator and
with large high altitude terrains, plays an important role. Using GIS-based multi-criteria
analysis about 195 sites with 6000 km2 area in total were selected as ideal locations for large
scale PV solar farms in Ethiopia taking into account various topographic, economic, social
and environmental factors. The factors used in this study have more of an effect at eliminating
non-suitable areas for large-scale solar farms. If the selected sites could be exploited properly,
they can generate more than 500 TWh electricity per annum using 10 percent PV efficiency,
which is many times the current annual production.

39
References
Aster, G. (2010). Validation Team: METI/ERSDAC, NASA/LPDAAC, USGS/EROS, 2009.
ASTER Global DEM Validation Summary Report.

Blumthaler, M., Ambach, W., & Ellinger, R. (1997). Increase in solar UV radiation with
altitude. Journal of photochemistry and Photobiology B: Biology, 39(2), 130-134.

Bosch, J., Batlles, F., Zarzalejo, L., & López, G. (2010). Solar resources estimation
combining digital terrain models and satellite images techniques. Renewable Energy,
35(12), 2853-2861.

Charabi, Y., & Gastli, A. (2010). GIS assessment of large CSP plant in Duqum, Oman.
Renewable and Sustainable Energy Reviews, 14(2), 835-841.

Charabi, Y., & Gastli, A. (2011). PV site suitability analysis using GIS-based spatial fuzzy
multi-criteria evaluation. Renewable Energy, 36(9), 2554-2561.

Datta, A., & Karakoti, I. (2010). Solar Resource Assessment using GIS & Remote Sensing
techniques. Paper presented at the 11th ESRI India User Conference 2010.

Dickinson, W. C., & Cheremisinoff, P. N. (1980). Solar energy technology handbook (Vol.
2): Marcel Dekker New York.

Gastli, A., & Charabi, Y. (2010a). Siting of large PV farms in Al-Batinah region of Oman.
Paper presented at the Energy Conference and Exhibition (EnergyCon), 2010 IEEE
International.

Gastli, A., & Charabi, Y. (2010b). Solar electricity prospects in Oman using GIS-based solar
radiation maps. Renewable and Sustainable Energy Reviews, 14(2), 790-797.

Gaur, A., & Tiwari, G. (2013). Performance of Photovoltaic Modules of Different Solar Cells.
Journal of Solar Energy, 2013.

Green, M. A., Emery, K., Hishikawa, Y., Warta, W., & Dunlop, E. D. (2012). Solar cell
efficiency tables (version 39). Progress in photovoltaics: research and applications,
20(1), 12-20.

Hardy, J. T. (2003). Climate Change: Causes, effects, and solutions: Wiley.

Janke, J. R. (2010). Multicriteria GIS modeling of wind and solar farms in Colorado.
Renewable Energy, 35(10), 2228-2234.

40
Kaygusuz, K. (2001). Renewable energy: power for a sustainable future. Energy, Exploration
& Exploitation, 19(6), 603-626.

Kumar, L. (2004). Reliability of GIS-based solar radiation models and their utilization in
Agro-meteorlogical stations.

Kumar, L., Skidmore, A. K., & Knowles, E. (1997). Modelling topographic variation in solar
radiation in a GIS environment. International Journal of Geographical Information
Science, 11(5), 475-497.

Li, D. (2013). Using GIS and Remote Sensing Techniques for Solar Panel Installation Site
Selection.

Malczewski, J. (1999). GIS and multicriteria decision analysis: John Wiley & Sons.

Martínez-Durbán, M., Zarzalejo, L., Bosch, J., Rosiek, S., Polo, J., & Batlles, F. (2009).
Estimation of global daily irradiation in complex topography zones using digital
elevation models and meteosat images: Comparison of the results. Energy Conversion
and Management, 50(9), 2233-2238.

Mekonnen, S. A. (2007). Solar Energy Assessment in Ethiopia: Modeling and Measurement.

Addis Ababa University.

Rich, P., Dubayah, R., Hetrick, W., & Saving, S. (1994). Using viewshed models to calculate
intercepted solar radiation: applications in ecology. American Society for
Photogrammetry and Remote Sensing Technical Papers.

Saaty, T. L. (1980). The analytic hierarchy process: planning, priority setting, resources
allocation. M cGraw-Hill.

Saaty, T. L. (1988). What is the analytic hierarchy process? : Springer.

Šúri, M., Cebecauer, T., & Skoczek, A. (2011). SolarGIS: solar data and online applications
for PV planning and performance assessment. Paper presented at the 26th European
photovoltaics solar energy conference.

Šúri, M., Huld, T., Dunlop, E. D., & Hofierka, J. (2007). Solar resource modelling for energy
applications Digital Terrain Modelling (pp. 259-273): Springer.

Šúri, M., Huld, T. A., Dunlop, E. D., & Ossenbrink, H. A. (2007). Potential of solar electricity
generation in the European Union member states and candidate countries. Solar
energy, 81(10), 1295-1305.

41
Tabik, S., Villegas, A., Zapata, E. L., & Romero, L. F. (2012). A fast GIS-tool to compute the
maximum solar energy on very large terrains. Procedia Computer Science, 9, 364-372.

Tachikawa, T., Hato, M., Kaku, M., & Iwasaki, A. (2011). Characteristics of ASTER GDEM
version 2. Paper presented at the Geoscience and Remote Sensing Symposium
(IGARSS), 2011 IEEE International.

Tadesse, M., & Belay, K. (2004). Factors influencing adoption of soil conservation measures
in Southern Ethiopia: The case of Gununo area. Journal of Agriculture and Rural
Development in the Tropics and Subtropics (JARTS), 105(1), 49-62.

Taha, R. A., & Daim, T. (2013). Multi-Criteria Applications in Renewable Energy Analysis, a
Literature Review Research and Technology Management in the Electricity Industry
(pp. 17-30): Springer.

Toman, M. T., & Jemelkova, B. (2003). Energy and economic development: an assessment of
the state of knowledge. The Energy Journal(4), 93-112.

Tovar‐Pescador, J., Pozo‐Vázquez, D., Ruiz‐Arias, J., Batlles, J., López, G., & Bosch, J.
(2006). On the use of the digital elevation model to estimate the solar radiation in
areas of complex topography. Meteorological Applications, 13(3), 279-287.

Turney, D., & Fthenakis, V. (2011). Environmental impacts from the installation and
operation of large-scale solar power plants. Renewable and Sustainable Energy
Reviews, 15(6), 3261-3270.

Uyan, M. (2013). GIS-based solar farms site selection using analytic hierarchy process (AHP)
in Karapinar region, Konya/Turkey. Renewable & Sustainable Energy Reviews, 28,
11-17. doi: DOI 10.1016/j.rser.2013.07.042

42
Internet Sources
1. Environmental System Research Institute, ArcGIS resources ArcGIS help 10.2,
http://resources.arcgis.com/en/help/main/10.2/index.html#/Area_Solar_Radiation/009z
000000t5000000/ (accessed in November, 2014)

2. Environmental System Research Institute, ArcGIS resources ArcGIS help 10.2, An

overview of the Solar Radiation tools,
http://resources.arcgis.com/en/help/main/10.2/index.html#//009z000000t4000000
(accessed on October, 2014)

3. The Federal Democratic Republic of Ethiopia, Central Statistical Agency,

http://www.csa.gov.et (accessed in March, 2013)

4. The Federal Democratic Republic of Ethiopia, Ethiopian Government Portal,

http://www.ethiopia.gov.et (accessed in February, 2013)

5. The Federal Democratic Republic of Ethiopia, Ethiopian Meteorology Agency,

http://www.ethiomet.gov.et/ (accessed in March, 2014)

6. The US Central Intelligence Agency, The World Fact Book,

https://www.cia.gov/library/publications/the-world-factbook/geos/et.html (accessed in
October, 2013)

7. The US Environmental Protection Authority,

http://www.epa.gov/climatechange/science/future.html (accessed in May, 2014)

8. The US National Aeronautics and Space Administration, Earth Observing System Data
and Information System, https://earthdata.nasa.gov/ (accessed in October, 2013)

9. The US National Aeronautics and Space Administration, Earth Observing System Data
and Information System,
http://reverb.echo.nasa.gov/reverb/#utf8=%E2%9C%93&spatial_map=satellite&spati
al_type=rectangle (accessed in September, 2014)

10. The US National Renewable Energy Laboratory, Best Research Cell Efficiencies,

http://www.nrel.gov/ncpv/images/efficiency_chart.jpg (accessed in April, 2014)

11. United Nations, Food and Agricultural Organizations, GeoNetwork,

http://www.fao.org/geonetwork/srv/en/main.home (accessed in March 2014)

43
12. United Nations, Food and Agricultural Organizations, Global Land Cover Network,
http://www.glcn.org/activities/globcover_en.jsp (accessed in March, 2014)

13. Wolfram Alpha computational knowledge engine, https://www.wolframalpha.com/

(accessed in April, 2014)

14. The Federal Democratic Republic of Ethiopia, Ethiopian Electric Power Corporation,
http://www.eepco.gov.et/flyersandmagazines.php?rm=1 (accessed in December, 2013)

15. Environmental System Research Institute, Locating Sites for Photovoltaic Solar
http://www.esri.com/news/arcuser/1010/solarsiting.html (accessed in February, 2013)

44
Appendix A: Observed Daily Global Solar Radiation of Meteorological Stations
Appendix B: GIS Models for Solar Radiation Calculation

GIS Model to create 1° latitude zone by mosaic 1° by 1° DEM tile of a particular degree
latitude

(Repeated to create 13 major zones and 11 minor DEM zones)

GIS Model to calculate solar radiation for a major zone using Area Solar Radiation Tool and
appropriate parameters

(Repeated to calculate solar radiation for each major and minor zone)
GIS Model to calculate Monthly Global Radiation of Ethiopia by mosaic all major and minor
zones’ solar radiation of a particular month

(Repeated to calculate Global Solar Radiation for each month in a Year)

GIS Model to obtain Annual Global Solar Radiation of Ethiopia by Mosaic Monthly Global
Radiation of Months in a Year
GIS Model to calculate Global Solar radiation of a Meteorological Station using Point Solar
Radiation Tool and appropriate parameters

(Repeated to calculate Global Solar Radiation for each Sample Meteorological Station)
Appendix C: GIS Models for GIS-Based Multi-Criteria Site Selection
for Large-Scale PV Installations in Ethiopia

GIS Model to Project and Reclassify Annual Solar Radiation Raster into Five Suitability
Level

GIS Model to calculate Euclidean Distance around Road, River and Power Transmission
Lines and to Project and Reclassify the output Raster into Five Suitability Level

GIS Model to Derive Aspect and Slope from Mosaic ASTER DEM of Ethiopia and to Project
and Reclassify the Aspect and Slope Raster into Five Suitability Level
GIS Model to Convert Land Cover and Population Density Polygons to Raster and to Project
and Reclassify the output Raster into Five Suitability Level

GIS Model to Create Fuzzy Membership Raster of each Criterion and to obtain Suitability
Map for large-scale PV installations in Ethiopia using Weighted Sum tool

GIS Model to convert Suitability Score Raster to Polygons

Small polygons less than 2.5 km2 were filtered out using Selection by Attributes. Finally, 195
Ideal Sites (Polygons) with aggregate capacity of 500 TWh/year were selected for large-scale
PV installations in Ethiopia.
Appendix D: Locations and Characteristics of the Selected Sites for
Large-Scale PV Installations in Ethiopia

X-Coordinate (m)

Y-Coordinate (m)

EGP (Twh/year)
2
Area-km
Woreda

Region
Zone
ID

1 536436 1477309 Hintalo Wajirat South Tigray Tigray 21.8 2.1

2 503066 1456858 Abergele Wag Himra Amhara 11.2 1.1
2 Tanqua Abergele Central Tigray Tigray 0.0
3 634762 1344975 Awura Zone 04 Afar 21.5 2.1
4 826938 1322921 Elidar Zone 01 Afar 14.3 1.4
5 792357 1327212 Elidar Zone 01 Afar 92.3 8.9
6 607448 1328821 Awura Zone 04 Afar 120.5 11.6
7 647558 1322364 Awura Zone 04 Afar 11.8 1.1
8 659415 1302222 Dubti Zone 01 Afar 14.3 1.4
9 795609 1289915 Elidar Zone 01 Afar 11.8 1.1
10 692223 1283279 Mile Zone 01 Afar 15.8 1.5
11 685610 1280218 Mile Zone 01 Afar 16.0 1.5
12 671381 1277399 Mile Zone 01 Afar 5.9 0.6
13 681384 1274595 Mile Zone 01 Afar 36.7 3.5
14 639986 1268937 Adear Zone 01 Afar 16.9 1.6
15 644446 1264739 Adear Zone 02 Afar 20.1 1.9
16 633621 1263436 Adear Zone 03 Afar 41.0 4.0
17 652465 1257195 Adear Zone 04 Afar 113.1 10.9
18 661425 1244669 Adear Zone 05 Afar 6.7 0.6
19 655243 1244008 Adear Zone 06 Afar 2.5 0.2
20 650907 1233914 Telalak Zone 05 Afar 71.1 6.9
20 Adear Zone 07 0.0
21 725008 1225961 Erer Shinile Somali 2.7 0.3
22 647790 1224731 Telalak Zone 05 Afar 21.9 2.1
23 722650 1221251 Erer Shinile Somali 26.9 2.6
24 722347 1213971 Erer Shinile Somali 12.2 1.2
25 649158 1213023 Talalak Zone 05 Afar 37.3 3.6
26 726291 1208625 Erer Shinile Somali 12.8 1.2
27 724352 1206100 Erer Shinile Somali 2.8 0.3
28 648997 1202193 Telalak Zone 05 Afar 48.2 4.6
29 890558 1197066 Ayisha Shinile Somali 11.5 1.1
30 697492 1197681 Erer Shinile Somali 12.0 1.2
31 677181 1204628 Erer Shinile Somali 72.7 7.0
32 707685 1187773 Erer Shinile Somali 23.1 2.2
33 672648 1190763 Erer Shinile Somali 60.6 5.8
 X-Coordinate (m)

Y-Coordinate (m)

EGP (Twh/year)
2
Area-km
Woreda

Region
Zone
ID

34 873065 1186178 Ayisha Shinile Somali 55.0 5.3

35 899872 1183570 Ayisha Shinile Somali 15.8 1.5
36 894081 1182583 Ayisha Shinile Somali 23.3 2.2
37 887797 1183402 Ayisha Shinile Somali 28.4 2.7
38 819719 1171309 Ayisha Shinile Somali 19.2 1.9
39 714636 1174558 Erer Shinile Somali 47.6 4.6
40 838207 1167780 Ayisha Shinile Somali 24.0 2.3
41 848279 1164871 Ayisha Shinile Somali 12.3 1.2
42 887814 1167702 Ayisha Shinile Somali 34.8 3.4
43 642063 1157424 Hadeleala Zone 05 Afar 26.8 2.6
44 736127 1156292 Erer Shinile Somali 34.8 3.4
45 821787 1156472 Ayisha Shinile Somali 32.0 3.1
46 742632 1154335 Shinile Shinile Somali 15.4 1.5
47 835908 1152384 Ayisha Shinile Somali 12.4 1.2
48 785076 1152967 Shinile Shinile Somali 23.5 2.3
49 873357 1149013 Ayisha Shinile Somali 6.1 0.6
50 832818 1157162 Ayisha Shinile Somali 62.4 6.0
51 641038 1150522 Hadeleala Zone 05 Afar 56.9 5.5
52 771181 1148298 Shinile Shinile Somali 64.0 6.2
53 688041 1140149 Erer Shinile Somali 7.5 0.7
54 704273 1140978 Erer Shinile Somali 11.4 1.1
55 843460 1136945 Ayisha Shinile Somali 12.1 1.2
55 843460 1136945 Shinile Shinile Somali 12.1 1.2
56 812674 1135912 Shinile Shinile Somali 11.6 1.1
57 823172 1135714 Shinile Shinile Somali 87.9 8.5
58 903282 1128441 Ayisha Shinile Somali 22.6 2.2
59 700483 1126070 Erer Shinile Somali 19.3 1.9
60 836810 1126114 Shinile Shinile Somali 15.8 1.5
61 696511 1120570 Erer Shinile Somali 14.4 1.4
62 817736 1122159 Shinile Shinile Somali 26.2 2.5
63 648088 1119207 Hadeleala Zone 05 Afar 21.0 2.0
64 733351 1118157 Erer Shinile Somali 27.3 2.6
65 18867 1114785 Assosa Assosa Beneshangul 31.5 3.0
66 867722 1112529 Denbel Shinile Somali 22.5 2.2
67 12903 1107944 Assosa Assosa Beneshangul 11.8 1.1
68 23737 1109620 Assosa Assosa Beneshangul 35.6 3.4
69 642213 1104789 Semurobina Gelalo Zone 05 Afar 50.2 4.8
70 10841 1102488 Assosa Assosa Beneshangul 29.3 2.8
 X-Coordinate (m)

Y-Coordinate (m)

EGP (Twh/year)
2
Area-km
Woreda

Region
Zone
ID

71 682122 1101533 Afdem Shinile Somali 91.0 8.8

72 29431 1096072 Bambasi Assosa Beneshangul 29.4 2.8
73 883682 1092090 Denbel Shinile Somali 11.0 1.1
74 674097 1091898 Afdem Shinile Somali 10.9 1.0
75 873436 1091105 Denbel Shinile Somali 36.4 3.5
76 24441 1089619 Bambasi Assosa Beneshangul 26.1 2.5
77 678460 1087839 Afdem Shinile Somali 37.6 3.6
78 947738 1089690 Awubere Jijiga Somali 105.6 10.2
79 825067 1078226 Shinile Shinile Somali 13.3 1.3
80 905566 1077799 Denbel Shinile Somali 21.8 2.1
81 33773 1075058 Bambasi Assosa Beneshangul 19.4 1.9
82 339121 1077326 Hababo Guduru Horo Gidru Wellega Oromia 33.3 3.2
83 961243 1074366 Awubere Jijiga Somali 30.9 3.0
84 936195 1072423 Awubere Jijiga Somali 27.3 2.6
85 788855 1074021 Erer Shinile Somali 49.4 4.8
85 Shinile 0.0
86 338732 1068188 Hababo Guduru Horo Gidru Wellega Oromia 14.9 1.4
87 697771 1065031 Afdem Shinile Somali 13.0 1.3
88 690667 1070166 Afdem Shinile Somali 119.1 11.5
89 624269 1060374 Dulecha Zone 03 Afar 31.7 3.1
90 337757 1060746 Gudru Horo Gidru Wellega Oromia 15.0 1.4
90 Hababo Guduru 0.0
91 955871 1062175 Awubere Jijiga Somali 214.8 20.7
92 971043 1056303 Awubere Jijiga Somali 16.6 1.6
93 713123 1053776 Afdem Shinile Somali 17.0 1.6
94 719450 1060303 Afdem Shinile Somali 137.6 13.3
95 619403 1053929 Dulecha Zone 03 Afar 99.7 9.6
96 332530 1042801 Guduru Horo Gidru Wellega Oromia 41.0 4.0
97 1000191 1030430 Kebri Beyah Jijiga Somali 11.6 1.1
98 310969 1029476 Guduru Horo Gidru Wellega Oromia 13.0 1.3
98 Jima Rare 0.0
99 614435 1039696 Dulecha Zone 03 Afar 82.2 7.9
100 606513 1019058 Awash Fentale Zone 03 Afar 14.3 1.4
101 601719 1015581 Awash Fentale Zone 03 Afar 15.5 1.5
102 1031222 1012529 Harshin Jijiga Somali 59.2 5.7
103 614645 1011450 Awash Fentale Zone 03 Afar 23.6 2.3
104 623377 1008049 Awash Fentale Zone 03 Afar 66.3 6.4
105 1019318 997213 Harshin Jijiga Somali 38.7 3.7
 X-Coordinate (m)

Y-Coordinate (m)

EGP (Twh/year)
2
Area-km
Woreda

Region
Zone
ID

106 1005885 994087 Harshin Jijiga Somali 14.4 1.4

107 1062450 989750 Harshin Jijiga Somali 12.5 1.2
108 1066128 985031 Harshin Jijiga Somali 13.0 1.3
109 1003434 984785 Harshin Jijiga Somali 23.2 2.2
110 535053 982213 Minjar Shenkora North Shewa Amhara 7.8 0.8
111 773980 977415 Gole Oda East Harerge Oromia 29.8 2.9
112 1003150 971445 Degehabur Degehabur Somali 27.6 2.7
113 554593 968343 Minjar Shenkora North Shewa Amhara 19.5 1.9
114 985616 969534 Degehabur Degehabur Somali 89.2 8.6
115 1010020 956071 Degehabur Degehabur Somali 16.7 1.6
116 986920 954504 Degehabur Degehabur Somali 45.4 4.4
117 752861 933311 Gole Oda East Harerge Oromia 11.7 1.1
118 1013021 930338 Degehabur Degehabur Somali 3.9 0.4
119 710601 871107 Legehida Bale Oromia 13.9 1.3
120 597352 866791 Amigna Arsi Oromia 31.3 3.0
121 607666 865588 Amigna Arsi Oromia 8.8 0.8
122 595219 858135 Amigna Arsi Oromia 35.2 3.4
123 729515 837681 Seweyna Bale Oromia 13.3 1.3
124 111530 822861 Bita Kefa SNNP 6.2 0.6
125 722553 822450 Seweyna Bale Oromia 18.0 1.7
126 642859 815568 Gasera Bale Oromia 5.7 0.6
127 741390 812313 Seweyna Bale Oromia 10.2 1.0
128 88104 813286 Mengesh Mejenger Gambella 24.0 2.3
129 651764 812139 Gololcha Bale Oromia 16.9 1.6
130 697855 809550 Seweyna Bale Oromia 3.1 0.3
131 651266 805285 Ginir Bale Oromia 11.4 1.1
132 704580 800345 Seweyna Bale Oromia 19.3 1.9
133 652679 796021 Ginir Bale Oromia 43.3 4.2
134 1076063 790539 Kebridehar Korahe Somali 10.2 1.0
135 551265 786874 Adaba West Arsi Oromia 28.1 2.7
136 537852 782536 Adaba West Arsi Oromia 11.7 1.1
137 726568 781940 Raitu Bale Oromia 11.1 1.1
138 601572 778509 Goba Bale Oromia 3.1 0.3
139 545005 780636 Adaba West Arsi Oromia 47.8 4.6
140 604714 775416 Goba Bale Oromia 11.4 1.1
141 1070739 776444 Kebridehar Korahe Somali 21.0 2.0
142 567114 771253 Adaba West Arsi Oromia 12.4 1.2
143 617257 775155 Goba Bale Oromia 46.5 4.5
 X-Coordinate (m)

Y-Coordinate (m)

EGP (Twh/year)
2
Area-km
Woreda

Region
Zone
ID

144 644198 772508 Goro Bale Oromia 30.8 3.0

145 595906 767388 Goba Bale Oromia 37.8 3.7
146 604978 761894 Goba Bale Oromia 17.2 1.7
147 577597 760891 Adaba West Arsi Oromia 37.8 3.6
148 802535 757662 Raitu Bale Oromia 10.7 1.0
149 602804 752336 Goba Bale Oromia 11.4 1.1
150 580043 752525 Goba Bale Oromia 42.8 4.1
151 589541 744951 Goba Bale Oromia 20.7 2.0
152 189085 663569 Selamago South Omo SNNP 15.8 1.5
153 183807 647818 Selamago South Omo SNNP 13.6 1.3
154 544168 631409 Liben Guji Oromia 26.8 2.6
155 543986 623516 Liben Guji Oromia 34.3 3.3
156 550071 612588 Liben Guji Oromia 21.7 2.1
157 547568 607333 Liben Guji Oromia 22.1 2.1
158 561235 603308 Liben Guji Oromia 114.1 11.0
159 550248 595376 Liben Guji Oromia 127.2 12.3
160 179763 588442 Gnangatom South Omo SNNP 18.7 1.8
161 170445 586177 Gnangatom South Omo SNNP 14.7 1.4
162 409745 584949 Yabelo Borena Oromia 16.6 1.6
163 552710 585214 Liben Guji Oromia 36.6 3.5
164 537096 587240 Liben Guji Oromia 50.5 4.9
165 245872 577274 Hamer South Omo SNNP 10.6 1.0
166 418283 579452 Yabelo Borena Oromia 16.3 1.6
167 623884 577507 Filtu Liben Somali 26.0 2.5
168 396343 576484 Yabelo Borena Oromia 12.0 1.2
169 425808 573195 Yabelo Borena Oromia 14.0 1.3
170 434895 568440 Yabelo Borena Oromia 14.6 1.4
171 669395 564587 Filtu Liben Somali 12.8 1.2
172 441415 563269 Yabelo Borena Oromia 20.4 2.0
173 603949 560081 Filtu Liben Somali 11.6 1.1
174 424774 552994 Yabelo Borena Oromia 12.5 1.2
175 431927 558727 Yabelo Borena Oromia 132.8 12.8
176 428523 548938 Yabelo Borena Oromia 14.3 1.4
177 423767 548525 Yabelo Borena Oromia 5.6 0.5
178 449580 549389 Arero Borena Oromia 20.9 2.0
179 479336 541791 Arero Borena Oromia 26.0 2.5
180 347517 536717 Teltele Borena Oromia 12.0 1.2
181 428347 534962 Yabelo Borena Oromia 12.4 1.2
 X-Coordinate (m)

Y-Coordinate (m)

EGP (Twh/year)
2
Area-km
Woreda

Region
Zone
ID

182 265637 535639 Hamer South Omo SNNP 22.0 2.1

183 440281 532829 Arero Borena Oromia 12.2 1.2
183 440281 532829 Yabelo Borena Oromia 12.2 1.2
184 437854 527742 Arero Borena Oromia 15.4 1.5
184 420028 529872 Yabelo Borena Oromia 10.5 1.0
185 437854 527742 Yabelo Borena Oromia 15.4 1.5
186 426299 524363 Yabelo Borena Oromia 16.9 1.6
187 255416 519456 Teltele Borena Oromia 4.0 0.4
188 246432 514044 Teltele Borena Oromia 20.3 2.0
189 448901 503338 Arero Borena Oromia 11.7 1.1
190 422129 505634 Dire Borena Oromia 46.2 4.5
191 428838 501670 Dire Borena Oromia 19.4 1.9
192 436833 499924 Arero Borena Oromia 57.4 5.5
192 436833 499924 Dire Borena Oromia 57.4 5.5
193 450656 497255 Arero Borena Oromia 28.2 2.7
194 431529 482746 Dire Borena Oromia 5.5 0.5
195 311149 477451 Teltele Borena Oromia 46.7 4.5