Energy Conversion and Management 303 (2024) 118188

Contents lists available at ScienceDirect

Energy Conversion and Management

journal homepage: www.elsevier.com/locate/enconman

Solar energy resource mapping, site suitability and techno-economic

feasibility analysis for utility scale photovoltaic power plants
in Afghanistan
JaveedUllah Hamad , Momina Ahmad , Muhammad Zeeshan *
Institute of Environmental Science and Engineering (IESE), School of Civil and Environmental Engineering (SCEE), National University of Sciences and Technology
(NUST), Islamabad, Pakistan

A R T I C L E I N F O A B S T R A C T

Keywords: The rate of electrification in Afghanistan stands at 30.2 % and is heavily dominated by fossil fuels. Besides, the
Solar resource mapping potential of solar power remains largely unexplored in the region. Situated at the heart of the solar belt, it is
Photovoltaic power essential to identify feasible locations for solar power plant installation in the country. This study integrates
Analytic Hierarch Process
validated meteorological, social and environmental parameters with geospatial and techno-economic factors to
Techno-economic feasibility
Afghanistan
evaluate the potential and suitability of photovoltaic power plants in Afghanistan. Annual average Global
Horizontal Irradiance map is generated for Afghanistan after validation of Global Horizontal Irradiance dataset
by Modern-Era Retrospective analysis for Research and Applications, version-2. Technical potential analysis is
performed for two types of materials and three tracking systems incorporating techno-economic and meteoro­
logical parameters. Furthermore, site suitability assessment is performed based on Multi-Criteria Decision
Making, Analytic Hierarchy Process integrated with Geographic Information System, considering techno-
economic, social, and environmental parameters. It is observed that the extremely suitable sites are located
near the transmission lines where high solar resources and relatively low temperatures are available, besides low
water stress. Further, a sensitivity analysis showed that the site suitability is sensitive to the criteria weights, and
each of the considered criteria significantly impacts the output. Techno-economic feasibility assessment is per­
formed using RETScreen for a 50 MW PV system at selected site. It is found that the highest annual energy
generation potential (166336 MWh/year) is attributed with poly-si REC module equipped with dual-axis tracking
system and lowest levelized cost of electricity is estimated for same module equipped with single-axis tracking
system (0.031 $/kWh). Other economic metrics also conformed to the suitability of single axis tracking systems.

1. Introduction alternative energy resources [1].

Development and utilization of solar energy has accelerated
Rapid increase in human population, and advances in industrial tremendously in the past ten years, more than any other renewable
development are increasing demand for energy consumption day by day energy (RE) technology, including wind, biomass and geothermal [4].
globally [1]. Historically, inexistence and ill implementation of envi­ This is due to the abundant and widespread availability of solar radia­
ronmental policies and disregard for the sustainability has resulted in tions over the globe and decreasing price of solar systems, resulting in
energy production relying mainly on fossil fuels i.e., oil, coal, and nat­ economies of scale [5]. Currently, solar energy makes more than 45.9 %
ural gas [2]. share in global renewable energy mix (excluding hydel energy), domi­
Besides being limited and un-replenishable resource, fossil fuel- nated largely by developed world with exceptions of India and China
based energy results in huge amount of greenhouse gases emissions, [6]. Currently, Photovoltaics (PV) and Concentrated Solar Power (CSP)
consequently causing global warming and environmental degradation, are two possible technologies for harnessing energy from sun. CSP plant
affecting humans and ecosystems, altogether [3]. These limitations harness the solar thermal energy and employ a standard steam boiler
necessitate the development of reliable, sustainable, and affordable power cycle for energy conversion, whereas PV panels directly convert

* Corresponding author.
E-mail address: mzalikhan@gmail.com (M. Zeeshan).

https://doi.org/10.1016/j.enconman.2024.118188
Received 4 August 2023; Received in revised form 22 October 2023; Accepted 7 February 2024
Available online 14 February 2024
0196-8904/© 2024 Elsevier Ltd. All rights reserved.
J. Hamad et al. Energy Conversion and Management 303 (2024) 118188

Global Horizontal Irradiance (GHI) into electricity through the photo­ Table 1
voltaic effect. PV systems are based on mature technologies and have Literature review on site suitability analysis of solar PV power plants based on
developed economies of scale globally, as they require less maintenance, AHP coupled with GIS.
have a long service life, and cause less water and air pollution, partic­ Study area/Year Evaluation criteria Method/ Ref
ularly during operational phase [7]. In 2021, the amount of energy of publication Objectives
produced by solar PV contributed about 5 % of total energy generation Afghanistan Annual average GHI, Geo spatial [17]
in the world and this percentage tends to reach 27 % in 2032 [8]. Over 2016 protected areas, land cover, exclusion,
the past decade, solar PV technology demand increased and it became rivers, distance to technical potential
transmission lines, distance to
cost effective and a competitive option for electricity harnessing in
rivers, slope
certain countries. Kabul, Direct Duration radiation, Site suitability using [1]
A big challenge for feasible site selection of PV power plants is Afghanistan Land use, Distance from Analytic Network
lacking accurate datasets, because ground data is scarce around the 2021 roads, Precipitation, Distance Process (GIS_ANP)
globe. It is particularly scarcer in developing countries, like Afghanistan from Fault, Slope, Aerosol
Optical depth, Aspect
where meteorological stations are available in big cities only [9]. As an Cyprus land use and land cover, AHP [18]
alternative, satellite and reanalysis datasets are extensively used glob­ 2016 protected area, Land
ally, which provide long-term estimations However, the accuracy of transportation, distance to
these datasets needs to be assessed before utilization due to coarse airport and distance to water
bodies,
spatial resolution and reliance on mathematical models rather than
The Regional GHI, DNI. Water bodies, AHP [19]
actual observations. Multiple studies have been conducted for the vali­ Unit of power transmission network,
dation and investigation of global reanalysis datasets like Modern-Era Rethymno land use and land cover,
Retrospective analysis for Research and Applications, version 2 2019 coastline, elevation, slope,
(MERRA-2) [10]. However, the validation of reanalysis datasets is still aspect, road infrastructure,
and most visited sites,
limited especially for the developing countries like Afghanistan [11].
Malatya, GHI, land use and land cover, AHP [3]
Multi-Criteria Decision-Making (MCDM) techniques are utilized to Turkey human settlement area,
evaluate the importance of different criteria that impact site selection. 2020 aspect, power transmission
These criteria can differ based on expert judgments, data attributes, and network, roads, fault line,
water bodies, substation, and
specific project needs. Through the amalgamation of Geographic Infor­
natural gas lines,
mation Systems (GIS) and MCDM methodologies, decision makers gain a Indonesia GHI, land use and land cover, AHP [20]
streamlined ability to select the most suitable option from a range of 2020 temperature, humidity, road
potentially conflicting alternatives. infrastructure, elevation,
Various MCDM techniques have been reported previously. For slope, aspect, human
settlements area, and
example, Decision-Making Trial and Evaluation Laboratory (DEMATEL)
electricity grid station.
method has been utilized in many studies [12], it is burdened by the India Solar radiation, elevation, Fuzzy AHP [21]
requirement for an extensive number of pairwise comparisons and lacks 2021 slope, aspect, coastline, water
a consistency measurement. In contrast, the Best-Worst Method (BWM) bodies, airport, protected
area, land use/ land cover,
has gained popularity in a short span because it necessitates fewer
transmission network, cities,
comparisons as compared to other available options. However, its roads, and power plants.
application is notably intricate due to the substantial number of pairwise Chihuahua, Land use and land cover, road The Ranking and [22]
comparisons among criteria and the need to set restrictions for solving USA network, solar radiation, AHP
nonlinear models [13]. Methods like entropy [14] and criteria signifi­ 2021 temperature, wind speed,
vapor pressure, soil texture,
cance through intercriteria correlation [2] disregard the opinions of
slope, aspect, and landforms.
decision makers and rely on mathematical techniques to determine
criteria weights based on the information within the decision-making
matrix. The Step-Wise Weight Assessment Ratio Analysis (SWARA) performance of PV systems and hence LCoE represents the combined
method, while valuable, is not as widely used as other methods such as effect of the radiation intensity and temperature. While some of the
DEMATEL, primarily due to its inability to assess the consistency of aforementioned studies [1,3] employed satellite and reanalysis GHI
acquired comparisons [15]. Furthermore, the Full Consistency Method datasets as primary suitability criterion, prior validation of the data to
(FUCOM) does not emphasize flexibility, making it less suitable for cases evaluate its reliability was generally ignored.
requiring adjustments to weight coefficients based on individual pref­ Moreover, slope and elevation have been previously considered as
erences [16]. Due to its numerous advantages, robustness and versa­ suitability criteria in most studies [1,2]to prevent landslide and high
tility, this paper employs the Analytic Hierarchy Process (AHP) construction cost for plant installation respectively. Besides, to avoid
approach. Table 1 elaborates few of the recently published studies, environmental degradation of the ecosystems, designated protected area
which have utilized GIS based MCDM, AHP to analyze viable and suit­ (based on United Nations Environment Program World Conservation
able sites for PV power plants installation. Monitoring Centre (UNEP-WCMC)) [25] is considered as a criterion for,
Analyzing the above studies, it was observed that the criteria selec­ either exclusion or site suitability, although few of the classes of pro­
tion for exclusion and site suitability is random in these studies. For tected areas allow development of PV on them. Therefore, it is critical to
example, few of the above studies have considered land cover specif­ explore the validity of a parameter before it can be used as an exclusion
ically (agriculture land, forest cover and tree cover) as exclusion criteria criterion or a site suitability parameter.
and depended on already produced annual average resource maps for Furthermore, it is also possible that a region may have suitable
estimating generation potential without incorporation of technical and meteorological, topographical, and land-use conditions, but techno-
economic constraints [17]. In the studies conducted in China [23] and economic conditions may not go along with site specific conditions,
Tanzania [24], site suitability was determined primarily considering hence it will diminish the real potential and misconstrued site selection.
solar radiation intensity. It has been found that a more comprehensive Consequently, rigorous techno-economic analysis along with site suit­
alternative for assessing site suitability, could be the Levelized Cost of ability is vital for sustainable planning and decision-making related to
Electricity (LCoE). This is because higher insolation levels often coincide PV power plant specially in developing countries [26].
with elevated ambient temperatures, which can significantly impact the From the reviewed literature, it is evident that most of the studies

2
J. Hamad et al. Energy Conversion and Management 303 (2024) 118188

have focused on site suitability analysis while neglecting the consider­ • Dependency on already produced annual average resource maps for
ation of techno-economic potential assessment. For instance, commonly establishment of generation potential without incorporation of
used equations for technical potential assessment, as employed by Sun technical and economic constraints,
et al. [23] and Feng et al. [27], do not factor in year-long thermody­ • Studies conducted for Afghanistan only focused on site suitability
namic variations related to meteorological conditions such as temper­ and electricity generation potential and lacking economic and in­
ature. This oversight can potentially contribute to higher uncertainties vestment viability indicators like LCOE, NPV, IRR and PBP utilizing
in the results. RETScreen simulating tool.
Similarly, Anwarzai et al. [17] use annual average GHI for capacity • Limitation of research for sustainable development of renewable
factor estimations, which can lead to either overly conservative or energy technologies, especially on solar PV power in Afghanistan.
optimistic assessments of actual potential. Similarly, the method adop­
ted by Jing et al. [28] might introduce uncertainty into estimates by Contributions of this study to the determined research gap include:
using modeled GHI data and the temperature coefficient without prior
validation. Given the empirical evidence of meteorological parameters • Validation of MERRA-2 reanalysis datasets (available globally) over
significantly impacting PV panel performance, it is imperative to include multiple timescales and discussion on influencing variables,
these factors while deciding harnessable potential of PV powerplants in • Utilizing validated, high spatial and temporal frequency datasets for
a region [29]. energy resource mapping and concise simulation of PV systems en­
There are multiple open-access analysis software to perform techno- ergy generation potential,
economic feasibility assessment of PV systems like System Advisor • Area screening and site suitability importance is evaluated based on
Model (SAM), Hybrid Optimization of Multiple Energy Resources distinct social and techno-economic-environmental parameters,
(HOMER), RETScreen. Among these, RETScreen expert is a resourceful • Use of meteorological variables as influencing factors in techno-
tool to assess techno-economic feasibility of solar PV power plants [30] economic feasibility assessment,
Bakos and Soursos [31] utilized RETScreen to evaluate technical and • Assessment of economic profitability based on vital metrics like NPV,
economic feasibility of a stand-alone hybrid PV/diesel system for a IRR and PBP besides LCoE only.
tourist resort in Greece. They also examined the techno-economic
feasibility of on-grid PV systems for buildings in Greece and Paros Is­ This study serves as the first of its kind, multi-objective feasibility
land [23]. Kassahun Y. Kebede employed HOMER and RETScreen to assessment of grid-connected PV power plant in the study area at a
perform the viability study of 5 MW on-grid solar PV power plant in suitable location, selected after comprehensive site suitability analysis.
Ethiopia considering economic metrics of Levelized Cost of Electricity Besides, energy resource maps are generated based on long term vali­
(LCOE), Net Present Value (NPV), Benefit Cost ratio (B-C) and Internal dated data. In addition, it can give supporting information for policies
Rate of Return (IRR) [32]. Obeng et al. used RETScreen for assessment of implementation to decision makers to attract and convince investors for
the technical and economic feasibility of a 50 MW grid connected solar PV deployment.
PV plant for Nsoatre Campus in Ghana. They assessed economic viability In the first part of this study, a concise validation of GHI and other
of the project based on the economic indicators including LCOE, NPV, meteorological data of MERRA-2 at distinct timescales (hourly and
Pay Back Period (PBP) and Return on Investment (ROI) [33]. Besides, monthly) have been performed and annual averaged GHI map is pro­
studies conducted for Afghanistan only concentrated on site suitability duced using validated reanalysis dataset of MERRA-2. Land use, topo­
and electricity generation potential and lacked economic and invest­ graphical and environmental exclusion is employed to screen out study
ment viability indicators like LCOE, NPV, IRR and PBP utilizing estab­ area for PV power plants establishment in the country. RETScreen expert
lished tools like RETScreen. software is then utilized for technical potential and LCoE assessment for
The regulatory framework for energy in Afghanistan is spread among 109 equidistant points over the study area. The LCoEs were estimated
different ministries and organizations[1]. The energy sector has for three distinct tracking systems (fixed tilt, single axis, and dual axis)
remarkably expanded in the recent past and 30.2 % of the population for mono-crystalline silicon and poly crystalline silicon REC (poly-si
has access to the grid connected electricity [17]. Currently Afghanistan REC) panels. Further, MCDM (AHP) method integrated with GIS is
mostly depends upon imported electricity and indigenously generated applied for comprehensive spatial suitability analysis of grid connected,
electricity is only one quarter of the total supply, that too, mostly fossil utility-scale PV system considering six distinct techno-economic-
fuels [34]. As per the new national power system master plan, the net environmental criteria including LCoE, land use and land cover
electricity demand would rise 6-fold by 2032 and the planned hydel and (LULC), distance from transmission lines, road infrastructure, protected
fossil fuel-based energy production cannot meet the required load of areas, and water stress areas. The sensitivity of the site suitability to
18,120 GWh [17]. Afghanistan has huge potential for solar and wind these criteria was assessed by a comprehensive sensitivity analysis.
energy generation besides the hydropower, therefore it has been sug­ Lastly, to find a fair understanding regarding the viability of the project,
gested by Afghanistan National Development Strategy (ANDS) program the economic feasibility of PV system is assessed by means of the
to consume renewable energy in rural areas. In response to this sug­ following economic indicators: LCoE, NPV, PBP and IRR of the gener­
gestion, the 10 % target of electricity production from renewable energy ated electricity.
sources by (2030) has been set by National Renewable Energy Policy
(NREP) [35]. 2. Data and methods
After detailed and careful literature review, it is concluded that the
studies on solar PV establishment in the study area are limited and thus, The methodology is summarized in Fig. 1 as a case study for
the feasibility needs to be systematically explored, in terms of technical, Afghanistan.
economic, and environmental design and performance. The research As given in the research framework (Fig. 1), annual averaged GHI
gap determined through the literature is summarized as: map is generated by using MERRA-2 dataset after a comprehensive
validation process. Further, high elevation areas and areas with limited
• Limited studies on MERRA-2 reanalysis datasets validation globally resources falling below a specified cutoff limit were excluded. After­
and specifically in the region this study focuses on, wards, LCoE is estimated for the selected equidistant point over entire
• Non-incorporation of the influence of meteorological variables in study area, which serves as a suitability criterion for a combined effect of
feasibility assessment, GHI and temperature. Later a weighted overlay tool employed within a
• Random selection of criteria and sub-criteria for site exclusion and GIS to evaluate and identify suitable sites. This process allows us to
site suitability analysis, perform a detailed techno-economic feasibility analysis for the most

3
J. Hamad et al. Energy Conversion and Management 303 (2024) 118188

Fig. 1. Framework for data validation, site suitability and techno-economic feasibility analysis for solar PV power plant are employed.

promising sites within the study area. by NASA’s Global Modelling and Assimilation Office [42]. MERRA-2 is
an updated version of MERRA and is also the first long-term global
2.1. Data sets reanalysis datasets for different meteorological variables. The dataset
covers the globe with a spatial resolution of 0.5◦ × 0.625◦ (55.65 ×
All the exclusion criteria used for site suitability, along with their 69.57 km2), and a temporal resolution of minimum one hour. The GHI
characteristics and sources used in this study are elaborated in the data provided by MERRA-2 from 2007 to 2021 was used in this study to
following subsection. generate the annual average GHI map. The data was processed using
MATLAB 2020a. Besides, data sets of temperature at 2 m elevation and
2.1.1. Land use and land cover data sets wind speed at 10 m elevation for the same time period were also used in
Land use, environmental and socioeconomic parameters are shown this study for PV techno-economic analysis.
in Table 2.

2.1.1.1. Meteorological data. MERRA-2 reanalysis data was introduced

Table 2
Land use / land cover datasets, description, and sources.
Criteria Data Format Spatial Resolution Time Period Sources References

Administrative Maps .shp – – Diva GIS [36]

Land use and Land Cover .netcdf 10 m 2020 ESRI Land Cover [37]
Water stress areas .tiff 2019 Aqueduct by Water Resources Institute [38]
Transmission Network .shp 2021 Energydata.info [39]
Roads .shp 2021 Humanitarian Open Street Map [40]
Elevation .tiff 30 m 2014 SRTM [41]
Slope .tiff 30 m derived from Elevation
Protected Areas .shp 2021 United Nations Environment Program World Conservation Monitoring Centre [25]

4
J. Hamad et al. Energy Conversion and Management 303 (2024) 118188

2.2. Input data validation • Hourly global and diffuse irradiance on a horizontal surface is
calculated for a representative day with the same daily global radi­
Before its utilization, the datasets were validated by comparing them ation as the monthly average,
with ground measurements. Unfortunately, Afghanistan has limited • Hourly global irradiance values on a tilted or tracking surface are
operational ground stations that provide GHI data and authors could calculated for each hour of the day,
only get data from only one station for years 2020–2021. Besides, four • Finally, the hourly tilted values are aggregated to calculate the
ground stations distributed over the country which provided tempera­ average daily irradiance within the plane of the PV array [46].
ture and wind speed data for the years 2010 to 2014 were also included. • Besides, solar PV energy generation is affected by PV cell heat
To get reasonable and concise result of validation, two stations are tolerance capacity (determined by temperature), cell conversion ef­
selected from neighbor country of Pakistan. This was done because ficiency, wind speed, and weather change [5].
Pakistan is located in proximity to Afghanistan, and thus the countries
are partly identical in topography and climatic conditions [43]. Besides The energy produced (Ep) by the PV array, Ep, is simply depicted in
hourly data, the validation was performed at monthly temporal reso­ equation (1).
lutions as well, since RETScreen required monthly averaged data of GHI,
Ep = Sηp Ht (1)
temperature and wind speed for analysis. For a detailed analysis, three
statistical tests for validation namely Root Mean Squared Error (RMSE),
where S is the area of the array, ηp is the overall PV array efficiency, and
Mean Bias Error (MBE), and Pearson Correlation Coefficient.
Ht is the daily total of computed tilted irradiances.
Table 3 summarizes the safety conditions, efficiency, and losses of
2.3. Solar resource mapping
mono crystalline silicon materials and inverters for selected sites.
Panels miscellaneous losses consist of losses due to the presence of
Annual average GHI map produced utilizing MERRA-2 monthly
dirt or snow on the modules, or mismatch and wiring losses. Typical
averaged data from 2007 to 2021, extracted for 198 equal-distance
values range from a few percent to 15 % [47]. In some exceptional
points at the resolution of 0.5◦ × 0.625◦ (55.65 × 69.57 km2) in the
circumstances (e.g., very harsh environment) this value could be as high
study area. The multiple year data was used to minimize the un­
as 20 % [47]. Inverter efficiency is selected as 96 % based on data from
certainties induced by the historical variations in the data. Afterwards,
U. S. Solar Photovoltaic System and Energy Storage Cost Benchmarks
Inverse Distance Weighting (IDW) interpolation [3] method was
[48].
employed in ArcMap 10.8.1 software, to convert data points into a map.
It was then resampled to a specific cell size of 0.0027◦ (300 × 300 m2)
2.5.1. Levelized cost of electricity estimation
for homogenization with other datasets.
LCoE is most important indicator of economic potential which de­
scribes the ratio of the total cost of the project during its lifetime (con­
2.4. Application of exclusion criteria
sisting of capital, periodic, feasibility study, development, and
engineering cost and O&M cost) to the electricity produced over lifetime
After consulting literature and experts’ opinion, five exclusion
of the plant. It is often expressed in $/kWh or $/MWh shown in equation
criteria covering environmental and socio-economic aspects, are
(2) [49]. It’s a robust measure to compare economic viability of distinct
considered. The threshold limit of 1400 kWh/m2/year [44] was applied
PV module and tracking systems. Financial parameters and cost data is
upon annual GHI map to screen out sites of low resources. Areas with
taken from recently published literature in the region [50], and
slope higher than 5 % were excluded as constructions costs significantly
renewable energy reports [48]. The input values of financial parameters
higher on high slope surfaces and they are also more prone to flash
are shown in Table 4.
floods [17]. Solar power plant deployment near human settlements and
its deployment on water bodies and permanent snow cover may result in Total cost during lifetime
LCOE = (2)
lowered power output, therefore they are excluded. Moreover, 100 m Total electricity produced during lifetime
buffer was applied for settlement for reducing noise and inconvenient
conditions for the residents [17]. All the data sets of considered pa­ It is to be noted that the costs and other financial parameters input for
rameters are resampled to a specific resolution of 0.0027◦ × 0.0027◦ the calculation of LCoE, as given in Table 4, were taken same throughout
(300 × 300 m2) for further applications. the study region.

2.5. Techno-economic analysis

Table 3
In this study technical analysis has been conducted for a proposed 50 Electrical specifications of solar system investigated in this study.
MW PV power plant. For the analysis, a mono crystalline silicon module Modules Unit Mono-Si JKM-250 Poly-
rated 250 W manufactured by Jinko solar and poly-si REC module rated M-60 si–REC350PE72
350 W manufactured by REC solar have been chosen due to low costs,
Manufacturer – Jinko Solar REC solar
higher efficiency, and high heat tolerance capacity. Multiple tracking Model – Mono crystalline Poly-si REC
systems including fixed tilt, single axis, and dual axis are considered and silicon
investigated for total 109 points, which were selected from criteria Maximum Power W 250 350
excluded LCoE map, scattered at equal distances of 0.5◦ × 0.625◦ (55.65 Number of units – 200,000 110,000
PV Module Efficiency % 15.27 20.7
× 69.57 km2) spatial resolution. Nominal operating cell ◦
C 45 45
RETScreen expert clean energy management software is utilized to temperature
evaluate the techno-economic potential at each selected point consid­ Temperature coefficient % ◦C 0.4 0.4
ering various meteorological and financial parameters. MERRA-2 rean­ Module area m2 1.637 2.117
Bifacial cell adjustment % 0 70
alysis data of GHI, temperature and wind speed monthly averaged data
factor
of each point were used at monthly temporal resolution. Miscellaneous losses % 5 5
The energy production of PV technology depends on GHI; the PV Inverter
panels convert the GHI directly into electricity based on photoelectric Efficiency % 96 96
effect [45]. PV energy production in each pixel is calculated based on Capacity kW 2500 2500
Miscellaneous losses % 1 1
algorithm used in RETScreen, following these steps:

5
J. Hamad et al. Energy Conversion and Management 303 (2024) 118188

Table 4
Costs and financial parameters structure for performing feasibility analysis.
Parameter Module Tracking System Value Unit References

Initial cost Poly-si REC Fixed tilt 561 $/kW [51]

Single axis 610 $/kW
Dual axis 680 $/kW
Initial cost Monocrystalline Fixed tilt 611 $/kW
Single axis 670 $/kW [48]
Dual axis 780 $/kW
Operational & Maintenance cost (O&M) Fixed tilt 14.6 $/kW
Single axis 16.92 $/kW [48]
Dual axis 17 $/kW
Periodic cost (inverter replacement) 40 $/kW/13 years [48]
Feasibility study, Development and Engineering cost 0.6 % [32]
Inflation rate 4.6 % [52]
Discount rate 12.1 % [53]
Electricity export rate (Tariff) 0.083 $/kWh [54]
Project life 25 Years
Electricity export escalation rate 0 % [53]

2.6. Site suitability analysis using analytic hierarchy process infrastructure in analysis of solar PV power plant establishment. The
roads with multiple buffers map applied at 1000, 3000, 6000, 10,000
It is essential to apply methods to identify which parameter or cri­ and >10000 m distances [3] are shown in Fig. 2d.
terion is more important than others. In this regard, AHP was utilized to
identify the locations of highest economic and environmental signifi­ 2.6.1.4. Protected areas. To keep the ecosystem safe and protect the
cance among high technical and resource potential sites [1] in this study ancient assets from destruction and danger associated with power plant
six different parameters were utilized consist to LULC, LCOE (for installation in a region, protected areas designated based on (UNEP-
screened areas), transmission lines network, road network, protected WCMC) for Afghanistan [25] were considered in solar PV power plant
area, and water stress area are selected for site suitability analysis. A site suitability analysis as depicted in Fig. 2e.
pairwise comparison matrix was formulated based on consulting liter­
atures and experts’ opinions in the field to determine weights for each 2.6.1.5. Water stress areas. Solar PV panels require water in installation
criterion [17]. The parameters are described in detail below. stage and need water for panels washing on regular basis, that’s why it is
necessary to avoid installation of power plant in water stressed area to
2.6.1. Levelized cost of electricity possible extent. The water stress area map is depicted in Fig. 2f.
LCoE is an important descriptive indicator of the overall competi­ The pair-wise matrix formulated in this study is shown in Table 5. To
tiveness of various power generating technologies. It displays the overall check the coherence of the process, the consistency ratio (CR) for esti­
cost of setting up and running a project in terms of dollars for each unit mation of the criteria weights were determined. If the CR is less than 0.1
of electricity the system generates over its lifespan [49]. In this study it shows that the process is accurate, otherwise the binary comparison
instead of GHI resource map, LCoE is considered, which includes the may have some ambiguities and should be reevaluated [7].
power output as well as costs pertinent to selected PV system. LCoE map LCoE was allotted highest weightage as it holds cumulative impact of
is depicted in Fig. 2a. GHI resource, temperature, wind speed and tracking system at a certain
point. To minimize cost and losses of electricity, proximity to trans­
2.6.1.1. Distance from transmission line. Generated electricity in the mission network was given second highest weightage. LULC has been
solar power plant is transmitted to power substations via transmission given the third highest priority. In LULC classes highest priority was
network and then distributed to the consumers. The longer the distance assigned to barren land followed by sparse vegetation and grass land
between solar power plant and power substation, the higher will be cost whereas, croplands, tree cover and shrubland were assigned lees weights
of new construction of transmission lines, rendering the project to reduce potential conflicts among land use for power plant establish­
economically less practicable. Therefore, it is necessary to select a ment and other reasons like agricultural or structural expansion, as well
location for PV power plant installation near the transmission network as to prevent societal concerns and dangers to ecosystems [24]. After­
or transmission lines. The transmission lines map for Afghanistan is wards, distance from road infrastructure was given the fourth highest
shown in Fig. 2b. weight whereas to ensure the least possible harm to biodiversity and
other important environmental sites, protected areas were given least
2.6.1.2. Land use land cover. LULC is an environmental criterion, weightage. Moreover, annual water stress maps were used for catego­
derived from Global LULC with Sentianel-2 which was reclassified into rizing regions based on water availability since water scarcity can
six classes: agriculture land (12.4 %), tree cover (1.3 %), sparse vege­ impinge on prospects of powerplant development. Extreme water stress
tation (6.7 %), shrub land (3.6 %), grass land (37.4 %) and baren land areas were least prioritized since in this study, panels’ cleaning through
(38.6 %). For site suitability analysis of solar power plant land avail­ water was considered [55]. But overall water requirement of PV systems
ability is an important part. It is very crucial to avoid the installation of is low, thus the total weight of the factor in holistic site suitability was
solar power plants in regions where agricultural land availability is low. Weights for each of these criteria were assigned through a rigorous
meagre. Therefore, for sustainable development, solar power plants process, involving consultations with experts and a thorough review of
should be installed keeping the sustainability of agricultural sector in the literature [56].
view. LULC map is shown in Fig. 2c. Further, raster maps for all input criteria were resampled to a specific
spatial resolution of (0.0027◦ × 0.0027◦ (300 × 300) m2 and reclassi­
2.6.1.3. Distance from roads. Deployment of PV electricity power plant fied on a scale of 1 to 5, where 1 represents the least suitable class and 5
may need a huge amount of transport to the selected area. Construction denotes most suitable class. Afterwards, weighted overlay tool in ArcGIS
of new roads will generate much new costs in the region if there is no 10.8.2 [36] was utilized to assess extremely suitable sites using equation
proper transport network. Therefore, it is essential to incorporate roads (3).

6
J. Hamad et al. Energy Conversion and Management 303 (2024) 118188

Fig. 2. Input parameters utilized in PV site suitability analysis using AHP (a) Levelized cost of electricity, (b) Distance from transmission lines, (c) Land use/land
cover, (d) Road network, (e) Protected areas, and (f) Water stress areas.

∑
n ’CWi’ stands for the criterion weight (i.e., the percentage of influence) of
s = (CW i × SV i ) (3) raster ’i,’ and ’SVi’ indicates the scale value of raster ’i’ for that specific
i=1
pixel.
In this context, ’S’ represents the suitability score assigned to each pixel,

7
J. Hamad et al. Energy Conversion and Management 303 (2024) 118188

Table 5
Pair-wise Comparison Matrix.
Criteria Water Stress Areas Protected Areas Distance to road LULC* Distance to TL* LCOE*

Water Stress Areas 1.00 0.33 0.20 0.20 0.14 0.11

Protected Area 3.00 1.00 0.33 0.20 0.20 0.14
Distance from Roads 5.00 3.00 1.00 1.00 0.33 0.20
LULC 5.00 5.00 1.00 1.00 0.33 0.20
Distance from Transmission lines 7.00 5.00 3.00 3.00 1.00 0.33
LCOE 9.00 7.00 5.00 5.00 3.00 1.00
column sums 30.00 21.33 10.53 10.40 5.01 1.99
*
LULC refers to land use land cover, TL to transmission lines, and LCOE to levelized cost of electricity.

2.7. Sensitivity analysis IRR of a project is equal to or higher than the company’s required rate of
return, then the deployment of the project is highly favorable [57]. On
Followed by site suitability analysis, a sensitivity analysis (SA) was the other hand, PBP is determined as the length of time required for
carried out. Sensitivity analysis is the process of examining the variation investment to recover its initial cost. If the PBP is less than useful life of
in the outcomes by varying the input parameters systematically (indi­ the project, it shows that the project is economical viable [58]. It is to be
vidually or jointly) within a certain level of interest [21]. Initially, all six noted that while designing the PC systems for technoeconomic evalua­
criteria considered in this study were grouped into three different cat­ tion, the components were chosen from database offered by RETScreen.
egories i.e., technical aspect (LCoE), economic aspect (distance to
transmission network, and distance to roads infrastructures) and socio- 3. Results and discussion
environment aspect (land cover, water stressed areas and protected
areas) and subsequently four SA scenarios were generated which 3.1. Re-analysis data validation
included: technical focused scenario (SA1), economic focused scenario
(SA2), Socio-environment focused scenario (SA3) and scenario with The statistical validity of the MERRA-2 reanalysis dataset is
equal-weights represented as (SA4). The higher percentage of extremely explained, along with the explanations for the observed bias and cor­
suitable classes indicated the higher influence of the respective category. relation. Table 6 summarizes the validation results.

3.1.1. Global Horizontal Irradiance

2.8. Techno-economic feasibility analysis for selected site A good correlation of > 0.95 for all station between MERRA-2
reanalysis and ground GHI datasets is observed. In this study sporadic
Feasibility analysis and investigation of grid-connected PV system overestimation and underestimation of MBE and RMSE have been re­
has a critical role for policy and decision makers, government, and in­ ported which reveal huge differences between reanalysis and ground
vestors to find a fair understanding regarding the viability of the project. station data. It may be due to overestimation of aerosol factorization in
In this regard, the economic feasibility of two different PV modules poly- retrieval algorithm and due to GHI’s greater ability to be affected by
si REC and mono crystalline silicon equipped with three different humidity, cloudiness, and particles in the atmosphere. Since irradiation
tracking systems are assessed by means of the following economic in­ is rarely geographically uniform, the large error (RMSE) attributed to
dicators: LCoE, NPV, PBP and IRR of the generated electricity. LCoE is each data source and timescale may be the result of the gridded dataset’s
commonly used to compare various PV modules based on its total cost different pixel resolutions and the subsequent comparison to data from a
and electricity produced over lifetime of the plant [49]. NPV is the single point source [26].
discounted sum of the income from selling the produced electricity net
of all costs affiliated with the electricity transmission system. A positive 3.1.2. Ambient temperature
NPV indicates that the project is economically feasible [49]. IRR is an MERRA-2 hourly and monthly temperature strongly correlates with
interest rate which makes the NPV of all cash flows zero. Precisely, if the

Table 6
Results of statistical tests for meteorological parameters dataset validation between reanalysis and ground datasets at different timestamps.
Variables Time stamps Kabul Jalalabad Kandahar Herat Balkh Peshawar Quetta

R
Ambient Temperature Hourly 0.95 0.97 1.0 1.0
Monthly 0.97 0.98 0.95 0.97 0.96 0.7 1.0
GHI Hourly 0.95 0.95 0.97
Monthly 0.92 0.96
Wind Speed Hourly 0.8 0.9
Monthly 0.21 0.47 0.6 0.1
MBE (%)
Ambient Temperature Hourly 5.62 1.02 10.3 10.5
Monthly − 3.64 6.21 1.48 5.72 3.66 10.4 10.3
GHI Hourly − 11.64 44.68 16.7
Monthly –22.2 − 7.12
Wind Speed Hourly − 30.8 − 5.3
Monthly − 0.04 0.39 − 30.7 − 5.0
RMSE (%)
Ambient Temperature Hourly 3.4 2.5 16.1 17
Monthly 4.4 6.53 3.44 2.142 2.783 11.8 10.9
GHI Hourly 113 86.04 83.2
Monthly 28.7 12.3
Wind Speed Hourly 50.1 58.7
Monthly 2.38 1.88 33.3 18.6

8
J. Hamad et al. Energy Conversion and Management 303 (2024) 118188

ground temperature data. The R value is calculated to be higher than 0.9 and ’excellent’ radiation levels for photovoltaic (PV) systems, as defined
for almost all selected locations. Studies have claimed that this may by NREL [63]. In fact, the entire study area exhibits a GHI greater than
possibly be due to removal of fixed sub-surface temperature and 1900 kWh/m2/year. Temporally from April to October highest solar
rendering of power conduction via the surface in MERRA-2 algorithm radiation is received due to high solar angle and clearer atmosphere
[59]. However high values of MBE and RMSE are reported for Quetta whereas it reduces in the following and preceding months. National
and Peshawar whereas, in case of other locations observational and Renewable Energy Laboratory (NREL) also produced annual average
MERRA-2 values are not similar but very close to each other. The GHI map for Afghanistan for the years 1999–2018 which depicted
occurrence of high MBE and overestimation might be due to error in relatively lower GHI ranges (1461 – 2191) kWh/m2/year as compared to
measurement and data scarcity problem of hilly regions due to dispersed this study [34].
network, cold climate, and topography heterogeneity [60].

3.3. Levelized cost of electricity

3.1.3. Wind speed
‘R’ value ranges from 0.21 to 0.90 for all sites at monthly timescale.
Previously identified regions with abundant solar resources, pri­
The correlation differing remarkably from one location to another. High
proportionality is usually found between wind speeds and correlation marily in the southern and southwestern provinces of the country ach­
ieved elevated LCoE values as energy generation potential was impacted
coefficient in sites having higher wind speeds [61]. Mainly, sites with
simple topography and at lower altitude depict the highest correlation due to high ambient temperatures in those areas. Conversely, the central
region, especially districts situated at higher altitudes, emerged as
because of terrain surface feature utilized in retrieval algorithm. Overall,
underestimation and overestimation are reported for sites beyond promising areas for solar energy projects. Here, the combination of
consistently cooler temperatures and year-round winds creates favor­
Afghanistan boundary might be due to mountainous topography of sites
which constantly affects the drag coefficient, whereas low bias is able conditions for achieving the lowest LCoE values among the
viable locations.
observed for all sites in study area therefore, it is highly suggested for
effectively use in the study [62]. Moreover, it is found more than half of the study area achieved LCoE
values lower than global weighted average (0.057 $/kWh) for the year
2021 [64]. The estimated values for LCoE in Afghanistan are much
3.2. Solar resource mapping lower than solar power plants installed in year 2018 – 2022 in other
countries like Pakistan [26], Iran [49], Ghana [33] and Saudi Arabia
Annual average GHI map for Afghanistan, for the years 2007–2021, [65]. Overall, the estimated LCoE value is lower than country’s elec­
is depicted in Fig. 3. It is found that high resources prevail in the south tricity tariff (0.083 $/kWh) [54], which shows that the addition of solar
and south-eastern and south-western provinces of the country. PV energy generation systems to country energy mix is favorable.
Approximately 88 % of the land boasts an annual GHI value exceeding Among the three considered systems, the lowest LCoE value is
2000 kWh/m2/year, with approximately 62 % of it surpassing 2200 associated with the single-axis tracking system. This is primarily due to
kWh/m2/year. These levels are classified respectively as ’very good’ its cost-effective O&M compared to the dual-axis system, combined with

Fig. 3. Afghanistan’s annual average Global Horizontal Irradiance map.

9
J. Hamad et al. Energy Conversion and Management 303 (2024) 118188

its superior tracking capabilities in comparison to the fixed tilt system. Table 7
Consequently, the LCoE map for the single-axis tracking system is cho­ Final weights for main and sub criteria used in AHP.
sen as the primary criterion for site suitability analysis, given its Criteria Unit Final Sub-Criteria Local
compelling economic and technical advantages. Weights Weights
LCOE maps for fixed tilt, single axis and dual axis tracking systems LCOE $/kWh 0.4476 0.036–0.047 0.5028
are shown in Fig. 4 a, b, and c respectively. 0.047–0.055 0.2602
0.055–0.062 0.1344
0.062–0.070 0.0678
3.4. Suitable sites
0.070–0.075 0.0348
Distance from road km 0.1109 0.1 to 1 0.5028
The final weights for criteria and sub-criteria were obtained using 1 to 3 0.2602
the AHP method, incorporating expert judgments gathered to interviews 3 to 6 0.1344
and are given in Table 7. Initially, the techno-economic, social, and 6 to 10 0.0678
0.0348
environmental aspects are compared pairwise to determine their rela­
>10
LULC 0.1266 Barren Land 0.3710
tive importance. Sparse Vegetation 0.2788
The site suitability output map is illustrated in Fig. 5. The ’extremely Grassland 0.1995
suitable’ class signifies the prime locations for utility-scale PV power Cropland 0.0819
Tree Cover 0.0439
plant installation. These sites were chosen to be further assessed for
Shrub land 0.0249
socio-environmental viability, as given in the next section. Grass land Distance from km 0.2347 0 to 10 0.5028
and barren land collectively, constituting around 76 % of total land area, Transmission
comprises mainly of extremely suitable class. However, many sites, Lines
although having lower LCoE, are not identified as ’extremely suitable’ 10 to 25 0.2602
25 to 50 0.1344
due to high slope, distance from transmission lines and roads in­
5 to 100 0.0678
frastructures, and water stress is extremely high (>80 %). 100–150 0.0348
Moreover, approximately 26 % viable areas fall in marginally suit­ Water Stress Areas % 0.0286 Low (<10%) 0.5028
able categories primarily due to lack of access to transmission lines and Low-Medium 0.2602
(10–20%)
road infrastructures and because of high estimated values of LCoE, even
Medium-High 0.1344
though huge solar resource is available. Eastern, southern, south- (20–40%)
western, and central parts of the country lie in moderately suitable High (40–80%) 0.0678
category due to distance from transmission lines and crop land despite Extremely High 0.0348
having high GHI ranges and roads connectivity. Approximately 3.5 % (>80%)
Protected Areas 0.0500 VI: PA with 0.4909
area is identified as extremely suitable for PV power plant installation.
sustainable use of
This category encompasses roughly 30 districts scattered across natural resources
the entire country. V: Protected 0.2913
An ’extremely suitable’ site located in the Bamyan province is Landscape
IV: Habitat 0.1507
selected for further techno-economic feasibility analysis.
Management
II: National Park 0.0670
3.5. Sensitivity analysis

Final maps produced for all four-sensitivity analysis scenarios are of extremely suitable sites would increase most noticeably in final
shown in Fig. S1–S4, SI. The results revealed that the site suitability is suitability map. In equal weight scenario defined as SA4, a significant
sensitive to the criteria weights, and each of the considered criteria decrease (from 3.45 % to 0.62 %) for the “extremely suitable” sites
significantly impacts the output. By comparing the results of each of the shows that the results are quite sensitive to the criterion weights. In SA3
SA scenario with the base scenario, it was observed that technical as­ a significant decrease occurred in “extremely suitable class” from (3.45
pects named as SA1 criterion (LCoE) contribute significantly to the in­ % to 1.95. Moreover, the least changes in the “extremely suitable” sites
crease of the “extremely suitable” sites (from 3.45 % to 6.74 %). If this (from 3.45 % to 2.2 %) were observed for SA2 as compared to base
criterion is given high importance in site suitability analysis, percentage scenario. This can possibly be attributed to the general lack of road

Fig. 4. Levelized Cost of Electricity trends assessed for different tracking systems across Afghanistan. *Dots in the maps represent the location of sites for which LCoE
is estimated.

10
J. Hamad et al. Energy Conversion and Management 303 (2024) 118188

Fig. 5. Site suitability map of utility-scale photovoltaic power plant installation in Afghanistan.

infrastructures and transmission networks in the overall study area. potential (166336 MWh/year) is associated with poly-si REC PV module
(having high efficiency as compared to mono crystalline silicon)
3.6. Electricity generation potential equipped with dual axis tracking system.

Fig. 6 shows the annual potential electricity generation for a 50 MW 3.7. Economic feasibility metrics
PV power plant for all three tracking systems and two different PV
materials mono crystalline silicone and poly-si REC at the selected Table 8 summarizes the results of economic feasibility indicators of
location i.e., Bamyan province. It is found that the use of tracking sys­ the 50 MW PV power system for two types of PV materials mono crys­
tems, as opposed to fixed tilt options, significantly increases electricity talline silicone and poly-si REC equipped with all three tracking systems
potential. This improvement arises from the ability of tracking systems at selected site. The calculated LCoE values for each kind of PV modules
to maximize the quantity of absorbed GHI by PV modules by minimizing equipped with fixed tilt single axis and dual axis tracking system are
the angle of incidence between the surface of PV modules and incoming lower than global weighted average of 0.057 $/kWh for the year 2021
solar radiation throughout the day. The highest electricity generation [64], and much lower than the country’s electricity tariff (0.083 $/kWh)

Fig. 6. Electricity generation potential for a 50 MW PV power plant with different tracking systems at selected location (Bamyan Province).

11
J. Hamad et al. Energy Conversion and Management 303 (2024) 118188

Table 8 • Around 3.5 % of viable areas are found to be extremely suitable for
Economic feasibility metrics of 50 MW PV system for selected site. solar PV power plant installation, based on site-specific techno-eco­
Module Tracking Economic feasibility metrics nomic parameters.
system • Annual energy generation of 168,791 MWh can be achieved from 50
NPV (USD) IRR PBP LCoE
(%) (years) (USD) MW power plant using poly-si REC PV module equipped with dual
axis tracking system deployed at selected site (Bamyan Province).
Mono Fixed 33,846,858 27.7 3.5 0.045
crystalline Single axis 57,326,651 35.7 2.8 0.036
• It was found that all considered configurations of the PV power
silicon Dual axis 55,706,642 31.9 3.1 0.039 systems very economically feasible as values of considered economic
Poly crystalline Fixed 35,391,130 29.3 3.4 0.42 metrics were greater than the widely accepted threshold values.
silicon REC Single axis 60,389,488 37.8 2.5 0.031 • Dual axis tracking system offer highest annual energy generation
Dual axis 58,709,890 33.1 2.9 0.36
potential but LCoE values of this system is higher as compared to
single axis tracking systems due to high O&M costs.
[54]. Furthermore, other economic parameters, IRR, NPV and PBP with • Likewise, single axis tracking system is assessed to be the most
calculated values support financial viability of the proposed power economical option for the selected site because of highest values of
plant. All PV power systems at the selected site is promising due to Net Present Value, and Internal Rate of Return, and lowest values of
reasonable estimated values of the economic metrics. Besides, among Pay Back Period and LCoE as compared to fixed tilt and dual axis
different PV materials and tracking systems, use of single axis tracking tracking system systems.
poly-si REC PV system appears to be the most economical option for the
selected site because of highest values of NPV, and IRR, in addition to Overall, huge solar energy resources prevail in the country with
the lowest values of PBP and LCoE compared to fixed tilt and dual axis recent technological advances and substantial cost reductions in the PV
tracking systems, as it absorb more irradiance, reduce surface recom­ sector. Besides, new supportive policies developed by Afghanistan’s
bination and less reflection due to having SiNx as compared to mono ministry of energy and water, development of grid-connected PV power
crystalline silicone [51]. It is concluded that the investment in power plants in central and southern regions of Afghanistan show a huge po­
plant using poly-si REC PV module equipped with single axis tracking tential and abundant market opportunities for local and foreign
system in selected site is economically favorable and adequately investors.
attractive for commercial investors based on these economic feasibility It is recommended that the present analysis could be expanded to
indicators. encompass PV and other renewable energy technologies. This expansion
would involve taking into account the alterations in modules and
4. Conclusion factoring in the influence of additional meteorological parameters,
notably cloud cover.
In this study, Modern-Era Retrospective analysis for Research and Furthermore, Enhancing the proposed model with additional deci­
Applications, version-2 (MERRA-2) re-analysis datasets of Global Hori­ sion criteria, like population growth and electricity demand, could
zontal Irradiance (GHI) and other meteorological variables (tempera­ provide valuable insights. Conversely, the land property is a pivotal
ture and wind speed) were validated against ground measurements at consideration in the land acquisition process. Thus, it is essential to
different timescales. Afterwards annual average GHI resource map was identify all suitable areas for solar plant installation, regardless of
produced. In order to undertake a concise technical resource assessment whether they are privately or publicly owned. Public or government-
of photovoltaic (PV) power plants, the study area was screened out owned land may be favored due to the potentially lower land acquisi­
based on solar resource capacity, topography, and land use classes. tion costs associated with such areas.
Furthermore, Multi-Criteria Decision Making, Analytic Hierarchy Nonetheless, it’s important to acknowledge, potential uncertainties
Process method integrated with Geographic Information System was in this study arising from the use of reanalysis timeseries data for
applied for spatial suitability analysis of utility-scale PV system meteorological parameters. To enhance accuracy, it’s advisable to
considering techno-economic and environmental aspects. Following incorporate ground observations to further reduce these uncertainties.
conclusions can be drawn from the study:
CRediT authorship contribution statement
• MERRA-2 datasets of GHI and temperature strongly correlate (R >
95 %) with ground data whereas wind speed shows weak correlation. JaveedUllah Hamad: Conceptualization, Methodology, Software,
• High bias is observed for GHI data by MERRA-2 due to its ability to Investigation, Writing – Original Draft, Visualization. Momina Ahmad:
be influenced by cloudiness, humidity, and aerosols in atmosphere. A Conceptualization, Methodology, Software, Formal analysis, Investiga­
significant error assigned to each data source is also attributed to the tion. M. Zeeshan: Conceptualization, Resources, Writing – Review &
fact that ground observations represented a point, whereas the Editing, Supervision, Project administration.
reanalysis data represented area (of different sizes, depending on the
resolution of data set). Declaration of competing interest
• In general, 100 % of land area has GHI greater than 1900 kWh/m2/
year, which produces huge deployment potential of solar energy in The authors declare that they have no known competing financial
the study area. Temporally, from April to October, the highest solar interests or personal relationships that could have appeared to influence
radiation is received due to high solar angle whereas it reduces on the work reported in this paper.
other months.
• More than half of the study area was found to have Levelized Cost of Data availability
Electricity (LCoE) values lower than global weighted average (0.057
$/kWh) for the year 2021, and less than country’s residential elec­ Data will be made available on request.
tricity tariff (0.083 $/kWh).
• Eastern, southern, south-western, and central parts of the country lie Acknowledgements
in moderately suitable category due to distance from transmission
lines and crop lands in spite of the fact that they have high GHI The authors would like to thank National University of Science and
values. Technology (NUST), Pakistan for providing research facilities and
funding to perform this study. Furthermore, the authors appreciate the

12
J. Hamad et al. Energy Conversion and Management 303 (2024) 118188

Afghan Ministry of Energy and Water (MEW), Alternate Energy Devel­ [23] Sun L, et al. A GIS-based multi-criteria decision making method for the potential
assessment and suitable sites selection of PV and CSP plants. Resour Conserv
opment Board (AEDB), Pakistan, NASA’s Global Modeling and Assimi­
Recycl 2021;168:105306.
lation office, and National Renewable Energy Laboratory (NREL) for [24] Aly A, Jensen SS, Pedersen AB. Solar power potential of Tanzania: Identifying CSP
providing meteorological data. Moreover, land use datasets provided by and PV hot spots through a GIS multicriteria decision making analysis. Renew
ESRI, United Nation Environment Program and Open Street Map are Energy 2017;113:159–75.
[25] UNEP-WCMC I, protected planet: the world database on protected areas (WDPA).
highly acknowledged. UNEP-WCMC, IUCN, Cambridge, UK. [Online] Available from: http://www.
protectedplanetnet [Accessed 22 March 2022].
[26] Ahmad M, Zeeshan M. Validation of weather reanalysis datasets and geospatial and
Appendix A. Supplementary data
techno-economic viability and potential assessment of concentrated solar power
plants. Energy Convers Manag Mar. 2022;256:115366. https://doi.org/10.1016/J.
Supplementary data to this article can be found online at https://doi. ENCONMAN.2022.115366.
[27] Feng Y, et al. High-resolution assessment of solar radiation and energy potential in
org/10.1016/j.enconman.2024.118188.
China. Energy Convers Manag 2021;240:114265.
[28] Jing J, Zhou Y, Wang L, Liu Y, Wang D. The spatial distribution of China’s solar
References energy resources and the optimum tilt angle and power generation potential of PV
systems. Energy Convers Manag 2023;283:116912.
[29] Ziane A, et al. Photovoltaic output power performance assessment and forecasting:
[1] Naseri M, Hussaini MS, Iqbal MW, Jawadi HA, Puya M. Spatial modeling of solar
Impact of meteorological variables. Sol Energy 2021;220:745–57.
photovoltaic power plant in Kabul, Afghanistan. J Mt Sci 2021;18(12):3291–305.
[30] Mehmood A, Shaikh FA, Waqas A. Modeling of the solar photovoltaic systems to
https://doi.org/10.1007/s11629-021-7035-5.
fulfill the energy demand of the domestic sector of Pakistan using RETSCREEN
[2] Günen MA. A comprehensive framework based on GIS-AHP for the installation of
software. In: 2014 international conference and utility exhibition on green energy
solar PV farms in Kahramanmaraş, Turkey. Renew Energy Nov. 2021;178:212–25.
for sustainable development (ICUE), IEEE, 2014, pp. 1–7.
https://doi.org/10.1016/J.RENENE.2021.06.078.
[31] Bakos GC, Soursos M. Technical feasibility and economic viability of a grid-
[3] Colak HE, Memisoglu T, Gercek Y. Optimal site selection for solar photovoltaic
connected PV installation for low cost electricity production. Energy Build 2002;34
(PV) power plants using GIS and AHP: A case study of Malatya Province, Turkey.
(7):753–8. https://doi.org/10.1016/S0378-7788(01)00142-6.
Renew Energy Apr. 2020;149:565–76. https://doi.org/10.1016/J.
[32] Kebede KY. Viability study of grid-connected solar PV system in Ethiopia.
RENENE.2019.12.078.
Sustainable Energy Technol Assess 2015;10:63–70. https://doi.org/10.1016/j.
[4] Guaita-Pradas I, Marques-Perez I, Gallego A, Segura B. Analyzing territory for the
seta.2015.02.003.
sustainable development of solar photovoltaic power using GIS databases. Environ
[33] Obeng M, Gyam S, Sarfo N, Kabo-bah AT. “Technical and economic feasibility of a
Monit Assess Dec. 2019;191(12). https://doi.org/10.1007/s10661-019-7871-8.
50 MW grid-connected solar PV at UENR Nsoatre Campus. vol. 247, 2020, doi:
[5] Sindhu S, Nehra V, Luthra S. Investigation of feasibility study of solar farms
10.1016/j.jclepro.2019.119159.
deployment using hybrid AHP-TOPSIS analysis: Case study of India. Renew Sustain
[34] Mehrad AT. Assessment of solar energy potential and development in Afghanistan.
Energy Rev Jun. 2017;73:496–511. https://doi.org/10.1016/J.RSER.2017.01.135.
E3S Web of Conferences EDP Sciences Feb. 2021. https://doi.org/10.1051/
[6] REN21, RENEWABLES 2021 GLOBAL STATUS REPORT. 2021: Paris: REN21
e3sconf/202123900012.
Secretariat.
[35] ANREP: Afghanistan national renewable energy policy, Ministry of Energy and
[7] Elboshy B, Alwetaishi M, Aly RMH, Zalhaf AS. A suitability mapping for the PV
Water, Afghanistan, Final Draft; November 2015.
solar farms in Egypt based on GIS-AHP to optimize multi-criteria feasibility. Ain
[36] ESRI. ArcGIS Desktop: Release 10.8.2 2021: Redlands, CA: Environmental Systems
Shams Eng J 2022;13(3):101618. https://doi.org/10.1016/j.asej.2021.10.013.
Research Institute.[online] Available from: https://www.esri.com/en-us/arcgis/
[8] REN21, RENEWABLES 2022 GLOBAL STATUS REPORT. 2022: Paris: REN21
products/arcgis-desktop/overview, Accessed 20th Jun 2023.
Secretariat.
[37] Kontgis Karra et al. Global land use/land cover with Sentinel-2 and deep learning,
[9] S. Jamshidi, K. Pourhossein, and M. Asadi, “Size estimation of wind/solar hybrid
2020. [Online] Available from: https://livingatlas.arcgis.com/landcover/
renewable energy systems without detailed wind and irradiation data: A feasibility
[Accessed 25 Mar 2022].
study,” Energy Convers Manag, vol. 234, no. October 2020, p. 113905, 2021, doi:
[38] R. Hofste et al., Aqueduct 3.0: Updated Decision-Relevant Global Water Risk
10.1016/j.enconman.2021.113905.
Indicators, World Resources Institute, 2019. [Online] Available from: https://www.
[10] A. Khatibi and S. Krauter, “Validation and performance of satellite meteorological
wri.org/applications/aqueduct/water-risk-atlas [Accessed 15 Mar 2022].
dataset merra-2 for solar and wind applications,” Energies (Basel), vol. 14, no. 4,
[39] Krokovelos A. Afghanistan Transmission network, 2020. [Online] Available from:
Feb. 2021, doi: 10.3390/en14040882.
https://energydata.info/ [Accessed 23 Mar 2022].
[11] Yang D, Bright JM. Worldwide validation of 8 satellite-derived and reanalysis solar
[40] OpenStreetMap. Afghanistan Road Network (main road), 2021. [Online] Available
radiation products: A preliminary evaluation and overall metrics for hourly data
from: https://geonode.wfp.org/layers/geonode:afg_trs_roads_osm [Accessed 18
over 27 years. Sol Energy Nov. 2020;210:3–19. https://doi.org/10.1016/J.
Mar 2022].
SOLENER.2020.04.016.
[41] National Renewable Energy Laboratory. Elevation; 2020. [Online] Available from:
[12] Parezanović T, Petrović M, Bojković N, Pamučar D. One approach to evaluate the
http://www.nrel.gov/international/ra_afghanistan.html [Accessed 17 May 2022].
influence of engineering characteristics in QFD method. European Journal of
[42] Gelaro R, et al. The modern-era retrospective analysis for research and
Industrial Engineering 2019;13(3):299–331.
applications, version 2 (MERRA-2). J Clim 2017;30(14):5419–54. https://doi.org/
[13] Žižović M, Pamucar D. New model for determining criteria weights: Level Based
10.1175/JCLI-D-16-0758.1.
Weight Assessment (LBWA) model. Decision Making: Appl Manage Eng 2019;2(2):
[43] worlddata.climate-comparison, [No]. [Online] Available from: https://www.
126–37.
worlddata.info/climate-comparison.php?r1=afghanistan&r2=pakistan [Accessed
[14] Weaver W. The Mathematical Theory of Communication. University of Illinois
13 Mar 2023].
Press; 1963.
[44] Yang Q, et al. A GIS-based high spatial resolution assessment of large-scale PV
[15] Pamučar D, Stević Ž, Sremac S. A new model for determining weight coefficients of
generation potential in China. Appl Energy 2019;247(April):254–69. https://doi.
criteria in mcdm models: Full consistency method (fucom). Symmetry (Basel)
org/10.1016/j.apenergy.2019.04.005.
2018;10(9):393.
[45] Anser MK, Mohsin M, Abbas Q, Chaudhry IS. Assessing the integration of solar
[16] Saha A, Roy R. An integrated approach to identify suitable areas for built-up
power projects: SWOT-based AHP–F-TOPSIS case study of Turkey. Environ Sci
development using GIS-based multi-criteria analysis and AHP in Siliguri planning
Pollut Res 2020;27(25):31737–49.
area, India. SN Appl Sci 2021;3:1–17.
[46] Retscreen Expert, Clean energy project analysis: RETScreen® engineering & cases
[17] Anwarzai MA, Nagasaka K. Utility-scale implementable potential of wind and solar
textbook: natural resources Canada. 2020. [Online]. Available: http://www.
energies for Afghanistan using GIS multi-criteria decision analysis. Renew Sustain
retscreen.net [Accessed 15 Oct 2023].
Energy Rev May 2017;71:150–60. https://doi.org/10.1016/J.RSER.2016.12.048.
[47] RETScreen Expert, Clean Energy Management Software – Version 8.1 [2022].
[18] Georgiou A, Skarlatos D. Optimal site selection for sitting a solar park using multi-
[Online] Available from: https://natural-resources.canada.ca/maps-tools-and-
criteria decision analysis and geographical information systems. Geosci Instrum
publications/tools/modeling-tools/retscreen [Accessed 13 Feb 2023].
Methods Data Syst Jul. 2016;5(2):321–32. https://doi.org/10.5194/gi-5-321-
[48] Ramasamy V et al. U.S . Solar Photovoltaic System and Energy Storage Cost
2016.
Benchmarks. With Minimum Sustainable Price Analysis : Q1 2022,” no. September,
[19] Giamalaki M, Tsoutsos T. Sustainable siting of solar power installations in
2022.
Mediterranean using a GIS/AHP approach. Renew Energy 2019;141:64–75.
[49] Mohammadi K, Naderi M, Saghafifar M. Economic feasibility of developing grid-
https://doi.org/10.1016/j.renene.2019.03.100.
connected photovoltaic plants in the southern coast of Iran. Energy 2018;156:
[20] Ruiz HS, Sunarso A, Ibrahim-Bathis K, Murti SA, Budiarto I. GIS-AHP Multi Criteria
17–31. https://doi.org/10.1016/j.energy.2018.05.065.
Decision Analysis for the optimal location of solar energy plants at Indonesia.
[50] Said Z, Mehmood A, Waqas A, Amine Hachicha A, Loni R. Central versus off-grid
Energy Rep Nov. 2020;6:3249–63. https://doi.org/10.1016/J.EGYR.2020.11.198.
photovoltaic system, the optimum option for the domestic sector based on techno-
[21] Saraswat SK, Digalwar AK, Yadav SS, Kumar G. MCDM and GIS based modelling
economic-environmental assessment for United Arab Emirates. p. 100944
technique for assessment of solar and wind farm locations in India. Renew Energy
Sustainable Energy Technol Assess 2021;43(December):2020. https://doi.org/
2021;169:865–84. https://doi.org/10.1016/j.renene.2021.01.056.
10.1016/j.seta.2020.100944.
[22] Prieto-Amparán JA, Pinedo-Alvarez A, Morales-Nieto CR, Valles-Aragón MC,
[51] Fazal MA, Rubaiee S. Progress of PV cell technology: Feasibility of building
Álvarez-Holguín A, Villarreal-Guerrero F. A regional gis-assisted multi-criteria
materials, cost, performance, and stability. Sol Energy 2023;258:203–19.
evaluation of site-suitability for the development of solar farms. Land (Basel) 2021;
10(2):1–19. https://doi.org/10.3390/land10020217.

13
J. Hamad et al. Energy Conversion and Management 303 (2024) 118188

[52] National Statistic and Information Authority (NISA). Statistical yearbook. [59] Hearty TJ, et al. Intercomparison of surface temperatures from AIRS, MERRA, and
Afghanistan: NISA; 2021. MERRA-2 with NOAA and GC-Net weather stations at summit, Greenland. J Appl
[53] Asian Development Bank (ADB). Economic and Financial Analysis. Afghanistan: Meteorol Climatol 2018;57(5):1231–45. https://doi.org/10.1175/JAMC-D-17-
ADB; 2020. 0216.1.
[54] Asian Development Bank (ADB). Sector Assessment (Summary): Energy. [60] Gupta P, Verma S, Bhatla R, Chandel AS, Singh J, Payra S. Validation of surface
Afghanistan: ADB; 2014. temperature derived from MERRA-2 reanalysis against IMD gridded data set over
[55] Abbasi HN, Zeeshan M. An integrated Geographic Information System and India. Earth Space Sci 2020;7(1):1–13. https://doi.org/10.1029/2019EA000910.
Analytical Hierarchy process based approach for site suitability analysis of on-grid [61] Rabbani R, Zeeshan M. Exploring the suitability of MERRA-2 reanalysis data for
hybrid concentrated solar-biomass powerplant. Renew Sustain Energy Rev 2023; wind energy estimation, analysis of wind characteristics and energy potential
187:113786. assessment for selected sites in Pakistan. Renew Energy 2020;154:1240–51.
[56] Ahmad M, Zeeshan M, Khan JA. Life cycle multi-objective (geospatial, techno- https://doi.org/10.1016/j.renene.2020.03.100.
economic, and environmental) feasibility and potential assessment of utility scale [62] Baranowski P, Gos M, Krzyszczak J, Siwek K, Kieliszek A, Tkaczyk P.
photovoltaic power plants. Energy Convers Manage 2023;291:117260. https://doi. Multifractality of meteorological time series for Poland on the base of MERRA-2
org/10.1016/j.enconman.2023.117260. data. Chaos Solitons Fractals 2019;127:318–33.
[57] Honrubia-Escribano A, Ramirez FJ, Gómez-Lázaro E, Garcia-Villaverde PM, Ruiz- [63] Asakereh A, Soleymani M, Sheikhdavoodi MJ. A GIS-based Fuzzy-AHP method for
Ortega MJ, Parra-Requena G. Influence of solar technology in the economic the evaluation of solar farms locations: case study in Khuzestan province, Iran. Sol
performance of PV power plants in Europe. A comprehensive analysis. pp. 488–501 Energy 2017;155:342–53.
Renew Sustain Energy Rev 2018;82(October):2017. https://doi.org/10.1016/j. [64] Ramasamy V et al. U.S. Solar Photovoltaic System and Energy Storage Cost
rser.2017.09.061. Benchmarks : Q1 2021 U.S. Solar Photovoltaic System and Energy Storage Cost
[58] Oloya IT, Gutu TJL, Adaramola MS. Techno-economic assessment of 10 MW Benchmarks; 2021.
centralised grid-tied solar photovoltaic system in Uganda. Case Stud Therm Eng [65] Ahmed MA, Mohamed MH, Al-sulaiman FA. Feasibility study of the grid connected
2021;25:100928. https://doi.org/10.1016/j.csite.2021.100928. 10 MW installed capacity PV power plants in Saudi Arabia 2017;80(December
2016) 319–329, doi: 10.1016/j.rser.2017.05.218.

14