A Spatial Decision-Support System for Wind Farm Site Selection in Djibouti

Abstract

The escalating energy demand in Djibouti requires the investigation of renewable energy sources, with wind energy emerging as a promising solution. To ensure the long-term efficiency and sustainability of wind energy projects, it is imperative to determine suitable sites for wind farm construction. When selecting a suitable site for a wind farm, there are multiple criteria to consider, such as wind velocity, ground slope, and distance to urban areas. Nevertheless, the current body of the literature reveals that no previous research has been conducted to explore an approach which involves multiple criteria to determine suitable sites for wind farms in Djibouti, as opposed to solely considering wind energy potential. This study proposes a spatial decision-support system to address the research gap in the selection of wind farm sites. Seven criteria are simultaneously evaluated in this system, including wind velocity, changes in wind direction, ground slope, distance to urban areas, distance to road network, distance to energy transmission networks, and land use. The CRITIC (Criteria Importance Through Intercriteria Correlation) method is used to objectively calculate the weights of the criteria. According to the results of performing the CRITIC method, wind velocity and distance to energy transmission networks were determined to be the most important criteria, while ground slope and land use were determined to be the least important criteria in comparison to others. A final suitability map showing the possible locations of wind farms in Djibouti was generated by considering the said criteria and their respective weights. The final suitability map reveals that the most suitable sites for the development of wind farms in Djibouti are located in the northeastern area between Obock and Khor-Angor, the southeastern area encompassing Lakes Ghoubet and Bara, and the southwestern area stretching from Lake Abbe to the Hanlé region. Using the proposed spatial decision-support system, decision makers would be empowered to make strategic and well-informed decisions when selecting the most suitable site for a wind farm in Djibouti.

1. Introduction

The issue of meeting energy demand is a major challenge in a large number of African countries. Pillot et al. [1] state that Djibouti is one of the countries facing energy challenges in Sub-Saharan Africa, as approximately 80% of Djibouti’s electricity demand is met from the Ethiopian grid [2]. However, importing electricity from another country comes at a price. Indeed, the average cost of electricity is higher in Djibouti (i.e., US $0.30/kWh) than in other countries in Africa. Lack of access to low-cost electricity supply can be a disadvantage for Djibouti’s growing economy, because Dabar et al. [3,4] state that increased economic activity inevitably leads to a significant increase in the energy demand in Djibouti. The reason for this growth in the economy may be due to the increase in the population, a result of Djibouti’s strategic location in the Horn of Africa, serving as an important shipping portal for goods entering and leaving east Africa. Therefore, alternative solutions are needed in Djibouti to meet the increasing energy demand. It is commonly claimed that the development and use of renewable energy sources (e.g., wind, solar, geothermal) can contribute to meeting energy demand and securing long-term supply of energy [5,6]. Indeed, the Government of Djibouti has established an ambitious long-term development plan to improve energy access and energy security as a primary strategic focus by using alternative energy sources [4]. A report prepared by Helimax [7] shows that Djibouti is among the fifteen African countries which have the best wind resource potential in Africa [4]. Even though Djibouti has the potential for wind energy, a review of the extant literature indicates that only a limited number of researchers focus on the deployment of wind energy in Djibouti. Dabar et al. [3] analyzed the wind velocity data measured at eight meteorological stations located in different regions for three years. Idriss et al. [8] also analyzed the wind velocity data obtained by using micro wind turbines in the city center of Djibouti. They used the collected data to investigate the economic impact of energy produced by those micro wind turbines. Dabar et al. [4] investigated the economic risk of wind energy production for electricity generation and green hydrogen production in Djibouti. These studies provide invaluable information about the wind energy potential in Djibouti, which is an important criterion in locating wind farms. However, the selection of the most appropriate sites for wind farms depends on several criteria that should be considered in the decision-making process [9], as the efficiency of wind farms that are projected to be constructed on a specific site is influenced by a variety of factors that can be classified into technical (e.g., turbine type), economic (e.g., capital costs and operational expenses), environmental (e.g., local wind climate), social (e.g., local stakeholder engagement), and operational (e.g., maintenance practices and control systems) aspects [10]. To optimize energy production and improve the economic feasibility of wind energy projects, it is essential to handle these factors as thoroughly as possible. The literature indicates that an approach that makes use of multiple criteria rather than considering only wind energy potential to show the most appropriate sites for wind farms in Djibouti has never been investigated. However, studies in the renewable energy literature (e.g., [9,11]) commonly claim that the most appropriate sites for wind farms should be selected by the combined use of multiple criteria. This study was undertaken in response to the absence of such research.

After reviewing the extant literature thoroughly, it was observed that the majority of the models involve the combined use of Geographical Information Systems (GIS) and multi-criteria decision-making methods to locate the most appropriate sites for wind farms. The use of GIS provides decision-makers with a robust solution that excludes the unsuitable areas where wind farms cannot be located. The exclusion of these areas is based on factors such as distance from protected areas, geology. Once the unsuitable areas are discarded, multi-criteria decision-making methods are used to locate the most appropriate sites for wind farms in the remaining potential areas. The extant literature also shows that the majority of the researchers used the Analytical Hierarchy Process (AHP) in determining the weights of the criteria. However, it should be noted that AHP relies on the subjective judgments of “experts” who are somehow familiar with making such decisions. According to a number of studies (e.g., [12,13]), the CRITIC method, developed by Diakoulaki et al. [14] in 1995, outperforms not only AHP but also other methods such as the entropy method because it uses statistical methods (i.e., standard deviation and correlation analysis) to objectively determine the relative importance of the criteria. To put a finer point on it, the CRITIC method does not rely on subjective judgements as is the case with AHP. In addition, it is easier to apply than AHP because the number of pairwise comparisons in AHP increases exponentially in problems involving a large number of criteria [13].

In sum, the literature indicates that there is a need for an approach that makes use of multiple criteria rather than considering only wind energy potential to show the most appropriate sites for wind farms in Djibouti. Utilizing the CRITIC method in determining the weights of the criteria that are used to locate the most appropriate sites for wind farms can create a robust model that can be used by professionals with confidence. This study aims to develop a spatial decision-support system that concurrently evaluates multiple criteria for wind farm site selection. The proposed spatial decision-support system is unique in that it (1) employs multiple criteria to identify the most suitable sites for wind farms in Djibouti and (2) utilizes the CRITIC method that overcomes the limitation related to the necessity of using subjective judgments in AHP when calculating the weights of the criteria used. The proposed spatial decision-support system has substantial implications for the wind energy industry. First, it improves decision-making processes by simultaneously integrating and evaluating a variety of criteria, thereby enabling decision-makers to consider a wider range of factors that affect the suitability of a site. Armed with such an advanced tool, professionals should be able to assess the suitability of sites for wind farm development in an objective manner. This holistic approach promotes the adoption of more strategic and well-informed decisions, essential for the sustainable development of wind farm projects. Consequently, this decision-support system can contribute to the optimization of site selection for wind farms and mitigate the risks associated with the selection of unsuitable sites. Second, the implementation of the decision-support system by industry professionals may result in more efficient and sustainable wind farm projects, fostering long-term success while reducing environmental and operational challenges.

2. Research Methodology

The process of selecting the optimal location for renewable energy projects is of utmost importance and necessitates a thorough identification and assessment of multiple criteria. Therefore, the objective of this study is to develop a spatial decision-support system for the selection of a site for a wind farm that takes into account multiple criteria simultaneously. The following tasks were performed to achieve the objective of this study:

• Review of the literature on the selection of suitable sites for wind farms.
• Identification and evaluation of the criteria to be considered when selecting suitable sites for wind farms.
• Collection of data in accordance with the defined criteria.
• Development of a suitability map for each criterion in the study area that has not previously been evaluated by multiple criteria, aside from wind potential, to determine the most suitable sites for wind farms.
• Application of the CRITIC method to the collected data to determine the relative importance of the criteria, rather than relying on subjective judgments from experts such as in the AHP method.
• Development of a final suitability map of the study area for the construction of a wind farm, considering all relevant criteria and their respective levels of importance.
• Assessment and discussion of the results derived from the examination of the maps.
• Discussion of the constraints of the study and presentation of recommendations for further research.

2.1. CRITIC Method

The CRITIC method, formulated by Diakoulaki et al. [14] in 1994, is a weight assignment technique used to objectively determine the importance of criteria in Multi-Criteria Decision-Making (MCDM) problems. This method is particularly advantageous for analyzing data that demonstrate strong correlations among criteria, allowing for a more thorough examination of the relationships between different criteria. In situations where traditional methods might fail to accurately represent the complex relationships between criteria, the CRITIC method stands out for its ability to effectively address complex correlations between criteria [15]. The CRITIC method provides several advantages in comparison to alternative criteria weighting methods. One notable advantage is its ability to consider inter-criteria correlations, which enables a more comprehensive analysis of the relationships between criteria and leads to more accurate weight assignments [16]. As a result, this method enhances the reliability and objectivity of weight assignments in the decision-making process [17]. Decision-makers who use the CRITIC method end up having a thorough understanding of the complex relationships between the different criteria and the different relative importance of the criteria. Consequently, they can make well-informed decisions [18]. The calculation steps of the method are briefly explained below:

Step 1: Development of the decision matrix X with n number of alternatives (i = 1, …, n) and m number of criteria (j = 1, …, m).

$$X={\left[x_{ij}\right]}_{n\ast m}=\left[\begin{array}{ccc}{x}_{11}& \cdots & x_{1m}\\ ⋮& \ddots & ⋮\\ x_{n1}& \cdots & x_{nm}\end{array}\right]$$

where $x_{ij}$ represents the performance value of alternative i in criterion j.

Step 2: Normalization of the elements in the decision matrix according to the type of criteria (i.e., beneficial or cost).

if the maximum value of the criterion j is the desired value:

$$n_{ij}={\displaystyle \frac{x_{ij}-{x_{ij}}^{min}}{{x_{ij}}^{max}-{x_{ij}}^{min}}}$$

if the minimum value of the criterion j is the desired value:

$$n_{ij}={\displaystyle \frac{{{x_{ij}}^{max}-x}_{ij}}{{x_{ij}}^{max}-{x_{ij}}^{min}}}$$

where $n_{ij}$ denotes the normalized performance value of alternative i with respect to criterion j.

Step 3: Calculation of the contrast intensity for each criterion, which is known as the standard deviation (σ_j).

Step 4: Determination of the amount of information (C_j) in each criterion j.

$$C_j={\mathit{\sigma }}_j\sum _{k=1}^m\left(1-r_{jk}\right)$$

where $r_{jk}$ stands for the linear correlation coefficient between criterion j and criterion k.

Step 5: Computation of the weights of the criteria (w_j).

$$w_j={\displaystyle \frac{C_j}{\sum _{k=1}^m{C}_k}}$$

2.2. Characteristics of the Study Area

The Republic of Djibouti is a country located in East Africa at coordinates 43°00′ E, 11°30′ N (see Figure 1). The country spans 23,200 square kilometers and experiences a climate characterized by two distinct seasons: winter and summer. Winter is characterized by hot and humid conditions, while summer is characterized by extremely hot and humid conditions. In winter, the temperature fluctuates between 20 °C and 22 °C, while in summer, it ranges from 30 °C to 40 °C. Despite its small population size of 988,002 in 2020, the country is facing a shortage of energy supply, with only half of the population having access to electricity. Power is available only 60% of the time, resulting in significant instability in the distribution of electricity and a substantial number of remote areas experiencing power outages [19]. The distribution of electric energy in Djibouti is controlled by EDD, which stands for Electricity of Djibouti. EDD has a legal monopoly over the entire distribution and transportation of electric energy throughout the country. The EDD’s social price range for electricity is between $0.21 and $0.52/kWh. It is widely acknowledged that the cost of electricity in Djibouti is significantly higher compared to neighboring countries, as reported by the Directorate of Statistics and Demographic Studies in 2018. The country possesses significant potential for various forms of renewable energies that can support sustainable development as well as address the needs of remote and rural populations [20,21,22]. In this context, wind energy is a promising energy source in Djibouti. Although several studies have been conducted to estimate the potential of renewable resources in Djibouti, with a particular emphasis on solar and geothermal energy (e.g., [1,23]), there are only a limited number of studies that have been performed on wind energy.

2.3. Identification and Evaluation of Criteria

When conducting a site suitability analysis for the construction of wind farms, it is crucial to take into account the most significant criteria relative to environmental, economic, technical, and social considerations in order to minimize any potential adverse effects on the selection process. After an extensive review of the existing literature, seven essential criteria, including wind velocity (characterized by the hourly mean of the wind velocity, C₁), changes in wind direction (characterized by the standard deviation of wind direction, C₂), ground slope (C₃), distance to urban areas (C₄), distance to road network (C₅), distance to energy transmission networks (C₆), and land use (C₇), were identified that could potentially affect the selection of a suitable site for the installation of a wind farm.

Table 1. Site selection criteria for wind farms, along with relevant references.

Author(s)	Year	Criteria IDs
		C1	C2	C3	C4	C5	C6	C7
Tracy et al. [24]	1978	✓	✓	✓
Pennell et al. [25]	1980	✓	✓	✓
Kirchhoff and Kaminsky [26]	1981	✓	✓
Druyan [27]	1985	✓	✓	✓				✓
Mosetti et al. [28]	1994	✓	✓
Baban and Parry [29]	2001	✓	✓	✓	✓	✓		✓
Tegou et al. [30]	2010	✓					✓	✓
Al-Yahyai et al. [31]	2012	✓		✓	✓
Sanchez-Lozano et al. [32]	2014	✓		✓	✓	✓	✓	✓
Latinopoulus and Kechagia [33]	2015	✓		✓	✓	✓		✓
Atici et al. [34]	2015	✓		✓		✓	✓
Fetanat and Horasaninejad [35]	2015	✓			✓		✓
Noorollati et al. [36]	2016	✓		✓	✓	✓	✓	✓
Wu et al. [37]	2016	✓		✓	✓	✓	✓	✓
Sanchez-Lozano et al. [38]	2016	✓		✓		✓	✓	✓
Villacreses et al. [39]	2017	✓			✓	✓	✓	✓
Chaouachi et al. [40]	2017	✓		✓				✓
Chamanehpour et al. [41]	2017	✓		✓	✓			✓
Ayodele et al. [42]	2018	✓		✓	✓	✓	✓	✓
Değirmenci et al. [43]	2018	✓		✓	✓	✓		✓
Solangi et al. [44]	2018	✓			✓	✓		✓
Rehman et al. [45]	2019	✓		✓			✓
Abdel-Basset et al. [46]	2021	✓					✓
Deveci et al. [47]	2021	✓					✓
This study	2024	✓	✓	✓	✓	✓	✓	✓

presents the criteria that influence the selection of a suitable site for the construction of a wind farm, along with their corresponding references.

The following subsections provide a comprehensive explanation of each criterion.

2.3.1. Wind Velocity (C₁)

In the selection of wind farm sites, the wind velocity criterion is an essential determinant because it directly affects the power output of wind farms. The energy generation potential of a site is significantly influenced by wind velocity [18,48]. The mean long-term wind velocity is frequently used to evaluate the suitability of a wind farm location, emphasizing the significance of wind velocity in determining the feasibility of generating wind power at a specific site [49].

Evaluating the feasibility of wind farm sites necessitates a good understanding of the wind velocity probability distributions, which are determined by the annual mean wind velocity and standard deviation values [50]. The operating performance of wind farms is significantly influenced by the distribution of wind velocity, emphasizing the necessity for a comprehensive understanding of wind characteristics during the site selection process [51]. The importance of wind velocity in achieving optimal energy generation levels is underscored by the fact that high wind velocity is required to optimize the power output of wind farms [52]. As a consequence, locations with lower wind velocity hold less significance compared to those with higher wind velocity.

In this study, wind velocity data were acquired from ECMWF Reanalysis v5 (ERA5) datasets that are accessible through the Copernicus Climate Data Store (C3S). ERA5 is the fifth-generation climate reanalysis dataset of the ECMWF (European Centre for Medium-Range Weather Forecasts), which has been developed for over 20 years and has been validated by numerous studies [53]. The wind data obtained from ERA5 has been calibrated with the monthly mean wind data time series reported in Dabar et al. [3]. A spatially distributed hourly wind velocity dataset for the Republic of Djibouti was compiled, comprising 262,800 records that span a 30-year period from 1993 to 2022. Wind velocity exceeding 6 m/s at hub height is considered suitable; however, wind velocity lower than 3 m/s must be excluded from the analysis since the latter value is the cut-in velocity limit for the wind turbines [54].

2.3.2. Changes in Wind Direction (C₂)

Wind direction is the bearing of the wind velocity vector with respect to geographical North, measured in degrees clockwise direction. Changes in wind direction over time are characterized by the standard deviation of wind velocity direction and are a representative metric to quantify the turbulence level in wind velocity. Changes in wind direction significantly affect wind turbine loads, wake characteristics, and the overall performance of the wind farm. According to Shaler et al. [55], wind turbine loads are indeed susceptible to changes in wind direction, which may involve vertical wind shear and wind turbulence. Furthermore, the evaluation of wind velocity along with an analysis of wind direction is important for the optimization of micro-siting, the reduction of operational costs, and the enhancement of turbine efficiency [56]. The persistence of low wind velocity and frequent changes in the wind direction are negative factors in the search for wind farm sites [57]. Locations with frequent changes in wind direction hold less significance compared to those with infrequent changes in wind direction, or more steady wind [58]. Changes in wind direction are measured by calculating the standard deviation of the wind velocity vector as recorded in the field. In the scope of this study, standard deviations below 84° are considered suitable, while wind direction standard deviations exceeding 96° are excluded from the analysis.

2.3.3. Ground Slope (C₃)

The slope of the site significantly affects the efficiency, safety, and cost-effectiveness of wind power plant installations. Wind turbine installation and maintenance can be difficult on steep slopes, so it is desirable to construct wind farms on flat or gently sloping areas [59]. Site-specific criteria, such as ground slope, proximity to transportation networks and power grids, and social acceptance play an important role in determining the feasibility and long-term success of wind farm projects [60,61]. Most notably, the precision and efficiency of wind farm site selection processes have been substantially enhanced by incorporating round slope data into Geographic Information Systems (GIS) and decision-making models.

Wind turbines are usually situated on land that has slopes of less than 10–20% in order to maximize operational efficiency and minimize difficulties during installation [62]. Furthermore, higher ground slope is an indication of rough topography which would adversely influence transport and installation logistics. Locations with steeper slopes are of lesser significance than those with gentler slopes [33,63]. Areas with slopes less than 25% are considered suitable for this study, whereas areas with slopes exceeding 67% are excluded from the analysis.

2.3.4. Distance to Urban Areas (C₄)

The proximity to urban areas influences operational efficiency, environmental considerations, and community acceptance. When selecting sites for wind farms, it is important to take into account the proximity to urban areas in order to minimize potential conflicts with local communities and ensure adherence to regulations and noise standards [61,64]. In addition, the safety of the residents and the disruptions to their daily activities that may result from wind turbine operations must be carefully considered [65]. Incorporating the distance to urban areas into site selection processes enables stakeholders to make informed decisions that maximize energy generation potential while minimizing conflicts and costs associated with urban proximity [62]. A wind farm project’s social acceptance by the local population can be improved by strategically placing wind farms at a suitable distance from urban centers. Hence, sites located further away from urban areas hold more significance compared to sites in closer proximity to urban areas. Areas located more than 10 km away from urban areas are deemed to be suitable [54].

2.3.5. Distance to Road Network (C₅)

The proximity to road networks affects the costs associated with construction and maintenance as well as the effectiveness of operations. Wind farm developers attempt to minimize construction and maintenance costs, emphasizing the importance of taking into account the distance of potential sites to road networks [66]. Proximity to road infrastructure reduces transportation expenses, facilitates equipment delivery, simplifies maintenance operations, and enhances emergency response and evacuation capabilities ultimately improving the economic feasibility of wind farms [63,67]. Construction sites located in close proximity to road networks offer enhanced accessibility for construction equipment, personnel, and maintenance vehicles, resulting in improved project management and reduced operational downtime [68]. Furthermore, the closeness to road networks plays an important role in minimizing negative environmental effects and maximizing land utilization [39]. Developers who build wind farms near existing road infrastructure can minimize the fragmentation of all animal species’ habitats, the disturbance to wildlife, and the need for new road construction [18]. Consequently, sites in close proximity to road networks hold more importance compared to sites located farther away from road networks. Locations within a 5-km radius of road networks are regarded as suitable [33].

2.3.6. Distance to Energy Transmission Network (C₆)

The proximity to energy transmission networks has a substantial effect on operational efficiency, cost-effectiveness, and grid integration. Wind farms located in close proximity to existing energy transmission networks have distinct advantages, such as lower transmission losses, lower costs associated with cabling, and improved connectivity to the power grid, leading to enhanced economic feasibility [69,70,71,72]. The close proximity of wind farms to energy transmission networks also improves the reliability and stability of wind power generation, thereby promoting the long-term success of wind farm projects [73]. It is essential to take into account the distance to energy transmission networks in order to optimize energy delivery and reduce system losses in wind farm operations. Wind farms located in close proximity to transmission networks can efficiently integrate renewable energy into the grid, supporting the shift towards a greener and more sustainable energy blend [74]. Therefore, sites in close proximity to transmission networks hold greater importance compared to sites located far away from transmission networks [75]. Locations that fall within a 68-km radius of transmission networks are considered to be suitable.

2.3.7. Land Use (C₇)

The availability of land has a significant influence on the feasibility of a wind farm project. It is imperative not to create conflicts with existing land use, such as agriculture, residential areas, or protected lands [76]. An assessment of land use helps to determine suitable sites for wind turbines that minimize any adverse effects on local ecosystems or land functionality [77]. The inclusion of land use considerations in site selection processes enables developers to make well-informed decisions that achieve a balance between energy production objectives and the sustainability of the environment and land use [78]. In addition, land use helps to evaluate the social and economic consequences of wind farm projects. An assessment of the existing land use is beneficial in understanding the potential effects of a wind farm on local communities, wildlife habitats, and cultural landscapes [62]. Land use evaluation also facilitates the identification of suitable sites that do not stand in the way of environmental conservation objectives and that satisfy regulatory requirements [79].

Short vegetation is more suitable for the construction of a wind farm compared to tall vegetation. As a consequence, cultivated areas, grassland, shrubs, and bare land are deemed to be highly suitable, while forest areas are deemed to be unsuitable [80]. The utilization of barren regions and arid lands for renewable energy has the potential to transform unproductive areas into productive sites, thereby advancing both local and national energy objectives while simultaneously dealing with issues with other land uses and potential disputes with landowners.

3. Findings

Following the identification of seven criteria that can influence the selection of suitable sites for wind farms, actual data pertaining to the criteria was collected from the study area. The validity and reliability of the data are critical to the quality of the results of the study. Therefore, all the required data were meticulously collected and verified for accuracy prior to their utilization for analysis in the Geographic Information System (GIS). The elimination of undesirable alternatives is the initial step in the data collection process. Hence, all sites that were deemed ecologically, biologically, culturally, or spiritually important were deliberately omitted from the analysis. A total of 23,299,806 potential sites were evaluated in this study for the construction of wind farms in Djibouti. Real data was collected for 23,299,806 alternative sites using a variety of measurement units in seven criteria. Wind velocity data were obtained from the ECMWF Reanalysis v5 (ERA5) datasets. Land use data were acquired from the Global Land Data Assimilation System (GLDAS) datasets, while ground slope, distance to urban areas, distance to road networks, and distance to power transmission networks were sourced from NextGIS datasets. In order to objectively determine the relative importance of the criteria, the CRITIC method was implemented on the data matrix (i.e., decision matrix) that includes 23,299,806 alternative sites and seven criteria. After being coded in MATLAB 9.13 (R2022b), the CRITIC method was applied to the problem at hand. In this process, only the land use criterion was evaluated using a nominal scale (e.g., wetland: 1, bare land: 5), while the other criteria were evaluated using actual data. The weights of the criteria, determined through the implementation of the CRITIC method, are presented in

Table 2. The relative importance of the criteria affecting the selection of suitable sites for wind farms.

Criterion	Weight
Wind velocity (C1)	0.232
Changes in wind direction (C2)	0.130
Ground slope (C3)	0.061
Distance to urban areas (C4)	0.129
Distance to road network (C5)	0.145
Distance to energy transmission network (C6)	0.183
Land use (C7)	0.120

.

Based on the results of the CRITIC method presented in Table 2, the median hourly wind velocity and the distance to energy transmission networks were determined as the most important criteria. Conversely, ground slope and land use were determined to be the least important criteria in comparison to other criteria. Several previous studies, including [81,82,83] have yielded similar results. It should be noted that the weight of each criterion determined by the CRITIC method is site-dependent, meaning that for different application sites, the order of criterion weights different is subject to change. This indicates that the CRITIC method is adaptive to the study area it is applied.

The wind velocity, the criterion with the highest weight, is critical because it directly influences the energy output of wind turbines, with higher wind velocity typically resulting in increased electricity generation [84,85]. The distance to energy transmission networks, the criterion with the next highest weight, is essential for the efficient and cost-effective connection of wind power to the grid [18]. Wind farms that are situated at a considerable distance from transmission networks may experience increased expenses for connecting to the grid, which may affect the economic viability and overall feasibility of wind farm projects [86].

Land use and ground slope, criteria with the lowest weights, are important criteria in the selection of wind farm sites, but they are often less important compared to wind velocity and distance to energy transmission networks, which have a more direct impact on the economic viability, feasibility, and efficiency of wind farm projects (see Table 2). Wind turbines are typically installed on farms with slopes that are less than 10–20% [59]. This indicates that although ground slope is a factor to be considered, it may not be among the most important factors when selecting a site for wind turbine installation. In addition, the spatial advantages and constraints of land use and ground slope are not trivial when selecting suitable sites for wind farms [87], but they may not hold the same importance as other more influential criteria. The primary reasons that land use is less important than other criteria are (1) Djibouti is a small country in terms of land area, but its energy requirements are relatively low, and (2) a significant portion of the country’s land area, approximately 70%, is bare land.

A suitability map was developed for each criterion in the study area by using specific measurement units and the upper and lower limits for each criterion, which were determined based on previous studies in the relevant literature. To assess the suitability of an alternative site for wind farm installation, each criterion was categorized into five grades (e.g., “1: not suitable”, “5: highly suitable”) based on its level of suitability between the lower and upper limit values.

Table 3. Class intervals for the criteria and their corresponding grades.

ID	Criterion	Measurement Unit	Class Intervals	Grade	Area (km2)
C1	Wind velocity	m/s	3–4	1	5385
			4–5	2	8892
			5–6	3	4786
			6–7	4	2729
			7–8	5	1409
C2	Changes in wind direction	°	89–96	1	1111
			86–89	2	1863
			84–86	3	7223
			81–84	4	7883
			78–81	5	5120
C3	Ground slope	%	>67	1	418
			44–67	2	1555
			25–44	3	2766
			9–25	4	4727
			0–9	5	13733
C4	Distance to urban areas	km	0–2	1	8679
			2–6	2	6712
			6–10	3	3848
			10–16	4	2815
			16–26	5	1146
C5	Distance to road network	km	15–26	1	8679
			9–15	2	6712
			5–9	3	3848
			2–5	4	2815
			0–2	5	1146
C6	Distance to energy transmission networks	km	132–178	1	2926
			99–132	2	3992
			68–99	3	6526
			31–68	4	4677
			0–31	5	5079
C7	Land use	Nominal scale	Wetland/settlement	1	275
			Forest	2	2
			Shrub land	3	6520
			Agricultural land	4	18
			Bare land	5	16,385

provides detailed information on the seven criteria, including their IDs, measurement units, class intervals, grades, and area.

The class intervals and grades of each criterion were evaluated for each of the 23,299,806 potential sites in Djibouti, as follows:

3.1. Evaluation of the Suitability Map for Wind Velocity (C₁)

There is a significant variation in wind velocity across Djibouti, revealing that the southeastern and certain areas of the southwestern regions are particularly well-suited for the development of wind farms, as the wind velocity exceeds 6 m/s, as depicted in Figure 2. Conversely, western and northwestern regions with wind velocity below 3 m/s are not suitable for the installation of wind farms. The diverse climate of Djibouti, which is affected by its coastal position and distinct topographical characteristics, is the reason for the variation in wind velocity across different areas. In addition, the existence of flat or gently inclined land, specifically in the southeastern and certain areas of the southwest regions of Djibouti, provides a long and wide stretch of land that enables the unrestrained movement of wind, which leads to increased wind velocity. Table 3 indicates that over 38% of the study area falls within the range of suitability from moderate to very high wind velocity. Conversely, the remaining 62% of the study area is categorized as being less suitable to not suitable for installing a wind farm, as shown in Figure 2.

3.2. Evaluation of the Suitability Map for Changes in Wind Direction (C₂)

As illustrated in Figure 3, the southeastern and northwestern regions of Djibouti exhibit consistent wind patterns with lower changes in wind direction because of the lack of noticeable geographical obstacles that create consistent atmospheric conditions in these two regions. Conversely, the southwestern region higher changes in wind direction is not suitable. According to Table 3, more than 87% of the study area is classified as moderately to highly suitable. In contrast, Figure 3 demonstrates that the remaining 13% of the study area is classified as “less suitable to not suitable” for the installation of a wind farm.

3.3. Evaluation of the Suitability Map for Ground Slope (C₃)

As depicted in Figure 4, the northeastern areas, with the exception of those near the coast, as well as the western regions, are unsuitable for hosting wind farms due to very steep slopes. The topography of Djibouti, located in the Horn of Africa, displays a variety of ground slopes as a result of tectonic movements, weathering, and erosion that occurred over millennia. Dry regions exhibit steeper slopes, whereas areas with vegetation possess gentler slopes. Therefore, the vast majority of the study area is suitable for the installation and maintenance of wind farms due to its minimal slope. Table 3 indicates that over 91% of the study area is categorized as moderately to highly suitable. Conversely, Figure 5 illustrates that the remaining 9% of the study area is categorized as “less suitable to not suitable” for the construction of a wind farm.

3.4. Evaluation of the Suitability Map for Distance to Urban Areas (C₄)

The northeastern, western, and southwestern regions of Djibouti, which are located away from heavily populated areas, are suitable for the establishment of wind farms, as demonstrated in Figure 5. Constructing wind farms in rural areas decreases the likelihood of harm to human wellbeing caused by noise, turbine malfunctions, or accidents. According to Table 3, more than 34% of the study area is classified as moderately to highly suitable. On the other hand, the remaining 66% of the study area, is marked as “less suitable to not suitable” for the installation of a wind farm (see Figure 5).

3.5. Evaluation of the Suitability Map for Distance to Road Network (C₅)

The northern, central, and southeastern regions of Djibouti are home to suitable sites for wind farm construction, which are located less than 5 km from road networks, as shown in Figure 6. The optimized access to existing highways in these regions reduces the necessity for the construction of new access roads, thereby reducing costs and minimizing environmental impact. Moreover, in these regions, emergency services can promptly and efficiently handle accidents, equipment malfunctions, or natural disasters. On the other hand, the western, southern, and southwestern regions of Djibouti are unsuitable for wind farm construction due to their remote locations from road networks. Based on Table 3, over 34% of the study area is identified as moderately to highly suitable. However, as demonstrated in Figure 6, the remaining 66% of the study area is classified as “less suitable to not suitable” for the development of a wind farm, as far as the distances to road network are considered.

3.6. Evaluation of the Suitability Map for Distance from Energy Transmission Networks (C₆)

As indicated in Figure 7, wind farms located in the south-eastern region of Djibouti would benefit from their close proximity to energy transmission networks, resulting in minimal transmission losses and consequently high efficiency and effectiveness. The south-eastern region is therefore highly suitable to the installation of wind farms. However, the north-eastern region of Djibouti is unsuitable for the installation of wind farms due to the considerable distance to energy transmission networks. Table 3 reveals that more than 70% of the study area is classified as moderately to highly suitable. Nevertheless, as illustrated in Figure 7, 30% of the study area is classified as “less suitable to not suitable” for the construction of a wind farm given their distances to energy transmission networks.

3.7. Evaluation of the Suitability Map for Land Use (C₇)

The current land use of Djibouti, as presented in Table 3 and Figure 8, indicates that the vast majority of the study area (i.e., 99%) is classified as highly suitable.

The final suitability map of Djibouti was generated by combining the objectively obtained weights of the criteria, which were calculated using the CRITIC method, with the grading process of potential sites. The final suitability map was generated by assigning corresponding weights to the criteria within the ArcGIS program, wherein potential sites in Djibouti were graded based on each criterion. The final suitability map of Djibouti is presented in Figure 9.

4. Discussion on the Final Suitability Map

The final suitability map in Figure 9 demonstrates the most suitable sites for the construction of wind farms in the study area, as determined by the aggregate of all criteria and their respective weights. According to Figure 9, the most suitable sites for the construction of wind farms in Djibouti are the northeastern region between Obock and Khor-Angor, the southeastern region encompassing Lake Ghoubet and Bara, and the southwestern region stretching from Lake Abbe to the Hanlé region. The analysis results for the three most suitable regions for wind farm construction in Djibouti, evaluated individually based on seven criteria, are as follows:

• The following are the characteristics of the region between Obock and Khor-Angor:

  ○

  Wind velocity ranges from 5.4 to 7.4 m/s, which is considered moderately to very highly suitable.

  ○

  Changes in wind direction are infrequent, evidenced by a standard deviation less than the 84° limit for suitability.

  ○

  The region is characterized by gentle slopes of much less than the 25% limit for suitability.

  ○

  The region is situated over 6 km from urban areas.

  ○

  The region is situated within a 5-km radius of highways as required for a good location.

  ○

  The region is primarily composed of barren land and scrubland that easily lend themselves for wind farm construction and operation.

  ○

  The only caveat with this location is that it is situated over 132 km from any power transmission network.

• The following are the characteristics of the region surrounding Lake Ghoubet and Bara:

  ○

  Wind velocity ranges from 5.6 to 7 m/s, which is considered moderately to highly suitable.

  ○

  The wind farm area is stable, as the standard deviation is within 84° and changes in wind direction are rare.

  ○

  The slopes in the region are gentle and fall within the 25% threshold for suitability.

  ○

  The region is situated within 5 km of highways, ensuring efficient access for transportation and maintenance.

  ○

  The region is located within 31 km of energy transmission networks, facilitating connectivity to the grid.

  ○

  The land is primarily barren and covered in shrubs, making it a suitable site for the construction of a wind farm.

  ○

  Nonetheless, the drawback of this region is its closeness to urban areas, situated less than 10 km away, which may present challenges for community acceptance and adverse environmental effects.

• The following are the characteristics of the region stretching from Lake Abbe to the Hanlé region:

  ○

  Wind velocity ranges from 4.9 to 6.5 m/s, which is considered low to highly suitable.

  ○

  The region features gentle slopes that satisfy the suitability criterion of 25%.

  ○

  The region is located over 10 km from urban areas, thereby reducing the impact on human populations.

  ○

  The land is characterized by barren terrain and a small water body near Lake Abbe, which may restrict the amount of available land for constructing a wind farm.

  ○

  This region is characterized by a high degree of wind direction variability, with a standard deviation exceeding 84°. Additionally, it is situated 68 to 99 km from energy transmission networks and is more than 17 km from road networks, which complicates and increases the cost of grid integration and accessibility.

It is important to note that these regions may not be the most suitable sites for wind farm construction when evaluated individually according to the seven criteria. However, they are the most suitable sites when all seven criteria are evaluated simultaneously with their respective weights. The suitability of Djibouti for wind farm installation is outlined in

Table 4. Classification of potential sites for wind farm construction in Djibouti according to the level of their suitability.

Level of Suitability	Class Intervals	Grade	Area (km2)	Area (%)
Not suitable	0	0	69	0.30
Very low	0–1	1	2331	10.05
Low	1–2	2	6341	27.33
Moderate	2–3	3	7418	31.97
High	3–4	4	4492	19.36
Very high	4–5	5	2549	10.99

, providing area-specific information.

Table 4 reveals that moderately suitable, highly suitable, and very highly suitable areas for wind farm construction in Djibouti account for 62% of the total area. Conversely, the unsuitable, less suitable, and very less suitable areas comprise 38% of the total area. The results of the study indicate that Djibouti contains largely suitable sites (14,459 km²) for the generation of wind energy, indicating substantial wind energy potential in the country. However, Djibouti’s unique geographical characteristics, inadequate infrastructure and logistical difficulties, especially in remote areas deemed highly suitable for wind energy, raise substantial concerns about the feasibility of construction and maintenance of wind farms. Unless meticulously managed, the construction of large-scale wind farms in Djibouti could potentially disrupt traditional practices, as rural livelihoods are closely linked to land use. Therefore, this study proposes a spatial decision-support system that simultaneously evaluates seven critical criteria for site selection in wind farm development, thereby providing a robust framework for all stakeholders involved in the wind energy project planning process. This study proposes a spatial decision-support system that serves as an effective tool for stakeholders in the wind energy industry. The proposed spatial decision-support system allows decision makers to make strategic and well-informed decisions when selecting the most suitable sites for wind farm installation by evaluating multiple criteria at the same time. In conclusion, professionals who utilize the proposed model should be able to achieve successful and long-lasting wind farm projects.

5. Conclusions

Access to enough energy is problematic in several countries in Sub-Saharan Africa, including Djibouti. Given the scarcity of energy production and the growing energy demand in Djibouti, a comprehensive examination of renewable energy sources has become imperative. Wind energy has emerged as a potential solution. When it comes to Africa’s wind resources, Djibouti is one of the fifteen countries with the most promise [4,7]. Despite this promise, a thorough examination of the current literature has revealed a noticeable lack of research on the topic of wind energy production in Djibouti. Although several studies have provided valuable insights into the wind energy potential of Djibouti, there is an absence of research that consider a multi-criteria approach to the selection of suitable sites for wind farm installation in the country.

The main objective of this study is to fill this research gap by developing a spatial decision-support system for the selection of suitable wind farm sites in Djibouti. The proposed decision support system evaluates seven key criteria for selecting wind farm site locations, which include wind velocity, changes in wind direction, ground slope, distance to urban areas, distance to road networks, distance to energy transmission networks, and land use. The CRITIC method was used to determine the weights of these criteria.

• It was found that the most important criteria were wind velocity and distance to energy transmission networks. Conversely, ground slope and land use were determined to be the least important criteria. Although land use and ground slope are essential criteria in the location of wind farm sites, they are frequently overshadowed by other criteria, such as wind velocity and distance to energy transmission networks, which have a more direct influence on the economic viability, feasibility, and efficiency of wind farm projects.

A final suitability map for the installation of wind farms in Djibouti was developed by taking into account all criteria and their corresponding weights. The outcomes derived from the final suitability map are as follows:

• Wind farms in Djibouti would be most effectively located in three regions: the northeastern region (encompassing Obock and Khor-Angor), the southeast region (encompassing Lakes Ghoubet and Bara), and the southwest region (encompassing Lake Abbe and Hanlé). Although these three regions may not be suitable for the installation of wind farms based on each criterion taken separately, they are determined to be compromise locations when all criteria are evaluated together, taking into account their respective weights.
• The area in Djibouti that is suitable wind farm sites is significant (14,459 km²), which could be a sustainable solution to address a significant portion of the country’s increasing energy demands.

This study offers valuable insights to policymakers and stakeholders who are involved in the development of wind energy infrastructure in Djibouti, thereby facilitating the country’s transition to a more sustainable and environmentally friendly energy landscape. Through the simultaneous assessment of multiple criteria, the proposed spatial decision-support system enables decision makers to make strategic and well-informed decisions when selecting the most suitable sites for wind farm construction. To sum up, practitioners can use the outcomes of this research study to locate as many wind farms as necessary in the most suitable locations to meet the growing energy demand in Djibouti.

This study has three limitations. First, the proposed spatial decision-support system considers seven essential criteria to determine the most suitable sites for wind farms. The criteria for site selection in wind farm construction may differ based on the economic, social, and geographical conditions of the country in question. In future research, the proposed approach can be expanded to incorporate additional criteria, such as economic, social, and environmental impact. Second, the proposed decision support system was solely tested in Djibouti. The proposed approach can be tested in other countries where actual data on the site selection criteria, adjusted to take into account local geographical, economic, and social conditions, are available. Future research is recommended to explore possible sites for off-shore wind farms. Third, the weights of the criteria were objectively determined utilizing the CRITIC method. Future research could substitute the CRITIC method with alternative objective methods grounded in quantitative data analysis, such as Entropy.