1. Introduction
Solar energy is one of the most promising renewables because it is considered a consistent source of energy that is not significantly vulnerable to seasonal weather patterns changes [
1]. In addition, the output efficiency of solar technologies has been increasing in recent years and the ability to utilize them in a variety of locations is very favorable [
2]. The demand for solar energy is increasing worldwide as countries are following steps for sustainable development and CO
2 emissions reduction [
3]. Solar energy can be exploited through the solar photovoltaic (PV) and solar thermal technologies for various applications. Solar radiation can be converted directly into electricity by using photovoltaic (PV) technology, which is one of the potential methods that offer clean and renewable energy. Compared to CO
2 emissions from coal combustion which amounts to 975 g per kilowatt-hour (kWh), the emissions from using PV is about 50 g per kWh [
4]. It offers ongoing free energy and the life expectancy of solar products can be up to 30 years [
4]. In recent years extensive studies were conducted on the optimal use of solar energy [
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15,
16]. Studies have demonstrated that solar energy can roughly fulfill 1000 times the global energy requirement; although, nowadays only 0.02% of this energy is utilized [
17]. Such estimations are generally based on the physically available solar radiation on the Earth’s surface [
18]. Although the global utilization level of solar energy has been a small fraction of its actual capacity, in recent years the investments in solar energy have grown significantly [
19].
Europe, China, and the United States are the main investors in this field. A recent report by Renewable Energy Policy Network for the 21st Century (REN21), however, revealed that developing world invested more in renewable energy than rich countries for the first time in 2015 [
20]. Devabhaktuni et al. [
2] believe solar energy is a beneficial utility, especially for developing countries, for several reasons, including the locations of most developing countries in regions with high solar radiation.
Large-scale PV systems provide significant environmental advantages when compared to conventional energy sources. In addition, it facilitates the reuse of marginal lands [
21]. However, the required area for implementing large PV systems may cause undesirable impacts on landscape, land use, and biodiversity [
22]. Preferably, these implementations should be located on unused, low productivity lands to minimize such impacts [
23]. Non-ideal locations are forests, extreme remote areas, and the areas with instability and a high degree of existing development [
24]. In order to identify a suitable location for solar PV installation, spatial analysis with multi-criteria function and decision support system is required. In recent years, spatial analyses of renewable energy suitability became a popular research area [
25].
Massimo et al. [
26] developed a geographical information database system (GIS DB) for the integration of solar energy in the energy planning of a wide area of central Italy to evaluate the productive potential of the land use, estimate residential energy consumptions and assess the renewable energy sources. De Simón-Martín et al. [
27] developed an innovative Geographic Information System (GIS) application to locate and supervise the operation of a large PV plant of 108 kWp and a small PV plant of 9 kWp installed on a home rooftop. Their developed tool could increase control of the PV plant performance and could help to “evaluate several PV module replacement strategies in a preventive maintenance program”. Huld et al. [
28] developed a GIS-based tool to perform geospatial analysis and mapping of the energy output and reliability of PV mini-grid system for estimating the performance of photovoltaic (PV) mini-grid system over large geographical areas. Carrion et al. [
29] conducted a study aimed to determine the electricity generation capacity of solar photovoltaic power plants and to select solar energy site in Andalusia (Spain). They defined criteria with considering national parks, land use, urban planning, slope, shadows, accessibility, radiation, sunshine hours and temperature. Hofierka and Kanuk [
30] presented a method for evaluation of photovoltaic potential in urban areas using open source solar tools and three-dimensional modeling in an urban environment using GIS. This radiation tool provided R.Sun solar radiation model, which was estimated using PVGIS application. Janke [
31] identified the sites with high potential for solar and wind fields in Colorado and found suitable areas for development of solar and wind fields using multi-criteria modeling and such criteria as wind potential, solar potential, distance to transportation lines, distance to the city, population density, distance from heavy roads and state lands acquired using the GIS. Aragonés-Beltrán et al. [
32] selected the best variable among four variables using analytic network process (ANP) by minimizing the risk for photovoltaic power plant projects. They used twelve criteria for implementation of the decision-making process.
In order to select the most optimal sites for new development plans including solar power plants, it is necessary to develop a set of criteria and factors to facilitate the decision-making process. To assess the suitability of land for the construction of photovoltaic fields, Charabi and Gastli [
33] used technical criteria including solar radiation indices, access to land and land use; economic criteria including proximity to the distribution network and slope; and environmental criteria including sensitive areas, hydraulic lines, and risk of dunes. Sánchez-Lozano et al. [
34] categorized the criteria as environmental including agrological capacity; geomorphological including slope, orientation, and area; location including distance to roads, distance to power lines, distance to villages, distance to substations; and climatic including solar irradiation potential and average temperature. In another categorization of the criteria for photovoltaic site selection, Uyan [
35] suggested environmental criteria including distance from residential areas and land use; and economic criteria including distance from roads, slope and distance from transmission lines. Weights can be assigned to the above criteria according to their importance and each of the variable and their weights may have more or less favorability in the final decision than another [
33].
To facilitate the flow of environmental information from data sources to decision-makers there is a growing need for well-developed environmental information systems [
36]. In recent years, ArcGIS software has transformed the environmental decision-making process. It has significantly changed the organization and management of geographical data and has improved the spatial modeling and assessment capabilities across a range of disciplines [
37,
38,
39,
40,
41,
42]. The application of GIS software in the selection of the suitable sites for new development plans can highly accelerate the decision-making with combining different data layers in the form of different conceptual models. Based on the type of combination strategy in these models, the type, value, and the number of data layers will be different [
43].
The majority of GIS-based site suitability studies are built upon multi-criteria analysis (MCA) to synthesize complex problems with multiple variables [
44]. According to Hermann et al. [
45] MCA is “a decision-making tool used in environmental systems analysis to evaluate a problem by giving an order of preference for multiple alternatives on the basis of several criteria that may have different units” (p. 1788). Some methods for multi-criteria analysis include analytic hierarchical process (AHP), analytic network process (ANP), Boolean logic, weighted linear combination (WLC), and fuzzy logic. These methods have been used for identifying optimal sites for solar PV installation (e.g., [
33,
34,
35,
46,
47,
48]). AHP is one of the most comprehensive systems designed for multi-criteria decision-making developed by Saaty [
49]. Using this method, it is possible to formulate problems in a hierarchical fashion [
38]. AHP’s main characteristic is based on paired comparisons [
43]. Complex problems with multiple criteria can be broken down into a number of one-to-one comparisons [
50]. ANP, however, is a more general form of the AHP which considers the interdependence of the included criteria [
51]. “It takes the dependence and feedback among elements into account based on the actual situation which expands and improves AHP so that it can simulate the complex interrelationships of reality better” [
52] (p. 42). In Boolean logic, all values attributed to the criteria are reduced to either TRUE (1) or FALSE (0) meaning that in each criterion the land is either suitable or unsuitable [
53] for a particular development plan such as solar power plant installation. In WLC, instead of absolute values of 0 or 1, importance weights are assigned to criteria. Both Boolean and WLC methods employ discrete thresholds to define suitability, and the values and weights are usually defined by a group of experts [
54]. In both methods, many assumptions and uncertainties are involved, while no sensitivity analysis is performed [
44]. In contrast, fuzzy logic provides the possibility of more flexible analyses where there are no certain boundaries between suitable and non-suitable [
55]. The fuzzy concept in site selection studies involves classes with continuous grades of membership ranging from 0 to 1, with 0 as the indication of absolute falsehood showing that an area is not suitable for intended plan; 1 as the indication of the absolute truth showing that the area is suitable, and values between 0 and 1 showing “partial membership of suitability” [
44].
Although the use of fuzzy logic for geospatial analysis offers more realistic results for decision-making than the Boolean method [
56], combining these two methods would utilize the capabilities of both of them to provide a strong analytical tool for site selection studies using GIS. This research intends to develop a comprehensive GIS-based Boolean-fuzzy model for site selection of solar power plant in semi-arid regions of central Iran. Using such approach for PV siting has not been widely studied and a research gap is lying there. Therefore, this research can shed some light on the way the Boolean and fuzzy logics can be utilized together for solar power plant site selection.
There are great potentials for PV utilization in central and southern regions of Iran. Some studies have investigated and quantified these potentials. Khorasanizadeh and Mohammadi [
57] predicted daily global solar radiation in four cities of Iran: Tabass, Isfahan, Kerman, and Bandarabass, which are located in sunny regions. To assess the potential of solar energy in Iran, another study has been carried out by Besarati et al. [
58] through which they generated several solar maps of Iran and also, they investigated the viability of PV power plants in 50 Iranian cities. Based on the results of their work, Arak, the capital of Markazi Province, was identified as a suitable place for the installation of PV power plants. Estimated solar radiation in Iran is about 1800–2200 kWh/m
2 which is higher than global average [
59]. In terms of receiving solar radiation, Iran is a good area for deployment of solar power plants. Thus, it is essential to identify zones where are optimal locations for the PV development by considering economic, environmental, and technical criteria. The comprehensive framework used in this research may be helpful for solar energy planning and policy making in other semi-arid regions.