Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.
Advertisement
Scientific Reports volume 15, Article number: 43603 (2025) Cite this article
1077 Accesses
Metrics details
Developing photovoltaic (PV) projects in coal mining subsidence areas represents a strategic pathway to improving land use efficiency and accelerating the transition to renewable energy. Nevertheless, the siting of such projects entails complex challenges arising from climatic, geological, economic, and policy-related constraints. This study establishes a comprehensive evaluation framework comprising 20 key indicators and applies a fuzzy DEMATEL–ISM approach to an empirical case in Shanxi Province, China. The findings reveal that economic cost factors—specifically the levelized cost of energy (F12), payback period (F13), and operation and maintenance costs (F11)—serve as primary constraints in site selection. In contrast, climatic variables such as the frequency of extreme weather events (F15) and the number of dusty days (F5) function as fundamental driving forces that exert indirect influence by increasing equipment vulnerability, raising operation and maintenance requirements, and ultimately elevating energy production costs. Land use characteristics and grid accessibility are identified as high-level decision-making factors. The main contribution of this study lies in the innovative application of the fuzzy DEMATEL–ISM method to PV siting in subsidence-prone areas, which reveals a multi-level causal structure linking climatic, economic, and spatial dimensions. Based on these insights, the study proposes targeted policy recommendations, including differentiated financial incentives, the adoption of weather-resistant photovoltaic modules, and upgrades to grid infrastructure, thereby offering scientific support for the sustainable development of energy systems and ecological restoration in resource-dependent regions.
The overexploitation of coal resources has resulted in large-scale surface subsidence, leading to land abandonment and ecological degradation, and posing serious challenges to regional sustainable development. Against the backdrop of the “dual-carbon” strategy, the integration of photovoltaic (PV) development with subsidence land remediation has emerged as a promising approach not only to improve land use efficiency but also to facilitate energy structure optimization and green transformation in resource-based regions¹. Large coal mining subsidence areas—formed by excessive extraction—face significant difficulties in ecological restoration, widespread land resource wastage, and frequent geological hazards, all of which severely hinder the economic and social development of affected regions^2,3. Driven by the “dual-carbon” goal, the combined development of PV power generation and subsidence zone management—aimed at achieving synergistic progress in ecological restoration and renewable energy utilization—has become a key direction in energy structure adjustment^4,5.
In Huainan, Anhui Province, a pioneering floating photovoltaic (PV) power plant was developed atop water bodies formed by coal mining subsidence. This project demonstrates a practical integration of land reclamation and renewable energy generation, serving as a model for synergizing ecological restoration with sustainable energy development. In Shanxi’s Datong coal mining subsidence zone, a large-scale ground-mounted PV power station has been successfully connected to the grid, driving regional energy transition and economic development. Internationally, countries such as Poland and the United States have also implemented PV installations in abandoned mining pits to promote land reuse. These cases demonstrate the practical feasibility of PV development in coal mining subsidence areas⁶. Given the high land demand of PV systems, such subsidence zones provide an ideal spatial carrier for large-scale solar deployment⁷. Building PV power plants in these areas not only optimizes the energy mix but also contributes to ecological restoration and coordinated socio-economic development^8,9. However, due to the complex geological conditions of subsidence zones, comprehensive assessments of safety risks, solar irradiance efficiency, and economic costs are necessary. In particular, scientific site selection faces multidimensional challenges such as policy coordination, technological adaptation, ecological constraints, and economic feasibility. This highlights the urgent need to establish an integrated index system that combines both subjective expert judgment and objective data analysis^10,11. Despite successful pilot projects, the development of PV in coal mining subsidence zones still confronts issues such as unstable geological foundations, high operation and maintenance costs, complex land suitability evaluation, and insufficient alignment between policy and planning frameworks. Existing studies often focus on single dimensions, such as economic viability or engineering adaptability, lacking a systematic analysis of multi-dimensional influencing factors. Consequently, current approaches fall short in providing robust support for scientific PV siting and policy design.
This study employs the fuzzy DEMATEL–ISM method to construct a multidimensional indicator system encompassing climatic, environmental, economic, geological, social, and policy-related factors. Taking Shanxi Province as a representative case, the research systematically identifies the key driving factors and underlying mechanisms that influence photovoltaic (PV) site selection^12,13,14,15. The study not only extends the applicability of multi-criteria decision-making tools from a methodological perspective but also provides scientific evidence to support energy transition and subsidence land remediation in resource-based regions from a practical standpoint.Specifically, this research aims to address the following core questions:(1) What are the major factors affecting PV siting in coal mining subsidence areas, and how can they be scientifically screened? (2) How do these influencing factors interact with one another, and how can their interrelationships be quantitatively assessed?(3) How can the most critical factors be identified, and how can targeted siting strategies be developed based on the analysis results?
Synergistic development of coal mining subsidence area and photovoltaic power generation is an innovative model to alleviate land waste and promote energy transformation^16,17. Such areas, where the surface of the ground is sinking or waterlogged as a result of mining activities, have both idle land resources and new energy development potential, and their photovoltaic utilization can enhance land use efficiency and help resource-oriented cities transition to clean energy^18,19. Under the global energy crisis and low-carbon goals, PV siting research has gradually expanded to non-traditional spaces such as railroads, deserts and subsidence zones, of which coal mining subsidence zones have become the focus of research and practice due to their wide area and urgent need for treatment^20,21.
Coal mining subsidence zones are caused by the loss of support, displacement and subsidence of strata as a result of the extraction of underground resources^22,23. Its treatment includes engineering remediation (underground filling, surface loading, building reinforcement)¹¹ and ecological restoration (revegetation, land reclamation)^24,25. In this context, the construction of photovoltaic in subsidence zones can not only improve the efficiency of land use, but also promote the transition of resource-oriented cities to clean energy cities^26,27. In addition, photovoltaic projects can be combined with soil and water conservation measures to help with ecological restoration^28,29.
The development of photovoltaic in coal mining subsidence areas has been verified for its environmental and economic value in many countries, such as the United States, Poland through photovoltaic systems to achieve land reuse in mining areas, and Saudi Arabia to explore the model of “integration of light storage”^30,31. Such projects not only reduce the cost of ecological restoration, but also feed the regional economy through power generation revenues and promote the green transformation of industries^13,32. In the future, it is necessary to deepen the research on technology suitability, optimize the dynamic matching of photovoltaic and geological conditions, and build a synergistic policy-technology-capital framework in order to accelerate the scale-up of this model in the global mining legacy areas^9,33.
A large amount of land in China’s coal mining subsidence areas is in urgent need of regeneration and utilization, which is highly compatible with the new energy development strategy³⁴. The photovoltaic projects in such areas not only realize power generation by erecting photovoltaic panels, but also combine the planting of shade-loving crops under the panels and the use of cleaning water to moisten the soil and fix the sand, thus forming a composite model of “photovoltaic and ecological restoration”^35,36. Due to China’s vast territory and environmental differences, PV construction in different regions has gradually derived from a variety of establishment methods^37,38. To this end, floating photovoltaic has become an innovative direction, for example, Huainan, Anhui Province, utilized the coalfield sedimentary waters to build the world’s first floating power plant³⁹. Efficient use of water surface space and synergistic development of aquaculture for sustainable fisheries transition^40,41,42. This type of model takes into account the synergy between energy development and ecological economy, and provides new ideas for the intensive use of land resources.
The siting of photovoltaic power plants in coal mining subsidence areas requires comprehensive consideration of multiple factors such as natural conditions, economic costs, security risks and land restoration, and its complexity far exceeds that of conventional projects⁴³. China’s large amount of abandoned coal mining land is highly compatible with the new energy strategy, but the construction process needs to take into account ecological restoration and spatial optimization, coupled with significant regional climatic differences and various construction modes, exacerbating the difficulty of decision-making². The core contradiction focuses on the competition for land resources and geological adaptability, for example, floating photovoltaic (such as the case of Huainan, Anhui Province) can use the subsidence of the waters to reduce the cost of infrastructure, operation and maintenance, but the geological instability is still a threat to the long-term safety, the need to balance the efficiency of sunlight, economy and prevention and control of hidden dangers^35,44.
Existing photovoltaic siting studies mostly focus on economic indicators, ignoring environmental stability and social, political and other hidden factors, and the model is difficult to analyze the non-linear causal relationship between multiple criteria³⁷. Coal mining subsidence area site selection involves multidimensional conflicts in geography, technology, and policy, and comprehensive tools are urgently needed to quantify key factors and harmonize conflicts⁴⁵. Constructing a scientific decision-making framework, integrating expert logic and objective data, not only to avoid subjective bias, but also to optimize the path of new energy development in the mining area, providing theoretical and practical support for the regeneration of abandoned land and energy transformation⁴⁶. A growing body of research has begun to explore the feasibility of utilizing subsidence land in mining areas for photovoltaic (PV) development, focusing on aspects such as economic returns, geological stability, and ecological restoration benefits^47,48,49. However, most of these studies remain confined to a single dimension, lacking a systematic integration of climatic, geological, economic, social, and policy factors. As a result, they fall short in capturing the complex mechanisms arising from the interactions among multiple influencing variables.
Multi-criteria decision-making methods need to be targeted in PV site selection due to differences in characteristics. Hierarchical analysis (AHP) has a clear structure but relies on the experience of a large number of experts and is prone to information bias due to pairwise comparisons^50,51. Although ANP is able to characterize factor associations, the computational complexity is significantly higher^16,25. The best weight method (BWM) improves efficiency by simplifying comparisons, but it is difficult to handle nonlinear relationships⁵². TOPSIS calculations are transparent but ignore factor interactions and VIKOR is susceptible to subjective weights^53,54. Interpretive Structural Modeling (ISM) Intuitive and widely used, but computationally resource intensive⁵⁵. Decision Making Experimentation and Assessment Laboratory (DEMATEL) has become an effective tool for multi-attribute decision making due to the advantages of causality visualization, but its traditional form is overly reliant on expert qualitative judgments and runs the risk of subjectivity⁵⁶. To this end, the improved DEMATEL method incorporating triangular fuzzy numbers quantifies uncertainty by fuzzifying linguistic variables, which can retain expert experience and reduce information loss, and provides more objective decision support for the construction of the index system for photovoltaic siting under complex geological conditions^53,57.In recent years, multi-criteria decision-making (MCDM) methods—such as the Analytic Hierarchy Process (AHP), Analytic Network Process (ANP), Technique for Order Preference by Similarity to Ideal Solution (TOPSIS), and VIseKriterijumska Optimizacija I Kompromisno Resenje (VIKOR)—have been widely applied in PV siting studies to evaluate land suitability and project feasibility⁵⁸. These approaches have provided valuable methodological support for solar energy development. However, existing applications have primarily focused on conventional spaces such as deserts, farmlands, and building rooftops, with limited attention paid to more complex and constrained environments such as coal mining subsidence areas.
Although the integration of DEMATEL and ISM has been widely applied in fields such as safety management and environmental assessment, its systematic application in the context of photovoltaic (PV) siting in coal mining subsidence areas remains unprecedented. This study introduces two major methodological innovations based on the traditional framework:(1) Incorporation of triangular fuzzy numbers: Expert linguistic judgments are transformed into fuzzy quantitative indicators, reducing subjectivity and enhancing the objectivity and robustness of causal relationship identification.(2) Contextual innovation: For the first time, a multidimensional factor system encompassing climate, geology, economy, and policy dimensions is constructed specifically for the unique setting of coal mining subsidence zones. Through the DEMATEL–ISM method, the study reveals a hierarchical interaction pathway among climate, economy, and spatial factors.Therefore, the methodological novelty lies not only in technical enhancement, but also in its deep coupling with a context-specific application scenario.
Therefore, it is imperative to establish a multidimensional and systematic siting framework to scientifically identify key factors and uncover their causal mechanisms. In this study, a comprehensive indicator system—encompassing climatic, economic, geological, social, and policy dimensions—is constructed for the specific context of coal mining subsidence areas. By introducing the fuzzy DEMATEL–ISM method, the study systematically reveals the interaction pathway among climate, economy, and spatial factors, thereby providing a scientific decision-making basis for photovoltaic development and ecological restoration.Aiming at the problems of strong subjectivity and complicated calculation caused by the superposition of models in photovoltaic siting in coal mining subsidence area, this study innovatively introduces the DEMATEL-ISM model, and realizes the synergy between the micro-factor analysis and the macro-level delineation through the construction of the influence matrix and the reachable matrix¹¹. The method was first applied to the coal mining subsidence area scenario, which not only identifies the key influencing factors and their causal relationships, but also constructs a hierarchical influence pathway model⁵⁶. Its combination of DEMATEL’s causal quantification and ISM’s systematic hierarchical advantages provides a scientific decision-making framework for PV siting in large risky areas in the energy transition⁵⁹.
This study is divided into four stages. First, identifying PV siting influencing factors and constructing an index system. Second, the questionnaire is designed by combining expert scoring and triangular fuzzy method to obtain quantitative data and ensure the combination of subjective judgment and objective analysis. Third, DEMATEL-ISM method is used for calculation. DEMATEL is used to construct the direct influence matrix and calculate the total influence matrix, influence degree, centrality and causality to identify the key factors¹⁶. Based on the DEMATEL results, ISM further constructs the accessibility matrix, performs hierarchical decomposition, and forms a multilayer system structure model to reveal the hierarchical relationships among factors. Fourth, through the key factor identification and factor relationship analysis, the research conclusions are refined and optimization strategies are provided for PV siting in subsidence areas(Fig. 1).
Research framework.
Shanxi Province is an important energy base in China, with more than 40% of its 156,700 square kilometers of land in coal mines, and a variety of mineral reserves among the nation’s leading⁶⁰. As a pilot province for the management of coal mining subsidence, the province has formed more than a thousand villages damaged by long-term mining in the subsidence area, and there is an urgent need for land restoration⁶¹. The province has significant potential for photovoltaic development, with an average annual solar radiation of 4770–5500 MJ/m² (peaks in summer)¹⁰. Decentralized and centralized PV are suitable for development in an area of 6,453 and 13,900 square kilometers respectively, with a theoretical annual power generation of 562,000 GWh and a total installed capacity of over 627,000 MW^62,63. Since 2000, the government has been promoting dynamic monitoring and ecological restoration of subsidence zones, which provides a practical scenario for the “PV + ecological management” model with both resource base and policy support⁶⁴. Shanxi Province’s “Mineral Resources Master Plan (2021–2025)” proposes a “dual carbon” strategy and energy reform, whereby the synergistic development of coal and new energy requires simultaneous control of land subsidence and carbon emissions, optimization of the energy structure and enhanced ecological restoration⁶⁵. The subsidence zone in Shanxi is still characterized by geological hazards and constraints on sustainable development, and its carbon–neutral photovoltaic (PV) siting is both typical and practically valuable, and urgently needs in-depth study.From a macro perspective, China’s path toward carbon neutrality exhibits significant regional differences in emission characteristics.⁶⁶ found that national carbon emissions are driven by industrial structure and regional disparities, while Huang, et al.⁶⁷ revealed the evolving spatial network of land-use carbon emissions in Jiangxi Province. In contrast, Shanxi’s subsidence areas face the dual pressures of energy extraction and ecological degradation, making them a representative region for exploring the synergy between low-carbon transition and land restoration (Fig. 2).
Location of the study area(Map generated in ArcGIS Pro 3.31 (Esri, https://www.esri.com/arcgis) using the Esri Topographic (No Labels) basemap. Basemap sources: Esri, FAO, NOAA, USGS, EPA, NPS, © OpenStreetMap contributors, and the GIS User Community.Administrative boundaries were obtained from the Geospatial Data Cloud of the National Geomatics Center of China (https://www.gscloud.cn). All thematic layers and symbols were created by the authors.)
This paper utilizes a fuzzy DEMATEL-ISM method based on the influencing factors of PV siting. It first establishes a set of PV siting influencing factor indicator system, combines expert scores and quantifies fuzzy information, then analyzes the importance of each influencing factor according to the DEMATEL method, and combines it with the ISM method to derive the mechanism of action between the influencing factors, which will help decision makers to design targeted policies, and therefore adopts the method suitable for DEMATEL-ISM to design the matrix scale.
This study initially identified 35 potential influencing factors through an extensive literature review, covering aspects such as climatic conditions, environmental constraints, geological risks, economic costs, social factors, and policy support. Considering the local context of Shanxi Province and data availability, factors that were difficult to quantify or lacked consistent data—such as public acceptance—were excluded. Subsequently, the Delphi method was employed, involving two rounds of expert evaluation with 15 invited specialists. Indicators with excessively high correlation or low weight—such as overlapping economic and ecological variables—were eliminated. Ultimately, 20 key factors were retained, covering the full lifecycle of PV projects, from site selection and construction to operation and maintenance.
As shown in Table 1, each indicator is accompanied by detailed information on its data source and measurement unit. For instance, climatic indicators—such as solar radiation and average temperature—were obtained from the China Meteorological Administration (units: MJ/m², °C), while geological risk data were derived from geological survey reports issued by the Ministry of Natural Resources (units: meters, categorical levels). Economic indicators—including investment cost and payback period—were sourced from official records of the Shanxi Provincial Energy Bureau (units: RMB/kW, years). Policy-related indicators were extracted from national and provincial energy planning documents and include both qualitative and quantitative descriptions. This systematic documentation of data sources and standardized units ensures the scientific validity and practical applicability of the constructed indicator system.
In summary, the study identified 20 major influencing factors, including solar irradiation (F1), average temperature (F2), relative humidity (F3), sunlight clarity index (F4), average annual dust days (F5), land use and cover (F6), carbon emission reduction (F7), distance to urban areas (F8), grid infrastructure (F9), investment costs (F10-F13), geology and climate risk (F14-F18), grid absorption capacity (F19), and government subsidies (F20), as shown in (Table 1).
In this study, for the 20 factors affecting PV siting in coal mining subsidence areas, 20 experts in the field of photovoltaics in Shanxi Province were invited (10 valid samples with more than 8 years of experience were eventually retained) to construct a direct influence matrix, and a 0–4 scale was used to assess the intensity of the role between factors. The expert selection mechanism was as follows: a total of 20 experts with long-term experience in photovoltaic development, energy planning, and mining area management were invited to participate in the initial screening process. Ultimately, 10 valid questionnaires were retained, all from individuals with more than eight years of relevant professional experience. To minimize disciplinary bias, the expert panel was composed of scholars from multiple fields, including geotechnical engineering (3 experts), climatology and environmental science (2 experts), energy economics (3 experts), and policy and planning studies (2 experts). This interdisciplinary composition ensured a comprehensive and multi-dimensional understanding of PV siting in coal mining subsidence areas. Through the pre-survey released a few questionnaires to invite experts to modify the indicator system, and feedback preliminary matrix and experts to discuss and reach a consensus, to ensure the reliability of the data⁴⁶. The questionnaire design followed the principle of a diagonal of 0. After excluding seven inexperienced samples, professional group scores were used to quantify the causal associations and to support the scientific analysis of the subsequent DEMATEL-ISM models.
This study ultimately incorporated 10 valid expert questionnaires. Given the nature of multi-criteria decision-making (MCDM) and studies focused on identifying interactions among influencing factors, the representativeness and consistency of expert input are more crucial than the sample size itself⁶⁸. To mitigate potential biases from a small sample and improve judgment quality, several measures were adopted. First, the expert panel was composed of individuals from diverse disciplines—including geotechnical engineering, climate and environmental sciences, energy economics, and policy and planning—in order to reduce systemic bias arising from a single disciplinary perspective. Second, a two-round Delphi process was conducted to iteratively refine assessments of interaction strengths among factors through anonymous feedback and revisions, thereby fostering consensus and reducing the influence of outlier opinions. Third, the consistency and reliability of expert evaluations were tested using Kendall’s W for inter-rater agreement and Cronbach’s α for internal consistency, with commonly accepted thresholds (e.g., W and α values close to or exceeding 0.70) employed to validate the stability and coherence of the scoring. In addition, three types of sensitivity tests were carried out to assess the robustness of the findings: (1) a leave-one-out analysis, in which each expert was removed in turn and the rankings of causality, influence, and centrality were recalculated to examine changes in the set of key factors; (2) weight perturbation, where slight adjustments were made to expert weights and to the mapping intervals of triangular fuzzy numbers (e.g., modifying the right endpoint of “high impact” from 1.00 to 0.95 or 0.90, or adjusting from 0.75–1.00–1.00 to 0.80–1.00–1.00) to evaluate the stability of rankings; and (3) subgroup consistency testing, in which experts were grouped by discipline (engineering, climate, economics, policy) and the rank-order correlation between subgroup results and the full sample was assessed. Results across all tests indicated that the identification of key influencing factors was not sensitive to individual experts, weight variations, or fuzzy number interval settings, thereby confirming the robustness and reproducibility of the conclusions.
The data used in this study primarily cover the period from 2020 to 2023 to ensure temporal consistency across all variables. Climatic and environmental indicators—such as solar radiation and the number of dusty days—were derived from annual average data provided by the China Meteorological Administration and the Shanxi Provincial Environmental Monitoring Center, thereby minimizing the impact of outlier years. Economic data, including the levelized cost of energy (LCOE) and investment costs, were obtained from projects officially registered with the Energy Bureau in 2022, ensuring alignment with current market conditions. Policy-related indicators were based on two key documents that remained effective throughout the study period: the 14th Five-Year Plan for Energy Development of Shanxi Province and the 14th Five-Year Plan for National Renewable Energy Development. By restricting the data to a defined time window and referencing currently valid policy documents, the study aims to maintain both data consistency and policy relevance.
To avoid potential redundancy in the indicator system that could lead to biased model estimations, Pearson correlation analysis was conducted on the 20 indicators (Table 2). The results show that most correlation coefficients are below 0.7, indicating a generally low level of inter-indicator correlation and the absence of significant linear dependencies. Therefore, the indicator system does not exhibit notable multicollinearity and is capable of independently reflecting the characteristics of each dimension.
Triangular fuzzy number (math.) (widetilde{N}) can be represented by the ternary ((l,m,r)), Its affiliation function ({mu }_{widetilde{N}}(x)) It can be expressed as:
where (l) and (r) denote the lower and upper limits of the fuzzy number, respectively, (m) is the most probable value, and the larger the value of (r-l) indicates the greater the fuzziness of the fuzzy number.
When scoring the expert questionnaire, the linguistic evaluation can be converted to a triangular fuzzy number through Table 3.
CFCS method
1) Standardization
2) Calculate the normalized values for the left and right sides
3) Calculation of total standardized values
4) Calculate the defuzzification value of the first expert evaluation
5) Combine the evaluations of individual experts to obtain the defuzzified direct impact matrix
After obtaining the direct influence matrices for all the experts’ defuzzification, the DEAMTEL calculation can be performed.
When constructing the normalized influence matrix, the row maximum value method is chosen as the benchmark. The sum of each row of the Z matrix is first solved, the maximum value is taken, and then each element of the Z matrix is divided by that value to obtain the normalized matrix B.
Integrated impact matrix. The integrated system matrix reflects the combined effects of the influences among the various elements in the system.
where (I) is the unit matrix.
Calculate the degree of influence, degree of being influenced, degree of centrality, degree of cause and weight of each element.
The degree of influence is defined as the sum of the elements of each row in the matrix (T), which measures the overall degree of influence of the element represented by each row on all other elements, expressed in ({D}_{i}).
The degree of influence is the sum of the elements of the columns of the matrix T, reflecting the extent to which the elements represented by each column are influenced by the combination of all the other elements, expressed in ({C}_{i}).
Centrality is a measure of the centrality of a factor in the evaluation system and its importance, calculated as the sum of the factor’s influence and influenced degrees, expressed in ({M}_{i}).
The degree of cause is obtained by subtracting the degree of influence and the degree of being influenced by a certain element, expressed in ({R}_{i}).
The centrality is normalized to obtain the weights of the indicators.
Plotting cause and effect diagrams. Plot the causality graph by taking the center degree as the horizontal coordinate and the cause degree as the vertical coordinate.
(1) Compute the reachability matrix. Overall impact matrix (H=T+I), (T) for the integrated impact matrix, (I) is the unit matrix. Let the number less than the threshold (lambda) in the matrix (H) be 0, Other 1, Obtain the reachable matrix (F).
(2) Compute reachable sets, antecedent sets and intersection sets.
Reachable set ({R}_{i}={{f}_{i}|{F}_{ij}=1}). The factor corresponding to the column with a value of 1 in each row indicates the set of all factors that can be reached by departing from that factor.
Antecedent set ({S}_{i}={{f}_{i}|{F}_{ji}=1}). The rows in each column with a value of 1 correspond to factors that indicate the set of all factors that can reach that factor.
Then compute the intersection ({R}_{i}cap {S}_{i}) .
(3) Factor stratification. There are two algorithms for references:
Once a factor satisfies its condition ({a}_{i}) and when it satisfies its condition (R({a}_{i})=R({a}_{i})cap S({a}_{i})), Indicates that ({a}_{i}) is the highest level factor. Remove the rows and columns corresponding to factor ({a}_{i}) from the reachability matrix (M), and then count the new reachable sets, antecedent sets, and intersection sets until all factors have been divided to form the final factor hierarchy.
When a factor ({a}_{i}) satisfies its (R({a}_{i})=R({a}_{i})cap S({a}_{i})), it indicates that (R({a}_{i})) is the top level factor. Delete the rows and columns corresponding to factor (R({a}_{i})) from the reachable matrix (M), and then calculate the new reachable set, antecedent set, and intersection set, filtering layer by layer until the factors are all stratified to form the final factor stratification.
The evaluation data given by all experts are quantified by using the formula with the triangular fuzzy number conversion table, and then they are aggregated and defuzzified by the corresponding rules and given equal weights to all experts. The aggregated direct influence matrix and the calculated comprehensive influence matrix are shown in the table respectively. obtained as a direct impact matrix after deblurring (Table 4), with larger values representing larger impact magnitudes⁴ (Table 5).
The traditional DEMATEL categorizes influencing and non-influencing factors based on thresholds given by experts, which is arbitrary and lacks scientific basis. In this study, the DEMATEL method was improved by using the mean + standard deviation of the combined influence matrix to determine the threshold value⁵. At the same time, the interrelationships between the factors were graded from “no influence” to “high influence” based on threshold values. The significance of the combined influence matrix is the value of the importance of the role of the two factors in the system, with larger values indicating a higher status of the two factors in the system(Table 6)⁶⁹. It is clear that the F5 and F12 roles have the highest magnitude of importance in the system, followed by F5 and F14 (Table 5) .
The factors that are influenced to a greater extent are F15, F5, F4 and F7(Table 7) . they are more influenced by the other factors. The factors that are influenced to a greater extent are F11, F12 and F13. these factors are more influenced by other factors. The three highest ranked factors in the category of Awareness are F12, F13 and F11, where the higher the value the higher the importance. The cause set is {F1, F2, F3, F5, F7, F15, F16, F17} indicating that these factors have higher power over other factors in the system. The set of outcome factors is {F4, F6, F8, F9, F10, F11, F12, F13, F14, F17, F18, F19, F20}. The highest ranked weight is F12.
Combining the two indicators of centrality and causality to get the causal diagram (Fig. 3). The split value of horizontal and vertical coordinates are center degree and cause degree, respectively, combining the values of the two indicators can be drawn quadrant diagram, divided into four quadrants, according to the quadrant can be seen: F4, F5, F7, F15 of the center degree and cause degree are high, that is, the element is of high importance and for the cause elements; F1, F2, F3, F16, F17 of the center degree is low and the cause degree is high, that is, the element is of low importance and for the cause elements; F8, F9, F18, F19, and F20 indicate low centrality and low cause, i.e., the element is of low importance and is a result element; F6, F10, F11, F12, and F13 have high centrality and low cause, i.e., the element is of high importance and is a result element.
Cause and effect diagram.
The key to expressisng the DEMATEL methodology as ISM is to transform the integrated impact matrix into a reachability matrix (Table 8). The reachability matrix indicates whether there is a “reachability” relationship between elements in the system, i.e. whether there is a “path” between two elements, with the number 1 indicating that there is a path between two elements and the number 0 indicating that there is no path between two elements³. By row, the numbers indicate the impacts generated by the other elements of each element; in addition to the paths generated by itself, the first row, F1, has an impact on F6, F7, F10, F12, and F13; the second row, F2, has an impact on F11, F12, and F13; and the third row, F3, has an impact on F11, F12, and F13 (Table 8) . By column, the numbers indicate that other elements can have an impact on that element, e.g., F11, F12, F13 have an impact on F6.
After obtaining the reachability matrix, the reachable set and the antecedent set can be calculated separately. The reachable set is the set of elements that can be influenced by other elements, which is the number 1 when organized by rows¹¹. The reachable sets of F2 are F2, F11, F12, and F13(Table 9) .The antecedent set refers to the set of other elements that can have an effect on the current element, which is the set of the number 1 when it is listed by columns.The antecedent set of F2 is F2. Next, the intersection of the reachable set and the antecedent set is computed, and the intersection of F2 is F2.
Based on the derivation of the reachable set and antecedent set of each factor from the reachability matrix, the 20 factors affecting PV siting are divided into four layers, revealing the recursive relationship between the layers. The first layer of factors has high external dependence but weak driving effect, which needs to be linked to the underlying fundamental factors for intervention, and a single governance is easy to be superficial; the middle layer has the dual attributes of results and causal factors, which can be used as a short-term optimization target. The hierarchical analysis shows that the intermediate level strategy should connect the top-level phenomenon and the bottom-level motivation, and combine the medium- and long-term design to balance the treatment of symptoms and root causes, so as to avoid the limitations of isolated decision-making. The last layer is the root cause or core driver of the system, which usually has a high degree of causality, is more independent, and is less influenced by other factors, but has a profound impact on the system as a whole^13,52. The four levels of the factor are L1 = {F6, F8, F9, F10, F11, F12, F13, F17, F18, F19}, L2 = {F2, F3, F7, F14, F16, F20}, L3 = {F1, F5, F15}, L4 = {F4}(Fig. 4) .
Factor grading.
To evaluate the impact of triangular fuzzy number interval settings on the results, a multi-scenario sensitivity analysis was conducted. A total of 45 experimental scenarios were designed by varying the disturbance width (w = 0, 0.05, 0.10), α-cut levels (0.0–1.0), and the interval point selection parameter (p = 0.0, 0.5, 1.0), in order to comprehensively test the robustness of the direct-relation matrix. The results indicate that the rankings of key factors based on their total influence (D + R) and causality (D − R) remain highly consistent across all scenarios. The Top 5 indicators consistently appeared in all runs with a 100% frequency, and the causal grouping remained essentially unchanged (Table 10). These findings strongly validate the robustness of the research conclusions with respect to fuzzy interval settings, enhancing both the methodological reliability and the credibility of the results.
Building upon the existing analytical results, this study further validates the findings by referencing actual PV projects in Shanxi Province. For example, the Datong coal mining subsidence PV station and the comprehensive remediation project in the Yangquan subsidence zone illustrate the practical applicability of the model. In the early stages of project implementation, government subsidies (F20) played a pivotal role in lowering the threshold for initial investment, serving as a key enabling factor. However, during the long-term operational phase, levelized cost of energy (F12), payback period (F13), and operation and maintenance costs (F11) emerged as the dominant constraints determining the project’s economic sustainability. These findings indicate that policy incentives are critical in the initiation phase, while economic factors take precedence in the long-term performance of PV projects—consistent with the causal chain established in this study.
According to the DEMATEL-ISM model, the various factors affecting PV siting within the system constitute a complex multi-layered hierarchical structure. First, land use and coverage (F6), distance to urban areas (F8), distance to power lines and substations (F9), maintenance cost (F11), levelized energy cost (F12), payback period (F13), soil texture (F17), slope (F18), and grid capacity (F19) are the most direct influences on the entire PV siting system, which reflects the fact that the economic influences are the most significant in the PV construction process. The economic influencing factors in the process are the most numerous, i.e., maintenance cost, levelized energy cost, and payback period. Therefore, in the development of PV projects in subsidence zones, it is important to take economic evaluation as a primary consideration to ensure the smooth promotion of construction². In the second tier, average temperature (F2), relative humidity (F3), reduction of carbon emissions (F7), degree of settlement deformation (F14), soil reliability (F16), and government subsidy guidance (F20) are the next most important direct factors. According to the order of centrality, F9 occupies a more important position in the siting system. It can be seen that the initial investment of capital can have a considerable impact on enhancing the efficiency of growth. The third level also has three factors, namely solar irradiation (F1), the average annual number of dust days (F5), and weather extremes (F15), all of which have a direct influence on factor F8 in the second level, and factors F6 and F12 in the first level, while the fourth level has only the sunlight clarity index (F4), which is at the lowest level of the model. They do not interact with each other, but together they influence the levelized cost of energy (F13). These are the most basic structural elements, and as the PV siting industry develops, these four levels of elements are closely related to each other and interact with each other in a complex way.
Compared with existing studies, the results of this research demonstrate both consistency and regional specificity. Previous studies have frequently emphasized the central role of economic factors in PV site selection. For instance, systematic reviews by Rediske et al.¹³ and Spyridonidou & Vagiona⁴⁵ highlight that levelized cost of energy and payback period are almost universally recognized as key determinants in siting decisions across regions. Similarly, this study identifies economic cost factors (F12, F13, F11) as core constraints. However, it also reveals that climate-related risk factors—such as the number of dusty days (F5) and extreme weather events (F15)—carry greater importance, a finding that diverges from studies focused on more conventional settings such as deserts or rooftops.
The causal relationship between factors can be determined by the degree of causation, the degree of influence, and the degree of being influenced. Among the causal factors, the five factors with the highest scores in the degree of causality are extreme weather extremes (F15), average annual number of dust days (F5), reduction of carbon emissions (F7), sunshine clarity index (F4), and solar irradiation (F1). The five factors with the highest scores at the bottom of the hierarchy are levelized energy cost (F12), payback period (F13), maintenance cost (F11), land use and cover (F6), and initial investment (F10) are the fundamental factors affecting site selection. This is because the increase in energy demand is due to environmental factors on the one hand. On the other hand, the adverse effects of extreme weather account for a significant portion of the total and are categorized as risk factors³. In addition, the sunlight clarity index is also a key consideration for site selection, with the three factors with the highest impact scores corresponding to the three factors with the lowest impact scores resulting in the lowest impact scores being the distance to the urban area (F8) and the distance to power lines and substations (F9). These two factors are located at the surface level of the hierarchy diagram and have a direct impact on site selection. Slope (F18), grid consumption capacity (F19), and government subsidy guidance (F20) are all in Quadrant III, indicating that they have a higher level of popularity and are more critical factors. In addition, in Quadrants II and IV, land use and coverage (F6), initial investment (F10), maintenance cost (F11), levelized cost of energy (F12), and payback period (F13) are of higher importance and are key factors influencing PV site selection. The factors in quadrants I and III are carbon emission reduction (F7), government subsidy guidance (F20), etc., respectively. These factors are classified in the middle of the hierarchy, both influencing and being influenced by other factors. Although these factors are not ranked high in the middle of the hierarchy, they should not be ignored as they play a transitional role.
Furthermore, the DEMATEL–ISM analysis enables the identification of key causal pathways among influencing factors, revealing a structured “factor–pathway–outcome” mechanism. For example, one prominent causal chain begins with extreme weather events (F15), which can directly cause physical damage to photovoltaic modules. This damage increases operation and maintenance costs (F11), subsequently raising the levelized cost of energy (F12), and ultimately results in a prolonged investment payback period (F13). An indirect pathway starts with extreme weather leading to intensified subsidence deformation (F14), which necessitates additional foundation reinforcement investment (F10), thereby extending the payback period.In the case of dusty weather conditions (F5), frequent dust accumulation on photovoltaic panels reduces their power generation efficiency. This decrease in performance lowers electricity revenue and similarly leads to a longer payback period. In contrast, the benefits associated with carbon reduction (F7) can improve project feasibility by enhancing eligibility for policy subsidies (F20). These subsidies help offset initial investment costs (F10), thereby mitigating economic barriers.Overall, the results indicate that climatic and geological risks predominantly influence photovoltaic siting decisions through their impact on financial metrics, whereas policy interventions serve a moderating role by reducing investment burdens. This reflects a fundamental logic in siting decisions: climatic and geological drivers influence outcomes through economic constraints, which can be adjusted or offset through policy support and technological adaptation. Such findings not only clarify the dynamic interactions among environmental, economic, and policy variables but also offer a theoretical basis for designing more effective and targeted policy instruments.
This divergence is closely related to the regional characteristics of Shanxi Province. On the one hand, Shanxi is a typical resource-dependent region with a long-standing reliance on the coal industry, facing substantial pressure for economic transformation; hence, economic cost constraints are particularly prominent in site selection decisions. On the other hand, the geological environment of subsidence areas is fragile, and the region frequently experiences sandstorms and extreme weather, making the indirect impacts of climatic factors on photovoltaic operation more significant. For example, frequent dust events lead to reduced module efficiency and increased cleaning frequency, thereby elevating operation and maintenance costs (F11); extreme weather intensifies ground subsidence (F14), resulting in greater demand for foundation reinforcement and ultimately prolonging the investment payback period (F13). Consequently, the ranking of key factors in Shanxi differs from that in other regions, reflecting the crucial role that local context plays in shaping photovoltaic siting mechanisms.
This study reveals the key roles of climate adaptation, economic feasibility and infrastructure support in PV siting in subsidence zones, and puts forward corresponding policy recommendations to enhance the overall benefits of PV projects and promote the optimization of energy structure and sustainable regional development.
First, the policy should optimize the economic incentive mechanism, the levelized cost of energy (LCOE) and payback period as the core consideration for PV approval, and the establishment of regional differentiation of the subsidy system, to provide additional operation and maintenance subsidies for extreme weather-frequent areas to reduce the cost of environmental impacts⁸. Second, in terms of environmental adaptability, the application of weather-resistant PV modules should be guided and the design of PV power plants should be optimized in order to enhance the ability to resist extreme weather¹⁶. In addition, the “ecological restoration + photovoltaic” model should be implemented in subsidence zones, combined with vegetation restoration and soil treatment to improve land-use efficiency and realize the dual benefits of ecological management and new energy development. In terms of land use coordination, PV priority development zones should be delineated to reduce conflicts with urban areas and land around substations, and legislation should be adopted to ensure the rationality of PV project sites.
In addition, the policy level also needs to build an intelligent policy support system, improve grid consumption capacity, require PV projects to be supported by energy storage facilities, and optimize grid infrastructure in remote areas to reduce transmission losses. Further, reform the subsidy policy and set up an “Emission Reduction Contribution Incentive Fund” to provide additional support to projects with significant carbon emission reduction, and to promote synergy between economic benefits and environmental goals. At the same time, establish a dynamic monitoring mechanism, build a national big data platform for PV siting, integrate meteorological, geological and economic data to realize intelligent decision-making on siting, and optimize policy guidance on a regular basis.
Photovoltaic development conditions within Shanxi Province are not homogeneous, as significant differences exist across prefecture-level cities in terms of geological, climatic, and economic characteristics⁷⁰. For example, in the eastern regions (such as Yangquan and Jinzhong), coal mining subsidence is severe and land stability is relatively poor, necessitating greater investment in foundation reinforcement and ecological restoration during photovoltaic project development. In the northern areas (such as Datong and Shuozhou), solar radiation is abundant, but frequent sandstorms and extreme weather impose higher demands on component durability and operation and maintenance efforts. In contrast, the southern regions (such as Changzhi and Jincheng) enjoy higher average solar radiation intensity and relatively well-developed grid infrastructure, offering more favorable conditions for energy integration. These regional disparities suggest that photovoltaic siting strategies should be tailored to local conditions: reinforcing ground stability and ecological measures in the east, enhancing wind and sand protection along with O&M subsidies in the north, and focusing on economic returns and market-oriented operation in the south.
This study reveals the critical roles of climate adaptability, economic feasibility, and infrastructure support in the site selection of photovoltaic (PV) projects in mining subsidence areas⁷¹. In light of the regional differences within Shanxi Province, more targeted policy recommendations can be proposed. In the northern region (such as Datong and Shuozhou), where sandstorms are frequent and extreme weather events are common, it is advisable to prioritize the development of weather-resistant PV modules and increase subsidies for operation and maintenance. In the central region (such as Taiyuan and Yangquan), although grid infrastructure is relatively well-developed, grid curtailment remains a challenge, which calls for accelerated deployment of energy storage systems and grid expansion. In the southern region (such as Changzhi and Jincheng), where geological stability and topographic constraints are more prominent, policy subsidies play a more decisive role in promoting project implementation. Therefore, the subsidy mechanism should reflect regional differentiation to enhance the feasibility and attractiveness of PV development.
The academic value of this study lies not only in the use of the DEMATEL–ISM method to construct a multi-layered causal model, but also in its contextualized expansion of photovoltaic site selection in coal mine subsidence areas. Unlike Wu & Lee (2007), who applied the method in a management science context, this research integrates DEMATEL–ISM with fuzzy set theory under the background of energy transition and land reclamation, revealing how climatic, economic, and geological conditions interact and ultimately affect siting outcomes. This interdisciplinary and cross-domain methodological innovation expands the application boundaries of DEMATEL–ISM and provides new theoretical and methodological support for renewable energy site selection in high-risk areas.
Compared with existing studies, the findings of this research demonstrate both commonalities and distinct innovations. Previous literature has largely emphasized the pivotal role of economic feasibility in photovoltaic site selection. For instance, Rediske et al.¹³, through a systematic review, pointed out that levelized cost and payback period are nearly universal core factors in PV siting decisions, while Spyridonidou and Vagiona⁴⁵ also identified economic constraints as the most decisive factor in the siting of large-scale PV projects across Europe. This study is consistent with these conclusions, confirming that economic factors (F12, F13, F11) constitute core constraints. However, it further reveals that in the specific context of coal mine subsidence areas, climatic and geological risk factors—such as the number of dust days (F5) and extreme weather events (F15)—play an amplified role within the causal chain. This indicates that the study not only verifies the significance of economic considerations, but also innovatively uncovers the interlinkages between climate–geological conditions and economic costs in subsidence-affected regions.
This divergence stems from regional characteristics: compared to conventional sites, the geological environment of subsidence areas is more fragile, and climatic risks are more readily transformed into economic costs, thereby rendering the “climate–economy” causal chain a critical determinant of project success or failure. This study is the first to apply the DEMATEL–ISM method to photovoltaic site selection in coal mine subsidence areas, constructing a multi-layered causal model and verifying the causal pathways of region-specific factors, thus filling a research gap in the related field.
This study integrates DEMATEL and ISM methods, replaces the adjacency matrix with a comprehensive influence matrix, and combines factor association strength and hierarchical division to construct a four-layer influence factor system for photovoltaic siting in coal mining subsidence areas, which fills in the blanks of systematic research in the risk area²². At the theoretical level, it reveals for the first time the non-linear coupling mechanism of climate, economic and land restoration factors, forming a multidisciplinary decision-making framework for site selection; at the practical level, the proposed hierarchical model takes into account both short-term strategies (optimizing intermediate pivotal factors) and long-term paths (linking underlying drivers), providing quantitative tools for the government and enterprises to balance ecological restoration, security risks and energy transition, and promoting sustainable regeneration and low-carbon development of mining areas.
The present investigation covers both DEMATEL and ISM analytic processes, and the interactions between the influencing factors need to be assessed. However, too many factor judgments may burden the respondents and affect the accuracy of the assessment. In addition, since the global PV development is still in the rapid evolution stage, a mature theoretical framework for site selection has not yet been formed. Therefore, this study mainly refers to existing literature, known cases and public data, but may have some limitations due to unpublished information and the fact that the long-term impacts of some projects are not yet clear.
This study incorporates actual photovoltaic project cases in cities such as Datong and Yangquan in the discussion section to enhance regional relevance⁷². However, the overall number of cases remains limited, failing to fully capture the heterogeneity across all prefecture-level cities in Shanxi Province. As such, the conclusions mainly reflect patterns observed in typical regions. Future research should integrate more sub-regional data and real-world projects to conduct cross-regional comparative analyses, thereby improving the generalizability and practical applicability of the model. Although the expert panel sought to ensure interdisciplinary representation—including geotechnical, climatic, economic, and policy-related perspectives—the limited sample size (n = 10) constrained the comprehensiveness of disciplinary coverage. For example, assessments related to social acceptance and long-term policy stability may have been influenced by individual expert experiences. Future studies should consider expanding the expert pool and applying consistency tests, such as Cronbach’s α or the Kappa coefficient, to further enhance the robustness of the model. In addition, while this study controlled the temporal range of data (2020–2023) and selected valid policy documents within the study period, it is important to note that climate data are subject to long-term variability, and the policy landscape may change in the future⁷³. Therefore, the findings primarily reflect the conditions of a specific timeframe, and future research should incorporate dynamic monitoring data and policy scenario simulations to improve the temporal relevance and extrapolation capacity of the conclusions.
Based on the DEMATEL-ISM method, this study systematically analyzes 20 key influencing factors and their hierarchical relationships for PV siting in coal mining subsidence areas. The results show that: climatic environmental factors (e.g., extreme weather, number of dusty days, sunshine clarity index) constitute the bottom-level base driver, which determines the regional resource endowment; economic cost factors (levelized energy cost, payback period, and maintenance cost) act as the core constraints, which directly affect the project feasibility; and land use and infrastructure accessibility are located in the high-level hierarchy, which become the final decision variables, and are subject to the climatic and economic The study constructs a multilevel causal model. The multilevel causal model constructed in the study reveals the recursive logic of “resource-economy-space”, suggests that policy incentives should be used to balance climate adaptation, land restoration benefits and short-term economic gains, and suggests the integration of dynamic data analysis and GIS technology to optimize the accuracy of site selection. The framework provides interdisciplinary decision-making support for the energy transition of coal-sinking areas, and helps ecological restoration and low-carbon industrial restructuring under the goal of “dual-carbon”.
No datasets were generated or analysed during the current study.
Chen, M. et al. Green land use efficiency and influencing factors of resource-based cities in the yellow river basin under carbon emission constraints. Buildings 12, 551 (2022).
Article  Google Scholar 
Almeida, A. C. L. Multi actor multi criteria analysis (MAMCA) as a tool to build indicators and localize sustainable development goal 11 in Brazilian municipalities. Heliyon https://doi.org/10.1016/j.heliyon.2019.e02128 (2019).
Article  PubMed  PubMed Central  Google Scholar 
Zhao, J. et al. An investigation into the disturbance effects of coal mining on groundwater and surface ecosystems. Environ. Geochem. Health 45, 7011–7031. https://doi.org/10.1007/s10653-023-01658-w (2023).
Article  CAS  PubMed  Google Scholar 
Qin, T., Zhang, C., Ning, D. & Wang, H. A multi-criteria assessment system for the influence of mining on the geological environment: A case study of the Huodong mining area in Shanxi Province. China. All Earth 36, 1–15. https://doi.org/10.1080/27669645.2024.2380223 (2024).
Article  Google Scholar 
Shang, H. et al. Surface environmental evolution monitoring in coal mining subsidence area based on multi-source remote sensing data. Front. Earth Sci. https://doi.org/10.3389/feart.2022.790737 (2022).
Article  Google Scholar 
Mert, Y. Contribution to sustainable development: Re-development of post-mining brownfields. J. Clean. Prod. 240, 118212 (2019).
Article  Google Scholar 
Bai, W. et al. Assessment of the potential of salt mines for renewable energy peaking in China. Energy 300, 131577. https://doi.org/10.1016/j.energy.2024.131577 (2024).
Article  CAS  Google Scholar 
Mansouri Kouhestani, F. et al. A comprehensive assessment of solar and wind energy potential at the University of Lethbridge campus, a medium-sized western Canadian university. Int. J. Green Energy 16, 1246–1263, https://doi.org/10.1080/15435075.2019.1671400 (2019).
Hasti, F., Mamkhezri, J., McFerrin, R. & Pezhooli, N. Optimal solar photovoltaic site selection using geographic information system—based modeling techniques and assessing environmental and economic impacts: The case of Kurdistan. Sol. Energy 262, 111807. https://doi.org/10.1016/j.solener.2023.111807 (2023).
Article  CAS  Google Scholar 
Luo, X., Dong, C. & Dong, X. How economic transformation influence the employment of resource-based cities: Evidence from Shanxi Province. China. Resources Policy 89, 104553. https://doi.org/10.1016/j.resourpol.2023.104553 (2024).
Article  Google Scholar 
Saura Lacárcel, J. R., Yuan, B. & Fan, M. Research on identifying main factors affecting building carbon reaching the peak based on DEMATEL-ISM method. SHS Web Conf. https://doi.org/10.1051/shsconf/202317002011 (2023).
Article  Google Scholar 
Chen, X., Chen, B., Wang, Y., Zhou, N. & Zhou, Z. Response of vegetation and soil property changes by photovoltaic established stations based on a comprehensive meta-analysis. Land 13, 478 (2024).
Article  Google Scholar 
Rediske, G., Siluk, J. C. M., Gastaldo, N. G., Rigo, P. D. & Rosa, C. B. Determinant factors in site selection for photovoltaic projects: A systematic review. Int. J. Energy Res. 43, 1689–1701. https://doi.org/10.1002/er.4321 (2019).
Article  CAS  Google Scholar 
Sun, J. et al. The energy security risk assessment of inefficient wind and solar resources under carbon neutrality in China. Appl. Energy 360, 122889. https://doi.org/10.1016/j.apenergy.2024.122889 (2024).
Article  Google Scholar 
Wu, Y., Zhang, B., Xu, C. & Li, L. Site selection decision framework using fuzzy ANP-VIKOR for large commercial rooftop PV system based on sustainability perspective. Sustain. Cities Soc. 40, 454–470. https://doi.org/10.1016/j.scs.2018.04.024 (2018).
Article  Google Scholar 
Li, Y. & Saracoglu, B. O. Location and investment factors of hydropower plants. Energy Sources, Part A Recovery, Utilization, and Environmental Effects, https://doi.org/10.1080/15567036.2021.1963015 (2021).
Erixno, O., Ramadhani, F., Rahim, N. A. & Rivai, A. Realistic operation and strategic energy, economic and environmental analyses of a hybrid renewable energy based—Micro combined heat and power. Sol. Energy 272, 112467. https://doi.org/10.1016/j.solener.2024.112467 (2024).
Article  Google Scholar 
Ruan, T., Wang, F., Topel, M., Laumert, B. & Wang, W. A new optimal PV installation angle model in high-latitude cold regions based on historical weather big data. Appl. Energy 359, 122690. https://doi.org/10.1016/j.apenergy.2024.122690 (2024).
Article  Google Scholar 
Souayfane, F. et al. On the behavior of renewable energy systems in buildings of three Saudi cities: Winter variabilities and extremes are critical. J. Build. Eng. 70, 106408. https://doi.org/10.1016/j.jobe.2023.106408 (2023).
Article  Google Scholar 
Akther, A., Ahamed, T., Noguchi, R., Genkawa, T. & Takigawa, T. Site suitability analysis of biogas digester plant for municipal waste using GIS and multi-criteria analysis. Asia-Pacific J. Regional Sci. 3, 61–93. https://doi.org/10.1007/s41685-018-0084-2 (2018).
Article  Google Scholar 
Pimentel Da Silva, G. D., Magrini, A. & Branco, D. A. C. A multicriteria proposal for large-scale solar photovoltaic impact assessment. Impact Assess. Project Appraisal 38, 3–15, https://doi.org/10.1080/14615517.2019.1604938 (2019).
Ohunakin, O. S. & Saracoglu, B. O. A comparative study of selected multi-criteria decision-making methodologies for location selection of very large concentrated solar power plants in Nigeria. Afr. J. Sci. Technol. Innov. Dev. 10, 551–567. https://doi.org/10.1080/20421338.2018.1495305 (2018).
Article  Google Scholar 
Wang, J.-J., Jing, Y.-Y., Zhang, C.-F. & Zhao, J.-H. Review on multi-criteria decision analysis aid in sustainable energy decision-making. Renew. Sustain. Energy Rev. 13, 2263–2278. https://doi.org/10.1016/j.rser.2009.06.021 (2009).
Article  Google Scholar 
Pérez, N. S. & Alonso-Montesinos, J. Economic and environmental solutions for the PV solar energy potential in Spain. J. Clean. Prod. 413, 137489. https://doi.org/10.1016/j.jclepro.2023.137489 (2023).
Article  Google Scholar 
Xia, Y. et al. Combined PS-InSAR technology and high-resolution optical remote sensing for identifying illegal underground mining in the suburb of Yangquan City, Shanxi Province China. Remote Sens. 15, 3565 (2023).
Article  ADS  Google Scholar 
Al Garni, H. Z. & Awasthi, A. in Advances in Renewable Energies and Power Technologies (ed Imene Yahyaoui) 57–75 (Elsevier, 2018).
Chai, S., Kong, F., Liu, Y., Liang, M. & Liu, Q. Photovoltaic solar farms site selection through “policy constraints-construction suitability”: A case study of Qilian county. Qinghai. Land 13, 1420 (2024).
Article  Google Scholar 
Song, W., Song, W., Gu, H. & Li, F. Progress in the remote sensing monitoring of the ecological environment in mining areas. Int. J. Environ. Res. Public Health 17, 1846 (2020).
Article  PubMed  PubMed Central  Google Scholar 
Wen, L. & Diao, P. Simulation study on carbon emission of China’s electricity supply and demand under the dual-carbon target. J. Clean. Prod. 379, 134654. https://doi.org/10.1016/j.jclepro.2022.134654 (2022).
Article  Google Scholar 
Iweh, C. D., Gyamfi, S., Tanyi, E. & Effah-Donyina, E. Assessment of the optimum location and hosting capacity of distributed solar PV in the southern interconnected grid (SIG) of Cameroon. Int. J. Sustain. Energy https://doi.org/10.1080/14786451.2023.2168002 (2023).
Yimen, N., Hamandjoda, O., Meva’a, L., Ndzana, B. & Nganhou, J. Analyzing of a Photovoltaic/Wind/Biogas/Pumped-Hydro Off-Grid Hybrid System for Rural Electrification in Sub-Saharan Africa—Case Study of Djoundé in Northern Cameroon. Energies 11, 2644 (2018).
Wang, K. et al. Evaluating and simulating the impacts of land use patterns on carbon emissions in coal resource-based regions: A case study of Shanxi province China. J. Clean. Prod. 458, 142494. https://doi.org/10.1016/j.jclepro.2024.142494 (2024).
Article  CAS  Google Scholar 
He, T., Xiao, W., Zhao, Y., Deng, X. & Hu, Z. Identification of waterlogging in Eastern China induced by mining subsidence: A case study of google earth engine time-series analysis applied to the Huainan coal field. Remote Sens. Environ. 242, 111742. https://doi.org/10.1016/j.rse.2020.111742 (2020).
Article  Google Scholar 
Zhang, M. et al. A habitable earth and carbon neutrality: Mission and challenges facing resources and the environment in china—An overview. Int. J. Environ. Res. Public Health 20, 1045 (2023).
Article  PubMed  PubMed Central  Google Scholar 
Li, J. Theoretical research on three-dimensional cultivation of poultry and fruit and vegetable based on photovoltaic solar greenhouse under the background of new energy. Int. J. Manag. Edu. Hum. Dev. 3, 1016–1020 (2023).
Google Scholar 
Khazaelpour, P. & Zolfani, S. H. FUCOM-optimization based predictive maintenance strategy using expert elicitation and artificial neural network. Expert Syst. Appl. 238, 121322. https://doi.org/10.1016/j.eswa.2023.121322 (2024).
Article  Google Scholar 
Shi, F., Li, X., Cao, Y. & Bai, B. The feasibility analysis of “ecological photovoltaics” from coal gangue mountains. Sustainability 15, 8761 (2023).
Article  ADS  Google Scholar 
Kosorić, V., Huang, H., Tablada, A., Lau, S.-K. & Tan, H. T. W. Survey on the social acceptance of the productive façade concept integrating photovoltaic and farming systems in high-rise public housing blocks in Singapore. Renew. Sustain. Energy Rev. 111, 197–214. https://doi.org/10.1016/j.rser.2019.04.056 (2019).
Article  Google Scholar 
Bei, Y., Yuan, B., Wang, Z. & Tang, J. in 2022 5th International Conference on Energy, Electrical and Power Engineering (CEEPE). 1126–1130 (IEEE).
Song, C. et al. Application of photovoltaics on different types of land in China: Opportunities, status and challenges. Renew. Sustain. Energy Rev. 191, 114146. https://doi.org/10.1016/j.rser.2023.114146 (2024).
Article  Google Scholar 
Piyatadsananon, P. Spatial factors consideration in site selection of ground-mounted PV power plants. Energy Procedia 100, 78–85. https://doi.org/10.1016/j.egypro.2016.10.135 (2016).
Article  Google Scholar 
Pouran, H. M. From collapsed coal mines to floating solar farms, why China’s new power stations matter. Energy Policy 123, 414–420 (2018).
Article  Google Scholar 
Zhou, D., Chen, T., Ding, H. & Wang, Q. Tracking the provincial energy transition in China: A comprehensive index. Energy 304, 131879. https://doi.org/10.1016/j.energy.2024.131879 (2024).
Article  Google Scholar 
Yilmaz, F., Ozturk, M. & Selbas, R. Development and assessment of a newly developed renewable energy-based hybrid system with liquid hydrogen storage for sustainable development. Int. J. Hydrogen Energy 56, 406–417. https://doi.org/10.1016/j.ijhydene.2023.12.208 (2024).
Article  ADS  CAS  Google Scholar 
Spyridonidou, S. & Vagiona, D. G. A systematic review of site-selection procedures of PV and CSP technologies. Energy Rep. 9, 2947–2979. https://doi.org/10.1016/j.egyr.2023.01.132 (2023).
Article  Google Scholar 
Kara, K. et al. A spherical fuzzy-based DIBR II-AROMAN model for sustainability performance benchmarking of wind energy power plants. Expert Syst. Appl. 253, 124300. https://doi.org/10.1016/j.eswa.2024.124300 (2024).
Article  Google Scholar 
Zhang, Z. et al. Renew mineral resource-based cities: Assessment of PV potential in coal mining subsidence areas. Appl. Energy 329, 120296 (2023).
Article  Google Scholar 
He, T. et al. Harnessing the solar photovoltaic potential in global mining areas: Energy locations for the future. The Innovation (2025).
Xiao, Y., Chen, Y., Huang, Y. & Yan, Y. Surface protection technologies for earthen sites in the 21st century: Hotspots, evolution, and future trends in digitalization, intelligence, and sustainability. Coatings 15, 855 (2025).
Article  CAS  Google Scholar 
Rios, R. & Duarte, S. Selection of ideal sites for the development of large-scale solar photovoltaic projects through analytical hierarchical process—geographic information systems (AHP-GIS) in Peru. Renew. Sustain. Energy Rev. 149, 111310. https://doi.org/10.1016/j.rser.2021.111310 (2021).
Article  Google Scholar 
Seddiki, M. & Bennadji, A. Multi-criteria evaluation of renewable energy alternatives for electricity generation in a residential building. Renew. Sustain. Energy Rev. 110, 101–117. https://doi.org/10.1016/j.rser.2019.04.046 (2019).
Article  Google Scholar 
He, J. et al. Lifecycle risk assessment on the sustainable development of upgrading energy projects using abandoned mines: An ISM-BWM method. Sustain. Energy Technol. Assess. 54, 102833. https://doi.org/10.1016/j.seta.2022.102833 (2022).
Article  Google Scholar 
Jahanshahloo, G. R., Lotfi, F. H. & Izadikhah, M. Extension of the TOPSIS method for decision-making problems with fuzzy data. Appl. Math. Comput. 181, 1544–1551. https://doi.org/10.1016/j.amc.2006.02.057 (2006).
Article  Google Scholar 
Rehman, S., Khan, S. A. & Alhems, L. M. Application of TOPSIS approach to multi-criteria selection of wind turbines for on-shore sites. Appl. Sci. 10, 7595 (2020).
Article  CAS  Google Scholar 
Abdulla, H., Sleptchenko, A. & Nayfeh, A. Photovoltaic systems operation and maintenance: A review and future directions. Renew. Sustain. Energy Rev. 195, 114342. https://doi.org/10.1016/j.rser.2024.114342 (2024).
Article  Google Scholar 
Wang, L., Cao, Q. & Zhou, L. Research on the influencing factors in coal mine production safety based on the combination of DEMATEL and ISM. Saf. Sci. 103, 51–61. https://doi.org/10.1016/j.ssci.2017.11.007 (2018).
Article  Google Scholar 
Sonter, L. J., Moran, C. J., Barrett, D. J. & Soares-Filho, B. S. Processes of land use change in mining regions. J. Clean. Prod. 84, 494–501. https://doi.org/10.1016/j.jclepro.2014.03.084 (2014).
Article  Google Scholar 
Onuoha, H. et al. Integrating GIS and AHP for photovoltaic farm site selection: A case study of Ikorodu. Nigeria. Processes 13, 164 (2025).
Article  CAS  Google Scholar 
Wu, W.-W. & Lee, Y.-T. Developing global managers’ competencies using the fuzzy DEMATEL method. Expert Syst. Appl. 32, 499–507. https://doi.org/10.1016/j.eswa.2005.12.005 (2007).
Article  Google Scholar 
Wang, B. et al. A review of the photothermal-photovoltaic energy supply system for building in solar energy enrichment zones. Renew. Sustain. Energy Rev. 191, 114100. https://doi.org/10.1016/j.rser.2023.114100 (2024).
Article  Google Scholar 
Yi, Z. et al. Long-term Landsat monitoring of mining subsidence based on spatiotemporal variations in soil moisture: A case study of Shanxi Province, China. Int. J. Appl. Earth Obs. Geoinf. 102, 102447. https://doi.org/10.1016/j.jag.2021.102447 (2021).
Article  Google Scholar 
Ma, D. & Zhao, S. Quantitative analysis of land subsidence and its effect on vegetation in Xishan coalfield of Shanxi Province. ISPRS Int. J. Geo Inf. 11, 154 (2022).
Article  Google Scholar 
Wang, W. et al. Unveiling the implementation barriers to the digital transformation in the energy sector using the Fermatean cubic fuzzy method. Appl. Energy 360, 122756. https://doi.org/10.1016/j.apenergy.2024.122756 (2024).
Article  Google Scholar 
Cui, L., Zhang, J., Su, Y. & Li, S. A GIS-based multidimensional evaluation method for solar energy potential in Shanxi Province. China. Energies 16, 1305 (2023).
Article  Google Scholar 
Tan, M., Dong, J., Qu, J. & Hao, M. The patterns of migration of potentially toxic elements from coal mining subsidence areas and associated soils to waterlogged areas. Toxics 11, 888 (2023).
Article  CAS  PubMed  PubMed Central  Google Scholar 
Jia, J., Xin, L., Lu, C., Wu, B. & Zhong, Y. China’s CO₂ emissions: A systematical decomposition concurrently from multi-sectors and multi-stages since 1980 by an extended logarithmic mean divisia index. Energ. Strat. Rev. 49, 101141 (2023).
Article  Google Scholar 
Huang, H., Jia, J., Chen, D. & Liu, S. Evolution of spatial network structure for land-use carbon emissions and carbon balance zoning in Jiangxi Province: A social network analysis perspective. Ecol. Ind. 158, 111508. https://doi.org/10.1016/j.ecolind.2023.111508 (2024).
Article  CAS  Google Scholar 
Srđević, B. & Srđević, Z. Multi-model assessing and visualizing consistency and compatibility of experts in group decision-making. Mathematics 12, 1699 (2024).
Article  Google Scholar 
Feng, X., Li, E., Li, J. & Wei, C. Critical influencing factors of employees’ green behavior: Three-stage hybrid fuzzy DEMATEL–ISM–MICMAC approach. Environ. Dev. Sustain. 26, 17783–17811. https://doi.org/10.1007/s10668-023-03364-0 (2024).
Article  Google Scholar 
Ouchani, F.-Z., Jbaihi, O., Merrouni, A. A., Ghennioui, A. & Maaroufi, M. Geographic information system-based multi-criteria decision-making analysis for assessing prospective locations of pumped hydro energy storage plants in Morocco: Towards efficient management of variable renewables. J. Energy Storage 55, 105751 (2022).
Article  Google Scholar 
Huang, Y., Huang, Y., Chen, Y. & Yang, S. Spatial evolution of traditional waterside settlements south of the Yangtze River and the distribution of settlement heritage: evidence from the Nanxi River Basin. npj Heritage Sci. 13, 62 (2025).
Article  Google Scholar 
Huang, Y., Xia, J., Chen, Y. & Hong, J. Global research progress on rural settlement spatial patterns from 2000–2024: A comprehensive review of hotspots, frontiers, and emerging trends. Front. Earth Sci. 12, 1505570 (2024).
Article  ADS  Google Scholar 
Pan, Z., Chen, Y., Huang, Y. & Zheng, L. The vernacular cultural landscape in traditional villages: Global hotspots, emerging trends, and a case study of China’s Qilu cultural district. Front. Earth Sci. 12, 1511292 (2024).
Article  ADS  Google Scholar 
Download references
Institute of Urban and Sustainable Development, City University of Macau, Macau SAR, China
Yan Lu, Yu Yan, Mengyao Wang & Long Zhou
Search author on:PubMed Google Scholar
Search author on:PubMed Google Scholar
Search author on:PubMed Google Scholar
Search author on:PubMed Google Scholar
Yan Lu: Conceptualization, Data curation, Formal analysis, Visualization, Writing – original draft. Mengyao Wang: Methodology, Validation, Writing – review & editing, Software. Yu Yan: Supervision, Funding acquisition, Conceptual guidance, Project administration, Writing – review & editing. Long Zhou: Supervision, Theoretical guidance, Writing – review & editing.
Correspondence to Yu Yan.
The authors declare no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
Reprints and permissions
Lu, Y., Yan, Y., Wang, M. et al. Study on the influencing factors of photovoltaic siting in coal mining subsidence areas–taking Shanxi Province as an example. Sci Rep 15, 43603 (2025). https://doi.org/10.1038/s41598-025-26672-z
Download citation
Received: 30 June 2025
Accepted: 29 October 2025
Published: 11 December 2025
Version of record: 11 December 2025
DOI: https://doi.org/10.1038/s41598-025-26672-z
Anyone you share the following link with will be able to read this content:
Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative
Advertisement
Scientific Reports (Sci Rep)
ISSN 2045-2322 (online)
© 2026 Springer Nature Limited
Sign up for the Nature Briefing: Anthropocene newsletter — what matters in anthropocene research, free to your inbox weekly.

source