Global solar photovoltaic (PV) deployment has accelerated from less than 200 GW in 2015 to more than 1.6 TW in 2023, accounting for roughly one-third of all new generation capacity worldwide (IRENA and FAO, 2021; Bhadra et al., 2024). Cumulative capacity is projected to exceed 4 TW by 2030 under current national commitments, with China, the European Union, and the United States representing more than 70% of new additions, while neighboring Canada targets a net-zero electricity grid by 2035. This rapid expansion has intensified competition for arable and post-industrial land, prompting interest in dual-use configurations that maintain agricultural function. Agrivoltaic systems—the co-location of PV infrastructure with crop or livestock production—are identified by the Food and Agriculture Organization of the United Nations (FAO) and the International Renewable Energy Agency (2021). As a key strategy to advance both food and energy security while strengthening climate resilience and rural employment. For example, integrating PV with managed agriculture can raise total land-use efficiency by 60%–90%, reduce water consumption through moderated microclimates, and stabilize rural income streams (Gadhiya et al., 2024). Renewable-energy interventions in agri-food systems could further reduce global agricultural greenhouse-gas emissions by up to 10% while providing affordable energy access for nearly 2.8 billion rural people (IRENA and FAO, 2021). Recent agrivoltaic models employ geospatial analysis coupled with fuzzy-logic or multi-criteria decision frameworks to integrate variables such as slope, aspect, land cover, and solar irradiance (Gadhiya et al., 2024; Reher et al., 2025). These approaches routinely achieve land-equivalent ratios above 1.5 under optimized crop–panel configurations, illustrating the ecological and economic potential of renewable-energy generation and food production within shared landscapes.

The decline of coal mining in Appalachia has left the region economically and ecologically vulnerable, with many communities in West Virginia struggling to transition to new economic opportunities (Kent, 2016). In response, hydraulic fracturing (fracking) emerged in the early 2,000s as a potential economic boon due to the region’s location atop the Marcellus Shale, a geologic formation rich in natural gas (Energy Information Administration (US), 2017). However, while natural gas production in West Virginia exceeded projections, the economic benefits have largely bypassed local communities. Despite a 60% increase in economic output from gas-producing counties, job growth failed to materialize, and local communities experienced significant population declines, with some areas losing as much as 11% of their population (O’Leary, 2021). The environmental and health impacts associated with fracking further complicate the situation. The practice has been linked to water contamination, respiratory issues, and increased cancer risks, which exacerbate the public health challenges faced by Appalachian communities (Meng, 2017; Clark et al., 2022).

As many of the natural gas wells drilled during the fracking boom are near the end of their production cycles, West Virginia faces an urgent need to address the decommissioning of these wells. Natural gas wells typically have a lifespan of around 20 years, with many wells established in the mid-2000s now reaching their end (Canadian Association of Petroleum Producers (CAAP), 2022). The decline in production has left over 4,000 wells abandoned in the state (West Virginia Public Broadcasting (WVPB), 2022), with operators often postponing decommissioning due to the high costs and lack of financial incentive to properly close wells (Weber et al., 2015). At the end of a well’s lifespan, natural gas companies are often obligated to employ reclamation techniques, however use of reclamation varies widely and, if wells are not outright abandoned, often is not more than filling in retention ponds and spreading a grass-legume seed mix over the disturbed site (Skousen et al., 2012). Proper well decommissioning, while costly, is necessary to mitigate the environmental risks posed by abandoned wells and to reclaim the land for alternative uses, such as agriculture, recreation, or renewable energy projects. Given the widespread environmental concerns and economic challenges, this paper develops a model for identifying natural gas wells in West Virginia that are approaching or have reached the end of their lifecycle and assessing each well site for agrivoltaic reclamation suitability.

Agrivoltaics—the dual use of land for photovoltaic (PV) electricity generation and agricultural production—has emerged as a land-efficient strategy to meet improved ecological targets while maintaining or even enhancing ecosystem services. In terrain-limited regions like central Appalachia, where steep slopes, legacy degradation, and land fragmentation constrain intensive cropping systems, low-input ruminant livestock systems are often the most viable agricultural strategy. Grazing-based remediation, especially with rotational small ruminant or cattle systems, is particularly well-suited to these marginal lands and can be strategically integrated into utility-scale solar sites to suppress invasive species, recycle nutrients, and reduce vegetation management costs (Barron-Gafford et al., 2019; Toledo et al., 2024). Moreover, co-located systems have been shown to mitigate thermal extremes, increase soil moisture retention, and improve animal welfare and forage productivity in arid and temperate climates (Barron-Gafford et al., 2019; Higgins et al., 2022).

Despite being well-suited for agrivoltaics, Appalachia, overall, lags in agrivoltaic research. DeLong et al. (2023) estimated that current and planned utility-scale PV projects for the Appalachian region of Tennessee require only 0.076% to 0.137% of state farmland, with projected future land demands under Tennessee Valley Authority’s (TVA) solar goals totaling less than 1% of available agricultural land. These findings suggest that agrivoltaic expansion need not compete with core food production and may even help preserve agricultural landscapes under development pressure. As demonstrated in the Mar Menor coastal lagoon in Spain, PV installations can catalyze ecosystem recovery on degraded land, suggesting that agrivoltaic design can extend beyond co-location toward active ecological restoration (DeLong et al. (2023)). Despite these benefits, implementation requires careful planning to ensure compatibility between livestock behavior, PV infrastructure, and site hydrology (Goodrich et al., 2013).

Agrivoltaic site suitability models abound, with most using some combination of pairwise comparison method, analytical hierarchical process, and criterion weights (Reher et al., 2025; Nugroho et al., 2024). Most of these models focus on installation of solar arrays into existing agricultural land, particularly cropping systems (Dinçer et al., 2025; Spyridonidou and Vagiona, 2023; Reher et al., 2024). While these models are valuable, their applicability in the Appalachian region is limited. Steep slopes, dense forests, and the relative absence of row-crop agriculture constrain the types of agrivoltaic systems that are feasible in states like West Virginia. A framework that aligns land suitability analysis, ecological functionality, and grazing system design—particularly for marginal Appalachian landscapes—can inform sustainable solar deployment that meets proposed decarbonization goals while improving rural economies and land integrity.

Fuzzy logic is a mathematical framework for handling reasoning that is approximate rather than exact, allowing for degrees of truth between 0 and 1, rather than the binary true/false of classical logic (Zadeh, 1988). This makes it particularly suited for spatial analysis, where uncertainty and vagueness are common, such as in land suitability assessments or environmental impact evaluations (Kainz, 2007). The core concept of fuzzy logic is based on fuzzy sets, where an element can belong to a set to a certain degree, rather than simply being a member or non-member. A membership function maps elements in a universe of discourse to values between 0 and 1, representing their degree of membership in a fuzzy set. For instance, in spatial analysis, a fuzzy set for “suitable terrain” might assign a 0.8 membership degree to a region with moderate slope and a 0.3 to 1 with steeper slopes.

The key operations in fuzzy logic, such as fuzzy union (OR), intersection (AND), and complement (NOT), allow for flexible modeling of spatial relationships where boundaries are not crisp. For example, in land-use planning, fuzzy logic can combine multiple variables (e.g., slope, soil type, solar radiation) to determine the suitability of land for a particular use, incorporating uncertainty and vagueness in the definitions of “suitable” or “unsuitable.” Additionally, fuzzy inference systems can be used to generate fuzzy outputs from spatial data inputs (e.g., “If slope is moderate and soil quality is high, then suitability for agriculture is high”), which are then aggregated to produce a final, interpretable output (Mamdani and Assilian, 1975). Fuzzy logic uses membership functions for a fuzzy set. These functions allow fuzzy sets to model spatial data that are inherently imprecise. In spatial analysis, this can be used to assess the uncertainty of environmental impacts, predict suitability for land use, and integrate multiple criteria while considering spatial heterogeneity. Fuzzy logic modeling and fuzzy set theory are widely used in spatial analysis and have been effectively combined with analytical hierarchical process in site suitability analysis, highlighting how these models can be applied as agricultural and environmental decision tools (Badr et al., 2018; Zouari et al., 2023). Fuzzy logic has not been used as a method for site suitability analysis of agrivoltaic installations in remedial grazing systems.

This study contributes a novel, lightweight spatial decision tool for identifying post-production natural gas wells suitable for agrivoltaic reclamation. Unlike prior site-suitability models focused on cropland or purely techno-economic factors, the framework integrates fuzzy-logic spatial analysis with biologically informed grazing buffers that reflect the land-use dynamics of animal agriculture. By coupling ecological and infrastructural variables within a reproducible GIS workflow, the model extends agrivoltaic assessment into the largely unstudied domain of energy-site reclamation via animal agriculture and demonstrates a transferable approach for linking renewable-energy planning with livestock-based land restoration. This study also addresses the absence of a reproducible spatial framework for identifying end-of-lifecycle natural gas wells suitable for agrivoltaic reclamation, particularly within the topographically complex and policy-constrained landscapes of central Appalachia.

This study investigated the use of fuzzy logic modeling, combined with use-specific buffer analysis, to identify well sites suited for agrivoltaic remediation. The results demonstrate an effective, reproduceable fuzzy logic model that allows for more nuanced site selection based on multiple, sometimes complex, criteria. Several sites well-suited to agrivoltaic remediation were identified. Though many of these were spatially associated with the same county, this study did not find any other correlations not accounted for within the fuzzy membership layers. Though suitable sites were identified, several factors increased the number of unsuitable or only-marginally suitable sites. For example, the steepness of many slopes made much of the study area unsuitable for animal agriculture. Steep slopes and forest cover regularly pose hurdles for the installation of agrivoltaic arrays (Dongrong, 2013). The major factor in diminishing suitability was the dense forest cover in much of West Virginia. Though forests were treated as a somewhat suitable land cover type, the density of forest-type cells within many of the buffer zones greatly reduced their extracted mean fuzzy overlay value. Interestingly, the solar grazing results of this study closely align with Arnette and Zobel’s (2011) spatial analysis of renewable energy potential in the greater southern Appalachian Mountains. Sites suitable for pasture grazing reclamation were most densely located in the Northeastern portion of West Virginia proximal to the Maryland and Pennsylvania borders. Additional dense clusters of pasture grazing sites remained in the far west adjacent to the Ohio border.

The range in solar production for each well site was an adequate proxy for agrivoltaic installation productivity, however it does not address the wider capabilities of such installations or how differing well pads and sites may affect solar energy harvest. Solar power production is influenced by several design and environmental factors, including racking systems, panel technologies, and agrivoltaic applications. The angle and orientation of racking systems affect solar irradiance capture and thus impact energy yield throughout the year (Makrides et al., 2010a,b). Different photovoltaic technologies—such as monocrystalline, polycrystalline, and thin-film—also demonstrate varied performance under outdoor conditions, especially in response to temperature and spectral irradiance (Mulcué-Nieto J. G. et al., 2020; Mulcué-Nieto A. et al., 2020; Sharma et al., 2013a,b). Additionally, agrivoltaic systems, which integrate agriculture with photovoltaic infrastructure, can alter the microclimate around panels and improve land-use efficiency, though they may introduce shading trade-offs that influence power output (International Renewable Energy Agency, 2021).

West Virginia’s current solar energy policy landscape presents several barriers that may inhibit the adoption of agrivoltaic systems as a viable remediation strategy for balancing energy production with agricultural land conservation. Unlike neighboring Virginia’s relatively progressive framework—which includes net metering expansion, lifecycle facility analysis, and farmland impact mitigation (Virginia Clean Economy Act, 2020a, 2020b; Virginia House Bill 774, 2022; Virginia House Bill 894, 2022; McGowan E., 2021; McGowan K., 2021)—West Virginia lacks a coordinated state-level vision or policy mandates that meaningfully integrate solar development with agricultural land use objectives. Critically, the absence of a renewable portfolio standard and limited incentives for distributed generation in West Virginia place disproportionate emphasis on utility-scale installations. These projects often involve site clearing, grading, are unsightly and compaction practices that degrade soil health, increase runoff, and preclude future agricultural productivity, directly opposing the core tenets of agrivoltaic design (National Renewable Energy Laboratory, 2020a,b). National-scale assessments suggest that widespread agrivoltaic adoption could generate more than 100,000 rural jobs over two decades while enhancing farm-level resilience and energy independence (Proctor et al., 2020). In regions historically dependent on extractive industries, agrivoltaics are increasingly viewed as a just-transition mechanism, coupling distributed solar with grazing or crop production to diversify income rather than replace agricultural activity (Center for Agrivoltaic Innovation and Adoption, 2025). Case studies from Virginia and other Appalachian states emphasize that local engagement, clear permitting frameworks, and equitable grid-connection incentives are essential to realize these benefits (Center for Agrivoltaic Innovation and Adoption, 2025; Proctor et al., 2020). Strengthening cooperative extension programs and farmer-developer partnerships can further de-risk adoption and channel agrivoltaic investment into working-land restoration and rural reinvestment across central Appalachia (Center for Agrivoltaic Innovation and Adoption, 2025).

Moreover, West Virginia has yet to implement guardrails around the use of prime farmland for solar development, nor has it widely encouraged research or guidance on solar-agriculture co-location. Without regulatory mechanisms akin to Virginia’s Department of Environmental Quality (DEQ) permit-by-rule environmental reviews or strategic support for dual-use systems, developers have little incentive to pursue agrivoltaic designs that preserve or enhance rural landscapes (Virginia House Bill 894, 2022). This policy inertia not only risks further marginalization of working lands but also undercuts the potential for agrivoltaics to serve as a climate-smart land-use strategy in Appalachia. Until state energy policy explicitly supports multifunctional land uses, agrivoltaic innovation will remain an academic exercise rather than a widespread practice in West Virginia.

Intersections between well site grazing areas, which excluded well pad area, and land parcel boundaries were investigated. Though these intersections could have been incorporated into the fuzzy logic model via a Small fuzzy membership function, indicating that fewer parcel intersections were preferred, its inclusion was beyond the scope of this study. Proper incorporation of land parcel ownership and segmentation into land consolidation experiments requires robust models that can effectively account for the spatial, legal, and social aspects of land distribution. Various approaches have been proposed to analyze land parcel fragmentation and ownership patterns. For example, Janus (2018a,b) introduced a model that accounts for the shape of transportation networks, which can improve the accuracy of farm efficiency assessments by considering the segmentation of land based on access routes. Similarly, Demetriou et al. (2013) proposed a methodology for measuring land fragmentation that integrates adjacent plots, enhancing the precision of land use models based on agricultural parcel data.

However, challenges arise from the need to reconcile cadastral data, which often contain inconsistencies or incomplete ownership records, as highlighted by Ertunç E. et al. (2022) and Ertunç H. M. et al. (2022). Moreover, incorporating ownership data into GIS models often requires significant data preprocessing, as discussed by Basista A. (2015) and Basista I. (2015), who emphasizes the importance of accurate parcel segmentation in the GIS tools used for land consolidation. In practice, models must not only consider physical land distribution but also the socio-legal constraints imposed by land ownership structures, which complicates the segmentation process and the overall model accuracy, the added complication of natural gas company well pad rights notwithstanding. Similarly, because the model was developed as a GIS-based biophysical screening tool, it does not incorporate grid-interconnection costs, substation proximity, transmission constraints, or utility-territory differences that are central to techno-economic decision-making. These data are often unavailable at a statewide scale or fragmented across private utilities, making consistent integration difficult. Similarly, land-mineral rights conflicts and the absence of a unified decommissioning registry in West Virginia limit the ability to capture ownership barriers within the suitability model. Future extensions could incorporate publicly available grid datasets where feasible, couple the spatial framework with cost-based scoring, and refine the grazing buffer using field-based forage measurements. These additions would expand the model from a biophysical site-screening tool into a fuller decision-support framework. However, despite installation complexities wrought by shared ownership, the sites identified by this model still hold ample opportunity for agrivoltaic remediation.

This study developed a fuzzy-logic, GIS-based framework to identify natural gas well sites in central Appalachia suitable for agrivoltaic reclamation. By integrating topographic, solar, and accessibility criteria with biologically informed grazing buffers, the model departs from purely topographic or economic methods and centers ecological viability and land-management coordination. The approach provides a reproducible, lightweight decision tool for screening locations where solar development can coexist with rotational grazing, supporting both renewable-energy and agricultural productivity goals. The framework also aligns with key Sustainable Development Goals—advancing clean energy (SDG 7), climate action (SDG 13), and land restoration and rural livelihoods (SDGs 15 and 8). Its principal advantage lies in coupling ecological and infrastructural suitability within a transferable model that can guide post-extraction land recovery and dual-use energy planning throughout the Appalachian region. As extractive industries continue their decline, regenerative agriculture, paired with distributed solar, may serve as both an ecological restoration tool and a durable economic anchor for Appalachian communities seeking stability beyond the boom-and-bust cycle of fossil fuel dependency.