Abstract: The construction of photovoltaic (PV) arrays alters local hydrothermal conditions, light availability, and soil structure quality beneath the panels, creating heterogeneous micro-environmental gradients. This induces a “niche differentiation - dominant species reorganization” effect through microhabitat screening, resulting in differentiated plant community characteristics. Following the implementation of “PV + Ecological Remediation” project in semi-arid coal mining subsidence areas, there is an urgent need to elucidate the impact of PV construction on plant community composition, diversity, and stability, along with its driving mechanisms, and to apply microhabitat regulation techniques to enhance ecological restoration outcomes in subsided lands.In this paper, Photovoltaic construction areas were selected in coal mining subsidence land in Yushenfu coal mining area , employed field quadrat surveys and multivariate statistical analyses. It compares plant community composition, structure, and stability characteristics across distinct microhabitat units of the PV array (front eaves of the photovoltaic panel, underside of the photovoltaic panel, rear eaves of the photovoltaic panel and middle of the photovoltaic panel) with those in naturally recovered coal mining subsidence areas and in control areas (non-subsided lands). The driving mechanisms of community differentiation were analyzed by integrating monitored soil physicochemical properties and microclimatic factors across these units. The results showed that: (1) Compared with the undisturbed areas, the diversity and stability of plant communities significantly decreased in naturally recovered coal mining subsidence areas , However, paradoxically, plant density increased in the naturally recovered areas due to the establishment of one-year-old and two-year-old plants. Plant communities exhibited heterogeneity across microhabitat units within the photovoltaic array. Compared to the naturally recovered areas, front eaves and rear eaves of the PV panel units enhanced plant density, diversity, and stability, whereas underside and middle of the PV panel zones showed opposite trends. (2) Plant communities in naturally recovered subsidence areas, front eaves of the PV panel, rear eaves of the PV panel and middle of the PV panel zones demonstrated greater similarity in species composition to non-subsided lands, dominated by drought-tolerant perennial plants. In contrast, the underside of the PV panel microhabitat shaped by shaded conditions and high humidity was primarily dominated by mesophytic annual/biennial plants, showing reduced species similarity to control areas (3) Soil moisture and light intensity were common dominant factors affecting the community characteristics across the study sites. Meanwhile, surface temperature in the control areas, surface wind speed  in the naturally recovering areas, soil organic matter in front and rear eaves of the PV panels, air humidity underside of the PV panels and soil bulk density in the middle of the PV panels as statistically significant differential drivers of community divergence (P<0.05). Overall, based on plant adaptive differentiation to microhabitats, a differentiated community structure pattern has formed within the subsided land PV area, resulting in significant spatial heterogeneity in vegetation restoration effectiveness. Our findings demonstrate that the the front eaves zone effectively promotes vegetation restoration in the subsided land through significant improvements in microhabitat conditions and soil amelioration. In contrast, the rear eaves zone exhibits a limited promotive effect. Conversely, the middle and underside zones show significant inhibitory effects on vegetation restoration. These results provide a critical scientific basis for implementing targeted ecological optimization measures following the deployment of “PV + Ecological Remediation” projects.

Key words: coal mining subsidence, photovoltaic plant, vegetation restoration, plant community structure, ecological effects