Key factors and pathways through which photovoltaic arrays mediate changes in microclimate, vegetation, and soil in a rocky desertification ecosystem

Photovoltaic (PV) arrays substantially alter above- and belowground environmental factors as well as ecosystem functionality. However, the intrinsic mechanisms driving these changes remain poorly understood, particularly in fragile karst ecosystems. The present study examined the impacts of installing photovoltaic arrays on microclimate, vegetation, and soil properties in a rocky desertification ecosystem in southwestern China. Compared with adjacent reference areas, the photovoltaic arrays significantly modified the local microclimate, reducing light intensity by 49.39%–68.03% and wind speed by 76.53%–97.96% across seasons (measured at 60-min intervals with sensors at 1.5–2 m height). A seasonal alternation between “heat island” and “cold island” effects was observed: under the arrays, air temperature increased by 0.54 °C in summer but decreased by 0.23–0.42 °C in spring and winter, while relative humidity consistently increased (0.38%–10.82%) throughout the year. Regarding soil properties, the PV arrays caused a 22.8% to 28.1% increase in soil temperature and moisture during the rainy season but led to significant declines in bulk density, organic carbon, and hydrolyzed nitrogen during the dry season, with reductions ranging from 10% to 55% ( p < 0.05). The observed 15.3%–27.8% declines in carbon and nitrogen represent substantial nutrient losses; bulk density showed seasonal reversals (decreasing in the dry season due to soil shrinkage but increasing in the rainy season due to particle migration) based on repeated sampling at fixed plots. Conversely, available potassium increased by 32% in the dry season ( p < 0.05), attributed to reduced plant uptake and leaching losses under shading. Plant community diversity (Shannon index, Pielou evenness, and richness) significantly increased, while the dominance index decreased ( p < 0.05). This diversity change may be driven by PV shading, which suppresses dominant heliophilous species and promotes shade-tolerant species, thereby redistributing species dominance and creating more equitable resource competition. However, plant root systems became simplified, with a 6.73%–80.40% reduction in root tip number. This simplified root architecture may reduce resource acquisition efficiency from deeper soil layers, yet it does not significantly alter total above- or belowground biomass. This trade-off between aboveground gains (higher diversity) and belowground costs (simpler root structures) may have implications for long-term ecosystem stability and recovery, especially in shallow karst soils. Partial least squares path modeling (GOF = 0.67) revealed direct benefits to microclimate (path coefficient = 0.76) and diversity (0.89) but direct costs to roots (−1.22) and soil (−0.97), representing statistical associations rather than causation. The model accounted for 68%, 86%, 45%, and 82% of the variance in microclimate, plant diversity, root morphology, and soil properties, respectively. Indirectly, microclimate improved soil via plant diversity (∼0.26 path coefficient) and directly via microclimate (∼0.52), totaling approximately 80% of the direct negative effect, partially offsetting direct negative impacts. Mechanistically, PV shading initiates a cascade: altered microclimate reshapes plant communities by reducing heliophilous species and promoting shade-tolerant species, while plants simplify root architecture under low light and soil nutrients decline; these direct costs are partially offset by diversity-mediated feedback to soil properties. This trade-off is particularly acute in shallow karst soils with limited nutrient buffering, cautioning against extrapolation from other ecosystems and highlighting the need for site-specific assessment and long-term monitoring.