Depth-dependent responses of soil microbial communities and co-occurrence networks to solar photovoltaic panel installation along a soil profile in a hyper-arid desert

Introduction Photovoltaic (PV) power generation is expanding rapidly in arid regions, but its belowground ecological effects across soil profiles remain poorly understood. Methods Here, we examined soil bacterial and fungal communities across three microhabitats (bare land, between PV panels, under PV panels) and five soil layers (0–100 cm) at a PV station in Xinjiang, China. Bacterial 16S rRNA gene and fungal ITS amplicon sequencing was integrated with functional prediction, community assembly analysis, and co-occurrence networks. Results PV panel installation significantly altered microbial taxonomic composition and potential functions, with Pseudomonas and functions related to methanol oxidation and nitrogen fixation enriched in PV zones, while nitrate reduction and hydrocarbon degradation were more abundant in bare land (BL). These effects weakened with increasing soil depth. Bacterial α-diversity did not change significantly, but bacterial community structure differed significantly between PV zones and BL; fungal community structure differed only in the 0–20 cm and 20–40 cm soil layers. Bacterial assembly shifted from deterministic homogeneous selection in BL to more stochastic processes in PV zones, accompanied by increased dispersal limitation, suggesting weakened deterministic environmental filtering and reduced dispersal under PV panels, whereas fungal assembly remained stochastic-dominated. Bacterial co-occurrence networks decreased in size and complexity with soil depth. PV-zone co-occurrence networks were more modular but more vulnerable than BL, while fungal networks were smaller and relatively stable. Discussion These findings reveal depth-dependent and microhabitat-specific microbial responses to PV installation, with bacteria more sensitive than fungi, and provide a scientific basis for assessing the ecological impacts of large-scale PV deployment in arid desert ecosystems.