Effects of Different Photovoltaic Installation Modes on Soil Bacterial Community Structure, Diversity, and Assembly Mechanisms in Rocky Desertification Areas
Abstract
Soil bacteria regulate nutrient transformation and cyclingin the soil, thereby influencing ecosystem functionality and stability. However, the impact pathway and underlying mechanisms of photovoltaic array on soil bacterial community are imperfectly known.This study quantified the impacts of different photovoltaic (PV) installation modes on the structure, diversity, and succession of soil bacterial communities in rocky desertification areas . In contrast to control plot (CP), a two-row PV installation mode (TP) reduced ACE, Chao1, Simpson and Shannon indices of soil bacterial communities, whereas a flat-top PV installation mode (FP) increased these indices,FP treatment had significantly higher ACE, Chao1 and Shannon indices than TP. PV installations improved the relative abundances of Myxococcota and Bacteroidota, with FP treatment increasing by 27.9% and 213.4% respectively compared to CP treatment, but reduced the relative abundance of Acidobacteriota, Actinobacteriota and Gemmatimonadota, with the largest reduction in Actinobacteriota ranging from 36.0% to 32.3%.PV installations indirectly drove changes in bacterial communities by altering the climate and subsequently affecting the soil. Soil physicochemical properties and microclimate were identified as key predictors of bacterial community structure and diversity. Network analysis revealed that different PV installation modes significantly influenced co-occurrence network structure of soil bacterial communities. Compared with the control, both TP and FP enhanced associations among bacterial taxa, with TP showing the strongest network connectivity and FP exhibiting intermediate connectivity between the control and TP. The assembly process of soil bacterial communities was primarily driven by deterministic processes, which dominated in all plots (control, TP, FP). PV installations affected the assembly of soil bacterial communities, in the control, homogeneous selection entirely controlled community assemblages (100%), while in TP, its proportion decreased to 78%,under FP conditions, stochastic processes began to exert influence, with dispersal limitation accounting for 17%. Our data suggested that FP model increased the diversity and richness of soil bacterial community, enhanced stochastic process of community assembly, and had a milder impact on network connectivity in rocky desertification areas compared with TP.