Background Photovoltaic (PV) infrastructure generates microhabitat heterogeneity through light modulation, critically influencing plant community assembly. Understanding species turnover, functional reorganization, and phylogenetic patterns under PV shading is essential for ecological management of PV farms. The National PV & Storage Empirical Research Platform (Daqing Base), established on soda saline-alkali soils in Northeast China, provides a natural experiment with contrasting PV mounting systems. We investigated how two-year shading regimes reshape plant community structure, niche dynamics, and phylogenetic diversity, addressing knowledge gaps in renewable energy ecology within fragile marginal ecosystems.

Methods Plant communities in fixed tilt racks (gap of fixed tilt rack, FG; underneath of fixed tilt rack, FU) and horizontal single-axis trackers (gap of horizontal single-axis tracker, HG; underneath of horizontal single-axis tracker, HU) PV arrays were surveyed for species composition, cover, height, and abundance. Species were categorized into three functional groups: bunchgrasses, rhizomatous grasses, and forbs. We quantified α-diversity (Margalef diversity index, equitability index), phylogenetic metrics (Faith’s phylogenetic diversity, net relatedness index), β-diversity components (turnover/nestedness), and niche dynamics (Levins/Pianka indices). Statistical analyses compared microhabitat effects on community assembly, emphasizing functional and phylogenetic responses to shading regimes.

Results 1) Shading shifted species composition toward shade-tolerant plants beneath panels. While α-diversity remained stable (P > 0.05), phylogenetic diversity significantly decreased under panels (P < 0.05), indicating environmental filtering. Fixed tilt rack shading enhanced community evenness by 13.10% versus horizontal single-axis tracker. 2) Functional group responses diverged sharply: FU shading reduced bunchgrasses importance value by 55.74%, but increased rhizomatous grasses and forbs by 10.03% and 11.89%, respectively. In contrast, HU shading amplified bunchgrasses importance value by 401.52%, while suppressing rhizomatous grasses (−7.38%) and forbs (−23.76%). 3) Niche overlap intensified beneath panels, community in FU and HU showed higher niche overlap species pairs (O_ik > 0), mean niche overlap values and the species pairs with O_ik > 0.50. Phylogenetic clustering (net relatedness index > 0) underneath panels contrasted with overdispersion (net relatedness index < 0) in gap areas, indicating stronger competitive exclusion under shading, furthermore, fixed tilt rack systems exhibited greater phylogenetic clustering. β-diversity was dominated by species turnover (76.79% in fixed tilt rack systems, 83.10% in horizontal single-axis tracker systems), reflecting microhabitat-driven community differentiation rather than nestedness.

Conclusions PV shading drives functional and phylogenetic reorganization via light-mediated environmental filtering, with underneath panels systems selecting stress-tolerant species more effectively. Dominant species turnover underscores microhabitat divergence between gap and underneath panels zones. These findings establish a predictive framework for biodiversity management in PV farms, emphasizing fixed tilt systems’ ecological impacts in saline-alkali ecosystems. Strategic PV design could balance energy production with conservation in marginal lands.