Nonlinear effects of landscape pattern and forest structure on subtropical forest ecosystem services: Evidence from Jiangxi Province, China

Subtropical forests ecosystem services (FES) underpin global carbon cycle balance, biodiversity conservation, and regional ecological security. However, the nonlinear relationships between FES, landscape patterns, and forest community structure in subtropical forests have received relatively little attention. We selected subtropical forests in Jiangxi Province, China, surveyed 69 forest plots, and quantified FES combined with model simulation. We integrated Structural Equation Model (SEM) and Generalized Additive Mixed Model (GAMM) to reveal the nonlinear driving relationships of four FES, namely forest stock volume (FSV), carbon sequestration (CS), water yield (WY), and soil retention (SR). Results showed spatial differentiation in FES. There was a significant synergistic relationship between FSV and CS, and the same applies to SR and WY. Both landscape pattern and forest structure had significant direct or indirect impacts on FES. WY and SR were more affected by landscape patterns, while CS and FSV were more influenced by forest structure factors. FES exhibited nonlinear relationships with key influencing factors, with specific correlations: 1) A U-shaped relationship existed between WY and Simpson index; 2) A W-shaped correlation occurred between SR and Pielou index; 3) Percentage of landscape, class area, largest patch index showed approximate W-shaped links with CS; 4) Higher FSV was achieved at low biodiversity levels and high canopy closure extent. This study provides methodological references and scientific support for subtropical forest management and sustainable FES supply.