Unravelling tree diversity patterns and responses to environmental gradients in a tropical forest landscape of the Western Ghats.

Understanding spatial patterns of plant species diversity and the factors (e.g., climate and human) that drive these patterns is essential for biodiversity conservation. We used data from 170 0.1-ha forest plots in the Shettihalli tropical forest landscape of the Western Ghats biodiversity hotspot, India, to analyse tree community composition and the drivers of α-diversity (Shannon) and β-diversity (LCBD). Compositional patterns were visualized using Non-Metric Multidimensional Scaling (NMDS), and hybrid feature selection with structural equation modeling (SEM) was employed to evaluate the direct and indirect effects of environmental variables on diversity. NMDS identified four distinct forest types in the Shettihalli landscape: semi-evergreen, dry deciduous, moist deciduous, and plantation forests, each with distinct plant composition. Shannon diversity and ecological uniqueness was significantly higher in semi-evergreen forest than in deciduous forest plots. The SEMs explained about 79% and 39-45% of the variation in α-diversity and β-diversity. Our analysis indicated that current diversity patterns result from multiple processes, with structure, disturbance, and edaphic parameters exerting the strongest direct and indirect effects on α-diversity. β-diversity, in contrast, was largely influenced by climate, topography, stand structure, and edaphic factors. Overall, our findings indicate that various factors (e.g., climate, topography, and human disturbance) interact to shape tree diversity patterns in tropical forests. These findings will help develop unique conservation and management strategies for distinct forest types in tropical forest ecosystems.