Understanding the effects of forest types, vegetation structural diversity, and soil properties on above- and below-ground carbon stock of moist deciduous forest ecosystems in Western Ghats, India

The dynamics of tree biomass carbon (Tcs) and soil organic carbon (SOC) stocks vary significantly across forest types, shaped by vegetation structural diversity, soil properties, and nutrient availability. However, the relative influence of these ecological factors on Tcs and SOC stocks across different forest types remains poorly understood. This study examines the influence of forest types on tree diversity {Shannon index (H′), species richness (SR), and probability of interspecific encounter (PIE)}, structural attributes (basal area and density), soil properties {bulk density, pH, and N}, nutrients {P, k, Ca, Mg, Fe, Zn, Mn, and Cu}, and carbon stocks in Western Ghats. Data were collected from 12 stands across two forest types {5 permanent preservation plots (PPP) and 7 teak plantations (TP)} using field-based quadrat sampling and soil analysis (0–100 cm depth). Structural equation modeling (SEM) was employed to disentangle forest types’ direct and indirect effects on Tcs and SOC stocks. Results showed significantly higher tree diversity, nutrient enrichment, Tcs, and SOC stock in PPP compared to TP. Tcs ranged from ∼186–312 Mg C ha⁻¹, while SOC ranged from ∼16–84 Mg C ha⁻¹, jointly contributing 47–54 % of forest C storage. Tree density, H’, and soil nutrients (P, K, C/N, and Mn) emerged as major drivers of Tcs and SOC variation, explaining 73–90 % of the observed variance. SEM analysis confirmed structural diversity and soil nutrients as primary drivers influencing C stock, with soil properties having a lesser impact. Overall, the findings highlight that forest types and soil depth modulate the effects of vegetation and soil attributes on C dynamics. This study advances understanding of carbon stock regulation in tropical forests and provides valuable insights for conservation planning and sustainable forest management.