Context-dependent forest carbon storage: a shift in dominance from structural complexity to large-sized trees across disturbance gradients

Forest biomass stores vast amounts of carbon, playing a vital role in biodiversity maintenance, carbon sequestration, and climate change mitigation. However, the synergistic effects of forest attributes (e.g., plant diversity, community-level functional traits, and structural complexity) on biomass carbon stock remain unclear, especially across large ecological scales. By conducting a systematic forest survey encompassing 62,842 trees in 1217 plots (20 m*20 m) in southwest China, we quantified the relationships between forest attributes and carbon stock, and investigated the modulating effects of topography, climate and anthropogenic disturbance. We found that forest carbon stock increased with mean annual precipitation, soil nitrogen and elevation as well as decreased with disturbance intensity, but had no significant relationship with neither mean annual temperature or slope. All attributes had significant relationship with forest carbon stock. Specifically, tree species diversity (hill numbers: ⁰D, ¹D, ²D) had a positive relationship with forest carbon stock, indicating that both the richness of all species and that of common species were linked to forest carbon accumulation. Community-level maximum tree height exhibited a positive relationship with forest carbon stock, while community-level leaf nitrogen content exhibited a negative relationship with it Within stand structure factors, stand density, canopy closure, and structural complexity (DBH variance and height variance) exhibited positive relationships with forest carbon stock. Furthermore, these relationships were modulated by elevation, mean annual precipitation, and disturbance intensity. Interestingly, the best predictor of forest carbon stock shifted from structural complexity in undisturbed forests to maximum tree height in disturbed ecosystems, reflecting the decrease in structural complementarity effects while the intensification of dominant species effect under anthropogenic disturbance. Our study reveals the critical context-dependency of forest attribute-carbon stock relationships. This finding offers a framework for carbon- and climate-targeted forest restoration across tropical and subtropical regions under similar environmental conditions, highlighting the need for strategies that synergistically manage both diversity and large-sized trees.