Leaf trait network variations with woody species diversity and habitat heterogeneity in degraded karst forests

Leaf Trait Networks (LTNs) visually display the intricate trait connections in offering insights into how plants adapt to diverse environments. The pervasive environmental heterogeneity can influence plant diversity and functional traits. Various habitats sustain distinct species diversity and leaf trait differentiation, precisely as the degraded karst-heterogenous ecosystem caused by inhomogeneous microhabitats. However, how the LTNs respond to understory habitat heterogeneity and species diversity in degraded karst forests remains unclear. This study selected fifty-three forest plots for plant community survey and habitat investigation in a typical karst topography to visualize LTNs using twenty-one woody leaf traits. LTN architecture variations were analyzed in relation to habitat heterogeneity and species diversity through network-architectural parameters, including edge density, diameter, average path length, and average clustering coefficient, utilizing regression, principal component analysis, and structural equation model. The results showed a 35% variation in modularity across all plots, indicating distinct network architectural modularization differentiation. Network diameter and average path length displayed significant initial decreases followed by increases with ascending PC1 heterogeneity scores, indicating moderate heterogeneity sustained greater network connectivity. In addition, the network diameter, average path length, and average clustering coefficient significantly decreased with increasing species diversity, indicating that high species diversity enables greater connectivity and complexity. Moreover, the habitat heterogeneity significantly affected species diversity directly and affected the average clustering coefficient, diameter, average path length, and edge density indirectly through species diversity effects via SEM pathway analysis. Overall, habitat heterogeneity and woody plant diversity of karst forests significantly influence LTN architectures in degraded karst forests. In conclusion, moderate heterogeneity sustains greater network connectivity while species diversity jointly enhances connectivity and complexity by combining modular regulation of plant functional traits.