Species-Level Associations of Leaf Traits and Photosynthetic Capacity in Common Woody Plants of Subtropical Secondary Forest Vegetation

Leaf functional traits are widely used to describe interspecific variation in plant resource-use strategies and photosynthetic performance. We examined leaf traits and photosynthetic characteristics in 21 common woody species (11 tree and 10 shrub species; two individuals per species) sampled from subtropical secondary-forest vegetation in Chongqing, southwestern China. Maximum net photosynthetic rate (Pnmax), apparent quantum yield (AQY), light saturation point (LSP), and light compensation point (LCP) were derived from species-level fitted light-response curves. Among the 17 species with accepted fits, Pnmax ranged from 3.35 to 12.93 μmol m−2 s−1. Estimates for Jacaranda mimosifolia and Vitex negundo var. cannabifolia were extrapolated beyond the measured irradiance range (14.66 and 16.23 μmol m−2 s−1, respectively). No finite Pnmax was defined for Osmanthus fragrans or Rhododendron simsii because β was estimated at zero; their fitted asymptotic net photosynthetic rates were 3.33 and 4.29 μmol m−2 s−1, respectively. Nitrogen content per unit leaf area (Narea) ranged from 0.46 to 2.28 g m−2. Trait analyses used species means as independent observations. Spearman correlations were evaluated with Benjamini–Hochberg false discovery rate correction, and prespecified candidate regression models were compared using adjusted R2, AICc, and leave-one-species-out cross-validation. After FDR correction, Narea was positively associated with Pnmax (Spearman’s ρ = 0.673, q = 0.0137) and AQY (ρ = 0.666, q = 0.0137). No statistically significant differences in photosynthetic parameters remained between trees and shrubs after correction for multiple testing. Among the prespecified candidate models, the Narea-only model had the lowest AICc for Pnmax (adjusted R2 = 0.440). For AQY, the Narea-only and Narea + LA models had similar support (ΔAICc = 0.41), and the data did not clearly distinguish between them. Adding LA did not significantly improve either model in the incremental F-tests (p = 0.486 for Pnmax and p = 0.135 for AQY). The positive Pnmax–Narea association remained after controlling for LMA using partial Spearman correlation (partial ρ = 0.762, p < 0.001), indicating that shared LMA normalization alone did not explain the association in this dataset. These species-level associations are observational and do not establish causation.