Leaf spectroscopy detects site-associated functional differentiation in a dominant mangrove species near its northern range margin

Functional variation within species may be especially important at range margins, yet conventional trait measurements constrain population-level assessments. Leaf reflectance spectroscopy integrates signals associated with pigments, water, chemistry, and structure, potentially revealing functional differentiation. We analyzed leaf spectra from 48 Kandelia obovata trees across a southern reference population (SRP), a natural range-margin population (NRM), and a long-established planted population north of the historical range margin (PNR). We compared full spectra and 12 vegetation indices, partitioned wavelength-specific variance, measured leaf traits in 30 trees, and visualized population structure using principal component analysis (PCA) and t-distributed stochastic neighbor embedding (t-SNE). Spectral profiles differed among populations across visible, near-infrared, and shortwave-infrared wavelengths. Site-level differences accounted for a mean 19.7% of spectral variance, compared with 59.1% attributable to within-tree-plus-residual variation. Vegetation indices exhibited index-specific population differences. Trait measurements showed higher SPAD values in PNR and lower equivalent water thickness, leaf thickness, and leaf mass per area in SRP, with only partial correspondence to vegetation-index patterns. Both ordination methods revealed site-associated structure, although separation differed among datasets. Because spectral and trait records could not be reliably paired at the tree or leaf level, comparisons were restricted to population-level patterns. These findings indicate that leaf spectroscopy can detect site-associated differentiation within K. obovata , but effects of climate, population origin, stand history, and sampling date remain confounded.