Phenotyping seasonal photosynthetic energy-partitioning states in temperate evergreen conifers using hyperspectral imaging

Evergreen conifers retain needles and photosynthetic capacity during winter despite prolonged exposure to freezing temperatures and high irradiance, yet it remains unclear which aspects of this seasonal photosynthetic regulation can be reliably detected using passive optical sensing. Here, we phenotype seasonal photosynthetic energy-partitioning states in two temperate evergreen conifers, Pinus densiflora and Pinus koraiensis, by integrating chlorophyll (Chl) fluorescence imaging, pigment analyses, photosynthetic oxygen evolution, and hyperspectral reflectance imaging. Across three years, Chl fluorescence parameters (e.g., F_v/F_m, Y_PSII, Y_NPQ, and Y_NR) indicated reduced PSII photochemical efficiency and increased energy dissipation during winter, accompanied by changes in carotenoid-to-chlorophyll (Car/Chl) ratio, xanthophyll cycle de-epoxidation (DEPS), and photosynthetic oxygen evolution. Despite reduced PSII photochemical efficiency, photosynthetic oxygen evolution and pigment pools remained comparatively stable, indicating that reductions in PSII photochemistry were not accompanied by proportional declines in whole-leaf photosynthetic capacity. These seasonal physiological states were consistently reflected in hyperspectral vegetation indices. Photochemical Reflectance Index (PRI) showed correlations with both regulated (r = 0.57) and non-regulated light energy dissipation (r = − 0.67), consistent with its sensitivity to Car/Chl ratio (r = − 0.87) and DEPS (r = − 0.85). Indices including Fluorescence Curvature Index (FCI) and Ratio Analysis of Reflectance Spectra chlorophyll b (RARSb) showed consistent seasonal correlations with photosynthetic oxygen evolution (r = − 0.68 and − 0.68), while also reflecting variation in pigment composition (e.g., Chl content; r = 0.73 and 0.78). Drone-based hyperspectral imaging further indicated that these needle-level physiological patterns were detected at the canopy level under the conditions examined. Indices such as FCI exhibited clear seasonal profiles, with separation between summer and winter values. By framing seasonal photosynthetic regulation as a phenotyping problem rather than a mechanistic one, this study provides a data-supported proof-of-concept framework for non-invasive, scalable monitoring of evergreen forests using hyperspectral sensing.