Reflectance and needle functional traits in Scots pine seed orchards

The Scots pine plays an important ecological and economic role in Europe, especially under changing climate conditions, yet its physiological and spectral variability remains largely unexplored. Hyperspectral remote sensing provides a non-invasive method to investigate vegetation traits. By combining hyperspectral data from both needle-level (via spectroradiometer) and canopy-level (via UAVs) measurements, we explored needle functional traits (NFTs) to understand the physiological processes driving tree variability. We used partial least squares regression (PLSR) to predict photosynthetic pigments, needle water content, and morphology at both levels, achieving <i>R²</i> values up to 0.8. Our study, conducted on replicated clonal seed orchards over two seasons, highlights the efficacy of hyperspectral phenotyping. Using linear mixed models, we estimated broad-sense heritability, with the carotenoids-to-total chlorophylls ratio showing the highest heritability (0.29) of all the pigments. This result suggests a stronger genetic influence on a stress-related trait, offering the potential for breeding superior individuals. Significant genetic correlations between specific wavelengths and NFTs indicate the potential for indirect selection in breeding programs. Equivalent water thickness and total chlorophyll content also exhibited significant genetic correlations with visible and near-infrared spectral regions at the canopy level. The low genotype-by-environment interaction and strong year-to-year consistency suggest stable clonal rankings, especially in the visible spectrum. These results underscore the value of NFTs and needle reflectance spectra in identifying genetic variability in Scots pine. Integrating these data with machine learning provides promising tools for improving breeding programs.