Phylogenetic patterns in leaf thermal traits predicted from hyperspectral reflectance

Predicting climate change impacts on plant-atmosphere interactions requires a better understanding of the environmental and phylogenetic drivers of leaf thermal (energy balance and heat tolerance) traits. Our understanding of these drivers has been limited by resource-intensive field sampling methods. However, recent advances in remote sensing and field spectroscopy has enabled us to distinguish leaf chemical, physical and physiological traits with a high degree of accuracy. While there is potential for the use of field spectroscopy to quantify leaf thermal traits, this remains untested. We measured hyperspectral reflectance and leaf thermal traits at the leaf level in 205 species spanning 186 plant families in two common gardens that control for environmental variation. On each leaf, hyperspectral reflectance was measured using an SVC HR-1024i field spectroradiometer, and two thermal traits (the thermal time constant τ and photosystem II heat tolerance T50), were measured using standard methods. Partial Least Squares Regression (PLSR) was used to correlate reflectance spectra with τ and T50. We found that leaf-level reflectance predicted both leaf τ and T50. Further, phylogenetic signals were detected in thermal traits, and in various spectral regions. Our results demonstrate the potential for hyperspectral data to be used for monitoring phylogenetic diversity and leaf thermal traits.