Remotely sensed carotenoid dynamics predict photosynthetic phenology in conifer and deciduous forests

Abstract Crucially, the phenology of photosynthesis conveys the length of the growing season. Assessing the timing of photosynthetic phenology is key for terrestrial ecosystem models for constraining annual carbon uptake. However, model representation of photosynthetic phenology remains a major limitation. Recent advances in remote sensing allow detecting changes of foliar pigment composition that regulate photosynthetic activity. We used foliar pigments changes as proxies for light-use-efficiency (LUE) to model gross primary productivity (GPP) from remote sensing data. We evaluated the performance of LUE-models with GPP from eddy covariance and against MODerate Resolution Imaging Spectroradiometer (MODIS) GPP, a conventional LUE model, and a process-based dynamic global vegetation model at an evergreen needleleaf and a deciduous broadleaf forest. Overall, the LUE-models using foliar pigment information best captured the start and end of season, demonstrating that using regulatory carotenoids and photosynthetic efficiency in LUE models can improve remote monitoring of the phenology of forest vegetation.