LAUSR

Satellite remote sensing provides unmatched spatiotemporal information on vegetation gross primary productivity (GPP). Yet understanding of the relationship between GPP and remote sensing observations and how it changes with factors such as scale, biophysical constraint, and vegetation type remains limited. This knowledge gap is especially apparent for dryland ecosystems, which have characteristic high spatiotemporal variability and are under-represented by long-term field measurements. Here we utilize an eddy covariance (EC) data synthesis for southwestern North America in an assessment of how accurately satellite-derived vegetation proxies capture seasonal to interannual GPP dynamics across dryland gradients. We evaluate the enhanced vegetation index, solar-induced fluorescence (SIF), and the photochemical reflectivity index. We find evidence that SIF is more accurately capturing seasonal GPP dynamics particularly for evergreen-dominated EC sites andmore accurately estimating the full magnitude of interannual GPP dynamics for all dryland EC sites. These results suggest that incorporation of SIF could significantly improve satellite-based GPP estimates.