Near-surface remote sensing techniques provide unmatched spatiotemporal information on ecosystem photosynthesis, termed gross primary productivity (GPP). Yet, our understanding of the relationship between remote sensing proxies and observed GPP - and how this relationship changes with space and time, biophysical constraint, vegetation type, etc. - remains limited. This knowledge gap is especially apparent for dryland ecosystems, which have high spatial and temporal variability and are under-represented by long-term, continuous field measurements. Here, we assess the ability of multiple remote sensing vegetation proxies to accurately capture diurnal to seasonal GPP dynamics in three dryland eddy covariance tower sites in southern Arizona: a grassland (US-WKG), a savanna (US-SRM), and a mixed conifer forest (Mt. Bigelow, Arizona). We specifically evaluate the long established normalized difference vegetation index (NDVI), and compare against the photochemical reflectivity index (PRI) and solar-induced fluorescence (SIF), which are vegetation proxies linked more directly to plant physiological function. Our study sites offer unique opportunity to study seasonal sensitivity to drought stress, as the influence of the North American Monsoon drives strong bimodal patterns in seasonal productivity. We divided our observations into pre-monsoon, monsoon, and post-monsoon periods, and examined the sensitivity of tower mounted NDVI, PRI, and SIF to changes in environmental drivers at both diurnal and seasonal time scales.
Results/Conclusions
Preliminary results suggest that PRI and SIF are more sensitive than NDVI to drought-induced reduction of GPP for both diurnal and seasonal time periods. This is partly due to decoupling of NDVI from GPP that is especially apparent at the savanna and conifer sites during periods with low soil moisture and/or high vapor pressure deficit. Analyses related to how these responses change across sites and time periods are ongoing. Our initial findings indicate that the use of PRI and SIF in tandom could yield dramatic improvements in remote sensing-based estimates of GPP, particularly in dryland systems.