Evaluating the Effects of Fire on Semi-Arid Savanna Ecosystem Productivity Using Integrated Spectral and Gas Exchange Measurements

Drylands account for 40% of the land surface and have been identified as increasingly important in driving interannual variability of the land carbon sink. Yet, understanding of dryland seasonal ecosystem productivity dynamics - termed Gross Primary Productivity (GPP) - is limited due to complex interactions between vegetation health, seasonal drought dynamics, a paucity of long-term measurements across these under-studied regions, and unanticipated disturbances from varying fire regimes. For instance, fire disturbance has been found to either greatly reduce post-fire GPP through vegetation mortality or enhance post-fire GPP though increased resource availability (e.g., water, light, nutrients, etc.). Here, we explore post-fire ecosystem recovery by evaluating seasonal GPP dynamics for two Ameriflux eddy covariance flux tower sites within the Santa Rita Experimental Range of southeastern Arizona: 1) the US-SRG savanna site dominated by a mix of grass and woody mesquite vegetation that was burned in May 2017, and 2) the US-SRM savanna site dominated by similar vegetation but unburned for the full measurement record. For each site, we collected leaf-level spectral and gas exchange measurements, as well as leaf-level chemistry and soil chemistry to characterize differences in nutrient availability and microbial activity throughout the 2017 growing season. From spectral data, we derived and evaluated multiple common vegetation metrics, including normalized difference vegetation index (NDVI), photochemical reflectivity index (PRI), near-infrared reflectance (NIRv), and MERIS terrestrial chlorophyll index (MTCI). Early results suggest rates of photosynthesis were enhanced at the burned site, with productivity increasing immediately following the onset of monsoonal precipitation; whereas initial photosynthesis at the unburned site remained relatively low following first monsoonal rains. MTCI values for burned vegetation appear to track higher levels of leaf-level nitrogen content upon monsoonal onset, but requires further validation by leaf-level chemistry. This work suggests that the integration of spectral, gas exchange, and soil measurements could be a powerful framework toward advancing our understanding of fire-ecosystem productivity feedbacks across spatiotemporal scales.