COS 31-4 - Biophysically-Informed Indicators of Concurrent and Impending Drought

Abstract: Plants play a dynamic role in the Earth system as active participants that respond to changing conditions through rapid physiological adjustments that impact downstream processes like photosynthesis and evapotranspiration. Despite technological advancements and improved data resolution, detecting drought as it unfolds remains challenging, particularly rapid-onset ‘flash’ droughts. Conventional drought monitoring efforts often overlook the critical role of vegetation in the water cycle and instead focus primarily on meteorological variables. However, sub-daily variations in vegetation activity have the potential to serve as valuable bioindicators of impending drought stress. Here, we use sub-daily satellite data linked to plant physiological processes to detect vegetation stress prior to and during drought. Building on biophysical theory and prior research, we introduce a novel, biophysically-informed approach to incorporating vegetation into drought monitoring and forecasting efforts. Our objectives are to compare spaceborne and airborne afternoon vs. morning Land Surface Temperature (LST) to in situ measurements of evapotranspiration and water potential, with the broader goal of developing a sub-daily morning vs. afternoon LST product using data from NASA’s Moderate Resolution Imaging Spectroradiometer (MODIS) sensors. This will help us assess the potential for a remotely-sensed, biophysically-informed drought signal based on sub-daily variation in LST across biomes.

This project is based on prior research where we developed an in-situ early drought warning signal based on biophysical theory. We found that diurnal changes in LST, which are influenced and impacted by changes in evapotranspiration (ET), correspond with shifts in the timing of daily maximum ET during drought. Using anomalously low soil water content as a proxy for dry conditions, we uncovered 338 dry periods during the growing season from eddy covariance sites. We found 18 out of 24 sites, covering eight ecoregions, showed that the timing of peak ET shifted to earlier in the day, temporally decoupling from the timing of peak vapor pressure deficit. We extended these results by leveraging diurnal temporal decoupling of LST, rather than ET, as a signal of plant stress during drought conditions. Preliminary analysis of MODIS morning and afternoon overpasses show a similar pattern. Drone flight LST data further demonstrated a correlation between midday water potential and LST. These findings provide support for a spaceborne metric of sub-daily LST indicative of biophysical responses to stressful conditions, paving the way for a new, biophysically-informed drought monitoring and forecasting framework that provides sub-daily information at spatial scales relevant to land management decisions.