Towards biophysically-informed early drought warning systems

Kesondra Key, Environmental Science, Indiana University, Bloomington, IN and Mallory L. Barnes, O'Neill School of Public and Environmental Affairs, Indiana University, Bloomington, IN

Background/Question/Methods
Drought, one of the most expensive and environmentally damaging natural disasters on the planet, is projected to increase in both intensity and duration by 2050 in the United States. Despite causing potentially severe and long-lasting impacts to ecosystems and human societies, we still lack reliable methods of detecting drought as it develops. Currently, several organizations around the world rely on both drought forecasting models and drought indices to detect drought signals. These metrics are well-equipped to detect drought once it is under way, but do not detect drought at spatial and temporal scales conducive to early detection and mitigation. While existing methods have improved our ability to detect drought, it is still difficult to predict and detect early warning signs. This study works to develop an early drought warning signal that can be incorporated into an Early Drought Warning system to assist in informing at risk populations, such as farmers, stakeholders, and landowners, of impending drought. Our work focuses on quantifying physiological plant stress, as a result of drought intolerance, prior to drought through the decoupling of evapotranspiration and vapor pressure deficit via remotely sensed variables that can be detected at a near real time scale.
Results/Conclusions
Several plant-climate interactions aggravate drought conditions. The interaction of interest for this study is the decoupling of evapotranspiration (ET), a key ecosystem variable controlled by plant transpiration and the evaporation of water, and atmospheric vapor pressure deficit (VPD), an indicator of moisture saturation of the air. As drought develops, ET becomes decoupled from its drivers, including VPD. We propose that this decoupling is indicative of physiological plant stress, such as drought stress. This study aims at capturing this effect of preceding drought conditions in the Midwest of the United States during the drought of 2012 while comparing to non-drought years through Flux Tower data. We find that during drought conditions, peak ET shifts to earlier in the day, temporally decoupling from peak VPD and peak solar radiation. This effect occurs at a critical point preceding drought conditions, aptly named “the Critical Climate Period.” Our goal for this study is to identify the temporal nature of this decoupling effect to uncover its relationship with both the timing of drought and the severity of its impacts. Therefore, we propose that this work, when coupled with near real time satellite data, will allow for improved drought detection in existing drought monitoring and detection frameworks.