Skip to main content
ARS Home » Northeast Area » Beltsville, Maryland (BARC) » Beltsville Agricultural Research Center » Hydrology and Remote Sensing Laboratory » Research » Publications at this Location » Publication #380466
Research Project: Integrating Remote Sensing, Measurements and Modeling for Multi-Scale Assessment of Water Availability, Use, and Quality in Agroecosystems

Location: Hydrology and Remote Sensing Laboratory

Title: Predicting rapid changes in Evaporative Stress Index (ESI) and soil moisture anomalies over the continental United States

Author
item LORENZ, D. - University Of Wisconsin
item OTKIN, J. - University Of Wisconsin
item ZAITCHIK, B. - Johns Hopkins University
item HAIN, C. - Nasa Marshall Space Flight Center
item Anderson, Martha

Submitted to: Journal of Hydrometeorology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 8/30/2021
Publication Date: 9/1/2021
Citation: Lorenz, D., Otkin, J., Zaitchik, B.F., Hain, C., Anderson, M.C. 2021. Predicting rapid changes in Evaporative Stress Index (ESI) and soil moisture anomalies over the continental United States. Journal of Hydrometeorology. 22(11):3017–3036. https://doi.org/10.1175/JHM-D-20-0289.1.
DOI: https://doi.org/10.1175/JHM-D-20-0289.1

Interpretive Summary: The ability to accurately predict degradation in moisture conditions in agricultural lands one to four weeks out will be of significant value in activating drought mitigation responses, both at the farm and regional level. However, drought forecasting has proven to be complex and unreliable, especially for “flash drought” events such as occurred in the U.S. Corn Belt in 2012. This paper uses a hybrid statistical dynamical approach to explore the predictability of rapid changes in soil moisture and evapotranspiration (ET) or crop water use, both of which are widely used in monitoring flash drought. The results indicate that soil moisture changes are reasonably predictable out to two weeks. Changes in ET were less skillfully forecasted, being a more integrative measure of both soil water availability (the soil evaporation component) and crop health (the transpiration component). However, if ET can be partitioned into these components, then forecasts in rapid changes can be improved. These results will inform development of improved drought forecasting systems that are better tuned for agricultural applications.

Technical Abstract: Probabilistic forecasts of changes in soil moisture and an Evaporative Stress Index (ESI) on sub-seasonal time scales over the contiguous U.S. are developed. The forecasts use the current land surface conditions together with numerical weather prediction forecasts from the Sub-seasonal to Seasonal (S2S) Prediction Project. Changes in soil moisture are quite predictable two weeks in advance with 50% or more of the variance explained over the majority of the contiguous U.S.; however, changes in ESI are significantly less predictable. A simple red noise model of predictability shows that the spatial variations in forecast skill are primarily a result of variations in the autocorrelation, or persistence, of the predicted variable, especially for the ESI. The difference in overall skill between soil moisture and ESI, on the other hand, is due to the greater soil moisture predictability by the S2S model forecasts. As the forecast lead time increases from 8-14 days to 15-28 days, however, the autocorrelation dominates the soil moisture and ESI differences as well. An analysis of modelled transpiration, and bare soil and canopy water evaporation contributions to total evaporation, suggests improvements to the ESI forecasts can be achieved by estimating the relative contributions of these components to the initial ESI state. The importance of probabilistic forecasts for reproducing the correct probability of anomaly intensification is also shown.