Probabilistic estimates of drought-induced yield loss in the Southeastern United States

Authors: KHEDUN, PRAKASH - Clemson University; Sohoulande Djebou, Dagbegnon

Submitted to: Agricultural Systems
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 6/2/2025
Publication Date: 6/20/2025
Citation: Khedun, P., Sohoulande Djebou, D.C. 2025. Probabilistic estimates of drought-induced yield loss in the Southeastern United States. Agricultural Systems. 229. https://doi.org/10.1016/j.agsy.2025.104418.
DOI: https://doi.org/10.1016/j.agsy.2025.104418

Interpretive Summary: Rainfed agriculture feeds the world; but it is vulnerable to an increasing drought severity and frequency under climate changes. A better understanding of drought effects on crop yields is required to plan mitigation strategies. However, the linkages between drought and crop yield variability are complex, suggesting the need of various approaches to investigate the phenomenon. Here, we use a probabilistic method to analyze the effect that drought poses to four major cash crops in the Southeastern US including corn, cotton, peanuts, and soybean. We use the probabilistic method to model the linkages between a drought indicator and crop yield at the county level. We then quantified the drought-induced yield loss in each county and determined the joint and conditional probability of yield loss for different drought intensities. We show that corn is the most vulnerable—around 90% of counties lost at least 1 t/ha during drought—while the loss in yield for the other crops were lower. Further, we found that counties in southern GA suffered lower losses comparatively, possibly because a growing number of farmers have been adopting irrigation practices following the 2005/06 drought episode in the US. We also computed the conditional probability of yield loss given an impending drought to show which counties are more vulnerable. The methodology developed in this study can be incorporated in a dashboard or decision support system that farmers, planners, and economists can use to make informed decisions on the risk that drought poses in their counties and state and develop adequate mitigation plans including the estimation of insurance and compensation.

Technical Abstract: Rainfed agriculture feeds the world; but it is vulnerable to drought, and drought severity and frequency are increasing as the climate changes. Therefore, a better understanding of their effects on yield is required to guide the design of appropriate mitigation strategies. The linkages between drought and its drivers and crop health and yield are complex, thus warranting a multivariate approach. Here, we use a probabilistic method to analyze the threat that drought poses to four major cash crops: corn, cotton, peanuts, and soybean, in the Southeastern US. We use copula to model the dependence, at the county level, between the Standardized Precipitation and Evapotranspiration Index (SPEI), a meteorological drought indicator, and crop yield. We then quantified the drought-induced yield loss in each county and determined the joint and conditional probability of yield loss for different drought intensities. We show that corn is the most vulnerable—around 90% of counties lost at least 1 t/ha during drought—while the loss in yield for the other crops were lower. Further, we found that counties in southern GA suffered lower losses comparatively, possibly because a growing number of farmers have been adopting irrigation practices following the 2005/06 drought. We also computed the conditional probability of yield loss given an impending drought to show which counties are more vulnerable. The methodology developed in this study can be incorporated in a dashboard or decision support system that farmers, planners, and economists can use to make informed decisions on the risk that drought poses in their counties and state and develop adequate mitigation plans including the estimation of insurance and compensation.