Observed trends in daily rainfall variability result in more severe climate change impacts to agriculture

Abstract

There is increasing evidence that climate change is impacting not just total amounts of precipitation, but its temporal dynamics as well. While previous studies have identified the importance that the temporal distribution of daily rainfall has on crop production, climate models often do not represent these distributions consistently with observed trends. The objectives of this study are to (1) evaluate the relationship between rainfall variability and yields for economically important crops in the upper Southeastern United States and (2) assess the potential impact that incorporating observed trends in rainfall variability has on yield projections under future climate change. This study develops statistical models of historic crop yields for five crops, finding that an explanatory variable related to daily rainfall variability, the wet-day Gini coefficient (GC), has a statistically significant negative relationship with crop yields in all cases. These models are then used to estimate the impacts of climate change using an ensemble of downscaled general circulation model (GCM) projections and scenarios that include a continuation of observed GC trends. Most downscaled GCMs evaluated do not project changes in GC consistent with observed trends, and scenarios that assume a continuation of observed trends result in projected yields that are up to 5.8% lower than those directly based on GCM projections. While additional research is needed in the climate science community to better understand how rainfall variability may change in the future, this should be mirrored in the impacts community so that agricultural impact assessments incorporate these potentially important changes.