Abstract
Understanding the climatic influences on the interannual variability in crop yields and how they vary spatially across climate gradients, can be of significant value for forecasting yields. County level climate-yield relationships for winter wheat were examined across a moisture gradient over primarily rainfed agricultural systems in the Columbia Basin of the United States. Wheat yields were most strongly correlated with energy and moisture availability during the latter stages of crop development. Estimated actual evapotranspiration was typically the best predictor of interannual yield variability, with the strongest relationships for counties with intermediate amounts of mean annual precipitation. Crop yields were negatively impacted by warmer temperatures during the latter stages of crop development, particularly in the climatologically cooler counties as delayed crop phenology results in warmer phenostage temperatures when crops are most sensitive. A variety of multi-variate statistical models explain an average of 29-37% of interannual county-level yield variance over the Columbia Basin, yet show spatial heterogeneities in climate yield relationships suggesting the importance of subregional climate-crop modeling.