资源环境科技发展态势分析平台

Climate-yield relationships for winter wheat were examined across primarily dryland agricultural systems for counties in the Columbia Plateau of the northwestern USA from 1980 to 2014. Interannual linear climate-yield relationships were assessed at subregional scales with climate variables of energy and moisture using temperature, precipitation, heat stress, and water balance metrics for varying wheat phenostages. Interannual variability in moisture availability exhibited significant relationships with wheat yields across the Columbia Plateau, with the strongest relationships during the latter phenostages. Actual evapotranspiration generally exhibited the highest correlation with interannual yield variability, with the strongest relationships occurring in counties with intermediate amounts of precipitation. Crop yields were negatively affected by warmer temperatures during latter phenostages, particularly in clima to logically cooler counties where temperature-sensitive phenostages occur later in the calendar year when temperatures are higher. Linear stepwise regression models were developed separately for each county, as well as using a single pooled model that is universal for all counties. County-level models explained an average of 37% of the interannual variability in yield, with more explained variance in counties without widespread irrigation. The pooled model yielded generally similar results, but explained an average of 10% less variance than county-level models across the study region, with substantially less explained variance in counties that exhibited strong relationships with temperature. These findings demonstrate the utility of evaluating climate-yield relationships at subregional scales across agricultural regions that traverse climatic gradients.