Changes in diurnal temperature range and national cereal yields

Abstract

Models of yield responses to temperature change have often considered only changes in average temperature ( T avg) with the implicit assumption that changes in the diurnal temperature range (DTR; the difference between daily maximum and minimum temperature) can safely be ignored. The goal of this study was to evaluate this assumption using a combination of historical datasets and climate model projections. Data on national crop yields for 1961–2002 in the ten leading producers of wheat ( Triticum spp.), rice ( Oryza spp.) and maize ( Zea mays) were combined with datasets on climate and crop locations to evaluate the empirical relationships between T avg, DTR and crop yields. In several rice and maize growing regions, including the two major nations for each crop, there was a clear negative response of yields to increased DTR. This finding reflects a nonlinear response of yields to temperature, which likely results from greater water and heat stress during hot days. In many other cases, the effects of DTR were not statistically significant, in part because correlations of DTR with other climate variables, and the relatively short length of the time series resulted in wide confidence intervals for the estimates. To evaluate whether future changes in DTR are relevant to crop impact assessments, yield responses to projected changes in T avg and DTR by 2046–2065 from 11 climate models were estimated. The mean of climate model projections indicated an increase in DTR in most seasons and locations where wheat is grown, mixed projections for maize, and a general decrease in DTR for rice. These mean projections were associated with wide ranges that included zero in nearly all cases. The estimated impacts of DTR changes on yields were generally small (<5% change in yields) relative to the consistently negative impact of projected warming of T avg. However, DTR changes did significantly affect yield responses in several cases, such as in reducing US maize yields and increasing India rice yields. Because DTR projections tend to be positively correlated with T avg, estimates of yield changes for extreme warming were particularly affected by including DTR (up to 10%). Finally, based on the relatively poor performance of climate models in reproducing the magnitude of past DTR trends, it is possible that future DTR changes and associated yield responses will exceed the ranges considered here.