Communicated by V. Kerry Smith, Arizona State University, Tempe, AZ, July 1, 2009 (received for review October 13, 2008)
The United States produces 41% of the world's corn and 38% of the world's soybeans. These crops comprise two of the four largest sources of caloric energy produced and are thus critical for world food supply. We pair a panel of county-level yields for these two crops, plus cotton (a warmer-weather crop), with a new fine-scale weather dataset that incorporates the whole distribution of temperatures within each day and across all days in the growing season. We find that yields increase with temperature up to 29° C for corn, 30° C for soybeans, and 32° C for cotton but that temperatures above these thresholds are very harmful. The slope of the decline above the optimum is significantly steeper than the incline below it. The same nonlinear and asymmetric relationship is found when we isolate either time-series or cross-sectional variations in temperatures and yields. This suggests limited historical adaptation of seed varieties or management practices to warmer temperatures because the cross-section includes farmers' adaptations to warmer climates and the time-series does not. Holding current growing regions fixed, area-weighted average yields are predicted to decrease by 30–46% before the end of the century under the slowest (B1) warming scenario and decrease by 63–82% under the most rapid warming scenario (A1FI) under the Hadley III model.
Author contributions: W.S. and M.J.R. designed research, performed research, contributed new reagents/analytic tools, analyzed data, and wrote the paper.
The authors declare no conflict of interest.
This article contains supporting information online at www.pnas.org/cgi/content/full/0906865106/DCSupplemental.
↵* We estimated models with interactions between temperature and rainfall on a shorter time scale, but these models do not predict out-of-sample significantly better than the additively separable model reported above. It is possible that the relatively poor predictive power of precipitation, in comparison with temperature, stems from greater measurement error in the precipitation variable as spatial interpolation is more difficult.
↵† In the SI Appendix, we show that shifting the growing season one month earlier does little to mitigate yield loss.
↵‡ Because access to subsidized water rights is correlated with climate, omitting these variables, which vary on the subcounty level of irrigation districts, will result in biased coefficient estimates on the climatic variables in a cross-sectional analysis (18). Across all counties, 58% of cotton acres, 18% of corn acres and 9% of soybean acres are irrigated.
↵§ An alternative method to approximate the distribution of daily temperatures from the distribution of monthly average temperatures is developed by ref. 23. This method appears appropriate for predicting the average frequency that a certain weather outcome will be realized but less appropriate in predicting a specific frequency of a weather outcome in a particular year. Thom's method works well in a cross-sectional analysis where the dependent variable is tied to expected weather outcomes (for example, the link between land values and climate). The method is less well suited to our analysis where the dependent variable (yield) linked to specific weather outcomes.
↵¶ The northern subset includes counties in Illinois, Indiana, Iowa, Michigan, Minnesota, New Jersey, New York, North Dakota, Ohio, Pennsylvania, South Dakota, and Wisconsin. Interior counties are in Delaware, Kansas, Kentucky, Maryland, Missouri, Nebraska, Virginia, and West Virginia. The southern counties are in Alabama, Arkansas, Georgia, Louisiana, Mississippi, North Carolina, Oklahoma, South Carolina, Tennessee, and Texas.
