Weather and Yield, 1950-94: Relationships, Distributions, and Data

Nonlinear functions of weather explain almost all significant variations in crop yield. Corn, soybean, and spring wheat responses were estimated. Yields respond to monthly temperature and precipitation during May - September. Farm size is also important. Farms that harvest more acreage get higher yields. The yield distribution depends on the probability distribution of weather and the nonlinear response function. Temperature data follow a normal (symmetric) probability distribution, while precipitation data are better described by a (nonsymmetric) gamma distribution. The probability distribution for crop yields contains components that are normal, log-normal, chi-squared, and other random variables. Monthly temperature and precipitation data for 1950-94 are aggregated to State and regional estimates, using either harvested cropland or geographic area in 344 sub-State climatic divisions.