Agriculture is highly susceptible to changes in weather systems, particularly sub-annual changes in temperature and precipitation. Thus, impacts of meteorological variables on crop yields have been widely studied in the empirical economic literature and by process modelers via global gridded crop models (GGCMs). Both approaches have limitations. Past empirical efforts have not been able to model differentiated yield responses to temperature and precipitation for rainfed versus irrigated crops and GGCMs can be difficult to calibrate to field conditions in a broad range of climates and are computationally expensive to run.

We present the results of an empirical modeling approach to differentiate the year-to-year yield responses of rainfed and irrigated maize, soy, and winter and spring wheat to temperature and precipitation (via soil moisture) in key crop-specific sub-annual growth stages, using historical U.S. county-level data. One key novelty of this work is the use of meteorological values and soil moisture at biophysically important growth stages. This differs from prior work that has taken a purely reduced form approach, using measures such as growing degree days, maximum, minimum, or mean values over the course of a year or growing season, and results in a greater understanding of annual yield response to sub-annual weather conditions for a variety of crops and management practices. This approach also allows users to define their own adaptation growth stage calendars appropriate for testing various hypotheses. Finally, the logarithm of yield was used in fitting the model so that temperature and precipitation impacts on yearly yield can be isolated for use in future projections of yield impacts under changing climate patterns, including climate variability.

This work joins past crop modeling efforts as a valuable tool to impacts modelers for empirically estimating differentiated annual yield responses to sub-annual temperature and precipitation values for rainfed versus irrigated crops. It highlights that crop-specific biophysical information can allow for more flexible crop response models for use in uncertainty analysis, model linkage studies, and integrated assessment models, particularly along different adaptation pathways.