Representing Extreme Temperature Events and Resolving Their Implications for Yield

Abstract:

Although it is well recognized that extreme temperatures occurring at particular growth stages are destructive to yield, there appears substantial scope for improved empirical assessment and simulation of the relationship between temperature and yield. Several anecdotes are discussed.

First, a statistical analysis of historical U.S. extreme temperatures is provided. It is demonstrated that both reanalysis and model simulations significantly differ from near-surface temperature observations in the frequency and magnitude of extremes. This finding supports empirical assessment using near-surface instrumental records and underscores present difficulties in simulating past and predicting future changes.

Second, an analysis of the implications of extreme temperatures on U.S. maize yield is provided where the response is resolved regionally and according to growth stage. Sensitivity to extreme temperatures during silking is found to be uniformly high across the U.S., but the response during grain filling varies spatially, with higher sensitivity in the North. This regional and growth-stage dependent sensitivity implies the importance of representing cultivar, planting times, and development rates, and is also indicative of the potential for future changes according to the combined effects of climate and technology.

Finally, interaction between extreme temperatures and agriculture is indicated by analysis showing that historical extreme temperatures in the U.S. Midwest have cooled in relation to changes in regional productivity, possibly because of greater potential for cooling through evapotranspiration. This interpretation is consistent with changes in crop physiology and management, though also noteworthy is that the moderating influence of increased evapotranspiration on extreme temperatures appears to be lost during severe drought.

Together, these findings indicate that a more accurate assessment of the historical relationship between extreme temperatures and yield may be obtained through judicious use of instrumental data, resolving regional and stage-based responses, and dynamically accounting for the interactions between temperature extremes and crop type. These considerations should also be informative for accurate simulation and prediction of future changes in yield.