A strategy is needed for translating changes and variability of global climate into impacts on agriculture. Global models, which are needed to evaluate global consequences of anthropogenic influences such as changes in greenhouse gas concentrations, do not supply climate information of sufficient resolution to meet this need. Regional climate modeling is conceptually more appealing than statistical downscaling, and captures some fine-scale dynamical processes that are unresolved by global models. However, inaccurate information on the timing and magnitude of precipitation events, and biases in temperature produced by the regional models are amplified by a crop model to produce large discrepancies between simulated and observed yields. If, as data seem to show, crops are more efficient than native perennial vegetation at recycling moisture to the atmosphere, then the physiological processes represented within crop models are needed within the climate model to more accurately represent the linkage between soil and atmosphere for simulating climates in intensively cultivated areas. For areas of intensive agriculture, such as the U.S. Midwest, these advances in climate modeling will be particularly beneficial.
Was this article helpful?