Assessing the impact of climate change on crops requires an overview of the various aspects of climate that are projected to change and an evaluation of the individual and combined effects of these changes. Uncertainty in our understanding of the impacts of climate change on crop production arises from our uncertainty about change in the climate itself and also from uncertainty in translating those changes into changes in crop production. In this chapter, we provide an overview of our ability to simulate future scenario climates at regional scales with a particular eye toward the climate needs for crop modeling.
Food production is a global enterprise that links production and crop failures in one region with global markets and farmer choices in far distant regions. However, for this chapter we will focus on the United States, where climate data of relatively good spatial and temporal resolution are available over a large region and where regional climate modeling has been done with applications to crop modeling and yields.
Current efforts to study impacts of changing climate on sustainability and productivity of intensive agriculture in the U.S. Midwest are hampered by scale of resolution of future climate information, by use of a single routine to represent all crops, and by limited representation of hydrological processes to account for water and nutrient transport from individual fields, slopes, and watersheds (Mearns et al., 2001; Tsvetsinskaya et al., 2001). In particular, Mearns et al. (2001) evaluated the impact of spatial variability of both climate and soil properties on crop yield, and concluded that, compared to climate, the spatial scale of soil properties had a larger impact on variance and autocorrelation of yields, but a smaller impact on mean yields. From these results on yields, we can infer that the spatial variability of soil parameters also will have large impact on sustainability factors such as nitrogen loss, sediment loss, and change in soil organic matter content.
The impact on crops of changes in temperature and precipitation cannot be separated from impacts of other global changes. Shaw et al. (2003) evaluated the combined effects of warming, increased precipitation, increased nitrogen deposition, and increased CO2 on California annual grassland, and found that the first three alone and together increased net primary production. Increased CO2 alone also increased net primary productivity (NPP). However, in combination with other simulated changes, increased CO2 generally decreased the enhancing impact of other factors, suggesting that single factor changes on ecosystems are not simply additive. This may be a consequence of the reduced root allocation by grasses grown in enhanced CO2 environments. Shaw et al. (2003) concluded that multifactor responses must be evaluated. Analogous studies need to be conducted on agricultural crops to evaluate whether enhanced yields projected under increases in atmospheric CO2 are realistic when taken in combination with other global change factors. Mearns et al. (1997) examined the impact of changes in both mean and variance of climate on output of a crop model, and demonstrated the importance of including variability.
Changnon and Hollinger (2003) analyzed data from an experiment in Illinois that artificially enhanced rainfall for each precipitation event during the growing season. Rains enhanced by 10% to 40% had little effect except in dry years. Furthermore, all increases had positive effects in a dry year and negative effects in a wet year. They also pointed out the important role of timing of precipitation events and temporal sequences of temperature throughout the growing season. Extrapolating their results to explore impacts of possible future increases in precipitation, they conclude that rainfall increases of 10% will have little impact on corn yields except in wet years when such increases may actually lower yields. They project rainfall increases of 25% and 40% to increase yields by 3% and 9%, respectively, except in unusually dry years. These increases are much less than the 15% yield increase for corn projected for 2030 by the National Assessment Synthesis Team (2000).
From these and other recent studies, we conclude that significant challenges lie ahead as we attempt to quantify food-production security and sustainability more realistically in future scenarios of climate. In this chapter, we present a brief overview of the current status and limitations of our ability to construct accurate simulations of future scenario climates at the scale of resolution needed to assess regional impacts. We also examine how inaccuracies in some critical climate inputs to crop models are reflected in the yields they produce.
Was this article helpful?