Crop production and future climate change in a high latitude region: A case study for the Upper Great Lakes region of the United States

Agriculture is particularly susceptible to climate change, as inferred from large historical variations in crop production in response to past climate variability. The major goal of this dissertation is to evaluate the spatial variability of the impacts of projected future climate change on crop production in a high latitude region. Corn and soybean production in the Upper Great Lakes Region (UGLR) of the United States, encompassing the states of Michigan, Wisconsin and Minnesota, serves as the case study. The CERES-Maize and CROPGRO-Soybean models included in the Decision Support System for Agrotechnology Transfer (DSSAT) were employed to simulate county-level crop production for current and future time slices.;A first step in the analysis was to assign individual counties to nearby climate stations, as climate data (e.g., temperature and precipitation) are primary drivers for the DSSAT simulations. A climate regionalization procedure was applied to group climate stations from the United States Historical Climate Network into subgregions with similar characteristics of the annual cycle of temperature and precipitation. Representative stations were chosen from each climate region, and counties were assigned to the closest representative station. The regionalization ensured that the spatial gradients of temperature and precipitation across the UGLR were captured in the analysis.;Daily solar radiation, also an important variable for the DSSAT simulation but one which is infrequently recorded, was calculated using a mechanistic solar radiation model parameterized at a central site within the study area for which concurrent observations of solar radiation, temperature and precipitation were available. This model was selected after systematically evaluating the sensitivity of simulated corn and soybean yield to different solar radiation sources.;Climate change projections for a mid century (2041--2070) time slice were derived from eight combinations of regional climate models and global climate models released by the North American Climate Change Assessment Program. The impact assessment revealed that by the mid century, assuming the current CO2 level, corn and soybean yield for the northern part of the region likely will increase due to more favorable growing season conditions than at present; whereas, a slight decrease in yield due to shorter time to maturity may occur in the southern part of the UGLR, which is currently a major crop production region. However, under elevated CO2 concentrations, the number of counties with a projected decrease in yield is smaller due to the positive impacts of higher CO2 levels, particularly for soybean production.;As the DSSAT simulations do not capture the effects of pest and disease infestations or economic factors on crop yield, a prototype interdisciplinary model for corn yield was developed using an asymmetric production function to integrate the DSSAT simulated yields with economic determinants (i.e., costs of pesticides, machinery, labor and fertilizer) to estimate observed county-level yield. The interdisciplinary model was shown to be a viable alternative to classical yield functions, and this model framework merits future study and refinement.;In general, this study suggests that more favorable growing conditions by the mid century will benefit the northern UGLR where some areas may produce relatively high corn and soybean yields especially under elevated CO 2 concentrations. Despite a potentially small reduction in crop yields, current crop production in the southern UGLR will remain large, partially due to the positive impact of elevated CO2 concentrations.