Hunter, Luyi2020-02-262020-02-262019-11https://hdl.handle.net/11299/211761University of Minnesota Ph.D. dissertation. November 2019. Major: Bioproducts/Biosystems Science Engineering and Management. Advisor: Timothy Smith. 1 computer file (PDF); x, 123 pages.Intensive crop farming enables the food production system to meet the growing food needs from the boomed population. Globally, crops cultivated with synthetic nitrogen fertilizers feed more than 50% of the people, and the 16% irrigated croplands provides approximately 36% of global harvest. At the same time, millions of arable acres are being converted to urban areas every year, and global warming keeps disrupting the crop system. Existing croplands must guarantee consistent yield increase to keep up with the further increase of food demands under challenging climatic conditions, which leads to more nutrient inputs and increases risks of soil erosion, fresh water contamination, and excessive atmospheric greenhouse gas emissions. The goal for agricultural development in 21st century is increasing crop productivity in parallel with reducing adverse impacts on the natural environment, as defined in the concept of sustainable agriculture. Corn (Zea mays) will potentially play a leading role in the battle against hunger because of its high productivity and broad versatility. The United States is the largest corn producer and exporter in the world. While generating billions of dollars of revenue, the US corn production system has been questioned for its efficiency and blamed for environmental & societal impacts it generates. To enhance the resilience and sustainability of US corn production, practitioners need an integrated tool for: 1) consolidating publicly available information on corn, soil, climate, and agricultural practices to to approximate the current environmental conditions and management strategies on US corn fields; 2) predicting field performance on yields and environmental impacts to shed light on the future state of the US corn system; 3) simulating bio-chemical processes of corn fields in sufficient resolution and scalability to allow comparison and optimization of farming practices; 4) translating farming inputs, crop outputs, environmental and societal externalities into dollar values for easier interpretation and better application in decision making. This dissertation designed and implemented metamodels to tackle these four tasks by investigating the status quo of US corn systems, as well as proposing spatially specific agricultural practices for sustainable corn farming.enCrop System SimulationDNDCMachine LearningMaizeMetamodelThesis or Dissertation