Nitrogen (N) leaching to surface waters from grain farms in the Mississippi River Basin (MRB) is the primary cause of hypoxia in the Gulf of Mexico. I constructed N mass balances for a gradient of farm types to explore the relationship between agroecosystem management practices and potential N loss on 95 farms in the upper MRB. Nitrogen balances ranged from high average annual surpluses (149 kg N ha-1yr-1) to large deficits (80 kg N ha-1yr-1). Fields with greater than 50% of total N additions from legume N sources, and fields with crop rotations that included both annual and perennial species, were approximately in balance, compared to fertilizer-based practices in corn-soybean rotations with surpluses of 35 kg N ha-1yr-1. To explore how a subset of farmers in Iowa transitioned to practices with the greatest promise for reducing N losses, I analyzed qualitative interviews conducted with farmers between 2008 and 2010. I identified resources and strategies they harnessed to develop opportunities for, and overcome barriers to, transitioning to alternative practices within the context of the industrialization of agriculture in the MRB. To enhance resilience and mitigate risk, alternative farmers increased farm-level biodiversity and enterprise diversity. They developed new competencies such as ecological thinking and cultivated external network linkages with peers, knowledge organizations, and federal policy. Finally, I conducted a 15N-tracer experiment in Illinois to investigate the biogeochemical mechanisms of one ecological practice that has promise for reducing N losses, use of winter annual cover crops. I applied 15N-labeled ammonium sulfate fertilizer to corn, and compared winter rye (Secale cereal) cover and bare fallow treatments following corn harvest. After one year, total recovery of 15N ranged from 37-45%. Due to unfavorable weather conditions rye biomass was low and little 15N was recovered in the rye. However, the cover crop significantly reduced soil inorganic N pools in the spring (11.1 kg N ha-1 in bare fallow compared to 1.9 kg N ha-1 with rye cover), by an amount similar in magnitude to total rye N uptake (23.7 kg N ha-1), indicating that cover crops scavenge inorganic N mineralized from soil organic matter pools.