A study of the Boone River Watershed: linking an agroecosystems model with a process-based Hydrologic Model

Modern agricultural practices in Iowa have maximized production, often at the expense of flood resiliency and water quality. Various practices, including planting cover crops and small grains, or applying fertilizer at recommended reduced rates, have been proven to reduce flooding and nitrogen loads. However, adoption of these strategies has been slow, flooding continues to threaten Iowa communities, and water quality has worsened. It can also be difficult to predict the watershed-scale impact of implementing mitigation strategies. Therefore, we linked the Agricultural Production Systems Simulator (APSIM) with the Generic Hydrologic Overland-Subsurface Toolkit (GHOST) to use the crop and nutrient dynamics of an agroecosystems model to inform the land use inputs of a process-based hydrologic model. We modeled the Boone River Watershed from 2002 to 2018, and calibration achieved a Nash-Sutcliffe Efficiency of 0.70 and R2 of 0.72. Next, we modified the model to incorporate alternative farming scenarios, and observed the hydrologic and water quality impacts. Nitrate loads were estimated using an end-member mixing analysis (EMMA) with nitrate concentrations from APSIM and discharges from GHOST. Adding cover crops to a corn-soybeans rotation reduced the total nitrate load by 41%. This reduction increased to 47% and 53% in four-year (corn-soybeans-oats-alfalfa) and three-year (corn-soybeans-oats) rotations, respectively. Applying fertilizer at the recommended rate reduced total nitrate load 22%. All three scenarios halved the number of floods exceeding the 2-year threshold, reduced the June 2008 flow below the 50-year threshold, and decreased the exceedance probability of moderate and large floods.