The watersheds of the Des Moines lobe in north central Iowa have fundamentally changed in the last 170 years. Where there was once prairie, row crop agriculture now dominates. This progress has enabled this small region of the Midwest to provide food, fuel, feed, and fiber for millions, but with recent flooding events of 2008 and 2010 questions have been raised about the hydrological impacts of these lands. These flood events are driven by peak flow and concerns about the effect of drainage on peak flow should be investigated. MIKE SHE, a watershed scale model, was used to simulate daily streamflow in a multisite comparison to determine if the model is suitable to simulate streamflow in heavily drained agricultural land. The model was tested for five years (2007-2011) in two similar watersheds (1127 ha and 1356 ha) in Palo Alto County, Iowa. In the testing watershed, the simulated streamflow correlated well with the observed streamflow, as shown by a daily Nash-Sutcliffe coefficient of 0.62 and a coefficient of determination of 0.66. Likewise, the model performed well in the validation watershed with a daily Nash-Sutcliffe coefficient of 0.73 and coefficient of determination of 0.79. This shows that the model can be used in the future to simulate flow in similar agricultural regions throughout the Midwest that employ tile drainage to maintain suitable water table levels needed for crop growth.

Changes in land use management and drainage design were simulated to better understand the hydrological impact that land use and tile drainage has on the landscape. It was shown that if row crops are converted to pasture or prairie, with drainage infrastructure intact, evapotranspiration would increase by 25% and the magnitude of peak events would decrease by over 50% in some cases. Likewise, if the drainage infrastructure was removed and only perennial grassland remained, similar to likely pre-settlement conditions for the region, then water table height becomes the main driver of surface flow and overall flow from both watersheds would decrease by 55%. Alternatively, if the depth of tile drains were decreased from 1.2m to 0.75m the effect would allow for 7 to 20 mm of extra surface runoff, while decreasing subsurface flow and maintaining the total flow. Lastly, if all drainage infrastructures were removed from the watersheds and row crop monoculture were to be maintained there would be an increased frequency of peak flow that may lead to damaging flood events. These results show that MIKE SHE could be used in land use management decisions and assessment of drainage design for mitigation of hydrological impacts downstream of heavily drained agricultural watersheds. This may help target land areas for wetland placement by showing the effects of eliminating drainage structure will have on the watershed.