We use watershed modeling and field sampling to investigate how changes in land use might effect runoff and river water quality in intensively managed agricultural watersheds.  We are currently investigating how the location of a conservation practice (such as a constructed wetland or field cover crops) might change its effectiveness at reducing nutrient or sediment runoff. Current research projects, funded by USDA, are using machine learning methods to upscale and emulate models from smaller watersheds to improve predictions for larger watersheds in the Upper Mississippi River Basin. In addition to watershed modeling, we are investigating hysteresis between water flow rates and concentrations of nitrate and suspended sediment using environmental sensor data. From this analysis we aim to better understand the contributions of hydrological and biogeochemical processes at an event time scale.