CONSTRUCTING A SWAT MODEL OF THE ST. CROIX BASIN

James E. Almendinger, St. Croix Watershed Research Station, Science Museum of Minnesota, Marine on St. Croix, MN
Xuesong Zhang, Texas A&M University, College Station, TX and Joint Global Change Research Institute, Pacific Northwest National Laboratory, MD

The ecological integrity of the St. Croix River has been compromised by inputs of sediment and nutrients, especially phosphorus, from the drainage basin. To partially restore the river, both states have agreed to reduce phosphorus loads by 20% relative to those in the 1990s. Previous studies have shown that anthropogenic nonpoint-source inputs of phosphorus are several times larger than point-source inputs and thus where greater reductions in phosphorus could be achieved. The purpose of this project, funded by the National Park Service, is to construct a computerized watershed model that can help resource managers choose the most effective means of reducing nonpoint-source phosphorus. This model of the St. Croix basin will provide a framework for making sound, science-based decisions in how best to restore and protect the St. Croix River, thus securing the nation's investment in the St. Croix National Scenic Riverway. We here report current progress in constructing such a watershed model for the 20,000-km² St. Croix basin using the Soil and Water Assessment Tool (SWAT), a program developed by the USDA to predict nonpoint loads of sediment and nutrients from large basins over long periods of time. SWAT can handle both point and nonpoint sources, in both urban and rural landscapes; its strength, however, is in simulating sediment and nutrient exports from agricultural lands. Data sets for the model of the St. Croix are now being assembled. Based on topographic data, the stream network, and monitoring locations, the basin has been subdivided into 212 subbasins of about 100 km² each. Climate data have been compiled from the cooperative-station network (73 precipitation and 59 temperature stations) and spatially interpolated across the basin, to construct a record of daily precipitation and temperature from 1960 through 2008 for pseudo-stations at the centroid of each delineated subbasin. The model will also include 48 point sources and the 29 largest reservoirs in the basin. The next steps will be to configure the soils and land-use data into "hydrologic response units" (HRUs). Then the model will be tested for parameter sensitivity and calibrated to monitoring data.