Applying a SWAT model of the Sunrise River watershed, eastern Minnesota, to predict water-quality impacts from land-use changes. Report to the Minnesota Pollution Control Agency.

The Sunrise River watershed has at least four river reaches and ten lakes listed as impaired by the Minnesota Pollution Control Agency. These impairments are likely the result of nonpoint-source loads of sediment and nutrients, among other constituents. To better identify the sources of nonpoint loads, how they are transported to the receiving waters, and how they might be reduced, a computer watershed model of the Sunrise River watershed was constructed with the Soil and Water Assessment Tool (SWAT). The purpose of this project was to apply the SWAT model (revised in autumn of 2011) to selected land-use change scenarios in the Sunrise River watershed and quantify the resulting sediment and phosphorus loads. Four sets of scenarios were modeled: (1) changes from projected population growth, (2) changes in agricultural practices, (3) changes in urban practices, and (4) changes from wetland mitigation. By the year 2030, population in the Sunrise watershed could increase from 66,000 to 120,000, causing an increase in developed lands from 16% (current) to 24% of the total watershed area. Phosphorus loads to rivers and lakes within the watershed would increase by 7%, and the phosphorus load from the Sunrise to its receiving water, the St. Croix River, would increase by 5%. Lakes nearest expanding urban centers would receive the largest phosphorus-load increases, commonly exceeding 10%. These lakes would benefit from urban best-management practices (BMPs); however, SWAT was not effective in simulating such BMPs. The model was more suited to simulating agricultural BMPs, especially those that reduced phosphorus content in runoff by reducing soil-test phosphorus levels (up to 20% reduction in loads) and those that treated runoff in grassed waterways (18% reduction) or vegetated filter strips (11% reduction). These reductions assume full implementation on every corn, soybean, and alfalfa field, which is unlikely, but partial implementation could still result in substantial load reductions. No-till scenarios were much more effective at reducing sediment loads than phosphorus. Wetland restoration or routing more runoff through existing wetlands could result in substantial phosphorus load reductions, up to nearly 20% at the watershed outlet and within the Chisago Lakes Improvement District. Overall we conclude that reducing nonpoint loads of phosphorus is feasible, but that there is no easy solution. To attain the largest reductions in phosphorus load would require substantial land modification, either as agricultural BMPs or wetland restoration, or both. The highly valued lakes adjacent to developed areas would benefit from all BMPs in their contributing areas, especially in the face of projected increases in population and development pressure. Even if these increases do not occur by the year 2030, we presume they will occur eventually.