SWAT Modeling of Sediment Yields for Selected Watersheds in the Laurentian Great Lakes Basin

Abstract

In the Laurentian Great Lakes Basin (GLB), corn acreage has been expanding since 2005 in response to high demand for corn as an ethanol feedstock. This study integrated remote sensing-derived products and the Soil and Water Assessment Tool (SWAT) within a GIS modeling environment to assess the impacts of cropland change on the sediment yield within four selected watersheds. The SWAT models were calibrated over a 6 years period (2000–2005) and predicted stream flows were validated. The R2 values were 0.76, 0.80, 0.72, and 0.81 for the St. Joseph River, St. Maryʼs, the Peshtigo River, and the Cattaraugus Creek Watersheds, respectively. The corresponding E (Nash and Sutcliffe model efficiency coefficient) values ranged from 0.24 to 0.79. The average annual sediment yields (tons/ha/yr) ranged from 0.12 to 4.44 for the baseline (2000−2008) condition. Sediment yields were predicted to increase for possible future cropland change scenarios. The first scenario was to convert all “other” agricultural row crop types (i.e., sorghum) to corn fields and switch the current/baseline crop rotation into continuous corn. The average annual sediment yields increased 7%−42% for different watersheds. The second scenario was to further expand the corn planting to hay/pasture fields. The average annual sediment yields increased 33%–127% compared to the baseline conditions