Applying a Model to Predict the Location of Land Drained by Subsurface Drainage Systems in Central Minnesota

Agricultural drain tile systems are a significant influence on the condition of wetlands and waterways. The influence of these systems is often difficult to determine since installation records are incomplete or were never kept. Using a modified decision class tree and raster analysis in ArcGIS, a model for predicting the location of land drained by subsurface systems was evaluated. The three-county study site in the agricultural region of central Minnesota provided an area of known drain tile systems so that the model predictions could be compared to locations of existing systems and drained land. The model criteria incorporated publically available data including agricultural land use data identified by the National Land Cover Dataset (NLCD) and the National Agricultural Statistics Service (NASS), soil characteristics obtained through the Soil Survey Geographic Database (SSURGO), and slope characteristics developed from the National Elevation Dataset (NED). Results indicate that with the best combination of criteria the model predictions correspond nearly 80% with the actual drain tile data. The potential to incorporate the influence of drain tile areas into land-use based assessments of wetland and waterway health is an important outcome of being able to identify land drained by artificial subsurface drainage features.