The results show land area predicted to be drained by artificial subsurface systems corresponded well with areas that are known to have these systems. The best combinations of land use, soil characteristics, and slope criteria resulted in the identification of nearly 80% of the actual drain tile area.
The ability of the model to identify land drained by subsurface tile could be evaluated better if additional data about the locations of existing tile systems was obtained. This study is limited by the drain tile data since it includes only the subsurface drainage features under the jurisdiction of the counties within the study area. It is unknown whether the other areas predicted to be drained in this model actually are drained. But because the model predictions correspond well with the areas that are truly drained, the likelihood that these are drained is good.