Site-specific decision-making based on RTK GPS survey and six alternative elevation data sources: Soil erosion predictions

Precision farming equipment based on Global Positioning Systems (GPS) enables landowners to gather spatially distributed topographic data in real-time kinematic (RTK) mode, which has the potential to be used in addition to or as substitute for commonly available topographic data sources (e.g., U.S. Geological Survey (USGS) topographic contour lines and/or digital elevation models). The latter are considered insufficiently accurate in their topographical representation of watershed boundaries, slopes, and upslope contributing areas to be able to meaningfully apply detailed process-based soil erosion assessment tools at the field scale. In this second of two articles discussing the usefulness of the available data sets from a decision-maker's perspective, the same comprehensive accuracy tests that were used for these topographical parameters are applied to the spatially distributed soil erosion assessment results simulated by the Water Erosion Prediction Project (WEPP) model supported by Geographic Information Systems (GIS). The impact of the accuracy of six alternative topographical data sources on predicting soil erosion rates using WEPP is compared to on-site soil erosion predictions using elevation measurements from a survey-grade RTK GPS with centimeter accuracy. Results show that the more precise topographic measurements with a photogrammetric survey or any differential GPS units yield more precise on-site soil loss predictions for individual raster cells (0.01 ha) and hillslope areas of interest (0.5 ha). However, the best WEPP predictions for average annual off-site runoff (-18.3% error) and sediment yield (-2.7% error) from upslope contributing areas of about 4 ha within the 30 ha watershed were achieved using the USGS 10 ft contours. These results demonstrate that in this case, the runoff and sediment yield predictions using DEMs based on the commonly available contour lines can be even better than those from the more precise and costly topographic data sets. The contours also allowed successful application of the WEPP model to identify all 11 hillslope areas of interest (0.5 ha) with soil loss (10) or deposition (1) problems that were initially mapped in the field as larger rills and sedimentation areas, respectively.