Submitted to: Soil Erosion for 21st Century Symposium
Publication Type: Proceedings
Publication Acceptance Date: 1/5/2000
Publication Date: N/A
Interpretive Summary: Excessive sediment movement from farm fields fills water reservoirs and lakes, increases water treatment costs, and affects the habitat of many aquatic organisms. Mathematical models have been developed to predict sediment transport from farm fields varying in topography, soil, cropping and tillage practices, and precipitation characteristics. The impact of changing cropping and/or tillage practices on sediment transport can be easily evaluated once the model has been properly tested and evaluated. In this study, we compared measured soil losses from a small field located in western IA to soil losses predicted by the Water Erosion Prediction Project (WEPP) watershed model for a 17-year period. Two approaches were taken. The first was to compare measured annual soil losses averaged over the 17-year period to those predicted by the WEPP model. For this comparison, the model over predicted surface runoff by 10% and soil loss by 45%. Most of the over prediction occurred in years when measured surface runoff and soil losses were low. The second approach was to compare soil losses for the 10 largest events that occurred over the 17-year period. These results indicate that future model revisions need to focus on improving model predictions for years with low measured surface runoff and soil losses. The results will be useful to erosion scientists and action agency personnel who are continually working with and improving the WEPP watershed model. Official release of the WEPP watershed model to the USDA-Natural resources Conservation Service (NRCS) is scheduled for January 2003.
Technical Abstract: Improved mathematical models are needed to better predict runoff and sediment transport from field-sized areas. This research was conducted on a 5.6-ha corn-cropped watershed located in the deep-loess hills region of western Iowa. Measured runoff and sediment losses for a 17-year period (1975-1991) were compared to those predicted by the water Erosion Prediction Project (WEPP) watershed model version 99.5. A digital Elevation Model (DEM) was used to divide the watershed into 8 hillslope and 3 channel elements. The WEPP watershed model requires 6 input files, including climate, cropping and management, soil, topography, channel, and structure. Measured daily precipitation and minimum and maximum temperatures were used, while other climatic variables were generated from WEPP's climate generator. Cropping and management inputs were based on actual tillage, planting, and harvesting dates and implements. Mean annual lprecipitation for the 17-year period was 818 mm. Measured and predicted 17-yr mean annual surface runoff and soil losses were 51.9 and 57.3 mm and 16.5 and 23.9 Mg/ha, respectively. The model over predicted mean annual surface runoff and soil losses by 10.4 and 44.9%, respectively. Although measured and predicted mean annual values compared reasonably well, the differences on an annual basis were often high for soil loss. The model greatly overestimated annual sediment loss for years when measured loss was low. Ten large events accounted for about 60% of the total soil loss measured over the 17-year period. Predicted soil losses for 9 of these 10 events agreed closely with measured amounts. Results indicate that predicted soil losses for small runoff events need to be improved to increase the prediction accuracy of the WEPP watershed model.