Soil erosion resulting from agricultural land use is associated with environmental impacts and crop productivity loss. Erosion prediction is currently based on models, like USLE or WEPP. In Brazil, traditionally, the scale for soil conservation planning is the watershed (1000 to 5000 ha) and the Brazilian “small watershed programs” are international references for successful soil conservation. The quantification of erosion as a georeferenced variable is essential for its analysis together with other important data treated in Geographic Information Systems (GIS). This paper describes a computer program named Erosion Database Interface (EDI) that allows georeferenced applications of USLE and WEPP at the watershed scale. EDI operates as a pure interface, allowing the connection of EDI to any GIS that work with vector formats. The georeferenced vector-based design, the action as a pure interface and the flexibility to fit to different scales are aspects that distinguish EDI from other spatial erosion prediction tools. For use with USLE, each soil type has its own K-value and each land use has its own CP-value. The R factor is constant and LS-values are calculated from topographic data. For use with WEPP, EDI builds a soil, management and slope input file for each transect after separating the transect in the Overland Flow Elements, which are used by WEPP for erosion rate calculation. Based on input data from either USLE or WEPP, EDI will generate output files with longitude and latitude in meters and soil loss. To illustrate EDI’s performance in a scale compatible with land use planning for the Brazilian small watershed programs a practical example of a sugarcane area (2,000 ha), located at the south-eastern part of Brazil (Piracicaba county) was used. The resulting maps and some statistics are shown and it is concluded that EDI is an efficient program for georeferenced erosion prediction for both WEPP and USLE. The presentation of erosion rates as a georeferenced information gives a new dimension to the erosion process as compared to its presentation as mean rates.
Keywords GIS, WATERSHED, EROSION MODELS.