Evaluating Impacts of Detailed Land Use and Management Inputs on the Accuracy and Resolution of SWAT Predictions in an Experimental Watershed

Land use and management practice inputs to the Soil Water Assessment Tool (SWAT) are critical for evaluating impact of land change best practices on soil erosion water quality in watersheds. We developed an algorithm this study maximize usage records during setup SWAT a small experimental watershed New Brunswick, Canada. In algorithm, hydrologic response units (HRUs) were delineated based field boundaries associated with long-term records. The model was further calibrated validated respect flow sediment nutrient (nitrate soluble phosphorus) loadings at outlet. As comparison, baseline version also set up using conventional way HRU delineation limited information practices. These two versions compared input output resolution prediction accuracy monthly annual loadings. Results show that new method had much higher accuracies generating areas crops, fertilizer application, tillage operation, diversion terraces (FDT), grassed waterways watershed. Compared method, improved predicting loading due better representation FDT watershed, it successfully estimated spatial across However, there no definite increase simulation high temporal inputs, though simulations sensitive configuration. examination showed comparable between models. results indicate SWAT, although information, is able provide quantity outlet, as long data can reasonably represent average condition over target period.