Modifying and Validating the SWAT Model to Determine Landuse Effects on Watershed Water Quality: Using a Dual Level of Model Performance Based on Subbasin Size

The North Bosque River (NBR) was included in the CleanWater Act § 303(d) impaired list. The Texas Institute for Applied Environmental Research used the SoilWater Assessment Tool (SWAT) to develop a phosphorus (P) Total Maximum Daily Load. SWAT was modified to dynamically change manure application rates based on simulated annual soil P, change areas receiving manure, alter manure quantities each year, apply liquid and solid manure pools separately, move manure between subbasins, improve landscape P processes, model contributions of dairy lagoon discharges and improve in-stream water quality kinetics. Data was refined to increase spatial resolution of subbasins and include Public Law (PL)-566 flood retardation reservoirs. A dual level of model performance was established: one level for large drainage areas and a reduced performance level for all other sites. Main stem sites were to have streamflow, sediment, total nutrients achieve a “good” rating. For secondary sites and constituent parts of total nutrients besides PO4 a “satisfactory” rating was acceptable model performance. This dual level ofmodel performancewas developed in recognition of uncertainties in model input and measured data that resulted in better model calibration performance for larger (primary) drainage areas as compared to smaller drainage areas, and for total nutrients as opposed to their constituent parts. The refined SWAT model was satisfactorily calibrated for both historical long-term (30-year) base, surface and total streamflow data (Nash-Sutcliff Efficiency (NSE) > 0.71 and %Error < 12.7) and monthly streamflow and total nutrient loads (primary sitesNSE > 0.66 and%E < 14).Average daily load and concentrationweremore problematic but satisfactorily predicted at three out of five sites and four out of five sites respectively.