SWAT Literature Database for Peer-Reviewed Journal Articles

Title:Modelling of point and non-point source pollution of nitrate with SWAT in the River Dill, Germany 
Authors:Pohlert, T., J.A. Huisman, L. Breuer and H.-G. Freude 
Journal:Advances in Geosciences 
Volume (Issue):
Article ID: 
URL (non-DOI journals): 
Broad Application Category:hydrologic and pollutant 
Primary Application Category:pollutant cycling/loss and transport 
Secondary Application Category:calibration, sensitivity, and/or uncertainty analysis 
Watershed Description:692 km^2 Dill River, which is part of the Lahn-Dill-Bergland region located in the middle of the State of Hessen in Germany. 
Calibration Summary:Daily (Nov. 1990 to Dec. 1993) E = 0.81 --------------------------- monthly (April 2000 - March 2002) nitrate r2/E = .69/.66 
Validation Summary:Daily (April 2000 to Jan. 2003) r2/E = .60/.56 --------------------------- April 2001 - March 2002 nitrate values: monthly r2/E = .77/.77 daily E = 0.15 
General Comments:Automatic flow calbiration was performed with the Shuffled Complex Evolution Algorithm developed at the Univ. of Arizona (SCE-UA); optimization procedure and selected parameters identical to Huisman et al. (2004). Manual calibration was peformed for nitrates. Flow inaccuracies were attributed to several factors including subdaily processes that could not be simulated accurately and timing of collection of 24-hour rainfall periods. Monthly nitrate loads were generally accurately predicted; overpredictions were attributed to a simple in-stream mixing submodel that ignored degradation processes. Denitrification was greatly overpredicted and nitrate uptake kinetics were also not simulated well. The authors state these problems must be corrected before the model is used for scenario analyses. Ongoing research is being conducted to improve nutrient cycling in the model. 
Abstract:We used the Soil and Water Assessment Tool (SWAT) to simulate point and non-point source pollution of nitrate in a mesoscale mountainous catchment. The results show that the model efficiency for daily discharge is 0.81 for the calibration period (November 1990 to December 1993) and 0.56 for the validation period (April 2000 to January 2003). The model efficiency for monthly nitrate load is 0.66 and 0.77 for the calibration period (April 2000 to March 2002) and validation period (April 2002 to January 2003), respectively. However, the model efficiency for daily loads is low (0.15), which cannot only be attributed to the quality of input data of point source effluents. An analysis of the internal fluxes and cycles of nitrogen pointed out considerable weaknesses in the models conceptualisation of the nitrogen modules which will be improved in future research.