SWAT Literature Database for Peer-Reviewed Journal Articles

Title:Sensitivity and uncertainty anaylsis coupled with automatic calibration for a distributed watershed model 
Authors:Muleta, M.K. and J.W. Nicklow 
Year:2005 
Journal:Journal of Hydrology 
Volume:306(1-4) 
Pages:127-145 
Article ID: 
DOI:10.1016/j.jhydrol.2004.09.005 
URL (non-DOI journals): 
Model:SWAT 
Broad Application Category:hydrologic & pollutant 
Primary Application Category:calibration, sensitivity, and/or uncertainty analysis 
Secondary Application Category:pollutant cycling/loss and transport 
Watershed Description:133 km^2 Big Creek (southern Illinois) 
Calibration Summary:Daily (June, 25 1999 to Aug. 26, 2001; PRS station): streamflow E = .74 sediment E = .46 ---------------------------- daily graphical comparisons between simulated and measured values also shown 
Validation Summary:Daily (April 20, 2000 - August 26, 2001; CRS station): streamflow E = .23 sediment E = -.005 ---------------------------- daily graphical comparisons between simulated and measured values also shown 
General Comments:An automatic calibration approach using a hierarchy of 3 techniques (screening, parameterization of physically immeasurable parameters, & parameter sensitivity analysis) was performed at the parameter identification stage. The sensitivity analysis was conducted using a stepwise regression approach for 35 different input parameters. Application of the 3 techniques reduced the calibrable parameters to 20: 12 for streamflow & 8 for sediment. Latin hypercube sampling was used to generate input data frm assigned disributions and ranges, and parameter estimation was performed using a genetic algorithm. A GLUE method was used to evaluate uncertaintity. The authors state that the poor verification results may be due in part to the relatively short period used for calibration. 
Language:English 
Keywords:Sensitivity analysis; Automatic calibration; Uncertainty analysis; Model verification; Genetic algorithms; Distributed watershed model