Title: | A global sensitivity analysis tool for the parameters of multi-variable catchment models |
Authors: | van Griensven, A., T. Meixner, S. Grunwald, T. Bishop, M. Di Luzio and R. Srinivasan |
Year: | 2006 |
Journal: | Journal of Hydrology |
Volume (Issue): | 324(1-4) |
Pages: | 10-23 |
Article ID: | |
DOI: | 10.1016/j.jhydrol.2005.09.008 |
URL (non-DOI journals): | |
Model: | SWAT |
Broad Application Category: | hydrologic and pollutant |
Primary Application Category: | sensitivity analysis |
Secondary Application Category: | pollutant cycling/loss and transport |
Watershed Description: | 3240 km^2 Sandusky River and 932 km^2 Upper North Bosque, which drain respectively portions of northern Ohio (to Lake Erie) and north central Texas, U.S. |
Calibration Summary: | |
Validation Summary: | |
General Comments: | Different local (single parameter) and global (multiple parameter) sensitivity analysis methods are described using SWAT2003 (earlier version of SWAT2005). A combined novel latin-hypercube & one factor-at-a-time sampling (OATS) approach is described and used, that facilitates a global sensitivity analysis for a large set of parameters with just a limited set of runs. Examples of applying the approach is described using 46 input parameters for two watersheds for flow, SS, TN, TP, nitrate, & ammonia at several locations. Hydrologic parameters were dominant & the results were found to not be transferable between the two watersheds. SWAT output is reported in terms of uncertainty measures; e.g., confidence bounds for daily flows and sediment loads. |
Language: | English |
Keywords: | Model parameters; River; Sensitivity analysis; Water quality |