Studying effects of spatial scale on model calibration using Soil Water Assessment Tool (SWAT)

摘要

Calibration of model parameters is a critical step in any successful watershedmodeling project. The calibrated values of model parameters, however, areinfluenced by sub-watershed divisions and resolution of input data. The objective ofthis research is to study the effects of sub-watershed divisions and scale of soil dataon calibration of SWAT model for the St. Joseph River Watershed (SJRW). Two setsof SJRW models one with SSURGO data (1:24,000 scale) and other with STATSGOdata (1:250,000 scale), each set having six different level of sub-watershed divisionvarying from 10 sub-watersheds to 97 sub-watersheds, are independently calibratedusing SCA-UA algorithm for daily streamflow output at watershed outlet. Calibratedresults show no significant difference in terms of model performance however goodparameter sets showed varied range of sensitivity to model spatial scale. This resultindicates towards redefining uncertainty range of sensitive model parameters.