Location: Southeast Watershed ResearchTitle: Effect of Watershed Subdivision and Filter Width on SWAT Simulation of a Coastal Plain Watershed) Author
Submitted to: Journal of the American Water Resources Association
Publication Type: Peer reviewed journal
Publication Acceptance Date: 11/16/2009
Publication Date: 6/14/2010
Citation: Cho, J., Lowrance, R.R., Bosch, D.D., Strickland, T.C., Vellidis, G., Her, Y.G. 2010. Effect of Watershed Subdivision and Filter Width on SWAT Simulation of a Coastal Plain Watershed. Journal of the American Water Resources Association (JAWRA). 46(3):586-602. DOI: 10.1111/j.1752-1688.2010.00436x. Interpretive Summary: Computer based hydrologic and water quality models have proven to be useful tools for examining alternative management scenarios and their impact on the environment. This examination can be an important component of watershed-scale evaluations. The Soil and Water Assessment Tool (SWAT), is watershed-scale model commonly used for these type evaluations. In many watersheds, riparian buffers play a dominant role in hydrologic and water quality functions. However, SWAT provides a simplified representation of these buffers. This analysis provides guidance to more accurately determine watershed configurations and to represent riparian buffers within SWAT. The watershed and buffer configuration were found to have a limited impact upon watershed hydrology, but a dramatic impact upon sediment and nutrient dynamics. Recommendations are made on optimum watershed and buffer configurations as a function of stream network. .
Technical Abstract: The Soil and Water Assessment Tool (SWAT) does not simulate riparian buffers, but has a simple filter function that may be useful to simulate buffer function. SWAT was calibrated on a subwatershed (15.7 km2) within the Little River Experimental Watershed, Georgia. The calibrated parameter set was applied to 32 different watershed configurations, eight different critical source areas (CSAs) with four different filter strip width (FILTERW) representations for each. Streamflow predictions were stable regardless of changes in watershed subdivision and FILTERW. Predicted sediment and nutrient loads from HRUs decreased as CSA increased when variable FILTERW was considered. Sediment and nutrient yield at the watershed outlet showed dynamic responses to different combinations of CSA and FILTERW depending on selected in-stream processes. Sediment and nutrient yields at the watershed outlet were less sensitive to the changes in CSA and FILTERW when in-stream processes were considered. A subdivision level which provides similar drainage density to the actual stream network was suggested as the lower CSA boundary. A subdivision level that does not change the stream length at the watershed outlet was suggested as the upper CSA boundary.