EVALUATING THE SWAT MODEL FOR A LOW-GRADIENT FORESTED WATERSHED IN COASTAL SOUTH CAROLINA

摘要

Modeling the hydrology of low-gradient forested watersheds on shallow, poorly drained soils of the coastal plain is a challenging task due to complexities in watershed delineation, microtopography, evapotranspiration, runoff generation processes and pathways including flooding and submergence caused by tropical storms, and complexity of vegetation species. The main objective of this study was to calibrate and validate the GIS-based spatially distributed hydrologic model SWAT for the 72.6 km(2) low-gradient, third-order Turkey Creek watershed within the Francis Marion National Forest in the South Carolina Coastal Plain. Model calibration used GIS spatial data of the watershed and 2.75 years (2005-2007) of streamflow and climate data, and the model was validated with 2.5 years (2008-2010) of data. Based on limited field measurements, results showed that the SWAT model with an improved one-parameter "depletion coefficient" for plant evapotranspiration in the SCS curve number (CN) estimate can predict the daily and monthly streamflow processes of this watershed reasonably well and better than the CN method. The model performance was "good" (E = 0.68; RSR = 0.56) to "very good" (E = 0.90; RSR = 0.31) for the monthly calibration and validation periods but only "satisfactory" (E = 0.59; RSR = 0.64) to "good" (E = 0.70; RSR = 0.55) for the daily calibration and validation periods. Better predictions were found for the validation period that included two wetter years than the calibration with two drier years. The model's predictions of the zero or near-zero flow days of summer were also in agreement with the measurements for 60% of the time. However, it was concluded that the refined SWAT model was still unable to accurately capture the flow dynamics of this forest ecosystem with shallow, high water table soils for events preceded by wet saturated conditions during the dry summer and wet winter periods, warranting further investigations on these forest systems. The five-year average annual runoff coefficient of 19% with a baseflow amount of 27%, on average, of the runoff (streamflow) and ET of 987 mm predicted by the model were found reasonable compared to the estimated values and other published data for the region. Further improvements in estimates of forest potential evapotranspiration, rainfall spatial variability, and antecedent moisture as a function of water table should reduce uncertainties in flow predictions, allowing the model to be used in hydrologic impact assessments of land use change, land management practices, and climate change in coastal landscapes.