SIMULATION OF A LOW-GRADIENT COASTAL PLAIN WATERSHED USING THE SWAT LANDSCAPE MODEL

Accurate simulation of landscape processes in natural resource models requires spatial distribution of basin hydrology and transport processes. To better represent these processes, a landscape version of the SWAT model has been developed to simulate the runoff, run-on, and infiltration processes that typically occur in different parts of the landscape. The model addresses flow and transport across hydrologic response units prior to concentration in streams, and is capable of simulating flow and transport from higher landscape positions to lower positions. The SWAT landscape model was tested using data collected from a heavily vegetated riparian buffer system in the Atlantic Coastal Plain near Tifton, Georgia. Simulations of surface runoff lateral subsurface runoff and groundwater flow for an upland field, a grass buffer, and a sub-divided forested buffer floodplain were generated. Model results and field data indicate that surface runoff was dominant in the upland field, while groundwater flow was dominant in the grass buffer and the floodplain. While average annual surface runoff agreed satisfactorily with observations from the site, annual and monthly simulated values varied considerably from observed values. Simulated surface runoff tracked general trends in the observed data, but winter months and extreme events were overestimated while summer months were underestimated. Annual surface runoff predictions at the edge of the upland field varied from the observed data by 11% to 44%. The Nash-Sutcliffe efficiency for annual estimates of surface runoff at the field edge was 0.83 for the three-year calibration period. The results demonstrate the ability of the model to simulate the surface runoff and enhanced infiltration typically associated with riparian buffer systems. Additional revision of the model will likely be necessary to adequately represent redistribution of water between surface, lateral subsurface flow, and groundwater flow.