A comparison of MIKE SHE and DRAINMOD for modeling forested wetland hydrology in coastal South Carolina, USA
Models are widely used to assess hydrologic impacts of land-management, land-use change and climate change. Two hydrologic models with different spatial scales, MIKE SHE (spatially distributed, watershed-scale) and DRAINMOD (lumped, fieldscale), were compared in terms of their performance in predicting stream flow and water table depth in a first-order forested watershed in coastal South Carolina. The model performance was evaluated using the coefficient of determination (R2) and Nash- Sutcliffe’s model efficiency (E). Although both models performed reasonably well in predicting monthly and annual average water table depths and stream flow with acceptable E values (0.55-0.99) for the five-year period (2003-2007), MIKE SHE yielded better results than DRAINMOD for daily hydrologic dynamics. Both models, however, showed relatively large uncertainty in simulating stream flow for dry years. The subsurface drainage predicted by MIKE SHE was lower than simulated by DRAINMOD for dry years, higher for extremely wet years and similar for normal climate years. The differences were likely that MIKE SHE employed distributed physical characteristics of the watershed, especially of soil and topography which can substantially affect the subsurface flow, but the spatial average condition was only used by DRAINMOD; the results from both models were, thus, similar for those average (e.g., normal climate) conditions, and different for varying conditions. This study suggests a lumped parameter model could perform equally well at the monthly temporal scale for modeling stream flow under average climatic conditions; however a distributed hydrological model provides more accurate prediction of daily stream flow and water table depth across varying climatic conditions.