Modeling the Monthly Water Balance of a First Order Coastal Forested Watershed
A study has been conducted to evaluate a spreadsheet-based conceptual Thornthwaite monthly water balance model and the process-based DRAINMOD model for their reliability in predicting monthly water budgets of a poorly drained, first order forested watershed at the Santee Experimental Forest located along the Lower Coastal Plain of South Carolina. Measured precipitation, weather, stream flow and soil hydraulic data from a 160 ha low-gradient, naturally drained watershed on a mixed pin hardwood forest (WS80) were used in the testing of the models. The Penman-Monteith and Thornthwaite based potential evapotranspiration (PET) methods were used in both models. While rooting depth and field capacity of the soil were the only parameters needed for the Thornthwaite model, complete data on soil water retention and saturated hydraulic conductivity by layers, root depths, surface storage, and drainage design parameters were needed for the DRAINMOD to simulate monthly and annual hydrology. As expected, results using two and half years (January 2003 – June 2005) of data showed that DRAINMOD was a better predictor (Nash-Sutcliffe E = 0.92 and absolute average monthly deviation, Eaamd = 11.0 mm) of monthly outflows compared to the Thornthwaite model (E = 0.83 and Eaamd = 16.8 mm). Although DRAMMOD can also be used to analyze the daily water table dynamics in contrast to the Thomthwaite model, it was not considered in this study. A DRAINMOD simulation was later conducted with a 46-year (1956-2002) weather dataset to evaluate the long-term hydrology of the watershed, specifically to describe the variability in predicted outflows as affected by the year-to-year climatic variations. Despite the limitations of these models, they can be useful tools for land managers and planners to estimate water tables, quantify monthly and seasonal water budgets as well as estimate nutrient and sediment loadings from poorly drained coastal forests.