Modeling hydrologic and water quality response of a mixed land use watershed in Piedmont physiographic region (Maryland)

Nonpoint source pollution of surface and groundwater has become recognized as an important environmental problem associated with agricultural production for any region. Mathematical watershed scale models are among the best tools available for analyzing water resources (quantity and quality) issues in spatially diverse watersheds. Existing models without proper validation have resulted in some misconceptions about such models.; A conceptual, continuous time, and watershed scale, distributed hydrologic/water quality model, SWAT, was applied to a 340 ha watershed with mixed land use in the Piedmont physiographic region of Maryland to test its performance in predicting hydrologic/water quality response. Six years (1994–1999) of hydrologic and nutrient loading data were used to calibrate and validate the capability of SWAT. Additionally, an artificial neural network approach based on a back propagation scheme (ASCE, 2000a and 2000b) was used to generate the incomplete flow data in 1998.; Results indicated that the SWAT model like many other similar models only handle subsurface flow bounded by the surface topography, thus neglecting the possible subsurface flow contribution from outside the watershed. A water budget analysis therefore was performed to quantify various components of the hydrologic cycle within the watershed. The imbalance in water budget analysis showed the ground water contribution from outside the watershed to be in the range of 20%–61% of the measured base flow. Therefore, accounting for such contribution is deemed to be necessary for fair evaluation of the model results. Adjustments of measured base flow and stream flow made to exclude the extra groundwater recharge showed a moderate improvement in model's predictions.; Finally, model's simulations of monthly sediment and nutrient loadings were poor. Despite the poor monthly predictions and some extreme events, the model showed a strong agreement between annual measured and simulated data for sediment, nitrate and soluble phosphorus loadings. It was concluded that SWAT is a reasonable watershed scale model on long-term simulations of the impact of different management practices on hydrologic and water quality response of mixed land use watersheds. However, its use on short-term or storm by storm basis may not be appropriate.; The uncertainty analysis of input parameters and its effect on model simulation was also investigated in this study. The large standard deviation (102.4 mm) and large range of stream flow (215.0 to 530.0 mm) imply significant uncertainty in stream flow prediction. This portion of the study concluded that assessing the uncertainty in model outputs resulting from uncertainty in input data may be very vital in evaluating the environmental risks associated from such uncertainties.