Develompent of an automated GIS-based modeling approach to support regional watershed assessments.

摘要

A comprehensive, GIS-based modeling approach was developed to enable accurate prediction of nutrient loads in watersheds throughout the state of Pennsylvania; particularly those watersheds for which historical stream monitoring data do not exist. This approach relies on the use of statewide GIS data sets for deriving reasonably good estimates for various critical model parameters that exhibit considerable spatial variability within the state. Data manipulation and subsequent simulation modeling is managed via an interface (called AVGWLF) between a popular desktop GIS software package (ArcView) and the Generalized Watershed Loading Function (GWLF) model.; An evaluation was first made of the accuracy of the AVGWLF approach for estimating mean annual nutrient loads. In this case, simulated and observed mean annual nutrient load values were pooled separately for both calibration and verification watersheds. Nash-Sutcliffe (N-S) coefficients were then calculated for each group for both N and P by comparing the simulated and observed mean annual nutrient loads for each watershed against the average (i.e., mean) value of the observed mean annual nutrient loads in each group. During the calibration step, an N-S value of 0.97 was calculated for both total nitrogen and total phosphorus loads. In the verification step, N-S values of 0.92 and 0.95 were calculated for total nitrogen and total phosphorus loads, respectively.; The utility of using AVGWLF for estimating nutrient loads for shorter time frames was also evaluated. As was done with the comparison of mean annual loads between watersheds, Nash-Sutcliffe coefficients for monthly, seasonal and yearly nitrogen and phosphorus loads were also calculated individually for each of the calibration and verification watersheds. On average, the median N-S values for both nutrients was about 0.67. The majority of the calculated N-S coefficients for both nutrients were consistently above 0, which means that AVGWLF almost always provides a better estimate than just the mean monthly, seasonal or annual load in any given watershed. Since historical water quality measurements are routinely not available for most watershed studies in Pennsylvania, the potential benefit of using AVGWLF in such situations cannot be under-estimated.