Integrating a Water Balance into the Soil & Water Assessment Tool (SWAT) to Better Model Hydrologically Sensitive Areas

EPA Grant Number: FP916984
Title: Integrating a Water Balance into the Soil & Water Assessment Tool (SWAT) to Better Model Hydrologically Sensitive Areas
Investigators: White, Eric Davidson
Institution: Cornell University
EPA Project Officer: Michaud, Jayne
Project Period: September 1, 2008 through August 31, 2009
RFA: STAR Graduate Fellowships (2008) RFA Text |  Recipients Lists
Research Category: Academic Fellowships


Non-point source pollution is an issue of concern for many water bodies in the United States; nutrient rich runoff from agricultural fields being a prime example of this pollution type. To aid in curbing the effects felt downstream by use of fertilizers, pesticides, and other chemicals upstream, watershed managers frequently utilize computer watershed models. One such model is the widely used Soil & Water Assessment Tool (SWAT), developed by the USDA. It has been shown that SWAT can be quite accurate in modeling the timing and quantity of pollutants coming from specific land uses defined in a modeled watershed. While useful at a larger, watershed scale, SWAT has no current mechanism to accurately determine where runoff and pollutants are being generated within each single land use type. By making modifications to calculations within the model, this project will improve SWAT’s ability to predict this spatial distribution of pollutant-laden runoff at the intra-watershed scale.


To effectively create an improved version of SWAT, three main steps will be followed. The first step includes making changes to the calculations that SWAT performs. This will be done by removing an existing process from SWAT’s source code and replacing it with a new process for determining runoff generation. The process to be replaced requires some assumptions that have been shown to be not as accurate as first thought. By no longer needing these assumptions, SWAT will be capable of determining runoff-generating areas as a function of physical processes occurring on the watershed, returning much more accurate results. The second step that must be taken is to actually run this new version of SWAT to determine if it performs as expected. To do this, a model will be created for one of the New York City drinking water source areas, located in the Catskills region of New York. The runoff-generating areas predicted by the model will be compared to actual locations of such areas. These comparisons will then be used to determine just how accurate the new version of SWAT is. The third, and final, step in this project is what is called model calibration. This involves adjusting input values to the model until the highest level of model accuracy has been obtained. After this point there will be two distinct products from this project. The first product will be a very accurate model of a New York City source watershed, which will be used by researchers and land use planners throughout New York State. The second, and most important, product will be a new version of SWAT with great accuracy in determining exact locations in a watershed where water-borne pollutants are being generated.

Supplemental Keywords:

hydrology; watershed modeling; land use planning; SWAT; non-point source pollution,

Progress and Final Reports:

  • Final