EVALUATING UNCERTAINTY IN E. COLI RETENTION IN VEGETATED FILTER STRIPS IN LOCATIONS SELECTED WITH SWAT SIMULATIONS Guber, A. K. a , Pachepsky Y. A. b , and Sadeghi, A. M. c a Department of Earth and Environmental Sciences University of California Riverside; b USDA-ARS Environmental Microbial safety Laboratory, Beltsville, MD; c USDA-ARS Hydrology and Remote Sensing Laboratory, Beltsville, MD. Abstract. Vegetated fulter strips (VFS), as a best management practice, have become an important component of the water quality improvement in watersheds. The SWAT model allows for a comprehensive description of agricultural practices, and has proven to be efficient in applications to watershed with a substantial agricultural component. The REMM, KINEROS, and other models have recently been suggested to use SWAT output to evaluate the VFS retention capacity with respect to nutrient and sediment loads. The concern about the manure-borne pathogens in waterways and water bodies has to be addressed in more details by estimating the efficiency of VFS in pathogen retention. The existing knowledge base shows that the functioning of VSF as barriers for manure-borne pathogens to the large extent depends on vegetation status, soil infiltration capacity in VFS, and rainfall intensity and duration. The effect of these factors on the pathogen breakthrough in VFS needs to be researched at the time scales smaller than the computation interval of one day that SWAT uses. The downscaling of the SWAT output to the smaller time scales, however, introduces an uncertainty. We have developed the model STIR to simulate the overland transport and loss to infiltration of manure-borne pathogens in VFS. This model was used in Monte Carlo simulations to evaluate the possible variability in pathogen breakthrough in VFS with inputs from SWAT. The simulations show that probabilistic characterization of the VFS efficiency with site-specific soil and weather properties can be useful in making decisions on VFS placement with respect to manure-borne pathogens. Keywords. SWAT, water quality, pathogens, indicator organisms, vegetated buffer strip, STIR model, uncertainty INTRODUCTION Vegetation filter strips (VFS) are commonly used to decrease the pollutant loads from manured fields and pastures. Using SWAT simulations has been suggested to make decisions about the VFS placement. One needs to select parameters of a particular VFS and assess its reliability after a decision about the VFS location is made. The REMM, the KINEROS, the VFSMOD and other models have recently been suggested to use SWAT output to evaluate the VSF retention capacity with respect to nutrient and sediment loads (Allison et al., 2006; EPA, 2005, Goodrich et al., 2006). By the frequency of being the cause of water quality impairment, pathogens rank first and second among five leading pollutants in estuaries and rivers, respectively, in the United States (EPA, 2004). A wide range of opinions exists on the VFS efficiency with respect to pathogens and/or indicator organisms (Pachepsky et al., 2006). Tools to evaluate the efficiency of VFS and select its parameters with respect to manure-borne pathogens have been developed in 80th. They include the Agricultural Runoff Management II: Animal Waste Version (ARM II) model (Overcash et al., 1983), the Utah State (UTAH) model (Springer et al., 1983), the MWASTE model (Moore et al., 1988), and the COLI model (Walker et al., 1990). Recent interest to the fate and transport of manure-borne pathogens has generated a substantial increase in data on fate and transport of pathogens and indicator organisms in VFS. Several excellent reviews have been published (Jamieson et al., 2002, Ferguson et al., 2003, Tyrrel and Quinton, 2003, Unc and Goss, 2004, Oliver et al., 2005). The existing knowledge base shows that the efficiency of VFSs as barriers for pollutants depends to large extent on soil infiltration capacity in a VFS and soil moisture content before the rain event, vegetation status and microtopography, and rainfall intensity and duration (Munoz-Carpena et al., 1999; Helmers et al., 2006). The effect of these factors on the pathogen breakthrough in VFS needs to be researched at time scales smaller than the computation interval of one day that SWAT uses. The downscaling of daily values of weather variables and SWAT outputs to the smaller time scales inevitably introduces an uncertainty. Another source of uncertainty for the VFS evaluation is the spatial scale of information about VFS soil and vegetation properties, and its management. The objective of this work was to research the opportunities for the VFS evaluation in the uncertainty framework using the model STIR (Solute Transport with Runoff and Infiltration) that we have under development (Kouznetsov et al., 2006). 286 This is not a peer-reviewed article. Watershed Management to Meet Water Quality Standards and TMDLS (Total Maximum Daily Load) Proceedings of the Fourth Conference 10-14 March 2007 (San Antonio, Texas USA) Publication Date 10 March 2007 ASABE Publication Number 701P0207. Eds A. McFarland and A. Saleh