Several methods have been developed for use in estimating the water quality loads associated with urban and agricultural landuses and practices. These include the use of sophisticated computer models, typically based on using pollutant loading and runoff functions, regression equations, load export coefficients (LECs), and event mean concentrations (EMCs). This research has examined the feasibility of using a simple EMC approach with the Kentucky Nutrient Model (KYNM). The thesis includes an extensive literature review of EMCs and a synthesis of recommended values for a range of typical urban and agricultural landuses. The thesis also includes an extensive literature review of potential BMPs along with a summary of the typical removal efficiencies and costs associated with each type of BMP. The research also explored the potential to use the results from multiple applications of site specific BMP models like the Source Loading and Management Model (WinSLAMM) in the development of general functional relationships that could then be used to evaluate BMP performance on a more site-specific basis. The developed EMC table and the associated BMP performance curves should provide valuable tools for use in better managing nutrient loads for urban and agricultural watersheds.

This study provides an evaluation of the performance of the WinSLAMM model in two cities of northern Utah, a region in which it has not been validated to date. Runoff volumes for a series of events from 2015 to 2017 were calculated for six study areas. These results were compared to model outputs to evaluate the prediction accuracy. The model performed poorly when default parameters where used, as expected. Thus, the calibration of runoff coefficient was done with local data. Total suspended solids (TSS), total phosphorus (TP) and total dissolved phosphorus (TDP) were also evaluated and calibrations were developed. The WinSLAMM model performed well for TSS prediction and TP and TDP estimations were calculated with the TP/TSS and TDP/TP ratios observed in the data for the Logan study areas. Three scenarios of GI implementation were analyzed; the treatment of 10%, 50% and 100% of connected impervious surfaces. For the 10% case, treating streets produced the greatest TSS reduction, while treating roofs yielded the greatest volume reduction. For the 50% case, there was a great difference in TSS reduction with varying surfaces. Implementing GI for streets produced the largest TSS reduction, followed by parking lots and then by roofs.