Quantification of nutrient loads in urban stormwater across Colorado municipal areas

Abstract

As nutrient pollution becomes a growing problem across the world, it becomes increasingly important to be able to quantify the amount of nutrients each of the different sources contributes. Understanding the quantity of nutrients each source is providing makes it possible to understand where and how to address nutrient pollution effectively. Increases in urbanization have resulted in the degradation of water quality specifically from urban stormwater. Due to regulations that have been adopted to address this problem, stormwater Best Management Practices (BMPs) have been implemented to provide water quality treatment and runoff reduction. Since these BMPs are altering the quality and quantity of urban stormwater they need to be included in modeling the urban watershed. As a first approach, a simple framework for which an order of magnitude estimate of nutrient load from stormwater could be determined was developed. The effects of BMPs were incorporated into the framework as well as a method for estimating the current amount of drainage area being treated by BMPs. The provided approach contains a simple model which can rapidly estimate nutrient loads from urban stormwater in any given watershed or area of interest using the most recently available national datasets. A simple approach was chosen because it is easy to understand regardless of a user's background in modeling. All of the inputs for the method are available through national datasets, allowing the approach to be applied anywhere within the United States. The proposed method of estimating nutrient loads was evaluated by comparing predicted event loads to observed event loads from two different study sites within Fort Collins, CO. From this analysis it was determined that the proposed approach would only be applicable for estimating the order of magnitude of nutrient loads. A Monte Carlo simulation discovered large bands of uncertainty surrounding the prediction. However, a Bayesian statistical analysis found that the uncertainty bands around the estimate could be substantially reduced by using observed data within an area of interest. Overall the proposed approach should be used for planning level purposes where order of magnitude estimates are appropriate. In order to consider BMPs into the method it was necessary to know how much drainage area was currently being treated by BMPs. An approach was developed that compared changes in land use as measured by National Land Cover Database (NLCD) products to provide a measurement of the estimated treatment drainage area provided by stormwater BMPs. This approach was evaluated in Fort Collins, CO and was found to be able to effectively predict a range of BMP implementation which captured the actual amount of BMP implementation for Fort Collins as a whole as well as for 11 of the 12 sub-basins within Fort Collins. Further research should include evaluating the proposed approach for load estimation in other urban watersheds of varying sizes utilizing different structural BMPs and green infrastructure. This would require a watershed to have readily available data regarding the nutrient load being exported by urban sources within the watershed. Further research should also include conducting the BMP treatment area estimation approach in other urban watersheds to determine whether the method is applicable in other regions and municipalities other than Fort Collins.