Assessing the Impact of Loading Rate Uncertainties in TMDL Allocations

Abstract

To account for uncertainty in the form of variability and incomplete knowledge in Total Maximum Daily Load (TMDL) studies, a Margin of Safety (MOS) is introduced, typically chosen arbitrarily to be 5%. This study characterizes significant uncertainties present in loading rates by using multiple scenarios that each act as real potential loads to the watershed. Loading rates are generated from specified parameter spaces using the Latin Hypercube Sampling technique, and sample distributions are chosen based on various published loading rates and the likelihood that particular loads would more closely represent loads applied to the watershed. The multiple loading rate scenarios, coupled with the multiple water quality and hydrologic parameter sets are then simulated with the watershed model, HSPF. Acceptance criteria are then applied to assess the likelihood of a particular parameter/load set to act as a real simulator of the watershed. A robust genetic algorithm model linked with a response matrix, is then used to optimize TMDL allocations and predict the reliability of compliance associated with each allocation scenario. This methodology is applied to fecal coliform load allocations in the Moore’s Creek Watershed in Virginia. The results are compared to a similar study that assumed loading rates to be deterministic, allowing for the assessment of the impact of uncertainty in loading rates.