Probabilistic Approaches for Assessment of Non-Point Source Pollutant Loads from Urban Watersheds

Estimation of non-point source pollution has been a major challenge for many urbanized and urbanizing watersheds. The estimation of loads from non-point sources particularly from urban watersheds is very complex due to many factors that include the random variations of rainfall, runoff, pollutant buildup and washoff and the overall complexity of the watershed. Thus, it is appropriate to analyze the urban stormwater pollution within a probabilistic setting. As an alternative approach to continuous simulation, analytical probabilistic models have been proposed for analyzing urban runoff quality control systems. Based on concepts from statistics and probability theory, these models have been derived from the probabilistic characteristics of system input using the functional relationship between inputs and outputs. Exponential probability density functions are used to express the probabilistic behavior of long-term rainfall characteristics. Buildup functions and washoff functions, analogous to those used in continuous simulation models, are used. The models are verified with available field data from an urban watershed located in the Greater Toronto Area and reasonable agreements are obtained between model results and observed data.