Probabilistic risk assessment of urban stormwater discharge

Urban stormwater is an important source of detrimental impacts for receiving waters. Discharges from separated or combined sewer systems are recognized as a major contamination input to the environment. Moreover, these discharges may impact the morphological aspect of receiving waters, like rivers and streams. But these impacts are difficult to handle. First, precipitation is the main driving force of the problem and has stochastic features, which must be taken into account. Second, some variables and parameters of the urban and receiving water systems are difficult to measure or to estimate (e.g., riverbed characteristics, number of inhabitants in a watershed, imperviousness, etc.) and therefore cannot be determined exactly. Finally, factors involved in the cause-effect relationship between urban stormwater discharges and the receiving water ecosystem, i.e., the assessment criteria, are difficult to describe with a linear relationship. To tackle this problem, we developed a new impact-based approach in Switzerland to allow engineers to plan adequately treatment devices and technical solutions. To support this approach, a probabilistic software tool has been proposed to assess impacts of urban drainage to receiving waters based on new impact-based criteria. The software includes probabilistic modeling facilities (Monte Carlo simulation routine) and sensitivity analysis to focus the data acquisition on the most important parameters. Long-term simulations (10-year duration) are performed using historical rain data from different Swiss locations for the simulation of an urban watershed, upstream watershed, and the receiving water’s hydrology and water quality. This long period allows us to account for variability in precipitation. The computation typically runs 200-500 iterations. Most variables and parameters’ uncertainties are taken into consideration by generating random values of a given multivariate distribution. The results, after sorting, are plotted as a cumulative distribution function that represents the probability (on the Y-axis) complying with a criterion (on the X-axis). The guidelines and the software were proposed to authorities and to experts active in the field of urban wet-weather and ecosystem management. This new approach was accepted by the majority of the experts, even if the complexity of the planning procedure generated skepticism from some users. The concepts of uncertainty and variability were well understood by users. In fact, they are commonly used in urban wet-weather and water ecosystem management, even if not explicitly presented. For example, the definition of watershed imperviousness, the choice of rain series, the initial rain losses, the impacts of chemicals on aquatic species, etc., are parameters that are commonly estimated in a given range of reliabilities. Examples of applications and presentations of scenarios were the best way to illustrate the benefits of using the new approach. Nevertheless, the interpretation of results is quite difficult and a communication strategy, such as organization of workshops and courses is a need for the acceptance of the methodology.