Uncertainty in measuring the event pollutant removal performance of online detention tanks with permanent outflow

Abstract

Detention tanks are frequently used in today's urban stormwater management. It is a recognized fact that these structures are efficient in order to reduce the frequency of flooding. By settling suspended solids and specific particle pollutants, detention tanks also may contribute to reduce pollutant loads discharged into the receiving waters. However, some recent studies report the measured event removal performance of tanks functioning with permanent outflow (online) to be quite low and relatively variable from one site to another. Even if the studies do not refer to it, the measured performance is probably impaired by considerable uncertainty. The aim of this paper is to ascertain to which point uncertainty could explain the weakness and the variability of the measured performance. Initially, the theoretical formulation of relative uncertainty in the pollutant removal performance of any tank is given. Then, this relationship is applied to the case study of an online tank functioning with permanent outflow: the Vénissieux tank. Some original methods are presented in order to assess uncertainty components. Final results obtained show that absolute uncertainty in the measured performance is far from negligible (between 13% and 37% whilst removal rates range between 25% and 73%). In conclusion, a coherent plan of action is proposed that can be used on other sites to reduce uncertainty to a minimum. A critical look is also taken at the general problem of uncertainty assessment.