Receiving water quality assessment: comparison between simplified and detailed integrated urban modelling approaches

摘要

Urban water quality management often requires use of numerical models allowing the evaluationnof the cause–effect relationship between the input(s) (i.e. rainfall, pollutant concentrations onncatchment surface and in sewer system) and the resulting water quality response.nThe conventional approach to the system (i.e. sewer system, wastewater treatment plant andnreceiving water body), considering each component separately, does not enable optimisation ofnthe whole system. However, recent gains in understanding and modelling make it possible tonrepresent the system as a whole and optimise its overall performance. Indeed, integrated urbanndrainage modelling is of growing interest for tools to cope with Water Framework Directivenrequirements. Two different approaches can be employed for modelling the whole urban drainagensystem: detailed and simplified. Each has its advantages and disadvantages. Specifically, detailednapproaches can offer a higher level of reliability in the model results, but can be very timenconsuming from the computational point of view. Simplified approaches are faster but may leadnto greater model uncertainty due to an over-simplification. To gain insight into the abovenproblem, two different modelling approaches have been compared with respect to theirnuncertainty. The first urban drainage integrated model approach uses the Saint-Venant equationsnand the 1D advection-dispersion equations, for the quantity and for the quality aspects,nrespectively. The second model approach consists of the simplified reservoir model. The analysisnused a parsimonious bespoke model developed in previous studies. For the uncertainty analysis,nthe Generalised Likelihood Uncertainty Estimation (GLUE) procedure was used. Model reliabilitynwas evaluated on the basis of capacity of globally limiting the uncertainty. Both models have angood capability to fit the experimental data, suggesting that all adopted approaches arenequivalent both for quantity and quality. The detailed model approach is more robust andnpresents less uncertainty in terms of uncertainty bands. On the other hand, the simplified rivernwater quality model approach shows higher uncertainty and may be unsuitable for receivingnwater body quality assessment.