Uncertainty assessment of sewer sediment erosion modelling

Abstract

Urban stormwater quality modelling has become a fundamental issue in the evaluation of the receiving water bodies' quality state. Laboratory study as well as field campaigns have widely demonstrated that combined sewer sediments present cohesive-like properties which increase their resistance to erosion. However, only few models take sewer sediments into account considering their rheological properties. In the present paper different sewer sediment erosion models have been tested and their uncertainties have been assessed. The main goal was to discriminate the algorithms with respect to their robustness and the reduction of uncertainty. In order to accomplish such objective the GLUE methodology has been used. The model has been tested using the quantity-quality data gathered for the Fossolo catchment (Bologna, Italy). Results show a general tendency of sewer sediment erosion models to become overparameterised when trying to fit the complexity of the physical processes thus increasing modelling uncertainty.