Validating Stormwater system simulations in Edmonton Using MIKE URBAN

Many municipalities use chloramination to disinfect drinking water so as to avert the production of the disinfection by-products (DBPs) that result from conventional chlorination processes and the consequential public health risks. However, the long-lasting monochloramine disinfectant (NH2Cl) can pose a significant risk to the environment. As, it can be introduced into stormwater sewers and thus freshwater sources. This study was intended to investigate decay of NH2Cl in stormwater networks starting by building a stormwater model and validating its hydraulic and hydrologic computations, and then modelling water quality in the storm sewers. The presented work here is only the first stage of this study. The 30th Avenue basin in Edmonton was chosen as a case study, because it has various land-use types including commercial, industrial, residential and parks. The City of Edmonton has already built a MIKE-URBAN stormwater model for modelling floods. However, this model was built to the trunk level where only the main drainage features were presented. Also, this model was not calibrated and known to consistently compute pipe flows higher than the observed values; not to the benefit of studying water quality. So the first goal was to complete modelling and updating the real stormwater network. Then, available GIS Data was used to calculate different catchment properties such as slope, length and imperviousness. To calibrate and validate this model, data of two temporary pipe flow monitoring stations was used along with records of two other permanent stations available for eight consecutive summer seasons. The effect of various hydrological parameters on model results was investigated. It was found that model results were affected by the ratio of impervious areas. The catchment length was tested, however calculated, because it is approximate representation of the catchment shape. Surface roughness coefficients were calibrated using. Consequently, computed flows at the two temporary locations had correlation coefficients of values 0.846 and 0.815, where the lower value pertained to the larger attached catchment area. Other statistical measures, such as peak error of 0.65%, volume error of 5.6%, maximum positive and negative differences of 2.17 and -1.63 respectively, were all found in acceptable ranges.