Performance of ILSAX and SWMM urban drainage models

The performance of two urban event runoff models commonly used by the urban water authorities in Australia, the ILSAX (ILLUDAS-SA and something eXtra) and SWMM (Storm Water Management Model) models, were assessed and results are discussed in this paper. Data from the Giralang catchment in Canberra, Australia were used, and errors produced by the models in simulating runoff volume, peak discharge and time to peak discharge of event hydrographs were examined by comparing simulated and recorded hydrographs. The calibration of the ILSAX model parameters was done by trial and error. The SWMM model parameter calibration was done by using the pattern search technique with different initial values selected by trial and error to determine the global optimum parameters. Analysis of the ten storms from Giralang catchment in Canberra, indicated that both the ILSAX and SWMM models gave higher percentage error in simulating runoff volume (14-95%), peak discharge (13-108%) and time to peak discharge (0-29%) for small rainfall events which were less than 1 year average recurrence interval (ARI). Both models were equally good in simulating runoff volume (2-58%), peak discharge (1-54%) and time to peak discharge (0-20%) for medium size rainfall storm events between 2 and 10 years ARI. The results of this study showed that both parameter values and model estimation results were storm dependent. Simulated plots for the best and worst estimations of individual storm events by the ILSAX and SWMM models indicated that both models were capable of simulating the shape of observed hydrograph in all the selected events.