Probabilistic analysis of runoff simulations in a small urban catchment

摘要

In Urban Hydrology, a basic question is whether or not the common methods involving the use of design storms bring to the the some results obtained by those methods that make use of real storms. In general one on say that different design storms give good results when used with the appropriate model, or, conversely, that good results can be achieved through careful model calibration. On the basis of 51 rainfall-runoff recordings obtained from the experimental catchment of Luzzi (Cosenza, Italy), the frequency distribution of the observed peak discharges was initially computed. Then the runoff events were simulated using Wallrus, a well known simulation model, taking as input the observed precipitations. The frequency distribution of the simulated peak discharges was compared to that of the observed ones, with the aim of calibrating the model on a statistical basis. After that the rainfall events were analysed, obtaining the frequency distributions of the observed intensities over several durations and developing IDF curves of given frequencies and then the Chicago design storms. The plotting positions of the peak disharges simulated by this way show a good agreement with the distribution of both the observed peak discharges and the peak discharges simulated through the real storms.