Simple approach to characterizing a pollution event return period in urbanhydrology, by means of hydrological parameters

The constant concern for reduction of impact of urban discharges during rai ny episodes has led to the investigation of pollution induced by urban runo ff. Many studies have shown that pollution is a function of several climato logical factors. The main factor is related to the rainy episode and to the dry episode preceding the rain. To evaluate better the natural link existi ng between pollution and the couple of variables studied (duration of dry e pisode, intensity of subsequent rain), a statistical study has been conduct ed in the urban catchment of Entzheim (France). A first step consisted of t he separate study of the behaviour of each random variable: rain and dry pe riod, using Gauss, Galton, Gumbel, Fuller and Weibull laws. Fitting was mad e using data sampled over a period of 124 months. The second step of this r esearch consisted of processing pollution data based on 38 events. The stat istical analysis made it possible to define the most efficient descriptive parameters of the couple of variables studied for the prediction of polluti on and to establish linear regressions. The fitting of a global model on po llution data compared to the regression models should apply to several catc hments. It is now possible to determine the occurrence of each event knowin g only a few parameters. The third step consisted of the formulation of a f requency law for the couple of variables studied and comparing it to the la w obtained in the second step. The objective is to formulate a simple relat ionship between the frequencies so as to rapidly predict the order of magni tude of the pollution event return period.