Sensibilité des modèles de calcul des flux polluants en réseaux d'assainissement par temps de pluie aux données expérimentales de calage

Sensitivity of quality models to experimental data used for calibration Three types of stormwater quality models were tested to evaluate the sensitivity of calibration results to experimental data. The study was carried out on « Le Marais » catcment for suspended solids where 40 reliable measured events and good knowledge of the sewer system are available. The method used is based on resampling calibration subsets of measured events among the 40 available ones. For each calibration, the model was validated against the remaining events to stand upon the quality of the model. It was found that the models are quite sensitive to calibration data, a problem neglected in practical studies. The variability of calibration results decreased when more events were used. It decreased dramatically between 1 and 10 events used for calibration. It also appeared that the number of events used in practice by sewer system managers and model users for calibration is insufficient and that a special effort has to be done to make longer data series of better quality available. In addition to variability, systematic errors occurred in the estimation of the total load of the available events. Hence, relative application of models is preferable.