Predicting the impact of climate change on urban drainage systems in northwestern Italy by a copula-based approach

Study region: Milan, northwestern Italy. Study focus: The impact of expected trends in storm temporal structures is analyzed with reference to urban drainage systems, featuring catchment areas spanning from 10 ha to 100 ha. A bivariate stochastic model for the derivation of flood frequency is developed, accounting for the seasonality of storm volumes, durations and their mutual dependence structure. Its reliability is verified by comparing it to continuous hydrodynamic simulations. To do so, a 21-year long series observed at Milan-Monviso raingauge was used. Model comparison evidences a satisfactory agreement between models. New hydrological insights for the region: Although the total annual precipitation is not expected to change, relevant increases in flood frequencies are predicted. Such increases vary between 10-20% and appear to be independent of the return period. Thus, great concerns arise for the existing urban drainage systems located in northwestern Italy, which should basically be unable to face these flood frequency changes. A leading role is played by the intensification of summer and spring storms, both in terms of increase in volumes and decrease in durations. Moreover, changes in the dependence structure have a significant impact when summer storms are considered. Conversely, flood frequency curves are far less sensitive to the storm temporal structures featuring other seasons. These results can be explained by considering the seasonal distribution of storms critical for urban drainage systems.