LAM Multi-model Precipitation Forecasts driving discharge predictions

In the present study, an approach to ensemble prediction is proposed, based on multiple precipitation scenarios generated by different high-resolution NWP models driving the same hydrological model. Thus, the uncertainty associated with the meteorological forecasts provided by this multi-model approach (model errors, i.e. that fraction of the global uncertainty related to the use of different models) can propagate into the rainfall-runoff model, providing a more informative hydrological prediction. Also, this approach represents a complementary tool to validate and interpret QPF for the verification of NWP model performances over the medium-sized basin scale. The proposed methodology has been applied to an episode of intense precipitation that affected the Reno river basin, a medium-sized catchment in Emilia-Romagna Region in northern Italy. The ground effects of the precipitation event are evaluated in terms of streamflow evolution over the basin.