A meteo-hydrological forecasting chain: performance of the downscaling and rainfall-runoff steps in a small catchment

Forecasting ground effects of severe meteorological events with an adequate lead time is fundamental for civil protection scopes and is therefore an important challenge for the scientific community. The paper focuses on the performance of some steps of a meteo-hydrological forecasting chain that can be applied in small watersheds to assess hydrological risk deriving by an intense storm predicted at the large meteorological scale. The proposed procedure integrates large-scale rainfall fields, as those produced by numerical weather prediction (NWP) models, with statistical rainfall downscaling and hydrological modelling. More in details, assuming a large scale rain rate as the input of the process, the forecasting chain produces an ensemble of hydrographs that are post-processed in order to give a probabilistic representation of mean streamflow maxima for different time windows. The outcome of this procedure can be thus applied to assess the risk that some critical streamflow thresholds may be exceeded. The procedure has been tested on more than one thousand recorded events in the Araxisi catchment in Sardinia, Italy. Results and performances are presented and discussed.