Multiple transfer plus noise models for real time flow forecast: application to the Upper Tiber River basin, Italy

This work deals with the problem of real time flood forecasting by stochastic models of transfer function noise type. Nonlinear effects are represented by using cumulated rainfall series together with ‘intervention’ or ‘perlog’ series. This model, which was used for flow forecast in a large number of events on different basins, is applied here to the Upper Tiber River basin (Central Italy). The hydrometeorological data acquisition is performed in this basin by a telemetering system using radio waves.