Incorporation into the system of the largest number of meteorological data available for the detection of situations that could lead to flooding episodes, as well as the use and incorporation of a series of station data that complement and contribute to improving the available data.

Use of probabilistic predictions based on neural networks for the early detection of potentially dangerous situations, calibration of models, adaptation of last generation models and prediction by “ensembles”, as well as development and integration of a reservoir management model that allows the taking of opportune decisions on the discharges that are necessary to cushion floods.

Communications protocol by means of which both administrations and the public can be informed of possible alert situations, addressed to competent authorities (regional and local), civil protection and citizens.