A fuzzy classification of the hydrodynamic forcings of the Rhone River plume: an application in case of accidental release of radionuclides
Résumé
Assessing and modelling the coastal plume dispersion of nuclearized rivers is strategic in case of accidental releases for the protection of vulnerable areas, but taking into account all the possible hydrodynamical conditions is challenging. River plumes are mostly affected by wind and river discharge, but the variability of these two forcings suggest that data mining methods may be particularly effective to define their major trends and influences on the plume behavior.This study uses fuzzy c-mean clustering on Rhone River (France) discharge and wind speed at its estuary for two objectives: explaining the variability of the riverplume by defining scenarios of hydrodynamic forcings, and relating these scenarios with the spatial extension of the plume . The application to the Rhone River, the most nuclearized European river, highlighted the ability of this method to classify a 10 years serie of wind and discharge into 6 scenarios with remarkable characteristics. These scenarios correspond to different surface currents and plume behaviors, and they were used to simulate the extension and dilution of a radioactive release. These simulations can be used as a quick decision tool, and a decisional tree is also proposed to identify in real time which climatological scenario occurs at the river mouth and the potential plume pattern.
Origine : Fichiers produits par l'(les) auteur(s)
