Analyse hydrologique de la crue-éclair catastrophique du 15 juin 2010 dans la région de Draguignan (VAR, France)

The 15th of June 2010 an exceptional flash flood occurred in the Var region in southeastern France. This flood caused 25?casualties and catastrophic damages in the area of Draguignan. After this event several researchers and engineers teams conducted field surveys in order to gather information on the dynamics of the flood: peak discharge estimates for the main Argens River and all the tributaries hit by the event, and information on the chronology of the flood based on interviews of eyewitnesses. The SCHAPI and the DREAL PACA organised exchanges between the different teams involved in order to ensure a global coherence of discharge estimates, and to examine the question of the possible return period of the flood. This work of comparison revealed that despite the several sources of uncertainty, the discharge estimates proposed by the different teams were relatively close to each other. This work finally led to a complete cartography of peak discharges including almost 40?local estimates. The spatial repartition of peak discharges appears coherent with rainfall accumulations estimated based on radar data. A more detailed hydrological analysis of the flood was then conducted based on the comparison of estimated peak discharges and hydrographs simulated using a simple distributed rainfall-runoff model. This analysis revealed a great disparity of the hydrological answer of the tributaries of the Argens River. The strongest reactions were observed on the Florieye and the Real Rivers for which rainfall accumulations exceeded 350?mm in 24?h. The Aille River also had a strong reaction despite of much more limited rainfall accumulations on this watershed (up to 250?mm). On the other hand, the Nartuby River showed a very attenuated answer despite of very elevated rainfall accumulations (exceeding 300?mm): this could be attributed to the presence of large karst formations in this last watershed. Lastly, the discussion on the return periods of the flood enabled to identify the river reaches where a 100-year return period had probably been exceeded.