Statistico-dynamical downscaling for Mediterranean heavy precipitation

Mediterranean regions are regularly affected by high-precipitating events (HPEs) that often lead to devastating flash floods. The evolution of the occurrence and severity of such extreme events in the frame of climate change remains an open question. To address this issue, we have designed a statistico-dynamical downscaling method to apply to climate-model outputs. This statistico-dynamical downscaling is based on a two-step method: first, representative situations are selected among an ensemble of propitious synoptic environments to HPEs for the present (1961-2000) and future (2070-2099) climates and second, for these situations simulations are carried out with a fine-scale non-hydrostatic model. Large-scale circulation patterns (LSCs) propitious to high-impact weather events have been identified by a statistical method. This has been applied to the coupled atmosphere-ocean regional climate model ARPEGE Climate/OPAMED simulation considering an enhanced greenhouse climate following the IPCC-SRES-A2 scenario. Then dynamical downscaling is performed by simulating the 21 selected cases with the French MESO-NH numerical model at 2.5 km horizontal resolution. Initial and boundary conditions are provided by the climate-model scenarios. The high-resolution simulations for the 21 cases give accumulated rainfall exceeding 100 mm, often 300 mm and even 500 mm. Results highlight that the statistico-downscaling method is able to select LSCs leading to HPEs and to reproduce precipitating systems typical of Mediterranean events. This study also shows that it is possible to infer some conclusions about the evolution of the geographical location, occurrence and intensity of these intense events. The number of days with LSCs propitious to heavy precipitation significantly increases in the future climate. In particular, the simulated rainfall for the selected 21 cases shows an increase in the spatial variability of such events in the future climate associated with a slight increase of rainfall maxima. Copyright (C) 2011 Royal Meteorological Society