[1] The groundwater-induced flooding that occurred in the Somme Basin during April 2001 damaged numerous dwellings and communication routes, and economic activity of the region was flood-bound for more than 2 months. It was the first time that such a sudden event was recognized as resulting from groundwater discharge, despite the Somme valley not being prone to flooding. Because of a dual porosity of the chalk in the basin, nonlinear processes, involving a hydraulic continuity between the macropores of the unsaturated zone and the chalk groundwater, govern water migration through the unsaturated zone. Such a process is the result of switching behavior of groundwater recharge from matrix flow to macropore flow due to accumulated wetness over several years. There is much evidence to support that the flood probability model is climatedependent for the studied region because nonlinear processes amplify the effects of nonstationarities of climatic inputs. An estimation of the return period of catastrophic flooding depends on the long-term precipitation fluctuations. This has implications for flood risk assessment requiring the need to distinguish between short-and long-term flooding risks. Other basins that may not appear particularly prone to flooding could also be subjected to similar groundwater-induced flooding should the long-term precipitation fluctuations observed in the north of France since the beginning of the 1980s persist. Similar extraordinary situations can occur in Belgium and England, whereby significant flooding results in substantial contribution of groundwater flows.

[1] Within the framework of the Global Energy and Water Cycle Experiment (GEWEX)/ Rhone project, a system has been built to estimate the hydrological budget of the Rhone, one of the major European rivers (with a 86,500 km 2 surface area for the French part of the catchment). The methodology is based on three models, one for each component of the hydrometeorological system: The atmospheric parameter analysis, the snow cover, the surface water and energy budget, and the underground water transfer and discharge estimation. This tool has been validated for 14 years (from 1981 to 1994) by comparing the daily river flows simulated by the models with measurements from 145 gauging stations. In this study, the results of the ARPEGE-Climat general circulation model (GCM) have been used to estimate the climate of the Rhone catchment in 60 years. The perturbation of the air temperature and precipitation amount has been used to modify the actual set of meteorological parameters in order to simulate the hydrological budget of the Rhone river. Vegetation and soil structure are supposed to be identical to current values, which is a strong hypothesis. The river discharge and soil water resources under the climatic scenario are compared with the results of the actual simulation. Strong contrasts in the hydrological response of the catchment are noticeable, depending on the location of the subcatchments and the amount of precipitation. Snow cover is the most sensitive hydrological component to the air temperature increase, and the high mountainous river regimes are strongly modified. When considering the soil water content, it appears that the northern part of the domain stays quite wet, whereas drying is enhanced in the south. The uncertainties in the results are estimated by analyzing the model sensitivity to different simple climatic scenarios. In particular, the analysis brings into light the impact of the downscaling of the GCM results.

No abstract

No abstract