Skip to Main content Skip to Navigation

Prévisibilité des ressources en eau à l'échelle saisonnière en France

Singla Stéphanie 1
1 GMME - Groupe de Météorologie à Moyenne Échelle
CNRM - Centre national de recherches météorologiques
Abstract : Although seasonal meteorological forecasts are operational for several years, theirapplication to hydrology remain less developed. However, seasonal hydrological forecasts can be a useful tool to forecast a few months in advance hydrological characteristics like soil moisture or river flow. The aim of this thesis is thus to evaluate the potential of the Hydro-SF hydrometeorological forecasting suite to forecast river flow and soil moisture at seasonal time scale over France, for the water resources management, especially for the anticipation of droughts and low flows. To address this goal, the different sources of predictability of the hydrologic system, and the added value of seasonal forecasts compared with climatologic forecasts, are evaluated for the spring season (March-April-May) from 1960 to 2005. The results, already published, show that an important part of the predictability comes from: the snow cover over high mountains, the large and complex aquifer simulated for the Seine catchment, and the atmospheric forcing over most plains. Moreover, differents atmospheric forcings from the climate model ARPEGE are confronted, and the added value of seasonal forecasts compared with climatologic forecasts is observed over the North-East of France. Secondly, as the importance of atmospheric forcing over most plains is seen with the precedent results, a specific work is done on downscaling. The meteorological surface parameters from seasonal forecasts are disaggregated using a complex statistic method : weather types classification and analogs with DSCLIM. This disaggregation is compared with the method implemented before with a simple spatial interpolation and calculation of standardized anomalies. This work is always realized for the spring season, and bring out that the contribution of DSCLIM compared with the simple method used before, is ambivalent for atmospheric forcing and hydrological variables. Several others investigations about downscaling of seasonal forecast are thus needed, and suggestions are done for future work. Finally, in the context of water resources management, seasonal hydrological forecasts are done for summer (June-July-August), when water uses conflicts are at a maximum because of low flows and droughts. This topic is addressed using four forecasting experiments with differents initializations (from February to May), in order to assess the potential of the Hydro- SF suite over the period of summer low flow over France, but also in order to determine the optimal initialization forecast for the best anticipation of possible droughts. Results are interesting as they show significant scores with an initialization of seasonal forecasts from April, especially for : catchment located downstream from mountains where the hydrological predictability depends on snow cover, and the Seine catchment where the importance of the contribution of the groundwater for river flow increase in summer compared with spring. In the same way that for spring, the added value of seasonal forecasting for the hydrological system is evaluated and located, for summer, in the South of France.
Complete list of metadata
Contributor : Eric Martin <>
Submitted on : Friday, March 8, 2013 - 11:46:03 AM
Last modification on : Monday, April 5, 2021 - 2:26:18 PM
Long-term archiving on: : Sunday, June 9, 2013 - 6:15:23 AM


  • HAL Id : tel-00798280, version 1



Singla Stéphanie. Prévisibilité des ressources en eau à l'échelle saisonnière en France. Interfaces continentales, environnement. Institut National Polytechnique de Toulouse - INPT, 2012. Français. ⟨tel-00798280⟩



Record views


Files downloads