La Houille Blanche
Number 1, Février 2000
|Page(s)||51 - 56|
|Published online||01 July 2009|
La prévision des précipitations : progrès depuis dix ans, perspectives à dix ans
Precipitation forecasting: 10 years of progress, perspectives for the 10 next years
Météo-France, CNRM, GMAP, MOD
Meteo-France has two main tools for forecasting precipitations: 2PIR, as a nowcasting tool, and a panoply of numerical models (IFS, ARPEGE, Aladin) for short to medium range forecast. As in every National Meteorologkal Services the quality of the numerically forecasted precipitations is routinely evaluated at Meteo-France. The particular variability of this field imposes to evaluate the accumulated precipitations against observed values themselves. The quality of the forecasted precipitations has been improved significantly since ten years, and is comparable roughly with the quality obtained at ECMWF (Reading, UK), one of the leading Centres in the world. These improvements can mainly be explained by the refinement of the horizontal resolution allowed by the available computers capacity, and in a lesser extend by progresses in the forecast technique itself (dynamical and physical representation, data assimilation). The perspectives to ten years are far from being vanishing. The future better results should come from progresses in the same three main ingredients (further realism in the physical description, further refinement of the horinzontal resolution, and above all, implementation of 4D-Var data assimilation with a possibility of real-time modification of the initial state).
© Société Hydrotechnique de France, 2000
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.