Méthodes de prévision d’ensemble pour l’étude de la prévisibilité à l’échelle convective des épisodes de pluies intenses en Méditerranée

Abstract : This PhD thesis aims at quantifying the uncertainty of convection-permitting numerical weather forecasts, with a particular interest in the predictability of Mediterranean heavy precipitating events. Four uncertainty sources, which impact the predictability of these events, were investigated : the description of the synoptic-scale circulation, the representation of meso-scale atmospheric conditions (especially the low-level jet feeding the convective systems with moist and unstable air), the impact of complex physical processes such as the setting up of a cold pool, and the definition of surface conditions. To quantify the impact of these four uncertainty sources, the ensemble forecasting technique was chosen, using the AROME model. Each uncertainty source is studied separately through the definition of dedicated perturbations, and the resulting ensembles are first evaluated over heavy precipitation case studies. We then proceed to a statistical evaluation of the ensembles for 2- and 4-week long forecast periods. This evaluation, completed with the design of ensembles sampling several uncertainty sources together, allows us to draw some practical tips for the design of an operational convective scale ensemble forecasting system at Météo-France
Document type :
Theses
Complete list of metadatas

https://pastel.archives-ouvertes.fr/pastel-00805613
Contributor : Abes Star <>
Submitted on : Thursday, March 28, 2013 - 1:57:31 PM
Last modification on : Tuesday, March 6, 2018 - 3:55:57 PM
Long-term archiving on : Sunday, April 2, 2017 - 9:51:05 PM

File

TH2012PEST1166_complete.pdf
Version validated by the jury (STAR)

Identifiers

  • HAL Id : pastel-00805613, version 1

Collections

Citation

Benoît Vié. Méthodes de prévision d’ensemble pour l’étude de la prévisibilité à l’échelle convective des épisodes de pluies intenses en Méditerranée. Sciences de la Terre. Université Paris-Est, 2012. Français. ⟨NNT : 2012PEST1166⟩. ⟨pastel-00805613⟩

Share

Metrics

Record views

823

Files downloads

766