About the Reliability of Manual Model PV Corrections to Improve Forecasts

Abstract

he National Weather Forecast Centre of Météo-France has developed a tool that corrects the state of the atmosphere within the Action de Recherche Petite Echelle Grande Echelle (ARPEGE) operational global model by adjusting the potential vorticity when the initial conditions and available observations disagree. Among observational datasets, geostationary satellite data are the primary source of information. Here, the representation of the coherent structures of the tropopause in the model is assessed with Meteosat ozone and water vapor images. Modifications to the initial conditions of the potential vorticity (PV) in areas where the model fails can be applied using a potential vorticity inversion, thus providing a new balanced distribution of the wind and temperature that will then be used as a new initial state. The purpose of the paper is to investigate the degrees of weakness of the present qualitative approach. To this end, PV modifications to the initial conditions are applied by different experts on the eve of the windstorm Klaus (24 January 2009) that hit southwestern France. The different initial PV fields and the subsequent forecasts show significant differences in terms of wind and mean sea level pressure, while sharing some common features. The human modification process is therefore partially reproducible and skillful since the forecast is improved most of the time.