Statistical correction for model prediction on winter circulation in the extra-tropical of Northern Hemisphere

Abstract

The performance of ensemble-forecast system on winter 500hPa height field in the Northern Hemispheric Extratropics (NHE) is studied by using the Meteo France model data of DEMETER project, with analysis on performance of the model modes based on empirical orthogonal function (EOF) of observations. Both optimum subset regression (OSR) and analogue method are used to advance the model prediction on ‘bad modes’. Results suggest that the prediction ability of the mode accounting for less variance may be higher than the mode with more variance. The OSR failed, while the analogue method based on OSR shows a possibility of improving the prediction techniques by correcting the bad modes of model. However, since the model has a poor capability in representing the second and third EOF modes of the observation which account for a large percentage of the total variance, the forecast ability can not be improved effectively where the model prediction information is not enough or incorrect. So it is necessary to make a further analysis on the samples of the ‘bad modes’ and the corresponding external forcing which might better realize the correction for such ‘bad modes’.