¿Cómo distintas clasificaciones de tipos de circulación discriminan las fases de la NAO?

Abstract

In this study, a ranking between a variety of atmospheric circulation classifications from the European Action COST733 is established attending to the best discrimination of positive and negative North Atlantic Oscillation (NAO) phases. As an evaluation measure, the x2 –Pearson statistics has been used. Secondly, the impact of using different number of circulation types in the discrimination of NAO phases has been analysed. The main interest of the study is in the joint use of two general circulation approaches: modes of variability and circulation classification methods, which usually are studied in a separate way. Their relationship will allow us to establish a ranking among classifications. As a general rule, the classifications discriminate the negative NAO phase better than the positive NAO phase, and the larger the number of circulation types, the better the discrimination is. The classifications that best discriminate positive NAO phase belong to the same family of classification algorithms (optimization) and turn out to be independent of the number of circulation types considered. By contrast, the classifications that best discriminate negative NAO phase encompass two families of algorithms (optimization and principal component analysis) showing a more clear impact with the number of circulation types.