Fuzzy rule‐based classification of atmospheric circulation patterns

Abstract A fuzzy rule‐based methodology is applied to the problem of classifying daily atmospheric circulation patterns (CP). The purpose of the approach is to produce a semi‐automated classification that combines the expert knowledge of the meteorologist and the speed and objectivity of the computer. Rules are defined corresponding to the geographical location of pressure anomalies. A CP is described by the location of four different possible pressure anomalies. The rules are formulated with fuzzy sets, allowing a certain flexibility because slightly different pressure maps may correspond to a given CP. Accordingly the degree of fulfilment of a rule is defined in order to measure the extent to which a pressure map may indeed belong to a CP type. As an output of the analysis, the CP on any given day is assigned to one, and only one, CP type to a varying degree of credibility. The methodology is applied to a European case study. The subjective classification of European CPs given by Hess and Brezowsky provides a basis for constructing the rules. The classification obtained can be used, for example, to simulate local precipitation conditioned on the 700 hPa pressure field. The information content of the fuzzy classification as measured by precipitation‐related indices is similar to that of existing subjective classifications. The fuzzy rule‐based approach thus has potential to be applicable to the classification of GCM produced daily CPs for the purpose of predicting the effect of climate change on space‐time precipitation over areas where only a rudimentary classification exists or where none at all exists.