By using the monthly data from 1951 through 1984, empirical orthogonal ex-pansion is performed for the 500 hPa geopotential height north of 65°N and the canonicalfields are clustered by fuzzy classification. It is noted that both the mean monthly polarvortex fields and the large-scale anomaly fields fall into three regimes, with those ofthe January mean field and the April anomaly field having characteristic features. In addition,the relationship between the time weighing coefficients of the canonical fields andEl Nino/SO is examined,showing significant anomalies in the large-scale polar anomaly fields during April and October of the year when E1 Nino occurs. These polar circulation anomalies havd considerably influenced the temperature fields in China during Apriland October. Thus, we may conclude that this is one of the most important reasons for arelatively cool April and a warm October in China during the El Nino year.