An area-weighted k-means clustering method based on pattern correlations is proposed and used to explore the relationship between the Siberian High (SH) and Arctic Oscillation (AO) during the winter months (December-January-February) of... more
An area-weighted k-means clustering method based on pattern correlations is proposed and used to explore the relationship between the Siberian High (SH) and Arctic Oscillation (AO) during the winter months (December-January-February) of 1948–2014. Five regimes are identified. Four of these five regimes (comprising 171 of 201 months) show a negative correlation between the SH and AO indices, while the last regime (30 months) shows a positive correlation. The location of the SH shifts southward into China under two of the four negative correlation regimes (117 months), with pressure variations over the center of activity for the SH opposite to pressure variations over the climatological center of the SH (which is used to define the SH index). Adjusting the SH index to account for these spatial shifts suggests positive rather than negative correlations between major variations in the SH and AO under these regimes. Under one of the two remaining negative-correlation regimes, pressure anomalies are weak over the Arctic Ocean. In total, only one regime comprising 21 of 201 months strictly obeys the negative correlation between the SH and AO reported by previous studies. The climate regime characterized by an intensified SH is associated with a greater frequency of cold surges over northern and southeastern China, and the weakening of the East Asian winter monsoon during the 1980s was accompanied by a sharp reduction in the occurrence of this regime.
