Winter 2008 in the North Pacific - post-mortem analysis

Fig. 1. SST anomalies in the winter (DJF) of 2008.

The SST anomaly pattern in the winter of 2008 (Fig. 1) closely resembled the negative phase of the Pacific Decadal Oscillation (PDO), which was in line with our forecasts issued in June and September 2007. The PDO index averaged for the months of December through February was -0.86, the third lowest value since the climate regime shift in 1976/77 (Fig. 2).

Atmospheric circulation over the North Pacific featured a strong Subtropical high (Fig. 3). When the regime shift detector is applied to the time series of its central pressure, the results clearly shows a regime shift in 1999 (Fig. 4). The values of the North Pacific index, which is frequently used to characterize atmospheric circulation over the North Pacific, were positive in December and January, but turned negative in February (not shown).

Fig. 2. Winter (DJF) PDO index, 1901-2008.

Although the general patterns of atmospheric circulation and SST anomaly distribution were predicted correctly, there were two specific regions where the forecast failed. The first region was the Bering Sea, where, contrary to our latest forecast from September 2007, air and sea temperatures in its eastern part were colder than average. In this sense, our earlier forecast issued in June 2007 was better. The forecast was changed later, primarily because of rule 98, which shifted the odds in favor of a mild winter in the Bering Sea. Based on that rule, it was expected that the mid-tropospheric high pressure cell in the central North Pacific (Fig. 5) would be shifted a bit farther to the north and deflect more storms to the Bering Sea. Instead, during the winter of 2008, storms maintained a more zonal track, often traveling south of the Aleutian Islands into the Gulf of Alaska. An advection of cold air behind their cold fronts resulted in anomalously cold winter in the eastern Bering Sea. 

Fig. 3. SLP anomalies in the winter (DJF) of 2008.

Another region were the forecast failed was south of Japan. Typically during El Niño events, the West Pacific (WP) index tends to be positive (rule 208), which is accompanied by a warming in the area south of Japan (Fig. 6). The opposite is true for La Niña events. Although the WP index in the winter of 2008 decreased from the previous winter, it still remained in the positive territory. The winter of 2008 was extremely cold in China, North Vietnam and their offshore waters (Fig. 7; see also media reports). Some cooling in the area south of Japan was observed in February 2008 (not shown), but it was not enough to turn the mean winter temperature values negative.

Generally speaking, both the Bering Sea and the Kuroshio Current area south of Japan are among the most difficult regions for seasonal climate forecasting. Even seemingly minor variations in atmospheric circulation may result in different temperature anomalies in those regions, which is particularly true for the Bering Sea.

Fig. 4. Winter (DJF) central pressure in the subtropical high, 1949-2008.	Fig. 5. 500-hPa height anomalies in the winter (DJF) of 2008.
Fig. 6. Correlation between the winter (DJF) WP index and SAT anomalies.	Fig. 7. SAT anomalies in the winter (DJF) of 2008.