Summary: A negative phase of the Pacific/North American (PNA) pattern is forecast, which translates into an anomalously cold winter over much of the western part of North America. Anomalously mild winter is expected for the southeastern and (with less confidence) southwestern United States.

Fig. 1. The forecast of surface air temperature anomalies in North America in the winter (DJF) of 2008.  The blue (red) contours and the letter C (W) indicate colder (warmer) than normal temperatures. CL means near normal. The base period for calculating anomalies is 1971-2000. The numbers at the contours are confidence factors of the forecast. 

Discussion

The atmospheric circulation over North America will resemble a negative phase of the Pacific/North American (PNA) pattern. This phase of the PNA pattern is characterized by a frequent upper atmospheric troughing in the west and ridging in the east of the continent. This leads to relatively more frequent outbreaks of cold air in the west and less frequent in the east. See more information about the PNA at the CPC website.

Two major contributors to this state of the PNA pattern are: 1) the La Niña event in the equatorial Pacific and 2) the negative phase of the Pacific Decadal Oscillation (PDO) (see the latest forecasts for the ENSO and PDO). An analysis of a cumulative effect of other factors (briefly discussed below) also suggests that the negative phase of the PNA is more likely. 

One of the rules that were used in this forecast represents a tree-structured statistical model that links the quasi-biennial oscillation (QBO) in the stratospheric winds and year-to-year changes in the PNA index (Rule 202). This model suggests a high probability of a decrease in the PNA index compared to the last winter. A similar conclusion can be made based on an extrapolation, which makes use of an internal structure of the PNA time series. When the index decreases, there is an almost 80% chance that it will become negative.

A number of rules, however, point to a possibility of a positive PNA index in 2008. One of those rules describes a statistically significant correlation found between the PNA and the North Atlantic Oscillation (NAO) indices, with the latter leading the former by 3 years (Rule 35). In 1995, the NAO index was positive, and this suggests the PNA index may be positive in 2008. Two other factors, surface air temperature (SAT) in Greenland (Rule 199) and the Gulf of Mexico (Rule 200), act in favor of a positive PNA index as well. Due to the low confidence factors for these rules, however, their effect of the PNA forecast is relatively weak, and the evidence remains quite solid that the negative PNA index is more likely in 2008. 

This PNA forecast is consistent with the forecast of winter SAT anomalies in the southeastern US.  The latter forecast takes into account the state of the NAO, which is expected to be positive (see the winter 2008 forecast for the North Atlantic). Also, a quite strong relationship for the winter temperature in the Southwest was found with the QBO (Rule 31) and a weaker relationship with the Eurasian snow cover (Rule 27). Both rules suggest an anomalously mild winter for the region.

In the early fall of 2007, the distribution of sea surface temperature (SST) anomalies in the North Atlantic will be reviewed, which may correct the outlook for the winter of 2008. Currently, the so-called “horseshoe” pattern is forming there, and if this pattern remains unchanged through the summer and early fall, a warmer than normal winter can be expected in the southwestern US.

Significant uncertainty exists in the forecast for the Midwest and Ohio Valley region. The Climate Prediction Center (CPC) outlook calls for positive SAT anomalies in these regions calculated as departures from the 1971-2000 climatology. This is a relatively safe forecast given a strong positive temperature trend since the late 1970s. For example, in the central climatological region (click here to see the region map), only 4 out of 19 winters since 1989 were colder than the average temperature for the period 1971-2000. Our forecast predicts some cooling in this region, with SAT anomalies below the average for the past ten years and probably close to the average for the period 1971-2000.