Despite the slow but steady advances in weather forecasting, seasonal outlooks for winter are not likely to keep up with this improving trend. In fact, I would contend the opposite has been and will be occurring: a deterioration in accuracy from what was already a less than impressive track record.
This deterioration will be episodic or periodic within the context of an entire winter season. These episodes may only cover a small part of a season, but they may be the most memorable periods of extreme winter weather over 90 to 120 days. The episodic nature of these extremes can be coupled with climate change, particularly the marked warming in the arctic region. The problem is these episodes cannot be foreseen months in advance; it’s more a matter of a couple of weeks’ predictability. So, long-range forecasters trying to predict trends for an entire winter season are facing a potentially greater probability for “short fuse” pattern changes occurring along the way.
A classic example of this occurred early in the winter of 2001-2002. The focus was on Christmas week, which brought lake-effect snow amounts to the heart of the metro area never matched before or since on record.
About 82 inches of snow fell that week in Buffalo. But by the end of the cold weather season, the seasonal total was around 132 inches, meaning only around 50 inches fell the rest of the season outside of that amazing week. This abrupt extreme period could only be seen a week or less in advance, though seven-day forecasting during that set of lake snow events was excellent. On the other hand, what long-range seasonal forecaster could see such an anomalous week back in October or even November? The rest of the cold weather season was less snowy and milder than average. If a winter outlook made in late October had predicted milder than average temperatures, it would have verified. Yet that one week took snow from a distinct lower trend to above average for the season. Catch my drift? (Sorry.)
Climate models certainly pick up on a warming climatology, with warming that continues to accelerate over recent decades on a global basis. On this basis alone, long-range models reflect ongoing trends. Last week, a seasonal ensemble of the European model (an ensemble with a collection of 51 separate runs of the model) already was projecting warmer than average winter temperatures over much of the middle and eastern parts of North America. This was posted by Boston’s WBZ meteorologist Eric Fisher, courtesy of weathermodels.com.
There are many other factors going into model output beyond climatology. There are current sea surface temperatures in the Pacific, atmospheric/ocean surface interacting oscillations that include the El Nino Southern Oscillation (ENSO) as well as other shorter fuse oscillations. (Sidebar: ENSO is expected to remain mainly neutral – neither favoring el nino nor la nina phrases. Taken by itself, a neutral ENSO has often been associated with colder than average and, sometimes, snowier than average conditions in Western New York. But we have learned not to take a single factor by itself in most cases, so don’t draw any conclusions from this.)
According to a long-range European model ensemble, all factors considered on a regional and global basis currently favor a milder than average winter over the course of the cold weather season.
In addition to the European ensemble above, there are other seasonal models, such as the American CFS model. It favors a dominant mild Pacific flow in November.
Unlike the European, however, the CFS brings a cooling trend to our part of the country in December, with a similar trend in January before winter moderates in February and March.
The Canadian seasonal model more closely agrees with the regionally colder American model for mid-winter.
The National Weather Service Climate Prediction Center (CPC) looks at all models. CPC favors a milder than average start to winter, from November through January.
By mid-winter, CPC comes around to replace the milder probabilities with “EC” for the north central and Great Lakes region. EC means, essentially, equal chances temps could go either way. This suggests CPC is not confident the predominantly milder trend will hold up.
The hypothesis underlying this article is the short fuse more extreme wintry outbreaks will become more likely in this long-term arctic warming. The lessening of the temperature contrast between high arctic latitudes and the mid-latitudes periodically weakens the jet stream and the polar vortex. When the polar vortex is strong, it tends to bottle up the polar air in the high latitudes. When the vortex weakens, it tends to buckle southward in a temporarily “blocky” pattern, which can deliver harsh winter extremes to several regions in the mid-latitudes.
Because these occasional episodes have become more likely with arctic warming, and because the blocky patterns can’t be foreseen more than a couple of weeks in advance, in many years this will mean less consistent winter patterns, with more ups and downs. Sometimes these ups and downs may tend toward the extreme. Over the long haul, the ups will probably outnumber the downs and, in a few decades, may overwhelm the downs.
In the meantime, here in The Buffalo News I will steer shy of going much beyond a couple of weeks for trends most of the time, unless I’m feeling wild and crazy because some peer-reviewed study proves my hypothesis is hooey.