A Record-Early Stratospheric Warming, But No Eastern US Chill? AI Models Discussion 10-29-2026

If you’ve been following long-range forecasts lately, you may have seen some chatter about a “sudden stratospheric warming” showing up in early November. These events have a reputation. Some of the most memorable cold outbreaks in recent years, like the 2018 “Beast from the East” in Europe, came in the weeks after one. So today we’ll look at what our Brightband-SUBS (BB-SUBS) AI ensemble is showing, whether we should believe it, and, most importantly, whether it means cold weather for the Eastern US.

The short version: the model is very confident that a stratospheric warming will happen, peaking around November 9. But it looks weak and short-lived compared with famous past events, and the forecast keeps the Eastern US warmer than normal for the rest of November.

What is a sudden stratospheric warming?

High above the weather we feel, about 20 miles up in the stratosphere, a ring of strong west-to-east winds circles the Arctic every winter. This is the stratospheric polar vortex, and it helps keep the coldest air bottled up near the pole. Sometimes large waves in the atmosphere push upward from below and knock the vortex off balance. The air over the pole warms, and the winds around it slow down or even reverse direction. In some cases the effects then work their way down toward the surface over the following weeks, loosening the “lid” on the Arctic air and letting it spill south into North America, Europe, or Asia.

What is the model showing?

The BB-SUBS run from October 28 (100 members, out to 32 days) shows the vortex getting stretched out over the next week (Figure 1). By November 2–6 it is nearly pulled into two pieces, one over Siberia and one over northeastern Canada. The Canadian piece then fades, and the vortex shifts toward the North Atlantic and Europe before settling back near the pole around November 23.

The standard way to call an event like this is to check whether the winds about 20 miles up near 60°N (roughly the latitude of southern Alaska and Oslo) flip from westerly to easterly. In this run they do (Figure 2). The winds start around 43 mph, slow steadily, and reverse in the ensemble mean around November 6. They bottom out on November 10 at about 11 mph from the east and are blowing from the west again by November 16. 95 of the 100 members reverse. Confidence has also grown quickly: the October 23 run had only a 9% chance of a reversal, and the five runs since have climbed through 49%, 73%, 81%, 89%, and now 95%.

The warming itself is more gradual than the “sudden” name suggests. Temperatures high above the pole rise by about 18°F over 12 days. That sounds like a lot, but the big midwinter events warm by 54–90°F in just a few days.

Should we believe it?

If this happens, it will be record-breaking. In 47 years of records (1979–2025), these winds have never reversed before December 5 (1981), and the only other case before December 10 was 1987 (Figure 3). A reversal in early-to-mid November would be about four weeks earlier than anything on record.

A record-breaking forecast naturally raises the question of whether the model has a habit of doing this. To check, we looked at 112 past BB-SUBS forecasts (reforecasts) for the same time of year from 2003 to 2025. Those forecasts almost never produced a reversal in early-to-mid November (well under 1% of members), and if anything the model tends to make the vortex a little too strong later in its forecasts. So this doesn’t look like a model quirk. Instead it appears to be driven by a real burst of wave energy pushing up into the stratosphere from late October into early November, about four times stronger than normal for this time of year.

How does it compare with past events?

To put this event in context, we compared it with three well-known stratospheric warmings: February 2018 (followed by the “Beast from the East” in Europe), January 2009 (the strongest on record), and January 2013 (followed by a cold March in Europe and eastern North America). On the day the winds reverse, the forecast vortex looks a lot like those events, with two lobes and a high-pressure area squeezing in between (Figure 4). The big difference is strength. It’s early in the season, so the vortex is still fairly weak to begin with: its winds never get above about 43 mph in this run, compared with roughly 80–150 mph in the month before each of the past events.

Lining the events up by the day the winds reverse makes the difference clear (Figure 5). The forecast winds only reach about 11 mph from the east and stay reversed for about 10 days. By comparison, the 2018 event reached 54 mph for 17 days, 2009 reached 67 mph for 29 days, and 2013 reached 29 mph for 22 days. The forecast vortex also recovers within about two weeks, while the past events stayed disrupted for a month or more. By most measures, this looks like a quarter- to third-strength event that crosses the reversal line mainly because the early-November vortex is still weak.

There is one place where this event stands out. In the lower stratosphere, about 10 miles up, the model shows a strong and persistent signal that is as big as 2018’s in absolute terms and very large for November. But that signal fades quickly on the way down. At the surface, there’s no sign of the classic “negative Arctic Oscillation” pattern, the setup that lets cold air escape the Arctic. In all three past events, that surface pattern was already in place about a week after the winds reversed.

What does this mean for the Eastern US?

This is the part most people care about, and the answer for November is: probably not much. Through the rest of the month, the model keeps a deep trough over the Gulf of Alaska and a strong ridge over western and central Canada that extends into the Eastern US (Figure 6). That’s basically the reverse of the classic East Coast cold pattern, which has a ridge over Alaska and a trough in the East. Eastern North America averages about 4–9°F above normal each week, and only 7% of members are colder than normal in mid-to-late November. The chance of a week landing among the coldest 20% for this time of year (based on 1990–2019) is below 5% everywhere in North America in every week of the forecast. The only notable cold in the whole forecast is over eastern Siberia during the first week of November, before the winds even reverse.

Part of the reason is that there just isn’t much cold air to tap right now (Figure 7). At the start of the forecast, the Arctic as a whole is about 8°F warmer than the 1990–2019 normal, Alaska and the Yukon are about 13°F warmer, and western Siberia is about 9°F warmer. Eastern Siberia is the only real pool of cold air, which is exactly where the model’s one cold week shows up. (Some of that Arctic warmth reflects the long-term warming trend compared with a 1990–2019 baseline, but the overall pattern still matters.)

Finally, Figure 8 compares surface temperatures in the weeks after the winds reversed. Within three weeks, the 2018 event had Europe about 11°F below normal, 2009 had Siberia about 14°F below normal (and later eastern North America about 4°F below), and 2013 had both Europe and eastern North America about 5°F below normal. All three show the “warm Arctic, cold continents” pattern you’d expect. The forecast has the warm Arctic but not the cold continents, with eastern North America running about 4–5°F above normal.

The Bottom Line

• The model is very confident (95% of members) that the stratospheric winds will reverse in early-to-mid November, which would be the earliest on record by about four weeks.
• This doesn’t look like a model quirk. The model rarely does this in past forecasts, and the event is driven by a real burst of upward wave energy.
• The event looks weak, shallow, and short-lived compared with high-impact events like 2018, 2009, and 2013, and the vortex rebuilds within about two weeks.
• So far there’s no sign of it reaching the surface. The Eastern US looks warmer than normal through the end of November.
• The thing to watch in future runs is the burst of upward wave energy driving this event: whether it holds up, and whether another one follows.

We’ll keep an eye on the stratosphere and be back with another update soon!

Thanks to Hans Mohrmann, Alistair White, and Ryan Keisler for the analysis behind this post.