donsutherland1

The CFSv2 has now moved into the early stages of its skillful range. Its monthly forecast for December is below. I assessed this idea against the currently expected state of the teleconnections. ENSO was omitted, as the sample size became so small that is was useless (n=2). Based on the expected predominant state of the teleconnections for December (WPO-/EPO-/PNA- and AO neutral) it appears reasonable. The AO should start negative but could go positive as December progresses based on the current guidance. The second half outcome will depend on whether or not the stratospheric warming event propagates to the troposphere to promote a redevelopment or persistence of AO blocking. So, for purposes of conservatism, I treated the AO as neutral (monthly average above -0.50 but below +0.50) for looking at the monthly CFSv2 idea. Should an AO+ regime develop (SSW fails to propagate in a fashion that promotes AO blocking), warm anomalies could cover much of the Eastern Seaboard with the exception of central or northern New England and have the risk of spreading into at least the Ohio Valley and perhaps farther westward depending on the strength of the AO+. For now, this is an identified risk, but not the most likely outcome. On the other hand, should strong AO blocking develop, most of the eastern half of the U.S. would likely see cold anomalies. This scenario would also have the highest probability of seeing a strong Arctic outbreak reach the East Coast. Given the present guidance, this seems to be less likely than an earlier breakdown of the AO-. But the stratospheric warming event is a wildcard and how such events ultimately play out over long timeframes is not a high-skill undertaking. It should also be noted that teleconnections cannot be forecast skillfully beyond 10-14 days. All in all, the combination of the CFSv2 and teleconnections highlights the Great Lakes Region and Upper Midwest as having the greatest probability of experiencing cold anomalies. They also highlight the Great Lakes Region as having the highest probability of above normal monthly snowfall. As for my thinking using the framework I illustrated yesterday, I've highlighted what has changed in my thinking overnight in blue. The emergence of a neutral AO on the guidance was noted (there's a very large spread among the ensemble members). No changes were made beyond December 10 as signals remain mixed and there remains no clear signal for a change.

A generally milder than normal pattern will likely continue through next Wednesday. However, exceptional warmth appears unlikely. The closing days of November will likely turn colder. There is potential for New York City's Central Park to experience its first freeze of 2025. The 1991-2020 normal first data is November 21. The 1961-1990 baseline was November 11. Last winter's first freeze occurred on November 30. The opening days of December will turn milder. However, colder conditions could begin to develop during the second half of that week and continue through the second week of December. Severe cold appears unlikely through at least the first 10 days of December. Afterward, the pattern evolution will depend, in part, on how the imminent stratospheric warming event propagates. Meanwhile, today will be Central Park's 1,393rd consecutive day without daily snowfall of 4" or more. The record of 1,394 days was set during February 22, 1929 through December 16, 1932. That stretch ended with 6.7" daily snowfall on December 17, 1932. The ENSO Region 1+2 anomaly was -0.7°C and the Region 3.4 anomaly was -0.7°C for the week centered around November 12. For the past six weeks, the ENSO Region 1+2 anomaly has averaged -0.16°C and the ENSO Region 3.4 anomaly has averaged -0.65°C. La Niña conditions will likely continue through at least mid-winter. The SOI was +31.56 today. The preliminary Arctic Oscillation (AO) was +1.436 today. Based on sensitivity analysis applied to the latest guidance, there is an implied 86% probability that New York City will have a cooler than normal November (1991-2020 normal). November will likely finish with a mean temperature near 46.6° (1.4° below normal). Supplemental Information: The projected mean would be 1.1° below the 1981-2010 normal monthly value.

That’s true. Unfortunately, there’s no daily TNH index to quantify things.

Below is an illustration of an extreme forecast that is being pushed on Twitter/X despite the lack of evidence to actually support the idea (e.g., small pool of deep cold in the Northern Hemisphere) and the lack of skill in guidance from this far out: "Lower Rio Grande Valley" doesn't specify a specific point or location for purposes of verification, so it will be excluded. Here's the forecast coupled with December 2000-2024 statistics: This illustrative example will be verified at the end of December. Given the estimated probabilities for various outcomes, the odds are strongly against the extreme forecast verifying in more than one of the above locations.

Several things: 1) There remains a lot of uncertainty how the stratospheric warming will propagate. Propagation could be a big factor in prolonging/strengthening or bringing back Atlantic blocking. 2) Even as stratospheric warming often (not always) propagates favorably to build or prolong/strengthen Atlantic blocking, one should not immediately dismiss hints of evidence to the contrary. Confirmation bias is bad for forecasting. Such evidence exists e.g., here's the AO forecast from the EPS 46-day forecast: Although the below table talks about considering climate change-related impacts on forecasts, I highlighted the section that relates to how and why confirmation bias skews forecasts: A systematic forecasting framework allows one to acknowledge such evidence while avoiding rash decisions from model noise. As I don't have a crystal ball to have 100% confidence in whether or not the above AO forecast will verify, I have made a note to continue to monitor developments. Needless to say, there is more to it than such a framework. Each scenario has clusters of outcomes (not discussed above). It should be noted that such frameworks don't guarantee accuracy, but they do reduce the risks that can otherwise lead to bad forecasts, especially when one is dealing with timeframes over which model skill is limited or worse. Finally, the 11/22 12z guidance has reaffirmed the December 1-10 idea. Indeed, more than half of the 11/22 0z EPS ensembles suggested 1" or more snow in Detroit and Toronto (a good signal for measurable snow prospects from this far out). I expect those figures to hold up with the 12z cycle.

Yes. I used a full progression through Phase 2 for fuller insight. I plan to take a closer look at the MJO cases.

The teleconnection idea (predominant state during the periods) is based on the GEFS and EPS, including the weekly EPS ideas. That's why I noted that the probability for Scenario 3 has increased somewhat, as some of the guidance now moves toward a more robust AO+. The consensus remains a persistent PNA- (closer to neutral but not positive). Here's the latest GEFS forecast going into December (preponderance of ensemble members show a PNA<0):

And going beyond the lack of skill in MJO forecasts beyond 10 days, other factors can overwhelm it. Sometimes even high amplitude passages through Phase 8 during passage through Phases 7-2 don't generate a wintry response in North America. December 1-12, 1990 is one example. March 1-25, 2023 is another example, even as it was an off-the-scales passage through Phase 8.

Here's my thinking based on the latest guidance and some of the longer-range guidance. I focus on what I believe are the three biggest potential scenarios for much of December.

The 0z EPS rolls the cold forward pretty much as one would expect with a WPO-/EPO-/AO-/PNA- pattern evolution. One sees how the Great Lakes/Upper Midwest is the focus of the cold while warmth proves very persistent in the Southeast. 288 hours: 312 Hours: 336 Hours: 360 Hours: The usual caveats concerning skill apply that these timeframes. Moreover, how the stratospheric warming propagates will influence the pattern evolution for the second half of the month.

I wanted to illustrate that CPC isn’t panicking. This map at that timeframe is along the lines of what many of us are thinking. It’s very close to my baseline scenario.

The latest week 3-4 CPC outlook:

That's a great map.

A generally milder than normal pattern will likely continue through next Wednesday. However, exceptional warmth appears unlikely. The closing days of November will likely turn colder. Meanwhile, today will be Central Park's 1,392nd consecutive day without daily snowfall of 4" or more. The record of 1,394 days was set during February 22, 1929 through December 16, 1932. That stretch ended with 6.7" daily snowfall on December 17, 1932. The ENSO Region 1+2 anomaly was -0.7°C and the Region 3.4 anomaly was -0.7°C for the week centered around November 12. For the past six weeks, the ENSO Region 1+2 anomaly has averaged -0.16°C and the ENSO Region 3.4 anomaly has averaged -0.65°C. La Niña conditions will likely continue through at least mid-winter. The SOI was +26.47 today. The preliminary Arctic Oscillation (AO) was +0.707 today. Based on sensitivity analysis applied to the latest guidance, there is an implied 83% probability that New York City will have a cooler than normal November (1991-2020 normal). November will likely finish with a mean temperature near 46.6° (1.4° below normal). Supplemental Information: The projected mean would be 1.1° below the 1981-2010 normal monthly value.

A number of boundary conditions e.g., SSTAs, and teleconnections this year differ from those in recent La Niña winters that saw extremely low snowfall. None of this guarantees a “big” winter, but it does mean the range of plausible outcomes is perhaps broader than what we’ve seen the last few years.

Interesting. You should ask it to follow up with links to related scientific papers on which its assessment was based.

Yes. That’s true.

He’s probably right regarding social media. But it has been discussed fairly widely here.

Yes. There were near misses.

The probabilistic maps from the latest CFSv2 monthly forecasts (just outside its skillful range) appear to be consistent with my baseline thinking (Upper Midwest/Great Lakes-focused cold that spreads eastward, above normal snowfall in the Upper Midwest-Great Lakes Region ( including Chicago to Toronto). Finally, there continues to be no evidence of the kind of buildup of expansive severe cold in the Northern Hemisphere that would produce a severely cold December in the CONUS. This does not preclude the possibility of 1-2 Arctic outbreaks, but those are synoptic events that can't be forecast from this far out. But the path toward an extremely cold month is one that is statistically very unlikely and for which there is no evidence currently to support such claims. Therefore, the Social Media chatter toward that end is pure speculation. The idea that December would rival 1983 is nonsensical. December 1983 (CONUS mean temperature of 25.48°) was the coldest December on record in the CONUS. The last December with a CONUS mean temperature below 30° was December 2009 (29.64°) and that outcome was made possible by extraordinary and persistent blocking in the WPO/EPO/AO domains. Both those frigid Decembers were preceded by Novembers with a much larger deep cold pool in the Northern Hemisphere. So, a logical question would be how such cold would materialize if the deep cold pool is small and there is no indication of 2009-style extreme blocking on the guidance (which can't reliably be forecast from this far out)? If the ingredients aren't present or can't yet be determined to be present, one can't credibly call for such outcomes, especially when they are rare statistical events. Finally, the odds are further tilted against such an outcome by the warming that has occurred in the Northern Hemisphere (especially the Arctic) since 1983. Below are the GISS Arctic region temperature anomalies since 1980.

Simon is highly credible with deep expertise. He doesn't make outlandish statements for clicks or subscriptions. He has done important research. I am aware of his concerns and take them seriously. With regard to the upcoming SSWE, I have outlined what is my current baseline thinking. Overall, I have three general scenarios: Scenario 1: Great Lakes Focused Cold, East Gets a Window (EPO-/AO-/PNA-) -- My Baseline Scenario 2: Colder/Blockier; Classic Eastern Winter Period (EPO-/AO-/PNA+) Scenario 3: Delayed or Short Impact/Persistent SE Ridge; Cold Underperforms in the East; reduced snow opportunities in the Mid-Atlantic Region (EPO-/AO+/PNA-) Under my current baseline, cold would first move into the Great Lakes Region and then spread farther east (except for the Southeast; the cold will be fleeting there). The late November cold shot is unrelated to the stratospheric event. Should the forecast AO- break down relatively quickly, the odds of Scenario 3 would increase. Given guidance skill (lack thereof) at long timeframes, I am sticking with the baseline until there is credible evidence it is going off track. Interestingly, during Winter 1996-1997, January was the coldest month relative to normal for such cities as Detroit and New York. That outcome was suggested by my final winter analog set.

It will turn milder tomorrow. Once in place, the generally milder than normal pattern will likely continue into the closing days of November. However, exceptional warmth appears unlikely. The closing days of November could turn cooler. Meanwhile, today will be Central Park's 1,391st consecutive day without daily snowfall of 4" or more. The record of 1,394 days was set during February 22, 1929 through December 16, 1932. That stretch ended with 6.7" daily snowfall on December 17, 1932. The ENSO Region 1+2 anomaly was -0.7°C and the Region 3.4 anomaly was -0.7°C for the week centered around November 12. For the past six weeks, the ENSO Region 1+2 anomaly has averaged -0.16°C and the ENSO Region 3.4 anomaly has averaged -0.65°C. La Niña conditions will likely continue through at least mid-winter. The SOI was +31.43 today. The preliminary Arctic Oscillation (AO) was -0.316 today. Based on sensitivity analysis applied to the latest guidance, there is an implied 78% probability that New York City will have a cooler than normal November (1991-2020 normal). November will likely finish with a mean temperature near 46.9° (1.1° below normal). Supplemental Information: The projected mean would be 0.8° below the 1981-2010 normal monthly value.

It's something we should be watching closely. The SSWE is imminent. Considering a typical 10-30-day lag, I suspect that the December 10-25 period might see the coldest anomalies relative to normal for such cities as Chicago, New York, and Philadelphia. That's still far out and we'll have to see how things evolve (does the PV split, where do the twin vortices wind up, etc.).

Yes. Since 2020, Pittsburgh has had 7 such days while New York City has had 6. Temperatures have not warmed so sharply and abruptly that such storms are precluded. I suspect that this winter will see at least one such day in both cities. It should also be noted that New York City has gone through two long stretches where 4" or above snow days were infrequent.