Plain Language Summary
I use October Eurasian snow cover extent (SCE) as one of our main predictors for winter temperature anomalies across the Northern Hemisphere (NH). Eurasian SCE was slightly above normal for the month of October coupled with La Niña is biasing the AER winter forecast on the cold side of normal for western and central North America, Northern Europe. The winter forecasts from the government centers are much warmer (at least in extent). One wild car could be the water temperatures in the North Pacific, if strong enough they can favor the core of cold temperatures further to west across North America.
Impacts
I have been promising a preliminary winter forecast for surface temperature anomalies for the Northern Hemisphere and today I deliver. The forecast was generated the last week of October using my estimate of October Eurasian snow cover extent (SCE), the dominant sea level pressure anomaly across Northern Eurasia for October, the September Arctic sea ice extent anomaly and the predicted winter (December-February) Niño 3.4 index (a common indicator for El Niño/Southern Oscillation (ENSO)). Also I included in this winter’s forecast my own very rough estimate for the winter Pacific Decadal Oscillation (PDO) index. Typically, I don’t include the PDO index in my winter forecasts, but the strong negative PDO pattern has emerged, and it could very well be a factor. in Figure i, I present a preliminary AER winter forecast that was shared with clients at the end of October. I will share an update to the winter forecast on the blog at the end of the month. Please keep in mind since the forecast was generated the observed values for October SCE and SLP are now known and will be used in updated forecasts. Also, I used as my climatology the values from 1991-2020.
Figure i. Forecasted surface temperature anomalies for December 2021 - Fwbruary 2022 from the AER model.
I have to say when the forecast was first generated, I was surprised by how cold the forecast appears. At this point I don’t have a lot of confidence in the forecast, and it will certainly be revised. But just to emphasize how cold the forecast appears, at least visually (and to be honest the anomalies are not large), all I need to do is present in Figure ii a consensus of the dynamical model temperature anomaly forecast from the North American Model Ensemble (NNME). The dynamical models are predicting almost universal relative warmth across the Northern Hemisphere continents except for Alaska and Western Canada. Across North America it is a canonical La Niña pattern and across Eurasia is it a global warming signal. To be fair there is probably some global warming influence included in the North American forecast as well.
Figure ii. Forecasted surface temperature anomalies for December 2021 - Fwbruary 2022 from the NMME models.
I have discussed this before, but the lack of variability among the model forecasts is noteworthy and one should look at all the NMME more forecasts to see just how much they all resemble each other (see NMME winter forecasts). The European C3S ensemble also looks similar but has not been updated for November yet (see C3S winter forecast). In my opinion observational analysis between ENSO and NH winter temperatures does not support such a high degree of confidence in the forecast from the dynamical models. I have argued in my papers that the dynamical models are insensitive to Arctic forcing and overly sensitive to tropical forcing (e.g. Cohen et al. 2016). The large amount of variability or noise in the climate system is used to criticize arguments of Arctic influence (including my own), i.e., the large amount of noise makes it impossible to distinguish or identify the signal of Arctic influence, i.e., all noise and no signal. Yet when it comes to tropical influence it is all signal and no noise? Anyway, enough of my griping.
But as I have argued many times before even if the AER winter forecast is completely wrong, I strongly believe that it provides value in at least it is an alternative forecast and not a herd forecast. And all we need to do is look to last winter to see how wrong the herd forecast can be (see the blog post from March 8, 2021). But I readily admit if the polar vortex remains strong this winter, the dynamical models will look like champs in March 2022, which is one takeaway from winter 2019/20.
I do think that Arctic surface conditions, especially with the recent snow blitz across Asia are well suited for disrupting the polar vortex. Negative sea anomalies heading into the winter are focused in two regions. The first is the Barents-Kara Seas and low sea ice in this region is thought to be most favorable for disrupting the polar vortex. The other region is Baffin Bay. As I showed last week, negative sea ice anomalies in this region are associated with relatively cold temperatures in the Eastern US and Northern Europe. Whether sea ice anomalies in this region can force anything I don’t know, and they could just be a lagging indicator (they indicate recent Greenland blocking that favors relatively cold temperatures in the Eastern US and Northern Europe but warm temperatures in Baffin Bay, which inhibits sea ice formation).
Though October SCE was not impressive and only slightly above normal but so far November SCE is very impressive. But for full disclosure, I have not shown any analysis that November SCE has any influence on winter weather, but other studies have (see blog from October 25, 2021). Also, there are three weeks left and SCE could still change dramatically.
But snow and sea ice can remain very favorable for disrupting the polar vortex from now until March, but they alone are not sufficient to force a polar vortex disruption. Whatever signal they provide will never make it to the polar vortex without blocking/high pressure in the Barents-Kara Seas/Ural/Scandinavia region. Right now, there are no signs of a meaningful return of blocking/high pressure to this region. Based on the model forecasts I believe you have to bias the probability of a relatively strong polar vortex and a relatively mild pattern across the NH including the Eastern US and Northern Europe heading into December. The best possibility of resulting in meaningful cold, especially to the Eastern US, in my opinion would be another stretched polar vortex in December as is occurring now. A larger polar vortex disruption that is associated with a sudden stratospheric warming takes longer to occur and without dramatic changes I don’t see one occurring in December.
But it is early, and I do believe that Arctic snow and ice favor a return of Barents-Kara Seas/Ural/Scandinavia region more so than its continuous absence. But also in my opinion, the blocking needs to be persistent and not transitory to have enough influence to impact the winter means. Early in my career all I wanted to see was Greenland blocking but now I am much more fixated on Ural blocking. I see little evidence that Greenland blocking can perturb the polar vortex, instead Ural blocking is critical. And I am more and more convinced that without some type of disruption to the polar vortex notable (or from my perspective memorable) winter weather is nearly impossible. But to be fair to Greenland blocking it may be an early precursor to polar vortex stretching as we showed in Cohen et al. 2021. And I don’t intend to diminish that Greenland blocking is strong signal for a Northeastern US snowstorm and certainly cold and snow for Northern Europe. Whether the new obsession is warranted is still an open question.
Finally, I did see the October value for the PDO and it was deep into negative territory; seems to me that it has been a while since we have observed a number that negative and my PDO index used as a predictor in the winter forecast may be too conservative. A more negative PDO would favor pulling the cold temperatures across North America further west with more warming in the Eastern US. I wish that I could be as confident as the dynamical models but as I see it, there are many mixed signals so far and therefore lots of uncertainty. AO Blog Update | AER | Weather Risk Assessment
----------------
Engels niet machtig? Probeer eens translate google.