Director, Seasonal Forecasting
Atmospheric and Environmental Research
JANUS Research Group
October 20, 2025
Summary
The Arctic Oscillation (AO) is currently negative to near neutral and is predicted to oscillate around neutral and possibly go negative again the next two weeks as pressure/geopotential height anomalies across the Arctic are currently mostly mixed and are predicted to remain mostly mixed to mostly positive the next two weeks. The North Atlantic Oscillation (NAO) is currently negative as positive pressure/geopotential height anomalies dominate across Greenland, and the NAO is predicted to oscillate around neutral the next two weeks as pressure/geopotential height anomalies are predicted to mostly mixed to positive across Greenland.
The next two weeks, weak ridging/positive geopotential height anomalies across Greenland will support transient troughing/negative geopotential height anomalies downstream across much of Europe. The progressive flow pattern will support normal to above normal temperatures across Europe including the United Kingdom (UK) with a brief flip to normal to below normal temperatures first across much of Europe including the UK next week.
Ridging/positive geopotential height anomalies are predicted to dominate Asia with the exception of troughing/negative geopotential height anomalies across Eastern Asia the next two weeks. This pattern favors normal to above normal temperatures across most of Asia with the exception of normal to below normal temperatures regionally across Eastern Asia the next two weeks.
The general predicted pattern across North America the next two weeks is widespread ridging/positive geopotential height anomalies centered on Eastern Canada with troughing/negative geopotential height anomalies along the west coast of North America and the Eastern United States (US). This patten will favor widespread normal to above normal temperatures across Alaska, much of Canada and the US with the exceptions of normal to below normal temperatures in the Eastern US this week and then next week normal to below normal temperatures will sweep into the West Coasts of Canada and the US and then eventually in the Eastern US.
I continue discussing my thoughts about the upcoming pattern across the Northern Hemisphere (NH) and continue with some preliminary predictors for the upcoming winter polar vortex (PV) and weather.
Plain Language Summary
So far for October, it is a tale of two continents, extremely mild in North America and fairly widespread cold across Eurasia, especially in Southeast Asia, Siberia and focused now in Mongolia (see Figure). It has also been chilly in Central Asia and in Europe, especially in Southeast Europe (see Figure). No joke but Mongolia has the most resilient winter weather in the Northern Hemisphere. It has been a forgettable month of October for me here in Boston but at least for me the cold is where it should be in October. And the forecast can be summarized as mild weather reigns supreme (see Figures 3, 6 and 9). Longer term Greenland blocking could become established in early November that could finally bring cooler weather to North America and re-establish some colder weather across northern Eurasia. I am also closely watching the snow cover advance across Siberia this month, a rapid advance suggests a more severe winter and a slower advance a milder winter.
Figure.
Estimate of the observed surface temperatures (°C; shading) from 01 Oct to 19 Oct 2025 based on GFS initializations and the GFS forecast from the 20 October 2025 run.
Impacts
Once again, first some business that I need to take care of. Unfortunately, readers should come to expect disruptions to the blog going forward. We are working on providing the blog on a best-available effort at the moment, and look forward to resuming regular publication soon.
In addition, the AWS outage impacted the blog graphics today. The graphics are from yesterday though the discussion in the text is mostly from the 00Z suite of the GFS run from 20 October 2025. So there may be slight discrepencies between the discussion and the graphics.
This week, I will start off looking forward to the winter polar vortex and related weather. Seems to me that there has been plenty of early hype on the polar vortex (PV). Maybe the lack of hurricanes created a weather void that needed to be filled but some of the hype is probably deserved but most of it not. There are multiple reasons to be excited for an active PV season related to multiple episodic, regional and maybe even widespread bouts of severe winter weather, which I will soon list, but the same could be said in fall 2019 and as far as I am concerned that was a dud of a winter as far as interesting PV behavior and the lack of winter weather was quite notable.
The first reason to start taking seriously the idea of notable PV disruptions this winter is October Siberian snow cover. Siberian snow cover got off to a quick start in September and that has continued through the first half of October (see Figure i). My daily snow cover plot is from Friday (and unfortunately due to the AWS outage, I can’t access my code) so somewhat dated but it does appear that the snow cover advance has slowed. Last October, also after a very quick start, the snow cover advance went sideways for the remainder of the month in what may have been an unprecedented mid-month reversal. I don’t foresee a repeat but at least for this week the advance seems to have taken a pause. However, if overall the remainder of the month is close to normal, the snow cover extent (SCE) for October should come in above normal and higher than the last four Octobers.
Figure i.
Daily snow cover extent across Eurasia for 1 through 17 October 2025 shown in red. All Octobers since 2015 are also shown as well as the long-term average shown in dashed black line.
I also compute the snow advance index (SAI). I mostly use the SAI currently to anticipate the timing of a potential sudden stratospheric warming (SSW). And in that regard, I would couple that with the quasi-biennial oscillation (QBO), which I currently use also as an indicator of the timing of a possible winter SSW. If the SAI is positive and the QBO is easterly, this favors an earlier SSW, if one were to occur in winter. If the SAI is negative and the QBO westerly that would favor a later SSW, if one were to occur in winter. Last October the SAI was negative and the QBO westerly and for all intents and purposes we had an SSW in early March (it is labeled a Final Warming but dynamically it was indistinguishable from an SSW). That is very late for an SSW so those predictors worked well last winter as far as I am concerned. The QBO is definitely easterly, and the SAI is currently positive. Not a given, but I expect the SAI to finish the month positive. These two factors would favor an earlier SSW (again if it were to occur) making any SSW much more of a player in winter weather, certainly much more so than last winter whose impact was delayed well into spring.
I am also watching Arctic sea ice. Current conditions are shown in Figure ii. It has been shown that less sea ice in the North Atlantic sector of the Arctic weakens the polar vortex while less sea ice in the North Pacific sector strengthens the polar vortex. The latest Arctic sea ice anomalies have taken a dramatic and decided shift or weighting towards the North Atlantic relative to the North Pacific sector. I think the deficit on the North Atlantic side is too great to be overcome probably through the end of the year and this too would favor an overall weaker PV this winter.
Figure ii.
Arctic sea ice cover extent for 19 October 2025. White depicts ice covered areas and the orange contour the climatological extent of Arctic sea ice for the date. Plot taken from: https://nsidc.org/sea-ice-today
Another winter predictor that I use in the AER seasonal forecast is the dominant monthly mean sea level pressure (SLP) anomaly across northern Eurasia. Again, there has been no lack of high latitude blocking across Eurasia however right now it is over Central Siberia. That is in a position that favors a stronger PV and not a weaker one, but this is the predictor that is the most fluid this month and I won’t make a call until the month is over.
On a related note, I am working with a team mostly from Microsoft to develop an AI subseasonal model and we are competing in the ECMWF sponsored AI Quest. It is very early but our model has so far been highly competitive and consistently outperforms the AIFS. I thought that it was interesting that our model is predicting a negative AO and Greenland blocking for early November (see Figure iii) much more so than the AIFS. I also included in the figure the contemporary forecast from the AIFS. Not optimal for weakening the PV but would support a colder pattern in Northern Europe and/or Eastern US. Anyway, I thought something of interest to watch.
Figure iii.
Forecast of the probability of the top 20% of observed mean sea level pressure for 8 November through 9 November 2025. Forecast issued 16 October 2025. Plot taken from https://aiweatherquest.ecmwf.int/
My final point about the winter PV and the possibility of an earlier more impactful SSW, the entire forecasting community has been vexed predicting the correct tropospheric response to SSWs. A right or wrong forecast has profound implications for anticipating and preparing for severe winter weather. I myself have been frustrated waiting for the response from an SSW to reach the lower troposphere and many times it can feel like “Waiting for Godot.” Thanks to my collaboration with a talented younger scientist Marlene Kretschmer, I may have seemed obsessed with stretched PVs (and I believe with good reason) but I haven’t “taken my eye off the ball” on SSWs. And with my talented MIT postdoc Mostafa Hamouda (now at GFZ in Potsdam) I feel that we have made great progress in improving whether Greenland blocking will follow an SSW within the first 30 days. The accuracy of the ECMWF S2S model is less than 50% whereas our ML based model is 80% and even higher. We are trying to write-up the results for publication, but I am excited to try out the model in real-time this winter. So, stay tuned.
In conclusion, in my opinion, three important Arctic predictors that I rely on for the winter forecast: the Eurasian SCE and SAI, the Arctic sea ice so far all favor an earlier and/or weaker polar vortex this winter and more widespread severe winter weather across at least parts of the mid-latitude continents. Add you can also include the QBO as a fourth predictor. But as Orson Welles said about Paul Masson wine (and Beethoven), “I will issue no forecast of the PV before its time,” and it is still not time.
For much of October there has been plenty of high-latitude blocking, though at many times not in the places that I would prefer to see. Currently the high latitude blocking is located where it is mostly contributing to mild weather across most of the Northern Hemisphere (NH) continents. Once again in general, the two-week forecast for the mid-tropospheric circulation is characterized by low pressure centered near the North Pole and high latitude ridging along the periphery of the Arctic (see Figure i). This week, two dominant high-pressure centers are predicted one over eastern Canada and the other over Siberia (see Figure ii). Then next week the Siberian high pressure is predicted to slide west closer to the Urals.
Figure iv.
Initialized 500 mb geopotential heights (dam; contours) and decameter anomalies (dam; shading) across the Northern Hemisphere for 20 Oct 2025 and forecasted from 21 Oct to 04 Nov 2025. The forecasts are from the 00Z 20 Oct 2025 GFS model ensemble
Which brings me to the final point of today’s discussion, the possibility of a stretched PV (and really how did I make it this far into the blog without bringing up my two favorite atmospheric features Ural blocking and a stretched PV; my impulse control is impressive). If you live east of the Rockies in North America, you couldn’t be faulted in believing that the Hudson Bay high pressure ridging is an atmospheric feature that has even preceded the most famous celestial permanent atmospheric anticyclone – Jupiter’s read spot. This is eerily reminiscent to me of a Canadian warming. And lo and behold many (but not all) PV forecasts resemble to me at least, a Canadian warming (see Figure v). And if you read last winter’s blog, then you surely remember that the smart money is on that a stretched PV will follow. So maybe a boring October will be followed by a more interesting November?
Figure v.
Forecasted average 10 mb geopotential heights (dam; contours) and geopotential height anomalies (m; shading) across the Northern Hemisphere for November 2025. The forecasts are from the 00Z 20 Oct 2025 CFS.
Which brings me to the final point of today’s discussion, the possibility of a stretched PV (and really how did I make it this far into the blog without bringing up my two favorite atmospheric features Ural blocking and a stretched PV; my impulse control is impressive). If you live east of the Rockies in North America, you couldn’t be faulted in believing that the Hudson Bay high pressure ridging is an atmospheric feature that has even preceded the most famous celestial permanent atmospheric anticyclone – Jupiter’s read spot. This is eerily reminiscent to me of a Canadian warming. And lo and behold many (but not all) PV forecasts resemble to me at least, a Canadian warming (see Figure v). And if you read last winter’s blog, then you surely remember that the smart money is on that a stretched PV will follow. So maybe a boring October will be followed by a more interesting November?
Near Term
This Week
The AO is predicted to be slightly negative this week (Figure 1) with mostly positive geopotential height anomalies currently across the Arctic and mixed geopotential height anomalies across the mid-latitudes of the NH (Figure 2). With predicted positive geopotential height anomalies across Greenland (Figure 2), the NAO is predicted to be negative this week.
Figure 1.
The predicted daily-mean AO at 1000 hPa from the 00Z 20 Oct 2025 GFS ensemble. Gray lines indicate the AO index from each individual ensemble member, with the ensemble mean AO index given by the red line with squares.
This week ridging/positive geopotential height anomalies centered on Greenland will support troughing/negative geopotential height anomalies across Northern Europe with more ridging/positive geopotential height anomalies across Southern Europe (Figure 2). This mostly zonal flow pattern will support widespread normal to above normal temperatures across Europe including the UK this period (Figure 3). This week ridging/positive geopotential height anomalies centered over Western Siberia forcing troughing/negative geopotential height anomalies across Eastern Siberia and East Asia (Figure 2). This pattern favors normal to above normal temperatures across much of Asia with normal to below normal temperatures across Eastern Siberia, East Asia and especially Mongolia this period (Figure 3).
Figure 2.
Forecasted average 500 mb geopotential heights (dam; contours) and geopotential height anomalies (m; shading) across the Northern Hemisphere from 21 Oct to 25 Oct 2025. The forecasts are from the 00Z 20 Oct 2025 GFS ensemble.
This week ridging/positive geopotential height anomalies will dominate North America centered on Eastern Canada with troughing/negative geopotential height anomalies across mostly limited to the Eastern US this week (Figure 2). This patten will favor widespread normal to above normal temperatures across the Alaska, Canada and the Western US with the biggest exception of normal to below normal temperatures in the Eastern US this week (Figure 3).
Figure 3.
Forecasted surface temperature anomalies (°C; shading) 21 Oct to 25 Oct 2025. The forecasts are from the 00Z 20 Oct 2025 GFS ensemble.
Troughing will support new rainfall across France, Scandinavia, Northeastern Siberia and parts of East Asia and India with otherwise mostly dry conditions widespread across Europe and Asia, with near normal precipitation across Afghanistan and Pakistan this week (Figure 4). Troughing will support new rainfall across the panhandle of Alaska, the West Coast fo Canada and Quebec with otherwise mostly dry conditions widespread across Canada and the US this week (Figure 4).
Figure 4.
Forecasted rainfall (mm/day; shading) from 21 Oct to 25 Oct 2025. The forecasts are from the 00Z 20 Oct 2025 GFS ensemble.
Near-Mid Term
Next Week
With geopotential height anomalies becoming mostly mixed across the Arctic and with mixed geopotential height anomalies across the mid-latitudes this period (Figure 5), the AO will likely remain close to neutral to possibly slightly positive this period (Figure 1). With predicted weak but positive pressure/geopotential height anomalies across Greenland (Figure 5), the NAO will likely be neutral to negative this period.
Figure 5.
Forecasted average 500 mb geopotential heights (dam; contours) and geopotential height anomalies (m; shading) across the Northern Hemisphere from 26 Oct to 30 Oct 2025. The forecasts are from the 00Z 20 Oct 2025 GFS ensemble.
Ridging/positive geopotential height anomalies centered south of Iceland and across the Iberian Peninsula will force troughing/negative geopotential height anomalies to the east across Europe (Figure 5). The pattern will support widespread normal to below normal temperatures across much of Europe including the UK with normal to above normal temperatures across far Eastern Europe and the Iberian Peninsula this period (Figure 6). Persistent ridging/positive geopotential height anomalies will dominate Asia centered near the Urals supporting troughing/negative geopotential height anomalies across Eastern Asia this period (Figure 5). This pattern favors widespread normal to above normal temperatures across much of Asia including Pakistan and Afghanistan with normal to below normal temperatures across limited to Eastern Siberia and parts of East Asia this period (Figure 6).
Figure 6.
Forecasted surface temperature anomalies (°C; shading) from 26 Oct to 30 Oct 2025. The forecasts are from the 00Z 20 Oct 2025 GFS ensemble.
Ridging/positive geopotential height anomalies are predicted to continue to dominate North America centered on Eastern Canada with troughing/negative geopotential height anomalies mostly limited to Western Canada and the Western US this period (Figure 5). This pattern will favor widespread normal to above normal temperatures across Alaska and much of Canada and the US from the Rockies eastward with normal to below normal temperatures limited to the West Coasts of Canada and the US (Figure 6).
Figure 7.
Forecasted precipitation rate (mm/day; shading) from 26 Oct to 30 Oct 2025. The forecasts are from the 00Z 20 Oct 2025 GFS ensemble.
Troughing will support new rainfall in Southeastern Europe, Japan and parts of Southeast Asia especially Bangladesh with otherwise mostly dry conditions widespread across Europe and Asia and near normal across Afghanistan and Pakistan this period (Figure 7). Troughing will support new rainfall across the Pacific Northwest and the Mississippi River valley with otherwise mostly dry conditions widespread across Canada and the US this period (Figure 7).
Mid Term
Week Two
With predicted mostly positive geopotential height anomalies across the Arctic, especially the North Atlantic sector, and mixed geopotential height anomalies across the mid-latitudes this period (Figure 8), the AO will likely remain near neutral to negative this period (Figure 1). With predicted weak but positive pressure/geopotential height anomalies across Greenland (Figure 8), the NAO will likely be negative this period.
Figure 8.
Forecasted average 500 mb geopotential heights (dam; contours) and geopotential height anomalies (m; shading) across the Northern Hemisphere from 31 Oct to 04 Nov 2025. The forecasts are from the 00Z 20 Oct 2025 GFS ensemble.
Ridging/positive geopotential height anomalies across the North Atlantic sector of the Arctic are predicted to support troughing/negative geopotential height anomalies across Northern Europe extending southeastward into Southeastern Europe with more ridging/positive geopotential height anomalies across the Iberian Peninsula this period (Figure 8). This mostly zonal flow pattern should favor widespread normal to above normal temperatures across Europe including the UK with normal to below temperatures mostly limited to Southeastern Europe this period (Figures 9). Persistent ridging/positive geopotential height anomalies will remain centered over the Urals with troughing/negative geopotential height anomalies across Eastern Asia with more ridging across Southern and Eastern Asia this period (Figure 8). The predicted pattern favors widespread normal to above normal temperatures across much of Asia including Pakistan and Afghanistan with normal to below normal temperatures limited to parts of Eastern Siberia and East Asia this period (Figure 9).
Figure 9.
Forecasted surface temperature anomalies (°C; shading) from 31 Oct to 04 Nov 2025. The forecasts are from the 00Z 20 Oct 2025 GFS ensemble.
Ridging/positive geopotential height anomalies are predicted to continue to dominate North America centered on Eastern Canada with troughing/negative geopotential height anomalies across western North America and ridging/positive geopotential height anomalies across eastern North America this period (Figure 8). This pattern supports widespread normal to above normal temperatures across most of Canada and the Western US with normal to below normal temperatures mostly limited to the West Coast of Canada and the Eastern US this period (Figure 9).
Figure 10.
Forecasted precipitation rate (mm/day; shading) from 31 Oct to 04 Nov 2025. The forecasts are from the 00Z 20 Oct 2025 GFS ensemble.
Troughing will support new rainfall across the UK and parts of Southeast Asia especially Bangladesh with otherwise mostly dry conditions widespread across Europe and Asia and near normal precipitation in Afghanistan and Pakistan this period (Figure 10). Troughing will support new rainfall along the coastal mountains of Western Canada and the Northeastern US with otherwise mostly dry conditions widespread across Canada and the US this period (Figure 10).
Longer Term
30-day
The latest plot of the polar cap geopotential height anomalies (PCHs) currently shows cold/negative PCHs in the upper stratosphere with warm/positive PCHs in the lower stratosphere and the troposphere (Figure 11). The cold/negative PCHs are predicted to persist in the upper stratosphere over the next two weeks and possibly include the and lower troposphere next week with warm/positive PCHs in the lower stratosphere and the mid to upper troposphere.
Figure 11.
Observed and predicted daily polar cap height (i.e., area-averaged geopotential heights poleward of 60°N) standardized anomalies. The forecast is from the 00Z 20 Oct 2025 GFS ensemble.
The predicted warm/positive PCHs in the lower troposphere this week (Figure 11) are consistent with the predicted negative to near neutral surface AO this week (Figure 1). Then next week with the forecast of the PCHs possibly flipping cold/negative in the lower troposphere (Figure 11) could favor a positive surface AO bias next week (Figure 1).
Figure 12.
Forecasted average 500 mb geopotential heights (dam; contours) and geopotential height anomalies (m; shading) across the Northern Hemisphere for November 2025. The forecasts are from the 00Z 20 Oct 2025 CFS.
I include in this week’s blog the monthly 500 hPa geopotential heights (Figure 12) and surface temperatures for November (Figure 13) from the Climate Forecast System (CFS; the plots represent yesterday’s four ensemble members). The forecast for the troposphere is ridging centered over Hudson Bay, Europe, East Asia and the Gulf of Alaska with troughing across Greenland to Iceland, the Urals, Eastern Siberia, the Dateline and the Canadian Maritimes (Figure 12). This pattern favors seasonable to relatively warm temperatures across Europe, much of Asia, the Tibetan Plateau, Pakistan, Afghanistan, Alaska, much of Canada and the Central US with seasonable to relatively cool temperatures across Northern Siberia, Northeast Asia, India, the West Coasts of Canada and the US and the Eastern US (Figure 13).
Figure 13.
Forecasted average surface temperature anomalies (°C; shading) across the Northern Hemisphere for November 2025. The forecasts are from the CFS 00Z 20 Oct 2025.
Boundary Forcings
SSTs/El Niño/Southern Oscillation
Equatorial Pacific sea surface temperatures (SSTs) anomalies are slightly below normal, along the equatorial Pacific (Figure 14) consistent with neutral conditions but suggestive that La Niña could return once again this winter and current forecasts show large spread and plenty of uncertainty but now mostly favor a return of La Niña conditions this winter. Observed SSTs across the NH remain well above normal especially in the North Pacific and much of the North Atlantic with the exception east of the Canadian Maritimes and extending south of Iceland, though below normal SSTs exist regionally especially in the South Pacific. The “warm blob” in the Northeastern North Pacific has received some hype this fall. Warranted or not, something that I am watching as well.
Figure 14.
The latest daily-mean global SST anomalies (ending 20 Oct 2025). Data from NOAA OI High-Resolution dataset.
Madden Julian Oscillation
Currently the Madden Julian Oscillation (MJO) is in phase three (Figure 15) and the forecast are for the MJO to quickly move into phases four and five and then weaken to where no phase is favored (Figure 15). Phases three through five favor ridging in western North America and troughing in Eastern Canada. Therefore, it seems hard for me to see that the MJO is having much influence on North American weather the next two weeks. But admittedly this is outside of my expertise.
Figure 15.
Past and forecast values of the MJO index. Forecast values from the 00Z 20 Oct 2025 ECMWF model. Yellow lines indicate individual ensemble-member forecasts, with the green line showing the ensemble-mean. A measure of the model 'spread' is denoted by the gray shading. Sector numbers indicate the phase of the MJO, with geographical labels indicating where anomalous convection occurs during that phase. Image source https://www.cpc.ncep.noaa.gov/products/precip/CWlink/MJO/CLIVAR/ecmf.shtml
