mariuscomper.uk Română

The polar vortex · updated 6 November 2026

In early November the polar vortex almost stopped. What follows in Bucharest?

The polar vortex is a ring of westerly wind that spins around the pole in winter, about 30 kilometres up. Averaged all the way round the latitude of Oslo, its wind fell from 83 km/h on 24 October to 0.3 km/h on 5 November: the weakest start to November since 1958. When the ring stops and reverses, the air above the pole warms suddenly by tens of degrees. Headlines say such a stratospheric warming brings a freeze.

In short: in Bucharest, the weeks after a stratospheric warming were no colder than usual. And across 65 winters, a weak early-November vortex did not foretell a colder winter. A tentative, unproven signal shows only in northern Europe.

0.3 km/h

the vortex wind on 5 November 2026, from the American GFS model; the European, Canadian and British models: 0.8–2.1 km/h

45 km/h

the weakest 5 November before this one, in 1960 (NCEP reconstruction from measurements, 1958–2025)

−0.04 °C

Bucharest, the average departure from normal after the 42 warmings, over the 51 days starting 10 days after each

Below: what it looks like from above, what followed in Bucharest after the 42 warmings from 1958 to 2023, and what can be said in advance about the coming weeks.

The vortex, seen from above

The picture shows the northern hemisphere from above, about 30 kilometres up, where air pressure is 10 hectopascals. Colour is temperature: violet and blue below −70 °C, yellow and red above −40 °C. The thin lines join points where that layer sits at the same height. Where they crowd together the wind is strong, and the threads of light show where it would carry the air.

On 24 October, a ring of wind averaging 83 km/h held a core of cold air over the pole. On 26 October it was −67 °C there.

1–5 November: pushed off the pole

Large waves of air rising from the lower atmosphere pushed the vortex off the pole and stretched it into two lobes: one over western Siberia, the other over the Labrador Sea, which by 5 November had become the deeper. Above the pole, sinking air is compressed and warms: by 5 November it had reached −55 °C.

The average wind along the 60° parallel was down to 0.3 km/h, because the wind circling the lobes blows in opposite directions and cancels out on average. On the edges of the lobes it still blows hard, in places at more than 150 km/h.

What the forecasts of 6 November show

GFS and ECMWF show the average wind dropping just below zero on 9 November, and all four models, American, European, Canadian and British, on 11 November: along the 60° parallel the average wind would blow from the east. That is how a major stratospheric warming is defined. In the NCEP reconstruction it would be the earliest since 1958, 16–18 days ahead of the one on 27 November 1968.

In the GFS forecast the cold core then moves over Scandinavia and northern Russia, and by 20 November the wind would be back to 41 km/h. From day to day the models differ from each other and from the reconstruction by typically 6–9 km/h, sometimes tens, so the exact date will be known only once it has passed.

February 2018, from above

This is what a major warming looked like in 2018: the vortex split in two, and the air above the pole went from −69 °C to −31 °C in 14 days (4–18 February 2018). Days 10–60 in Bucharest averaged 2.3 °C below normal, one of the five coldest of the 42 cases.

January 1977: another shape, another outcome

In January 1977 the air above the pole went from −65 °C to −46 °C (2–5 January 1977). The vortex was pushed aside, not split. On NOAA’s list, using the ERA5 reconstruction, it is a major warming; in the NCEP reconstruction the picture is drawn from, the average wind did not drop below zero. In Bucharest, days 10–60 were 3.4 °C above normal, the opposite of 2018.

68 winters at 30 kilometres

Each grey line is one winter, from 1958–59 to 2025–26: the mean wind along the 60° north parallel at 10 hPa, day by day, from 1 October to 30 April. In autumn the wind strengthens. In some winters it then drops suddenly below zero, and the dots on the zero line mark the first reversal of each winter. The orange line is 2026.

From 2 to 5 November 2026, every day was the weakest yet for its date, by more than the usual gap between GFS and the reconstruction (31 October and 1 November also came in below the old record, but by less than that gap). The 1–5 November average was 11.9 km/h; the weakest before it was 48 km/h, in 1960.

The polar vortex wind in every winter since 1958, and in 2026Mean wind at 60° north and 10 hPa, in km/h, from 1 October to 30 April: one grey line for each of the 68 winters from 1958–59 to 2025–26 and the band holding 8 of 10 winters in 1991–2020. The 2026 line falls from 83 km/h on 24 October to 0.3 km/h on 5 November, below every other winter, and the forecast takes it below zero. −100−50050100150200250OctNovDecJanFebMarApr 8 in 10 winters, 1991–2020 below zero the wind reverses: a major warming 30 Nov 195816 Jan 196023 Mar 19658 Dec 196527 Nov 19682 Jan 197017 Jan 19712 Feb 197322 Feb 197929 Feb 19804 Dec 198124 Feb 19842 Jan 198523 Jan 19878 Dec 198722 Feb 198915 Dec 199820 Mar 200011 Feb 20012 Jan 200218 Jan 20037 Jan 200421 Jan 200624 Feb 200722 Feb 200824 Jan 20099 Feb 20107 Jan 201312 Feb 20182 Jan 20195 Jan 202116 Feb 202317 Jan 20244 Mar 2026 earliest: 27 Nov 1968 5 November: 0.3 km/h 2026 km/h, westerly wind at 60° north, 30 km
Solid line: GFS analysis to 5 November. Dashed: the GFS forecast of 6 November; the band around it: the four models. Winter 2025–26 ends on 17 March 2026.

The grey lines come from the NCEP reconstruction, built from all available measurements; the data used here stop on 17 March 2026. The 2026 line comes from the American GFS model’s analysis, which can differ from the reconstruction by a few km/h from day to day. On 5 November, even the highest of the four models (2.1 km/h) is 43 km/h below the old record.

NOAA’s list of warmings ends in February 2023. The same rule, applied to the wind, finds two more since then, on 17 January 2024 and 4 March 2024, and a reversal on 4 March 2026 that the data, which end on 17 March 2026, cannot yet confirm. The earliest so far was on 27 November 1968, in the NCEP reconstruction. NOAA’s list, which uses the ERA5 reconstruction, puts the same warming on 28 November 1968 and has no November 1958 event.

The 42 stratospheric warmings and the days after, in BucharestOne row for each stratospheric warming on NOAA’s list, from 1958 to 2023. Each square is one day in Bucharest, from day 10 before the warming to day 60 after it, coloured by how far the temperature was from normal, from dark blue (much colder) to dark red (much warmer). The average of days 10–60 is written at the right of each row. Days from the warming −100+10+30+60 Feb 1958−0.8Jan 1960+0.3Jan 1963−2.5Dec 1965+1.6Feb 1966+1.7Jan 1968+0.6Nov 1968−1.8Mar 1969−0.6Jan 1970−0.6Jan 1971+0.5Mar 1971+0.5Jan 1973−0.3Jan 1977+3.4Feb 1979+0.5Feb 1980−1.7Mar 1981−0.2Dec 1981+0.9Feb 1984−1.6Jan 1985−7.9Jan 1987−4.4Dec 1987+1.9Mar 1988−1.0Feb 1989+3.2Dec 1998−0.2Feb 1999+0.2Mar 2000+1.9Feb 2001+1.6Dec 2001+3.3Feb 2002+1.3Jan 2003−3.9Jan 2004−0.1Jan 2006−1.7Feb 2007+0.9Feb 2008+2.2Jan 2009+0.5Feb 20100.0Mar 2010−0.4Jan 2013+1.2Feb 2018−2.3Jan 2019+0.8Jan 2021+0.9Feb 2023+0.1 The 51 days measured Mean, °C −10 °C0+10 °Cagainst normal

Rows are in chronological order. Pick a row to see its figures.

What it means for the polar vortex to break

Sometimes the ring of wind around the pole weakens abruptly, stops and reverses, and the air above the pole warms by tens of degrees within days. Meteorologists call this a major sudden stratospheric warming. NOAA’s list, on the ERA5 reanalysis, holds 42 from 1958 to 2023, about six a decade.

The figure, a wall of squares, has one row for each. Each square is one day in Bucharest, from the tenth day before the warming to the sixtieth after it: blue when it was colder than normal, red when it was warmer. The box around days 10–60 is the window measured.

12 Feb 2018: the weeks after were cold

After the warming of 12 February 2018, the average temperature in Bucharest stayed 2.3 °C below normal over days 10–60 after it. The coldest week was 10.9 °C below normal. It is one of the five coldest of the 42 cases.

−100+100+10+30+60
Daily Bucharest temperature departure from normal, °C, by days from the warming. Box: days 10–60.

January 1985: the coldest case in the set

The warming of 1 January 1985 was followed by 51 days (days 10–60) averaging 7.9 °C below normal. The coldest week was 13.6 °C below normal, and the coldest day averaged −20.2 °C.

Next come January 1987 (−4.4 °C) and January 2003 (−3.9 °C). These three cases, with 1963 and 2018, are the only five that passed 2 °C below normal.

−100+100+10+30+60
Daily Bucharest temperature departure from normal, °C, by days from the warming. Box: days 10–60.

The same kind of warming, mild weeks

After the warming of 9 January 1977, Bucharest averaged 3.4 °C above normal over days 10–60. The warmings of December 2001 (+3.3 °C) and February 1989 (+3.2 °C) ended the same way.

All meet the same major-warming criterion, though the vortex’s shape differs from case to case. In Bucharest these 42 cases show no consistent cooling, and the data do not say why outcomes differ.

−100+100+10+30+60
Daily Bucharest temperature departure from normal, °C, by days from the warming. Box: days 10–60.

Sorted by outcome, there is no threshold

The rows have settled from the coldest case to the warmest. 18 were colder than normal and 24 were warmer. Five passed 2 °C below normal, four passed 2 °C above, and 33 stayed near normal.

In these data, a stratospheric warming was followed consistently neither by a freeze nor by mild weather. The two ends of the column are far apart.

What can be said in advance about the coming weeks

If the wind reverses, the 2026 warming becomes the 43rd row of the wall above. The rule was fixed on 6 November, before any of its days were counted. The warming’s date is the first day the GFS analysis, not the forecast, shows the wind below zero; for the other 42 the date comes from NOAA’s list. The temperature is calculated exactly as for the other 42: the average Bucharest departure over days 10–60 after it. If the wind has not reversed by 31 March 2027, no 43rd row is added.

43%

of the 42 cases were colder than normal in Bucharest over days 10–60

48%

after any day from 28 November to 24 March, same measure

28–30

of 42, in Oslo, Helsinki and Stockholm

2026Row 43: empty on 6 November 2026; the days counted would start on 19 November 2026.

The 42 are the only guide. Their average is −0.04 °C. Half came out between −0.8 and +1.1 °C, and nine in ten between −3.8 and +3.1 °C. Of the 42, 18 (43%) were colder than normal. After any other day in the same part of the year the share is 48%. In other words, the coming 51 days have about the same chance of being colder than normal as any others.

Only six warmings came in November or December. After them, in Bucharest: December 1965 +1.6 °C, November 1968 −1.8 °C, December 1981 +0.9 °C, December 1987 +1.9 °C, December 1998 −0.2 °C, December 2001 +3.3 °C. Two of six were colder. They are too few to add anything.

In the north the numbers look different. In Oslo, Helsinki and Stockholm, 28 to 30 of the 42 cases were colder than normal, by about a degree on average. There history leans towards cold, but the signal does not pass the test across 20 cities (below).

If the wind reverses on 9 November, as the GFS forecast of 6 November shows, the days counted are 19 November – 8 January 2027, and the result can be calculated in January 2027. A single case can neither confirm nor overturn an average. Whatever it shows, it joins the other 42.

Does a weak vortex in November foretell a cold winter? In 65 winters, no

The second test was also fixed on 6 November, before the calculation. For each winter from 1958–59 to 2022–23 (the ERA5 temperatures used here stop in 2023), the vortex’s mean wind over 1–10 November was compared with Bucharest’s temperature in December, January and February.

The vortex wind in early November and the Bucharest winterOne dot for each winter from 1958–59 to 2022–23: across, the mean wind over 1–10 November in km/h; up, the Bucharest temperature departure in December–February. The dots follow no slope. The dotted line marks where 2026 would sit, at 6 km/h, outside the series.04080120160−6−4−20+2+4+6 2026 1958–1959: 60 km/h, +1.9 °C1959–1960: 74 km/h, +2.2 °C1960–1961: 39 km/h, +3.0 °C1961–1962: 66 km/h, 0.0 °C1962–1963: 75 km/h, −3.3 °C1963–1964: 60 km/h, −1.1 °C1964–1965: 62 km/h, +0.1 °C1965–1966: 87 km/h, +2.5 °C1966–1967: 83 km/h, −1.0 °C1967–1968: 104 km/h, +0.7 °C1968–1969: 78 km/h, −2.0 °C1969–1970: 100 km/h, +0.2 °C1970–1971: 79 km/h, +2.1 °C1971–1972: 95 km/h, +1.1 °C1972–1973: 77 km/h, +0.8 °C1973–1974: 89 km/h, +0.3 °C1974–1975: 73 km/h, +2.1 °C1975–1976: 108 km/h, −0.2 °C1976–1977: 89 km/h, +2.6 °C1977–1978: 65 km/h, −0.7 °C1978–1979: 100 km/h, −0.1 °C1979–1980: 64 km/h, −0.3 °C1980–1981: 83 km/h, −0.4 °C1981–1982: 95 km/h, +0.1 °C1982–1983: 110 km/h, +2.9 °C1983–1984: 110 km/h, +0.8 °C1984–1985: 106 km/h, −5.0 °C1985–1986: 79 km/h, +0.5 °C1986–1987: 105 km/h, −2.3 °C1987–1988: 78 km/h, +1.1 °C1988–1989: 96 km/h, +1.5 °C1989–1990: 87 km/h, 0.0 °C1990–1991: 121 km/h, −0.2 °C1991–1992: 97 km/h, −1.3 °C1992–1993: 95 km/h, −0.8 °C1993–1994: 69 km/h, +2.1 °C1994–1995: 108 km/h, +1.3 °C1995–1996: 102 km/h, −2.6 °C1996–1997: 65 km/h, −0.4 °C1997–1998: 109 km/h, +1.4 °C1998–1999: 73 km/h, −1.3 °C1999–2000: 101 km/h, −0.1 °C2000–2001: 94 km/h, +2.2 °C2001–2002: 117 km/h, +0.3 °C2002–2003: 116 km/h, −3.5 °C2003–2004: 93 km/h, −0.8 °C2004–2005: 58 km/h, +0.3 °C2005–2006: 58 km/h, −0.8 °C2006–2007: 80 km/h, +3.5 °C2007–2008: 86 km/h, −0.5 °C2008–2009: 74 km/h, +1.5 °C2009–2010: 64 km/h, −1.2 °C2010–2011: 78 km/h, −1.9 °C2011–2012: 105 km/h, −1.9 °C2012–2013: 114 km/h, −0.1 °C2013–2014: 102 km/h, −0.2 °C2014–2015: 88 km/h, +0.2 °C2015–2016: 121 km/h, +1.3 °C2016–2017: 51 km/h, −2.6 °C2017–2018: 92 km/h, +1.2 °C2018–2019: 97 km/h, −0.2 °C2019–2020: 122 km/h, +2.4 °C2020–2021: 107 km/h, +2.0 °C2021–2022: 82 km/h, +1.8 °C2022–2023: 109 km/h, +2.7 °C vortex wind, 1–10 November, km/hBucharest, Dec–Feb, °C against normal

There is practically no link: the correlation is 0.06 (0 means no link, 1 a perfect one). If the winters are shuffled at random, a value at least as large turns up in 62% of cases. The ten winters with the weakest early-November vortex were 0.4 °C colder than the rest on average, but the 95% interval runs from −1.4 to +0.7 °C. With E-OBS temperatures, which come from weather stations, the result is the same: 0.07 over the same 65 winters, 0.03 over 76 winters, 1950–51 to 2025–26 (the wind before 1958 is less reliable). The same calculation for the 20 European cities in the northern-Europe section finds the link in none of them.

The year with the weakest start to November until now, 1960, was followed in Bucharest by a winter 3.0 °C milder than normal. And 2026, at 6 km/h over 1–10 November (partly forecast), lies far below any year in the series, whose weakest was 39 km/h. No line drawn through past years can be stretched that far.

The average of the 42 cases: almost zero

The average of the 42 cases is −0.04 °C against normal, with a 95% uncertainty interval from −0.75 to +0.59 °C. The interval includes zero, so the 42 cases do not show whether the mean is below or above normal.

The comparison: 7,605 start dates spread over every year 1958–2022, between 28 November and 24 March, the season in which the warmings occur, each with the same 51-day average. 48% of them were colder than normal; after stratospheric warmings, 18 of 42 (43%). More than 2 °C below normal came out for 16% of the random dates and for 5 of 42 (12%) after a warming.

If 42 start dates are drawn at random, an average as far from the all-dates average appears in 85% of the tries. By this approximate comparison, the observed average is not unusual; it does not prove, though, that warmings have no effect at all.

The 42 cases after warmings, against randomly chosen datesTop: 42 dots, one for each warming, ordered by the 51-day average, from −7.9 to +3.4 °C. Bottom: the distribution of the same averages for 7,605 randomly chosen dates, mean +0.02 °C and standard deviation 2.0 °C. The two overlap almost completely. −10−8−6−4−20+2+4+6+8 42 cases after a warming (one dot = one case) 7,605 randomly chosen dates, same 51-day average average after warmings: −0.04 °C 51-day average, °C against normal colder

Where it shows more: northern Europe

The same calculation, repeated for 20 cities as an exploration. The coldest results are Oslo (−1.17 °C), Helsinki (−1.07 °C) and Stockholm (−1.04 °C). In each of them, 28 to 30 of the 42 cases were colder than normal, and the 95% interval stays below zero. But the search covered 20 cities, and after correcting for multiple comparisons (Holm) none clears the 5% threshold: adjusted p 0.08 for Stockholm, 0.10 for Oslo, 0.20 for Helsinki. The signal is suggestive, not established.

In Bucharest, 18 of 42 cases were colder than normal; in Cluj-Napoca 19, in Iași 17, in Timișoara 18, in Constanța 19. Temperatures in the five cities rise and fall together (correlation between 0.89 and 0.97), so they confirm practically one result, not five separate ones.

Two cautions. The three northern cities are neighbours, so they confirm one another only in part. And Istanbul comes out slightly warmer (+0.39 °C); across 20 cities, one in twenty is expected to look “significant” by chance alone, so it should not be read as an effect.

The 51-day average after the 42 warmings, by cityFor 20 cities, the average departure from normal over the 51 days after warmings, with the 95% interval. Oslo, Helsinki and Stockholm sit near −1 °C with intervals below zero. The Romanian cities sit at zero, with intervals that contain it.−2−10+1Oslo28/42Helsinki29/42Stockholm30/42Berlin23/42Moscow24/42Warsaw23/42Kyiv22/42London27/42Budapest21/42Iași17/42Madrid26/42Paris22/42Rome22/42Bucharest18/42Cluj-Napoca19/42Timișoara18/42Constanța19/42Sofia18/42Athens19/42Istanbul12/42average departure from normal, °C (line: 95% interval) · right: colder cases out of 42

And El Niño?

NOAA’s list records the El Niño phase of each winter. Of the 42 cases, 17 fell in El Niño winters, 16 in La Niña and 9 in neutral ones. In Bucharest, the average after a warming was −0.30 °C in El Niño winters, −0.09 °C in La Niña and +0.53 °C in neutral ones.

The uncertainty intervals are wide, over a degree and a half, and overlap. From these 42 cases it cannot be said that El Niño changes what follows a stratospheric warming in Bucharest. That does not show the effect is absent: 42 cases split into three groups are too few to see it.

Bucharest: average after a warming, by El Niño phaseThree groups with wide, overlapping 95% intervals that all contain zero.−2−10+1+2El Niño (17)10/17La Niña (16)5/16neutral (9)3/9average departure from normal, °C · right: colder cases

What to watch when you wonder whether a freeze is coming

A headline about the polar vortex is not a forecast for Bucharest. In these 42 cases, the average temperature on days 10–60 after a stratospheric warming was below normal in 43% of cases, nearly as often as after a random date (48%).

For the cold of the coming days, the right source is the official forecast of Romania’s national meteorological administration. The state of the stratosphere, day by day, is in the charts of NOAA’s Climate Prediction Center.

How it was calculated

The vortex wind

The mean westerly wind along the 60° north parallel at 10 hPa (about 30 km), from the NCEP/NCAR reanalysis: daily means on its 2.5° grid, from 1958 to 17 March 2026. For autumn 2026, the same mean over 36 points, one every 10° of longitude, from the GFS analysis and forecast, via Open-Meteo; IFS (ECMWF), GEM and the Met Office on the same points for comparison. Over 318 days in two winters covered by both, GFS came out 0.5–1.1 km/h above the reconstruction on average; the daily difference was typically 6–9 km/h, and the largest 51 km/h, on 30 December 2023. Records are compared with the winters 1958–2025: before 1958, measurements at this height were sparse.

The warming rule

Charlton and Polvani (2007), as in NOAA’s list: the first day between 1 November and 31 March on which that wind turns negative; a new warming counts only after 20 consecutive days of positive wind; and one not followed, by 30 April, by 10 consecutive days of positive wind is the spring’s final warming and does not count. Applied to NCEP/NCAR, the rule recovers to the exact day all 39 NOAA events in that reanalysis’s column, and two after February 2023, plus a reversal on 4 March 2026 not yet confirmed.

The view from above

Height and temperature of the 10 hPa surface, day by day, north of 30°: NCEP/NCAR for past warmings (2.5° grid), GFS for 2026 (points every 5° of latitude and 15° of longitude, interpolated to the same grid). The threads of light follow the wind computed from the heights (the geostrophic wind), a good approximation at high latitudes. Each day in the picture is a 24-hour mean.

The temperature

ERA5 daily mean at 2 m, for the grid point nearest each city, via Open-Meteo. It is a reanalysis, not a weather station: values can differ from what a thermometer showed. For Bucharest, a second record comes from E-OBS (the mean of daily minimum and maximum); the two give the same winters (correlation 0.98).

The departure

Each day is compared with the 1991–2020 average for the same day of the year (smoothed over ±7 days). The long-term drift towards milder winters is then removed (0.34 °C per decade in Bucharest), so recent mild winters are not put down to the vortex.

The window

Each event’s result is the average departure over days 10–60 after the central date in NOAA’s list (ERA5 column). The roughly 60-day horizon comes from the literature (Baldwin and Dunkerton, 2001); starting at day 10 was chosen before the calculation. The table below shows what other windows give.

The uncertainty

The 42 warmings fall in 35 winter seasons; 7 seasons hold two warmings, with overlapping windows. The 95% intervals come from 10,000 resamples of whole seasons, not of single events. The comparison uses 7,605 start dates between 28 November and 24 March in 1958–2022; neighbouring windows overlap, so they are not independent, and the comparison shows the shape of the distribution rather than an exact test.

Tests fixed in advance

On 6 November 2026, before the calculation, the rule for row 43 and the November test were written down: the Spearman correlation between the 1–10 November wind and the December–February departure, p from 10,000 permutations, Holm correction across the 20 cities, Bucharest as the main test.

The limits

42 cases are few. A 51-day average hides short cold snaps: the “coldest week” column in the table shows them. Precipitation, blizzards and snow cover were not measured. The five Romanian cities give similar results but remain five grid points in one region.

WindowMean, °C95% intervalColder
days 10–60 (the main window)−0.04−0.75 … +0.5918/42
days 0–60+0.01−0.69 … +0.6119/42
days 0–30+0.36−0.37 … +1.0416/42
days 10–40+0.37−0.33 … +1.0015/42
days 20–60−0.09−0.81 … +0.5421/42
days 30–60−0.30−1.08 … +0.3722/42

In Bucharest, no window’s 95% interval excludes zero.

All 42 warmings and what followed in Bucharest

NOAA’s list, ERA5 column. Departures from the 1991–2020 normal after removing the trend. A week is a seven-day average within the 51-day window. After February 2023 the same rule applied to the NCEP/NCAR wind finds two more warmings (17 Jan 2024, 4 Mar 2024) and an unconfirmed reversal (4 Mar 2026).

Central dateEl NiñoMean of days 10–60, °CColdest week, °CWarmest week, °C
8 Feb 1958El Niño−0.8−4.5+3.1
17 Jan 1960neutral+0.3−7.0+5.8
27 Jan 1963neutral−2.5−10.0+2.3
16 Dec 1965El Niño+1.6−2.2+6.5
22 Feb 1966El Niño+1.7−1.8+4.2
7 Jan 1968La Niña+0.6−3.9+5.5
28 Nov 1968El Niño−1.8−7.9+4.8
13 Mar 1969El Niño−0.6−5.0+2.6
2 Jan 1970El Niño−0.6−6.0+4.7
18 Jan 1971La Niña+0.5−5.1+6.0
20 Mar 1971La Niña+0.5−2.9+2.8
31 Jan 1973El Niño−0.3−4.9+5.0
9 Jan 1977El Niño+3.4−0.6+8.0
22 Feb 1979neutral+0.5−3.3+5.5
29 Feb 1980El Niño−1.7−4.5+2.1
4 Mar 1981neutral−0.2−4.6+3.6
4 Dec 1981neutral+0.9−3.0+6.6
24 Feb 1984La Niña−1.6−5.6+2.1
1 Jan 1985La Niña−7.9−13.6+0.1
23 Jan 1987El Niño−4.4−11.1+1.4
8 Dec 1987El Niño+1.9−3.1+6.3
14 Mar 1988El Niño−1.0−5.4+1.4
21 Feb 1989La Niña+3.2+1.3+5.7
15 Dec 1998La Niña−0.2−8.4+4.1
26 Feb 1999La Niña+0.2−3.6+3.2
20 Mar 2000La Niña+1.9−1.7+5.5
11 Feb 2001La Niña+1.6−3.5+6.2
30 Dec 2001neutral+3.3−4.5+9.5
17 Feb 2002neutral+1.3−5.2+7.8
18 Jan 2003El Niño−3.9−9.7−0.3
5 Jan 2004neutral−0.1−4.3+7.4
21 Jan 2006La Niña−1.7−4.5+2.7
24 Feb 2007El Niño+0.9−1.5+4.3
22 Feb 2008La Niña+2.2−0.6+4.2
24 Jan 2009La Niña+0.5−2.5+6.8
9 Feb 2010El Niño0.0−7.2+3.9
24 Mar 2010El Niño−0.4−4.1+1.4
6 Jan 2013neutral+1.2−1.3+4.2
12 Feb 2018La Niña−2.3−10.9+4.0
1 Jan 2019El Niño+0.8−2.7+5.2
5 Jan 2021La Niña+0.9−4.1+5.2
16 Feb 2023La Niña+0.1−4.1+3.6

Data for the 42 events (JSON) The vortex wind, 1958–2026 (JSON)

Sources

The wind and the views from above come from the NCEP/NCAR reanalysis and the forecast models; the list of warmings from NOAA; the city temperatures from ERA5. The calculation can be redone from them.

  • NOAA PSL · Kalnay et al., 1996

    NCEP/NCAR Reanalysis 1

    Wind, height and temperature at 10 hPa, daily means, used here from 1958 to 17 March 2026. The source of the 68 winters and of the views from above of past warmings.

    Open the source
  • Open-Meteo

    Forecast API and Historical Weather API

    Analysis and forecast at 10 hPa from the GFS (NOAA), IFS (ECMWF), GEM (Canada) and Met Office models for autumn 2026; ERA5 daily temperature for the cities.

    Open the source
  • NOAA CSL

    Sudden Stratospheric Warming Compendium: major events

    The list of major warmings by reanalysis, with the El Niño phase of each winter. The wall uses the ERA5 column; the rule applied to the wind is checked against the NCEP-NCAR column.

    Open the source
  • Butler et al., 2017

    A sudden stratospheric warming compendium

    The paper that describes the NOAA list and its criterion.

    Open the source
  • Charlton and Polvani, 2007

    A new look at stratospheric sudden warmings, part I

    The rule: the zonal-mean wind at 10 hPa and 60° north turns easterly in winter, then returns.

    Open the source
  • Baldwin and Dunkerton, 2001

    Stratospheric harbingers of anomalous weather regimes

    Average effects are followed for about 60 days after an event begins; hence the 60-day window.

    Open the source
  • Domeisen and Butler, 2020

    Stratospheric drivers of extreme events at the Earth’s surface

    A synthesis of how the stratosphere connects to cold air outbreaks.

    Open the source
  • Hersbach et al., 2020

    The ERA5 global reanalysis

    The reanalysis behind the city temperatures: a model fitted to observations on a grid, from 1940 to today.

    Open the source
  • Cornes et al., 2018

    E-OBS, daily station temperatures on a grid

    The second Bucharest record in the November test, independent of ERA5.

    Open the source
  • Natural Earth

    Land outlines, 1:110 million

    The coastlines drawn under the vortex.

    Open the source
  • NOAA CPC

    Stratospheric monitoring

    Daily charts of stratospheric wind and temperature.

    Open the source
  • ANM

    Official forecast for Romania

    Romania’s national meteorological administration.

    Open the source