Roger Andrews: The Cold War Years 1940 to 1942

Posted: January 18, 2013 by tallbloke in Analysis, atmosphere, climate, general circulation, Ocean dynamics, weather

THE COLD WAR YEARS – 1940, 1941 AND 1942

With all the recent interest in cold and snow I thought it might be interesting to revisit the three coldest years that North-Central and Eastern Europe experienced in the 20th century, which occurred one after the other in 1940, 1941 and 1942. The graph below shows the 20th century temperature records for three widely-separated cities in the affected region – Helsinki, Kiev and Vienna. The three successive cold years show up as an anomalous feature on all of them:


I briefly toyed with the idea that these cold years might have been caused by the war (blackouts, curfews, fuel rationing etc =  lower UHI impact) but it didn’t fit. Then with the graph below the cause became apparent. The jumbo-sized 1940-41 El Niño was the culprit:


I never went into the question of exactly how warm water on the surface in the Equatorial Pacific translated into freezing feet in Finland, but Brönniman (2005, here) has put a story together based on surface temperatures, radiosonde records, ozone concentrations, upper air circulation, polar vortices, stratospheric warming, planetary waves, sea level pressure, a coupled climate model (which doesn’t use any GHG forcing, so I guess that’s okay) and a few other things besides. I don’t know whether the story hangs together, but it seems to dovetail with some of the things people have been saying on other threads.

Here’s a figure from the report to give a foretaste:


And the paper isn’t of purely scientific interest. If the 1940-41 El Niño really did cause the abnormally cold 1941-42 Russian winter it may well have altered the outcome of the war.

  1. colliemum says:

    “If the 1940-41 El Niño really did cause the abnormally cold 1941-42 Russian winter it may well have altered the outcome of the war.”
    As history buff, I can state that this El Nino actually did alter the outcome of the war. German troops got stuck, i.e. frozen, 30 km from Moscow. Had it been warmer, they might have overrun Moscow.
    It should also be noted that winter came very early in 1941 – with deep freeze and heavy snow starting in October already.

    This is extremely interesting, and it would be great to investigate how anomalies influenced historic events.

  2. Roger Andrews says:

    Chapter 3 of Al Gore’s “Earth in the Balance” lists a number of examples. Although it is the only Chapter in the book worth reading.

  3. colliemum says:

    @ Roger Andrews:
    Thanks for that information – when the snow is gone, I may venture to the local library and have a look.
    I won’t buy that book – Algore’s just got 100 Million $$$ from Big Oil, he doesn’t need my money.

  4. Scute says:

    Is there a ‘paleo’ record for El Niño conditions in 1812? Napoleon had the same problem.

    I always thought that Hitler was considered rash to underestimate the Russian winter. Perhaps not. It seems there was a large element of bad luck and that he was probably basing his prognostications on the winter climate of the past few years, although, I believe the intention was to achieve his goal before winter set in- Operation Barbarossa was slated to begin in May, delayed to June and suffered setbacks along the way.

  5. michael hart says:

    He didn’t have much fun in Stalingrad, either.

  6. Roger Andrews says:


    Two reconstructions at:

    Neither shows 1812 as an El Niño year. I guess Napoleon just had it coming.

  7. Greg Goodman says:

    Though the cold years in N. Europe are not in doubt I would be very cautious with SST (ie Nino 3.4 etc ) from that period.

    Once again I draw attention to the aberrations to the data at that time and the reliability of the Met. Office’s attempts to “correct” them:

    The real data shows a step change from one month to the next when U.S entered the war, another opposite drop when U.S Nave demobilised.

    Between ICOADS 2.4 that I analysed, and 2.5 that Hadley used for hadSST3 a lot of UK admiralty data was merged into ICOADS. Sadly this simply muddied the waters rather than fixing the problem.

    The steps are less clearly defined in 2.5 though clearly the error is still there. Rather than finding the cause and correcting it they have diluted it.

    So be careful not to put too much weight on what the mid Pacific did or did not do during the war years.

  8. Roger Andrews says:

    The global SST data for WWII are indeed heavily distorted by observational and other biases, but the Niño3.4 Index seems to be OK. The Pacific SST data remained largely unaffected by wartime events before Pearl Harbor in December 1941 and the El Niño was close to being over by then.

  9. tchannon says:

    This ought to be cross checkable via proxy even though there is supposed to be actual data.

    Does this help?

  10. Greg Goodman says:

    I’d be as worried about Hadley “bias corrections” in this period as any sampling bias caused by the war.

    Here is one check I just did for other reasons looking at Nino regions 1 and 2 (also comparing to icoads did not show any major differences , so that’s encouraging.)

    flat bit post was looks like missing data but it covers the years you are interested in here. The correlation is not that tight that this should be taken as proof of a problem around 1941 but it does look like that peak may be exaggerated.

  11. Greg Goodman says:

    comparing had ISST and icoads for el Nino 1+2 and Nino 3.4

    Nothing too anomalous. Looks like hadley homogenisation processing slightly truncated the 1939 and attenuated the drop in 1942.

    differences in LOD may be due to other factors, since it’s not a 1:1 fit.

    This was not a huge El Nino, so it seems unlikely it was the cause of the Russian winter , although it may have contributed.

  12. Greg Goodman says:

    Due to the synchronicity of the two events I think using the idea of El Nino “causing” anything that far away may be misplaced.

    It is probably more an indication of a mutual common cause. Cold north pacific tends to coincide with warmer tropics. This is the PDO (as opposed to the PDO index).

    Whatever causes these oscillations probably caused excessively cold N.P and N. Atl. as well as warmer tropical Pacific.

    I’m sure Vucevik would be talking about Bz by now 😉

  13. Roger Andrews says:


    “I’d be as worried about Hadley “bias corrections” in this period as any sampling bias caused by the war”.

    A few years ago I wrote a long expose of these bogus “bias corrections” It’s here if you want to wade through it.

    Most relevant to the present discussion is Figure 6, link below. The SST, Tair and cloud cover records are clearly heavily distorted by wartime biases. The ENSO Index, however, shows no obvious sign of distortion, so I think we can accept Niño3.4 as correct within normal limits of error over this period.

    “I think using the idea of El Nino “causing” anything that far away may be misplaced … It is probably more an indication of a mutual common cause.” Don’t think so. The match between temperatures in Vienna and the 1940-41 El Niño is too close to be coincidence, and Niño3.4 leads Vienna temperatures by four months as closely as I can figure it.

    And yeah, Vuk, where are you? Or you, Tallbloke, or Stephen Wilde?

  14. Adam Gallon says:

    Interestingly enough, I’ve just watched a programme on the Beeb about the 62-63 winter.
    The programme featured the programme broadcast at the time (Cliff Mitchelmore chaired it) and their “Weather Frog” showed a series of charts (ie big sheets of cardboard with the isobars etc on them!) the last one was the explanation for the extreme, prolonged, cold by “American scientists”, being warm water around the Hawaiian Islands.

  15. tallbloke says:

    Adam, they must have spent all their time on field trips. It would explain the shorts and loud flowery shirts.

  16. Interesting that there was a spike of CO2 in the northern hemisphere in 1940. see here or the partial translation of Kreutz’s paper on my higher temperature and solar radiation arriving at the surface leading to outgassing of CO2 from the ocean followed by shifts in the jetstream. Looking at some random weather stations in Australia it seems that the summer of 1939 was very hot (eg a max of 47.8C/118F at Cobar NSW in Jan) but that the summer of 1941 was relatively cool (eg Cobar max temps in Jan around 30C and none over 40C)

  17. Ulric Lyons says:

    1940-42 winters show a -NAO

    Though it is hard to see much drop in the monthly Ap index which is well curious. (?)

    On 1812:
    “1811 was an El Niño year that transitioned to a La Niña 1812. Such transitions are marked in Eastern Europe and Russia by bitterly cold winters,”

    And 1814 appears to have had an El Nino episode.

  18. Roger Andrews says:

    Ulric Lyons:

    Thanks for the data link. The NAO looks like it correlates at least as well with the cold years as Niño3.4 and might even lead it, which would be interesting. I’ll take another look tomorrow

  19. tallbloke says:

    Roger A: Take a look at Erl Happ’s work for NAO leading ENSO.

  20. Roger Andrews says:


    I took a look.

    Erl Happ actually bases most of his work on the Arctic Oscillation, not the NAO, but since both are calculated using the pressure differential between high and low NH latitudes they show basically the same thing.

    Anyway, I compared the NAO/AO with ENSO (Niño3.4) and found that sometimes NAO/AO leads ENSO, sometimes ENSO leads NAO/AO, sometimes the two are positively correlated, sometimes they’re negatively correlated and sometimes they show no correlation at all. So if there is any kind of causative link between NAO/AO and ENSO it’s pretty darned subtle.

    During 1940, 1941 and 1942, however, and for the only extended period in the 20th century, they all came together, with a strong El Niño coinciding almost exactly with a strongly negative NAO/AO (i.e. high pressure in the Arctic, low pressure farther south). There’s not much doubt that this persistent pressure differential was the direct cause of the abnormal cold in Europe in those years, but did the El Niño cause the pressure differential? Or was it just coincidence that the El Niño occurred when it did?

    Back to you, if you’re still there.