Must Read: Arctic Sea Ice Death Spiral And Cold Weather

The media are debating if the decrease in  Arctic ice  is related to this winter’s cold weather in Germany. This post discusses the most recent current research about this including the most important figures from relevant studies.

Translated from an article by Stefan Rahmstorf [] are translation notes via Rabett Run

First, what does the unusual temperature distribution observed this March actually look like? Here is a map showing the data (up to and including March 25, NCEP / NCAR data plotted with KNMI Climate Explorer):

Freezing cold in Siberia, reaching across northwestern Europe, unusually mild temperatures over the Labrador Sea and parts of Greenland and a cold band diagonally across North America, from Alaska to Florida. Averaged over the northern hemisphere the anomaly disappears – the average is close to the long-term average. Of course, the distribution of hot and cold is related to atmospheric circulation, and thus the air pressure distribution. The air pressure anomaly looks like this:
There was unusually high air pressure between Scandinavia and Greenland. Since circulation around a high is clockwise [anticyclone], this explains the influx of arctic cold air in Europe and the warm Labrador Sea.

Arctic sea ice

Let us now discuss the Arctic sea ice.  The summer minimum in September set a new record low, but also at the recent winter maximum there was unusually little ice (ranking 6th lowest – the ten years with the lowest ice extent were all in the last decade). The ice cover in the Barents sea was particularly low this winter.  All in all until March the deficit was  about the size of Germany compared  to the long-term average.

Is there a connection with the winter weather?  Does the shrinking ice cover influence the atmospheric circulation, because the open ocean strongly heats the Arctic atmosphere from below?  (The water is much warmer than the overlying cold polar air.) Did the resulting evaporation of sea water moisten the air and thus lead to more snow? These questions have been investigated by several studies in recent years.

Petoukhov and Semenov, JGR 2010

These two scientists from the Potsdam Institute and the Leibniz Institute of Marine Sciences in Kiel used the ECHAM5 atmospheric model.  As a boundary ondition in a series of simulations they reduced ice cover in the Barents and the adjacent Kara Seas. With a medium sized reduction in ice volume similar to what we now see they calculated the anomaly in the pressure distribution shown below (they also examined what would happen for a complete loss):

Thus, abnormally high air pressure between Scandinavia and Greenland. Leading to the following temperature anomaly:
Cold from Siberia to Western Europe and a heat bubble centered on Labrador – quite similar to the temperatures this March.

Jaiser et al., Tellus 2012

Our colleagues from the Alfred Wegener Institute, together with U.S. researchers, examined the question of whether one finds this relationship in observational data. For this purpose, they investigated the correlation between the distribution of air pressure and ice cover in an analysis of their covariance. They concluded:

Our analysis suggests that Arctic sea ice concentration changes exert a remote impact on the large-scale atmospheric circulation during winter

The following graph from this work shows how the distribution of air pressure in winter changes between phases of low and high fall-ice cover in the Arctic:

You can see that low ice cover correlates with high pressure in the Arctic winter, again between Scandinavia and Greenland – just as in the model calculations of Petoukhov and Semenov. (Jaiser et al do not show temperature maps)

Liu et al., PNAS 2012

Next on the dance card is a study which appeared shortly thereafter by researchers from the U.S. and China, which approached the problem from both sides: model simulations and data analysis. The following map shows the linear regression between less ice cover in autumn and the temperature in winter:

Again, these temperature anomalies are similar to the two shown above, the temperature map for this March and that from the model of Petoukhov and Semenov.

An observed correlation (as in this diagram and in Jaiser et al.) is still not a causal connection, therefore Liu et al. also made a series of model calculations, but with a different model (the American model from NCAR). Further, they calculated a larger ensemble (20 runs with varying initial conditions), which makes the results statistically more robust. As a boundary condition in the model they took the observed reduction of the ice cover in the Arctic and calculated ensembles with or without the ice decrement. The result of these simulations for the temperature is as follows:

Again cold in Siberia and Europe, warmer over the Labrador Sea, and a cold band across North America. From the model, we know that it’s not just a correlation, but a causal relationship: In the model setting the prescribed ice cover results in cold temperatures over Europe, just as in Petoukhov and Semenov. The associated air pressure maps are also similar.

Liu et al. also have dealt with whether alternatively the Arctic Oscillation (AO) or the North Atlantic Oscillation (NAO) could explain the temperature anomalies.  They showed that these natural oscillations lead to different spatial patterns and inter-annual fluctuations – they ruled out the AO and NAO as causes.

They also have examined the relationship with snow and come to the conclusion that additional open water in the Arctic humidifies the air, leading to increased snowfall.  The snowpack is important because even the strong March sun can hardly heat up snow-covered land; the white surface simply reflects the sun’s rays. According to Liu, et al.

We conclude that the recent decline of Arctic sea ice has played a critical role in the recent cold and snowy winters.

However, the taz quoted [German paper] yesterday the spokesman of the German Weather Service [DWD in German] as saying that if there was a direct relationship with the sea ice cover,  the entire winter would have to be very cold in Germany.  I think this trivial argument with which he would like to wipe from the table the climate research results shown above is pretty embarrassing for the DWD.  Of course open water in the Arctic  does not prevent stochastic weather variability.  There will always be warm and cold periods. In all these studies it comes down to changing probabilities in the prevailing weather patterns: Petoukhov and Semenov estimate that the probability of cold winter extremes could triple, that is even in the Abstract. One wonders whether the DWD representative has read the relevant studies at all – and if not, why he feels the urge to comment on them in the media. Unfortunately, it has a certain tradition that meteorologists dealing with weather, are not familiar with climate science.

In my view, the above studies provide strong evidence for a link between Arctic ice loss due to global warming, more frequent winter high pressure, especially over the Atlantic-European part of the Arctic, and an associated influx of cold air to Europe. As we have often seen in recent winters – for example in a spectacular way in the first half of February 2012.

Still this is still not settled science – the studies are relatively new and need to be discussed intensively in the professional community and confirmed by further research, or perhaps called into question. This is the normal process of scientific debate, through which at the end findings are distilled which are robust and widely accepted, such as the fact that our emissions of greenhouse gases warm the climate.

P.S. (29.3.) I see now that the climate lie detector has put its finger on the beautiful article in Welt Online, which a few days ago once again presented the proverbial “Russian scientists who predict an ice age” warning of a “freezing pig cycle”.  There is no better way of actually disproving that solar fluctuations are to blame, when even diehard “climate skeptics” can not find any more respectable arguments. This time, even sponsored by Gazprom (for real!).  Anyone interested in the subject of climate and sun: could read What role does the sun play? [Maybe Eli will translate some day for some carrots]


Jaiser R, Dethloff K, Handorf D, Rinke A, Cohen J (2012) Impact of sea ice cover changes on the Northern Hemisphere winter atmospheric circulation.Tellus Series A-Dynamic Meteorology and Oceanography 64th doi: 10.3402/tellusa.v64i0.11595

Liu JP, Curry JA, Wang HJ, Song MR, Horton RM (2012) Impact of declining Arctic sea ice on winter snowfall. Proceedings of the National Academy of Sciences of the United States of America 109 (11) :4074-4079th doi: 10.1073/pnas.1114910109

Petoukhov V, Semenov VA (2010) A link between reduced Barents-Kara sea ice and cold winter extremes over northern continents. Journal of Geophysical Research-Atmospheres 115th doi: 10.1029/2009jd013568

8 Responses to Must Read: Arctic Sea Ice Death Spiral And Cold Weather

  1. tomwys says:

    Siberian, Asian, and Eurasian snow-cover albedo eclipsed that of 2007-2008 this year, initiated with record open Arctic and attendant Lake/Ocean Effect snows on land surrounding the Arctic.

    Snowpack, covering Asia all the way to the Himalayas and Istanbul, reflected GigaWatts of sunlight at all wavelengths, most of which CO2 does NOT intercept and re-radiate back down to the surface.

    The resulting colder temperatures assured that moisture entering Eurasia from the North Sea, Baltic, Atlantic, and Mediterranean would fall mostly as snow, adding to albedo as winter ended. [snip]

  2. Dick Smith says:

    What a fine piece of work linking several threads of evidence. The great ones make it look effortless.

  3. Icarus says:

    Having found a causal relationship between declining Arctic sea ice cover and winter weather, the obvious next step is to run the models forward to predict what will happen as the ice continues to decline and the Arctic ocean becomes ice-free. Has this been done? I have a keen interest in this as a UK gardener… and in general, as someone who likes there to be food in the shops. The current weather is giving farmers a pretty hard time as I understand it – too much precipitation, followed by an unusually cold spring.

  4. nice translation Eli. There are earlier studies making similar linkages – everyone knew losing Arctic ice would affect NH but just didn’t know details of how.

    From an Arctic science conference June 2010:

  5. Neven says:

    Tom Wysmuller, when will the increase in snowfall in fall and winter start to linger in spring and even summer, and thus become a negative feedback for Arctic sea ice melt? Because there still aren’t any signs of that.

  6. Mulga Mumblebrain says:

    It is unlikely, surely, with the general rise in global temperatures soon overwhelming the short-term influence of Arctic sea ice disappearance. It seems like a denialist argument that there is some sort of negative feedback kicking in that will restore Arctic sea ice safely, and with no untoward repercussions.

  7. Artful Dodger says:

    Indeed, as snow is an excellent insulator, more reduces permafrost freezing, making increased snowfall a POSITIVE feedback.

    BTW, more snow also leads to thinner sea ice by the end of Winter.

  8. Sourabh says:

    Hey Guys,

    Thanks for an awesome post. I have been following climate science for a while now. I really appreciate work you guys are doing.

    I don’t know anything about meteorology. Therefore, I have been having trouble understanding how declining see ice causes high pressure over Arctic? I know that open ocean absorbs more heat. But, how does it affect jet-stream? Can you provide any link/info about it?