Midlatitudes unaffected by sea ice loss

Climate scientists cannot agree on what caused a recent spate of severe winters over North America and Eurasia. Now, a simple yet powerful physics-based approach makes it clear that record-low Arctic sea ice coverage was not the root cause.

Global warming and cold winters do not belong together — or do they? For years, climate scientists have struggled to explain a link in observations between record decreases in Arctic sea ice cover and record-cold winters over northern midlatitudes. The Arctic to midlatitude link could be explained by a cause-and-effect sequence as follows: sea ice loss transfers heat from the exposed ocean surface to the atmosphere, and this added energy drives a change in atmospheric circulation that carries cold air from the Arctic into the midlatitudes. A problem with this sequence is that climate models do not robustly support it. Writing in Nature Climate Change, Russell Blackport and colleagues1 put to rest the notion that Arctic sea ice loss caused the cold midlatitude winters, showing instead that atmospheric circulation changes preceded — and then simultaneously drove — both sea ice loss and midlatitude cooling. Atmospheric circulation was the culprit, not sea ice loss.

Since satellite observations became available in the late 1970s, the coverage of Arctic sea ice in winter has decreased by about half a million square kilometres per decade2. Riding on top of this trend are winter-to-winter fluctuations that some scientists argue have affected temperatures over North America and Eurasia3,4,5. Compare, for example, the evolution of sea ice cover averaged over the Barents–Kara Sea and surface air temperature averaged over Eurasia in winter (Fig. 1). (Here, linear trends have been removed to accentuate winter-to-winter fluctuations.) Large drops in sea ice coverage over the Barents–Kara Sea often coincide with unusually cold winters over Eurasia (consider the periods starting in 1982, 1999 and 2010). Observations similarly suggest a link between low sea ice coverage over the Chukchi–Bering Sea and cold conditions over North America. Given the striking and globally disproportionate changes occurring in the Arctic over recent decades, these links have captivated scientists, policymakers and journalists alike. As one prominent Arctic scientist told the New York Times (23 March 2018)6, “What happens in the Arctic doesn’t stay in the Arctic.”

Fig. 1: Periods of large sea ice loss in the Arctic coincident with large midlatitude cooling.

Time series of winter average (December, January and February) sea ice extent averaged over the Barents–Kara Sea (black curve; left axis) and surface air temperature averaged over Eurasia (blue curve; right axis). Linear trends have been removed to accentuate winter-to-winter fluctuations. The grey shadings and thicker lines illustrate periods within which large Arctic sea ice loss is coincident with large midlatitude cooling.

But does the proposed connection between Arctic sea ice decline and cold midlatitude winters stand up to scrutiny? One way to answer this question is to impose observed sea ice loss corresponding to a particular winter, or over several winters, in a climate model and assess its impact on midlatitude temperature. Here, results have been disappointing. In one modelling study7, for example, weak Eurasian cooling was obtained in response to reduced sea ice over the Barents–Kara Sea; in other studies8,9, no cooling was found. The fact that climate models do not robustly support the sequence of Arctic-to-midlatitude cause-and-effect described above may suggest that the state of climate modelling is insufficiently advanced to address the problem. However, Blackport and colleagues, while recognizing that climate models have their shortcomings, argue that the problem is not with the climate models but with the interpretation of what is driving what.

The breakthrough that Blackport and colleagues make comes from a physics-based recognition that when the atmosphere is driving sea ice loss, anomalous heat is transported from the atmosphere to the surface, whereas when sea ice loss is driving the atmosphere, anomalous heat is transported from the surface to the atmosphere: incoming heat in the one case, outgoing heat in the other. When the authors considered winters since the late 1970s when sea ice loss was accompanied by incoming heat — the atmosphere driving sea ice loss — they found midlatitude cooling. On the other hand, in winters when sea ice loss was accompanied by outgoing heat — sea ice loss driving the atmosphere — midlatitude cooling was absent. Evidently, atmospheric circulation, not sea ice loss, is the call, and midlatitude cooling the response. That these relationships are seen in observations and two independent climate models gives Blackport and colleagues confidence in both their physics-based analysis and their climate model simulations. To further bolster their case, the authors considered conditions in the month before, during and after cold midlatitude winters. If sea ice loss were in the driver’s seat, then it would lead and midlatitude cooling would lag; but such is not the case in either the observations or their model simulations. Blackport and colleagues conclude their analysis by showing that the sea ice loss in the near future is also unlikely to lead to winter cooling.

In my opinion, the evidence presented by Blackport and colleagues1 brings the case to a close. Midlatitude cooling in winter is not caused by Arctic sea ice loss. Rather, it is a side effect of regional circulation changes that precede and then simultaneously drive Arctic sea ice loss and midlatitude cooling. A fair question remaining to be answered is: what is behind these cooling-favourable circulation changes? Are they chance manifestations of a chaotic atmosphere, or are they due to some other remote forcing? It has been suggested10 that west tropical Pacific warming has influenced midlatitude circulation and increased the incidence of cold winters over North America. So far, climate models have been unable to show this11. Perhaps some of the deep physics-based thinking of Blackport and colleagues1 would be helpful here as well.


  1. 1.

    Blackport, R., Screen, J. A., van der Wiel, K. & Bintanja, R. Nat. Clim. Change, (2019).

  2. 2.

    Vaughan, D. G. et al. in Climate Change 2013: The Physical Science Basis (eds Stocker, T. F. et al.) Ch. 4 (Cambridge Univ. Press, 2013).

  3. 3.

    Inoue, J., Hori, M. E. & Takaya, K. J. Clim. 25, 2561–2568 (2012).

    Article  Google Scholar 

  4. 4.

    Tang, Q., Zhang, X., Yang, X. & Francis, J. A. Environ. Res. Lett. 8, 014036 (2013).

  5. 5.

    Kug, J.-S. et al. Nat. Geosci. 8, 759–762 (2015).

    CAS  Article  Google Scholar 

  6. 6.

    Pierre-Louis, K., Popovich, N. & Pearce, A. New York Times, (2018).

  7. 7.

    Mori, M., Wantabe, M., Shiogama, H., Inoue, J. & Kimoto, M. Nat. Geosci. 7, 869–873 (2014).

    CAS  Article  Google Scholar 

  8. 8.

    McCusker, K. E., Fyfe, J. C. & Sigmond, M. Nat. Geosci. 9, 838–842 (2016).

    CAS  Article  Google Scholar 

  9. 9.

    Ogawa, F. et al. Geophys. Res. Lett. 45, 3255–3263 (2018).

    Article  Google Scholar 

  10. 10.

    Palmer, T. Science 344, 803–804 (2014).

    CAS  Article  Google Scholar 

  11. 11.

    Sigmond, M. & Fyfe, J. C. Nat. Clim. Change 6, 970–974 (2016).

    Article  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to John C. Fyfe.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Fyfe, J.C. Midlatitudes unaffected by sea ice loss. Nat. Clim. Chang. 9, 649–650 (2019).

Download citation

Further reading


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing