Insights & News

Last Winter’s Temperature Patterns Forebode Strong Arctic Sea Ice Melt

6th March 2020

Relatively speaking, it’s been a sweltering winter season for the Northern Hemisphere. Depending on which dataset you look at, the Dec-Feb averaged surface temperatures rank 1st or 2nd highest in recorded history.

The winter that it comes second to in some datasets is 2015-2016. Due to the combined effects of a very strong El Niño event in the tropical Pacific and an exceptionally low amount of sea ice cover in the Arctic Ocean*, that winter saw by far the highest globally averaged sea temperature on record (see below, left).

Plots of Dec-Feb averaged surface air temperature for the Northern Hemisphere, oversea (left) and overland (right), 1951-2020.

Atmospheric circulations drove some of the associated additional lower-atmospheric heat and moisture across the Arctic during the following six months. There, it really did a number on the sea ice, melting away ~77% of its area between March and September – the 3rd highest proportional loss on record (1979-2019).

The 2019-20 winter wasn’t as warm over the oceans, but overland it almost ties with the record-holder. Regionally, Europe and Asia had by far their warmest winters on record (see below).

Plots of Dec-Feb averaged surface temperatures for Europe (left) and Asia (right), 1951 to 2020

Are the cooler oceans likely to prevent this year’s Arctic sea ice (ASI) melting season (Apr-Sep) from being as severe as 2016’s was, or will the pattern of exceptional overland warmth be the dominant decider?

To explore this, I’ve sought out the most similar winters, by temperature patterns, within all the past years we have good quality ASI data for; 1979-2019. Below, you can see how the top ten most similar compare to 2019-20.

The similarity was calculated for 60 degrees latitude northward. It’s the warmth of land areas here, encircling the Arctic Ocean, which has the greatest potential influence on how much ASI melt occurs Apr-Sep. Notice that the strongest similarity is between Scandinavia and Siberia – this is important for reasons I explain later.

The similarity was calculated for 60 degrees latitude northward. It’s the warmth of land areas here, encircling the Arctic Ocean, which has the greatest potential influence on how much ASI melt occurs Apr-Sep. Notice that the strongest similarity is between Scandinavia and Siberia – this is important for reasons I explain later.

History Indicates:
High Chance of Top-Three Arctic Sea Ice Melt Season

Having identified the ten most and ten least similar years, it was then simply a matter of comparing the statistics for the following ASI melting seasons.

The results are more dramatic than I ever imagined. It’s worth taking a few moments to absorb all the information from the following two plots and summary table.

In these plots, the most similar ten years to 2020 (for Dec-Feb mean 925 mb temperature within 60-90°N) have orange markers and the least similar have blue markers. For these, the numbers above are the rank of that year relative to those before it (not the whole series). This is to reduce the sensitivity of the results to climate change.

In these plots, the most similar ten years to 2020 (for Dec-Feb mean 925 mb temperature within 60-90°N) have orange markers and the least similar have blue markers. For these, the numbers above are the rank of that year relative to those before it (not the whole series). This is to reduce the sensitivity of the results to climate change.

 

Right back through to the mid-1980s, there have been some markedly similar winters, the majority of which were followed by hard-hitting ASI melting seasons.

Eighty percent of them were in the top three, by extent loss, up to that point in history. By area loss, the figure is a little lower but still suggests a nearly a one-in-four chance that 2020 will stand on a very undesirable podium.

 

“My overall take from these results is that 2020 is likely to see very high sea ice losses in the peripheral regions of the Arctic Ocean, with the central region faring better – but still seeing a proportional loss within the top five years on record.”

 

These numbers stand out even more when considering the contrast with the least similar years.

In my experience, the ASI extent measure tends to be more sensitive to sea ice loss at the periphery of the Arctic, while area responds more to losses within the middle.

My overall take from these results is that 2020 is likely to see very high sea ice losses in the peripheral regions of the Arctic Ocean, with the central region faring better – but still seeing a proportional loss within the top five years on record.

How Does This Connection Work?
‘Thermal Inertia’ and Albedo Feedback…

To understand this connection between winter overland temps and the following ASI melting season, we need to think about snow and ice.

Winter snowfall is usually a common occurrence across many northern parts of Europe, Asia and North America. By January, wide areas of lying snow and ice are the expected norm.

This serves as a vast, highly reflective blanket across the land.

Mid-Nov through mid-Feb, this shiny status (high albedo) matters little – but late Feb through mid-Mar, the sun makes its return to the Arctic regions and the albedo starts to make a big difference.

Diagram illustrating how a snow or ice layer keeps the ground beneath colder by reflecting the sun's rays skyward.

The more snow cover there is, the more the sun’s rays are reflected skyward.

In this way, extensive snow cover can ‘self-help’ by keeping the air colder, reducing the rate of melting. This can delay the springtime warmup by many weeks, even a month or so in places.

Most Likely 2020:
Strong Assault on Sea Ice from the Eurasian Side

As of 3rd March 2020, there are large deficits of snow cover in northern Europe and western Russia (see below). More energy from the sun is being absorbed by the land, raising ground temperatures faster as spring gets going.

Map showing observed snow depth anomalies for the Northern Hemisphere for 3rd March 2020.

Source: https://ccin.ca/index.php/ccw/snow/current

At times in the coming weeks, it should still be cold enough for snowfall, but this will likely struggle to stay on the warmer than usual ground for long, especially as the sun shines down more strongly each new day.

During the following few months, the predominant west-east movement of weather systems across the region will periodically transport the anomalous warmth as far east as Siberia. Odds are the snow cover will melt away sooner than usual there.

That anomalous heat will then be in a prime position to move across the Eurasia-side peripheral Arctic sea ice from time to time, leading to strong melt there. Remember – this is the region with the most similarity to Dec-Feb 2020 in those top-ten most similar years. The hardest-hitting ASI melt seasons tend to be those that see a strong ‘assault’ from the Eurasian side.

There’s Still Hope

The damaging sequence of events I just outlined is the most likely outcome, not a guaranteed one. As the stats show – not all the most similar years saw extreme ASI loss.

Sometimes, external factors force weather patterns that can overcome the snow albedo feedback and lay down a good amount of snow during the first half of spring.

As of writing this, guidance tools aren’t suggesting such developments within the first few weeks of March 2020, largely due to an exceptionally strong polar vortex. This makes a lot of sense, so I can’t deny that I’m not feeling hopeful – but I’m not yet prepared to rule out a ‘revenge of the snow’ within the next five weeks.

 

James Peacock MSc
Head Meteorologist at MetSwift

 

* Open ocean releases heat into the atmosphere, while sea ice can block this transfer and keep the air much colder.

Speak to one of our experts and discover how our analytic platform can revolutionise your weather & NatCat insight. Request a meeting, a chat & demo, or a free trial. Developed via consultation & collaboration.