Arctic sea ice has reached its summer minimum extent for the year, clocking in at 4.59m square kilometres (sq km), which puts it joint sixth lowest in the 40-year satellite record alongside 2008 and 2010.

The twelve smallest summer lows in the satellite record have all occurred in the last twelve years.

The announcement, from the National Snow and Ice Data Center (NSIDC) in the US, is provisional as “changing winds or late-season melt could still reduce the Arctic ice extent”.

Melt season

As the northern hemisphere summer comes to an end, the annual retreat of Arctic sea ice slows to a stop before ice starts accumulating again. This point is marked each year as the summer minimum, and used as an indicator for the health of Arctic sea ice.

The 2018 summer low is the joint sixth smallest on record, beaten only by 2015 in fifth, 2011 in fourth, 2007 and 2016 in joint second, and 2012 in first with a record low of 3.39 sq km. The 2018 figure is 1.63m sq km below the average summer minimum for 1981-2010.

The chart below shows how 2018 (blue line) compares with the previous four years, the record summer of 2012 (brown dotted line), and the long-term average (grey line).

Arctic sea ice extent up to 23 September 2018, along with daily ice extent data for four previous years and the record low year (2012). The 1981-2010 median is in dark grey, and the grey areas around the median line show the interquartile and interdecile ranges.
Image: NSIDC

Satellite data show that Arctic sea ice hit its minimum on two dates this year – 19 and 23 September. The latter is one of the latest dates for the summer low in the satellite record and around 5-9 days later than the 1981-2010 average.

The NSIDC says the lateness was “at least partially caused by southerly winds from the East Siberian Sea, which brought warm air into the region and prevented ice from drifting or growing southward”.

While this year isn’t record-breaking, it is still cause for concern, says Prof Julienne Stroeve, professor of polar observation and modelling at University College London, senior research scientist at the NSIDC, and member of the Arctic Basecamp consortium. She tells Carbon Brief:

“Put simply, in the last ten years the Arctic is melting faster than it ever has previously since records began. We have lost over half of the summer sea ice coverage since the late 1970’s and it is realistic to expect an ice-free Arctic sea in summer in the next few decades.”

Cool and cloudy

The summer melt season kicked off at the end of March after reaching the second-lowest winter maximum on record. Sea ice extent remained low through spring, with the average extent in April tied with 2016 as the lowest on record and the average for May being the second lowest on record.

Spring saw some unusually warm temperatures in parts of the Arctic – temperatures for the month of May in Svalbard, for example, were 6C higher than average.

However, cooler and windier conditions from June onwards saw the pace of sea ice melt slow, despite short periods of rapid decline. This is reflected in the average extents for June, July and August, which clocked in as the fourth, ninth and seventh lowest on record, respectively.

The weather over the Arctic during the summer has been dominated by low pressure, explains Zack Labe, a University of California, Irvine PhD student studying sea ice. This meant a “tendency toward cooler and cloudier weather” over the summer months.

The Arctic Oscillation (AO) has also had an impact, says Labe. The AO is a natural climate pattern characterised by westerly winds circulating around the Arctic. When the AO is positive – as it predominantly has been from June to September – the winds tend to be stronger, holding cold air over the Arctic.

This “is generally a favourable weather pattern for less sea ice export and melt during the summer”, says Labe. It also contributed to an unusual summer for the Atlantic side of the Arctic, he adds:

“The movement of ice, due to the positive Arctic Oscillation, favoured less sea ice exporting into the Atlantic. This contributed to unusually low sea ice in this region, especially in the Greenland Sea. Also, sea ice around Svalbard has been near record lows for the last few months.”

The map below shows how sea ice on the 23 September (shaded white) compares to the average position at that time of the year (orange line).

 Arctic sea ice extent on 23 September 2018. The orange line shows the 1981-2010 average extent for that day.
Image: NSIDC


Simulations of Arctic sea ice thickness from the University of Washington’s PIOMAS model suggest that this summer’s average thickness is also close to the long-term declining trend, says Labe:

“Sea ice is much thinner than average around Greenland, where typically the thickest ice of the Arctic is found. Overall, nearly the entire Arctic Ocean is thinner than the 1981-2010 average.”

This is illustrated in the maps below, which show the estimated thickness of Arctic sea ice in August 2018 (left) and how that compares to the 1981-2010 average (right). The dark blue shading in the right-hand map shows areas where ice is as much as two metres thinner than average for this time of year.

Current simulated (PIOMAS; Zhang and Rothrock, 2003) sea ice thickness (left) and anomalies relative to a 1981-2010 baseline (right) for August 2018.
Image: Zack Labe

Both Arctic sea ice thickness and extent are continuing on the steady descent seen over the last four decades, says Labe. This is down to a combination of natural variability and human-caused warming.

But even with relatively cool conditions for some of this summer, sea ice extent still dropped below 5m sq km, notes Labe:

“This is probably a result of the long-term loss of multi-year, thicker ice leaving the Arctic a very different place than 30 years ago. Unfortunately, these dramatic changes are now happening in all seasons.”

As Carbon Brief reported in its “State of the Climate” analysis in August, 2018 has seen record low amounts of “multiyear ice” – ice that has survived without melting for multiple years. In the first half of 2018, multiyear ice comprised just 34% of Arctic sea ice, with only 2% at least five years old. In the 1980s, upwards of 60% of Arctic sea ice was multiyear ice.

The lack of multi-year ice is even affecting how scientists carry out their fieldwork, adds Stroeve:

“For us to conduct our experiments and measure Arctic changes effectively, we need to be able to deploy our instruments on mature, multi-year ice so that they have a chance to survive at least one year. This sort of ice is harder and harder to find and we are moving further north to find it. For us scientists, we might need to think about redesigning our instruments to cope with the rapid changes we are witnessing.”