Sea Ice Retreat: More than Meets the Eye

(from Lloyd’s Report 2012: Arctic Opening – Opportunity and Risk in the High North)


The reduction in the extent of summer sea ice is the most high-profile indicator of Arctic climate change. The processes driving this retreat are complex: sea ice dynamics, air temperature, sea temperature, weather patterns and the physical geography of the Arctic as an ocean enclosed by land all play a part.

Although there is some variability in ice extent from year to year, and although the annual cycle of melting and freezing continues, the overall downward trend in the September sea ice extent, recorded by the US National Snow and Ice Data Center (NSIDC) since 1979, is strong and unambiguous. Historical data from other sources – such as the number of days when particular harbours have been iced up or ice-free, or century-old ice records of scientific expeditions – support the picture of sharply reduced ice extent compared with earlier periods.

In September 2011, the month when Arctic sea ice extent is typically at its lowest, ice coverage fell to a low of 4.33 million square kilometres (1.67 million square miles), some 2.38 million square kilometres less than the 1979–2000 average (see Figure 3)1 . The NSIDC records show ice extent lower in only one other year – 2007, when it reached 4.17 million square kilometres. Using a slightly different methodology, scientists at the University of Bremen reported that Arctic sea ice extent actually reached a minimum of 4.24 million square kilometres on 8 September 2011 – 27,000 square kilometres below the Bremen team’s estimate for summer 20072. According to their estimates, Arctic sea ice cover last reached this minimum 8,000 years ago.

Dramatic as it is, the reduction in the extent of Arctic sea ice cover is only half the picture. Arctic ice is also both thinner and younger than previously. In the early 1980s, the NSIDC estimated that as much as 40% of Arctic September ice was more than five years old. In 2011, that proportion had declined to 5%. This shift has important ramifications, both climatic (e.g. the dynamics of the ice cover) and socioeconomic (e.g. the location of multi-year ice has a significant impact on the viability of various Arctic shipping routes).

Estimating ice thickness – and therefore the overall volume of Arctic ice – is more complicated than measuring surface ice extent. Ice thickness varies across the Arctic depending on a range of conditions, and cannot be continuously assessed. Most Arctic ice is constantly moving (vi).

However, the picture built up by a combination of modelling, on- and under-ice data collection from the Arctic and satellite remote sensing suggests that Arctic ice thickness – and volume – is declining even more rapidly than ice extent. The monthly average ice volume estimated in September 2011 was 4,300 km3, 66% below the mean for 1979–2010 (see Figure 4)3.

Ice extent, age and thickness are all relevant to the likely future of Arctic sea ice. Recent research suggests that most models have underestimated the importance of these and other factors in predicting the trajectory of Arctic ice extent4. Younger and thinner Arctic ice is more prone to melting, and more prone to break-up – including by ships. The formation of sea ice will be affected by a relatively more open Arctic Ocean, as waves tend to become stronger and more frequent (viii). In short, the less ice there is in one year, the harder it is for ice extent and volume to recover over the winter months. The demise of Arctic sea ice – to the extent of ice-free Arctic summers – could be more abrupt than the trend lines suggest.

Projections of the date when the Arctic Ocean will first be free of sea ice in summer have been brought forward in recent years. The 2007 IPCC report suggested that this might occur by the end of the 21st century. Since then the record of actual reductions in sea ice extent have led most scientists to conclude that the first ice-free summer in the Arctic Ocean will be within the next 25 to 40 years, while some claim it could conceivably occur within the next decade (ix). Reductions in summer sea ice allowing for essentially unimpeded maritime traffic, will occur before the Arctic Ocean becomes fully ice-free in summer.

The Arctic Ocean will continue to freeze up in winter. Ice extent will remain unpredictable, hampering regular traffic without ice-capable vessels and complicating planning for oil and gas exploration. Sea ice will continue to be a challenge to navigation in large parts of the Arctic for much of the year, particularly where broken ice clogs narrow waterways, or where sea ice is flushed out of the Arctic through the Davis and Fram Straits. In some places, climate change may result in an accelerated rate of calving of icebergs from glaciers, which will in turn increase the number and size of icebergs (x). This is likely to present additional challenges for maritime activity on the sea surface and raise the risk of scouring along the seabed (xi).

Figure 3: Decline in average sea ice extent in September, 1979-2011 Image  

Figure 4: Decline in average estimated sea ice volume, 1979-now (vii) 


(vi) Exception to this is land-fast ice which is sea ice that has frozen over shallow parts of the continental shelf.

(vii) Shaded areas show one or two standard deviations from the trend. Error bars indicate the uncertainty of the monthly anomaly plotted once per year.

(viii) The enclosure of the Arctic Ocean by the land masses of North America and Eurasia have tended to reduce the fetch of waves, and thereby lead to different sea ice dynamics to those around Antarctica.

(ix) There is a wide range of projections for when the first ice-free Arctic summer will occur. See, for example, Muyin Wang and James E. Overland, “A sea ice free summer Arctic within 30 years?”, Geophysical Research Letters, Vol. 36, 2009; and Julienne Stroeve, Marika M. Holland, Walt Meier, Ted Scambos and Mark Serreze, “Arctic sea ice decline: Faster than forecast”, Geophysical Research Letters, Vol. 34, 2007. The most aggressive projections suggest this could occur before 2020 (see, for example, Professor Wieslaw Maslowski, Naval Postgraduate School, or Professor Peter Wadhams, University of Cambridge).

(x) Calving occurs when an iceberg breaks off from an ice-shelf (in the Antarctic) or from a glacier as it runs into the sea (for example, off the coast of Greenland).

(xi) Scouring occurs when the bottom of a glacier drags along the seabed. In relatively shallow waters this is potentially a risk for sub-sea infrastructure, such as cables, pipelines and sub-sea oil and gas installations.


  •  1. http://nsidc.org/arcticseaicenews/2011/10/
  •  2. http://www.iup.uni-bremen.de:8084/amsr/#Arctic
  •  3. http://psc.apl. washington.edu/wordpress/research/projects/projectionsof- an-ice-diminished-arctic-ocean/
  •  4. IPCC climate models do not capture Arctic sea ice drift acceleration: Consequences in terms of projected sea ice thinning and decline Vol. 116, 2011 P. Rampal J. Weiss C. Dubois J.-M. Campin Journal of Geophysical Research

Charles Emmerson, Glada Lahn, 2012, Sea Ice Retreat: More than Meets the Eye © Lloyd’s 

Search Guide [toggle]