Sea ice in the Arctic region has declined in recent decades in area and volume due to climate change. It has been melting more in summer than it refreezes in winter. Global warming, caused by greenhouse gas forcing is responsible for the decline in Arctic sea ice. The decline of sea ice in the Arctic has been accelerating during the early twenty-first century, with a decline rate of 4.7% per decade (it has declined over 50% since the first satellite records). Summertime sea ice will likely cease to exist sometime during the 21st century.
The region is at its warmest in at least 4,000 years. Furthermore, the Arctic-wide melt season has lengthened at a rate of five days per decade (from 1979 to 2013), dominated by a later autumn freeze-up. The IPCC Sixth Assessment Report (2021) stated that Arctic sea ice area will likely drop below 1 million km2 in at least some Septembers before 2050.Fox-Kemper, B., H.T. Hewitt, C. Xiao, G. Aðalgeirsdóttir, S.S. Drijfhout, T.L. Edwards, N.R. Golledge, M. Hemer, R.E. Kopp, G. Krinner, A. Mix, D. Notz, S. Nowicki, I.S. Nurhati, L. Ruiz, J.-B. Sallée, A.B.A. Slangen, and Y. Yu, 2021: Chapter 9: Ocean, Cryosphere and Sea Level Change. In Climate Change 2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change [Masson-Delmotte, V., P. Zhai, A. Pirani, S.L. Connors, C. Péan, S. Berger, N. Caud, Y. Chen, L. Goldfarb, M.I. Gomis, M. Huang, K. Leitzell, E. Lonnoy, J.B.R. Matthews, T.K. Maycock, T. Waterfield, O. Yelekçi, R. Yu, and B. Zhou (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, pp. 1211–1362, doi:10.1017/9781009157896.011. In September 2020, the US National Snow and Ice Data Center reported that the Arctic sea ice in 2020 had melted to an extent of 3.74 million km2, its second-smallest extent since records began in 1979. Earth lost 28 trillion tonnes of ice between 1994 and 2017, with Arctic sea ice accounting for 7.6 trillion tonnes of this loss. The rate of ice loss has risen by 57% since the 1990s.
Sea ice loss is one of the main drivers of Arctic amplification, the phenomenon that the Arctic warms faster than the rest of the world under climate change. It is plausible that sea ice decline also makes the jet stream weaker, which would cause more persistent and extreme weather in mid-latitudes. Shipping is more often possible in the Arctic now, and will likely increase further. Both the disappearance of sea ice and the resulting possibility of more human activity in the Arctic Ocean pose a risk to local wildlife such as polar bears.
One important aspect in understanding sea ice decline is the Arctic dipole anomaly. This phenomenon appears to have slowed down the overall loss of sea ice between 2007 and 2021, but such a trend will probably not continue. In March 2026, the US National Snow and Ice Data Center reported that Arctic sea ice reached about 14.22 million km2, which was among the lowest yearly high points recorded in more than four decades of satellite monitoring.
Definitions
The Arctic Ocean is the mass of water positioned approximately above latitude 65° N. Arctic Sea Ice refers to the area of the Arctic Ocean covered by ice. The Arctic sea ice minimum is the day in a given year when Arctic sea ice reaches its smallest extent, occurring at the end of the summer melting season, normally during September. Arctic Sea ice maximum is the day of a year when Arctic sea ice reaches its largest extent near the end of the Arctic cold season, normally during March. Typical data visualizations for Arctic sea ice include average monthly measurements or graphs for the annual minimum or maximum extent, as shown in the adjacent images.
Sea ice extent is defined as the area with at least 15% of sea ice cover; it is more often used as a metric than simple total sea ice area. This metric is used to address uncertainty in distinguishing open sea water from melted water on top of solid ice, which satellite detection methods have difficulty differentiating. This is primarily an issue in summer months.
Observations
A 2007 study found the decline to be "faster than forecasted" by model simulations. A 2011 study suggested that it could be reconciled by internal variability enhancing the greenhouse gas-forced sea ice decline over the last few decades. A 2012 study, with a newer set of simulations, also projected rates of retreat that were somewhat less than that actually observed.
Satellite era
Observation with satellites shows that Arctic sea ice area, extent, and volume have been in decline for a few decades. The amount of multi-year sea ice in the Arctic has declined considerably in recent decades. In 1988, ice that was at least 4 years old accounted for 26% of the Arctic's sea ice. By 2013, ice that age was only 7% of all Arctic sea ice.Watch 27 years of 'old' Arctic ice melt away in seconds The Guardian 21 February 2014
Scientists recently measured sixteen-foot (five-meter) wave heights during a storm in the Beaufort Sea in mid-August until late October 2012. This is a new phenomenon for the region, since a permanent sea ice cover normally prevents wave formation. Wave action breaks up sea ice, and thus could become a feedback mechanism, driving sea ice decline.
For January 2016, the satellite-based data showed the lowest overall Arctic sea ice extent of any January since records began in 1979. Bob Henson from Wunderground noted:
January 2016's remarkable phase transition of Arctic oscillation was driven by a rapid tropospheric warming in the Arctic, a pattern that appears to have increased surpassing the so-called stratospheric sudden warming. The previous record of the lowest extent of the Arctic Ocean covered by ice in 2012 saw a low of 1.31 million square miles (3.387 million square kilometers). This replaced the previous record set on September 18, 2007, at 1.61 million square miles (4.16 million square kilometers). The minimum extent on 18th Sept 2019 was 1.60 million square miles (4.153 million square kilometers).
A 2018 study of the thickness of sea ice found a decrease of 66% or 2.0 m over the last six decades and a shift from permanent ice to largely seasonal ice cover.
Earlier data The overall trend indicated in the passive microwave record from 1978 through mid-1995 shows that the extent of Arctic sea ice is decreasing 2.7% per decade.Johannessen, O. M., M. Miles, and E. Bjørgo, 1995: The Arctic's shrinking sea ice. Nature, 376, 126–127. Subsequent work with the satellite passive-microwave data indicates that from late October 1978 through the end of 1996 the extent of Arctic sea ice decreased by 2.9% per decade.Cavalieri, D. J., P. Gloersen, C. L. Parkinson, J. C. Comiso, and H. J. Zwally, 1997: Observed hemispheric asymmetry in global sea ice changes. Science, 278, 1104–1106. Sea ice extent for the Northern Hemisphere showed a decrease of 3.8% ± 0.3% per decade from November 1978 to December 2012.
Future ice loss
An "ice-free" Arctic Ocean, sometimes referred to as a "blue ocean event" (BOE), is often defined as "having less than 1 million square kilometers of sea ice", because it is very difficult to melt the thick ice around the Canadian Arctic Archipelago. The IPCC AR5 defines "nearly ice-free conditions" as a sea ice extent of less than 106 km2 for at least five consecutive years.
Estimating the exact year when the Arctic Ocean will become "ice-free" is very difficult, due to the large role of interannual variability in sea ice trends. In Overland and Wang (2013), the authors investigated three different ways of predicting future sea ice levels. They noted that the average of all models used in 2013 was decades behind the observations, and only the subset of models with the most aggressive ice loss was able to match the observations. However, the authors cautioned that there is no guarantee those models would continue to match the observations, and hence that their estimate of ice-free conditions first appearing in 2040s may still be flawed. Thus, they advocated for the use of expert judgement in addition to models to help predict ice-free Arctic events, but they noted that expert judgement could also be done in two different ways: directly extrapolating ice loss trends (which would suggest an ice-free Arctic in 2020) or assuming a slower decline trend punctuated by the occasional "big melt" seasons (such as those of 2007 and 2012) which pushes back the date to 2028 or further into 2030s, depending on the starting assumptions about the timing and the extent of the next "big melt". Consequently, there has been a recent history of competing projections from climate models and from individual experts.
Climate models
A 2006 paper examined projections from the Community Climate System Model and predicted "near ice-free September conditions by 2040".
A 2009 paper from Muyin Wang and James E. Overland applied observational constraints to the projections from six CMIP3 climate models and estimated nearly ice-free Arctic Ocean around September 2037, with a chance it could happen as early as 2028. In 2012, this pair of researchers repeated the exercise with CMIP5 models and found that under the highest-emission scenario in CMIP5, Representative Concentration Pathway 8.5, ice-free September first occurs between 14 and 36 years after the baseline year of 2007, with the median of 28 years (i.e. around 2035).
In 2009, a study using 18 CMIP3 climate models found that they project ice-free Arctic a little before 2100 under a scenario of medium future greenhouse gas emissions. In 2012, a different team used CMIP5 models and their moderate emission scenario, RCP 4.5 (which represents somewhat lower emissions than the scenario in CMIP3), and found that while their mean estimate avoids ice-free Arctic before the end of the century, ice-free conditions in 2045 were within one standard deviation of the mean.
In 2013, a study compared projections from the best-performing subset of CMIP5 models with the output from all 30 models after it was constrained by the historical ice conditions, and found good agreement between these approaches. Altogether, it projected ice-free September between 2054 and 2058 under RCP 8.5, while under RCP 4.5, Arctic ice gets very close to the ice-free threshold in 2060s, but does not cross it by the end of the century, and stays at an extent of 1.7 million km2.
In 2014, IPCC Fifth Assessment Report indicated a risk of ice-free summer around 2050 under the scenario of highest possible emissions.
The Third U.S. National Climate Assessment (NCA), released May 6, 2014, reported that the Arctic Ocean is expected to be ice free in summer before mid-century. Models that best match historical trends project a nearly ice-free Arctic in the summer by the 2030s.
In 2021, the IPCC Sixth Assessment Report assessed that there is "high confidence" that the Arctic Ocean will likely become practically ice-free in September before the year 2050 under all SSP scenarios.
A paper published in 2021 shows that the CMIP6 models which perform the best at simulating Arctic sea ice trends project the first ice-free conditions around 2035 under SSP5-8.5, which is the scenario of continually accelerating greenhouse gas emissions.
By weighting multiple CMIP6 projections, the first year of an ice-free Arctic is likely to occur during 2040–2072 under the SSP3-7.0 scenario.
Impacts on the physical environment Global climate change