Recent changes in Arctic sea ice: observations and linkages to the climate system

This is a paper that I wrote last year as part of my Post Grad Diploma in Geography. It's a little general rather than detailed but gives a reasonably good overview of the state of the Arctic last year around May/June/July (in my somewhat biased view). Although I've mostly kept away from it in this paper I think that methane is important in this mix but there's a contrast between Russian and North American views on how important it is, with the Russians giving it more emphasis. I sit in the middle a little and probably wouldn't mind spending a summer on the Arctic to look at it more closely. Svalbard and the Russian Arctic being my preference for destinations although my Russian definitely would need work beforehand.

As the news is showing currently the warming in this part of the world is a lot more than elsewhere.


Arctic sea ice has been decreasing in extent by about 12% per decade since the beginning of the satellite record in 1979 (Fetterer, F. 2018). Recently this deterioration has sped up with the 14 lowest minimum extents (normally in September) all being after 2004 and the major falls in extent being in 2007, 2011/2012 and in 2016. These all set much lower sea ice level minimum boundaries that then last only a few years before the next precipice is jumped. So as expected although there has been a small recovery in the last few years in the minimum sea ice levels it is not enough to take these years out of the top six lowest minimum Arctic sea ice extents to date. At the other end of the ice growth season the three lowest maximum Arctic sea ice extents (normally in March) are all in the last 3 years (2016, 2017 & 2018 – Perovich, D. et al, 2018). Due to this increased melt current Arctic sea ice is also now a lot younger than it used to be and therefore less stable and more prone to melt (Barber, D.G. et al., 2009). Increased snowfalls from the changes in weather patterns that are occurring due to the increased greenhouse gases (GHGs) in the atmosphere are also restricting ice growth in the Arctic winter. It is clear therefore that we have entered an unstable era with a new summer ice extent norm yet to be set. In fact on our current trajectory one should expect the new norm to be no summer sea ice within decades (or sooner) (Overland, J.E. and Wung, M., 2013), a longer summer melt season (if melt remains an operative word) and the possibility of no Arctic sea ice at all (in any season) at some stage in the not too distant future.

Click to read rest of paper In this paper I attempt to define the most likely timeframe for these events and some of the possible global consequences of these rapid changes in Arctic sea ice extent.

Figure 1: Depiction of Arctic sea ice that’s at least four years old as of September 1984 (left) vs. September 2016 (right). Credit: NASA

Figure 2: Annual Arctic Sea Ice extent 1979 (beginning of Satellite record) to 2018. Credit: Zack Labe.

Arctic Amplification

Foster, G.L. al. (2017) state that “the climatic evolution of the Earth over geological time is largely a function of the concentration of the non-condensing greenhouse gases, planetary albedo and the TSI“ (TSI is total solar irradiance). Although, as is stated by Foster, G.L. al. (2017), TSI is very slowly increasing (4% over last 400 million years) it is the greenhouse gases (GHGs), specifically the non-condensing ones such as CO2 and CH4, which have been the main forcing agent of our climate for most of this period. For the most part their contribution has been beneficial, as they have kept Earth warmer than it otherwise would have been and therefore allowed life to develop. In addition it is thought that their decreasing atmospheric concentrations have compensated for increased TSI (Foster, G.L. al., 2017) over time thereby maintaining this beneficial climate. But the recent rapid increase of these GHGs in the Earth’s atmosphere due to the burning of stored carbon sources (such as coal and oil) since industrialization is now a level of climate forcing that is unprecedented in the recent, and quite possibly distant, past.

The Arctic is in fact currently warming at a rate that is up to twice the rate of the warming that is occurring at lower latitudes (Arctic Monitoring and Assessment Programme, 2017, Overland, J.E. et al 2014). This exaggerated warming trend in the Arctic over the last several decades, in comparison to the warming at lower lattitudes, is known as Arctic Amplification and is related to such feedbacks (amongst others) as lowered albedo and increased heat retention in what are increasingly ice-free seas (or ocean) during the Arctic summer. This then allows for strong heat transfers to the atmosphere in early winter (Serreze,M.C. et al., 2008). In addition there are changes in the constitution of the sea ice in the Arctic with less multi year ice (MYI) and more first-year sea ice (FYI) as well as ice deterioration.

In the last several decades declining albedo has also become an agent of a warming climate due mostly to ice sheet and sea ice collapse. Albedo itself is specifically a measurement between 0 and 1 of planet surface and cloud reflectivity and in the study of climate is associated with the sun’s radiative impact. It is especially relevant in the study of climate due to the ice preserved in the cryosphere and most especially that of the two poles where its melt can significantly impact the Earth’s climate system. This comes to pass as less albedo leads to higher melt rates due to the sun’s radiation being absorbed rather than reflected. Typically the values of albedo range from 0.07 for open water to 0.65 for bare ice and up to 0.85 for snow-covered sea ice (Stroeve, J.C. et al., 2012). Lowered albedo from the reduced extent of sea ice and snow is basically a feedback loop initiated by anthropogenic climate forcing via such events as warmer temperatures of sea and air, more moisture in the air, the spread of ash from more frequent forest fires to the cryosphere and an increased cover of low albedo micro-organisms especially in the meltponds that form on ice. But once set in motion is a climate forcing of its own and may itself initiate other forcings in the future such as methane release (a controversial topic in climate science) from the East Siberian Arctic Shelf (Shakhova, N. et al., 2010) or ocean current changes such as the the Atlantic meridional overturning circulation (AMOC) which is likely impacted by declining Arctic sea ice (Sévellec, F. et al., 2017).

The collapse in extent and area (area does not include the parts of the extent that are free of ice, so extent is the larger measure), and the changing nature, of the sea ice in the Arctic has resulted in less overall albedo as currently there is more open water, as opposed to ice, than there has been in the Arctic for at least 1457 years (Kinnard, C. et al., 2011). It is not a uniform change though as more snow, resulting from the higher moisture levels (IPCC AR5 Chapter 2, Hartmann, J. et al., 2013) in the atmosphere due to anthropogenic forcing, means that a specific location’s albedo may actually increase as part of an increase in snowfall such as it did in Greenland in 2017 (Tedesco, M. et al., 2017, Box, J.E., 2017). However it is more likely, in the current Arctic environment, that increased levels of snowfall will settle mostly on seawater or FYI where it can restrict ice growth and undermines the potential for multi year ice (MYI) growth (Sato, K. and Inoue, J., 2017).

Collecting the data for Arctic sea ice observations: In Situ.

In Situ observations in the Arctic are helpful for both past and present observations. They come from a variety of sources and include those collected from the likes of buoys, atmospheric and tropospheric profiling, Arctic radiosonde (rawinsonde) stations, field campaigns, fixed moorings and floats. They can include sea, land, under ice, submarine, aircraft and ship based observations as well as profiling of paleo data. Sub ice upward looking sonars (ULS), for example those deployed by the Beaufort Gyre Exploration Project (, can measure sea ice thickness as well as temperature, currents and salinity. But there is an issue regarding coverage of these observations in that there is a tendency to a location bias to the Arctic coast along with a lack of representation at higher latitudes, as in the case of rawinsonde stations in the Arctic (Figure 3).

Figure 3: Locations of stations represented in The Historical Arctic Rawinsonde Archive. Symbols denote the number of years (NYR) for which data are available (see legend). Records of at least 5 yr in length are available for 80 stations; records of at least 10 yr are available for 69 stations. Credit Kahl et al (1992).

There are of course good accessibility reasons for such bias, thick MYI areas can restrict access, the North Pole is (normally) less hospitable than Svalbard or Murmansk are, field campaigns may not have the resources to access the higher latitudes (nor to do so in the winter), ice floes migrate and so on. Nevertheless the end result is a bias towards the more accessible and popular locations that has to be compensated for. The easiest way of doing this is in comparison to the satellite record.

Collecting the data for Arctic sea ice observations: Satellite

Although satellite based sensoring equipment, and the associated algorithms that manipulate and analyse the data, have all improved +much since 1979, when they first became available, and do not suffer from the frequent location bias of in situ observations (other than directly over the North Pole) there are still some issues with the data especially in relation to measuring the thickness of sea ice. Satellite measuring of sea ice thickness and volume is a very recent innovation and as Lang et al. (2017; p.409) state in their 2017 paper “pan-Arctic thickness measurements only exist since the launch of NASAs Ice, Cloud, and land Elevation Satellite (ICESat) in 2003”. The 2010 launch of CryoSat-2 further improved on the ice thickness measuring ability of satellite based recorders. The relevant instrument onboard this satellite for sea ice thickness is a synthetic aperture radar (SAR)/Interferometric Radar Altimeter (SIRAL) (Kwok, R. & Cunningham, G.F., 2015).

But as NSIDC themselves state “to the sensor, surface melt appears to be open water rather than water on top of sea ice” (NSIDC, 2018). There is also a fairly well known error in the ENVISAT (short for “Environmental Satellite” which was launched in 2002 by the European Space Agency and retired in 2012) data for 2002/2003 due to some unsolved issues with ENVISAT during its first years of operation. Guerreiro et al. (2017; p.2059) address this and other matters regarding the datasets from this satellite in relation to ice thickness noting that “the current estimates obtained with the Envisat mission (2002–2012) still require some large improvements”. Sea ice thickness, it seems, is not easily tamed.

The status of the current satellite regime is also at some risk as the three current United States satellites are past their end of life dates and a reliable replacement has yet to be launched. In fact the Defense Meteorological Satellite Program (DMSP) F-20 satellite, which was intended as a replacement, was cancelled in early 2016 during the Obama administration (Witze, A., 2017). It would seem unlikely to be re-launched under the Trump administration other than for military purposes.

Nevertheless until now there has been a reliable satellite record from 1979 to the present giving scientists access to accurate data on sea ice extent, area and (from 2003) ice thickness and volume albeit there is ongoing work in improving the interpretation of this data.

Some issues regarding satellite data collection and calibration

• Sonar data on ice thickness from the US Navy is available on the NSIDC website ( but currently ends at November 2005. Further data after this date is available but it is not generally accessible. This data would be very useful as comparison to satellite records (Laxon, S.W. et al, 2013). • Radar penetration into the overlying snow layer remains a subject of investigation and may introduce errors into thickness calculations (Armitage, T.W.K., 2016). • Differentiating ponding on ice flows from general seawater has been an issue with satellite-based instruments (Kwok, R. and Cunningham, G.F., 2015).

The Arctic in previous epochs

There is some debate about the extent of Arctic sea ice going back as far as the Miocene. There were several warm periods (warmer than present) from this epoch to our own despite CO2 levels either being lower or only slightly higher than current readings (mostly in the range of 290-450 ppm). Nevertheless, recent scholarship is indicating that Arctic sea ice survived these warmer periods, including at least one in the Holocene (Stiegg, E.J. & Nieff, P.D., 2018), even if its extent was significantly reduced from present levels (Polyak, L. et al, 2010). Indeed on the basis of current CO2 atmospheric readings, if we assume the ‘business as usual’ Representative Concentration Pathway 8.5 (RCP 8.5, Riahi, K. et al 2007) scenario, we may in fact be forcing the Earth’s climate to a state not has not existed since 48 million years BP when both poles were ice free and sea levels in the vicinity of 70 metres higher than present day. There is even some suggestion that the current forcing may push us beyond 420 million years ago (Foster, G.L. al, 2017) for a parallel due to the combination of equally high GHGs but with the association of a higher TSI.

Figure 4: Compilation of available CO2 data for the last 450 million years. Proxy records are colour coded and labelled in the relevant panel. Greenhouse gas emission scenarios (RCP – Representative Concentration Pathways) used in IPCC AR5 are shown in the right hand panel. Note the variable log scale for time. For the geological data a smoothed line has been fit to the data with an uncertainty accounting for uncertainty in age and CO2. The black line describes the most probable long-term CO2 with 68% confidence limits in red, and 95% confidence in pink. Credit: Gavin Foster, Dana Royer and Dan Lunt.

This ice record before 1850 relies on proxies such as ice cores and sediment records. Therefore although the basic outline is accepted there is still some debate around the specifics and future paleo research will undoubtedly further help us to understand this record.

Anthropogenic Forcing

The Earth, compared to its near neighbours Mars and Venus, has a climate that is very beneficial to life (sometimes referred to as a Goldilocks climate, or as Earth being in the Goldilocks zone). This is is a result of the level of greenhouse gases in Earth’s atmosphere and especially that of CO2 (Johnson, C. W., 1998). Earth’s carbon cycle which both deposits and removes CO2 from the atmosphere has developed over billions of years and for at least the last billion years or so it has cycled CO2 through the Earth’s atmosphere, soil, ocean, rock and mantle in a way that has balanced the climate in a zone that allows life to continue and evolve. The main actors on climate before the industrial age were forcings such as solar activity, orbital cycles and volcanic activity as well as CO2.

Figure 5: Holocene forcings. Credit: Wanner et al (2011)

But the Earth’s carbon cycle has both not evolved to balance for either another agent of releasing CO2 to the atmosphere (the burning of coal, oil and gas by anthropomorphic agents) nor the speed at which it is being released at.

The most recent cooling period, the Neoglacial, was likely terminated by Anthropogenic CO2 emissions at the end of the 19th/beginning of the 20th Centuries. These emissions in fact have continued to increase, even if somewhat more slowly recently (although recent increases suggest that they were underreported in the middle of the current decade) and are currently in the vicinity of 36 gigatonnes per year (Olivier, J.G.J., Schure, K.M. and Peters, J.A.H.W., 2017).

The Arctic ice record: 1850 to 1979

In the present epoch the first satellite records, the most accurate of our records of sea ice extent and area, date to 1979. Since then there have been many improvements that allow more accuracy of such observations than there previously were, especially in relation to measuring mass of sea ice rather than just its extent. Nevertheless it is still an area in development (Quartly, G.D. et al., 2018). Before 1979 we rely almost exclusively on in situ measurements of various types.

There has been significant improvement in the digitising of records that date to before 1979 and therefore providing Arctic researchers access to accurate data from the pre-satellite record (this data correlates very well with the satellite record after 1979 (Figure 6).

Figure 6: Time series of Northern Hemisphere sea ice extent (solid line) and area (dotted) for 1890-2007, for the Met Office Hadley Centre datasets HadISST. (black), HadISST1.1 (red), and the NASA Team dataset (blue). Monthly average values are shown for a) January and b) July. Credit: Holly Titchner, UK Met Office. Credit: Fetterer/Carbon Brief

The first of these was the work of Professor John Walsh and Dr Mick Kelly (retired) who transcribed data in the 1980s from the US Navy and UK Meteorological Office (here and here)1 as well as from the Danish Meteorological Institute’s yearbook maps. More recently there was an international co-operative effort by many individuals to digitise charts, maps, logbooks, observations and surveys from various diverse sources back to 1850 (Fetterer, F. 2016). This has resulted in the ‘Gridded Monthly Sea Ice Extent and Concentration, 1850 Onward (Version 1)’ available from the National Snow and Ice Data Center (NSIDC) in the United States. Extrapolating this data to sea ice extents from 1850 onwards as well as incorporating the satellite records from 1979 onwards shows clearly that the current sea ice extents are the lowest on record and that the collapse of sea ice extent (especially that of the September minimum) in the Arctic has been rather phenomenal in the last few decades, and unprecedented in this record.

Figure 7: Sea ice cover maps for the annual minimum in September, for the periods 1850-1900, 1901-1950, 1951-2000, and 2001-2013. The maps show the sea ice extent in the lowest minimum during each period, which are in years: 1879, 1943, 1995, and 2012. Credit: Fetterer/Carbon Brief

A recent study (Kinnard, C. et al 2011) utilizing proxies such as ice cores, dendrology, lake sediments and historical Arctic sea ice observations would seem to extend the uniqueness of this collapse in Arctic Ice extent in the Holocene out even further, to 561 CE or 1,457 years ago (1,450 years at the time that the paper was written).

Recent changes in the Arctic

Figure 8: Near surface temperature changes derived from observations. Credit NASA/Jason Box. Sourced from

The Arctic is a region where thinner atmosphere, changes in albedo, cloud cover, air and sea/ocean temperatures along with changes in circulation patterns can all have an impact (Swivastrava, A.N. & Stroeve, J.C., 2005) in the process referred to as Arctic Amplification. This is clearly seen in Arctic sea-ice which has been increasing in its melt extent during the northern summer (September being the month of its lowest extent) for several decades and is both younger and less consistent in its substance and therefore more prone to future melt (Barber et al, 2009).

This year the sea ice extent in the Arctic has alternated between its lowest extents on record for early winter and its position as at July 17th 2018 being, as the NSIDC states, “1.24 million square kilometers (479,000 square miles) below the 1981 to 2010 average, but 670,000 square kilometers (259,000 square miles) above the record low for this day in 2011”* (NSIDC). But the underlying trend is down and according to Stroeve, J.C. et al (2012) the September (summer minimum) extent is collapsing at a rate of 12.4% per decade. In that same paper the authors also state that “These negative trends in ice concentration imply a corresponding decrease in surface albedo since open water has a much lower albedo than sea ice” (page 1102). In this case the disappearance of part of the Arctic ice extent then contributes to further sea ice loss via the loss of albedo, a case of Arctic Amplification. In addition, as Stroeve, J.C. et al (2018) point out, the date of the winter re-freeze is happening later in the year (specifically in the Chukchi, East Siberian, Laptev and Barents seas: page 1792). So the albedo of the Arctic ocean is not only being tempered by the collapse of Arctic sea ice, especially during the summer, but it is also being tempered by a longer Arctic summer. Indeed since July 17th, 2018, the sea ice extent has fallen quite remarkably again.

Early winter flux of Arctic ocean heat to the atmosphere can be up to two times what it is through an ice cover (Sandven, S. & Johannessen, O.M., 2010).

Clearly the Arctic is now contributing to the demise of its own sea ice extent via its decreased albedo stocks of reflective sea ice and snow.

Sea ice thickness and volume

The thickness of the Arctic sea ice cover has also seen a very large reduction over the last several decades. According to Kwok, R. and Rothrock, D.A. (2009), and based on declassified submarine sonar measurements of the United States Navy, this reduction was in the order of almost a halving of the ice thickness between 1980 and 2008, from a mean of 3.64 metres to 1.89 metres. This is quite variable and the thickest ice is near Ellesmere Island, the Greenland Coast and at the North Pole (Stroeve, J.C. et al, 2011, Kwok, R. and Rothrock, D.A., 2009, Barber, D.G. et al, 2018).

Figure 9: Bar chart shows the mean thicknesses of six Arctic regions for the three periods (1958– 1976, 1993 – 1997, 2003 – 2007). Thicknesses have been seasonally adjusted to September 15. Credit: Krok and Rothrock.

Kwok, R. and Rothrock, D.A. (2009) also note that from 2000 through to 2008 there appeared to be decreased rate of reduction in the thickness of the sea ice, based exclusively on the satellite record as the declassified sonar records were not available after 2000. Kwok, R. and Rothrock, D.A. (2009; L15501) in noting this make the statement that “Barring any systematic biases, possible reasons for the more moderate differences in the ICESat data could be the muting of the annual cycle during the 2003–2008 period due to the warmer winters and shorter growth seasons, or the fastest growth rate in the early fall unaccounted for by the ICESat periods.”. In a later 2015 paper by Kwok, R. and Cunningham, G.F. (2015; p.14) one of the sections is headed “Uncertainty in volume estimates” and this in itself clearly indicates that the measurement of thickness, although improving, is continuing to cause problems mostly due to ridging, ponding on ice and overlaying snow thickness which pose the largest issues for satellite based instrumentation. As Kwok, R. and Cunningham, G.F. (2015; p.14-15) note “the correct way to use the limited data on ice density remains an issue to be resolved”.

Professor David Barber, after an excursion to the Beaufort Sea on the ice-breaker Amundsen in September 2009, along with other staff from the Centre for Earth Observation Science at the University of Manitoba, noted after their return (University of Manitoba, 2009) that “thin, “rotten” ice can electromagnetically masquerade as thick, multiyear sea ice” (a fact that NSIDC now acknowledges) and that this sea ice was able to be traversed at a rate of knots that was very close to that of the open sea. His conclusions included;

• That sea ice was continuing to decline in 2009 rather than rebounding from the lows of 2007 (which was a significant collapse of sea ice extent over previous levels at time but exceeded several times since) as the satellite record was indicating. • That “near-surface physical properties were found to be sufficiently alike that their radiometric and scattering characteristics were almost identical” [i.e. the inability to discern differences between MYI and this ‘rotten ice’ in the satellite data.]. This was put down to “both ice regimes had similar temperature and salinity profiles in the near surface volume, both ice types existed with a similar amount of open water between and within the floes, and finally both ice regimes were overlain by similar, recently formed new sea ice in areas of negative freeboard and in open water area”. • “Ship navigation across the pole is imminent as the type of ice which resides there is no longer a barrier to ships in the late summer and fall” (University of Manitoba, 2009).

Kwok, R. and Cunningham, G.F. (2015) also note something very similar in stating that “the uncertainty in ρi [bulk density of sea ice] could be a more significant error source in the estimate of hi [ice thickness] due to large expected difference in ρi between first-year ice [FYI] and multi year ice [MYI]” (2009; p.6) and also note that “With the younger MY ice cover (less than 3 years) found in the Arctic in recent years, the mean ρiMY may perhaps be different from that obtained from older MY ice from earlier decades” (2009; p.6) in which the last part of this statement can (arguably) be correlated with David Barbers term ‘rotten ice’.

Another issue that both Kwok, R. and Cunningham, G.F. (2015) and Laxon, S.W. et al (2013) cover is the bias between different satellites and they have both attempted to calibrate these with in situ measurements to ascertain what these bias' may be. In the case of Kwok, R. and Cunningham, G.F. (2015) they did this by calibrating with Beaufort Gyre mooring readings, noting their accuracy and consistency over time. This yielded a difference between 0.02 and 0.06 metres with Laxon, S.W. et al (2013) but in doing so they acknowledge their agreement with the latter in that present understanding of in situ and satellite measurements of thickness was not enough to ascertain the accuracy of the Satellite bias anymore than Laxon, S.W. et al (2013) had previously noted (calibrating in situ with the European Space Agency CryoSat-2 satellite).

Kwok, R. and Cunningham, G.F. (2015) estimate the yearly loss as 500 km3 lower than it was during the previous ICESat (2003-2010) era (but it could be noted that this is off a smaller ice sea volume) whereas Laxon, S.W. et al (2013) had estimated it at ~800 km3. Somewhat in contrast they offer up that the ice thickness is more or less the same as in Kwok, R. and Rothrock, D.A. (2009) for the earlier 2003-2007 period although the growth of ice thickness in the 2013/2014 season and the collapse of the same in 2007 may have skewed this comparison. Nevertheless the volume of sea ice trend continues downwards (just as the extent and area trends are) albeit that they seem to have temporarily stalled in the years 2013/2014.

Figure 10: Arctic sea ice volume anomaly from PIOMAS updated once a month. Daily Sea Ice volume anomalies for each day are computed relative to the 1979 to 2017 average for that day of the year. Tickmarks on time axis refer to 1st day of year. The trend for the period 1979- present is shown in blue. Shaded areas show one and twostandard deviations from the trend. Error bars indicate the uncertainty of the monthly anomaly plotted once per year. Credit: Polar Science Center

As should be expected this thinning of the sea ice has its correlation in the average age of sea ice as well. This has fallen from 57% of the ice pack being at least five years old in 1987 (with a quarter of that ice at least nine years old) to only 7% being of that age (at least 5 years old) by 2007 and practically none of that 2007 ice being 9 or more years old (Maslanik, J.A. et al 2007). And as per Perovich, D. et al (2017; p.3) “Compared to younger ice, older ice tends to be thicker, stronger and more resilient to changes in atmospheric and oceanic forcing”. In plain language younger ice makes summer melt of the same much more likely and undermines ice growth in the winter in addition to making it more likely that sea ice will be blown around the Arctic and beyond, especially as a result of Arctic storms

“Within the central Arctic Ocean, the coverage of old ice has declined by 88%, and ice that is at least 9 years old (ice that tends to be sequestered in the Beaufort Gyre) has essentially disappeared.” (Polyak, L. et al, 2010; p.1759).

Figure 11: Age of Arctic Sea ice at the end of summer 1983-2010. Credit Stroeve, J.C. et al (2012)

Arctic Storms

In early August 2012 a strong storm (the Great Arctic Cyclone of 2012) did indeed enter the central Arctic and lasted for 13 days (Parkinson, C.L. et al, 2013). It managed, in this timeframe, to separate an 106 km2 area of ice that then melted. At the end of the summer melt seaston (September) minimum ice extent was at a lower level than that of the extraordinary year of 2007. The 106 km2 area of ice that had been separated by this storm was integral to this record as it made up 57% of the lower extent in comparison to 2007.

Potential wider impacts on worlds climate

Although models can indicate possible outcomes initiated by these new sea ice states (because the new equilibrium has not been reached yet it is likely there will be many) there are still many unknowns due to our imperfect knowledge of all the different agents acting on the Earth’s climate. But they are likely to include changes to air, water currents, temperatures, weather patterns, precipitation and cloud cover in addition to sea ice extent and volume.

Possibly the most obvious indication of a changing Arctic climate is the appearance of large icebergs off the coasts of Newfoundland or Greenland and this has indeed occurred over the last few years, albeit such one off events may not individually be attributable to a general change in the Arctic climate. They can also be a result of carvings of formerly land grounded ice as well as detached MYI or FYI from the Arctic Ocean. But it is no accident that this is where they have appeared as these locations are where Arctic ice often exits the Arctic with the Fram Strait being the other, and the largest, exit point. Baffin Bay/Davis Strait in fact exports ~70% of the Arctic sea ice export that Fram Strait does (Kwok, R., 2007) making it the second largest. The other sea ice export regions are the Bering Strait (the smallest) and those between Svalbard and Russia/Siberia.

Fran Strait and Baffin Bay/Davis Strait are the main exit points for both any Greenland ice sheet melt as well as any changes in the Beaufort Gyre that could agitate cold water and sea ice removal from that region of the Arctic in their directions.

Beaufort gyre

The Beaufort Gyre (BG), in the words of those who know the most about it, is “known as a unique sea ice and water circulation component of the Arctic Ocean rotating clockwise under the influence of prevailing winds. These winds act as a giant pump collecting low salinity water from the Arctic Ocean surface layer in the BG center” (Jones, D. and Proshutinsky, A., 2018 – webpage).

When the BG is strong the transport of ice from the western to the eastern Arctic is enhanced and results in thicker ice there as a result of ridging and rafting processes up against the Siberian Coast. More or less the same thing happens in the Canada basin (Stroeve, J.C. et al, 2011). As such the BG has been a large contributer to the manufacture of sea ice in the Arctic in the last century and probably beyond. The regular temporary reversal, every 5-7 years, of its normal clockwise direction meant that ice and fresh water were expelled out into the North Atlantic mostly via the Fram Strait. But recently it seems to have stopped this temporary reversal process and has been collecting fresh water since at least 2003.

“Increased wind combined with increased geostrophic ocean currents around the Beaufort Gyre since 2007 are linked to increased ice drift rates in the Beaufort Sea, and to pooling of freshwater in the upper water column, increasing stratification, in the Canada Basin.” (Hock, R. et al, 2017; p. 7).

The volume of this fresh water at ~23,000 km3 is now about equal to that of Lake Baikal or to that of all of the Great Lakes combined and when the BG eventually reverses direction the resulting outflow of cold fresh water and sea ice to the North Atlantic could result in a dramatic shift of climate on both sides of this ocean (Jones, D. and Proshutinsky, A., 2018).

Something similar to such a reversal has happened in the recent past in what is called “The great salinity anomaly in the northern North Atlantic” which cooled this ocean from 1968 to 1982 (Proshutinsky, A. et al, 2015). One theory for why this has yet to happen in the current period is that the increased Greenland ice sheet melt is putting a block on cyclones developing on the northern end of the Gulf Stream and that this is inhibiting the change in circulation of the BG (Proshutinsky, A. et al, 2015). Recently though the BG seems to have ceased collection of further fresh water leading to expectations by some that it may be about to reverse.

Warm Arctic, Cold continent

In the last few years several papers have appeared suggesting that the collapse in sea ice extent in the Arctic is leading to colder temperatures on the continents bordering the Arctic, in particular Asia and North America (Zou, Y. et al, 2017). It is known as ‘Warm Arctic, Cold Continent’ (WACC) and is a much debated topic with some scientists aligning in opposition to it (Sun, L. et al, 2016) and others proposing that it applies only to Central Asia (Screen, J. et al, 2015). Indeed Screen, J. et al (2015; p.7) in noting the results of their modelling essentially state the opposite will happen in North America as “reductions in Hudson Bay sea ice, which occur at relatively low latitudes compared to other longitudes, may enhance the response over eastern North America. Hence, the decline in cold extremes is more pronounced and robust here than in other mid-latitude regions”. Further Sun, L. et al (2016; p. 5348) conclude that recent events do not fit the WACC explanation and suggest that “observed recent cooling in central North America likely resulted from a mode of circulation variability that is symptomatic of internal dynamics, rather than being a symptom of sea ice forcing due to effects of Arctic amplification overall”. and that “Warm Arctic, Warm Continents” (p. 5437). is a more likely outcome (Sun, L. et al 2016). There is a similar theory in regard to Northern Europe and the North Atlantic Oscillation which James Screen, J. also argues against (Screen, 2017).

It is this writer’s conclusion that both Sun, L. et al (2016) and Screen, J. et al (2015) are correct and that recent climate trends in snow accumulation in North America and Europe (Kunkel, K.E. et al, 2016) and winter temperatures in the northern latitudes are indicating that WACC is unlikely to hold, especially in North America.

Other impacts

In addition to the potential impacts discussed above, there are several others that are worthy of note and would have been expanded on in a longer work. This includes changes to thermohaline circulations, noteworthy in its potential impact on the Atlantic Meriodanal Overturning Circulation. It also includes the potential for methane release especially from the East Siberian shelf (i.e. Shakhova, N. et al, various papers). Significant increase in melt from the Greenland Ice Sheet (and further away in the Antarctic Ice Sheet) with its attendant potential for sea level rise may play a role in the near future. Currently ice sheet gain (in snow fall) is more or less offsetting ice sheet loss (mostly it seems at the grounding line) in both sheets (although less so at Greenland). All of these are in turn capable of precipitating their own climatic feedbacks. There is however not enough space in this project to discuss these in detail.

Recent events

Svalbard airport temperatures are showing significant signs of now being out of the IPCC RCP ranges and temperatures there have been setting records every month since 2010 (NOAA, Global Climate Report – May 2018).

Svalbard, Greenland and the North Pole recorded temperatures that were up to 30 °C higher than expected in the January just past (2018) with all being above freezing point at various times during that month (Meyer, A. et al, 2018). These were all unprecedented in the respective records to that date.

Many maximum temperature records have been set in the last few weeks on or near the Arctic coast in Scandanavia, Russia and Siberia (UNRIC, 2018 for Scandanavia and Russia, various reports in news media sourcing the University of Maine’s climate reanalyzer at for Siberia, eg Samenow, 2018).

It is worth noting therefore that the last IPCC report (AR5) scenarios are conservative, even the RCP8.5 one, and that there are opposing views from climate scientists who expect impacts to be both more extensive and rapid (e.g. Hansen, J.E. et al 2016, Matthews, J.B. and Matthews, J.B.R., 2014). It may be, as it has been before, that the lower boundaries of the IPCC scenarios are tested. In regards to modelling a summer ice free Arctic there are many and some of them have that state (a summer ice free Arctic) happening as early as 2020.

Models: The end of summer sea ice extent.

The date at which the Arctic will be free of summer sea ice is much modelled in the climate science community concerned with it. In Overland, J.E.and Wung (2013) after observing that “Observations and citations support the conclusion that most global climate model results in the CMIP5 [World Climate Research Programme Coupled Model Intercomparison Project Phase 5] archive are too conservative in their sea ice projections.” divide the modelling into three categories;

1) Extrapolation of sea ice volume data. 2) Assumption a short series of rapid loss events. 3) Climate model projections.

The three methods, based on literature that attend to these different methods, all come back with a different timeframe for a Summer sea ice free Arctic but even the most optimistic (in regards to the continued existence of sea ice) of the outcomes as outlined by Overland, J.E. and Wung, M. (the third one – ‘Climate model projections’) has the final collapse of summer sea ice at ~2040. The least optimistic has it the collapse of summer sea ice at ~2020 (this being the first one – ‘Extrapolation of sea ice volume data’).

Stroeve, J.C. et al (2012) in their analysis of the trends of the participants in the CMIP3 modeled on the ICPP RCP4.5 emission scenario that itself is looking increasing unlikely (and may require negative emissions) had 33% of the participant’s models showing no summer sea ice by 2100, with one of the models, as in Overland, J.E.and Wung. M.’s,, shows a collapse of summer sea ice at 2020. Perhaps more interestingly currently observed extents were trending to the earlier losses of summer sea ice extent rather than to the later ones.

Figure 12: Time-series of modeled (colored lines) and observed (solid red line) September sea ice extent from 1900 to 2100. All 56 individual ensemble members from 20 CMIP5 models are included as dotted colored lines, with their individual model ensemble means in solid color lines. The multi-model ensemble mean is based on 38 ensemble members from 17 CMIP5 models (shown in black), with +/ 1 standard deviation shown as dotted black lines. Credit: Stroeve, J.C. et al (2012).

Political will

In the last decade restrictions (2011) were placed on United States NASA scientists working with Chinese counterparts (Witze, A., 2017) and public servants in the United States were told not to mention ‘climate change’ (NBC News, 2015, Brook, B., 2017). Recent elections in the United States have resulted in that country withdrawing from the Paris agreement (although one could argue that this just formalizes the lack of action of the former administration). Other countries have similar outcomes even if presenting differently.

It is therefore difficult to see the level of changes that are needed (which will soon require negative emissions) being put in place. As a result the outlook for Arctic sea ice remains bleak as the anthropogenic forcing remains in place and even looks to be at record levels as evidenced at Mauna Loa.

Figure 13: Annual mean growth rate of CO2 (ppm)at Mauna Loa. Credit: NOAA.


It seems, lacking politcal and societal changes, that we are on track for what is commonly called the ‘Business as Usual’ scenario under our present structures (the RCP8.5 scenario). It implies that several degrees of warming by the end of the current century is the most likely outcome of such forcing (in the range of 2.3 C to 5.0 C). This though is a global average and as the Arctic is warming roughly twice as fast as anywhere else it seems that the Arctic is therefore very likely to see even larger increases in temperatures. Indeed with decreasing ice and increasing heat retention in the ocean, especially during the summer, the potential late autumn/early winter heat fluxes from the ocean to the atmosphere will only increase. And the possibilty of even higher temperatures both in the Arctic and elsewhere are definitely not out of the question.

The final collapse of Arctic summer sea ice in the end may not appear as an extension of a linear line on a graph and instead could appear suddenly, dropping off the chart. Such an event that would likely play out as part of an Artic storm. In fact the Kara Sea sea ice is currently in a fairly bad state and may not even need a storm to achieve such a sea ice free status this year.

It seems likely that the Beaufort Gyre will reverse in the next few years. But the ice and cold fresh water that it would eject into the North Atlantic may just play an ameliorating effect on a much warmer climate in both North America and Northern Europe.

The chances of WACC appearing seem unlikely in the face of a warming that is equal to it or, more likely, stronger. If it does become a feature it is most likely to be in Central Asia.

But with an upwards trend in cyclones expected in the Arctic it becomes entirely plausible that the first sea ice free summer, expected in the next few decades, will come about as a result of a similar storm and that such an outcome would likely be in a timeframe that is unexpected. It is the writers view that such an event is becoming more and more likely and that therefore we should expect an summer ice free Arctic (much) sooner rather than later.

Acknowledgements Nicolas Cullen for his help and guidance on this project. My daughter for her excellent help and suggestions on grammar and style. My wife for putting up with a lack of a second income over the last 5 months. Sean Fitzsimons and Michelle Thompson-Fawcett for their patience.


#climatechange #science #arctic #richardseagernz #otagogeography

Enter your email to subscribe to updates.