Sea ice tracking low in both hemispheres

January of 2018 began and ended with satellite-era record lows in Arctic sea ice extent, resulting in a new record low for the month. Combined with low ice extent in the Antarctic, global sea ice extent is also at a record low.

Overview of conditions

sea ice extent map

Figure 1. Arctic sea ice extent for January 2018 was 13.06 million square kilometers (5.04 million square miles). The magenta line shows the 1981 to 2010 average extent for that month. Sea Ice Index data. About the data

Credit: National Snow and Ice Data Center
High-resolution image

The new year was heralded by a week of record low daily ice extents, with the January average beating out 2017 for a new record low. Ice grew through the month at near-average rates, and in the middle of the month daily extents were higher than for 2017. However, by the end of January, extent was again tracking below 2017. The monthly average extent of 13.06 million square kilometers (5.04 million square miles) was 1.36 million square kilometers (525,000 square miles) below the 1981 to 2010 average, and 110,000 square kilometers (42,500 square miles) below the previous record low monthly average in 2017.

The pattern seen in previous months continued, with below average extent in the Barents and Kara Seas, as well as within the Bering Sea. The ice edge remained nearly constant throughout the month within the Barents Sea, and slightly retreated in the East Greenland Sea. By contrast, extent increased in the Gulf of St. Lawrence, off the coast of Newfoundland, in the eastern Bering Sea and the Sea of Okhotsk. Compared to 2017, at the end of the month, ice was less extensive in the western Bering Sea, the Sea of Okhotsk and north of Svalbard, more extensive in the eastern Bering Sea and in the Gulf of St. Lawrence. Overall, the Arctic gained 1.42 million square kilometers (548,000 square miles) of ice during January 2018.

Conditions in context

extent timeseries

Figure 2a. The graph above shows Arctic sea ice extent as of February 5, 2018, along with daily ice extent data for five previous years. 2017 to 2018 is shown in blue, 2016 to 2017 in green, 2015 to 2016 in orange, 2014 to 2015 in brown, 2013 to 2014 in purple, and 2012 to 2013 in dotted brown. The 1981 to 2010 median is in dark gray. The gray areas around the median line show the interquartile and interdecile ranges of the data. Sea Ice Index data.

Credit: National Snow and Ice Data Center
High-resolution image

airtemp

Figure 2b. The plot shows air temperatures in the Arctic as difference from average for January 2018. Yellows, oranges, and reds indicate higher than average temperatures; blues and purples indicate lower than average temperatures.

Credit: NSIDC courtesy NOAA/ESRL Physical Sciences Division
High-resolution image

Air temperatures at the 925 hPa level (about 2,500 feet above sea level) remained unusually high over the Arctic Ocean (Figure 2b). Nearly all of the region was at least 3 degrees Celsius (5 degrees Fahrenheit) or more above average. The largest departures from average of more than 9 degrees Celsius (16 degrees Fahrenheit) were over the Kara and Barents Seas, centered near Svalbard. On the Pacific side, air temperatures were about 5 degrees Celsius (9 degrees Fahrenheit) above average. By contrast, 925 hPa temperatures over Siberia were up to 4 degrees Celsius (7 degrees Fahrenheit) below average. The warmth over the Arctic Ocean appears to result partly from a pattern of atmospheric circulation bringing in southerly air, and partly from the release of heat into the atmosphere from open water areas. Sea level pressure was higher than average over the central Arctic Ocean, stretching towards Siberia. This pattern, coupled with below average sea level pressure over the Chukchi and Bering seas, helped to move warm air from Eurasia over the central Arctic Ocean.

Ice growth for January averaged 37,000 square kilometers (14,000 square miles) per day, close to the average rate for the month of 42,700 square kilometers per day (16,486 square miles per day). In the Barents Sea, the ice extent was the second lowest during the satellite data record. Ice conditions in this region of the Arctic are increasingly viewed as important in having downstream effects on atmospheric circulation. These proposed links include northward expansion of the Siberian High and cooling over northern Eurasia.

January 2018 compared to previous years

extent trend graph

Figure 3. Monthly January ice extent for 1979 to 2018 shows a decline of 3.3 percent per decade.

Credit: National Snow and Ice Data Center
High-resolution image

The linear rate of decline for January is 47,700 square kilometers (18,400 square miles) per year, or 3.3 percent per decade.

Engaging stakeholders in sea ice forecasting

tourism graphs

Figure 4. These graphs show changes in polar tourism based on membership in the Association of Arctic Expedition Cruise Operators (AECO, top), and by the number and type of Arctic vessels operated or managed (bottom).

Credit: Kelvin Murray, Director, Expedition Operations EYOS Expeditions
High-resolution image

Uncertainty about future sea ice conditions presents challenges to industry, policymakers, and planners responsible for economic, safety, and risk mitigation decisions. The ability to accurately forecast the extent and duration of sea ice on different timescales is relevant to a wide range of Arctic maritime activities. While there have been considerable advances in sea ice forecasting over the past decade, it remains unclear how well end users are able to utilize these products and services in their planning. In response, the Sea Ice Prediction Network, in collaboration with several sponsors, held a workshop at the Arctic Frontiers Conference in Tromsø, Norway to foster dialogue between stakeholders and sea ice forecasters.

Conference attendees recognized that the sea ice forecasting community and users of these forecasts need a common language. Often forecast users do not understand the data presented by forecasters, nor do they have the skills to interpret the complex data products. Most marine operators in the Arctic require accurate daily to short-term (< 72 hours) information on the sea ice edge and near-ice-edge concentration. Forecast users often want additional information such as ice strength, ice thickness and ice drift. These data need to be accessed in a user-friendly format that can be easily downloaded (e.g., to a ship at sea). Typically, ice charts from national ice centers or high-resolution Synthetic Aperture Radar image maps are used for describing and analyzing sea ice for real-time navigation.

Longer-term seasonal ice forecasts are potentially useful to the polar marine industry but are not yet being relied upon. While improving, the uncertainty in these forecasts has not been clearly communicated. Nevertheless, logistics planners are interested in using longer-term forecasts, mostly to augment or extend more timely data or in-house diagnostics. Tour operators in particular desire seasonal and even two- to three-year forecasts so that they can plan what to offer their customers. Along with the increase in polar tourism (Figure 4), there is also significant industry traffic in the European Arctic, the Northwest Passage and some areas in the Northern Sea Route. Due to the decreasing ice cover, we can expect an extension of the seasonal activity, with ships embarking earlier and ending their journeys later than in previous years. This underscores the need for accurate forecasting, extending to the more variable shoulder seasons of Arctic sea ice.

Importance of ice drift

Figure 5. The top figure shows the location of the R/V Lance during the N-ICE2015 expedition (pink lines) with aircraft flight lines shown in black and blue. The bottom figure shows a time series of wind speed and direction, together with rates of ice divergence (blue line) and shear (purple line). Figure from Itkin et al. 2017.

Credit: Norwegian Polar Institute
High-resolution image

As the Arctic sea ice cover continues to thin, convergent sea ice motion can more readily pile up ice into large ridges. Such ridges can be hazardous to marine activities in the Arctic. Divergent ice motion produces openings in the ice called leads, where new ice can readily grow. Winds are the main driver for both ridging and lead formation. A single storm event can lead to significant redistribution of sea ice mass through ridging and new leads. As part of the Norwegian Young Sea ICE (N-ICE2015) expedition, colleagues at the Norwegian Polar Institute made detailed sea ice thickness and ice drift observations before and after a storm in an area north of Svalbard (Figure 5). Results showed that about 1.3 percent of the level sea ice volume was pressed together into ridges. Combined with new ice formation in leads, the overall ice volume increased by 0.5 percent. While this is a small number, sea ice in the North Atlantic is typically impacted by 10 to 20 storms each winter, which could account for 5 to 10 percent of ice volume each year.

Antarctic sea ice also low, leading to low global sea ice extent

In the Southern Hemisphere, after January 11 sea ice began tracking low, leading to a January average extent that was the second lowest on record. The lowest extent for this time of year was in 2017. Extent is below average in the Ross Sea and the West Amundsen Seas, while elsewhere extent remains close to average. The low ice extent is puzzling, given that air temperatures at the 925 hPa level are near average or below average (relative to the 1981 to 2010 period) over much of the Southern Ocean. The Weddell and Amundsen Seas were 1 to 2 degrees Celsius (2 to 4 degrees Fahrenheit) below average. Slightly above-average temperatures were the rule in the northwestern Ross Sea.

Further reading

Itkin, P., Spreen, G., Hvidegaard, S. M., Skourup, H., Wilkinson, J., Gerland, S., & Granskog, M. A. 2018. Contribution of deformation to sea ice mass balance: A case study from an N-ICE2015 storm. Geophysical Research Letters, 45. https://doi.org/10.1002/2017GL076056.

 

Baked Alaska and 2017 in review

Arctic sea ice extent in December 2017 was below average in both the far northern Atlantic and the Bering Sea, and notably high temperatures prevailed over most of the Arctic, especially over Central Alaska. We look back at the year’s events, and examine Arctic sea ice trends since 1850 based on a new compilation of data from maps, ship reports, and other records.

Overview of conditions

Figure 1. Arctic sea ice extent for December 2016 was 11.75 million square kilometers (4.54 million square miles). The magenta line shows the 1981 to 2010 average extent for that month.

Figure 1. Arctic sea ice extent for December 2017 was 11.75 million square kilometers (4.54 million square miles). The magenta line shows the 1981 to 2010 average extent for that month. Sea Ice Index data. About the data

Credit: National Snow and Ice Data Center
High-resolution image

Arctic sea ice extent for December 2017 averaged 11.75 million square kilometers (4.54 million square miles), the second lowest in the 1979 to 2017 satellite record. This was 1.09 million square kilometers (420,900 square miles) below the 1981 to 2010 average and 280,000 square kilometers (108,100 square miles) above the record low December extent recorded in 2016. Extent at the end of the month was below average in the far northern Atlantic Ocean and Barents Sea, slightly above average in western Hudson Bay, and continued to be below average in the Bering and Chukchi Seas. Near-average conditions prevailed along the eastern coast of Greenland and in the Sea of Okhotsk.

Conditions in context

Figure 2. The graph above shows Arctic sea ice extent as of January 2, 2018, along with daily ice extent data for four previous years. 2017 to 2018 is shown in blue, 2016 to 2017 in green, 2015 to 2014 in orange, 2014 to 2015 in brown, 2013 to 2014 in purple, and 2012 to 2012 in dotted brown. The 1981 to 2010 median is in dark gray. The gray areas around the median line show the interquartile and interdecile ranges of the data.

Figure 2a. The graph above shows Arctic sea ice extent as of January 2, 2018, along with daily ice extent data for five previous years. 2017 to 2018 is shown in blue, 2016 to 2017 in green, 2015 to 2014 in orange, 2014 to 2015 in brown, 2013 to 2014 in purple, and 2012 to 2012 in dotted brown. The 1981 to 2010 median is in dark gray. The gray areas around the median line show the interquartile and interdecile ranges of the data. Sea Ice Index data.

Credit: National Snow and Ice Data Center
High-resolution image

Figure 2b. This plot shows the departure from average air temperatures at the 925 hPa level in degrees Celsius for December 2017. Yellows and reds indicate higher than average temperatures; blues and purples indicate lower than average temperatures.||Credit: NSIDC courtesy NOAA Earth System Research Laboratory Physical Sciences Division

Figure 2b. This plot shows the departure from average air temperatures at the 925 hPa level in degrees Celsius for December 2017. Yellows and reds indicate higher than average temperatures; blues and purples indicate lower than average temperatures.

Credit: NSIDC courtesy NOAA Earth System Research Laboratory Physical Sciences Division
High-resolution image

Figure 2c. This plot shows the departure from average sea level pressures at the 925 hPa level in degrees Celsius for December 2018. Yellows and reds indicate higher than average air pressures; blues and purples indicate lower than average air pressures.|

Figure 2c. This plot shows the departure from average sea level pressures at the 925 hPa level in degrees Celsius for December 2018. Yellows and reds indicate higher than average air pressures; blues and purples indicate lower than average air pressures.

Credit: NSIDC courtesy NOAA Earth System Research Laboratory Physical Sciences Division
High-resolution image

Ice growth during December 2017 averaged 59,800 square kilometers (23,100 square miles) per day. This was fairly close to the average rate for the month of 64,100 square kilometers (24,800 square miles) per day. Ice growth in the Chukchi Sea (very late compared to previous years), the Kara Sea, and the eastern Hudson Bay areas were the main regions of change in December. In contrast, the ice edge slightly retreated in the Barents Sea near Franz Josef Land.

December air temperatures at the 925 hPa level (about 2,500 feet above sea level) throughout the Arctic Ocean were 2 to 6 degrees Celsius (4 to 11 degrees Fahrenheit) above average. Prominent warm spots were found over north Central Asia and Central Alaska (more than 10 degrees Celsius, or 18 degrees Fahrenheit above average), as well as over Svalbard and Central Siberia (nearly 6 degrees Celsius or 11 degrees Fahrenheit above average). Temperatures were 2 to 3 degrees Celsius (4 to 5 degrees Fahrenheit) below average in Eastern Siberia.

The air temperature pattern in December was similar to that seen in November, driven in part by the arrangement of high and low air pressure regions surrounding the Arctic. Below-average pressure over easternmost Siberia and above-average pressure over the Gulf of Alaska drove southwesterly winds into Central Alaska and the Yukon region. Warmth in the Central Arctic and in Svalbard was consistent with southerly winds arising from low pressure over Scandinavia and higher pressure in the Laptev Sea and Central Siberia.

The Arctic Oscillation (AO) is a key climate indicator for general wind, precipitation, and temperature patterns in the Arctic. The AO index was moderately positive through most of 2017, indicating a tendency toward strong circumpolar winds at high latitude and warm conditions in the mid-latitudes. December 2017 had a mix of conditions, resulting in a near-neutral AO state (as measured by the index).

December 2017 compared to previous years

Figure 3. Monthly December ice extent for 1979 to 2017 shows a decline of 3.7 percent per decade.

Figure 3. Monthly December ice extent for 1979 to 2017 shows a decline of 3.7 percent per decade.

Credit: National Snow and Ice Data Center
High-resolution image

The linear rate of sea ice decline for December is 47,400 square kilometers (18,300 square miles) per year or 3.7 percent per decade. Recall from our previous post that NSIDC has revised the way in which monthly average extents are computed, which has some impacts on computed trends.

2017 year in review

Figure 4. These figures show trends for sea ice-over dates in the Beaufort (top) and Chukchi (bottom) Seas.

Figure 4. These figures show trends for ice-over dates in the Beaufort (top) and Chukchi (bottom) Seas. Sea Ice Index data.

Credit: R. Thoman, NOAA
High-resolution image

The winter of 2016 to 2017 saw record low winter sea ice extent and higher than average temperatures. Indeed, the first four months of 2017 set or tied record low extents for the month. However, the melt season progressed somewhat slowly from May through July, as storminess and relatively cool conditions began to prevail. As such, sea ice extent at the seasonal minimum, on September 13, ended up as eighth lowest.

Assessments of sea ice thickness modeled by the Pan-Arctic Ice Ocean Modeling and Assimilation System (PIOMAS), as well as sea ice age near the seasonal minimum extent indicate that Arctic sea ice remains very low in overall volume. As the year ended, ice extent remained especially low in the Chukchi and Bering Seas. As discussed in an earlier post, the unusually early seasonal ice retreat in the Chukchi Sea this past summer likely relates to a strong inflow of oceanic heat into the region via the Bering Strait. With more heat in the upper ocean at summer’s end, it takes longer for sea ice to form in autumn and winter. Colleague Rick Thoman of the National Oceanic and Atmospheric Administration (NOAA) National Weather Service has assembled a time series of the ice-over dates in both the Chukchi and Beaufort Seas based on the satellite passive microwave record (Figure 4). The ice-over date is defined as the first day that the ice concentration exceeds 95 percent in the region. The trends towards later freeze up in both seas is striking. This has an impact on sea ice thickness as the growth season is shortened, which may lead to thinner ice by the end of winter. On the other hand, later freeze up also means less time for snow accumulation on the sea ice. Since sea ice grows faster for a thinner snowpack, this may partially offset the impacts of late ice formation.

A longer record of Arctic sea ice extent

Figure 5a. This figure shows departures from 1850 to 2013 calendar-month averages of Arctic sea ice extent as a function of year (x-axis) and calendar month (y-axis). The color bar at the right shows magnitudes of departures from the average.

Figure 5a. This figure shows departures from 1850 to 2013 calendar-month averages of Arctic sea ice extent as a function of year (x-axis) and calendar month (y-axis). The color bar at the right shows magnitudes of departures from the average.

Credit: J. E. Walsh, F. Fetterer, J. S. Stewart, W. L. Chapman. 2016. Geographical Review; after a figure by J. Stroeve, National Snow and Ice Data Center
High-resolution image

Figure 5b. These sea ice concentration maps compare the lowest September minimum Arctic sea ice extents for the periods 1850 to 1900, 1901 to 1950, 1951 to 2000, and 2000 to 2013.||Credit: F. Fetterer/National Snow and Ice Data Center, NOAA

Figure 5b. These sea ice concentration maps compare the lowest September minimum Arctic sea ice extents for the periods 1850 to 1900, 1901 to 1950, 1951 to 2000, and 2000 to 2013.

Credit: F. Fetterer/National Snow and Ice Data Center, NOAA
High-resolution image

Using a compilation of maps, ship reports, and other records, NOAA has published monthly estimates of Arctic sea ice extent spanning 1850 to 2013. While data in the earlier part of the record is limited, the carefully constructed time series helps to put the more recent satellite record in a longer-term context. Figure 5a shows the decline in extent over the period of satellite observations standing out prominently in comparison with the rest of the record, especially in late summer and early autumn. An earlier period of unusually low summer sea ice extent around 1937 to 1943 (as compared to the 1850 to 2013 average) did not extend to the winter season, and was followed by a few years of significantly higher-than-average summer ice extents. Early in the record (1850 to 1900), winter ice extent was not particularly elevated relative to the 1850 to 2013 average, but summer sea ice extent was quite a bit higher higher than the average. As another way to place recent conditions into a longer context using this data set, we show the years of the lowest September extent recorded within the 50-year periods 1850 to 1900, 1901 to 1950, 1951 to 2000, along with the lowest over the period 2000 to 2013 (Figure 5b). The decline in extent is apparent.

Low sea ice extent in the Antarctic

Figure 6. Antarctic sea ice extent for December 2017 was 9.34 million square kilometers (3.61 million square miles). The magenta line shows the 1981 to 2010 average extent for that month.

Figure 6. Antarctic sea ice extent for December 2017 was 9.34 million square kilometers (3.61 million square miles). The magenta line shows the 1981 to 2010 average extent for that month. Sea Ice Index data. About the data

Credit: National Snow and Ice Data Center
High-resolution image

In the Southern Hemisphere, sea ice for December 2017 averaged 9.34 million square kilometers (3.61 million square miles) and was the fourth lowest in the satellite record. Sea ice extent was far below average in the eastern Weddell Sea, but above average in the northwestern Weddell Sea. The East Antarctic coastline had near-average ice extent. As the Southern Hemisphere entered into the summer months, sea ice declined steeply. Temperatures at the 925 hPa level were 1 to 3 degrees Celsius (2 to 5 degrees Fahrenheit) higher than average in Dronning Maud Land and the northern Ross Sea, and generally lower than average over the ice sheet. Near-average temperatures have prevailed over the fringing Southern Ocean. Pressures were slightly above average over the continent and below average in the surrounding ocean. Consistent with this pattern, the Southern Annular Mode index, a measure of the strength of westerly winds, was moderately positive for December.

Further reading

Walsh, J. E., F. Fetterer, J. S. Stewart, and W. L. Chapman. 2016. A database for depicting Arctic sea ice variations back to 1850. Geographical Review. doi: 10.1111/j.1931-0846.2016.12195.x.

Record low extent in the Chukchi Sea

November 2017 will be remembered not for total Arctic ice extent, which was the third lowest recorded over the period of satellite observations, but for the record low extent in the Chukchi Sea. This is a key area for Arctic Ocean access, and is an indicator of oceanographic influences on sea ice extent.

Overview of conditions

Figure 1. Arctic sea ice extent for XXXX 20XX was X.XX million square kilometers (X.XX million square miles). The magenta line shows the 1981 to 2010 average extent for that month. Sea Ice Index data. About the data||Credit: National Snow and Ice Data Center|High-resolution image

Figure 1. Arctic sea ice extent for November 2017 was 9.46 million square kilometers (3.65 million square miles). The magenta line shows the 1981 to 2010 average extent for that month. Sea Ice Index data. About the data

Credit: National Snow and Ice Data Center
High-resolution image

Arctic sea ice extent for November 2017 averaged 9.46 million square kilometers (3.65 million square miles), the third lowest in the 1979 to 2017 satellite record. This was 1.24 million square kilometers (479,000 square miles) below the 1981 to 2010 average and 830,000 square kilometers (321,000 square miles) above the record low November extent recorded in 2016. Extent at the end of the month was below average over the Atlantic side of the Arctic, primarily in the Barents and Kara Seas, slightly above average in western Hudson Bay, but far below average in the Chukchi Sea. This continues a pattern of below-average extent in this region that has persisted for the last year.

Conditions in context

timeseries graph

Figure 2. The graph above shows Arctic sea ice extent as of December 3, 2017, along with daily ice extent data for five previous years. 2017 is shown in blue, 2016 in green, 2015 in orange, 2014 in brown, 2013 in purple, and 2012 in dotted red. The 1981 to 2010 median is in dark gray. The gray areas around the median line show the interquartile and interdecile ranges of the data. Sea Ice Index data.

Credit: National Snow and Ice Data Center
High-resolution image

Ice growth during November 2017 averaged 80,100 square kilometers (30,900 square miles) per day. This was stronger than the average rate for the month of 69,600 square kilometers (26,900 square miles) per day. Ice growth was particularly rapid within Hudson Bay, Baffin Bay, and the Kara Sea.

November air temperatures at 925 hPa (about 3,000 feet above sea level) were above average over essentially all of the Arctic Ocean, with prominent warm spots (more than 6 degrees Celsius, or 11 degrees Fahrenheit above the 1981 to 2010 average) over the Chukchi Sea and north of Svalbard. The unusual warmth in the Chukchi Sea at least in part manifests the extensive open water in this region, but a pattern of winds blowing in from the southwest also appears to have had an influence. The warmth north of Svalbard is more clearly related to the average pattern of atmospheric circulation over the month, with an area of low pressure centered over the Norwegian Sea and an area of high pressure centered north of the Taymyr Peninsula combining to transport warm air into the region.

November 2017 compared to previous years

ice extent trend

Figure 3. Monthly November ice extent for 1979 to 2017 shows a decline of 5.14 percent per decade.

Credit: National Snow and Ice Data Center
High-resolution image

The linear rate of sea ice decline for November is 55,000 square kilometers (21,200 square miles) per year, or 5.14 percent per decade. Recall from our previous post that NSIDC recently revised the way in which monthly average extents are computed, which has minor impacts on computed trends.

Open water in the Chukchi Sea

sea ice concentration

Figure 4. The map at top shows an analysis of sea ice concentration on November 30, 2017 in the area of the Bering and Chukchi Seas. The graph at bottom shows the combined sea ice concentration from 1978 to 2017, based on Sea Ice Index data.

Credit: Rick Thoman of the NOAA National Weather Service Alaska Region
High-resolution image

Based on an analysis by Rick Thoman of the NOAA National Weather Service, as of 19 November, ice extent in the combined Beaufort and Chukchi Seas sector was the lowest ever observed in the sea ice record (Figure 4). This was largely driven by the lack of sea ice within the Chukchi Sea. By the end of November, the Beaufort Sea was completely ice-covered. The NOAA analysis makes use of the NSIDC Sea Ice Index data set. As discussed in our June 7 post, the current state of the ice cover in this region likely has its origin as far back as last year, when warm conditions favored the persistence of open water in the Chukchi Sea into December of 2016.

Strong winds from the north occurred for a few days at the end of March and early April, pushing ice southward in the Bering Sea, breaking up the ice in the Chukchi Sea, and even flushing some ice out through the Bering Strait. We also suggested a possible role of a strong oceanic heat inflow to the Chukchi Sea via Bering Strait. In support of this view, in the summer of 2017, Rebecca Woodgate of the University of Washington, Seattle, sailing on the research vessel Norseman II, recovered mooring data that indicated an early arrival of warm ocean water in the strait, about a month earlier than the average. This resulted in June ocean temperatures that were 3 degrees Celsius (5 degrees Fahrenheit) above average. Higher ocean temperatures in summer plays a large role in the timing of when the ice will form again in winter. There is likely a considerable amount of heat remaining in the top layer of the ocean, which will need to be lost to the atmosphere and outer space before the region becomes fully ice covered.

Low Antarctic sea ice extent

Figure 5a. Antarctic sea ice concentration from AMSR2, in percent, for November 28, 2017. The Maud Rise polynya is seen at top.

Credit: University of Bremen
High-resolution image

Figure 5b. Small tabular icebergs are seen in the marginal ice zone of the northern Weddell Sea on November 22, 2017 during a NASA Operation IceBridge flight.

Credit: NASA/John Sonntag
High-resolution image

In the Southern Hemisphere, where it is late spring, sea ice declined at a faster-than-average pace after the very late-season October 12 maximum extent. This led to the third-lowest November average monthly extent in the satellite record, behind 1986 and 2016. Sea ice extent was near-average in all regions except the Weddell Sea, where extent is at a satellite-era record low.

The atmospheric circulation for November exhibited a very strong wave-3 pattern. In a wave-3 pattern, there are three major low-pressure areas around the continent separated by three high-pressure areas. Air temperatures for the month were near-average in most regions except for the eastern Weddell Sea, consistent with the reduced sea ice extent there.

The Maud Rise Polynya (Figure 5a) continued to grow through November, as increased sunshine and air temperatures allowed the upwelling warm water to expand the opening in the floating sea ice cover. At the beginning of December,  retreat of the sea ice edge converted the polynya to a large embayment in the sea ice cover.

Freezing in the dark

Rapid expansion of the Arctic sea ice cover is the norm for October as solar input dwindles and the remaining heat in the upper ocean is released upwards, warming the lower atmosphere and escaping to space. Because of late season growth, the seasonal Antarctic maximum we previously reported as occurring on September 15 was exceeded, with a new maximum set on October 11 and 12. This is the second-lowest and second-latest seasonal maximum extent in the satellite record.

Overview of conditions

Figure 1. Arctic sea ice extent for October 2017 was6.71 million square kilometers (2.60 million square miles). The magenta line shows the 1981 to 2010 average extent for that month.

Figure 1. Arctic sea ice extent for October 2017 was 6.71 million square kilometers (2.60 million square miles). The magenta line shows the 1981 to 2010 average extent for that month. Sea Ice Index data. About the data

Credit: National Snow and Ice Data Center
High-resolution image

Arctic sea ice extent for October 2017 averaged 6.71 million square kilometers (2.60 million square miles), the fifth lowest in the 1979 to 2017 satellite record. This was 1.64 million square kilometers (633,000 square miles) below the 1981 to 2010 average and 820,000 square kilometers (317,000 square miles) above the record low October extent recorded in 2012. By the end of October, extent remained below average throughout most of the Arctic except within the Laptev Sea, which is fully ice covered. Ice growth over the month was most prominent within the Beaufort, East Siberian, and Laptev Seas and within Baffin Bay. In the Chukchi, Kara, and Barents Seas, the rate of ice growth was slower. Ice extent also remains far below average in the East Greenland Sea.

Conditions in context

Figure 2a. The graph above shows Arctic sea ice extent as of November 2, 2017, along with daily ice extent data for five previous years. 2017 is shown in blue, 2016 in green, 2015 in orange, 2014 in brown, 2013 in purple, and 2012 in dotted brown. The 1981 to 2010 median is in dark gray. The gray areas around the median line show the interquartile and interdecile ranges of the data.

Figure 2a. The graph above shows Arctic sea ice extent as of November 2, 2017 along with daily ice extent data for five previous years. 2017 is shown in blue, 2016 in green, 2015 in orange, 2014 in brown, 2013 in purple, and 2012 in dotted brown. The 1981 to 2010 median is in dark gray. The gray areas around the median line show the interquartile and interdecile ranges of the data. Sea Ice Index data.

Credit: National Snow and Ice Data Center
High-resolution image

Figure 2b. This plot shows Arctic air temperature anomalies at the 925 hPa level in degrees Celsius for October 2017. Yellows and reds indicate higher than average temperatures; blues and purples indicate lower than average temperatures.

Figure 2b. This plot shows Arctic air temperature anomalies at the 925 hPa level in degrees Celsius for October 2017. Yellows and reds indicate higher than average temperatures; blues and purples indicate lower than average temperatures.

Credit: NSIDC courtesy NOAA Earth System Research Laboratory Physical Sciences Division
High-resolution image

Figure 2c. This plot shows Arctic air temperatures as a function of both height and latitudes. Above average air temperatures for the Arctic as a whole extend up to approximately 9,200 meters (30,000 feet) high in the atmosphere. Colors indicate temperatures in degrees Celsius. Yellows and reds indicate higher than average temperatures; blues and purples indicate lower than average temperatures.

Figure 2c. This plot shows Arctic air temperatures as a function of height and latitudes. Above average air temperatures for the Arctic as a whole extend up to approximately 9,200 meters (30,000 feet) altitude. Yellows and reds indicate higher than average temperatures; blues and purples indicate lower than average temperatures.

Credit: NSIDC courtesy NOAA Earth System Research Laboratory Physical Sciences Division
High-resolution image

Ice growth during October 2017 averaged 94,200 square kilometers (36,000 square miles) per day. This was 5,100 square kilometers (2,000 square miles) per day faster than the average rate of ice growth for the month. Total ice extent for the month remains more than 2 standard deviations below the 1981 to 2010 average.

October air temperatures at 925 hPa (about 3,000 feet above sea level) were 2 to 4 degrees Celsius (4 to 7 degrees Fahrenheit) above average over most of the Arctic Ocean and up to 7 degrees Celsius (13 degrees Fahrenheit) above average over the East Greenland Sea. Unusually high temperatures over the East Greenland Sea appear to largely reflect the transport of warm air from Eurasia, driven by the combination of above average sea level pressure over the Kara and Barents Seas, and below average pressure over the North Atlantic and Greenland. Elsewhere, above average near surface air temperatures reflect in part the exchange of heat from the ocean to the atmosphere as the ocean cools and sea ice forms, such as within the Chukchi Sea. A plot of temperatures as a function of both height and latitudes shows that the above average air temperatures for the Arctic as a whole extend up to approximately 9,200 meters (30,000 feet) high in the atmosphere.

October 2017 compared to previous years

Figure 3. Monthly October ice extent for 1979 to 2017 shows a decline of 9.3 percent per decade.

Figure 3. Monthly October ice extent for 1979 to 2017 shows a decline of 9.3 percent per decade.

Credit: National Snow and Ice Data Center
High-resolution image

The linear rate of sea ice decline for October is 77,600 square kilometers (30,000 square miles) per year, or 9.3 percent per decade relative to the 1981 to 2010 average. While this appears as an increase in the rate of October ice retreat compared to the trend reported last year, it is not a climate signal but is rather largely a result of using a different averaging method to derive the monthly average sea ice extent values (see below).

Effects of snow salinity on CryoSat-2 ice freeboard estimates

Figure 4. This schematic illustrates how salinity shifts the source of the radar signature in the icepack.

Figure 4. This schematic illustrates how salinity shifts the source of the radar signature in the icepack. Ice thickness can be over-estimated by radar satellites (CryoSat-2) when snow conditions are more saline.

Credit: V. Nandan
High-resolution image

After the end of sea ice melt season, the ocean cools and new sea ice forms. The ice crystals that form expel salt into the water. Some of this salt, or brine, is also expelled upwards to the surface of the ice or into snow that has fallen since the ice formed. The brine is then wicked upwards into the snowpack, leading to a slightly saline snowpack, ranging from 1 to 20 parts per thousand (standard seawater is about 35 parts per thousand). This saline snow is a strong reflector of radar energy.

A recent study led by the Cryosphere Climate Research Group at the University of Calgary investigated the impact of snow salinity on retrieving sea ice thickness from radar altimeters, such as CryoSat-2. The study shows that the snow layers observed over much of the Arctic’s first-year ice are salty enough to reflect the radar pulse from CryoSat-2, a radar altimeter used to measure sea ice thickness and ice sheet elevation. They calculate that a correction factor could compensate for this effect, and improve sea ice thickness measurements. While snow salinity is important, other factors, such as surface roughness and ice density also contribute to uncertainties in ice thickness, and they can potentially cancel each other out. Continued comparisons to observed thickness data is crucial to better quantify these uncertainties.

Antarctica’s double-humped sea ice maximum

Figure 5. This graph shows the first and second peaks in extent during the 2017 Antarctic sea ice freeze up.

Figure 5a. This graph shows the first and second peaks in extent during the 2017 Antarctic sea ice freeze up. The extent line for the year 2002 is also shown and has a similar pattern to 2017. Sea Ice Index data. About the data

Credit: National Snow and Ice Data Center
High-resolution image

Figure 5b. This map shows Antarctic sea ice concentration on October 31, 2017. Note the Maud Rise polynya at the top of the image. Data are from the Advanced Microwave Scannig Radiometer 2 (AMSR2)||Credit: Institute of Environmental Physics, University of Bremen|High-resolution image

Figure 5b. This map shows Antarctic sea ice concentration on October 31, 2017. Note the Maud Rise polynya at the top of the image. Data are from the Advanced Microwave Scanning Radiometer 2 (AMSR2).

Credit: Institute of Environmental Physics, University of Bremen
High-resolution image

In our last post, we noted that Antarctic sea ice may have reached its maximum extent for the year on September 15, at 17.98 million square kilometers (6.94 million square miles). However, after two weeks of decline, extent increased again reaching a second and final maximum of 18.03 million square kilometers (6.96 million square miles) on October 11 and 12. This is tied with 2002 for the latest maximum on record and is the second lowest Antarctic maximum extent in the satellite data record, slightly higher than 1986. Interestingly, 2002 had a similar Bactrian maximum pattern.

The Maud Rise polynya (or Weddell Sea polynya) continues to be a significant feature of the sea ice cover near 5°E longitude and 65°S. The feature appeared around September 13 and grew to its approximate current extent by September 17. Its current size remains about 30,000 square kilometers (12,000 square miles).

Winds and ocean temperatures continue to drive Antarctic sea ice variability. Since there is no land boundary to the north of the Antarctic continent, sea ice in the Southern Hemisphere is free to expand toward the equator until it reaches water temperatures that are high enough to melt sea ice. As a result, changes in winds or ocean temperatures can have a large influence on the amount of sea ice year to year. Changes in winds related to the positive phase of the Southern Annular Mode (SAM) appear to explain the positive trend in total Antarctic sea ice extent. When the SAM is in a positive phase during austral summer, stronger than average westerly winds blow around the Antarctic continent, and sea ice is pushed both westward and slightly northward due to the Coriolis effect. In addition, below average sea surface temperatures persist through the summer and lead to increased sea ice growth the following autumn, while the negative phase precedes higher sea surface temperatures and reduced sea ice growth. A new study suggests the negative SAM mode during 2016/2017 austral summer largely explained the record minimum Antarctic sea ice extent observed in March 2017.

Revised computation of the monthly mean extent

Figure 6. This chart compares the monthly October Arctic sea ice extents generated from the old (black dashed line) and the new (solid black line) averaging method.

Figure 6. This chart compares the monthly October Arctic sea ice extents generated from the old (black dashed line) and the new (solid black line) averaging method. Sea Ice Index data. About the data

Credit: National Snow and Ice Data Center
High-resolution image

We have updated the way the monthly average sea ice extent is calculated in the NSIDC Sea Ice Index, the source for our sea ice extent estimates. The monthly average total extent (and area) are now computed as an average of the daily values over the month. Historically, the monthly mean sea ice extent has been calculated based on the monthly mean averaged sea ice concentration field. While there is a rationale for both approaches, the new method is more intuitive and eliminates unusual and unexpected results in months when there is rapid ice growth and retreat. Most of the new monthly mean extents are smaller than the previous values with a mean extent difference between -0.45+0.24 and -0.23+0.16 million square kilometers for the Arctic and Antarctic, respectively. The largest differences for the Arctic occur during the month of October due to the rapid ice growth rates typical at that time of year, with the largest difference of -1.20 million square kilometers in October 2012. Changes in rankings and trends were much smaller because the new method tends to affect all years of a given month in a similar manner. October is also the month with the largest trend difference, increasing in magnitude from -7.4 percent per decade to -9.3 percent per decade. Changes in Arctic trends for other months are much smaller.

Similarly, in the Antarctic, differences in averaging methods results in the largest changes during the month of December when the ice cover is rapidly receding. The largest difference of -1.27 million square kilometers occurs in December 1981. The largest changes in the trends are for January and December with a change in value from +2.7 to +3.5 and +1.2 to +1.9 percent per decade, respectively. For more detailed information on the impacts of the revised averaging methods on trends and rankings, please see NSIDC Special Report 19.

Further reading

Nandan, V., T. Geldsetzer, J. Yackel, M. Mahmud, R. Scharien, S. Howell, J. King, R. Ricker, and B. Else. 2017. Effect of snow salinity on CryoSat-2 Arctic first-year sea ice freeboard measurements: Sea ice brine-snow effect on CryoSat-2. Geophysical Research Lettersdoi:10.1002/2017GL074506.

Doddridge, E. W. and J. Marshall. 2017. Modulation of the seasonal cycle of Antarctic sea ice extent related to the Southern Annular Mode. Geophysical Research Letters, 44, 9761–9768. doi: 10.1002/2017GL074319.

Windnagel, A., M. Brandt, F. Fetterer, and W. Meier. 2017. Sea Ice Index Version 3 Analysis. NSIDC Special Report 19. https://nsidc.org/sites/nsidc.org/files/files/NSIDC-special-report-19.pdf.

Arctic sea ice 2017: Tapping the brakes in September

After setting a record low seasonal maximum in early March, Arctic sea ice extent continued to track low through July. However, the rate of ice loss slowed in August and September. The daily minimum extent, reached on September 13, was the eighth lowest on record, while the monthly average extent was seventh lowest. In Antarctica, sea ice extent may have reached its annual winter maximum.

Overview of conditions

ice extent image

Figure 1. Arctic sea ice extent for September 2017 was 4.87 million square kilometers (1.88 million square miles). The magenta line shows the 1981 to 2010 average extent for that month. Sea Ice Index data. About the data

Credit: National Snow and Ice Data Center
High-resolution image

Arctic sea ice extent for September 2017 averaged 4.87 million square kilometers (1.88 million square miles), the seventh lowest in the 1979 to 2017 satellite record. This was 1.67 million square kilometers (645,000 square miles) below the 1981 to 2010 average, and 1.24 million square kilometers (479,000 square miles) above the record low September set in 2012.

After reaching the minimum on September 13 (eighth lowest on record), extent initially increased slowly (about 20,000 square kilometers, or 8,000 square miles, per day). However, starting September 26 and persisting through the end of the month, ice growth rates increased to about 60,000 square kilometers (23,000 square miles) per day. During the second half of the month, extent increased in all sectors except in the Beaufort Sea, where some local ice retreat persisted. The most rapid growth occurred along the Siberian side of the Arctic Ocean, where the ice edge advanced as much as 150 kilometers (90 miles) over the latter half of September. At the end of September, the ice edge in the Beaufort and Chukchi Seas remained considerably further north than is typical.

Conditions in context

extent timeseries

Figure 2a. The graph above shows Arctic sea ice extent as of October 4, 2017, along with daily ice extent data for five previous years. 2017 is shown in blue, 2016 in green, 2015 in orange, 2014 in brown, 2013 in purple, and 2012 in dotted brown. The 1981 to 2010 median is in dark gray. The gray areas around the median line show the interquartile and interdecile ranges of the data. Sea Ice Index data.

Credit: National Snow and Ice Data Center
High-resolution image

pressure anomaly

Figure 2b. This image shows the departure from average sea level pressure in millibars over the Arctic for June, July, and August in 2017. Yellows and reds indicate higher than average pressures; blues and purples indicate lower than average pressures.

Credit: NSIDC courtesy NOAA/ESRL Physical Sciences Division
High-resolution image

September air temperatures at the 925 hPa level (approximately 2,500 feet above sea level) were above average over much of the Arctic Ocean. Temperatures ranged from 5 degrees Celsius (9 degrees Fahrenheit) above the 1981 to 2010 long term average in the far northern Atlantic east of Greenland, to 1 to 2 degrees Celsius (2 to 4 degrees Fahrenheit) above the reference period in the western Arctic. Cooler conditions (1 degree Celsius or 2 degrees Fahrenheit below average) were present in Baffin Bay. Part of the above average temperatures over the coastal areas of the Arctic Ocean and in the northern North Atlantic likely results from heat fluxes from open water.

Looking back at this past summer (June through August), air temperatures at the 925 hPa level averaged for June through August were near or below the 1981 to 2010 average over much of the Arctic Ocean, notably along the Siberian side centered over the Laptev Sea (1 degree Celsius or 1.8 degrees Fahrenheit below the 1981 to 2010 average). By contrast, temperatures were slightly above average over much of the East Siberian, Chukchi and Beaufort Seas (1 degree Celsius, or 1.8 degrees Fahrenheit above average).

Like 2016, the summer of 2017 was characterized by persistently stormy patterns over the central Arctic Ocean, reflected in the summer average sea level pressure field (Figure 2b) as an area of low pressure centered just south of the North Pole in the Siberian sector of the Arctic. As has been shown in past studies, low pressure systems found over the central Arctic Ocean in summer are typically “cold cored.” This helps to explain the cool summer temperatures noted above. The cyclonic (counterclockwise) winds associated with the stormy pattern also tend to spread out the sea ice. Both processes likely helped to slow sea ice loss this summer.

September 2017 compared to previous years

ice trend

Figure 3. Monthly September ice extent for 1979 to 2017 shows a decline of 13.2 percent per decade.

Credit: National Snow and Ice Data Center
High-resolution image

The linear rate of sea ice decline for September is 86,100 square kilometers (33,200 square miles) per year, or 13.2 percent per decade relative to the 1981 to 2010 average. For comparison, the decline rate was calculated at 13.7 percent after the 2013 minimum, and 13.4 percent in 2016. Although sea ice shows significant year-to-year variability, the overall trend of decline remains strong.

Thickness and age trends in Arctic sea ice from models and data

Figure 4a. This image from the Pan-Arctic Ice Ocean Modeling and Assimilation System (PIOMAS) shows Arctic sea ice thickness departures from average (anomaly) in meters for September 2017, relative to the 2000 to 2015 average. Reds indicate thicker than average ice; blues indicate thinner than average ice.

Credit: NSIDC courtesy University of Washington Polar Science Center
High-resolution image

ice age

Figure 4b. Sea ice age distribution at the annual minimum extent for 1985 (upper left) and 2017 (upper right). Time series (bottom) of different age categories the minimum extent for 1985 to 2017. Note that the ice age product does not include ice in the Canadian Archipelago. Data from Tschudi et al., EASE-Grid Sea Ice Age, Version 3

Credit: W. Meier/National Snow and Ice Data Center, M. Tschudi et al.
High-resolution image

According to estimates from the University of Washington Polar Science Center’s PIOMAS, which assimilates observational data into a coupled ice-ocean model, sea ice volume was at record low levels from January through June of 2017. However, the generally cool summer conditions slowed the rate of ice melt, and the ice volume for September ended up fourth lowest in the PIOMAS record, above 2010, 2011, and 2012.

Another way to assess the volume of the ice, at least in a qualitative sense, is through tracking sea ice age (Figure 4b). Older ice is generally thicker ice. Over the satellite record, there has been a significant decline in coverage of the oldest, thickest ice. While this year’s minimum sea ice extent is higher than in 2016, the marginal gain can be largely attributed to younger ice types: first-year ice (0 to 1 years old) and second-year ice (1 to 2 years old). The oldest ice, that which is over 4 years old, is only slightly higher than last year and remains almost non-existent within the Arctic. At the minimum this year, ice older than 4 years constituted only ~150,000 square kilometers (~58,000 square miles), compared to over 2 million square kilometers (~770,000 square miles) during the mid-1980s.

Antarctic maximum extent

antarctic sea ice

Figure 5. The graph above shows Antarctic sea ice extent as of October 4, 2017, along with daily ice extent data for 2017 (aqua), 2016 (red), 2013 (dotted green), and 1986 (yellow). The 1981 to 2010 median is in dark gray. The gray areas around the median line show the interquartile and interdecile ranges of the data. Sea Ice Index data.

Credit: National Snow and Ice Data Center
High-resolution image

Antarctic sea ice may have reached its maximum extent on September 15, at 17.98 million square kilometers (6.94 million square miles), among the earliest maxima on record. If this date and extent hold, it will be the second-lowest daily maximum in the satellite record, 20,000 square kilometers (7,700 square miles) above 1986. Antarctic sea ice extent has been at record or near-record lows since September 2016. A series of recent studies have explored causes of the sudden decline in extent that occurred in austral late winter and spring of 2016. Most studies conclude that an unusual period of strong meridional winds—consistent with a very pronounced negative phase of the Southern Annular Mode index, coupled with a significant ‘wave-3 pattern’ in the atmospheric circulation—were the cause. A ‘wave-3 pattern’ refers to a tendency for circulation around the southern continent to resemble a three-leaf clover, rather than the more typical near-zonal (along lines of latitude) pattern.

The Maud Rise polynya, discussed in our last post, continues to grow and is now at about 35,000 square kilometers (14,000 square miles). A recent study (see Further reading, below) discusses how its formation is related to climate patterns and natural variability, and that the recent reappearance supports a forecast by an updated climate model.

Driftwood and long-term changes in Arctic ice movement

circulation

Figure 6. The maps show two modes of wintertime Arctic sea ice circulation patterns. (a) shows the Low Arctic Oscillation (AO) index has a strong Beaufort Gyre which supports ice re-circulation within the Arctic. (b) shows the High AO index in which the Beaufort Gyre is weak and the Transpolar Drift expands, leading to Arctic ice exported in a shorter time interval. Bold numbers show the average time in years for ice starting from various locations to be exported through Fram Strait under the illustrated patterns. The red dashed lines encircle the region of ice recirculation and persistence (Rigor et al., 2002). Over continents, light blue lines show watersheds with named major rivers (shown as bold blue lines) that export driftwood into the Arctic Ocean. Green letters indicate driftwood sample regions: CAA, Canadian Arctic Archipelago; EG, East Greenland; JM, Jan Mayen; NG, North Greenland; FJL, Franz Josef Land; NZ, Novaya Zemlya; SB, Svalbard. Circulation patterns compiled and modified from Rigor et al. (2002).

Credit: G. Hole and M. Macias-Fauria, The Cryosphere Discuss.
High-resolution image

While the satellite record has been key in documenting large declines in the Arctic sea ice cover during the past four decades, the data record is still relatively short. To provide a longer record, scientists turn to the geologic record and proxy data. One approach is to analyze the age, transport, and deposition of driftwood. Driftwood distribution depends strongly on past sea ice conditions and ocean currents. New research using 913 driftwood samples collected across the western Arctic (Figure 6) has shed new insight on sea ice changes during the Holocene, between 12,000 years ago to present. During the early Holocene (12,000 to 8,000 years ago), the analysis suggests that the clockwise Beaufort Gyre dominated Arctic Ocean circulation, allowing more sea ice to stay within the Arctic Ocean. In the mid-Holocene (8,000 to 4,000 years ago), temperatures were higher and the Transpolar Drift dominated, leading to more ice export out of the Arctic Ocean through Fram Strait and less sea ice in the Arctic Ocean. In the late Holocene (4,000 years ago to present), the Beaufort Gyre once again strengthened as temperatures slowly cooled until the most recent several decades.

Further reading

Hole, G. M. and M. Macias-Fauria. 2017. Out of the woods: Driftwood insights into Holocene pan-Arctic sea ice dynamic., J. Geophys. Res. Oceans, 122, doi:10.1002/2017JC013126.

Reintges, A., T. Martin, M. Latif, and W. Park. 2017. Physical controls of Southern Ocean deep-convection variability in CMIP5 models and the Kiel Climate Model. Geophys. Res. Lett., 44 (13), 6951-6958, doi:10.1002/2017GL074087.

Rigor, I.G., Wallace, J.M. and Colony, R.L. 2002. Response of sea ice to the Arctic Oscillation. Journal of Climate,15 (18), 2648-2663, doi:10.1175/1520-0442(2002)015<2648:ROSITT>2.0.CO;2.

Schlosser, E., Haumann, F. A., and Raphael, M. N. Atmospheric influences on the anomalous 2016 Antarctic sea ice decay. The Cryosphere Discuss., doi:10.5194/tc-2017-192, in review, 2017.

Arctic sea ice at minimum extent

On September 13, Arctic sea ice appears to have reached its seasonal minimum extent of 4.64 million square kilometers (1.79 million square miles), the eighth lowest in the 38-year satellite record. The overall rate of ice loss this summer was slowed by a persistent pattern of low sea level pressure focused over the central Arctic Ocean.

Please note that this is a preliminary announcement. Changing winds or late-season melt could still reduce the Arctic ice extent, as happened in 2005 and 2010. NSIDC scientists will release a full analysis of the Arctic melt season, and discuss the Antarctic winter sea ice growth, in early October.

Overview of conditions

Figure 1. Arctic sea ice extent for September 13, 2017 was 4.64 million square kilometers (1.79 million square miles), the eighth lowest in the satellite record. The orange line shows the 1981 to 2010 average extent for that day.

Figure 1. Arctic sea ice extent for September 13, 2017 was 4.64 million square kilometers (1.79 million square miles), the eighth lowest in the satellite record. The orange line shows the 1981 to 2010 average extent for that day. Sea Ice Index data. About the data

Credit: National Snow and Ice Data Center
High-resolution image

On September 13, 2017, sea ice extent reached an annual minimum of 4.64 million square kilometers (1.79 million square miles). This was 1.58 million square kilometers (610,000 square miles) below the 1981 to 2010 median extent for the same day, and 1.25 million square kilometers (483,000 square miles) and 500,000 square kilometers (193,000 square miles) above the 2012 and 2016 extents for the same day, respectively.

During the first two weeks of September, the ice edge continued to retreat in the Chukchi, East Siberian, and Kara Seas, whereas it slightly expanded in the Beaufort and Laptev Seas. The ice edge remains far to the north of its average position in the Chukchi Sea. The Northern Sea Route is largely open; Amundsen’s Northwest Passage (the southern route) has up to 50 percent ice cover in some places, though as noted in our last post, ships have successfully navigated through the southern route with icebreaker assistance. The northern Northwest Passage route, entered from the west via McClure Strait, remains choked by consolidated, thick, multi-year ice.

Conditions in context

Figure 2a. The graph above shows Arctic sea ice extent as of September 17, 2017, along with daily ice extent data for five previous years. 2017 is shown in blue, 2016 in green, 2015 in orange, 2014 in brown, 2013 in purple, and 2012 in dashed brown. The 1981 to 2010 median is in dark gray. The gray areas around the median line show the interquartile and interdecile ranges of the data. Sea Ice Index data.||Credit: National Snow and Ice Data Center|High-resolution image

Figure 2a. The graph above shows Arctic sea ice extent as of September 17, 2017, along with daily ice extent data for five previous years. 2017 is shown in blue, 2016 in green, 2015 in orange, 2014 in brown, 2013 in purple, and 2012 in dashed brown. The 1981 to 2010 median is in dark gray. The gray areas around the median line show the interquartile and interdecile ranges of the data. Sea Ice Index data.

Credit: National Snow and Ice Data Center
High-resolution image

Figure 2b. This image shows average sea level pressure over the Arctic Ocean for the period of September 1 to 16, 2017.

Figure 2b. This image shows average sea level pressure over the Arctic Ocean for the period of September 1 to 16, 2017.

Credit: NSIDC courtesy NOAA/ESRL Physical Sciences Division
High-resolution image

The date of the minimum ice extent for 2017 was two days earlier than the average minimum date of September 15. The earliest annual sea ice minimum in the satellite record occurred on September 5 in the years 1980 and 1987, and the latest on September 23, 1997.

As is typical of this time of year when the solar radiation received at the surface is quickly waning, the rate of ice loss slowed during the first half of September. Ice retreat from the beginning of September until the minimum averaged 25,300 square kilometers (9,770 square miles) per day, slightly faster than the 1981 to 2010 average for the same period of 22,800 square kilometers (8,800 square miles) per day.

The pattern of low sea level pressure over the central Arctic Ocean that persisted through this summer and inhibited summer ice loss has broken down. For the first half of September, the pattern has instead been one of above-average sea level pressure centered over the Barents Sea and extending across part of the Arctic Ocean (Figure 2b). Corresponding air temperatures at the 925 hPa level (about 2,500 feet above sea level) were above average over most of the Arctic Ocean. Above average temperatures over some parts of the Arctic Ocean likely reflect heat transfer to the atmosphere from areas of open water, hence cooling the ocean.

Ten lowest minimum Arctic sea ice extents (satellite record, 1979 to present)

Table 1.  Ten lowest minimum Arctic sea ice extents (satellite record, 1979 to present)
 RANK  YEAR MINIMUM ICE EXTENT DATE
IN MILLIONS OF SQUARE KILOMETERS IN MILLIONS OF SQUARE MILES
1 2012 3.39 1.31 Sept. 17
2 2016
2007
4.14
4.15
1.60
1.60
Sept. 10
Sept. 18
4 2011 4.34 1.67 Sept. 11
5 2015 4.43 1.71 Sept. 9
6 2008 4.59 1.77 Sept. 19
7 2010 4.62 1.78 Sept. 21
8 2017 4.64 1.79 Sept. 13
9 2014 5.03 1.94 Sept. 17
10 2013 5.05 1.94 Sept. 13

Effects of seasonal ice retreat in the Beaufort and Chukchi Seas

Figure 3. This chart shows combined sea ice extent in the Chukchi and Beaufort Seas from August 15 to October 7 for the years 2006 to 2016, including the extent so far for 2017. The colored dots show the day the minimum occurred in the region during a specific year. ||Credit: Courtesy R. Thoman/National Weather Service Alaska Region Environmental and Scientific Services Division| High-resolution image

Figure 3. This chart shows combined sea ice extent in the Chukchi and Beaufort Seas from August 15 to October 7 for the years 2006 to 2016, including the extent so far for 2017. The colored dots show the day the minimum occurred in the region during a specific year. Data are from the Multisensor Analyzed Sea Ice Extent (MASIE) product.

Credit: Courtesy R. Thoman/National Weather Service Alaska Region Environmental and Scientific Services Division
High-resolution image

According to a report by the Alaska Dispatch News, the lack of sea ice forced walruses to the shore of Alaska’s Chukchi Sea earlier than any time on record. The lack of ice also forced biologists monitoring Alaska polar bears to cut short their spring field season. In turn, the NOAA National Weather Service Climate Prediction Center states that because of the extensive open water, air temperatures over the Beaufort and Chukchi Seas and along the North Slope of Alaska will likely be far above average through this autumn.

Rick Thoman of the National Weather Service in Fairbanks, Alaska compiled an analysis of the combined Chukchi and Beaufort Seas ice extent from the Multisensor Analyzed Sea Ice Extent (MASIE) product. MASIE is based on operational ice analyses at the U.S. National Ice Center and is archived and distributed by NSIDC. It shows that 2017 tracked near record lows for the region through much of the summer, but after mid-August the pace of ice loss slowed relative to recent years. While it appears unlikely that extent in the Beaufort and Chukchi Seas will reach a record low (set in 2012), it will still be among the four or five lowest in the MASIE record (Figure 3). Note that the range in dates for the minimum extent in the region differs from those for the Arctic as a whole and tend to be later, ranging from September 10 in 2015 to September 25 in 2007 and 2008. In other words, the Chukchi and Beaufort Seas may continue to lose ice even after the overall Arctic minimum extent is reached. From the passive microwave data (not shown), the Chukchi/Beaufort minimum has occurred as early as August 14 in 1980 to as late as October 2 in 1991.

Antarctic sea ice approaching winter maximum

Figure 4a. The graph above shows Antarctic sea ice extent as of September 17, 2017, along with daily ice extent data for four previous years.

Figure 4a. The graph above shows Antarctic sea ice extent as of September 17, 2017, along with daily ice extent data for four previous years. 2017 is shown in blue, 2016 in green, 2015 in orange, 2014 in dashed brown, 2013 in purple. The 1981 to 2010 median is in dark gray. The gray areas around the median line show the interquartile and interdecile ranges of the data. Sea Ice Index data.

Credit: National Snow and Ice Data Center
High-resolution image

Figure 4b: This map shows Antarctic sea ice concentration on September 16, 2017. Note the Maud Rise polynya at the top of the image. Data are from the Advanced Microwave Scannig Radiometer 2 (AMSR2).

Figure 4b: This map shows Antarctic sea ice concentration on September 16, 2017. Note the Maud Rise polynya at the top of the image. Data are from the Advanced Microwave Scannig Radiometer 2 (AMSR2).

Credit: G. Heygster, C. Melsheimer, J. Notholt/Institute of Environmental Physics, University of Bremen
High-resolution image

Following the record low summer minimum extent in March, Antarctic sea ice extent is now nearing its winter maximum. This will likely be among the five lowest winter extents in the satellite era. As of mid-September, Antarctic ice extent was just under 18 million square kilometers (7 million square miles), which is approximately half a million square kilometers below the 1981 to 2010 median ice extent. Sea ice is below the typical extent in the Indian Ocean sector, the northern Ross Sea, and the northern Weddell Sea, and slightly above average extent in the northern Amundsen Sea region.

Between September 9 and September 17 of 2016, Antarctic sea ice lost nearly 100,000 square kilometers (38,600 square miles) of sea ice per day, and sea ice extent moved from near-average to a near-record-daily low by September 17. For the next 12 months Antarctic sea ice remained extremely low. Record low ice extents were set every day from November 5, 2016 to April 10, 2017. Extents averaged for November and December of 2016 were five standard deviations below average. No other 12-month period (September 2016 to August 2017) has had such persistently low sea ice extent. The year 1986 had near-record low extent for the winter period (June to December), but there were periods of near-average and even above-average ice extent earlier in the calendar year.

Beginning around September 2, an opening in the Antarctic sea ice pack formed north of Dronning Maud Land in the easternmost Weddell Sea (near 64°S, 5°E). By mid-September, this opening, or polynya, had grown to about 12,000 square kilometers (4,600 square miles). This feature has been observed intermittently in the Antarctic pack ice since the first satellite data became available in the 1970s. In 1974, 1975, and 1976, the polynya was much larger, averaging 250,000 square kilometers (96,500 square miles). It was absent for many years in the 1980s and 1990s. In recent years the feature has been observed sporadically and has been much smaller.

The polynya is formed when ocean currents uplift deep warm ocean water to the surface where it melts the sea ice. An oceanic plateau called the Maud Rise is responsible for forcing the vertical movement of the water. The persistence of certain atmospheric patterns, such as the southern annular mode, or SAM, is thought to play a role in driving the deep water layer against the Maud Rise.

2017 Arctic sea ice minimum animation

See the NASA animation of Arctic sea ice extent from the beginning of the melt season on March 8, 2017 to the day of the annual minimum on September 13, 2017 here.

Further reading

Gordon, A.L., Visbeck, M. and Comiso, J.C. 2007. A possible link between the Weddell Polynya and the Southern Annular Mode. Journal of Climate20(11), 2558-2571. doi:10.1175/JCLI4046.1

Holland, D.M. 2001. Explaining the Weddell Polynya–a large ocean eddy shed at Maud Rise. Science, 292(5522), 1697-1700. doi:10.1126/science.1059322.

Erratum

In Table 1, years 2014 and 2013 were both ranked ninth lowest. They should have been ninth and tenth respectively. This has been corrected.

 

 

The end of summer nears

Average sea ice extent for August 2017 ended up third lowest in the satellite record. Ice loss rates through August were variable, but slower overall than in recent years. Extensive areas of low concentration ice cover (40 to 70 percent) are still present across much of the Eurasian side of the Arctic Ocean.

Overview of conditions

Figure 1. Arctic sea ice extent for August 2017 was 5.51 million square kilometers (2.13 million square miles). The magenta line shows the 1981 to 2010 average extent for that month.

Figure 1. Arctic sea ice extent for August 2017 was 5.51 million square kilometers (2.13 million square miles). The magenta line shows the 1981 to 2010 average extent for that month. Sea Ice Index data. About the data

Credit: National Snow and Ice Data Center
High-resolution image

Arctic sea ice extent for August 2017 averaged 5.51 million square kilometers (2.13 million square miles), the third lowest August in the 1979 to 2017 satellite record. This was 1.77 million square kilometers (683,000 square miles) below the 1981 to 2010 average, and 800,000 square kilometers (309,000 square miles) above the record low August set in 2012.

Ice retreat was most pronounced in the western Beaufort Sea. A large region in the Beaufort Sea and East Siberian Sea has low concentration sea ice (40 to 70 percent). Patches of low concentration sea ice and some open water northeast of the Taymyr Peninsula are also present.

While a record low minimum extent in the Arctic is unlikely this year, the ice edge in the Beaufort Sea is extremely far north. In parts of this region, the ice edge is farther north than at any time since the satellite record began in 1979. This highlights the pronounced regional variability in ice conditions from year to year. A couple of the models that contribute to the Sea Ice Prediction Network Sea Ice Outlooks forecasted significantly less ice in the Beaufort Sea in July this year compared to average conditions.

Conditions in context

Figure 2a. The graph above shows Arctic sea ice extent as of September 5, 2017, along with daily ice extent data for five previous years. 2017 is shown in blue, 2016 in green, 2015 in orange, 2014 in brown, 2013 in purple, and 2012 in dotted brown. The 1981 to 2010 median is in dark gray. The gray areas around the median line show the interquartile and interdecile ranges of the data. Sea Ice Index data.||Credit: National Snow and Ice Data Center|High-resolution image

Figure 2a. The graph above shows Arctic sea ice extent as of September 5, 2017, along with daily ice extent data for five previous years. 2017 is shown in blue, 2016 in green, 2015 in orange, 2014 in brown, 2013 in purple, and 2012 in dotted brown. The 1981 to 2010 median is in dark gray. The gray areas around the median line show the interquartile and interdecile ranges of the data. Sea Ice Index data.

Credit: National Snow and Ice Data Center
High-resolution image

Figure 2b. This image shows average sea level pressure over the Arctic for the month of August 2017.

Figure 2b. This image shows average sea level pressure in millibars over the Arctic for the month of August 2017.

Credit: NSIDC courtesy NOAA Earth System Research Laboratory Physical Sciences Division
High-resolution image

Over the month, low atmospheric pressure prevailed over much of the Arctic Ocean, centered over the northern Beaufort Sea (Figure 2b). This helped to push the ice edge in that region northward. Summers dominated by low pressure, as has been the case for 2017, are generally not conducive to ice loss. Low pressure brings generally cool conditions and the cyclonic (counterclockwise) winds help spread the ice over a larger area. However, it appears that strong individual storms, such as the intense summer cyclone of 2012, may break up the ice and mix warm ocean waters into the sea ice, contributing to ice loss.

August 2017 compared to previous years

Figure 3. Monthly August ice extent for 1979 to 2017 shows a decline of 10.48 percent per decade.

Figure 3. Monthly August ice extent for 1979 to 2017 shows a decline of 10.5 percent per decade.

Credit: National Snow and Ice Data Center
High-resolution image

The linear rate of decline for August 2017 is 76,300 square kilometers (29,000 square miles) per year, or 10.5 percent per decade. The daily ice loss rate was variable over the month, slowing considerably during the middle and end of the month. Overall, the loss rate during August 2017 was slower than recent years, particularly 2012, when extent decreased rapidly during the first half of the month after a the passage of a strong storm.

A report on the Northwest Passage

Figure 4a. This shows the route of the Crystal Serenity (yellow line) overlaid on a map of ice cover in the Northwest Passage region.

Figure 4a. This shows the route of the Crystal Serenity (yellow line) overlaid on a map of ice cover in the Northwest Passage region for September 5, 2017.

Credit: Canadian Ice Service Daily and Regional Ice Charts
High-resolution image

Figure 4b. This chart shows sea ice area in the Parry Channel region of the Canadian Archipelago from April 30 to November 26 for the years 2010 (magenta), 2011 (dashed green), 2014 (blue green), 2016 (purple, and 2017 (red). The black line shows the 1980 to 2010 long-term average. ||Credit: Environment and Climate Change Canada|

Figure 4b. This chart shows sea ice area in the Parry Channel region of the Canadian Archipelago from April 30 to November 26 for the years 2010 (magenta), 2011 (dashed green), 2014 (blue green), 2016 (purple), and 2017 (red). The black line shows the 1981 to 2010 long-term average.

Credit: Environment and Climate Change Canada
High-resolution image

Figure 4c. This photo shows sea ice conditions along the route of the Crystal Serenity.||Credit: C. Haas|High-resolution image

Figure 4c. This photo shows sea ice conditions along the route of the Crystal Serenity.

Credit: C. Haas, York University
High-resolution image

Of particular interest each year are the ice conditions in the Northwest Passage. Our colleagues, Stephen Howell and Michael Brady of the Climate Research Division at Environment and Climate Change Canada, provided a status update based on sea ice charts produced by the Canadian Ice Service.

As of August 28, the northern route of the Northwest Passage was blocked by high concentration ice at the western (Parry Channel) entrance (Figure 4a), but overall extent in the passage is still tracking below the 1981 to 2010 average (Figure 4b). The record low seasonal extent in the northern route was in 2011. Low ice years in the northern route are typically the result of early breakup associated with above-average sea level pressure over the Beaufort Sea and Canadian Basin; this pattern tends to displace ice away from the western entrance. Conversely, low sea level pressure over the Beaufort Sea, such as seen this summer, packs ice up against the western entrance. It is unlikely the northern route will open in this year. Ice concentrations are much milder in the southern route (Amundsen’s Passage) but some ice still blocks the eastern part of that passage. This is primarily first year ice that may melt out in the coming weeks. During low ice summers, thick multi-year ice originating from the Arctic Ocean is typically advected into the channels of both the northern and southern routes, representing a significant hazard to transiting ships.

The Crystal Serenity luxury cruise ship, accompanied by an icebreaker, is attempting a repeat of last year’s cruise through the passage. Professor Christian Haas of York University in Toronto, who is providing sea ice expertise on board the ship, sent us a first-hand account of conditions:

“On August 29 and 30, we encountered some high concentration ice, up to 90 percent, just before entering Bellot Strait. The ice was mostly first-year ice, with about 30 percent multi-year ice having thicknesses greater than 2 meters (6.5 feet). Due to the relatively severe ice conditions, we were supported by the Canadian Coast Guard icebreaker Des Groseilliers, which helped clear some ice on our route into Franklin Strait. Overall, ice conditions have been much more severe than during the first transit of the Northwest Passage of the Crystal Serenity in 2016, when no ice was encountered at all.”

Figure 4c shows conditions along the route. The ship’s latest position is available at the cruise’s blog.

Also of note, the Finnish icebreaker MSV Nordica, set a new record for the earliest transit through the passage on July 29, traveling 10,000 kilometers (6,000 miles) in 24 days.

On the other side of the Arctic, the Northeast Passage (or Northern Sea Route) along the coast of Siberia appears largely open as shown by NSIDC data, though operational analysis by the U.S. National Ice Center shows some remaining ice along the coast of the Taymyr Peninsula.

A radar view of the Arctic

Figure 5. Composite image from Sentinel-1 SAR imagery for August 29, 2017 overlaid with passive microwave sea ice concentration from the JAXA AMSR2 sensor. Red/pink shades indicate regions of high (>90 percent) concentration, while green/brown shades indicate lower (~30-70 percent) concentration. Open water is in black to blue/gray shades. Open water is evident between floes of ice in the low concentration regions.

Figure 5. This figure shows a composite image from Sentinel-1 SAR imagery for August 29, 2017 overlaid with passive microwave sea ice concentration from the Japan Aerospace Exploration Agency (JAXA) Advanced Microwave Scanning Radiometer 2 (AMSR2) sensor. Red/pink shades indicate regions of high (more than 90 percent) concentration, while green/brown shades indicate lower (~30-70 percent) concentration. Open water is in black to blue/gray shades. Open water is evident between floes of ice in the low concentration regions.

Credit: Image from Danish Technical University, courtesy Leif Pedersen and Roberto Saldo
High-resolution image

The passive microwave satellite imagery that NSIDC uses is ideal for showing long-term changes in sea ice because it provides a continuous record extending back to 1979. Its drawback is low spatial resolution. Satellite radar images, though more limited in spatial and temporal coverage, provide a more detailed picture of the ice cover. A daily composite of images from the European Space Agency’s Sentinel-1 satellite, with a synthetic aperture radar (SAR) instrument, shows substantial open water well within the ice pack, north of 80° N latitude.

Cooler conditions, slower melt

A cooler than average first half of the month kept ice loss at a sluggish pace with little change in the ice edge within the eastern Arctic. Retreat was mostly confined to the western Beaufort and northern Chukchi seas. Ice extent remains above that seen in 2012 and 2007.

Overview of conditions

Figure 1. Arctic sea ice extent for August 21, 2017 was 5.27 million square kilometers (2.03 million square miles). The orange line shows the 1981 to 2010 average extent for that day. Sea Ice Index data. About the data

Credit: National Snow and Ice Data Center
High-resolution image

On August 21, 2017, ice extent stood at 5.27 million square kilometers (2.03 million square miles). This was 1.82 million square kilometers (703,000 square miles) below the 1981 to 2010 median extent for the same day, and 804,000 square kilometers (310,000 square miles) and 221,000 square kilometers (85,000 square miles) above the 2012 and 2007 extents for the same day, respectively. The ice edge remained nearly constant through the first half of the month in the Barents and Kara Seas, and retreated only slightly within the East Greenland Sea. The ice edge also remained stable in the Laptev and East Siberian Seas through the first half of the month. Ice retreat occurred primarily within the Chukchi and western Beaufort Seas as well as near the New Siberian Islands. Some ice continues to block the Northern Sea Route near Severnaya Zemlya. Both McClure Strait and the Amundsen Gulf routes within the Northwest Passage remain blocked by ice. On August 17, the Russian nuclear powered icebreaker 50 let pobedy reached the North Pole in just 79 hours, the fastest time yet.

Conditions in context

Figure 2. The graph above shows Arctic sea ice extent as of August 21, 2017, along with daily ice extent data for five previous years. 2017 is shown in blue, 2016 in green, 2015 in orange, 2014 in brown, 2013 in purple, and 2012 as a dashed line. The 1981 to 2010 median is in dark gray. The gray areas around the median line show the interquartile and interdecile ranges of the data. Sea Ice Index data.

Credit: National Snow and Ice Data Center
High-resolution image

Ice retreat from August 1 to August 21 averaged 73,000 square kilometers (28,000 square miles) per day. This was faster than the 1981 to 2010 average rates of ice loss of 57,300 square kilometers (22,000 square miles) per day, but slower than in 2012, which exhibited the fastest rate of ice loss compared to any other August in the passive microwave satellite data record. Normally the rate of ice retreat slows in August as the sun starts to dip lower in the sky. The rate of ice loss was more rapid at the beginning of August, slowing down considerably starting on August 17.

Air temperatures the first two weeks of August were 1 to 3 degrees Celsius (2 to 5 degrees Fahrenheit) cooler than the 1981 to 2010 average throughout the Arctic Ocean and over Greenland and the North Atlantic. The lowest air temperatures relative to the long-term average were found in coastal regions of the Kara and Barents Seas, continuing the pattern seen throughout much of this summer. Cooler than average conditions within the Central Arctic were a result of persistent cold-core cyclones. These cyclones have not been as large or as strong as the Great Arctic Cyclones of 2012 and 2016, despite the central pressure of one of these systems dropping down to 974 hPa on August 10. In addition, these cyclones are located closer towards the pole within the consolidated ice pack, where they are less likely to cause significant ice loss, as did the 2012 Great Arctic Cyclone in the Chukchi Sea.

While air temperatures start to drop in August, ice melt continues through the month as heat gained in the ocean mixed layer during summer continues to melt the ice from below and from the sides. Sea surface temperatures have been up to 5 degrees Celsius (9 degrees Fahrenheit) above average near the coastal regions, but generally near average or slightly below average along the ice edge in the Beaufort and Chukchi Seas.

NASA Operation IceBridge conducts summer flights

Figure 3. This photograph, taken during NASA Operation IceBridge on July 25, 2017, shows melt ponds on the surface of Arctic sea ice.

Credit: Eric Fraim/NASA
High-resolution image

NASA’s Operation IceBridge (OIB) airborne campaign flew several missions over the Greenland ice sheet this summer to study changes in Greenland outlet glaciers, as well as to observe sea ice. A recent mission collected laser altimeter data to investigate sea ice thickness changes resulting from the piling up of sea ice (or convergence) as ice motion pushes the ice up against the coast. Flights were completed on July 17 and July 25. High-resolution visible imagery collected on the flights also provides close-up looks of melt pond development.

Influence of warm Pacific water

Figure 4. This plot shows measurements of sea surface temperature from drifting buoys, along with satellite-derived sea surface temperature from NOAA and ice concentration from NSIDC for August 6, 2017. Buoy positions as of August 6 are indicated with circles. A gray dot indicates that the buoy is reporting a temperature value outside the range of -2 to 10 degrees Celsius. Red, orange, and yellow indicate higher temperatures, while blues and purples indicate lower temperatures. Whites indicate higher sea ice concentration, and grays indicate lower concentration.

Credit: University of Washington Polar Science Center
High-resolution image

This May, sea ice in the Chukchi Sea was at a record low for the satellite data record. The early retreat of ice in this region may partially be a result of unusually warm ocean temperatures in the region. As reported by Rebecca Woodgate of the University of Washington, Seattle, the Research Vessel Norseman II spent eight days in the Bering Strait and southern Chukchi Sea region to recover oceanographic moorings and whale acoustic instruments, in addition to deploying new instruments. The mooring data indicated early arrival of warm water in the strait, about a month earlier than the average, resulting in June ocean temperatures that were 3 degrees Celsius (5 degrees Fahrenheit) above average. Early intrusion of warm water in the Bering Strait back in May helped to melt sea ice from below, and may have been one of the factors for early development of open water in the region.

Arctic air temperatures and the Paris Climate Accord target

Figure 5. The bar graph, top, shows the Berkeley Earth evaluation of the ten warmest years since 1979 in the Arctic north of 80°N; the plot, middle, shows Arctic average temperatures for the period 1900 to 2016, relative to a 1951 to 1980 reference period; bottom, a map of Arctic temperature differences, in degrees Celsius, for the 2012 to 2016 period (5 years) relative to a 1981 to 2010 reference period.

Credit: National Snow and Ice Data Center
High-resolution image

Our past reports, and many other sources, have noted that the Arctic region is warming faster than the rest of the globe. This warming has accelerated in recent years, particularly since 2005. The ten warmest years on record for the Arctic are within the past twelve years, and 2016 was by far the warmest in the record since 1900. These observations are supported by both NOAA National Centers for Environmental Prediction (NCEP) reanalysis climate data, and by our colleagues at Berkeley Earth. Berkeley Earth is an independent climate fact- and analysis-checking group dedicated to an objective evaluation of the main claims and data sets used to support climate trends and forecasts.

One of the major statements of the recent Paris Climate Accord, dealing with heat-trapping ga reductions, is a target to hold the increase in the global average temperature to well below 2 degrees Celsius (3.6 degrees Fahrenheit) above the pre-industrial average. While this reference for the increase (pre-industrial average) is somewhat ambiguous, using reference average temperatures of either 1951 to 1980 or 1980 to 2010 for the Arctic shows that much of the area north of 80°N is already above this guideline over the past five years (2012 to 2016). As the Arctic will likely continue to warm above 2 degrees Celsius, other areas will need to warm less than that if the threshold is not to be exceeded. In general, land warms about 30 percent faster than oceans in the models, so in a global-average 2 degree Celsius warmer world, much of the global land area would have warmed more than 2 degrees Celsius.

The annual average air temperature for 2016 for the Arctic north of 80°N was more than 3.5 degrees Celsius (6.3 degrees Fahrenheit) above the 1951 to 1980 reference period, the warmest year yet, and most years during the past decade had annual average temperatures between 2 to  2.5 degrees Celsius (3.6 to 4.5 degrees Fahrenheit) above the reference period. Geographically, the NOAA NCEP reanalysis shows that recent warming is primarily located over the Arctic Ocean, and smaller warming trends are seen in the circum-Arctic land areas.

Further reading

Hawkins, E., Ortega, P., Suckling, E., Schurer, A., Hegerl, G., Jones, P., Joshi, M., Osborn, T.J., Masson-Delmotte, V., Mignot, J. and Thorne, P. 2017. Estimating changes in global temperature since the pre-industrial period. Bulletin of the American Meteorological Society, doi:10.1175/BAMS-D-16-0007.1.

Which August will we get?

Average sea ice extent for July 2017 ended up fifth lowest in the satellite record. This reflects weather conditions that were not favorable for ice loss. It will be important to monitor August 2017, as weather conditions and storm events during this month have been closely related to the seasonal minimum sea ice extent in the recent years.

Overview of conditions

Figure 1. Arctic sea ice extent for July 2017 was 8.2 million square kilometers (3.2 million square miles). The magenta line shows the 1981 to 2010 average extent for that month.

Figure 1. Arctic sea ice extent for July 2017 was 8.21 million square kilometers (3.17 million square miles). The magenta line shows the 1981 to 2010 average extent for that month. Sea Ice Index data. About the data

Credit: National Snow and Ice Data Center
High-resolution image

Arctic sea ice extent for July 2017 averaged 8.21 million square kilometers (3.17 million square miles), the fifth lowest July in the 1979 to 2017 satellite record. The average July extent was 1.58 million square kilometers (610,000 square miles) below the 1981 to 2010 long-term average, and 270,000 square kilometers (104,000 square miles) above the previous record low July set in 2011. July 2017 tracked 250,000 square kilometers (97,000 square miles) above the July 2012 extent and 20,000 square kilometers (7,700 square miles) above the July 2007 extent.

Ice extent was lower than average over most of the Arctic, particularly on the Pacific side where the ice retreated throughout July in the Beaufort, Chukchi, and East Siberian Seas. In the eastern Beaufort Sea on the other hand, extent slightly expanded during July. This may relate to the cyclonic (counterclockwise) pattern of winds favoring the drift of sea ice into the region.

Conditions in context

Figure 2a. The graph above shows Arctic sea ice extent as of August 1, 2017, along with daily ice extent data for five previous years. 2017 is shown in blue, 2016 in green, 2015 in orange, 2014 in brown, 2013 in purple, and 2012 in dotted red. The 1981 to 2010 median is in dark gray. The gray areas around the median line show the interquartile and interdecile ranges of the data

Figure 2a. The graph above shows Arctic sea ice extent as of August 1, 2017, along with daily ice extent data for five previous years. 2017 is shown in blue, 2016 in green, 2015 in orange, 2014 in brown, 2013 in purple, and 2012 in dotted red. The 1981 to 2010 median is in dark gray. The gray areas around the median line show the interquartile and interdecile ranges of the data. Sea Ice Index data.

Credit: National Snow and Ice Data Center
High-resolution image

Figure 2b. The plot shows differences from average for Arctic air temperatures at the 925 hPa level (about 2,500 feet above sea level) in degrees Celsius. Yellows and reds indicate higher than average temperatures; blues and purples indicate lower than average temperatures.

Figure 2b. The plot shows Arctic air temperature differences relative to the 1981 to 2010 long-term average at the 925 hPa level (about 2,500 feet above sea level) in degrees Celsius. Yellows and reds indicate higher than average temperatures; blues and purples indicate lower than average temperatures.

Credit: NSIDC courtesy NOAA Earth System Research Laboratory Physical Sciences Division
High-resolution image

The air temperature pattern over the Arctic was rather complex in July. Temperatures were above average over Alaska, extending into the Beaufort Sea (1 to 2 degrees Celsius or 2 to 4 degrees Fahrenheit) and the Kara and Barents Seas (2 to 4 degrees Celsius or 4 to 7 degrees Fahrenheit). By contrast, temperatures were 2 to 4 degrees Celsius (4 to 7 degrees Fahrenheit) lower than average over Greenland, East Central Siberia, and the Laptev Sea. The air pressure pattern at sea level was dominated by a broad area of low pressure covering most of the Arctic Ocean, with the lowest pressures centered just south of the Pole and west of the date line. Another locus of low pressure was centered over the southern Canadian Arctic Archipelago.

A cyclonic circulation over the central Arctic Ocean is generally viewed as unfavorable for rapid summer ice loss. Ice loss rates tend to be higher when the central Arctic Ocean is dominated by high pressure during summer.

July 2017 compared to previous years

Figure 3. Monthly July ice extent for 1979 to 2017 shows a decline of 7.4 percent per decade.

Figure 3. Monthly July ice extent for 1979 to 2017 shows a decline of 7.4 percent per decade.

Credit: National Snow and Ice Data Center
High-resolution image

The linear rate of decline for July 2017 was 72,500 square kilometers (28,000 square miles) per year, or 7.4 percent per decade.

Onset of surface melt

Figure 4. The plot on the left shows melt onset dates in day of the year (left). Warm colors indicate early melt onset and cooler colors indicate a later melt onset. The plot on the right shows departures from the average melt onset dates in number of days. Warmer colors indicate later than average melt onset and cooler colors indicate later than average melt onset.

Figure 4. The plot on the left shows the average melt onset dates in day of the year. Warm colors indicate early melt onset and cooler colors indicate a later melt onset. The plot on the right shows departures from the average melt onset dates in number of days. Warmer colors indicate later than average melt onset and cooler colors indicate earlier than average melt onset.

Credit: National Snow and Ice Data Center
High-resolution image

One important influence on the pace of summer sea ice retreat is the timing of the onset of surface melt. Surface melt drops the albedo, allowing more solar radiation to be absorbed at the surface. The onset of surface melt in 2017 was quite early in the Chukchi and Eastern Beaufort Seas—as much as 35 days earlier than the 1981 to 2010 average. Early melt was also seen in the Kara Sea, Baffin Bay, and Canadian Arctic Archipelago (10 to 20 days earlier than average). However, over a large portion of the Arctic Ocean’s central ice pack, melt onset was a few days later than average. Melt onset was up to ten days later than average in the polar areas at the northern extents of the Beaufort and East Siberian Seas. The spatial pattern in the timing of melt onset reflects the combination of the warm winter and early spring conditions along the edge of the pack on the Pacific side, very warm winter conditions in the Barents and Kara Sea areas, and relatively cool late spring to early summer conditions over the central Arctic Ocean.

Sea ice predictions for the 2017 minimum

Figure 5. This chart summarizes the 2017 summer September minimum forecasts from 36 different models and approaches to predicting the evolution of the Arctic pack.

Figure 5. This chart summarizes the 2017 summer September minimum forecasts from 36 different models and approaches to predicting the evolution of the Arctic pack.

Credit: SIPN/ARCUS
High-resolution image

A report released by the Sea Ice Prediction Network (SIPN), an activity of the Arctic Research Council of the United States (ARCUS), compiled 36 forecasts of September average sea ice extent that were submitted during July. Additionally, SIPN produces estimates for sea ice extent in the Alaska region and, new this year, for the Antarctic sea ice maximum (not shown). The September Arctic extent forecasts range from a new record low of 3.1 million square kilometers (1.2 million square miles) to an eleventh lowest extent of 5.5 million square kilometers (2.1 million square miles). The median of the estimates is slightly below the September 2016 value, currently the fifth lowest. A variety of methods are used to make these forecasts, ranging from coupled ice-ocean-atmospheric models to statistical approaches and heuristic guesses. NSIDC scientists Julienne Stroeve, Walt Meier, Andrew Barrett, Mark Serreze, and the late Drew Slater have regularly contributed to several separate estimations for the SIPN.

Sea ice retreat may be changing the AMOC

In the far northern Atlantic, warm water flowing northward from the tropics is cooled by the atmosphere, becomes denser, and eventually sinks to great depths. The descending water is key in driving a sub-surface and surface ocean circulation system called the Atlantic Meridional Overturning Circulation (AMOC), which is part of the global ocean conveyor belt of heat and salinity. Where the Atlantic water sinks has a very important effect on the climate of Northern Europe; the heat that the ocean loses to the atmosphere is what keeps Northern Europe quite warm relative to its latitude. For example, Amsterdam is at the same latitude as Winnipeg, Canada, but experiences much warmer winters.

Based on a recent modeling study, Florian Sévellec and colleagues propose that the ongoing loss of Arctic sea ice may disrupt the AMOC. The sea ice loss leads to a freshening of the northern North Atlantic and stronger heat absorption at the surface. This means that waters in the northern North Atlantic are less dense than they used to be, which has the effect of providing a cap, or lid, that may inhibit the northward flow of warm waters at the surface and the eventual sinking of these waters. The authors suggest that the Arctic sea ice decline may help to explain observations suggesting that the AMOC may be slowing down, and why there is a regional minimum in warming (sometimes called the Warming Hole) over the subpolar North Atlantic.

Further reading

Sea Ice Prediction Network. “2017: July Report.” Arctic Research Council of the United States. https://www.arcus.org/sipn/sea-ice-outlook/2017/july.

Sévellec, F., A. V. Fedorov, and W. Liu. 2017. Arctic sea-ice decline weakens the Atlantic Meridional Overturning Circulation. Nature Climate Change, doi:10.1038/nclimate3353.

A recent slowdown

Arctic extent nearly matched 2012 values through the first week of July, but the rate of decline slowed during the second week. Weather patterns were unremarkable during the first half of July.

Overview of conditions

Figure 1. Arctic sea ice extent for July 17, 2017 was 7.88 million square kilometers (3.04 million square miles). The orange line shows the 1981 to 2010 average extent for that day. Sea Ice Index data. About the data

Credit: National Snow and Ice Data Center
High-resolution image

As of July 17, Arctic sea ice extent stood at 7.88 million square kilometers (3.04 million square miles). This is 1.69 million square kilometers (653,000 square miles) below the 1981 to 2010 average, and 714,000 square kilometers (276,000 million square miles) below the interdecile range. Extent was lower than average over most of the Arctic, except for the East Greenland Sea (Figure 1). Hudson Bay was nearly ice free by mid July, much earlier than is typical, but in line with what has been observed in recent years.

Conditions in context

Figure 2a. The graph above shows Arctic sea ice extent as of July 17, 2017, along with daily ice extent data for five previous years. 2017 is shown in blue, 2016 in green, 2015 in orange, 2014 in brown, 2013 in purple, and 2012 as a dotted brown line. The 1981 to 2010 median is in dark gray. The gray areas around the median line show the interquartile and interdecile ranges of the data. Sea Ice Index data.

Credit: National Snow and Ice Data Center
High-resolution image

Figure 2b. This map compares sea ice extent for July 11 in 2017 and in 2012. White shows where ice occurred only in 2012, medium blue is where ice occurred only in 2017, and light blue is where ice occurred in both years.

Credit: National Snow and Ice Data Center
High-resolution image

Through the first week of July, extent closely tracked 2012 levels. The rate of decline then slowed, so that as of July 17, extent was 169,000 square kilometers (65,300 square miles) above 2012 for the same date (Figure 2a). The spatial pattern of ice extent differs from 2012, with less ice in the Chukchi and East Siberian Seas in 2017, but more in the Beaufort, Kara, and Barents Seas and in Baffin Bay (Figure 2b).

Visible imagery provides up close details

Figure 4a. Sea ice in the Canadian Archipelago on July 3, 2017. The blue hues indicate areas of widespread melt ponds on the surface of the ice. ||Credit: RESEARCHER'S NAME/ORGANIZATION *or * National Snow and Ice Data Center| High-resolution image

Figure 3a. This image from the NASA Moderate Resolution Imaging Spectroradiometer (MODIS) shows sea ice in the Canadian Archipelago on July 3, 2017. The blue hues indicate areas of widespread melt ponds on the surface of the ice.

Credit: Land Atmosphere Near-Real Time Capability for EOS (LANCE) System, NASA/GSFC
High-resolution image

sea ice floes

Figure 3b. The Sentinel-2 satellite captured this image of large sea ice floes in Nares Strait on July 8, 2017.

Credit: European Space Agency
High-resolution image

MODIS image of arctic

Figure 3c. This false-color composite image of the Arctic is based on NASA MODIS imagery from July 4 to 10. Most clouds are eliminated by using several images over a week, but some clouds remain, particularly over the ocean areas.

Credit: NASA/Canadian Ice Service
High-resolution image

NSIDC primarily relies on passive microwave data because it provides complete coverage—night and day, and through clouds—and because it is consistent over its long data record. However, other types of satellite data, for example visible imagery from the NASA MODIS instrument on the Aqua and Terra satellites or from the European Space Agency Sentinel 2 satellite, can sometimes provide more detail. When skies clear, details of the ice cover can be seen, including leads, individual ice floes and melt ponds. For example, on July 3 in the Canadian Archipelago, 1-kilometer resolution MODIS imagery shows that the ice surface has a distinctive blueish hue due to the presence of melt ponds on the surface (Figure 3a). Higher resolution Sentinel-2 imagery (10 meters, Figure 3b) on the other hand provides up close detail on individual melt ponds on the ice floes.

The Arctic is a cloudy place, and generally, it is difficult to obtain a clear-sky image of the entire region. However, if images are compiled, or composited, over several days, most of the region may have at least some clear sky. This approach can yield a composite image that is mostly cloud-free. The Canadian Ice Service uses this approach to create a weekly nearly cloud-free composite image of the Arctic (Figure 3c). However, because the ice cover moves (typically several kilometers per day) and melts (during the summer), over the week-long composite period, fine details that can be seen in the daily imagery are not as evident because they have been “smeared” out over the week.

An ice-diminished Arctic

In response to diminishing ice extent, the US Navy has been holding a semi-annual symposium to bring together scientists, policy makers, and others to discuss the sea ice changes and their impacts. The seventh Symposium is taking place this week in Washington, DC, and will be attended by NSIDC scientists Mark Serreze, Walt Meier, Florence Fetterer, and Ted Scambos.

Tendency for warmer winters is increasing

A new study published this week in Geophysical Research Letters by Robert Graham at the Norwegian Polar Institute shows that warm winters in the Arctic are becoming more frequent and lasting for longer periods of time than they used to. Warm events were defined by when the air temperatures rose above -10 degrees Celsius (14 degrees  Fahrenheit). While this is still well below the freezing point, it is 20 degrees Celsius (36 degrees Fahrenheit) higher than average. The last two winters have seen temperatures near the North Pole rising to 0 degrees Celsius. While an earlier study showed that winter 2015/2016 was the warmest recorded at that time, the winter of 2016/2017 was even warmer.

Reference

Graham, R. M., L. Cohen, A. A. Petty, L. N. Boisvert, A. Rinke, S. R. Hudson, M. Nicolaus, and M. A. Granskog. 2017. Increasing frequency and duration of Arctic winter warming events, Geophys. Res. Lett., 44, doi:10.1002/2017GL073395.