Chapter 11: Arctic Changes and their Effects on Alaska and the Rest of the United States

11.1: Introduction

Climate changes in Alaska and across the Arctic continue to outpace changes occurring across the globe. The Arctic, defined as the area north of the Arctic Circle, is a vulnerable and complex system integral to Earth’s climate. The vulnerability stems in part from the extensive cover of ice and snow, where the freezing point marks a critical threshold that when crossed has the potential to transform the region. Because of its high sensitivity to radiative forcing and its role in amplifying warming,¹ the arctic cryosphere is a key indicator of the global climate state. Accelerated melting of multiyear sea ice, mass loss from the Greenland Ice Sheet (GrIS), reduction of terrestrial snow cover, and permafrost degradation are stark examples of the rapid Arctic-wide response to global warming. These local arctic changes influence global sea level, ocean salinity, the carbon cycle, and potentially atmospheric and oceanic circulation patterns. Arctic climate change has altered the global climate in the past² and will influence climate in the future.

As an arctic nation, United States’ decisions regarding climate change adaptation and mitigation, resource development, trade, national security, transportation, etc., depend on projections of future Alaskan and arctic climate. Aside from uncertainties due to natural variability, scientific uncertainty, and human activities including greenhouse gas emissions (see Ch. 4: Projections), additional unique uncertainties in our understanding of arctic processes thwart projections, including mixed-phase cloud processes;³ boundary layer processes;⁴ sea ice mechanics;⁴ and ocean currents, eddies, and tides that affect the advection of heat into and around the Arctic Ocean.^{5 ,6} The inaccessibility of the Arctic has made it difficult to sustain the high-quality observations of the atmosphere, ocean, land, and ice required to improve physically-based models. Improved data quality and increased observational coverage would help address societally relevant arctic science questions.

Despite these challenges, our scientific knowledge is sufficiently advanced to effectively inform policy. This chapter documents significant scientific progress and knowledge about how the Alaskan and arctic climate has changed and will continue to change.

11.2: Arctic Changes

11.2.1 Alaska and Arctic Temperature

Surface temperature—an essential component of the arctic climate system—drives and signifies change, fundamentally controlling the melting of ice and snow. Further, the vertical profile of boundary layer temperature modulates the exchange of mass, energy, and momentum between the surface and atmosphere, influencing other components such as clouds.^{7 ,8} Arctic temperatures exhibit spatial and interannual variability due to interactions and feedbacks between sea ice, snow cover, atmospheric heat transports, vegetation, clouds, water vapor, and the surface energy budget.^{9 ,10 ,11} Interannual variations in Alaskan temperatures are strongly influenced by decadal variability like the Pacific Decadal Oscillation (Ch. 5: Circulation and Variability).^{12 ,13} However, observed temperature trends exceed this variability.

Arctic surface and atmospheric temperatures have substantially increased in the observational record. Multiple observation sources, including land-based surface stations since at least 1950 and available meteorological reanalysis datasets, provide evidence that arctic near-surface air temperatures have increased more than twice as fast as the global average.^{14 ,15 ,16 ,17 ,18} Showing enhanced arctic warming since 1981, satellite-observed arctic average surface skin temperatures have increased by 1.08° ± 0.13°F (+0.60° ± 0.07°C) per decade.¹⁹ As analyzed in Chapter 6: Temperature Change (Figure 6.1), strong near-surface air temperature warming has occurred across Alaska exceeding 1.5°F (0.8°C) over the last 30 years. Especially strong warming has occurred over Alaska’s North Slope during autumn. For example, Utqiagvik’s (formally Barrow) warming since 1979 exceeds 7°F (3.8°C) in September, 12°F (6.6°C) in October, and 10°F (5.5°C) in November.²⁰

Enhanced arctic warming is a robust feature of the climate response to anthropogenic forcing.^{21 ,22} An anthropogenic contribution to arctic and Alaskan surface temperature warming over the past 50 years is very likely.^{23 ,24 ,25 ,26 ,27} One study argues that the natural forcing has not contributed to the long-term arctic warming in a discernable way.²⁷ Also, other anthropogenic forcings (mostly aerosols) have likely offset up to 60% of the high-latitude greenhouse gas warming since 1913,²⁷ suggesting that arctic warming to date would have been larger without the offsetting influence of aerosols. Other studies argue for a more significant contribution of natural variability to observed arctic temperature trends^{24 ,28} and indicate that natural variability alone cannot explain observed warming. It is very likely that arctic surface temperatures will continue to increase faster than the global mean through the 21st century.^{25 ,26 ,27 ,29}

11.2.2 Arctic Sea Ice Change

Arctic sea ice strongly influences Alaskan, arctic, and global climate by modulating exchanges of mass, energy, and momentum between the ocean and the atmosphere. Variations in arctic sea ice cover also influence atmospheric temperature and humidity, wind patterns, clouds, ocean temperature, thermal stratification, and ecosystem productivity.^{7 ,10 ,30 ,31 ,32 ,33 ,34 ,35 ,36 ,37} Arctic sea ice exhibits significant interannual, spatial, and seasonal variability driven by atmospheric wind patterns and cyclones, atmospheric temperature and humidity structure, clouds, radiation, sea ice dynamics, and the ocean. ^{38 ,39 ,40 ,41 ,42 ,43 ,44}

Overwhelming evidence indicates that the character of arctic sea ice is rapidly changing. Observational evidence shows Arctic-wide sea ice decline since 1979, accelerating ice loss since 2000, and some of the fastest loss along the Alaskan coast.^{19 ,20 ,45 ,46} Although sea ice loss is found in all months, satellite observations show the fastest loss in late summer and autumn.⁴⁵ Since 1979, the annual average arctic sea ice extent has very likely decreased at a rate of 3.5%–4.1% per decade.^{19 ,37} Regional sea ice melt along the Alaskan coasts exceeds the arctic average rates with declines in the Beaufort and Chukchi Seas of −4.1% and −4.7% per decade, respectively.²⁰ The annual minimum and maximum sea ice extent have decreased over the last 35 years by −13.3% ± 2.6% and −2.7% ± 0.5% per decade, respectively.⁴⁷ The ten lowest September sea ice extents over the satellite period have all occurred in the last ten years, the lowest in 2012. The 2016 September sea ice minimum tied with 2007 for the second lowest on record, but rapid refreezing resulted in the 2016 September monthly average extent being the fifth lowest. Despite the rapid initial refreezing, sea ice extent was again in record low territory during fall–winter ²⁰¹⁶⁄₂₀₁₇ due to anomalously warm temperatures in the marginal seas around Alaska,⁴⁷ contributing to a new record low in winter ice-volume (see http://psc.apl.uw.edu/research/projects/arctic-sea-ice-volume-anomaly).⁴⁸

 

Figure 11.1

VIEW

September sea ice extent and age shown for (a) 1984 and (b) 2016, illustrating significant reductions in sea ice extent and age (thickness). Bar graph in the lower right of each panel illustrates the sea ice area (unit: million km²) covered within each age category (>1 year), and the green bars represent the maximum value for each age range during the record. The year 1984 is representative of September sea ice characteristics during the 1980s. The years 1984 and 2016 are selected as endpoints in the time series; a movie of the complete time series is available at http://svs.gsfc.nasa.gov/cgi-bin/details.cgi?aid=4489. (c) Shows the satellite-era arctic sea ice areal extent trend from 1979 to 2016 for September (unit: million mi²). [Figure source: Panels (a),(b): NASA Science Visualization Studio; data: Tschudi et al. 2016;⁴⁹ Panel (c) data: Fetterer et al. 2016²⁰⁹ ].

Other important characteristics of arctic sea ice have also changed, including thickness, age, and volume. Sea ice thickness is monitored using an array of satellite, aircraft, and vessel measurements.^{37 ,45} The mean thickness of the arctic sea ice during winter between 1980 and 2008 has decreased between 4.3 and 7.5 feet (1.3 and 2.3 meters).³⁷ The age distribution of sea ice has become younger since 1988. In March 2016, first-year (multi-year) sea ice accounted for 78% (22%) of the total extent, whereas in the 1980s first-year (multi-year) sea ice accounted for 55% (45%).⁴⁷ Moreover, ice older than four years accounted for 16% of the March 1985 icepack but accounted for only 1.2% of the icepack in March 2016, indicating significant changes in sea ice volume.⁴⁷ The top two panels in Figure 11.1 show the September sea ice extent and age in 1984 and 2016, illustrating significant reductions in sea ice age.⁴⁹ While these panels show only two years (beginning point and ending point) of the complete time series, these two years are representative of the overall trends discussed and shown in the September sea ice extent time series in the bottom panel of Fig 11.1. Younger, thinner sea ice is more susceptible to melt, therefore reductions in age and thickness imply a larger interannual variability of extent.

Sea ice melt season—defined as the number of days between spring melt onset and fall freeze-up—has lengthened Arctic-wide by at least five days per decade since 1979, with larger regional changes. ^{46 ,50} Some of the largest observed changes in sea ice melt season (Figure 11.2) are found along Alaska’s northern and western coasts, lengthening the melt season by 20–30 days per decade and increasing the annual number of ice-free days by more than 90.⁵⁰ Summer sea ice retreat along coastal Alaska has led to longer open water seasons, making the Alaskan coastline more vulnerable to erosion.^{51 ,52} Increased melt season length corresponds to increased absorption of solar radiation by the Arctic Ocean during summer and increases upper ocean temperature, delaying fall freeze-up. Overall, this process significantly contributes to reductions in arctic sea ice.^{42 ,46} Wind-driven sea ice export through the Fram Strait has not increased over the last 80 years;³⁷ however, one recent study suggests that it may have increased since 1979.⁵³

 

Figure 11.2

VIEW

A 35-year trend in arctic sea ice melt season length, in days per decade, from passive microwave satellite observations, illustrating that the sea ice season has shortened by more than 60 days in coastal Alaska over the last 30 years. (Figure source: adapted from Parkinson 2014⁵⁰ ).

It is very likely that there is an anthropogenic contribution to the observed arctic sea ice decline since 1979. A range of modeling studies analyzing the September sea ice extent trends in simulations with and without anthropogenic forcing conclude that these declines cannot be explained by natural variability alone.^{54 ,55 ,56 ,57 ,58 ,59} Further, observational-based analyses considering a range of anthropogenic and natural forcing mechanisms for September sea ice loss reach the same conclusion.⁶⁰ Considering the occurrence of individual September sea ice anomalies, internal climate variability alone very likely could not have caused recently observed record low arctic sea ice extents, such as in September 2012.^{61 ,62} The potential contribution of natural variability to arctic sea ice trends is significant.^{55 ,63 ,64} One recent study²⁸ indicates that internal variability dominates arctic atmospheric circulation trends, accounting for 30%–50% of the sea ice reductions since 1979, and up to 60% in September. However, previous studies indicate that the contributions from internal variability are smaller than 50%.^{54 ,55} This apparent significant contribution of natural variability to sea ice decline indicates that natural variability alone cannot explain the observed sea ice decline and is consistent with the statement that it is very likely there is an anthropogenic contribution to the observed arctic sea ice decline since 1979.

Continued sea ice loss is expected across the Arctic, which is very likely to result in late summers becoming nearly ice-free (areal extent less than 10⁶ km² or approximately 3.9 × 10⁵ mi²) by the 2040s.^{21 ,65} Natural variability,⁶⁶ future scenarios, and model uncertainties^{64 ,67 ,68} all influence sea ice projections. One study suggests that internal variability alone accounts for a 20-year prediction uncertainty in the timing of the first occurrence of an ice-free summer, whereas differences between a higher scenario (RCP8.5) and a lower scenario (RCP4.5) add only 5 years.⁶³ Projected September sea ice reductions by 2081–2100 range from 43% for an even lower scenario (RCP2.6) to 94% for RCP8.5.²¹ However, September sea ice projections over the next few decades are similar for the different anthropogenic forcing associated with these scenarios; scenario dependent sea ice loss only becomes apparent after 2050. Another study⁶⁹ indicates that the total sea ice loss scales roughly linearly with CO₂ emissions, such that an additional 1,000 GtC from present day levels corresponds to ice-free conditions in September. A key message from the Third National Climate Assessment (NCA3)⁷⁰ was that arctic sea ice is disappearing. The fundamental conclusion of this assessment is unchanged; additional research corroborates the NCA3 statement.

11.2.3 Arctic Ocean and Marginal Seas

SEA SURFACE TEMPERATURE

Arctic Ocean sea surface temperatures (SSTs) have increased since comprehensive records became available in 1982. Satellite-observed Arctic Ocean SSTs, poleward of 60°N, exhibit a trend of 0.16° ± 0.02°F (0.09° ± 0.01°C) per decade.¹⁹ Arctic Ocean SST is controlled by a combination of factors, including solar radiation and energy transport from ocean currents and atmospheric winds. Summertime Arctic Ocean SST trends and patterns strongly couple with sea ice extent; however, clouds, ocean color, upper-ocean thermal structure, and atmospheric circulation also play a role.^{40 ,71} Along coastal Alaska, SSTs in the Chukchi Sea exhibit a statistically significant (95% confidence) trend of 0.9° ± 0.5°F (0.5° ± 0.3°C) per decade.⁷²

Arctic Ocean temperatures also increased at depth.^{71 ,73} Since 1970, Arctic Ocean Intermediate Atlantic Water—located between 150 and 900 meters—has warmed by 0.86° ± 0.09°F (0.48° ± 0.05°C) per decade; the most recent decade being the warmest.⁷³ The observed temperature level is unprecedented in the last 1,150 years for which proxy indicators provide records.^{74 ,75} The influence of Intermediate Atlantic Water warming on future Alaska and arctic sea ice loss is unclear.^{38 ,76}

ALASKAN SEA LEVEL RISE

The Alaskan coastline is vulnerable to sea level rise (SLR); however, strong regional variability exists in current trends and future projections. Some regions are experiencing relative sea level fall, whereas others are experiencing relative sea level rise, as measured by tide gauges that are part of NOAA’s National Water Level Observation Network. These tide gauge data show sea levels rising fastest along the northern coast of Alaska but still slower than the global average, due to isostatic rebound (Ch. 12: Sea Level Rise).⁷⁷ However, considerable uncertainty in relative sea level rise exists due to a lack of tide gauges; for example, no tide gauges are located between Bristol Bay and Norton Sound or between Cape Lisburne and Prudhoe Bay. Under almost all future scenarios, SLR along most of the Alaskan coastline is projected to be less than the global average (Ch. 12: Sea Level Rise).

SALINITY

Arctic Ocean salinity influences the freezing temperature of sea ice (less salty water freezes more readily) and the density profile representing the integrated effects of freshwater transport, river runoff, evaporation, and sea ice processes. Arctic Ocean salinity exhibits multidecadal variability, hampering the assessment of long-term trends.⁷⁸ Emerging evidence suggests that the Arctic Ocean and marginal sea salinity has decreased in recent years despite short-lived regional salinity increases between 2000 and 2005.⁷¹ Increased river runoff, rapid melting of sea and land ice, and changes in freshwater transport have influenced observed Arctic Ocean salinity.^{71 ,79}

OCEAN ACIDIFICATION

Arctic Ocean acidification is occurring at a faster rate than the rest of the globe (see also Ch. 13: Ocean Changes).⁸⁰ Coastal Alaska and its ecosystems are especially vulnerable to ocean acidification because of the high sensitivity of Arctic Ocean water chemistry to changes in sea ice, respiration of organic matter, upwelling, and increasing river runoff.⁸⁰ Sea ice loss and a longer melt season contribute to increased vulnerability of the Arctic Ocean to acidification by lowering total alkalinity, permitting greater upwelling, and influencing the primary production characteristics in coastal Alaska.^{81 ,82 ,83 ,84 ,85 ,86} Global-scale modeling studies suggest that the largest and most rapid changes in pH will continue along Alaska’s coast, indicating that ocean acidification may increase enough by the 2030s to significantly influence coastal ecosystems.⁸⁰

11.2.4 Boreal Wildfires

Alaskan wildfire activity has increased in recent decades. This increase has occurred both in the boreal forest⁸⁷ and in the arctic tundra,⁸⁸ where fires historically were smaller and less frequent. A shortened snow cover season and higher temperatures over the last 50 years⁸⁹ make the Arctic more vulnerable to wildfire.^{87 ,88 ,90} Total area burned and the number of large fires (those with area greater than 1,000 km² or 386 mi²) in Alaska exhibit significant interannual and decadal variability, from influences of atmospheric circulation patterns and controlled burns, but have likely increased since 1959.⁹¹ The most recent decade has seen an unusually large number of years with anomalously large wildfires in Alaska.⁹² Studies indicate that anthropogenic climate change has likely lengthened the wildfire season and increased the risk of severe fires.⁹³ Further, wildfire risks are expected to increase through the end of the century due to warmer, drier conditions.^{90 ,94} Using climate simulations to force an ecosystem model over Alaska (Alaska Frame-Based Ecosystem Code, ALFRESCO), the total area burned is projected to increase between 25% and 53% by 2100.⁹⁵ A transition into a regime of fire activity unprecedented in the last 10,000 years is possible.⁹⁶ We conclude that there is medium confidence for a human-caused climate change contribution to increased forest fire activity in Alaska in recent decades. See Chapter 8: Drought, Floods, and Wildfires for more details.

A significant amount of the total global soil carbon is found in the boreal forest and tundra ecosystems, including permafrost.^{97 ,98 ,99} Increased fire activity could deplete these stores, releasing them to the atmosphere to serve as an additional source of atmospheric CO₂.^{97 ,100} Increased fires may also enhance the degradation of Alaska’s permafrost by blackening the ground, reducing surface albedo, and removing protective vegetation.^{101 ,102 ,103 ,104}

11.2.5 Snow Cover

Snow cover extent has significantly decreased across the Northern Hemisphere and Alaska over the last decade (see also Ch. 7: Precipitation Change and Ch. 10: Land Cover).^{105 ,106} Northern Hemisphere June snow cover decreased by more than 65% between 1967 and 2012,^{37 ,107} at a trend of −17.2% per decade since 1979.⁸⁹ June snow cover dipped below 3 million square km (approximately 1.16 million square miles) for the fifth time in six years between 2010 and 2015, a threshold not crossed in the previous 43 years of record.⁸⁹ Early season snow cover in May, which affects the accumulation of solar insolation through the summer, has also declined at −7.3% per decade, due to reduced winter accumulation from warmer temperatures. Regional trends in snow cover duration vary, with some showing earlier onsets while others show later onsets.⁸⁹ In Alaska, the 2016 May statewide snow coverage of 595,000 square km (approximately 372,000 square miles) was the lowest on record dating back to 1967; the snow coverage of 2015 was the second lowest, and 2014 was the fourth lowest.

Human activities have very likely contributed to observed snow cover declines over the last 50 years. Attribution studies indicate that observed trends in Northern Hemisphere snow cover cannot be explained by natural forcing alone, but instead require anthropogenic forcing.^{24 ,106 ,108} Declining snow cover is expected to continue and will be affected by both the anthropogenic forcing and evolution of arctic ecosystems. The observed tundra shrub expansion and greening^{109 ,110} affects melt by influencing snow depth, melt dynamics, and the local surface energy budget. Nevertheless, model simulations show that future reductions in snow cover influence biogeochemical feedbacks and warming more strongly than changes in vegetation cover and fire in the North American Arctic.¹¹¹

11.2.6 Continental Ice Sheets and Mountain Glaciers

Mass loss from ice sheets and glaciers influences sea level rise, the oceanic thermohaline circulation, and the global energy budget. Moreover, the relative contribution of GrIS to global sea level rise continues to increase, exceeding the contribution from thermal expansion (see Ch. 12: Sea Level Rise). Observational and modeling studies indicate that GrIS and glaciers in Alaska are out of mass balance with current climate conditions and are rapidly losing mass.^{37 ,112} In recent years, mass loss has accelerated and is expected to continue.^{112 ,113}

 

Figure 11.3

VIEW

Time series of the cumulative climatic mass balance (units: kg/m²) in five arctic regions and for the Pan-Arctic from the World Glacier Monitoring Service (WGMS;²¹⁰ Wolken et al.;²¹¹ solid lines, left y-axis), plus Alaskan glacial mass loss observed from NASA GRACE¹¹³ (dashed blue line, right y-axis). (Figure source: Harig and Simons 2016¹¹³ and Wolken et al. 2016;²¹¹ © American Meteorological Society, used with permission.)

Dramatic changes have occurred across GrIS, particularly at its margins. GrIS average annual mass loss from January 2003 to May 2013 was −244 ± 6 Gt per year (approximately 0.26 inches per decade sea level equivalent).¹¹³ One study indicates that ice mass loss from Greenland was −269 Gt per year between April 2002 and April 2016.⁴⁷ Increased surface melt, runoff, and increased outlet glacier discharge from warmer air temperatures are primary contributing factors.^{114 ,115 ,116 ,117 ,118} The effects of warmer air and ocean temperatures on GrIS can be amplified by ice dynamical feedbacks, such as faster sliding, greater calving, and increased submarine melting.^{116 ,119 ,120 ,121} Shallow ocean warming and regional ocean and atmospheric circulation changes also contribute to mass loss.^{122 ,123 ,124} The underlying mechanisms of the recent discharge speed-up remain unclear;^{125 ,126} however, warmer subsurface ocean and atmospheric temperatures^{118 ,127 ,128} and meltwater penetration to the glacier bed^{125 ,129} very likely contribute.

 

Figure 11.4

VIEW

Two northeast-looking photographs of the Muir Glacier located in southeastern Alaska taken from a Glacier Bay Photo station in (a) 1941 and (b) 2004. U.S. Geological Survey repeat photography allows the tracking of glacier changes, illustrating that between 1941 and 2004 the Muir Glacier has retreated more than 4 miles to the northwest and out of view. Riggs Glacier (in view) is a tributary to Muir Glacier and has retreated by as much as 0.37 miles and thinned by more than 0.16 miles. The photographs also illustrate a significant change in the surface type between 1941 and 2004 as bare rock in the foreground has been replaced by dense vegetation (Figure source: USGS 2004²¹² ).

Annual average ice mass from Arctic-wide glaciers has decreased every year since 1984,^{112 ,130 ,131} with significant losses in Alaska, especially over the past two decades (Figure 11.3).^{37 ,132} Figure 11.4 illustrates observed changes from U.S. Geological Survey repeat photography of Alaska’s Muir Glacier, retreating more than 4 miles between 1941 and 2004, and its tributary the Riggs Glacier. Total glacial ice mass in the Gulf of Alaska region has declined steadily since 2003.¹¹³ NASA’s Gravity Recovery and Climate Experiment (GRACE) indicates mass loss from the northern and southern parts of the Gulf of Alaska region of −36 ± 4 Gt per year and −4 ± 3 Gt per year, respectively.¹¹³ Studies suggest an anthropogenic imprint on imbalances in Alaskan glaciers, indicating that melt will continue through the 21st century.^{112 ,133 ,134} Multiple datasets indicate that it is virtually that Alaskan glaciers have lost mass over the last 50 years and will continue to do so.¹³⁵

11.3: Arctic Feedbacks on the Lower 48 and Globally

11.3.1 Linkages between Arctic Warming and Lower Latitudes

Midlatitude circulation influences arctic climate and climate change.^{11 ,136 ,137 ,138 ,139 ,140 ,141 ,142 ,143 ,144 ,145} Record warm arctic temperatures in winter 2016 resulted primarily from the transport of midlatitude air into the Arctic, demonstrating the significant midlatitude influence.¹⁴⁶ Emerging science demonstrates that warm, moist air intrusions from midlatitudes results in increased downwelling longwave radiation, warming the arctic surface and hindering wintertime sea ice growth.^{139 ,141 ,147 ,148}

The extent to which enhanced arctic surface warming and sea ice loss influence the large-scale atmospheric circulation and midlatitude weather and climate extremes has become an active research area.^{137 ,146} Several pathways have been proposed (see references in Cohen et al.¹⁴⁹ and Barnes and Screen¹⁵⁰ ): reduced meridional temperature gradient, a more sinuous jet-stream, trapped atmospheric waves, modified storm tracks, weakened stratospheric polar vortex. While modeling studies link a reduced meridional temperature gradient to fewer cold temperature extremes in the continental United States,^{151 ,152 ,153 ,154} other studies hypothesize that a slower jet stream may amplify Rossby waves and increase the frequency of atmospheric blocking, causing more persistent and extreme weather in midlatitudes.¹⁵⁵

Multiple observational studies suggest that the concurrent changes in the Arctic and Northern Hemisphere large-scale circulation since the 1990s did not occur by chance, but were caused by arctic amplification.^{149 ,150 ,156} Reanalysis data suggest a relationship between arctic amplification and observed changes in persistent circulation phenomena like blocking and planetary wave amplitude.^{155 ,157 ,158} The recent multi-year California drought serves as an example of an event caused by persistent circulation phenomena (see Ch. 5: Circulation and Variability and Ch. 8: Drought, Floods, and Wildfires).^{159 ,160 ,161} Robust empirical evidence is lacking because the arctic sea ice observational record is too short¹⁶² or because the atmospheric response to arctic amplification depends on the prior state of the atmospheric circulation, reducing detectability.¹⁴⁶ Furthermore, it is not possible to draw conclusions regarding the direction of the relationship between arctic warming and midlatitude circulation based on empirical correlation and covariance analyses alone. Observational analyses have been combined with modeling studies to test causality statements.

Studies with simple models and Atmospheric General Circulation Models (AGCMs) provide evidence that arctic warming can affect midlatitude jet streams and location of storm tracks.^{137 ,146 ,150} In addition, analysis of CMIP5 models forced with increasing greenhouse gases suggests that the magnitude of arctic amplification affects the future midlatitude jet position, specifically during boreal winter.¹⁶³ However, the effect of arctic amplification on blocking is not clear (Ch. 5: Circulation and Variability).¹⁶⁴

Regarding attribution, AGCM simulations forced with observed changes in arctic sea ice suggest that the sea ice loss effect on observed recent midlatitude circulation changes and winter climate in the continental United States is small compared to natural large-scale atmospheric variability.^{142 ,144 ,154 ,165} It is argued, however, that climate models do not properly reproduce the linkages between arctic amplification and lower latitude climate due to model errors, including incorrect sea ice–atmosphere coupling and poor representation of stratospheric processes.^{137 ,166}

In summary, emerging science demonstrates a strong influence of the midlatitude circulation on the Arctic, affecting temperatures and sea ice (high confidence). The influence of arctic changes on the midlatitude circulation and weather patterns are an area of active research. Currently, confidence is low regarding whether or by what mechanisms observed arctic warming may have influenced midlatitude circulation and weather patterns over the continental United States. The nature and magnitude of arctic amplification’s influence on U.S. weather over the coming decades remains an open question.

11.3.2 Freshwater Effects on Ocean Circulation

The addition of freshwater to the Arctic Ocean from melting sea ice and land ice can influence important arctic climate system characteristics, including ocean salinity, altering ocean circulation, density stratification, and sea ice characteristics. Observations indicate that river runoff is increasing, driven by land ice melt, adding freshwater to the Arctic Ocean.¹⁶⁷ Melting arctic sea and land ice combined with time-varying atmospheric forcing^{79 ,168} control Arctic Ocean freshwater export to the North Atlantic. Large-scale circulation variability in the central Arctic not only controls the redistribution and storage of freshwater in the Arctic⁷⁹ but also the export volume.¹⁶⁹ Increased freshwater fluxes can weaken open ocean convection and deep water formation in the Labrador and Irminger seas, weakening the Atlantic meridional overturning circulation (AMOC).^{170 ,171} AMOC-associated poleward heat transport substantially contributes to North American and continental European climate; any AMOC slowdown could have implications for global climate change as well (see Ch. 15: Potential Surprises).^{172 ,173} Connections to subarctic ocean variations and the Atlantic Meridional Overturning Circulation have not been conclusively established and require further investigation (see Ch. 13: Ocean Changes).

11.3.3 Permafrost–Carbon Feedback

Alaska and arctic permafrost characteristics have responded to increased temperatures and reduced snow cover in most regions since the 1980s.¹³⁰ The permafrost warming rate varies regionally; however, colder permafrost is warming faster than warmer permafrost.^{37 ,174} This feature is most evident across Alaska, where permafrost on the North Slope is warming more rapidly than in the interior. Permafrost temperatures across the North Slope at various depths ranging from 39 to 65 feet (12 to 20 meters) have warmed between 0.3° and 1.3°F (0.2° and 0.7°C) per decade over the observational period (Figure 11.5).¹⁷⁵ Permafrost active layer thickness increased across much of the Arctic while showing strong regional variations.^{37 ,130 ,176} Further, recent geologic survey data indicate significant permafrost thaw slumping in northwestern Canada and across the circumpolar Arctic that indicate significant ongoing permafrost thaw, potentially priming the region for more rapid thaw in the future.¹⁷⁷ Continued degradation of permafrost and a transition from continuous to discontinuous permafrost is expected over the 21st century.^{37 ,178 ,179}

 

Figure 11.5

VIEW

Time series of annual mean permafrost temperatures (units: °F) at various depths from 39 to 65 feet (12 to 20 meters) from 1977 through 2015 at several sites across Alaska, including the North Slope continuous permafrost region (purple/blue/green shades), and the discontinuous permafrost (orange/pink/red shades) in Alaska and northwestern Canada. Solid lines represent the linear trends drawn to highlight that permafrost temperatures are warming faster in the colder, coastal permafrost regions than the warmer interior regions. (Figure Source: adapted from Romanovsky et al. 2016;¹⁷⁵ © American Meteorological Society, used with permission.)

Permafrost contains large stores of carbon. Though the total contribution of these carbon stores to global methane emission is uncertain, Alaska’s permafrost contains rich and vulnerable organic carbon soils.^{99 ,179 ,180} Thus, warming Alaska permafrost is a concern for the global carbon cycle as it provides a possibility for a significant and potentially uncontrollable release of carbon, complicating the ability to limit global temperature increases. Current methane emissions from Alaskan arctic tundra and boreal forests contribute a small fraction of the global methane (CH₄) budget.¹⁸¹ However, gas flux measurements have directly measured the release of CO₂ and CH₄ from arctic permafrost.¹⁸² Recent measurements indicate that cold season methane emissions (after snowfall) are greater than summer emissions in Alaska, and methane emissions in upland tundra are greater than in wetland tundra.¹⁸³

The permafrost–carbon feedback represents the additional release of CO₂ and CH₄ from thawing permafrost soils providing additional radiative forcing, a source of a potential surprise (Ch. 15: Potential Surprises).¹⁸⁴ Thawing permafrost makes previously frozen organic matter available for microbial decomposition, producing CO₂ and CH₄. The specific condition under which microbial decomposition occurs, aerobic or anaerobic, determines the proportion of CO₂ and CH₄ released. This distinction has potentially significant implications, as CH₄ has a 100-year global warming potential 35 times that of CO₂.¹⁸⁵ Emerging science indicates that 3.4 times more carbon is released under aerobic conditions than anaerobic conditions, and 2.3 times more carbon after accounting for the stronger greenhouse effect of CH₄.¹⁸⁶ Additionally, CO₂ and CH₄ production strongly depends on vegetation and soil properties.¹⁸⁴

Combined data and modeling studies indicate a positive permafrost–carbon feedback with a global sensitivity between −14 and −19 GtC per °C (approximately −25 to −34 GtC per °F) soil carbon loss^{187 ,188} resulting in a total 120 ± 85 GtC release from permafrost by 2100 and an additional global temperature increase of 0.52° ± 0.38°F (0.29° ± 0.21°C) by the permafrost–carbon feedback.¹⁸⁹ More recently, Chadburn et al.¹⁹⁰ infer a −4 million km² per °C (or approximately 858,000 mi² per °F) reduction in permafrost area to globally averaged warming at stabilization by constraining climate models with the observed spatial distribution of permafrost; this sensitivity is 20% higher than previous studies. In the coming decades, enhanced high-latitude plant growth and its associated CO₂ sink should partially offset the increased emissions from permafrost thaw;^{179 ,189 ,191} thereafter, decomposition is expected to dominate uptake. Permafrost thaw is occurring faster than models predict due to poorly understood deep soil, ice wedge, and thermokarst processes.^{188 ,192 ,193} Additionally, uncertainty stems from the surprising uptake of methane from mineral soils.¹⁹⁴ There is high confidence in the positive sign of the permafrost–carbon feedback, but low confidence in the feedback magnitude.

11.3.4 Methane Hydrate Instability

Significant stores of CH₄, in the form of methane hydrates (also called clathrates), lie within and below permafrost and under the global ocean on continental margins. The estimated total global inventory of methane hydrates ranges from 500 to 3,000 GtC^{195 ,196 ,197} with a central estimate of 1,800 GtC.¹⁹⁸ Methane hydrates are solid compounds formed at high pressures and cold temperatures, trapping methane gas within the crystalline structure of water. Methane hydrates within upper continental slopes of the Pacific, Atlantic, and Gulf of Mexico margins and beneath the Alaskan arctic continental shelf may be vulnerable to small increases in ocean temperature.^{197 ,198 ,199 ,200 ,201 ,202 ,203}

Rising sea levels and warming oceans have a competing influence on methane hydrate stability.^{199 ,204} Studies indicate that the temperature effect dominates and that the overall influence is very likely a destabilizing effect.¹⁹⁸ Projected warming rates for the 21st century Arctic Ocean are not expected to lead to sudden or catastrophic destabilization of seafloor methane hydrates.²⁰⁵ Recent observations indicate increased CH₄ emission from the arctic seafloor near Svalbard; however, these emissions are not reaching the atmosphere.^{198 ,206}

References