Recent Arctic amplification and extreme mid-latitude weather

Abstract

The Arctic cryosphere is an integral part of Earth's climate sys-tem and has undergone unprecedented changes within the past few decades. Rapid warming and sea-ice loss has had significant impacts locally, particularly in late summer and early autumn. September sea ice has declined at a rate of 12.4% per dec-ade since 1979 (ref.1), so that by summer 2012, nearly half of the areal coverage had disappeared. This decrease in ice extent has been accompanied by an approximately 1.8 m (40%) decrease in mean winter ice thickness since 1980 (ref.2) and a 75–80% loss in volume 3 . Though sea-ice loss has received most of the research and media attention, snow cover in spring and summer has decreased at an even greater rate than sea ice. June snow cover alone has decreased at nearly double the rate of September sea ice 4 . The decrease in spring snow cover has contributed to both the rise in warm season surface temperatures over the Northern Hemisphere extratropical landmasses and the decrease in summer Arctic sea ice 5 . The com-bined rapid loss of sea ice and snow cover in the spring and sum-mer has played a role in amplifying Arctic warming. However, snow cover and sea-ice trends diverge in the autumn and winter with sea ice decreasing in all months while snow cover has exhibited a neutral to positive trend in autumn and winter 6

Full text

NATURE GEOSCIENCE | ADVANCE ONLINE PUBLICATION | www.nature.com/naturegeoscience 1
The Arctic cryosphere is an integral part of Earth’s climate sys-
tem and has undergone unprecedented changes within the
past few decades. Rapid warming and sea-ice loss has had
signicant impacts locally, particularly in late summer and early
autumn. September sea ice has declined at a rate of 12.4% per dec-
ade since 1979 (ref.1), so that by summer 2012, nearly half of the
areal coverage had disappeared. is decrease in ice extent has been
accompanied by an approximately 1.8m (40%) decrease in mean
winter ice thickness since 1980 (ref.2) and a 75–80% loss in volume3.
ough sea-ice loss has received most of the research and media
attention, snow cover in spring and summer has decreased at an
even greater rate than sea ice. June snow cover alone has decreased
at nearly double the rate of September sea ice4. e decrease in
spring snow cover has contributed to both the rise in warm season
surface temperatures over the Northern Hemisphere extratropical
landmasses and the decrease in summer Arctic sea ice5. e com-
bined rapid loss of sea ice and snow cover in the spring and sum-
mer has played a role in amplifying Arctic warming. However, snow
cover and sea-ice trends diverge in the autumn and winter with
sea ice decreasing in all months while snow cover has exhibited a
neutral to positive trend in autumn and winter6.
Climate change and Arctic amplification
While the global-mean surface temperature has unequivocally risen
over the instrumental record7, spatial heterogeneity of this warming
plays an important role in the resulting climate impacts. In particu-
lar, the near-surface of the Northern Hemisphere high latitudes are
warming at rates double that of lower latitudes8–10. is observed
Recent Arctic amplification and extreme
mid-latitude weather
Judah Cohen1*, James A. Screen2, Jason C. Furtado1, Mathew Barlow3,4, David Whittleston5,
Dim Coumou6, Jennifer Francis7, Klaus Dethlo8, Dara Entekhabi5, James Overland9 and Justin Jones1
The Arctic region has warmed more than twice as fast as the global average — a phenomenon known as Arctic amplification.
The rapid Arctic warming has contributed to dramatic melting of Arctic sea ice and spring snow cover, at a pace greater than
that simulated by climate models. These profound changes to the Arctic system have coincided with a period of ostensibly more
frequent extreme weather events across the Northern Hemisphere mid-latitudes, including severe winters. The possibility of
a link between Arctic change and mid-latitude weather has spurred research activities that reveal three potential dynamical
pathways linking Arctic amplification to mid-latitude weather: changes in storm tracks, the jet stream, and planetary waves
and their associated energy propagation. Through changes in these key atmospheric features, it is possible, in principle, for sea
ice and snow cover to jointly influence mid-latitude weather. However, because of incomplete knowledge of how high-latitude
climate change influences these phenomena, combined with sparse and short data records, and imperfect models, large uncer-
tainties regarding the magnitude of such an influence remain. We conclude that improved process understanding, sustained
and additional Arctic observations, and better coordinated modelling studies will be needed to advance our understanding of
the influences on mid-latitude weather and extreme events.
phenomenon (Figs1 and 2a,b) is termed polar or Arctic amplica-
tion. Arctic amplication occurs in all seasons, but is strongest in
autumn and winter. It is also a consistent feature in coupled climate
model simulations of the recent past and future projections forced
with increased greenhouse-gas concentrations11,12. Several pro-
cesses are thought to contribute to Arctic amplication, including
local radiative eects from increased greenhouse-gas forcing12,13,
changes in the snow- and ice-albedo feedback induced by a dimin-
ishing cryosphere14–16, aerosol concentration changes and deposits
of black carbon on snow and ice surfaces17, changes in Arctic cloud
cover and water vapour content18,19, and a relatively smaller increase
in emission of longwave radiation to space in the Arctic compared
with the tropics for the same temperature increase20. In addition
to these local drivers of Arctic amplication, Arctic temperature
change is sensitive to variations in the poleward transport of heat
and moisture into the Arctic from lower latitudes16,21.
Rapid Arctic warming has been accompanied by extensive loss
of sea ice9. Arctic sea ice strongly modulates near-surface conditions
at high latitudes, which then inuences regional and, potentially,
remote climate. Because open water has a much lower albedo than
ice, more sunlight is absorbed at the ocean surface, where sea ice has
recently receded in the Arctic. More absorbed energy has resulted
in 4–5°C sea surface temperature anomalies in these newly ice-free
regions22. However, during autumn when the air cools to tempera-
tures lower than the ocean surface, the excess heat absorbed during
summer is transferred from the ocean to the atmosphere via radia-
tive and turbulent uxes, which strongly warms the lower Arctic
troposphere. e additional heat in the system slows the formation
1Atmospheric and Environmental Research, Inc., Lexington, Massachusetts 02421, USA, 2College of Engineering, Mathematicsand Physical Sciences,
University of Exeter, Exeter, Devon EX4 4QF, UK, 3Department of Environmental, Earth, and Atmospheric Sciences, University of Massachusetts Lowell,
Lowell, Massachusetts 01854, USA, 4The Climate Change Initiative, University of Massachusetts Lowell, Lowell, Massachusetts 01854, USA, 5Department
of Civil and Environmental Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts, 02139, USA, 6Potsdam Institute for Climate
Impact Research — Earth System Analysis, 14412 Potsdam, Germany, 7Institute for Marine and Coastal Sciences, Rutgers University, New Brunswick,
New Jersey 08901, USA, 8Alfred Wegener Institute, Helmholtz Centre for Polar and Marine Research, AWI Potsdam 14473, Germany, 9Pacific Marine
Environmental Laboratory, Seattle, Washington 98115, USA. *e-mail: jcohen@aer.com
REVIEW ARTICLE
PUBLISHED ONLINE: 17 AUGUST 2014 | DOI: 10.1038/NGEO2234
© 2014 Macmillan Publishers Limited. All rights reserved

2 NATURE GEOSCIENCE | ADVANCE ONLINE PUBLICATION | www.nature.com/naturegeoscience
of sea ice through winter, both in extent and, especially, thick-
ness23,24. Hence, winter sea ice has thinned2, enabling easier melting,
fracturing and/or mobility of the ice cover. e increased fraction of
open water in winter generates warmer, moister air masses over the
Arctic Ocean and nearby continents15,25, weakening the meridional
near-surface temperature gradient. erefore, these feedbacks indi-
cate that observed Arctic sea-ice loss acts as both a response to and
a driver of Arctic amplication.
Mid-latitude extreme weather
A large number of extreme heat and rainfall events have been
reported over the past decade, especially in the Northern Hemisphere
mid-latitudes26–31. Figure3 illustrates that several standard extreme
temperature and precipitation indices have increased in frequency
and intensity over mid-latitude land areas (20–50°N) with espe-
cially rapid changes since the 1990s. For example, the amount of
precipitation on very wet days (exceeding the 95th percentile) has
increased from 160 to 185mm, and the percentage of warm days
(exceeding the 90th percentile) has increased from 10% before 1980
to 16% at present32.
Extreme weather has not been limited to heavy rainfall and warm
temperatures and recently has included cold extremes as well. Winter
temperatures have generally warmed since 1960 (Fig.2a), and the fre-
quency of anomalously cold winter days has decreased over mid-to-
high latitudes, but primarily north of 50°N, since 1979 in response to
mean warming and decreased variability33. However, also evident in
Fig.3d,f is that the number of days continuously below freezing has
increased and the minimum temperatures have decreased since 1990.
Figure3h also indicates that the frequency of unusually cold winter
months (colder than two standard deviations below the 1951–1980
mean30) had reversed its longer-term downward trend by the end of
the 1990s. is trend reversal in cold extremes has coincided with an
acceleration in the rate of warming at high latitudes relative to the rest
of the Northern Hemisphere starting approximately in 1990 (Fig.2b).
As seen in Fig.2c, continental winter temperature trends since 1990
exhibit cooling over the mid-latitudes, replacing the warming trends
observed over the longer period since 1960 (Fig.2a). e winter tem-
perature trends shown in Fig.2c start in 1990 but are not sensitive to
the exact start date. However, on average, daily winter cold extremes
were less severe over this period than they have been historically33.
e rapid Arctic warming implies that cold air outbreaks, when Arctic
air moves south into the mid-latitudes, are becoming less severe33.
e seven years between 2007 and 2013 have exhibited the low-
est minimum sea-ice extents recorded in September since satellite
observations began, with an all-time record low in 2007 followed by
another in 2012, when sea-ice extent fell below 4million km2 for the
rst time in the observational record. Several of these seven winters
following the low sea-ice minima have been unusually cold across
the Northern Hemisphere extratropical landmasses34–38. e recent
winter of 2013–2014 was characterized by record cold and wide-
spread snowstorms across the eastern United States and Canada
with the most intense cold-air outbreak in decades associated with
the weakening of the polar vortex39. e persistent and harsh cold
resulted in all-time record cold winters around the Great Lakes of
the United States since record keeping began in the 1870s.
e media and public have been quick to make the connection
between global, and in particular Arctic, warming and extreme
weather40. While global warming theory is consistent with record
warm temperatures and more intense precipitation events, it does
not directly explain cold extremes. Coupled models project boreal
winter amplication under greenhouse-gas forcing, where the
Northern Hemisphere landmasses would warm faster in winter rel-
ative to the other seasons11,41. Warming in the Arctic has continued
unabated since at least 1960. Longer-term observed temperature
trends in mid-latitudes are consistent with these projections, while
shorter-term trends are not. is highlights that results are sensi-
tive to the spatial extent of the analysis, the exact denition used
and especially the duration of an extreme, as extremes of diering
durations may be driven by dierent physical processes.
While cold extremes may be mostly due to natural variability, a
growing number of recent studies argue that recent extreme winter
weather is related to Arctic amplication. ree possible dynamical
pathways through which Arctic amplication may inuence mid-
latitude weather, including extreme weather, are summarized below.
We focus our discussion on Arctic linkages to mid-latitude weather
in the winter season for two reasons. First, most studies that have
linked Arctic amplication to mid-latitude weather have focused
on winter (a brief discussion of proposed linkages in other seasons,
mainly summer, is provided in the Supplementary Information).
Second, winter is the season in which mid-latitude temperature
trends have diverged most notably from both model projections
and from the other seasons42. To provide a focused review, we limit
our consideration to the literature concerning recent past (mid-
twentieth century onwards) and present-day climate variability and
trends. e implications of projected future Arctic amplication
(for example, at the end of the twenty-rst century) are likely large
and wide ranging, but are not considered here.
Arctic amplification influences and uncertainties
Whether to attribute severe winter weather to Arctic amplica-
tion or natural variability has emerged as a major debate among
 
1,000
800
600
400
90° 70° 50° 30° 10°


0
1 
1,000
800
600
400
90° 70° 50° 30° 10°


0
1
 
1,000
800
600
400
90° 70° 50° 30° 10°


0
1 
1,000
800
600
400
90° 70° 50° 30° 10°


0
1
-0.6 -0.3 0.0 0.30.60.9 1.2 1.5
Temperature trend (°C per decade)
Level (hPa)
Latitude (N)
ab
cd
Figure 1 | Polar amplification of temperature trends, 1979–2014. Zonally
averaged temperature trends averaged around circles of latitude for
a, winter (December–February), b, spring (March–May), c, summer
(June–August) and d, autumn (September–November). Trends are based
on ERA-Interim reanalysis data95 from March 1979 to February 2014.
The black contours indicate where trends dier significantly from zero at
the 99% (solid lines) and 95% (dotted lines) confidence levels. The line
graphs show trends (same units as in colour plots) averaged over the lower
part of the atmosphere (950–1,000hPa; solid lines) and over the entire
atmospheric column (300–1,000hPa; dotted lines)9.
REVIEW ARTICLE
NATURE GEOSCIENCE DOI: 10.1038/NGEO2234
© 2014 Macmillan Publishers Limited. All rights reserved

NATURE GEOSCIENCE | ADVANCE ONLINE PUBLICATION | www.nature.com/naturegeoscience 3
scientists43–45. In the observations, Arctic amplication has
separated from the noise of natural variability only in the past
approximately two decades (Fig. 2b), presenting a challenge for
the detection of robust atmospheric responses to Arctic ampli-
cation, including mid-latitude weather, over such a short time
period. In addition to the relatively short length of the observa-
tional record, the Arctic is poorly sampled. A major caveat of any
observational study is that correlation alone cannot demonstrate a
causal link. Cause and eect can be established through sensitiv-
ity or perturbation studies using climate models, but models are
subject to their own deciencies. Known model errors include
sea-ice–atmosphere coupling46,47, energy uxes and cloud proper-
ties47. Furthermore, modelling studies of the eects of sea-ice loss
on large-scale atmospheric circulation have produced conict-
ing results that make interpretation dicult. Finally, our under-
standing of fundamental driving forces of mid-latitude weather
is incomplete48.
Given these sources of uncertainty, a consensus on whether and
how Arctic amplication is inuencing mid-latitude weather is
lacking. To facilitate advancement on this important issue, there-
fore, we synthesize key ndings that argue for and against a signi-
cant link between Arctic amplication and mid-latitude weather.
All studies agree that the rst order impact of sea-ice melt is to
modify the boundary layer in the Arctic15,25. However, if and how
that signal propagates out of the Arctic to mid-latitudes diers
and can be loosely grouped under three broad dynamical frame-
works: (1) changes in storm tracks mainly in the North Atlantic
sector; (2) changes in the characteristics of the jet stream; and (3)
regional changes in the tropospheric circulation that trigger anom-
alous planetary wave congurations. In Fig.4, we show the known
primary inuences on mid-latitude weather, including the three
dynamical pathways introduced above and described in more detail
in the following sections. We recognize that these three pathways
are not distinct as they involve dynamical features of the atmos-
pheric circulation that are highly interconnected. Whilst imperfect,
our choice of this separation reects the dierent dynamical frame-
works that are commonly used — if not explicitly acknowledged —
to study the dynamics of mid-latitude weather.
Figure 2 | Winter temperature trends since 1960 and over the most recent period from 1990. a, Right: linear trend (°C per 10years) in December–
February (DJF) mean surface air temperatures from 1960–1961 to 2013–2014. Shading interval every 0.1°C per 10years. Dark grey indicates points with
insucient samples to calculate a trend. Left: The zonally averaged linear trend (°C per 10years). b, Area-average surface temperature anomalies (°C)
from 0° to 60°N (solid black line) and 60° to 90°N (solid red line) along with five-year smoothing (dashed black and red lines, respectively). c, As in
panel a but from 1960–1961 to 2013–2014. Shading interval every 0.2°C per 10years. Also note dierent scales between a and c. Data from the National
Aeronautics and Space Administration Goddard Institute for Space Studies temperature analysis (http://data.giss.nasa.gov/gistemp)96.
DJF surface temperature trends (1960–2013)
DJF surface temperature trends (1990–2013)
Latitude (N)Latitude (N)
1.2
2.4
1.6
0.8
–0.8
–1.6
–2.4
0
0.8
0.4
0
–0.4
°C per decade °C per decade
–0.8
–1.2
0°
3
2
1
0
–1
–2
0
–0.5
0.5
0
1.0
2.0
1.5
0.2
1960
Average temperature anomaly (°C)
1970 1980 1990 20002010
0.4
0.6
0–60° N
60–90° N
Trend (°C per decade)
Trend (°C per decade)
10°
20°
30°
40°
50°
60°
70°
80°
0°
10°
20°
30°
40°
50°
60°
70°
80°
a
b
c
DJF area-averaged temperature anomalies
Year
REVIEW ARTICLE
NATURE GEOSCIENCE DOI: 10.1038/NGEO2234
© 2014 Macmillan Publishers Limited. All rights reserved

4 NATURE GEOSCIENCE | ADVANCE ONLINE PUBLICATION | www.nature.com/naturegeoscience
Storm tracks
Large-scale and low-frequency variability in the extratropical
atmosphere is dominated by shis in storm tracks, oen expressed
by changes in large-scale atmospheric modes49. e dominant
atmospheric or climate mode that explains the greatest percent-
age of the mid- to high-latitude atmospheric variability, including
changes in the storm tracks, is the North Atlantic Oscillation/Arctic
Oscillation (NAO/AO). Changes in the storm tracks associated with
the NAO/AO have a strong inuence on the surface temperature
and precipitation variability in the North Atlantic sector50. When
the NAO/AO is in its positive phase, the storm tracks shi poleward
and winters are predominately mild across northern Eurasia and the
eastern United States but cold in the Arctic. When the NAO/AO is
in its negative phase, the storm tracks shi equatorward and win-
ters are predominantly more severe across northern Eurasia and the
eastern United States, but relatively mild in the Arctic. is tem-
perature pattern is sometimes referred to as the ‘warm Arctic–cold
continents’ pattern51. Recent observed wintertime temperature
trends across the Northern Hemisphere continents (Fig.2c) pro-
ject strongly on this temperature-anomaly pattern37, reecting a
negative trend in the NAO/AO over the past two decades37. Given
that climate models forced by regional and latitudinal variations in
atmospheric heating also exhibit changes in the NAO/AO50,52, it is
plausible that variability in sea ice and/or snow cover can inuence
the phase and amplitude of the NAO/AO, and consequently the
storm tracks.
e temperature pattern associated with variations in Eurasian
snow cover projects strongly onto the temperature pattern associ-
ated with the NAO/AO and recent temperature trends34,37,53. October
snow cover anomalies across Eurasia have been proposed as a skilful
Figure 3 | Temperature and precipitation extremes. Extreme indices in the mid-latitudes: trend maps for the 1951–2013 period and time series averaged
over the land area from 20° to 50°N. a, Trend in annual total wet-day precipitation. b, Annual very wet-day precipitation (that is, precipitation during days
exceeding the 95th percentile). c, Trend in annual very wet-day precipitation (that is, precipitation during days exceeding the 95th percentile). d, Coldest
daily minimum temperature. e, Trend in annual warm days (that is, percentage of days with temperatures exceeding the 90th percentile). f, Annual
number of icing days (days with maximum temperature <0°C). g, Percentage of land with summer months warmer than one standard deviation (solid)
and two standard deviations (dashed) above the 1951–1980 mean. h, Percentage of land with winter months colder than one standard deviation (solid) and
two standard deviations (dashed) below the 1951–1980 mean30. Stippling in the trend maps indicates significance at 95% confidence. The time series plot
yearly values (thin grey curves) and the long-term nonlinear trend (thick black curves). Panels a–f were created using the GHCNDEX global land gridded
dataset of climate extremes32 and definition of the extreme indices32.
−8 −6 −4 −2 02468
−6 −4 −2 0246
−0.4 −0.3 −0.2 −0.1 00.1 0.20.3 0.4
1950
0 0
10
10
20
20
30
30
30
–20
140
160
180
200
220
a b
c d
e f
g h
Total wet-day precipitation (mm yr–1)
Very wet-day precipitation (mm yr–1)
Warm days (% of days yr–1)
–18
–16
–14
40
35
40
50
60
1960 1970 1980
Year Year
Year
Year
Year
1990 2000 2010 1950
C
old winter months
(% of land)
W
arm summer months
(% of land)
Number of icing (d)
Coldest daily minimum
temperature (°C)
Very wet day
precipitation (mm)
1960 1970 1980 1990 2000 2010
1950 1960 1970 1980 1990 2000 2010
1950 1960 1970 1980 1990 2000 2010
1950 1960 1970 1980 1990 2000 2010
REVIEW ARTICLE
NATURE GEOSCIENCE DOI: 10.1038/NGEO2234
© 2014 Macmillan Publishers Limited. All rights reserved

NATURE GEOSCIENCE | ADVANCE ONLINE PUBLICATION | www.nature.com/naturegeoscience 5
predictor of the winter NAO/AO54,55, where extensive snow cover
is associated with the negative phase of the NAO/AO, though the
relationship may lack stationarity56. Satellite-based data indicate
a positive trend in Eurasian snow cover during October over the
past two to three decades6,37, though the veracity of these satellite-
based increases has recently been questioned57. A proposed physi-
cal mechanism to explain increased snow cover is that a warmer
Arctic atmosphere can hold more water vapour, which enhances
precipitation over the Eurasian continent. Additionally, the loss of
sea ice — and thus the increase in open water — has increased mois-
ture uxes to the atmosphere9. If near-surface atmospheric temper-
atures remain suciently cold — as is the case in Siberia during
autumn and winter — any additional precipitation will likely occur
as snow58,59. erefore, increasing October Eurasian snow cover
may have contributed to the recent tendency towards a negative
NAO/AO and cold Northern Hemisphere winters37. However, given
that the NAO/AO has considerable internal variability on multi-
ple timescales, the recent negative trend may be predominantly
internally driven.
e strong decline in sea ice during recent decades has intensi-
ed interest in the interactions between sea-ice conditions and the
atmosphere47,60. Most sea-ice–atmosphere coupled studies have dis-
cussed the atmospheric response in the context of NAO/AO vari-
ability. Observational analyses have shown signicant correlation
between reduced Arctic sea-ice cover and the negative phase of the
winter NAO/AO35,37,61–64, although it is unclear whether late sum-
mer and early autumn35 or late autumn and early winter38 sea-ice
anomalies are more skilful at predicting the winter weather patterns.
Modelling studies have also examined the NAO/AO response to
variations in Arctic sea ice35,65–74, by running simulations forced by
past sea-ice trends or case studies of years with large sea-ice anoma-
lies. ese studies have shown a full spectrum of NAO/AO responses
to reduced sea ice, from shis toward the positive phase68,71,73, the
negative phase35,65,74 or no signicant change73.
Furthermore, attributing NAO/AO changes and associated shis
in storm tracks to Arctic forcing has proved very dicult. e simu-
lated atmospheric circulation response to sea-ice loss is sensitive to
dierences in model physics, background atmospheric and oceanic
states, and the spatial patterns and magnitude of sea-ice anomalies.
Additionally, it has proven dicult to separate forced change due to
sea-ice loss from internal model variability. Large numbers of model
runs or ensembles are likely required to achieve statistically signi-
cant responses to forced sea-ice changes73. While these disparities
between studies preclude denitive conclusions, two general results
emerge. First, there are more studies that show a negative NAO/AO
response than a positive NAO/AO response. Second, the simulated
NAO/AO response to sea-ice loss is relatively small compared with
natural variability. is is consistent with the view that changes
in the NAO/AO are predominately internally driven and do not
necessarily require remote forcing75.
Jet stream
e second proposed dynamical pathway linking Arctic ampli-
cation to increased weather extremes is through its eects on the
behaviour of the polar jet stream. e dierence in temperature
between the Arctic and mid-latitudes is a fundamental driver of
the polar jet stream; therefore, a reduced poleward temperature
dierence could result in a weaker zonal jet with larger meanders.
A weaker and more meandering ow may cause weather systems
to travel eastward more slowly and thus, all other things being
equal, Arctic amplication could lead to more persistent weather
patterns76. Furthermore, Arctic amplication causes the thick-
ness of atmospheric layers to increase more to the north, such
that the peaks of atmospheric ridges may elongate northward and,
thus, increase the north–south amplitude of the ow76. Weather
extremes frequently occur when atmospheric circulation pat-
terns are persistent, which tends to occur with a strong meridional
wind component77,78.
Some aspects of this hypothesized linkage are supported by
observations and model simulations. A signicant decrease in
zonal-mean zonal wind at 500 hPa during autumn is observed
regionally76,79. is may be understood through the thermal wind
relationship, which states that vertical wind shear is proportional
to the meridional temperature gradient. Assuming that the winds
do not increase at the surface, the zonal wind at the jet-stream level
should slacken with a weaker meridional temperature gradient. In
other seasons when Arctic amplication is weaker, no signicant
trend in zonal-mean zonal wind is observed.
Northern Hemisphere
mid-latitude weather
Polar vortex
L
Northern Hemisphere cryosphere changes
• Summer and early fall Arctic sea-ice loss
• Fall Eurasian snow cover increase
• Late fall and winter Arctic sea-ice loss
Arctic
amplification
Changes in:
• Storm tracks
• Jet stream
• Planetary waves
Natural variability
• Internal climate modes
• Solar cycle
• Volcanic eruptions
Global climate
change
Figure 4 | Schematic of ways to influence Northern Hemisphere mid-latitude weather. Three major dynamical features for changing Northern
Hemisphere mid-latitude weather — changes in the storm tracks, the position and structure of the jet stream, and planetary wave activity — can be
altered in several ways. The pathway on the left and highlighted by double boxes is reviewed in this manuscript. Arctic amplification directly (by changing
the meridional temperature gradient) and/or indirectly (through feedbacks with changes in the cryosphere) alters tropospheric wave activity and the jet
stream in the mid- and high latitudes. Two other causes of changes in the storm tracks, jet stream and wave activity that do not involve Arctic amplification
are also presented: (1) natural modes of variability and (2) the direct influence of global climate change (that is, including influences outside the Arctic)
on the general circulation. The last two causes together present the current null hypothesis in the state of the science against which the influence of Arctic
amplification on mid-latitude weather is tested in both observational and modelling studies. Bidirectional arrows in the figure denote feedbacks (positive
or negative) between adjacent elements. Stratospheric polar vortex is represented by ‘L’ with anticlockwise flow.
REVIEW ARTICLE
NATURE GEOSCIENCE DOI: 10.1038/NGEO2234
© 2014 Macmillan Publishers Limited. All rights reserved

6 NATURE GEOSCIENCE | ADVANCE ONLINE PUBLICATION | www.nature.com/naturegeoscience
However, challenges remain in linking Arctic amplication
directly to changes in the speed and structure of the jet stream. For
example, other factors besides the near-surface meridional temper-
ature gradientinuence the zonal jet, including feedbacks from syn-
optic eddies or storms and the upper-level meridional temperature
gradient. Indeed, although Arctic amplication has weakened
the near-surface meridional temperature gradient, the tempera-
ture gradient between the tropics and mid-latitudes at higher alti-
tudes has strengthened80, which would increase jet stream-level
winds. Another challenge is identifying how much of the Arctic
e dierent components of a generalized mid-latitude jet are
illustrated in Fig. B1a. e proposed dynamical pathways linking
Arctic amplication to increased weather extremes are through
the highly nonlinear interaction between the jet stream, the
planetary waves and the storm tracks (Fig. 4). e wintertime
extratropical climate variability is aected by a complex set of
interactions and feedbacks between components, such as natural
variability modes, diabatic heating anomalies due to variations in
sea ice and snow cover, and atmospheric and oceanic heat trans-
port from tropical and subtropical latitudes. However, recently it
has been proposed that air–sea interaction in the Arctic could be
forcing teleconnection patterns and inuencing weather patterns
remotely in the mid-latitudes by heating the Arctic relative to the
rest of the globe36,76.
A change in the meridional temperature gradient, which pro-
jects onto the thermally driven component of the jet may or may
not result in a signicant change in the jet depending on how the
eddy-driven part of the jet varies. Complex interactions between
the mid-latitude wind jets, the planetary waves and baroclinic
weather systems is a nonlinear two-way feedback process, where
diabatic heating and cooling, orographic forcing and eddy wave
breakings drive the jets and teleconnection patterns. e yellow
arrow denotes the nal inuence, which is of synoptic variability
(jet eddies) on mid-latitude weather. e dynamical mechanisms
associated with each green arrow are as follows:
A. e temperature gradient, in this denition, inuences
the thermally driven jet (black solid circle) via the thermal-wind
balance (in combination with boundary conditions).
B. e temperature gradient inuences the eddy-driven jet
(black dashed circle) via changes in baroclinicity. e eddy-driven
jet inuences the temperature gradient via horizontal heat uxes.
C. e eddy-driven jet aects stratospheric winds (black
U shape) via vertical wave propagation. Stratospheric winds aect
the eddy-driven jet by altering the vertical wave-guide.
D. e thermally driven jet aects stratospheric winds via gen-
eration of orographically forced waves. Stratospheric winds aect
the thermally driven jet by altering the vertical wave guide.
E. e thermally driven jet aects the eddy-driven jet by act-
ing as a wave guide (the role of baroclinicity here directly associ-
ated with the temperature gradient). e eddy-driven jet aects the
thermally driven jet via energy uxes.
As can be seen from the gure, there are many feedbacks and
interactions involving mid-latitude jets, with the temperature gra-
dient being just one of them. erefore a weakening in the tem-
perature gradient may or may not result in a slowing down of the
jet depending on the net eect of other factors.
e North Atlantic Oscillation/Arctic Oscillation (NAO/AO) may
be considered a paradigm for the debate within the climate commu-
nity. Shown in Fig. B1b are the changes in the atmospheric circulation
associated with the negative phase of the NAO/AO. Positive (nega-
tive) zonal wind anomalies associated with the negative NAO/AO
are superimposed on the jet shown by a green solid (dashed) line.
Also shown are the temperature changes with warmer temperatures
in the Arctic (red) and colder temperatures in the mid-latitudes
(blue), increased high-latitude blocking (represented by clockwise
ow around a high) and a southward shi in the storm tracks (repre-
sented by a anticlockwise ow around a low), and increased meridi-
onal ow. All these dynamical changes are observed as the NAO/AO
shis from its positive phase to the negative phase. However, external
forcing, such as a reduced thermal gradient due to Arctic amplica-
tion, will project onto these dynamical patterns associated with the
negative NAO/AO: an equatorward shi in the zonal jet, increased
meridional ow, high-latitude blocking and a southward shi in
storm tracks. e yellow broken arrow denotes uncertainty whether
a change in the meridional temperature gradient can force all the
other changes depicted in the gure. Attributing observed changes
in mid-latitude weather to either Arctic amplication or internal
variability has proven challenging to date.
Box 1 | Jet-related dynamics.
Figure B1 | Schematic view of jet-related and negative North Atlantic Oscillation/Arctic Oscillation dynamics. a, Here, the tropospheric jet is divided
into two parts, a thermally driven part and an eddy-driven part. b, Changes in the atmospheric circulation associated with the negative phase of the
North Atlantic Oscillation/Arctic Oscillation. See Box text for detailed explanation.
Tropopause
Stratosphere
Equator North
pole
Polar night jet
Thermally
driven jet Eddy-driven jet
Temperature gradient
Surface
AB
C
D
ETropopause
Stratosphere
Equator North
pole
Mean jet
Anomalous temperatur
e
Surface
Mid-latitude
weather
Mid-latitude
weather
Arctic Oscillation
+/
Anomalous pressure
Anomalous zonal wind
a b
REVIEW ARTICLE NATURE GEOSCIENCE DOI: 10.1038/NGEO2234
© 2014 Macmillan Publishers Limited. All rights reserved

NATURE GEOSCIENCE | ADVANCE ONLINE PUBLICATION | www.nature.com/naturegeoscience 7
amplication is driven by local changes compared with remote
changes16. is distinction is highly relevant to the current debate
on possible Arctic–mid-latitude linkages, because if a signicant
portion of Arctic amplication is driven remotely, then Arctic
amplication may be partly viewed as a response to rather than a
forcing of mid-latitude weather. is highlights the importance of
As a summary of the studies presented, in Fig. B2 we synthesize
some common ideas about the atmospheric response to sea-ice
and snow cover variability that have until now been treated inde-
pendently. All sea-ice studies agree that sea-ice loss heats and mois-
tens the boundary layer of the Arctic atmosphere. It has also been
shown that a surface heat source in the extratropics induces down-
ward descent of air over the heat source, warming the atmospheric
column and raising heights in the mid-troposphere, while a trough
develops downstream inducing an equatorward ow of cold air97.
is is consistent with the result that reduced sea ice favours an
increase in mid- tropospheric heights in the Barents and Kara seas
region in winter51,88,92 with downstream troughing over Eurasia.
Studies also agree that increased snow cover cools the boundary
layer54. erefore a snow-induced surface cooling can lower heights
in the mid-troposphere, inducing enhanced ridging upstream.
In September and October, sea-ice loss has been most pro-
nounced in the Chukchi and East Siberian seas. Warming of the
atmosphere due to increased heating from newly ice-free ocean
causes geopotential heights to increase in the mid-troposphere,
which suppresses the jet stream southward over east Siberia. is
pattern, referred to as the Arctic Dipole, has strengthened during
the era of sea-ice loss61. A southward shi in the storm tracks over
East Asia allows for a more rapid advance of Eurasian snow cover
in October. Enlarged areas of open water north of Siberia also pro-
vide increased moisture ux to the atmosphere, which precipitates
as snow as the air mass is advected southward over Siberia58,71 (le
globe in Fig. B2).
In October, a more extensive snow cover cools the surface lead-
ing to lower heights and a trough in the mid-troposphere. Increased
troughing over East Asia favours upstream ridging near the Barents
and Kara seas and the Urals. Concurrently, the large sea-ice decits
and the associated strong surface heating anomalies migrate from
the Chukchi and East Siberian seas in September and October to
the Barents and Kara seas in November and December. is favours
mid-tropospheric ridging in the Barents and Kara seas region with
downstream troughing over East Asia. erefore, the extensive
snow cover over Siberia in October and November and the sea-ice
loss over the Barents and Kara seas in November and December
produce same-signed mid-tropospheric geopotential height pat-
terns over Eurasia. is planetary wave conguration is favour-
able for increased vertical propagation of Rossby waves from the
troposphere into the stratosphere98–100 (middle globe in Fig. B2).
Increased vertical propagation of Rossby wave energy from the
troposphere to the stratosphere weakens the polar vortex, resulting
in a stratospheric warming event. Circulation anomalies associated
with the warming event appear rst in the stratosphere and subse-
quently appear in the troposphere in January and February. ese
circulation anomalies resemble those associated with the negative
phase of the NAO/AO; that is, ridging over the Arctic especially near
Greenland, and a weaker, equatorward-shied polar jet stream. As
a result, warmer conditions prevail in the Arctic regions, but colder
and more severe winter weather occurs across the mid-latitude con-
tinents with a greater likelihood of snowstorms in the population
centres of the Northern Hemisphere mid-latitudes (right globe in
Fig. B2).
We propose a chain of events where less sea ice and increased
open water in the Arctic (that heats the atmosphere) and more
snow cover (that cools the atmosphere) both force the same pat-
tern, which results in a weakened polar vortex. Because the heat-
ing anomalies are displaced longitudinally, extensive Eurasian snow
cover and reduced Arctic sea ice can constructively interfere to
weaken the polar vortex and hence inuence surface weather.
Box 2 | Synthesis of cryospheric forcings.
Figure B2 | Synthesis of proposed cryospheric forcings. The schematic highlights a proposed way in which Arctic sea-ice loss in late summer through
early winter may work in concert with extensive Eurasian snow cover in the autumn to force the negative phase of the NAO/AO in winter. Snow is shown
in white, sea ice in white tinged with blue, sea-ice melt with blue waves, high and low geopotential heights with red ’H’ (red represents anomalous
warmth) and blue ’L’ (blue represents anomalous cold) respectively, tropospheric jet stream in light blue with arrows, and stratospheric jet or polar
vortex shown in purple with arrows. On the right globe, cold (warm) surface temperature anomalies associated with the negative phase of the winter
NAO/AO are shown in blue (brown). See Box text for detailed explanation.
HL
S
t
r
a
t
o
s
p
h
e
r
e
T
r
o
p
o
s
p
h
e
r
e
Vertical wave
activity flux
C
o
n
s
t
r
u
c
t
i
v
e
i
n
t
e
r
f
e
r
e
n
c
e
H
L
T
r
o
p
o
s
p
h
e
r
e
Moisture
flux
Sept. Oct. Nov. Dec. Jan. Feb.
S
t
r
a
t
o
s
p
h
e
r
e
T
r
o
p
o
s
p
h
e
r
e
H
H
LL
L
REVIEW ARTICLE
NATURE GEOSCIENCE DOI: 10.1038/NGEO2234
© 2014 Macmillan Publishers Limited. All rights reserved

8 NATURE GEOSCIENCE | ADVANCE ONLINE PUBLICATION | www.nature.com/naturegeoscience
considering the many ways in which mid-latitude jets are inu-
enced, including the meridional temperature gradient, which are
shown schematically in Fig. B1 in Box 1.
Observational support for the follow-on impacts of the hypoth-
esis related to a weakening zonal component of the jet76 is even
less strong — namely, whether Arctic amplication leads to larger
amplitude waves, slower wave propagation speeds and more per-
sistent weather patterns. Statistically robust evidence of increasing
north–south wave amplitude and slower propagation speed has not
been established79,81. is is not surprising given the recent emer-
gence of Arctic amplication and the large natural variability of the
atmosphere. Recent studies provide tentative evidence for increas-
ing amplitude in summer and autumn for some denitions of wave
amplitude, but not for others81. A signicant reduction in 500hPa
wave speeds during autumn was reported79, but the response was
not apparent in higher-level winds. e frequency of blocking-
high patterns is metric, region and time dependent, but as a whole
the observations do not support a signicant increase in blocking
occurrence over recent decades82.
e theory that Arctic amplication is resulting in a slower zonal
jet, increased meridional ow, amplied waves and more persistent
extreme weather has received a lot of attention from the media,
policymakers and climate sceintists83. In part due to the high prole,
this hypothesis has been scrutinized in the scientic literature more
extensively than other hypotheses linking Arctic climate change to
mid-latitude weather. However, it is worth noting that other stud-
ies on related topics, especially other observational studies, share
some of the same shortcomings35,37,38,61–64 (lack of statistical signi-
cance, causality unclear, incomplete mechanistic understanding,
and so on).
Planetary waves
Modication to large-scale Rossby waves over Eurasia is the third
proposed dynamical pathway linking Arctic amplication to mid-
latitude weather. Both observational analyses and modelling experi-
ments link more extensive snow cover across Eurasia, especially in
October, to changes in wave structure at high latitudes. Extensive
snow cover may lead to larger planetary waves that increase the ver-
tical propagation of wave energy into the stratosphere, favouring a
warmer and weakened stratospheric polar vortex84–87. It is proposed
that the atmospheric response lags the snow cover changes by a few
months because of the response time of the stratospheric circulation
and subsequent feedback to the troposphere.
Observed reductions in autumn–winter Arctic sea ice, especially
in the Barents and Kara seas, are also correlated with strengthened
anticyclonic circulation anomalies over the Arctic Ocean, which
tend to induce easterly ow and cold air advection over northern
Europe38,88–90, a link that may be sensitive to the timing of the sea-
ice anomalies. Winter anomalies trigger an immediate, local and
direct atmospheric response forced by increased turbulent heat
uxes locally over the Barents and Kara seas, which in turn changes
the baroclinicity and aects large-scale planetary or Rossby waves
in the atmosphere. Alternatively autumn sea-ice anomalies may
force a delayed, remote and indirect atmospheric response through
increased Eurasian snow cover46 or through altered baroclinicity
and high pressure over the Barents and Kara seas that force upward
propagating planetary waves into the stratosphere. Sucient wave
breaking in the polar stratosphere weakens the stratospheric polar
vortex and can trigger a stratospheric warming event. e circula-
tion anomalies associated with a stratospheric warming event prop-
agate back down to the surface in subsequent weeks, contributing to
a persistent negative NAO/AO and cold continental conditions90,91.
Several modelling studies have used prescribed Barents and
Kara sea-ice reductions to examine how the atmosphere responds.
Horizontal downstream propagation of the energy away from
anomalous, sea-ice-induced high pressure over the Barents and
Kara seas leads to the formation of a trough over Eurasia and sub-
sequent cold continental temperatures92. Such model experiments
have thus far only included the impact of sea-ice changes and not
the full extent of Arctic amplication.
e proposed response of planetary waves to reductions in both
snow cover and sea ice has inherent shortcomings. Free-running
(that is, without prescribed forcing) climate models do not simulate
well observations of the amplitude or the timing of wave changes to
more extensive snow cover86, resulting in a simulated weak relation-
ship found between October Eurasian snow cover and the winter
NAO/AO93. Regarding the response to sea-ice loss, caution is urged,
because strong trends in the sea-ice extent have made analyses
of the co-variability between sea ice and the atmosphere dicult
to interpret46. Furthermore the proposed atmospheric response
to sea-ice forcing is not robust and has yet to achieve statistical
signicance46, in part due to the shortness of the data record.
To conclude, variability in both sea ice and snow cover have been
hypothesized to independently force anomalously high geopotential
heights in the Barents and Kara seas. In Fig. B2 in Box 2, we pro-
vide a complementary perspective by proposing a synthesis of how
extensive snow cover and reduced sea ice in the autumn and early
winter can force local changes that constructively interfere to force
the same response in the planetary waves, which could inuence
winter weather patterns.
Synthesis of Arctic and mid-latitude linkages
Dramatic changes are occurring in the Arctic climate system, and
at the same time, the frequency of mid-latitude extreme weather
events appears to have increased. e potential link between Arctic
amplication and changes in extreme weather is a critical one, espe-
cially as Arctic amplication is robustly predicted to continue over
the coming decades. e climate dynamics literature concerning
Arctic–mid-latitude linkages is currently inconclusive, which may
help explain the media portrayal of a polarized view among scien-
tists81. Furthermore, the severe winter of 2013–2014 across eastern
North America focused the debate of whether extreme cold events
are attributable to climate change, including Arctic amplication, or
natural variability43,44. Cold winters such as the one experienced in
2013–2014 have occurred before and are expected as part of normal
weather variability even on a warmer planet94. Preliminary evidence
for a link between Arctic amplication and continental weather has
been presented, along with a range of dynamical hypotheses for
such a link. However, evidence demonstrating no robust statistical
or dynamical link between Arctic amplication and mid-latitude
climate variability has also been presented.
Nevertheless, dramatic changes to high-latitude sea ice and snow
cover have occurred, along with profound impacts at least locally in
the Arctic. e most robust atmospheric response to these changes
is an altered near-surface climate in the Arctic region. ere is
consensus that sea-ice loss enhances local warming, which weak-
ens near-surface meridional temperature gradients, moistens the
boundary layer and decreases the near-surface static stability. A
growing body of observational, modelling and theoretical evidence
suggests that the impact of high-latitude surface heating increases
upper-level geopotential heights, which aects the large-scale
atmospheric circulation beyond the Arctic.
To the rst order, amplied warming in the Arctic and a decrease
in the meridional temperature gradient should favour a weaker
zonal jet. However, whether weaker upper-level zonal winds causes
amplied and slower-moving planetary waves remains unclear.
Further evidence from modelling studies suggests that cryospheric
anomalies can alter the stratospheric polar vortex, storm tracks and
jet stream — all of which are key drivers of mid-latitude weather
and extremes. ese changes appear to be more likely in winter than
other seasons owing to the large Arctic amplication signal and
divergence of winter temperature trends from the other seasons.
REVIEW ARTICLE NATURE GEOSCIENCE DOI: 10.1038/NGEO2234
© 2014 Macmillan Publishers Limited. All rights reserved

NATURE GEOSCIENCE | ADVANCE ONLINE PUBLICATION | www.nature.com/naturegeoscience 9
e link between reduced Arctic sea ice and cold continental win-
ters is currently themost studied and arguably the best- supported
link between Arctic amplication and mid-latitude extreme
weather patterns.
Based on the research conducted to date, we oer a brief perspec-
tive on the challenges and research opportunities in the near future
(a more detailed list is included in the Supplementary Information).
Understanding the relative importance of dierent forcings mecha-
nisms, and how they interact with internally generated variability,
remains a key challenge. More and better observations (for example,
of ocean–ice–atmosphere energy exchange, cloud cover and tropo-
sphere–stratosphere coupling) would not only improve our under-
standing of the Arctic and its climate, but also help to elucidate
the mechanisms of atmospheric response to Arctic amplication
and better constrain the models. Better standardization of metrics
(extremes, blocking, wave amplitude, and so on) and coordination
of modelling experiments would allow results to be more directly
compared and the current disparities to be better understood.
Finally, testing hypotheses in a hierarchy of models of increasing
complexity, from simple dynamical models to state-of-the-art Earth
system models, would help to further our understanding and better
equip us to untangle the complexity of Arctic–mid-latitude linkages.
Methods
For Fig.1, we used the monthly mean elds from the ERA-Interim reanalysis95 to
compute seasonal means for the period March 1979 to February 2014. ese data
were averaged around circles of latitude (at 1.5° resolution). Standard seasonal
means were computed and used. We estimated trends using least-squares linear
regression. e statistical signicances of the regressions were calculated from a
two-tailed t-test.
Surface temperature anomalies for Fig.2 were taken from the NASA Goddard
Institute for Space Studies temperature record96. e decadal linear trends in sur-
face air temperature anomalies in Fig.2a are based on a least-squares regression
of the December–February (DJF) mean of monthly mean temperature anomalies
from 1960–1961 to 2013–2014. e corresponding time series of DJF temperatures
anomalies (Fig.2b) was constructed by weighting the anomalies by the cosine of
latitude.e same convention is used for Fig.2c except that the linear trends were
calculated based on DJF values during the period 1990–1991 to 2013–2014.
Figure3a–f was created using the GHCNDEX global land gridded dataset of
climate extremes32 available at www.climdex.org. e online data-visualization tool
was used to create linear trend maps and time series (over the period 1951–2014)
for dierent extreme indices provided in the GHCNDEX global land gridded
dataset. Time series are area-weighted averages of land regions within the latitu-
dinal belt from 20° to 50° N. Figure3g,h shows the percentage of land in the mid-
latitudes with unusually warm summer months or unusually cold winter months30.
For this, we used monthly gridded data from the NASA Goddard Institute for
Space Studies surface temperature dataset with a base period of 1951–1980. First,
we determined the local standard deviation due to natural variability at each grid
point in the latitudinal belt from 20° to 50°N for each calendar month of the boreal
winter (December–January–February) and boreal summer (June–July–August)
seasons. To do so, we applied a singular spectrum analysis to extract the long-term
(periods of 30years or greater) nonlinear trend over the twentieth century. Next,
we detrended the original time series by subtracting the long-term trend, which
gives the year-to-year variability. From this detrended signal, monthly standard
deviations were calculated using the 1951–2010 period, which were then seasonally
averaged. For boreal summer, we determined the percentage of land with tempera-
tures warmer than oneand two standard deviations beyond the mean (Fig.3g). For
boreal winter, we determined the percentage of land with temperatures colder than
oneand two standard deviations below the mean (Fig.3h).
Received 3 March 2014; accepted 22 July 2014;
published online 17 August 2014
References
1. Stroeve, J.C. etal. Sea ice response to an extreme negative phase of the Arctic
Oscillation during winter 2009/2010. Geophys. Res. Lett. 38, L02502 (2011).
2. Kwok, R. & Rothrock, D.A. Decline in Arctic sea ice thickness
from submarine and ICESat records: 1958–2008. Geophys. Res. Lett.
36, L15501 (2009).
3. Overland, J.E., Wang, M., Walsh, J.E. & Stroeve, J.C. Future Arctic
climate changes: adaptation and mitigation timescales. Earth’s Future
2, 68–74 (2014).
4. Derksen, C. & Brown, R. Spring snow cover extent reductions in the
2008–2012 period exceeding climate model projections. Geophys. Res. Lett.
39, L19504 (2012).
5. Matsumura, S., Zhang, X. & Yamazaki, K. Summer Arctic atmospheric
circulation response to spring Eurasian snow cover and its possible
linkage to accelerated sea ice decrease. J.Clim. http://dx.doi.org/10.1175/
JCLI-D-13-00549.1 (2014).
6. Mudryk, L.R., Kushner, P.J. & Derksen, C. Interpreting observed Northern
Hemisphere snow trends with large ensembles of climate simulations.
Clim. Dynam. 43, 345–359 (2013).
7. IPCC Summary for Policymakers in Climate Change 2013: e Physical
Science Basis (eds Stocker, T.F. et al.) 3–29 (Cambridge Univ, Press, 2013).
8. Serreze, M.C., Barrett, A.P., Stroeve, J.C., Kindig, D.M. & Holland,M.M.
e emergence of surface-based Arctic amplication. Cryosphere
3, 11–19 (2009).
9. Screen, J.A. & Simmonds, I. e central role of diminishing sea ice in recent
Arctic temperature amplication. Nature 464, 1334–1337 (2010).
10. Cowtan, K. & Way, R.G. Coverage bias in the HadCRUT4 temperature
series and its impact on recent temperature trends. Q.J.R.Meteorol. Soc.
133, 459–77 (2013).
11. Holland, M.M. & Bitz, C.M. Polar amplication of climate change in
coupled models. Clim. Dynam. 21, 221–232 (2003).
12. Stroeve, J.C. etal. e Arctic’s rapidly shrinking sea ice cover: a research
synthesis. Climatic Change 110, 1005–1027 (2012).
13. Gillett, N.P. etal. Attribution of polar warming to human inuence.
Nature Geosci. 1, 750–754 (2008).
14. Winton, M. Amplied Arctic climate change: What does surface albedo
feedback have to do with it? Geophys. Res. Lett. 33, L03701 (2006).
15. Serreze, M.C. & Barry, R.G. Processes and impacts of Arctic amplication:
a research synthesis. Glob. Planet. Change 77, 85–96 (2011).
16. Screen, J.A., Deser, C. & Simmonds, I. Local and remote controls on
observed Arctic warming. Geophys. Res Lett. 39, L10709 (2012).
17. Shindell, D. & Faluvegi, G. Climate response to regional radiative forcing
during the twentieth century. Nature Geosci. 2, 294–300 (2009).
18. Francis, J.A. & Hunter, E. New insight into the disappearing Arctic sea ice.
EOS Trans. Am. Geophys. Union 87, 509–511 (2006).
19. Graverson, R.G. & Wang, M. Polar amplication in a coupled climate model
with locked albedo. Clim. Dynam. 33, 629–643 (2009).
20. Pithan, F. & Mauritsen, T. Arctic amplication dominated by temperature
feedbacks in contemporary climate models. Nature Geosci. 7, 181–184 (2014).
21. Graversen, R.G., Mauritsen, T., Tjernstrom, M., Kallen, E. & Svensson, G.
Vertical structure of recent Arctic warming. Nature 451, 53–56 (2008).
22. Wood, K.R. etal. Is there a “new normal” climate in the Beaufort Sea?
Polar Res. 32, 19552 (2013).
23. Steele, M., Ermold, W. & Zhang, J. Arctic Ocean surface warming trends over
the past 100years. Geophys. Res. Lett. 35, L19715 (2008).
24. Inoue, J. & Hori, M.E. Arctic cyclogenesis at the marginal ice zone:
A contributory mechanism for the temperature amplication?
Geophys. Res. Lett. 38, L12502 (2011).
25. Screen, J.A. & Simmonds, I. Increasing fall-winter energy loss from the
Arctic Ocean and its role in Arctic amplication. Geophys. Res. Lett.
37, L16707 (2010).
26. Min, S.K., Zhang, X., Zwiers, F.W. & Hegerl, G.C. Human contribution to
more-intense precipitation extremes. Nature 470, 378–381 (2011).
27. Coumou, D. & Rahmstorf, S. A decade of weather extremes.
Nature Clim. Change 2, 491–496 (2012).
28. Westra, S., Alexander, L.V. & Zwiers, F.W. Global increasing trends in annual
maximum daily precipitation. J.Clim. 26, 3904–3918 (2013).
29. Coumou, D. & Robinson, A. Historic and future increase in the global land
area aected by monthly heat extremes. Environ. Res. Lett. 8, 034018 (2013).
30. Coumou, D., Robinson, A. & Rahmstorf, S. Global increase in record-
breaking monthly-mean temperatures. Climatic Change 118, 771–782 (2013).
31. Seneviratne, S.I., Donat, M.G., Mueller, B. & Alexander, L.V. No
pause in the increase of hot temperature extremes. Nature Clim. Change
4, 161–163 (2014).
32. Donat, M.G. etal. Global land-based datasets for monitoring climatic
extremes. Bull. Am. Meteorol. Soc. 94, 997–1006 (2013).
33. Screen, J.A. Arctic amplication decreases temperature variance in northern
mid- to high-latitudes. Nature Clim. Change 4, 577–582 (2014).
34. Cohen, J., Barlow, M. & Saito, K. Decadal uctuations in planetary
wave forcing modulate global warming in late boreal winter. J.Clim.
22, 4418–4426 (2009).
35. Liu, J., Curry, J.A., Wang, H., Song, M. & Horton, R. Impact of
declining Arctic sea ice on winter snow. Proc. Natl Acad. Sci. USA
109, 4074–4079 (2012).
36. Greene, C.H. & Monger, B.C. An Arctic wild card in the weather.
Oceanography 25, 7–9 (2012).
REVIEW ARTICLE
NATURE GEOSCIENCE DOI: 10.1038/NGEO2234
© 2014 Macmillan Publishers Limited. All rights reserved

10 NATURE GEOSCIENCE | ADVANCE ONLINE PUBLICATION | www.nature.com/naturegeoscience
37. Cohen, J., Furtado, J., Barlow, J.M., Alexeev, V. & Cherry, J. Arctic
warming, increasing fall snow cover and widespread boreal winter cooling.
Environ. Res. Lett. 7, 014007 (2012).
38. Tang, Q., Zhang, X., Yang, X. & Francis, J.A. Cold winter extremes in northern
continents linked to Arctic sea ice loss. Environ. Res. Lett. 8, 014036 (2013).
39. Tollefson, J. US cold snap fuels climate debate. Nature http://dx.doi.
org/10.1038/nature.2014.14485 (2014).
40. Hamilton, L.C. & Lemcke-Stampone, M. Arctic warming and your weather:
public belief in the connection. Int. J.Climatol. 34, 1723–1728 (2013).
41. Alexeev, V.A., Langen, P.L. & Bates, J.R. Polar amplication of surface
warming on an aquaplanet in “ghost forcing” experiments without sea ice
feedbacks. Clim. Dynam. 24, 655–666 (2005).
42. Cohen, J., Furtado, J., Barlow, M., Alexeev, V. &. Cherr y, J. Asymmetric
seasonal temperature trends. Geophys. Res. Lett. 39, L04705 (2012).
43. Wallace, J.M., Held, I.M., ompson, D.W. J., Trenberth, K.E. & Walsh, J.E.
Global warming and winter weather. Science 343, 729–730 (2014).
44. Palmer, T. Record-breaking winters and global climate change. Science
344, 803–804 (2014).
45. Fischer, E.M. & Knutti, R. Heated debate on cold weather.
Nature Clim. Change 4, 577–582 (2014).
46. Cohen, J., Jones, J., Furtado, J.C. &. Tziperman, E. Warm Arctic, cold
continents: a common pattern related to Arctic sea ice melt, snow advance,
and extreme winter weather. Oceanography 26, 150–160 (2013).
47. Vihma, T. Eects of Arctic sea ice decline on weather and climate: a review.
Surv. Geophys. http://dx.doi.org/10.1007/s10712-014-9284-0 (2014).
48. Hoskins, B. e potential for skill across the range of the seamless weather-
climate prediction problem: a stimulus for our science. Q.J.R.Meteorol. Soc.
139, 573–584 (2013).
49. Woollings, T. & Blackburn, M. e North Atlantic jet stream under
climate change and its relation to the NAO and EA patterns. J.Clim.
25, 886–902 (2012).
50. Bader, J. etal. A review on Northern Hemisphere sea-ice, storminess and the
North Atlantic Oscillation: observations and projected changes. Atmos. Res.
101, 809–834 (2011).
51. Overland, J.E., Wood, K.R. & Wang, M. Warm Arctic–cold continents:
impacts of the newly open Arctic Sea. Polar Res. 30, 15787 (2011).
52. Wu, A., Hsieh, W.W., Boer, G.J. & Zwiers, F.W. Changes in the Arctic
Oscillation under increased atmospheric greenhouse gases. Geophys. Res. Lett.
34, L12701 (2007).
53. Mote, T. & Kutney, E. Regions of autumn Eurasian snow cover
andassociations with North American winter temperatures. Int. J.Climatol.
32, 1164–1177 (2012).
54. Cohen, J. & Entekhabi, D. Eurasian snow cover variability and Northern
Hemisphere climate predictability. Geophys. Res. Lett. 26, 345–348 (1999).
55. Cohen, J. & Jones, J. A new index for more accurate winter predictions.
Geophys. Res. Lett. 38, L21701 (2011).
56. Peings, Y., Brun, E., Mauvais, V. & Douville, H. How stationary is the
relationship between Siberian snow and Arctic Oscillation over the 20th
century? Geophys. Res. Lett. 40, 183–188 (2013).
57. Brown, R.D. & Derksen, C. Is Eurasian October snow cover extent increasing?
Environ. Res. Lett. 8, 024006 (2013).
58. Ghatak, D., Frei, A., Gong, G., Stroeve, J. & Robinson, D. On the emergence of
an Arctic amplication signal in terrestrial Arctic snow extent. J.Geophys. Res.
115, D24105 (2010).
59. Ghatak, D. etal. Simulated Siberian snow cover response to observed Arctic
sea ice loss, 1979–2008.J.Geophys. Res. 117, D23108 (2012).
60. Budikova, D. Role of Arctic sea ice in global atmospheric circulation.
Glob. Planet. Change 68, 149–163 (2009).
61. Francis, J.A., Chan, W., Leathers, D.J., Miller, J.R. & Veron, D.E. Winter
Northern Hemisphere weather patterns remember summer Arctic sea-ice
extent. Geophys. Res. Lett. 36, L07503 (2009).
62. Overland, J.E. & Wang, M. Large-scale atmospheric circulation changes are
associated with the recent loss of Arctic sea ice. Tel lu s A 62, 1–9 (2010).
63. Strong, C., Magnusdottir, G. & Stern, H. Observed feedback between winter
sea ice and the North Atlantic Oscillation. J.Clim. 22, 6021–6032 (2009).
64. Hopsch, S., Cohen, J. & Dethlo, K. Analysis of a link between fall Arctic sea
ice concentration and atmospheric patterns in the following winter. Tellus A
64, 18624(2012).
65. Magnusdottir, G., Deser, C. & Saravanan, R. e eects of North Atlantic
SST and sea-ice anomalies on the winter circulation in CCM3. Part I:
Main features and storm track characteristics of the response. J.Clim.
17, 857–876 (2004).
66. Deser, C., Magnusdottir, G., Saravanan, R. & Phillips, A. e eects of North
Atlantic SST and sea-ice anomalies on the winter circulation in CCM3. Part II:
Direct and indirect components of the response. J.Clim. 17, 877–889 (2004).
67. Alexander, M.A. etal. e atmospheric response to realistic Arctic sea ice
anomalies in an AGCM during winter. J.Clim. 17, 890–905 (2004).
68. Strey, S.T., Chapman, W.L. & Walsh, J.E. e 2007sea ice minimum: impacts
on the Northern Hemisphere atmosphere in late autumn and early winter.
J.Geophys. Res. 115, D23103 (2010).
69. Porter, D.F., Cassano, J.J. & Serreze, M.C. Local and large-scale atmospheric
responses to reduced Arctic sea ice and ocean warming in the WRF model.
J.Geophys. Res. 117, D11115 (2012).
70. Bluthgen, J., Gerdes, R. & Werner, M. Atmospheric response to the extreme
Arctic sea ice conditions in 2007. Geophys. Res. Lett. 39, L02707 (2012).
71. Orsolini, Y., Senan, R., Benestad, R. & Melsom, A. Autumn atmospheric
response to the 2007low Arctic sea ice extent in coupled ocean–atmosphere
hindcasts. Clim. Dynam. 38, 2437–2448 (2012).
72. Singarayer, J.S., Valdes, P.J. & Bamber, J.L. e atmospheric impact
of uncertainties in recent Arctic sea-ice reconstructions. J.Clim.
18, 3996–4012 (2005).
73. Screen, J.A., Deser, C., Simmonds, I. & Tomas, R. Atmospheric impacts of
Arctic sea-ice loss, 1979–2009: separating forced change from atmospheric
internal variability. Clim. Dynam. 43, 333–344 (2013).
74. Peings, Y. & Magnusdottir, G. Response of the wintertime Northern
Hemispheric atmospheric circulation to current and projected Arctic sea ice
decline: a numerical study with CAM5. J.Clim. 27, 244–264 (2014).
75. Tanaka, H.L. & Seki, S. Development of a three-dimensional spectral linear
baroclinic model and its application to the baroclinic instability associated
with positive and negative Arctic Oscillation indices. J.Meteorol. Soc. Jpn
91, 193–213 (2013).
76. Francis, J.A. & Vavrus, S.J. Evidence linking Arctic amplication to extreme
weather in mid-latitudes. Geophys. Res. Lett. 39, L06801 (2012).
77. Petoukhov, V., Rahmstorf, S., Petri, S. & Schellnhuber, H.J. Quasiresonant
amplication of planetary waves and recent Northern Hemisphere weather
extremes. Proc. Natl Acad. Sci. USA 110, 5336–5341 (2013).
78. Screen, J.A. & Simmonds, I. Amplied mid-latitude planetary waves favour
particular regional weather extremes. Nature Clim. Change http://dx.doi.
org/10.1038/nclimate2271 (2014).
79. Barnes, E.A. Revisiting the evidence linking Arctic amplication to extreme
weather in midlatitudes. Geophys. Res. Lett. 40, 1–6 (2013).
80. Allen, R.J. & Sherwood, S.C. Warming maximum in the tropical upper
troposphere deduced from thermal winds. Nature Geosci. 1, 399–403 (2008).
81. Screen, J.A. & Simmonds, I. Exploring links between Arctic amplication and
mid-latitude weather. Geophys. Res. Lett. 40, 959–964 (2013).
82. Barnes, E.A., Dunn-Sigouin, E., Masato, G. & Woollings, T. Exploring
recent trends in Northern Hemisphere blocking. Geophys. Res. Lett.
41, 638–644 (2014).
83. Kintisch, E. Into the maelstrom. Science 344, 250–253 (2014).
84. Fletcher, C., Hardiman, S.C., Kushner, P.J. & Cohen, J. e dynamical
response to snow cover perturbations in a large ensemble of atmospheric
GCM integrations. J.Clim. 22, 1208–1222 (2009).
85. Allen, R.J. & Zender, C.S. Forcing of the Arctic Oscillation by Eurasian snow
cover. J.Clim. 24, 6528–6539 (2011).
86. Cohen, J., Furtado, J.C., Jones, J., Barlow, M., Whittleston, D. & Entekhabi, D.
Linking Siberian snow cover to precursors of stratospheric variability. J.Clim.
27, 5422–5432 (2014).
87. Peings Y., Saint-Martin, D. & Douville, H. A numerical sensitivity study
of the inuence of Siberian snow on the northern annular mode. J.Clim.
25,592–607 (2012).
88. Petoukhov, V. & Semenov, V.A. A link between reduced Barents-Kara sea
ice and cold winter extremes over northern continents. J.Geophys. Res.
115, D21111 (2010).
89. Inoue, J., Hori, M.E. & Takaya, K. e role of Barents Sea ice in the wintertime
cyclone track and emergence of a warm-Arctic cold-Siberian anomaly. J.Clim.
25, 2561–2568 (2012).
90. Jaiser, R., Dethlo, K., Handorf, D., Rinke, A. & Cohen, J. Impact of sea ice
cover changes on the Northern Hemisphere atmospheric winter circulation.
Tel lu s A 64, 11595 (2012).
91. Jaiser, R., Dethlo, K. & Handorf, D. Stratospheric response to Arctic sea
ice retreat and associated planetary wave propagation changes. Tel lu s A
65, 19375 (2013).
92. Honda, M., Inue, J. & Yamane, S. Inuence of low Arctic sea-ice minima on
anomalously cold Eurasian winters. Geophys. Res. Lett. 36, L08707 (2009).
93. Hardiman, S.C., Kushner, P.J. & Cohen, J. Investigating the ability of general
circulation models to capture the eects of Eurasian snow cover on winter
climate. J.Geophys. Res. 113, D21123 (2008).
94. Martin, S. & Dienbaugh, N.S. Transient twenty-rst century changes
in daily-scale temperature extremes in the United States. Clim. Dynam.
42, 1383–1404 (2014).
95. Dee, D.P. etal. e ERA-Interim reanalysis: conguration and performance of
the data assimilation system. Q.J.R.Meteorol. Soc. 137, 553–597 (2011).
96. Hansen, J.E. & Lebede, S. Global trends of measured surface air temperature.
J.Geophys. Res. 92, 13345–13372 (1987).
REVIEW ARTICLE NATURE GEOSCIENCE DOI: 10.1038/NGEO2234
© 2014 Macmillan Publishers Limited. All rights reserved

NATURE GEOSCIENCE | ADVANCE ONLINE PUBLICATION | www.nature.com/naturegeoscience 11
97. Hoskins, B.J. & Karoly, D.J. e steady linear response of a spherical
atmosphere to thermal and orographic forcing. J.Atmos. Sci.
38, 1179–1196 (1981).
98. Garnkel, C.I., Hartmann, D.L. & Sassi, F. Tropospheric precursors of
anomalous Northern Hemisphere stratospheric polar cortices. J.Clim.
23, 3282–3299 (2010).
99. Kolstad, E.W. & Charlton-Perez, A.J. Observed and simulated precursors
of stratospheric polar vortex anomalies in the Northern Hemisphere.
Clim. Dynam. 37, 1443–1456 (2010).
100. Cohen, J. & Jones, J. Tropospheric precursors and stratospheric warmings.
J.Clim. 24, 6562–6572 (2011).
Acknowledgements
We are grateful to E. Barnes for many helpful discussions and suggested revisions to the
manuscript. J.C. is supported by the National Science Foundation grants BCS-1060323
and AGS-1303647.J.S. is funded by Natural Environment Research Council grant
NE/J019585/1. M.B. received support from National Science Foundation grant
ARC-0909272 and NASA NNX13AN36G. J.O. receives support from the Arctic Research
Project of the National Oceanic and Atmospheric Administration Climate Program
Oce and the Oce of Naval Research, Code 322.
Author contributions
J.C. proposed and was the main author of the manuscript. All co-authors contributed to
the writing of the manuscript. J.S. created Fig.1, J.F. Figs2 & 4, D.C. Fig.3, J.F. and J.C.
Fig.4, M.B. and J.C. Fig. B1, and D.W. and J.C. Fig. B2.
Additional information
Supplementary information accompanies this paper on www.nature.com/ngeo.
Reprints and permissions information is available online at www.nature.com/reprints.
Correspondence and requests for materials should be addressed to J.C.
Competing financial interests
e authors declare no competing nancial interests.
REVIEW ARTICLE
NATURE GEOSCIENCE DOI: 10.1038/NGEO2234
© 2014 Macmillan Publishers Limited. All rights reserved