Abstract

This paper describes two dominant patterns of Asian winter climate variability: the Siberian high (SH) pattern and the Asia–Arctic (AA) pattern. The former depicts atmospheric variability closely associated with the intensity of the Siberian high, and the latter characterizes the teleconnection pattern of atmospheric variability between Asia and the Arctic, which is distinct from the Arctic Oscillation (AO). The AA pattern plays more important roles in regulating winter precipitation and the 850-hPa meridional wind component over East Asia than the SH pattern, which controls surface air temperature variability over East Asia.

In the Arctic Ocean and its marginal seas, sea ice loss in both autumn and winter could bring the positive phase of the SH pattern or cause the negative phase of the AA pattern. The latter corresponds to a weakened East Asian winter monsoon (EAWM) and enhanced winter precipitation in the midlatitudes of the Asian continent and East Asia. For the SH pattern, sea ice loss in the prior autumn emerges in the Siberian marginal seas, and winter loss mainly occurs in the Barents Sea, Labrador Sea, and Davis Strait. For the AA pattern, sea ice loss in the prior autumn is observed in the Barents–Kara Seas, the western Laptev Sea, and the Beaufort Sea, and winter loss only occurs in some areas of the Barents Sea, the Labrador Sea, and Davis Strait. Simulation experiments with observed sea ice forcing also support that Arctic sea ice loss may favor frequent occurrence of the negative phase of the AA pattern. The results also imply that the relationship between Arctic sea ice loss and winter atmospheric variability over East Asia is unstable, which is a challenge for predicting the EAWM based on Arctic sea ice loss.

1. Introduction

During boreal winter, the strongest continental anticyclone on Earth, known as the Siberian high (SH), covers the Asian continent. Intense cooling of the air’s surface layer and sinking motion induced by the mid- and upper-level convergence contribute to an enhancement of the SH (Ding and Krishnamurti 1987; Ding 1990). The SH strongly affects weather and climate over Asia and parts of Europe. Outbreaks of cold polar air westward from the SH pressure cell cause occasional severe cold spells over areas of Europe. An example is the winter of 2011/12, when more than 700 people died due to extreme cold conditions. The SH is an important part of the East Asian winter monsoon (EAWM) system. The EAWM is a highly significant feature of Asia’s winter circulation, closely associated with the development and southward propagation of cold surges over East Asia (Chang and Lau 1980; Ding 1990; Jhun and Lee 2004; Wu et al. 2006).

Recent studies have shown a strengthening trend in the SH over the past two decades (Jeong et al. 2011; Wu et al. 2011). The corresponding winter surface air temperature (SAT) exhibits a negative trend over the Asian continent (Cohen et al. 2009, 2012; Wu et al. 2011). Some regions of Eurasia have recently experienced exceptionally cold winters, such as in 2007/08, 2009/10, 2010/11, 2011/12, and 2012/13 (Fig. 1). It appears that cold winters have become more frequent over East Asia. It has been found that lower Arctic sea ice values from the previous autumn to winter may contribute to cold conditions over Eurasia and an enhanced SH, via large-scale dynamic and thermal processes (Dethloff et al. 2006; Francis et al. 2009; Honda et al. 2009; Petoukhov and Semenov 2010; Screen and Simmonds 2010; Overland and Wang 2010; Wu et al. 1999, 2011; Francis and Vavrus 2012; Liu et al. 2012; Jaiser et al. 2012; Hopsch et al. 2012; Tang et al. 2013; Vihma 2014; Peings and Magnusdottir 2014; Walsh 2014). Wu et al. (1999) showed that variability in winter sea ice in the Barents–Kara Seas is related to the intensity of the EAWM via the Eurasian teleconnection pattern. This study indicated that heavy sea ice in these seas excites the positive phase of the Eurasian pattern, with anomalously low 500-hPa heights over Siberia and positive height anomalies over East Asia. This pattern weakens the East Asian trough and the intensity of the EAWM, and decreases the frequency of cold air outbursts into China. Opposite effects are observed during light sea ice conditions in this area of the Arctic marginal seas (Petoukhov and Semenov 2010; Inoue et al. 2012). Based on simulation experiments with prescribed sea ice forcing in the Barents–Kara Seas, Petoukhov and Semenov (2010) suggested that sea ice loss may result in strong anticyclonic anomalies over the Arctic Ocean, leading to a continental-scale winter cooling, with more than a threefold increased probability of cold winter extremes over Eurasia. Inoue et al. (2012) showed that light sea ice conditions during winter in the Barents Sea could lead to an anticyclonic anomaly over the Siberian coast and cold advection over eastern Siberia.

Fig. 1.

SAT anomalies (°C) for six recent winters, with the linear trend for 1979–2013 period removed (NCEP–NCAR Reanalysis 1 data).

Fig. 1.

SAT anomalies (°C) for six recent winters, with the linear trend for 1979–2013 period removed (NCEP–NCAR Reanalysis 1 data).

Francis et al. (2009) and Honda et al. (2009) independently argued that summer or summer to autumn sea ice can impact the atmosphere in the ensuing wintertime. Wu et al. (2011) showed that persistent autumn to winter sea ice concentration (SIC) anomalies in the Barents–Kara Seas and the northern vicinity of these seas, with concurrent sea surface temperature (SST) anomalies, are responsible for the SH and SAT anomalies over the middle and high latitudes of Eurasia. However, recent observations and simulation experiments did not support significant impacts of Arctic sea ice loss on the midlatitudes (Screen et al. 2014; Peings and Magnusdottir 2014; Walsh 2014). On 16 September 2012, Arctic sea ice reached its minimum extent for the year, of 3.41 million square kilometers. This is the lowest seasonal minimum extent in the satellite record since 1979. However, in the ensuing winter (2012/13), the strength of the SH was nearly normal. Additionally, in the winter of 2006/07, a weakened SH (its standard deviation was below −1.0) corresponded to a negative SIC anomaly in the previous September (see Fig. 2 of Wu et al. 2011). These two cases imply that autumn sea ice loss does not always correspond to a strengthened SH (or enhanced EAWM). Indeed, in addition to Arctic sea ice, there are many factors that influence the SH, such as Eurasian snow cover (Cohen et al. 2012) and internal atmospheric variability. On the other hand, atmospheric circulation variability showed different regimes in the two abovementioned winters (see section 3 below). Thus, it is impossible to predict dominant patterns of winter atmospheric variability over the Asian continent in terms of a single external factor. The motivation of the present study is to explore dominant patterns of winter atmospheric variability over Asia and their possible linkages with Arctic sea ice loss. Our study demonstrates that Arctic sea ice loss also promotes the weakening of the EAWM.

2. Data and methods

The following datasets were used: 1) the Arctic SIC dataset (1° × 1°) from January 1979 to May 2013, obtained from the British Atmospheric Data Centre (BADC; http://badc.nerc.ac.uk/data/hadisst/); 2) the monthly mean sea level pressure (SLP), SAT, winds, and geopotential heights from January 1979 to March 2013, obtained from NCEP–NCAR Reanalysis 1; 3) monthly mean global land precipitation data from 1979 to 2013 (http://ftp.cpc.ncep.noaa.gov/precip/50yr/gauge/2.5deg/format_bin/; Chen et al. 2002); and 4) the monthly mean Arctic Oscillation (AO) index for the period from 1979 to May 2013 (http://www.cpc.ncep.noaa.gov/products/precip/CWlink/daily_ao_index/monthly.ao.index.b50.current.ascii).

Empirical orthogonal function (EOF) analysis was performed on winter [December–February (DJF)] mean SLP. The Monte Carlo method was applied to examine statistical field significance, as in Livezey and Chen (1983). For an anomalous field derived from linear regression, the percentage of grid points that are statistically significant at 0.05 (0.01) level is first identified over a domain. This process is then repeated 1000 times with different series of 34 (or 34 winters from 1979 to 2013) numbers randomly selected from a normal distribution. The anomalous field is deemed significant if the percentage of significant grid points exceeds that derived from 1000 experimental replications. Additionally, a Student’s t test was used to assess the statistical significance of atmospheric changes between different phases.

Additionally, the ECHAM5 (Roeckner et al. 2003) model (T63 spectral resolution and 19 pressure levels) was applied to explore the impacts of SIC on the model atmosphere. A simulation was performed with observed monthly SIC in the Northern Hemisphere from January 1978 to November 2012 (419 months) as the external forcing, while the SIC in the Southern Hemisphere and global SST were prescribed as their climatological monthly mean. The SST and SIC data were obtained through a spatially interpolation of observations taken from the BADC (http://badc.nerc.ac.uk/data/hadisst/); for detailed information, refer to the Atmospheric Model Intercomparison Project (AMIP) phase II SST and SIC boundary condition dataset (http://www-pcmdi.llnl.gov/projects/amip/AMIP2EXPDSN/BCS/bcsintro.php). This experiment was repeated with 40 different atmospheric initial conditions that were derived from a 50-yr control run. In regions where the SIC changes year to year, the SST was prescribed as its climatological value, unlike in Screen et al. (2014). In this study, all linear trends in the original data were first removed before performing EOF, regression, and composite differential analyses.

3. Two dominant patterns and their impacts

This study focuses on winter (DJF) SLP variability in the middle and high latitudes in order to reveal dominant patterns of winter atmospheric variability over the Asian continent. EOF analysis was applied to the normalized area-weighted winter mean SLP data (after detrending) over 30°–70°N, 80°–120°E and for 34 winters from 1979 to 2013. This domain contains the core region of the SH where the regionally averaged winter SLP over 40°–60°N, 80°–120°E is used as the SH index (SHI) to characterize the intensity of the SH (Wu and Wang 2002).

The first two EOFs (EOF1 and EOF2) respectively account for 50% and 26% of the variance. For the leading EOF, Figs. 2a–d show anomalies in winter mean SLP, 500-hPa height, SAT, and 850-hPa meridional wind components, derived from linear regressions on the normalized leading principal component (PC1). In the middle and high latitudes of Eurasia, winter SLP anomalies show a monopole structure, with the positive center located over the Ural Mountains, indicating a strengthened SH (Fig. 2a). Meanwhile, negative SLP anomalies are seen in the middle and low latitudes of the Asian continent. Winter mean 500-hPa height anomalies show a triple structure: the center of positive height anomalies is over the Kara Sea, and two negative centers are located over Europe and northeastern Asia. Thus, Fig. 2b indicates a westward shifted and strengthened 500-hPa East Asian trough. Height anomalies here closely resemble those in Jung et al. (2014; see their Fig. 3), who investigated the impact of the Arctic on winter 500-hPa height in the midlatitudes. Negative SAT anomalies over most of East Asia are dynamically consistent with the strengthened SH and deepened East Asian trough (Fig. 2c). Meanwhile, positive SAT anomalies cover the Arctic. Consequently, Figs. 2a–c characterize atmospheric circulation anomalies associated with the SH, supported by the correlation between the PC1 and the detrended SHI (r = 0.95; see Fig. 3). This systematic atmospheric circulation anomaly is herein termed the SH pattern. The 850-hPa meridional wind anomalies, however, do not exceed the level of statistical significance over eastern China south of 30°N (Fig. 2d). Strengthened northerlies are seen over the area from Lake Baikal extending southeastward to the northwestern Pacific. When accompanied by a strengthened SH, weak southerly anomalies emerge over parts of southern China. The SH pattern mainly reflects large-scale meridional circulation anomalies over the middle and high latitudes.

Fig. 2.

(a) Regression map of detrended winter mean SLP, regressed on the normalized PC1 of EOF analysis of detrended winter mean SLP variability over 30°–70°N, 80°–120°E (outlined in green); the yellow (light blue) and red (blue) areas indicate positive (negative) SLP anomalies at 0.05 and 0.01 significance levels, respectively. (b)–(d) As in (a), but for detrended winter 500-hPa height, SAT, and 850-hPa meridional wind components, respectively; contour intervals are 0.5 hPa in (a), 10 gpm in (b), 1°C in (c), and 0.3 m s−1 in (d).

Fig. 2.

(a) Regression map of detrended winter mean SLP, regressed on the normalized PC1 of EOF analysis of detrended winter mean SLP variability over 30°–70°N, 80°–120°E (outlined in green); the yellow (light blue) and red (blue) areas indicate positive (negative) SLP anomalies at 0.05 and 0.01 significance levels, respectively. (b)–(d) As in (a), but for detrended winter 500-hPa height, SAT, and 850-hPa meridional wind components, respectively; contour intervals are 0.5 hPa in (a), 10 gpm in (b), 1°C in (c), and 0.3 m s−1 in (d).

Fig. 3.

Normalized time series of the detrended winter SHI (red line) and the PC1s of EOF analyses of detrended winter mean SLP variability over four different domains: 30°–70°N, 80°–120°E (green); 30°–70°N, 50°–130°E (blue); 30°–80°N, 50°–130°E (black); and 20°–70°N, 60°–130°E (purple).

Fig. 3.

Normalized time series of the detrended winter SHI (red line) and the PC1s of EOF analyses of detrended winter mean SLP variability over four different domains: 30°–70°N, 80°–120°E (green); 30°–70°N, 50°–130°E (blue); 30°–80°N, 50°–130°E (black); and 20°–70°N, 60°–130°E (purple).

The same analysis process was carried out over three different domains: 1) 30°–70°N, 50°–130°E, 2) 30°–80°N, 50°–130°E, and 3) 20°–70°N, 60°–130°E. The leading EOFs over the three domains respectively account for 46%, 46%, and 44% of the variance. Corresponding winter SLP, 500-hPa height, SAT, and 850-hPa meridional wind anomalies, derived from linear regressions on their PC1s, closely resemble those shown in Figs. 2a–d (not shown). Their PC1s are significantly correlated with the detrended SHI, with correlations of 0.90, 0.75, and 0.81, respectively (Fig. 3).

For EOF2, the amplitudes of both positive SLP and 500-hPa height anomalies are weaker than those for EOF1 (Figs. 4a,b). Positive SLP anomalies appear over most of the Asian continent south of 50°N, with moderate negative anomalies in the north. Negative SLP and 500-hPa height anomalies mainly appear in the Arctic and northern North Pacific, making this pattern distinct from the positive phase of the AO. In fact, the correlation between EOF2 and the detrended AO is 0.28 (0.44) for winters of 1979/80–2008/09 (winters of 1979/80–2012/13). Positive SAT anomalies are observed in the middle and high latitudes of the Asian continent, with negative SAT anomalies to the south (Fig. 4c). Such a spatial distribution of SAT anomalies is dynamically consistent with that for the SLP anomalies. Significant anomalies in 850-hPa meridional winds are observed in East Asia, particularly in eastern and northeastern China (Fig. 4d).

Fig. 4.

As in Fig. 2, but regressed on the normalized PC2 of EOF analysis of detrended winter mean SLP variability over 30°–70°N, 80°–120°E [outlined by green lines in (a)]. Contour intervals are 0.5 hPa in (a), 5 gpm in (b), 0.5°C in (c), and 0.3 m s−1 in (d).

Fig. 4.

As in Fig. 2, but regressed on the normalized PC2 of EOF analysis of detrended winter mean SLP variability over 30°–70°N, 80°–120°E [outlined by green lines in (a)]. Contour intervals are 0.5 hPa in (a), 5 gpm in (b), 0.5°C in (c), and 0.3 m s−1 in (d).

The second EOFs over the three domains (30°–70°N, 50°–130°E; 30°–80°N, 50°–130°E; and 20°–70°N, 60°–130°E) respectively account for 22%, 23%, and 22% of the variance. Anomalies in detrended winter SLP, 500-hPa height, SAT, and 850-hPa meridional winds, derived from linear regressions on their PC2s, closely resemble those shown in Figs. 4a–d for the domain 30°–80°N, 50°–130°E, but with anomalies of opposing sign for the other domains (not shown). For two domains (30°–70°N, 50°–130°E and 20°–70°N, 60°–130°E), the PC2 time series are out of phase with that for 30°–70°N, 80°–120°E (Fig. 5); their correlations are −0.95 and −0.90, respectively. Over the domains 30°–70°N, 80°–120°E and 30°–80°N, 50°–130°E, the PC2 time series are in phase (Fig. 5; their correlation is 0.84). Although the domains are different for EOF analysis, the atmospheric circulation anomalies associated with EOF2s exhibit similar features over Asia and the Arctic. Thus, this systematic circulation anomaly is herein termed the Asia–Arctic (AA) pattern. The AA pattern predominantly exhibits large-scale zonal circulation anomalies, differing from the SH pattern. Because the intensity of the SH is closely related to the EAWM variability (Jhun and Lee 2004; Wu et al. 2006), the extracted leading SLP pattern should characterize SH variability as accurately as possible. If we selected a domain that covers more area of the Arctic Ocean, the correlation between the leading SLP pattern and the detrended SHI would decline. Consequently, the first two PCs of EOF analysis over 30°–70°N, 80°–120°E can be regarded as indices that depict the SH and AA patterns, respectively. It should be pointed out that the SH and AA pattern well represent the first two coupled patterns between winter mean SLP over Eurasia (30°–70°N, 50°–130°E) and any meridional vertical cross section of zonal winds over the Asian continent (extracted by the maximum covariance analysis; not shown). Thus, neither of the SH and AA patterns relies on the EOF method. They reflect different dynamic regimes, namely large-scale meridional and zonal circulation anomalies.

Fig. 5.

Normalized PC2s of EOF analyses of detrended winter mean SLP variability over four different domains: 30°–70°N, 80°–120°E (green); 30°–70°N, 50°–130°E (blue; multiplied by −1.0); 30°–80°N, 50°–130°E (black); and 20°–70°N, 60°–130°E (purple, multiplied by −1.0).

Fig. 5.

Normalized PC2s of EOF analyses of detrended winter mean SLP variability over four different domains: 30°–70°N, 80°–120°E (green); 30°–70°N, 50°–130°E (blue; multiplied by −1.0); 30°–80°N, 50°–130°E (black); and 20°–70°N, 60°–130°E (purple, multiplied by −1.0).

The regionally averaged winter 850-hPa meridional wind over eastern China (30°–40°N, 110°–120°E) is significantly correlated with the AA pattern (r = −0.56; after removing linear trends, the correlation is −0.60, at 0.01 significance level). In contrast to the AA pattern, the SH pattern does not show a significant relationship with the regionally averaged 850-hPa meridional wind (r = −0.20 after detrending). At the surface, however, the regionally averaged meridional wind (at 10 m) over 29.52°–40.95°N, 110.625°–120°E is significantly correlated with the SH and AA patterns after detrending at −0.51 and −0.57, respectively. This implies that compared with the SH pattern, the AA pattern shows a closer relationship with the EAWM.

To investigate dynamical connections to Arctic atmosphere variability, we first selected positive and negative phase winters for which their standard deviations are >0.8 or <−0.8, as shown in Table 1. We selected the latitude–pressure vertical cross section at 110°E to show wave activity fluxes (Fig. 6). In the middle and high troposphere south of 75°N, the wave activity fluxes show coherent propagations southward to 45°N, reflecting a dynamical linkage between the Arctic and the midlatitudes of East Asia (Fig. 6a). In the Arctic north of 75°N, the wave activity fluxes propagate northward over nearly the entire troposphere. For the AA pattern, southward propagation is mainly observed between 40° and 65°N (Fig. 6b), indicating that sub-Arctic atmospheric variability is directly linked with that over the midlatitudes of East Asia via atmospheric energy propagations. Over the Arctic, the wave activity fluxes display northward propagations. At 500 hPa, the wave activity fluxes originated from the Barents–Kara Seas propagate southeastward to the high latitudes of the Asian continent, and then propagate northeastward to the Arctic Ocean (Fig. 6c). Another branch propagates to the northwestern Pacific.

Table 1.

Winter cases with standard deviations >0.8 (positive phase) or <−0.8 (negative phase). Boldface indicates winters with both the SH and AA patterns.

Winter cases with standard deviations >0.8 (positive phase) or <−0.8 (negative phase). Boldface indicates winters with both the SH and AA patterns.
Winter cases with standard deviations >0.8 (positive phase) or <−0.8 (negative phase). Boldface indicates winters with both the SH and AA patterns.
Fig. 6.

(a) Differences in detrended mean geopotential heights between the positive and negative phases of the SH pattern along the latitude–pressure cross section at 110°E, superimposed on meridional and vertical (multiplied by 0.05) wave activity flux (vectors; m2 s−2) of Takaya and Nakamura (2001); light blue and blue areas indicate geopotential height differences at 0.05 and 0.01 significance levels, respectively. (b) As in (a), but for differences in detrended mean geopotential heights between the negative and positive phases of the AA pattern. (c) Differences in detrended mean geopotential heights between the negative and positive phases of the AA pattern at 500 hPa; contour intervals are 20 gpm; composite winter cases for the SH and AA patterns are shown in Table 1 (nonboldface winters).

Fig. 6.

(a) Differences in detrended mean geopotential heights between the positive and negative phases of the SH pattern along the latitude–pressure cross section at 110°E, superimposed on meridional and vertical (multiplied by 0.05) wave activity flux (vectors; m2 s−2) of Takaya and Nakamura (2001); light blue and blue areas indicate geopotential height differences at 0.05 and 0.01 significance levels, respectively. (b) As in (a), but for differences in detrended mean geopotential heights between the negative and positive phases of the AA pattern. (c) Differences in detrended mean geopotential heights between the negative and positive phases of the AA pattern at 500 hPa; contour intervals are 20 gpm; composite winter cases for the SH and AA patterns are shown in Table 1 (nonboldface winters).

Figure 7 shows the latitude–pressure vertical cross section of westerly anomalies associated with the SH and AA patterns along 110°E. Relative to the negative phase of the SH pattern, its positive phase corresponds to a strengthened westerly jet in the higher troposphere (Fig. 7a) (the center of the westerly jet at 110°E is around 30°N and 200 hPa; not shown). Meanwhile, over the middle and high latitudes, westerly winds are weakened significantly. Consequently, weakened westerlies over the middle and high latitudes favor cold air accumulation and outbreaks southward from the Arctic and high latitudes, which enhance strength of both of the SH and the East Asian trough (Figs. 2a,b), leading to negative SAT anomalies over East Asia (Fig. 2c), and vice versa for its negative phase. For the AA pattern, amplitudes of westerly anomalies are apparently weaker relative to the SH pattern (Fig. 7a) and coherently shift northward (Fig. 7b). Relative to the positive phase of the AA pattern, its negative phase corresponds to strengthened westerlies between 35° and 55°N, with weakened westerlies on both sides. The spatial distribution of westerly anomalies from 40° to 80°N is dynamically consistent with negative height anomalies in the middle and high latitudes in Fig. 6b. Weakened tropospheric westerlies over the high latitudes of the Asian continent favor Arctic cold air into the Asian continent and accumulation, resulting in positive SLP anomalies over the northern Asian continent (as in Figs. 4a,c, but with opposing sign). Meanwhile, strengthened tropospheric westerlies between 35° and 55°N obstruct cold air accumulation and outbreaks southward from the midlatitudes, leading to positive SAT and negative SLP anomalies over the middle and low latitudes of the Asian continent. Opposite anomalous fields are observed for the positive phase of the AA pattern.

Fig. 7.

(a) Differences in detrended mean westerly flows (m s−1) between the positive and negative phases of the SH pattern along the latitude–pressure cross section at 110°E. (b) As in (a), but for differences in detrended mean westerly flows between the negative and positive phases of the AA pattern; the composite cases for the SH and AA patterns and meanings for the color areas are as in Fig. 6 

Fig. 7.

(a) Differences in detrended mean westerly flows (m s−1) between the positive and negative phases of the SH pattern along the latitude–pressure cross section at 110°E. (b) As in (a), but for differences in detrended mean westerly flows between the negative and positive phases of the AA pattern; the composite cases for the SH and AA patterns and meanings for the color areas are as in Fig. 6 

Impacts of the two atmospheric patterns on winter precipitation differ, as shown in Fig. 8. The positive phase of the SH pattern causes decreases in winter precipitation over most of the Asian continent, particularly in the low latitudes east of 80°E, whereas increased precipitation is mainly observed in some areas of the midlatitudes and the Russian Far East (Fig. 8a). The AA pattern has more substantial impacts on winter precipitation than the SH pattern (Fig. 8b). Compared with the positive phase of the AA pattern, its negative phase significantly enhances precipitation over East Asia and central Asia. Enhanced precipitation is dynamically consistent with 500-hPa height anomalies in Fig. 6c. Negative 500-hPa height anomalies over the middle and high latitudes of Asia and positive height anomalies over the northwestern and northern Pacific favor increased precipitation between them. Meanwhile, decreased precipitation emerges in the high latitudes and between 70° and 110°E south of 40°N.

Fig. 8.

(a) Anomalous winter precipitation percentages, derived from differences in detrended winter mean precipitation between the positive and negative phases of the SH pattern divided by the winter mean precipitation averaged over the past 34 winters from 1979 to 2013; thin (dashed thin) and thick (dashed thick) purple contours indicate positive (negative) precipitation differences exceeding 0.05 and 0.01 significance levels, respectively. (b) As in (a), but differences in detrended winter mean precipitation between the negative and positive phases of the AA pattern; composite cases are listed in Table 1 (nonboldface winters).

Fig. 8.

(a) Anomalous winter precipitation percentages, derived from differences in detrended winter mean precipitation between the positive and negative phases of the SH pattern divided by the winter mean precipitation averaged over the past 34 winters from 1979 to 2013; thin (dashed thin) and thick (dashed thick) purple contours indicate positive (negative) precipitation differences exceeding 0.05 and 0.01 significance levels, respectively. (b) As in (a), but differences in detrended winter mean precipitation between the negative and positive phases of the AA pattern; composite cases are listed in Table 1 (nonboldface winters).

4. Possible associations with sea ice loss in autumn and winter

Both the SH and AA patterns are associated with Arctic SIC anomalies in the prior autumn [September–November (SON)] to winter (DJF) (Fig. 9). For the SH pattern, decreased autumn SIC is observed in the Siberian marginal seas (Fig. 9a), particularly from the northern Barents Sea across to the Kara Sea, extending eastward to the Laptev Sea and the Pacific part of the Arctic Ocean. In winter, negative SIC anomalies are mainly observed in the Greenland–Barents–Kara Seas, the Labrador Sea, and Davis Strait (Fig. 9b), dynamically consistent with the spatial distribution of winter surface wind anomalies (not shown). Wu et al. (2011) suggested that the regionally averaged (76.5°–83.5°N, 60.5°–149.5°E) September SIC is significantly correlated with the ensuing winter SIC averaged over the Barents–Kara Seas (67.5°–80.5°N, 20.5°–80.5°E) during the period from 1979 to 2010 (r = 0.66; correlation is 0.52 after detrending). Thus, the regionally averaged September SIC is a potential precursor for the ensuing winter SH that cannot be predicted using tropical SSTs alone [their correlation was −0.6 after detrending; see Fig. 2 of Wu et al. (2011)]. An enhanced SH is associated with persistent SIC negative anomalies from autumn to winter, and previous observations and simulation experiments support this association (Honda et al. 2009; Petoukhov and Semenov 2010; Wu et al. 1999; Liu et al. 2012; Inoue et al. 2012; Rinke et al. 2013; Jung et al. 2014).

Fig. 9.

Differences in detrended mean SIC between the positive and negative phases of the SH pattern in (a) the previous autumn (SON) and (b) winter (DJF). (c),(d) As in (a),(b), but for differences between the negative and positive phases of the AA pattern; green contours denote SIC differences at 0.05 significance level. The composite cases for the SH and AA patterns are shown in Table 1 (nonboldface winters).

Fig. 9.

Differences in detrended mean SIC between the positive and negative phases of the SH pattern in (a) the previous autumn (SON) and (b) winter (DJF). (c),(d) As in (a),(b), but for differences between the negative and positive phases of the AA pattern; green contours denote SIC differences at 0.05 significance level. The composite cases for the SH and AA patterns are shown in Table 1 (nonboldface winters).

For the negative phase of the AA pattern, negative SIC anomalies in the previous autumn are observed in some areas: the Barents–Kara Seas, the western Laptev Sea, the East Siberian Sea, and the Beaufort Sea (Fig. 9c). In winter, increased SIC in the Greenland Sea and southeastern Barents Sea is concurrent with decreased SIC in the Labrador Sea, Davis Strait, and some areas of the Barents Sea (Fig. 9d), unlike in Fig. 9b. Amplitudes and extents of SIC anomalies associated with the AA pattern are smaller relative to the SH pattern. Linear regression analyses further verify the associations between the two atmospheric patterns and previous autumn SIC anomalies (not shown).

In the data and methods section, simulation experiments forced by SIC forcing were introduced. Here we examine simulated winter atmospheric responses (Figs. 10 and 11) and corresponding autumn and winter SIC anomalies (Fig. 12). Data used here are original model output and SIC data rather than detrended data. Figure 10 shows differences in simulated winter mean atmospheric circulation between the positive and negative phases of the SH pattern during the period from 1978 to 2012. It is seen that positive SLP anomalies are mainly observed over the Arctic, much of North America, and the Tibetan Plateau, and significant negative SLP anomalies emerge over East Asia and the Russian Far East (Fig. 10a). At 500 hPa, positive geopotential height anomalies appear over the Arctic and are surrounded by negative anomalies (Fig. 10b). A shifted northward and deepened East Asian trough emerge over the East Asian coast, with positive height anomalies over the midlatitudes of the Asian continent and the northwestern Pacific sector. Significant positive SAT anomalies are mainly confined to the high latitudes, and positive SAT anomalies occupy most of Eurasia except for some areas of northern Eurasia and the Tibetan Plateau where negative SAT anomalies are visible (Fig. 10c). Thus, simulated winter atmospheric circulation and SAT anomalies indicate a weakened EAWM; to a great extent, they capture major characteristics of the negative phase of the AA pattern as shown in Fig. 4, but with opposing sign.

Fig. 10.

(a) Simulated differences in winter mean SLP between positive and negative phases of the SH pattern (see Table 1, nonboldface winters), derived from 40 experiments, thus positive and negative phases contain 120 and 280 winters, respectively; thin (dashed thin) and thick (dashed thick) purple contours denote positive (negative) SLP anomalies at 0.05 and 0.01 significance levels, respectively. (b),(c) As in (a), but for winter 500-hPa height and SAT differences, respectively; contour intervals are 0.2 hPa in (a), 2 gpm in (b), and 0.2°C in (c).

Fig. 10.

(a) Simulated differences in winter mean SLP between positive and negative phases of the SH pattern (see Table 1, nonboldface winters), derived from 40 experiments, thus positive and negative phases contain 120 and 280 winters, respectively; thin (dashed thin) and thick (dashed thick) purple contours denote positive (negative) SLP anomalies at 0.05 and 0.01 significance levels, respectively. (b),(c) As in (a), but for winter 500-hPa height and SAT differences, respectively; contour intervals are 0.2 hPa in (a), 2 gpm in (b), and 0.2°C in (c).

Fig. 11.

As in Fig. 10, but for the AA pattern (see Table 1, nonboldface winters and excluding the winter of 2012/13), derived from 40 experiments, thus negative and positive phases contain 160 and 240 winters, respectively.

Fig. 11.

As in Fig. 10, but for the AA pattern (see Table 1, nonboldface winters and excluding the winter of 2012/13), derived from 40 experiments, thus negative and positive phases contain 160 and 240 winters, respectively.

Fig. 12.

Differences in mean SIC in the forced model simulation between the positive and negative phases of the SH pattern in (a) the previous autumn (SON) and (b) winter (DJF). (c),(d) As in (a),(b), but for differences between the negative and positive phases of the AA pattern. The composite cases for the SH and AA patterns are same as those in Figs. 10 and 11, respectively.

Fig. 12.

Differences in mean SIC in the forced model simulation between the positive and negative phases of the SH pattern in (a) the previous autumn (SON) and (b) winter (DJF). (c),(d) As in (a),(b), but for differences between the negative and positive phases of the AA pattern. The composite cases for the SH and AA patterns are same as those in Figs. 10 and 11, respectively.

Figure 11 shows simulated differences between the negative and positive phases of the AA pattern. Significant positive SLP anomalies emerge over the Arctic and Siberia (Fig. 11a), and positive 500-hPa height anomalies appear over northern Eurasia and the Arctic, with concurrent negative height anomalies over Europe, East Asia, and the northern North Pacific, implying a strengthened East Asian trough (Fig. 11b). At the surface, positive SAT anomalies are observed over the Arctic and negative SAT anomalies occupy much of Eurasia, particularly over the central Asian continent and East Asia, where significant negative SAT anomalies are observed (Fig. 11c). Thus simulated atmospheric circulation anomalies, to a great extent, reflect the positive phase of the SH pattern rather than the AA pattern.

A stronger atmospheric response is seen in Fig. 11 relative to that in Fig. 10, indicated by amplitudes and extents of both of positive SLP and 500-hPa height anomalies. Differences in mean SIC forcing may be responsible for different atmospheric responses. It is seen that although SIC data used in Figs. 9a,b and 12a,b are derived from the BADC (http://badc.nerc.ac.uk/data/hadisst/) their differences are visible, particularly in the prior autumn, due mainly to detrended data used in Fig. 9. In Fig. 12a, negative SIC anomalies are mainly observed in from the Barents Sea eastward to the Laptev Sea, and positive SIC anomalies emerge from the East Siberian Sea eastward to the Beaufort Sea. Nearly opposite SIC anomalies are seen in part of the Laptev Sea eastward to the Beaufort Sea and the Arctic Ocean (Fig. 12c). The area with SIC anomalies ≤−5.0% is approximately 2.20 × 106 km2 in Fig. 12a and less than 3.04 × 106 km2 in Fig. 12c. Thus, compared to Fig. 12a, SIC anomalies in Fig. 12c correspond to more heating released to the atmosphere from the ocean, which enhances the feedback of sea ice loss on the winter atmosphere. Thus, simulated strengthening of SH (Fig. 11) is reasonable. Above analyses demonstrate that Arctic sea ice loss could either bring the positive phase of the SH pattern or produce the negative phase of the AA pattern, which corresponds to a weakened EAWM. Thus, it is a challenge to predict EAWM based on Arctic sea ice loss. Additionally, simulation results also show that Arctic sea ice loss also favors occurrences of cold winters in North America.

On the other hand, very weak differences in simulated winter SLP and 500-hPa height indicate low coherency of model results, and simulated time series of the SHI averaged over the 40 experiments are strictly confined to a very narrow range of 1029–1031 hPa (Fig. 13a; the observed SHI range was 1026.8–1033.5 hPa from 1979 to 2013). The simulated standard deviation of the SHI ranges from 1.1 to 2.1 hPa (Fig. 13b). Such low coherency reflects the combined effects of large internal variability and model uncertainties. Screen et al. (2014) also suggested that SLP and height responses are hard to detect and may be partially or totally masked by atmospheric internal variability. Additionally, the atmospheric response to sea ice loss may depend on the state of the atmosphere (Balmaseda et al. 2010).

Fig. 13.

(a) Time series of the simulated winter SHI (hPa) averaged over 40 experiments (solid blue line); dashed blue and red lines denote a mean of the simulated winter SHI averaged over the past 34 winters from 1978 to 2012 and 5-yr running means of the simulated winter SHI, respectively. (b) The standard deviation time series of the simulated winter SHI derived from 40 experiments.

Fig. 13.

(a) Time series of the simulated winter SHI (hPa) averaged over 40 experiments (solid blue line); dashed blue and red lines denote a mean of the simulated winter SHI averaged over the past 34 winters from 1978 to 2012 and 5-yr running means of the simulated winter SHI, respectively. (b) The standard deviation time series of the simulated winter SHI derived from 40 experiments.

It is seen that simulated SLP and 500-hPa height anomalies exhibit a quasi-barotropic structure over the Arctic (Figs. 10 and 11). This differs from the response of winter atmosphere to SIC loss in Screen et al. (2014) and Peings and Magnusdottir (2014) where a baroclinic response was evident over the Arctic Ocean. Screen et al. (2014) discussed possible reasons for the existence of negative winter SLP anomalies over the Arctic in response to sea ice loss, which differs from previous results (Alexander et al. 2004; Francis et al. 2009; Liu et al. 2012). They suggested that ensemble member numbers and prescription of SIC forcing in simulation experiments may be the reason for negative SLP anomalies in response to SIC loss.

Prior researchers discussed possible mechanisms for the impact of Arctic sea ice on the atmosphere: a decreased autumn Arctic sea ice would lead to an intensified heat loss from the ocean and a stronger heating effect to the overlying atmosphere, which would strengthen atmospheric baroclinicity and instability (Alexander et al. 2004; Jaiser et al. 2012; Porter et al. 2012). As the seasons progress, through a negative feedback process, baroclinic atmospheric processes diminish and barotropic interactions intensify, resulting in winter positive SLP and 500-hPa height anomalies over the high latitudes and the Arctic (Alexander et al. 2004; Deser et al. 2004, 2007; Magnusdottir et al. 2004; Jaiser et al. 2012; Wu et al. 2011; Walsh 2014), favoring a strengthened SH (the positive phase of the SH pattern). Additionally, Arctic sea ice loss in both autumn and winter would decrease the thermal gradient between the Arctic and the middle and high latitudes of Eurasia, leading to weakened westerlies in winter (Francis and Vavrus 2012; see Fig. 4 of Wu et al. 2011) and favoring cold air outbreaks southward from the Arctic. This is the possible mechanism responsible for the association between the SH pattern and Arctic sea ice loss. The similar mechanism may be at work for the linkage between the AA pattern and Arctic sea ice loss. Additionally, the simulated wave activity flux induced by sea ice loss also supports the dynamical connection between the Arctic and Asia (not shown). Recent studies, however, suggested that although sea ice loss affects atmospheric variability at northern midlatitudes, results show big differences in the magnitude, timing, and spatial extent of these effects (e.g., Vihma 2014). In addition to the state of the atmosphere (Balmaseda et al. 2010), differences in the magnitude and extent of SIC anomalies in Fig. 12 may be one of the reasons for different remote responses, supported by Petoukhov and Semenov (2010). We have not explored possible mechanisms for the impact of SIC loss on the SH and AA patterns herein because they are beyond the scope of the present study.

In the late 1990s, the Arctic surface wind fields experienced an interdecadal shift in both spring (April–June) and summer (July–September). An anomalous cyclone prevailed before 1997 and was then replaced by an anomalous anticyclone over the Arctic Ocean, which was consistent with the rapid decline in trend of September sea ice extent (Wu et al. 2012). Autumn Arctic SIC also experienced an interdecadal shift in the late 1990s, which is one of the possible reasons for interdecadal variability of winter SAT in East Asia (Yang and Wu 2013). Although the two AA patterns, derived respectively from the detrended and original winter mean SLP data over the same domain, are highly correlated (0.94; 0.99 after detrending), their low-frequency evolutions are different (Fig. 14). The AA pattern derived from the original data showed an interdecadal shift more clearly in the late 1990s. It is seen that positive phases of the low-frequency oscillations (>11 yr) were dominant before the winter of 1998/99 and were then replaced by frequent negative phases. This interdecadal shift was also reflected in a 7-yr running mean time series of the AA pattern. Since 2007 autumn SIC has maintained negative anomalies in the Arctic Ocean and Siberian marginal seas (not shown), favoring the occurrence of the negative phase of the AA pattern.

Fig. 14.

(a) The low-frequency oscillation (>11 yr) time series of the AA pattern (derived from detrended data; blue line), derived from a harmonic analysis; the red line is a 7-yr running mean time series of the AA pattern. (b) As in (a), but the AA pattern was derived from the original winter mean SLP data. Units are arbitrary.

Fig. 14.

(a) The low-frequency oscillation (>11 yr) time series of the AA pattern (derived from detrended data; blue line), derived from a harmonic analysis; the red line is a 7-yr running mean time series of the AA pattern. (b) As in (a), but the AA pattern was derived from the original winter mean SLP data. Units are arbitrary.

5. Conclusions and discussion

Using EOF analysis of winter mean SLP variability over the Asian continent, we have described two dominant patterns of winter atmospheric variability: the SH and AA patterns, which account for 76% of the variance over 30°–70°N, 80°–120°E. The SH pattern depicts well the dominant features of winter atmospheric circulation variability closely associated with the intensity of SH. The positive phase of the SH pattern corresponds to a systematic strengthening of both the SH and East Asian trough, leading to negative SAT anomalies over East Asia, and vice versa for its negative phase. Decreased autumn Arctic SIC along the Siberian marginal seas, particularly in the northern Barents–Kara Seas and the Pacific area of the Arctic Ocean, provides favorable external forcing for generating the positive phase of this pattern.

The AA pattern features a teleconnection pattern of atmospheric variability between Asia and the Arctic. The positive phase of the AA pattern describes positive SLP anomalies south of 55°N and negative SLP anomalies in the high latitudes and the Arctic, corresponding to a strengthened EAWM, and SAT anomalies with opposing sign emerge over the Asian continent. The AA pattern is more important than the SH pattern in regulating winter precipitation and the 850-hPa meridional wind component over East Asia. The negative phase of the AA pattern may be associated with decreased autumn Arctic SIC in the Barents–Kara Seas, the western Laptev Sea, and the Beaufort Sea. Simulation experiments with observed SIC forcing also support that Arctic sea ice loss could bring the positive phase of the SH pattern or produce the negative phase of the AA pattern, which corresponds to a weakened EAWM. Recently, the two dominant patterns have alternately occurred to influence East Asia. It appears that the AA pattern has become more frequent recently. Simulation experiments also indicate that Arctic sea ice loss favors occurrences of cold winters in North America.

This study has described a statistical association between Arctic SIC loss and the AA pattern, although physical details of their association need to be further investigated. Additionally, factors such as SSTs in the North Atlantic and subarctic seas and Eurasian snow cover also play roles in regulating East Asian winter climate variability (Li 2004; Peng et al. 2003; Cohen et al. 2012; Walsh 2014), and their relative contributions also warrant further study. The results herein imply that autumn Arctic sea ice loss plays an important role in regulating winter climate variability over East Asia.

Acknowledgments

The authors are grateful to all anonymous reviewers for their insight and constructive suggestions, which helped to significantly improve this paper. The authors thank the British Atmospheric Data Centre (BADC), NCEP–NCAR, and the NOAA/Climate Prediction Center for providing sea ice concentration data, atmospheric reanalysis data, and the global land precipitation data and AO index. This study was supported by the National Key Basic Research Project of China (2013CBA01804 and 2015CB453200), the National Natural Science Foundation of China (41475080, 41221064), and State Oceanic Administration Project (201205007).

REFERENCES

REFERENCES
Alexander
,
A.
,
S.
Bhatt
, and
J.
Walsh
,
2004
:
The atmospheric response to realistic Arctic sea ice anomalies in an AGCM during winter
.
J. Climate
,
17
,
890
905
, doi:.
Balmaseda
,
M. A.
,
L.
Ferranti
,
F.
Molteni
, and
T. N.
Palmer
,
2010
:
Impact of 2007 and 2008 Arctic ice anomalies on the atmospheric circulation: Implications for long-range predictions
.
Quart. J. Roy. Meteor. Soc.
,
136
,
1655
1664
, doi:.
Chang
,
C.-P.
, and
K.-M.
Lau
,
1980
:
Northeasterly cold surges and near-equatorial disturbances over the winter MONEX area during December 1974. Part II: Planetary-scale aspects
.
Mon. Wea. Rev.
,
108
,
298
312
, doi:.
Chen
,
M.
,
P.
Xie
,
J. E.
Janowiak
, and
P. A.
Arkin
,
2002
:
Global land precipitation: A 50-yr monthly analysis based on gauge observations
.
J. Hydrometeor.
,
3
,
249
266
, doi:.
Cohen
,
J.
,
M.
Barlow
, and
K.
Saito
,
2009
:
Decadal fluctuations in planetary wave forcing modulate global warming in late boreal winter
.
J. Climate
,
22
,
4418
4426
, doi:.
Cohen
,
J.
,
J.
Furtado
,
M.
Barlow
,
V.
Alexeev
, and
J.
Cherry
,
2012
:
Arctic warming, increasing snow cover and widespread boreal winter cooling
.
Environ. Res. Lett.
, 7, 014007, doi:.
Deser
,
C.
,
G.
Magnusdottir
,
R.
Saravanan
, and
A.
Phillips
,
2004
:
The effects of North Atlantic SST and sea ice anomalies on the winter circulation in CCM3. Part II: Direct and indirect components of the response
.
J. Climate
,
17
,
877
889
, doi:.
Deser
,
C.
,
A.
Tomas
, and
S.
Peng
,
2007
:
The transient atmospheric circulation response to North Atlantic SST and sea ice anomalies
.
J. Climate
,
20
,
4751
4767
, doi:.
Dethloff
,
K.
, and Coauthors
,
2006
:
A dynamical link between the Arctic and the global climate system
.
Geophys. Res. Lett.
,
33
,
L03703
, doi:.
Ding
,
Y.
,
1990
:
Build-up, air mass transformation and propagation of Siberian high and its relation to cold surge in East Asia
.
Meteor. Atmos. Phys.
,
44
,
281
292
, doi:.
Ding
,
Y.
, and
T. N.
Krishnamurti
,
1987
:
Heat budget of the Siberian high and the winter monsoon
.
Mon. Wea. Rev.
,
115
,
2428
2449
, doi:.
Francis
,
J. A.
, and
S. J.
Vavrus
,
2012
:
Evidence linking Arctic amplification to extreme weather in mid-latitudes
.
Geophys. Res. Lett.
,
39
,
L06801
, doi:.
Francis
,
J. A.
,
W.
Chan
,
D. J.
Leathers
,
J. R.
Miller
, and
D. E.
Veron
,
2009
:
Winter Northern Hemisphere weather patterns remember summer arctic sea-ice extent
.
Geophys. Res. Lett.
,
36
,
L07503
, doi:.
Honda
,
M.
,
J.
Inous
, and
S.
Yamane
,
2009
:
Influence of low Arctic sea-ice minima on anomalously cold Eurasian winters
.
Geophys. Res. Lett.
,
36
,
L08707
, doi:.
Hopsch
,
S.
,
J.
Cohen
, and
K.
Dethloff
,
2012
: Analysis of a link between fall Arctic sea ice concentration and atmospheric patterns in the following winter. Tellus,64A, 18624, doi:.
Inoue
,
J.
,
M. E.
Hori
, and
K.
Takaya
,
2012
:
The role of Barents Sea ice in the wintertime cyclone track and emergence of a warm-Arctic cold-Siberian anomaly
.
J. Climate
,
25
,
2561
2568
, doi:.
Jaiser
,
R.
,
K.
Dethloff
,
D.
Handorf
,
A.
Rinke
, and
J.
Cohen
,
2012
: Impact of sea ice cover changes on the Northern Hemisphere atmospheric winter circulation. Tellus,64A, 11595, doi:.
Jeong
,
J.
,
T.
Ou
,
H.
Linderholm
,
B.
Kim
,
S.
Kim
,
J.
Kug
, and
D.
Chen
,
2011
:
Recent recovery of the Siberian high intensity
.
J. Geophys. Res.
,
116
,
D23102
, doi:.
Jhun
,
J. G.
, and
E. J.
Lee
,
2004
:
A new East Asian winter monsoon index and associated characteristics of the winter monsoon
.
J. Climate
,
17
,
711
726
, doi:.
Jung
,
T.
,
M. A.
Kasper
,
T.
Semmler
, and
S.
Serrar
,
2014
:
Arctic influence on subseasonal midlatitude prediction
.
Geophys. Res. Lett.
,
41
,
3676
3680
, doi:.
Li
,
S.
,
2004
:
Impact of northwest Atlantic SST anomalies on the circulation over the Ural Mountains during early winter
.
J. Meteor. Soc. Japan
,
82
,
971
988
, doi:.
Liu
,
J. P.
,
J. A.
Curry
,
H. J.
Wang
,
M. R.
Song
, and
R. M.
Horton
,
2012
:
Impact of declining Arctic sea ice on winter snowfall
.
Proc. Natl. Acad. Sci. USA
,
109
,
4074
4079
, doi:.
Livezey
,
R.
, and
W. Y.
Chen
,
1983
:
Statistical field significance and its determination by Monte Carlo techniques
.
Mon. Wea. Rev.
,
111
,
46
59
, doi:.
Magnusdottir
,
G.
,
C.
Deser
, and
R.
Saravanan
,
2004
:
The effects 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. Climate
,
17
,
857
876
, doi:.
Overland
,
J. E.
, and
M.
Wang
,
2010
:
Large-scale atmospheric circulation changes are associated with the recent loss of Arctic sea ice
.
Tellus
,
62A
,
1
9
, doi:.
Peings
,
Y.
, and
G.
Magnusdottir
,
2014
:
Response of the wintertime Northern Hemisphere atmospheric circulation to current and projected Arctic sea ice decline: A numerical study with CAM5
.
J. Climate
,
27
,
244
264
, doi:.
Peng
,
S.
,
W. A.
Robinson
, and
S.
Li
,
2003
:
Mechanisms for the NAO response to the North Atlantic SST tripole
.
J. Climate
,
16
,
1987
2004
, doi:.
Petoukhov
,
V.
, and
V.
Semenov
,
2010
:
A link between reduced Barents-Kara sea ice and cold winter extremes over northern continents
.
J. Geophys. Res.
,
115
,
D21111
, doi:.
Porter
,
D.
,
J.
Cassano
, and
M.
Serreze
,
2012
:
Local and large-scale atmospheric responses to reduced Arctic sea ice and ocean warming in the WRF model
.
J. Geophys. Res.
,
117
,
D11115
, doi:.
Rinke
,
A.
,
K.
Dethloff
,
W.
Dorn
,
D.
Handorf
, and
J. C.
Moore
,
2013
:
Simulated Arctic atmospheric feedbacks associated with late summer sea ice anomalies
.
J. Geophys. Res. Atmos.
,
118
, 7698–7714, doi:.
Roeckner
,
E.
, and Coauthors
,
2003
: The atmospheric general circulation model ECHAM5. Part I: Model description. Max Planck Institute for Meteorology Rep. 349, 127 pp. [Available from MPI for Meteorology, Bundesstr. 53, 20146 Hamburg, Germany.]
Screen
,
J. A.
, and
I.
Simmonds
,
2010
:
The central role of diminishing sea ice in recent Arctic temperature amplification
.
Nature
,
464
,
1334
1337
, doi:.
Screen
,
J. A.
,
C.
Deser
,
I.
Simmonds
, and
R.
Tomas
,
2014
:
Atmospheric impacts of Arctic sea-ice loss, 1979–2009: Separating forced change from atmospheric internal variability
.
Climate Dyn.
, 43, 333–344, doi:.
Takaya
,
K.
, and
H.
Nakamura
,
2001
:
A formulation of a phase-independent wave-activity flux for stationary and migratory quasigeostrophic eddies on a zonally varying basic flow
.
J. Atmos. Sci.
, 58, 608–627, doi:.
Tang
,
Q.
,
X.
Zhang
,
X.
Yang
, and
J.
Francis
,
2013
:
Cold winter extremes in northern continents linked to Arctic sea ice loss
.
Environ. Res. Lett.
,
8
, 014036, doi:.
Vihma
,
T.
,
2014
: Effects of Arctic sea ice decline on weather and climate: A review. Surv. Geophys.,35, 1175–1214, doi:.
Walsh
,
J. E.
,
2014
: Intensified warming of the Arctic: Causes and impacts on middle latitudes. Global Planet. Change,117, 52–63, doi:.
Wu
,
B.
, and
J.
Wang
,
2002
:
Winter Arctic Oscillation, Siberian high and East Asian winter monsoon
.
Geophys. Res. Lett.
,
29
,
1897
, doi:.
Wu
,
B.
,
R.
Huang
, and
D.
Gao
,
1999
:
Effects of variation of winter sea-ice area in Kara and Barents Seas on East Asian winter monsoon
.
Acta Meteor. Sin.
,
13
,
141
153
.
Wu
,
B.
,
R.
Zhang
, and
R.
D’Arrigo
,
2006
:
Distinct modes of the East Asian winter monsoon
.
Mon. Wea. Rev.
,
134
,
2165
2179
, doi:.
Wu
,
B.
,
J.
Su
, and
R.
Zhang
,
2011
:
Effects of autumn–winter Arctic sea ice on winter Siberian high
.
Chin. Sci. Bull.
,
56
,
3220
3228
, doi:.
Wu
,
B.
,
J. E.
Overland
, and
R.
D’Arrigo
,
2012
:
Anomalous Arctic surface wind patterns and their impacts on September sea ice minima and trend
.
Tellus
,
64
,
18590
, doi:.
Yang
,
L.-N.
, and
B.
Wu
,
2013
:
Interdecadal variations of the East Asian winter surface air temperature and possible causes
.
Chin. Sci. Bull.
,
58
,
3969
3977
, doi:.