Haishan Li ,Ke Fan
a Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing, China
b School of Atmospheric Science, Sun Yat-sen University, and Southern Marine Science and Engineering Guangdong Laboratory (Zhuhai), Zhuhai, China
c University of the Chinese Academy of Sciences, Beijing, China
Keywords:Central asia Surface air temperature anomaly Arctic oscillation North atlantic oscillation Ural blocking
ABsTRACT The dominant patterns of the winter (December—February) surface air temperature anomalies (SATAs) over Central Asia (CA) are investigated in this study.The first two leading modes revealed by empirical orthogonal function(EOF) analysis represent the patterns by explaining 74% of the total variance.The positive phase of EOF1 is characterized by a monopole pattern,corresponding to cold SATAs over CA,while the positive phase of EOF2 shows a meridional dipole pattern with warm and cold SATAs over northern and southern CA.EOF1 is mainly modulated by the negative phase of the Arctic Oscillation (AO) in the troposphere,and the negative AO phase may be caused by the downward propagation of the precursory anomalies of the stratospheric polar vortex.EOF2 is mainly influenced by the Ural blocking pattern and the winter North Atlantic Oscillation (NAO).The SATAs associated with EOF2 can be attributed to a dipole-like pattern of geopotential height anomalies over CA.The dipole-like pattern is mainly caused by the Ural blocking pattern,and the NAO can also contribute to the northern part of the dipole.
Central Asia (35°—65°N,50°—100°E;CA) is one of the largest semiarid areas in the world and is characterized by a typical continental climate (Lioubimtseva and Henebry,2009 ;Mirzabaev,2013 ;Li et al.,2015).The surface air temperature (SAT) over CA during the winter months (December—February;DJF) reaches its minimum,and the low winter SAT may cause severe socioeconomic impacts.For example,as CA is well-known as a region of nomadic cattle breeding,low winter SAT in this region may cause high risks for agricultural employees and livestock in the open air (Begzsuren et al.,2004 ;Kerven et al.,2004 ;Nyssanbayeva et al.,2019).Thus,the winter SAT over CA and its variability are worthy of further study.
As part of the Eurasian continent,previous studies have revealed that the winter SAT over CA can be modulated by the large-scale atmospheric circulations and teleconnections over the whole continent.For example,the positive (negative) phase of the Arctic Oscillation (AO)/North Atlantic Oscillation (NAO) is associated with warm (cold) winter SAT(Thompson and Wallace,2000 ;Cattiaux et al.,2010);the strengthening of the winter Ural blocking pattern can lead to warming over the Ural Mountain region,and the outbreak of the blocking pattern can lead to strong cooling over downstream East Asia (Wang et al.,2010;Cheung et al.,2013 ;Luo et al.,2016);and a strengthened Siberian high is associated with cooling over most of the Eurasian continent(Cohen et al.,2001 ;Gong and Ho,2002).
Despite much research on the climatic systems that may influence the winter SAT over CA,studies focusing on the main features of the winter SAT over CA itself are relatively scarce.Thus,this study explores this issue by investigating the leading modes of the winter SAT and corresponding mechanisms.
Fig.1.(a,b) Spatial distributions of the first two leading modes for the winter SATAs over CA during 1979—2019,with the proportions of the explained variance given in the top-right corner,and (c,d) the associated normalized time series of the standardized PCs.
The atmospheric reanalysis data from the National Centers for Environmental Prediction/National Center for Atmosphere Research(Kalnay et al.,1996),with a 2.5° horizonal resolution from 1000 to 10 hPa comprising 17 pressure levels,including the daily and monthly mean SAT,geopotential height,and zonal and meridional winds in the winters of 1979—2019,are used in this study.The monthly AO and NAO indices of the Climate Prediction Center in the same period are also adopted.A two-dimensional blocking index developed by Scherrer et al.(2006),which can represent the blocking frequency in wintertime in the region 35°—65°N,is employed.The empirical orthogonal function (EOF) analysis method is used to distinguish the leading modes of the DJF surface air temperature anomalies (SATAs)over CA during 1979—2019.To illustrate the propagation of Rossby waves,the stationary wave activity flux defined by Takaya and Nakamura (2001),referred to as TN flux,is used.Other statistical methods employed include regression,composite analysis,and the Student’st-test.
3.1.Leading modes of DJF SATAs over CA
Fig.1 shows the spatial patterns of the leading two modes (EOF1 and EOF2) and their corresponding time series (PC1 and PC2).These two EOFs are both well separated (North et al.,1982) and account for 54.9% and 18.9% of the total variance,respectively (Fig.1 (a,b)).EOF1 is characterized by a monopole pattern,with the largest loading over northeastern CA,while EOF2 shows a meridional dipole pattern with opposite signs over northern and southern CA.The SATAs and atmospheric circulation anomalies corresponding to EOF1 and EOF2 are obtained by regressions upon PC1 and PC2 (Fig.2).For EOF1,significant cold SATAs appear over CA and correspond well to the significant northeasterlies at 850 hPa over northern CA (Fig.2 (a,c));while at 200 hPa,significant geopotential height anomalies over the mid-to-high latitudes(35°—90°N) display a negative AO-like pattern (Fig.2 (e)).For EOF2,the SATAs display a south—north dipole pattern over CA,with a warm center over the north side of CA and a cold center over the middle and south side of CA (Fig.2 (b)).Corresponding to the dipole SATA pattern,significant positive geopotential height anomalies at 850 hPa appear in the Urals region,accompanied by significant northeasterlies (southwesterlies) around the southern (northern) side of the anomalies (Fig.2 (d)).At 200 hPa,the geopotential height anomalies over CA display a distinct meridional dipole pattern,accompanied by a dipole structure over the North Atlantic but with opposite polarity (Fig.2 (f)).
3.2.Possible connection between the AO and EOF1
As mentioned above,the geopotential height anomalies at 200 hPa corresponding to EOF1 resemble a negative AO phase (Fig.2 (e)).The correlation coefficient between the winter AO index and PC1 is -0.67,and is statistically significant at the 99% confidence level.Considering that the winter AO and NAO are highly correlated,the correlation between the winter NAO index and PC1 should be further checked.The correlation coefficient between the winter AO index and PC1 is -0.65(significant at the 99% confidence level).However,the partial correlation between PC1 and the NAO index,in which the variability of the AO index is excluded,is insignificant at the 95% confidence level(-0.27),while the partial correlation between the AO index is still significant at the 99% confidence level (-0.42).Thus,it is reasonable to conclude that EOF1 may be linked to the negative phase of the AO rather than the NAO.The negative-AO-related SATAs and geopotential height anomalies over CA are analyzed (Fig.3).In Fig.3 (a),significant cold SATAs appear over northern Eurasia,which may be caused by the northeasterlies at 850 hPa over northern CA (Fig.3 (b)).The northeasterlies are caused by the positive geopotential height anomalies over Iceland,which comprise the northern part of a meridional dipole structure over the Atlantic-European sector,and are similar to the anomalies corresponding to EOF1.The geopotential height anomalies at 200 hPa resemble those at 850 hPa,but the intensity is stronger.Thus,the EOF1 mode may be largely influenced by the quasi-barotropic geopotential height anomalies corresponding to the negative phase of the AO.
Fig.2.Regressions of winter (a) SAT (shading;units: °C),(c) 850-hPa geopotential height (shading;units: m) and winds (vectors;units: m s -1),and (e) 200-hPa geopotential height (shading;units: m) onto the PC1 for the period 1979 -2019.(b,d,f) As in (a,c,e) but for PC2.The black dashed boxes show the region of CA;anomalies significant at the 95% confidence level are denoted by dots and purple contour lines,and vectors represent the anomalous winds significant at the 95%confidence level,as estimated by the Student’s t -test.
Baldwin and Dunkerton (1999) indicated that a strengthening (weakening) of the stratospheric polar vortex can propagate to the troposphere and cause a positive (negative) AO phase.Thus,the relationships between EOF1 and the precursory stratospheric anomalies is studied.The PC1-related polar vortex anomalies (represented by the area-averaged daily geopotential height anomalies over the polar cap (north of 60°N))are presented in Fig.4.It can be seen that significantly positive anomalies first appear in late November and early December at 10 hPa.The significantly positive anomalies from the stratosphere can propagate downwards and reach 1000 hPa in mid-December and result in a negative phase of the AO.Following Li and Gu (2010),a polar vortex index (PVI)is defined as the averaged geopotential height over the region north of 75°N to measure the intensity of the polar vortex.The correlation coefficient between the November PVI and PC1 is 0.35,and is statistically significant at the 95% confidence level.Thus,we conclude that the downward propagation of the positive signal from the stratosphere in November may cause the negative AO phase in the troposphere in the following winter,and consequently modulate the EOF1 mode.
3.3.Possible connection between the blocking pattern, NAO, and EOF2
To investigate the atmospheric anomalies associated with the EOF2 mode,the PC2-related geopotential height anomalies at 200 hPa,as well as the associated TN wave activity flux,are shown in Fig.5 (a).Significant geopotential height anomalies with a meridional dipole structure appear over CA,which closely resembles the Ural blocking pattern.Besides,significant anomalies also exist over the North Atlantic with an AO/NAO-like dipole pattern.The northern part of the dipole can excite an eastward-propagating Rossby wave,which can further contribute to the positive geopotential height anomalies over northern CA.The results indicate that the EOF2 mode may be associated with the Ural blocking and AO/NAO patterns.
Fig.3.Regressions of December (a) SAT (shading;units: °C) and (b) 850-hPa and (c) 200-hPa geopotential height (shading;units: m) and winds (vectors;units:m s -1) onto the negative AO index for the period 1979 -2019.The black dashed boxes show the region of CA;anomalies significant at the 95% confidence level are denoted by dots and purple contour lines,and vectors represent the anomalous winds significant at the 95% confidence level,as estimated by the Student’s t -test.
Fig.4.Regression of the daily geopotential height anomalies area-averaged over the polar cap poleward of 60°N (shading;units: m) onto PC1 for the period 1979—2019.Anomalies significant at the 95% confidence level are denoted by dots,as estimated by the Student’s t -test.
First,to investigate the relationship between EOF2 and the Ural blocking pattern,the PC2-related anomalous blocking frequency and 500-hPa geopotential height anomalies are given in Fig.5 (b).Significant high-frequency anomalies can be seen over the Urals region and corresponding to the local positive geopotential height anomalies.Following Wang et al.(2021),a Ural blocking index (UBI) is defined as the averaged blocking frequency over the region of (55°—65°N,40°—80°E).The correlation coefficient between UBI and PC2 is 0.72,and is statistically significant at the 99% confidence level.The UBI-related geopotential height anomalies are given in Fig.5 (c).Significant dipole-like anomalies can be seen over CA.The results imply that the Ural blocking can be a contributor to the dipole-like geopotential height anomalies associated with EOF2 over the CA region.However,only weak anomalies can be seen over the North Atlantic.Second,the connection between the EOF2 mode and the AO/NAO has also been studied.The correlation coefficients between PC2 with the AO and NAO are 0.29 (insignificant at the 95% level) and 0.44 (significant at the 99% level),respectively.Thus,EOF2 is mainly associated with the winter NAO.The NAO-related geopotential height anomalies are given in Fig.5 (d).Significant anomalies with a meridional dipole structure can be seen over the North Atlantic.The wave activity flux,emanating from the northern part of the dipole,can reach northern CA and cause significant positive geopotential height anomalies (Fig.5 (d)).Furthermore,the correlation coeffi-cient between the winter UBI and the NAO during 1979—2018 is 0.07,which implies that the combined effect of the Ural blocking pattern and the NAO may be treated as the linear combination of their own effects.Thus,we conclude that the dipole-like geopotential height anomalies over CA can be mainly related to the Ural blocking pattern;and the positive geopotential height anomalies over northern CA can be strengthened by the influence of NAO+(positive phase of the NAO).
To further investigate the combined influence of the Ural blocking pattern and the NAO on the atmospheric circulation over CA,four separate composites are constructed based on the winter UBI and NAO index.A year is taken to be an NAO+(NAO -) year if the NAO index is greater (less) than 0.6 (-0.6) standard deviations (where NAO -denotes the negative phase of the NAO).The NAO+(NAO -) years are furtherclassified as high-UBI (UBI +) years and low-UB (UBI -) years by ± 0.6 standard deviations.Table 1 lists the years selected for the four composites.For NAO+years,the numbers of UBI+and UBI -years are both 6;however,in NAO -years,the numbers of both UBI+and UBI -years are small —in particular,there is only one year for the NAO -/UBI -situation.Thus,three composites of 200-hPa geopotential height are constructed for NAO +/UBI +,NAO +/UBI -,and NAO -/UBI+years,and the corresponding wave activity fluxes are also analyzed (Fig.6).In the composite field for NAO +/UBI +,the geopotential height anomalies resemble those corresponding to EOF2 (Figs.2 (f) and 5(b)),with significant height anomalies over the Urals region and a positive NAO-like dipole pattern over the North Atlantic;and the wave train emanating from the northern part of the NAO-like dipole pattern over Greenland can propagate eastward into CA (Fig.6 (a)).The wave train may contribute directly to the positive geopotential height anomalies over northern CA.For NAO +/UBI -years,a strong positive-NAO-like dipole pattern appears over the North Atlantic,with negative anomalies over the region(50°—75°N,90°—0°W) and positive anomalies over the region (30°—45°N,90°—0°W);the anomalies over the North Atlantic can extend eastward and cause a weak meridional dipole pattern over CA,but with negative geopotential height anomalies over northern CA,which is unfavorable for the Ural blocking pattern (Fig.6 (b)).For NAO -/UBI +,a negative-NAO-like pattern exists over the North Atlantic,with significant positive geopotential height anomalies over the Scandinavian Peninsula and negative anomalies over eastern Europe;the wave activity flux emanating from the positive geopotential height anomalies can cause negative anomalies over nearly all of CA,which are also unfavorable for the Ural blocking pattern (Fig.6 (c)).Thus,the combined effect of the positive NAO phase and Ural blocking pattern can lead to the positive phase of the EOF2 mode.
Table 1 Years during 1979—2019 selected by the winter NAO index and the UBI.
Fig.5.(a) Regression of winter 200-hPa geopotential height (shading;units: m) onto the PC2 for the period 1979 -2019,and the associated horizonal components of the wave activity flux (vectors;units: m 2 s -2).(b) Regressions of winter blocking index (shading;units: %) and 500-hPa geopotential height (contours;units: m)onto the PC2 for the period 1979 -2019.Zero contours are omitted and negative values are dashed.(c,d) As in (a) but for the UBI and NAO,respectively.The black dashed boxes show the region of CA;anomalies significant at the 95% confidence level are denoted by dots and purple contour lines,as estimated by the Student’s t -test.
Fig.6.(a) Composite 200-hPa geopotential height anomalies (shading;units: m) for the NAO +/UBI+years,and the associated horizonal components of the wave activity flux (vectors;units: m 2 s -2).(b,c) As in (a) but for NAO +/UBI -years and NAO -/UBI+years,respectively.Anomalies significant at the 95% confidence level are denoted by purple contour lines,as estimated by the Student’s t -test.
This study investigates the first two leading EOF modes of the DJF SATAs over CA during the period 1979—2019.The two EOFs explain 54.9% and 18.9% of the total variance in the SATAs,respectively.EOF1 is characterized by a monopole pattern,corresponding to cold SATAs over CA,while EOF2 shows a meridional dipole pattern with warm and cold SATAs over northern and southern CA.EOF1 may be mainly modulated by the negative phase of the AO in the troposphere.In the negative AO phase,the positive height anomalies at 850 hPa located over the northwest of CA can cause significant northeasterlies over CA and the subsequently cold SATAs.Furthermore,the negative AO phase may be caused by the downward propagation of the precursory anomalies of the stratospheric polar vortex.EOF2 is associated with a dipole-like pattern of geopotential height anomalies,with positive and negative anomalies over northern and southern CA.The dipole-like pattern is mainly caused by the Ural blocking pattern,and the NAO can also contribute to the northern part of the dipole structure.The northeasterlies (southwesterlies) at 850 hPa around the southern (northern) side of the positive geopotential height anomalies (the northern part of the dipole-like structure) can lead to a dipole SATA pattern.
Funding
This work was funded by the National Natural Science Foundation of China [grant numbers 42088101 and 41730964 ] and an Innovation Group Project of the Southern Marine Science and Engineering Guangdong Laboratory (Zhuhai) [grant number 311021001].
Acknowledgements
We thank the two anonymous reviewers for their helpful comments that improved this article.