Certain large-scale flow patterns typically persistbeyond the periods associated with synoptic-scale variability in the extratropics(Dole,1986). Two of themost frequently described and discussed phenomenaare blocking(e.g.,Rex,1950; Sumner,1954) and teleconnection(e.g.,Wallace and Gutzler, 1981; Barnston and Livezey, 1987).

Atmospheric blocking is one of the most important weathers in midlatitudes and has long been recognized as a physical process of profound dynamical interest and of great practical relevance to operational forecasting(Tibaldi and Molteni, 1990). Theregion around the Ural Mountains is the third preferred region for blocking occurrence associated withthe Mediterranean storm track(Dole and Gordon, 1983; Wang et al., 2010). The circulation anomaliesover this sector have important effects on the EastAsian weather and climate(Tao,1957; Ye et al., 1962).A strong positive height anomaly over the Urals inwinter is related to a colder surface temperature inEast Asia(Li,2004). If the blocking high over theUrals persists,the long-lasting anomalies associatedwith the blocking would be responsible for the extreme events. For example,severe snowstorms occurred across China in January 2008 while a longlived blocking high occurred over the Urals(the "0801"event,hereafter). Recently,Wang et al.(2010)foundthat the Ural blocking has exerted more influences onthe East Asian winter climate following the climateshift in the mid 1970s. The circulation pattern overthe Urals is one of the critical factors in winter seasonal prediction for East Asia(Li,2004),making veryimportant to underst and it the variability of the Uralblocking high.

The relation between blocking and teleconnectionpatterns has been widely investigated(e.g.,Dole,1986;Wiedenmann et al., 2002; Barriopedro et al., 2006).As previously indicated,blocking high and teleconnection are the most prominent persistent anomalies, and both phenomena share some common features,including a two-week time period and intriguing structuralsimilarities. The Arctic Oscillation(AO) and NorthAtlantic Oscillation(NAO)are two of the most important teleconnection patterns in mid-high latitudes.The AO represents the leading empirical orthogonalfunction of a winter sea level pressure(SLP)field and has a zonal symmetrical appearance associated withthe polar vortex. The NAO represents the difference between the Icel and ic Low(IL) and the AzoresHigh(AH)as normalized monthly mean SLP anomalies and has a more local dipole structure. The regional weather and climate over East Asia are greatlyinfluenced by the variability of AO/NAO and by theUral blocking. Park et al.(2011)noted that the coldsurge over East Asia associated with blocking tends tooccur during negative AO periods. Because the Uralblocking is one of the most important weather systems over East Asia,the relation between this system and the AO/NAO deserves a further study. In recentyears,many investigations observed an eastward shiftof the center of action of the NAO since the late 1970s(Hilmer and Jung, 2000). We should concern not onlythe index and phase of the AO/NAO,but also thestructure of the teleconnections. As for the impactsof AO/NAO anomaly patterns,we have a paucity ofknowledge.

This study discusses the interannual variabilityof the Ural blocking and its relation to the AO/NAO.Section 2 describes the data and analysis methods usedin this study. Section 3 details the interannual variability of the Ural blocking and its influence on theEast Asian climate. Section 4 presents the statistical relationship between the AO/NAO and the Uralblocking, and then describes the circulation anomalyfeatures associated with the Ural blocking in some abnormal years. Section 5 documents some possible explanations for the circulation anomaly features. Thesummary and discussion are given in Section 6.2. Data and methods

The reanalysis data used in the present studyare from the National Centers for EnvironmentalPrediction-National Center for Atmospheric Research(NCEP-NCAR)(Kalnay et al., 1996). Variables analyzed include the monthly mean geopotential height,surface air temperature(SAT),zonal wind, and dailygeopotential height from January 1960 to December2008. The data are on a 2.5°×2.5° horizontal resolution and extend from 1000 to 10 hPa at 17 verticalpressure levels. The AO and NAO indices(AOI and NAOI)are taken from the Climate Prediction Centerof NCEP.

The blocking index of Tibaldi and Molteni(1990)(TM90,hereafter)is used here to examine the interannual variation of the Ural blocking. The TM90 index isdefined based upon two values of daily 500-hPa geopotential height gradient evaluated at each longitude,

To identify the characteristics of the quasistationary planetary wave,we apply zonal Fourier harmonics to the geopotential height. By exp and ing themonthly mean fields into their zonal Fourier harmonics,the zonal wavenumbers 1-3 are used to represent aquasi-stationary planetary wave following the methodof van Loon et al.(1973).

The time period analyzed is from 1960 to 2008.We use January monthly mean to represent the borealwinter and contrast it with the "0801" event, and thewinter of 2008 refers to only January 2008. In thisstudy,the climatology mean is defined as the averagefrom 1960 to 2008, and the anomaly is the departurefrom this climatology mean. Two-sided Student's t-tests are applied to test the significance of the composites and correlations.3. Interannual variability of the Ural blocking and its possible impact on the East Asian winter climate

The activity of the blocking high over theEurasian sector(0°-90°E)is characterized by obviousseasonal variability and interannual variability(Barriopedro et al., 2006). In this section,we first describethe definition and interannual variability of the Uralblocking high and then its possible impact on the winter climate over East Asia.

A positive(negative)anomaly of 500-hPa heightover the Ural Mountains represents enhanced(weakened)blocking activity over this sector(Li,2004).Li(2004)used a normalized height anomaly at thekey point(60°N,60°E)to represent the circulationanomaly over the Urals. Li and Gu(2010)usedarea averaged(45°-65°N,40°-70°E)monthly meangeopotential height anomalies at 500 hPa as the index of the Ural high(UHI). The geopotential height and geopotential height anomaly at 500 hPa are usedto estimate the activity of blocking. Many blockingdetection methods employ daily 500-hPa geopoten-tial height(e.g.,TM90)or daily 500-hPa geopotentialheight anomaly(e.g.,Dole and Gordon, 1983)as thebase fields.

In this study,we use the area averaged(50°-70°E)TM90 index to represent the blocking activity over theUrals. Figure 1 illustrates the interannual variationsof the Ural blocking frequency(URBF) and amplitude(URBA). The interannual variation of URBA ishighly consistent with that of URBF,the variation ofwhich is discussed subsequently. Meanwhile,the correlation coefficient of UHI and URBF is 0.415,whichexceeds the 99% confidence level. The high correlation implies that UHI can also express some featuresof the Ural blocking. If there was an open ridge or cyclone over the Urals in a nonblocked event,the TM90index would not distinguish the difference among atmospheric circulation anomaly patterns. The UHI isanother supplemental tool that identifies the characteristics of circulation anomalies over the Urals,especially for the nonblocked events.

Fig. 1. Interannual variations of the Ural blocking high(a)frequency and (b)amplitude(unit: m deg-1),using theTM90 blocking index.

Figure options

We then investigate the possible impact of theUral blocking on the winter climate over East Asia.Figure 2 displays the composite SAT of the months inwhich the URBF is greater than 0.16(approximately 5days). A significant negative SAT anomaly dominatesthe eastern part of China,possibly associated withthe activity of the Ural blocking. The Ural blockingwould induce anomalous cold temperature anomaliesdownstream of the blocking high due to the northerlyadvection by the anomalous meridional flow. Thus,itis very important to explore the relationship betweenthe Ural blocking and related driving factors.

Fig. 2. Composite of January surface air temperature(SAT)anomalies(K)for the Ural blocking frequency greaterthan 0.16(5 days). Light and dark shadings denote the regions that exceed the 90% and 95% confidence levels.

Figure options

Though many previous studies have explored theinfluence of teleconnection patterns on the blockinghigh,the relation remains unclear. Considering theimpact of the AO/NAO on the East Asian winter monsoon(e.g.,Gong et al., 2001; Wu and Wang, 2002;Chen et al., 2005),the relation between the AO/NAO and the Ural blocking deserves attention. In January,the correlation coefficient of the AOI and NAOI is0.728 during 1960-2008,exceeding the 99% confidencelevel. Although the correlation between the AO and NAO is very high,the relation between the AO/NAO and the Ural blocking may differ. In this section,we first survey the statistical relation between theAO/NAO and the Ural blocking high in January and then discuss the situation in certain individual years.4.1 The statistical relationship

In January,the 500-hPa geopotential heightanomalies and the AOI exhibit a negative correlationcoefficient of approximately -0.7 in the polar areas,related to the polar vortex(Fig. 3a). In the midlatitudearea,there exists a positive correlation b and with amaximum coefficient value of approximately 0.6. Thepositive correlation b and has two large-value centersover the northern Atlantic Ocean and northern PacificOcean,which coincide with the two main sectors ofblocking activities associated with the storm track overthe oceans. The b and breaks up over Eurasia and North America. This distribution of correlation reflects the zonal l and -sea thermal contrast. A negativecorrelation center over the Ural Mountains extendsto the polar areas. The 500-hPa geopotential heightanomaly over the Ural Mountains is positive duringthe negative AO phase,possibly favoring the blockinghigh activity, and vice versa.

Fig. 3. Distributions of the correlation between(a)AO,(b)NAO and 500-hPa geopotential height anomalies(in gpm)in January of 1960-2008. Contour intervals are 0.1, and the shadings denote regions that exceed the 95% confidencelevel.

Figure options

We also investigated the relationship between theAO and the Ural blocking using the TM90 blockingindex. When the AOI is greater than 0.6σ(st and arddeviation),a positive-AO-anomaly month is defined, and when less than -0.6σ,a negative-AO-anomalymonth is defined. According to this criterion,a total of 10(15)positive-(negative-)anomaly monthsis defined during 1960-2008(Table 1). The averageURBF and URBA in the negative-(positive-)AOanomaly month are 0.056 and 1.28 m deg^-1(0.002 and 0.013 m deg^-1),respectively. The difference ofURBF(URBA)between the negative and positiveAO phases exceeds the 99%(98%)confidence level.The preceding analyses indicate that the Ural blocking high tends to be more persistent and tense duringthe negative AO phase than during the positive phase,which is consistent with the finding of Li and Gu(2010).

Table 1. January AO anomaly years during 1960-2008

Table options

Figure 3b depicts the correlation of the JanuaryNAO index and 500-hPa geopotential height anomalies. A dipole pattern of correlation with the maximumcenters over the northern Atlantic Ocean and a zonalasymmetric structure are shown. The negative center over Greenl and represents the IL, and the positivecenter over the northern Atlantic Ocean(40°N)represents the AH. The IL center at 500 hPa is locatedover Greenl and ,consistent with that in Ulbrich and Christoph(1999). In addition,a weaker secondarydipole is located over the downstream of the northernAtlantic Ocean. The weaker dipole exhibits a positive center north of the Urals and a negative center south of the Urals which is not statistically significant. The relative low correlation coefficient overthe Urals(60°N,60°E)indicates that the relationshipbetween the NAO and the circulation over this sector is not very close statistically. From the precedingcorrelation analyses,the activity of the Ural blockingappears closely related to the different phases of theAO.4.2 The eastward extending of the IL and the Ural blocking

A severe freezing rain and snow storm attackedsouthern China in January 2008. One of the maincauses of the "0801" event is the activity of blocking over the Urals. The center of the positive heightanomaly related to the Ural blocking high(UR)waslocated at 65°N,60°E(Fig. 4),a slight northwardshift from the climatological location of the center ofthe Ural blocking(60°N). In January 2008,the Uralblocking occurred while the AO was in a positive period(hereafter referred to as a positive UR-AO eventfor brevity). Li and Gu(2010)indicated that the situation for January 2008 does not align with the statistical relationship between the AO and the UR. In contrast,there are some years in the negative AO phasein which height anomalies over the Urals are negative(negative UR-AO events,hereafter).

Fig. 4. Monthly mean 500-hPa geopotential heightanomalies(units: gpm)in January 2008. Contour intervalsare 20 gpm.

Figure options

Table 2 lists the UR-AO events during 1960-2008.One criterion for the UR-AO events is an absolutevalue of UHI greater than 40 gpm. The average valueof URBF is 0.261 for positive UR-AO events,indicating that the blocking is very persistent. On thecontrary,no blocking event is detected in the negativeAO-UR events using the TM90 blocking index.

Table 2. Quantitative characteristics of the two groups of UR-AO events

Table options

Figure 5 illustrates the composites of 500-hPageopotential height anomalies for the two groups ofUR-AO events and shows an asymmetrical zonal structure of the atmospheric circulation. The significant signatures are mainly concentrated over the UralMountains and the northern Atlantic Ocean,with insignificant anomalies over the northern Pacific Ocean.The common feature of both positive and negativeUR-AO events is that the IL exhibited eastward extending. Originally,the climatological center of theIL is located over Greenl and at 500 hPa(see Fig. 3b). When the center of the IL shifted to east of30°W,the NAO dipole anomaly pattern in upstreamregion exhibited a northeast-southwest tilting and exerted more impacts on the UR occurrence,undermining the statistical relationship between the AO and the UR. The AH,IL, and the anomaly center overthe Urals exhibit poleward(equatorward)tilting wavetrain-like characteristics in the positive(negative)URAO events. The case of the Ural blocking in January2008(Fig. 4)matches the composite of the positiveUR-AO events. Additionally,the centers of 500-hPageopotential height anomalies over the Urals in twogroups exhibit slight differences. The anomaly centerin the positive UR-AO events is located around 65°N,60°E, and the counter center in the negative UR-AOevents is located around 55°N,45°E.

Fig. 5. Composites of January 500-hPa geopotential height anomalies for the(a)positive and (b)negative UR-AOevents. Contour intervals are 20 gpm. Light and dark shadings denote regions that exceed the 90% and 95% confidencelevels.

Figure options

The eastward extending of the center of the IL and the blocking high over Eurasia in January 2008can also be presented in a synoptic view. Luo et al.(2007)noted that the positive-phase NAO favors theoccurrence of European blocking events. Consistentwith this result,a pre-existing anticyclonic anomalyoccurred over Sc and inavia/western Russia from lateDecember 2007 to early January 2008(Bueh et al., 2011). Bueh et al.(2011)analyzed the processes of aEuropean blocking extending eastward by using transient eddy feedback forcing. During January 2008,four heavy precipitation events occurred(Tao and Wei, 2008). The eastward shift of the center of the IL and the blocking high in the first precipitation eventwere the most prominent. Figure 6a indicates that theblocking was over the European continent(30°E) and the center of IL was located west of 30°W on 1 January 2008. Afterward,the blocking extended eastwardto the Urals(60°E), and the center of the IL shiftedeast of 30°W on 7 January 2008(Fig. 6b). The eastward extending of the center of the IL was alsoassociated with the reconstruction of the blocking high around the Urals in the subsequentthree precipitation episodes in January 2008(figureomitted).

Fig. 6. Daily 500-hPa geopotential height anomaly fields on(a)1 and (b)7 January 2008. Contour intervals are 75gpm.

Figure options

From the preceding analyses,we can see the importance of the eastward shift of the IL and the associated NAO dipole anomaly pattern in the upstreamregion for the activity of the Ural blocking in the URAO events. The cause and effect of the eastward shiftof the IL center are also noteworthy.

The cause of the eastward shift of NAO center hasbeen widely discussed in previous studies. Ulbrich and Christoph(1999)argued that increasing greenhousegas concentrations are the main cause for the eastwardshift of the center of action of the observed NAO pattern. Peterson et al.(2003)noted that this eastwardshift may be due to an increase in the strength of themean westerly wind in the Atlantic basin. Luo and Gong(2006)confirmed that in a strong mean westerly wind,the mean flow-induced eastward shift of theNAO exceeds the eddy-induced westward shift so thatthe NAO anomaly can undergo a net eastward shift.

Recently,Luo et al.(2010a,b)indicated that themeridional distribution of the jet may contribute tothe NAO dipole anomaly pattern according to the dynamical analytical solution and numerical model experiments. An initial symmetric dipole anomaly inthe meridional direction can evolve into a northeast-southwest(NE-SW)or northwest-southeast(NW-SE)tilted dipole structure if the core of this jet is in higherlatitudes(the north)or in lower latitudes(the south).The results predicted by the linear Rossby wave theoryin slowly varying media also confirm this. Climatologically,the maximum westerly wind core over the NorthAtlantic Ocean occurs at 45°N(figure omitted),deducing a faster zonal phase speed of the NAO dipolein midlatitudes than in high latitudes. Such a phasespeed distribution causes the NW-SE pattern of theNAO. The core of the northward(southward)jet shiftcan cause the zonal wind to intensify(weaken)in highlatitudes and to weaken(intensify)in midlatitudes(seeFig. 1 in Luo et al., 2010b).

Figure 7 illustrates the January zonal wind at 500hPa over the North Atlantic(60°W-0°). The solid linein Fig. 7a(Fig. 7b)represents the mean zonal windof the positive(negative)AO anomaly months in Table 1. The mean zonal wind in high latitudes(40°-60°N)of the positive UR-AO events is larger thanthe mean of the AO anomaly months with strongerbaroclinity. On the contrary,the mean zonal wind inmidlatitudes(20°-40°N)of the positive UR-AO eventsis smaller than the mean of the AO anomaly monthswith relaxed baroclinity, and the alteration latitudeis approximately 40°N. The meridional distributionof zonal wind of the positive UR-AO events may explain the eastward shift of the IL center and the NAOdipole anomaly pattern,according to the theory of Luo et al.(2010a,b). The negative AO phase is similar to the positive AO phase,except that the alteration latitude is approximately 44°N in the negativeAO phase,higher than its counterpart in the positiveUR-AO events. The meridional distribution of zonalwind over the North Atlantic in January 2008(thinline in Fig. 7a)is identical to the mean of the positiveUR-AO events. Thus,the anomaly of zonal wind inthe Atlantic sector may explain the IL eastward shift,which is very important for the Ural blocking high activity in January 2008.

Fig. 7. Meridional distributions of January zonal wind at 500 hPa(U500)in the Atlantic sector(60°W-0°)for(a)positive AO phase and (b)negative AO phase. Solid and dashed lines represent the mean of AO anomaly months inTable 1 and the UR-AO events during 1960-2008,respectively. The thin solid line in(a)is for January 2008.

Figure options

We further examine zonal wind at 500 hPa overthe Eurasian sector(20°-90°E,Fig. 8). The mostprominent differences of zonal wind between the URAO events and the mean of corresponding AO anomalymonths in Table 1 occur primarily in midlatitudes.The larger mean zonal wind in high latitudes(64°-80°N) and the smaller one in midlatitudes(40°-64°N)of the positive UR-AO events(Fig. 8a)induce greaterphase speed and thus disperse more energy downstream to high latitudes. This distribution of basicflow coincides with the poleward tilting wave-train-likeanomaly chain in the positive UR-AO events. Themeridional distribution of basic flow of the negativeUR-AO events(Fig. 8b)is opposite to the positiveUR-AO events. The mean zonal wind of the negative UR-AO events is greater in midlatitudes(40°-60°N). More energy disperses downstream to midlatitudes,which coincides with the equatorward tiltingwave-train-like anomaly chain in the negative UR-AOevents. The situation for January 2008 is more prominent than the mean of the positive UR-AO events,explaining the poleward shift of the Ural blocking highcenter to 65°N(Fig. 4).

Fig. 8. As in Fig. 7,but for the meridional distributions of January zonal wind in the Eurasian sector(20°-90°E).

Figure options

The effect of eastward shift of the IL on theUral blocking was examined from the perspective ofquasi-stationary planetary waves. Figure 9 shows theanomalous amplitude of quasi-stationary planetarywaves for the two groups of UR-AO events. Wavenumber 2 at 60°N at 500 hPa in the positive UR-AOevents is significantly strengthened, and wavenumber3 is weakened accordingly. The half wavelength is approximately 90 degrees for wavenumber 2. While thecenter of the IL in the upstream region shifts eastward to 30°W,the strengthened wavenumber 2 explains the positive height anomaly over the Urals withthe center at 60°E. The situation is different in thenegative UR-AO events. Wavenumber 3 is significantly strengthened, and wavenumber 2 is weakenedin the negative UR-AO events. The half of wavelength is approximately 60 degrees for wavenumber3. The shortened wavelength accounts for the negative height anomaly over the Urals with the center located near 45°E. The situation of the activityof the Ural blocking in January 2008 is the sameas the positive UR-AO events. The eastward shiftof the IL and the anomalous quasi-stationary planetary waves are responsible for the circulation anomalies over the Urals in the two groups of UR-AOevents.

Fig. 9. Composites of January amplitude anomalies of the quasi-stationary planetary waves for wavenumbers 2(a) and 3(b)in positive UR-AO events, and for wavenumber 2(c) and 3(d)in negative UR-AO events. Contour intervals are 5gpm. Light and dark shadings denote regions that exceed the 90% and 95% confidence levels.

Figure options

Based on the NCEP-NCAR reanalysis data,theinterannual variations of the Ural blocking high and associated impact on the East Asian climate in January have been investigated. The surface air temperature is lower than normal due to the activity of theUral blocking. One of the main causes of the severesnowstorms in China during January 2008 is the persistent maintanence of the Ural blocking high.To investigate the interannual variability of theUral blocking,its relation to the AO/NAO was alsoanalyzed. The Ural blocking high tends to be morefrequent and tense during the negative AO phase th and uring the positive AO phase,according to the statistical analyses. The relationship between the AO and the Ural blocking does not fit in some abnormal years,including January 2008.

We focused not only on the index and phase ofthe AO/NAO but also on the structure of the teleconnections. The zonal asymmetric structure of theAO may play a more important role in the UR-AOevents. The significant features in the compositesof 500-hPa geopotential height anomalies are mainlyover the Urals and the northern Atlantic Ocean. Theanomalies over the northern Pacific Ocean are not significant. The atmospheric circulation in the UR-AOevents exhibits more NAO dipole anomaly patterns.The center of the IL in the upstream region extendseast of 30°W, and the NAO dipole shows a northeast-southwest anomaly pattern in the UR-AO events.The influence of anomalous location of the IL on theUral blocking is investigated from the perspective ofquasi-stationary planetary waves. While the centerof the IL extends eastward to 30°W in the upstream,the strengthened amplitude of zonal wavenumber 2(wavenumber 3)explains the positive(negative)500hPa geopotential height anomalies at 60°E(45°E)inthe positive(negative)UR-AO events that favor(suppress)the activity of the Ural blocking. The situationof the Ural blocking in January 2008 matches the positive UR-AO events.

As being noted,the NAO and the Ural blocking high are both nonlinear problems. Some previousstudies have considered the interaction between basic flow,planetary waves, and synoptic-scale eddies(Luo et al., 2005,2010c). We use the theory of Luoet al.(2010a,b)that assumes the scale separationto explain the anomaly patterns of the NAO and theUral blocking high. The enhanced zonal wind withstronger baroclinity in high latitudes and attenuatedzonal wind with relaxed baroclinity in midlatitudesover the North Atlantic render the eastward shift ofthe center of the IL,according to the theory of Luoet al.(2010a,b). Intensified zonal wind in high latitudes and weakened zonal wind in midlatitudes inthe Eurasian sector render the poleward tilting wavetrain-like anomaly chain in the positive UR-AO events and explain the poleward shift of the Ural blockingin the positive UR-AO events. The distributions ofzonal wind in the Eurasian sector correspond to theequatorward tilting wave-train-like anomalies chain inthe negative UR-AO events. Thus,the NAO dipoleanomaly pattern and the anomalous quasi-stationaryplanetary waves are responsible for the circulationanomalies over the Urals in the two groups of UR-AOevents.

The analyses in this study are based on the January monthly mean. In fact,the interannual variationsof the Ural blocking can also be revealed by the meansof winter(December-January-February). Wang et al.(2010)noted that the Ural blocking underwent aneastward shift after 1976/1977, and the NAO patternsimultaneously exhibited an eastward shift(Hilmer and Jung, 2000). The interdecadal relationship between the Ural blocking and the NAO anomaly pattern in the upstream also deserves further discussion.This study focuses on the internal atmospheric variations,such as the AO/NAO,quasi-stationary waves and their effects on the Ural blocking. Some studiesalso indicate that external forcing,such as sea surfacetemperature anomalies(SSTA)in the Atlantic Ocean and snow mass,can modulate the AO/NAO pattern(Gong et al., 2002; Li et al., 2007). Therefore,external factors,such as ocean or l and processes,shouldbe considered in a general circulation model in thefuture.

Acknowledgments: The authors thank twoanonymous reviewers for their valuable comments and suggestions. Thanks also go to Prof. Dehai Luo and Prof. Cholaw Bueh for many interesting discussionsduring the course of this work.