Author's personal copy - Atmospheric and Oceanic Sciencenigam/Nigam-Baxter... · Encyclopedia of Atmospheric Sciences, Second Edition, 2015, 90–109 teleconnection patterns are uncovered

Provided for non-commercial research and educational use onlyNot for reproduction distribution or commercial use

This article was originally published in the book Encyclopedia of AtmosphericSciences 2nd edition The copy attached is provided by Elsevier for the authors

benefit and for the benefit of the authors institution for non-commercial research andeducational use This includes without limitation use in instruction at your institution

distribution to specific colleagues and providing a copy to your institutionsadministrator

All other uses reproduction and distribution including without limitation commercialreprints selling or licensing copies or access or posting on open internet sites

your personal or institutionrsquos website or repository are prohibited For exceptionspermission may be sought for such use through Elsevierrsquos permissions site at

httpwwwelseviercomlocatepermissionusematerial

From Nigam S Baxter S 2015 Teleconnections In Gerald R North (editor-in-chief)John Pyle and Fuqing Zhang (editors) Encyclopedia of Atmospheric Sciences 2nd edition

Vol 3 pp 90ndash109ISBN 9780123822253

Copyright copy 2015 Elsevier Ltd unless otherwise stated All rights reservedAcademic Press

Authors personal copy

TeleconnectionsS Nigam and S Baxter University of Maryland College Park MD USA

2015 Elsevier Ltd All rights reservedThis article is a revision of the previous edition article by S Nigam volume 6 pp 2243ndash2269 2003 Elsevier Ltd

Synopsis

Teleconnections refer to the climate variability links between non-contiguous geographic regions Teleconnection patterns areextracted from analysis of the sea-leveltropospheric pressure variations on monthly (and weekly) timescales The methodsused in extraction of teleconnection patterns are discussed and applied to recent period data Teleconnections are especiallywell-developed in Northern winter when they strongly influence subseasonal variability notably in surface temperature andprecipitation the stratospheric and SST links are also noted The patterns in boreal summer are shown as are SouthernHemisphere teleconnections Aspects of tropicalndashextratropical teleconnections are discussed along with the relationshipbetween annular modes and teleconnections Finally a teleconnection analysis with pentad (5-day averaged) data is pre-sented to initiate discussion of teleconnection evolution

Introduction

The term teleconnection is often used in atmospheric sciences todescribe the climate links between geographically separatedregions The remote region need not exhibit fluctuations of thesame sign in order to be lsquoteleconnectedrsquo In fact the interestingteleconnections often involve contemporaneous variations ofopposite signs Climate analysis is facilitated by the constructionof a teleconnection map which describes the linkage betweena region of interest (a base point) and all other points in thedomain that are farther than the decorrelation length scale of thevariable Teleconnection maps thus provide information aboutthe structure of recurrent climate variability especially itscorrelation-at-a-distance features The maps are useful becauseclimate variability is often manifest with such structure forexample winter variations in temperature and rainfall oversouthern Europe and the Iberian Peninsula are frequentlyopposite to those over northwestern Europe and Scandinavia

Teleconnection maps were first constructed for meteoro-logical parameters measured at the Earthrsquos surface such as theatmospheric pressure The selection of the base point is a criticalfirst step and was historically guided by the investigatorrsquosinsights and interests Today base points can be selected moreobjectively and the robustness of teleconnection maps can beascertained by independent analyses The statistical correlationof the fluctuations provides a measure of the teleconnectionstrength The structure and strength of the teleconnectionpatterns change with season altitude choice of variable andeven temporal averaging of data There are interesting differ-ences between the hemispheres too in part due to the presenceof extensive continents andmountainous zones in the NorthernHemisphere (NH) Teleconnectivity in the NH winter circula-tion has been extensively analyzed as teleconnections accountfor a significant portion of the winter variance in this region

Climate teleconnections are present in observations thathave been averaged in time over a period that is long enough tosuppress the day-to-day weather fluctuations but short enoughto retain the seasonal-to-interannual component of climatevariability Although teleconnection patterns often evolve onsubmonthly timescales (weeks) their spatial patterns arecharacterized in monthly and seasonal data Monthly averages

are typically used since a month is longer than the period ofmost large-scale synoptic waves in the troposphere Climateteleconnections thus highlight the lsquostandingrsquo component oflow-frequency variability ndash one with geographically fixed nodesand antinodes The connectivity of remote regions manifest inthe teleconnection map indicates the potential significance ofremote forcing in the generation of regional climate anomaliesA teleconnection map based on contemporaneous correlationshowever cannot by itself discriminate between the forcing andresponse regions

Although the structure of the prominent teleconnectionpatterns has been known for some time the reasons for theirorigin are not yet well understood For example themechanismsthat excite and sustain the North Atlantic Oscillation (NAO) ndashone of the notable and earliest discovered patterns ndash are stillbeing investigated In the context of such investigations it hasbeen questioned if the teleconnection patterns that are typicallyregional (eg NAO) robustly portray the spatial structure ofvariability from the viewpoint of elucidation of the underlyingdynamical processes

Such concerns are relevant since the canonical tele-connection patterns typically represent the mature phase ofvariability The mature-phase pattern however need notresemble the nascent-phase structure which may be morerevealing of the excitation mechanism Identification of theevolution process from analysis of the mature-phase structureis thus difficult The spatial imprint of variability captured bya teleconnection pattern can also be ineffective in revealing theunderlying mechanisms if the region in question is the locus oftwo temporally independent physical andor dynamicalprocesses While cautionary these remarks do not call fora radically new analysis paradigm Instead they point to theneed for more comprehensive analysis of variability particu-larly in the spatiotemporal domain to facilitate insights intothe evolutionary process While in-depth analysis of this kind isbeyond the scope of this article it is important to let the readerknow that contemporary teleconnection research is rooted inspatiotemporal analyses

Additionally while teleconnectivity has historically beenidentified from assessment of the correlation between somedefined base point and distant points in the domain not all

90 Encyclopedia of Atmospheric Sciences 2nd Edition Volume 3 httpdxdoiorg101016B978-0-12-382225-300400-X

Encyclopedia of Atmospheric Sciences Second Edition 2015 90ndash109


teleconnection patterns are uncovered from such analysisExamples on subseasonal to seasonal timescales include theMaddenndashJulian Oscillation (MJO) and the annular modesboth of which are briefly discussed in subsequent sections Anexample on interannual timescales is the El Nintildeo-SouthernOscillation (ENSO) which impacts climate around the globe

Analysis Method

Teleconnection patterns can be extracted from correlationanalysis and from the calculation of principal components(PCs) among other techniques Bothmethods have been widelyused in climate research and each offers some advantages

Correlation

Correlation analysis is the more straightforward of the twomethods Consider a meteorological field such as geopotentialheight that denotes the height of an isobaric surface in theatmosphere Geopotential height f is a function of longitudeand latitude and assume that it is defined at M grid points fi

represents height at the ith longitudendashlatitude grid pointGeopotential height is also a function of time and its monthlymean value is assumed to be available for several winters(Nwinter months) Interest in the variability of climate leads tothe consideration of departures of monthly mean heights fromtheir calendar month climatologies with prime (0) denotingthe departures f0

ik representing the height departure at the ithgrid point in the kth month The correlation in height depar-tures at two grid points i and j is denoted byHij and defined as

Hij frac14PN

kfrac141f0ikf

0jkPN

kfrac141f02ik

1=2PNkfrac141f

02jk

1=2Knowledge of the correlation matrix H can be used to

construct a teleconnectivity map (T) that objectively identifiesthe base points associated with various teleconnection patternsThe map is constructed by associating the magnitude of thestrongest negative correlation between a grid point and allothers with that grid point ie Ti frac14 j most negative member inthe ith row of the correlation matrix H j The local maxima inthis map (T) identify the potential base points Linkage betweenthe neighboring base points is assessed by examining the sites oftheir strongest negative correlations A cluster of linked basepoints constitutes the core of the teleconnection pattern andthree prominent patterns are identifiable using this technique

Teleconnection Map

Climate teleconnections were first investigated in the sea-levelpressure (SLP) field SLP is however an ill-defined quantityover land particularly near the mountains it can moreover beinfluenced by local meteorological processes Teleconnectivityis thus better analyzed in upper-air data which became avail-able since the mid to late 1940s The preferred analysis variablein recent decades has been the geopotential height ndash

a measured quantity whose horizontal and vertical gradientsare proportional to the wind and temperature respectivelyNorthern winter is the preferred season for teleconnection

analysis in view of the pronounced climatological stationarywaves (from impressive mountains and continents in the NH)that impart a significant lsquostandingrsquo component to the interan-nual fluctuations

Figure 1(a) shows the teleconnectivity in the 500-mb geo-potential height field during Northern winter The map isconstructed from correlation analysis of December Januaryand February (DJF) height anomalies from 0 to 90 N for the1979ndash2008 period the data are from NOAArsquos Climate ForecastSystem Reanalysis (CFSR) This recent reanalysis takes advan-tage of the latest advances in data assimilation and incorporatesimportant data from the satellite era and thus only extendsback to 1979 This figure is comparable to Figure 7(b) in thepioneering study by Wallace and Gutzler (1981) It identifiesthree major teleconnection patterns in the NH the NAO theNorth Pacific OscillationWest Pacific pattern (NPOWP) andthe PacificndashNorth America (PNA) pattern The NAO and NPOWP each consist of two base points over the Atlantic and Pacificbasins respectively The PNA as identified by Wallace andGutzler (1981) consists of four base points the subtropicalPacific near the dateline the Aleutians interior northwesternNorth America and the southeastern United States Notehowever that the fourth base point over southeastern UnitedStates is missing in Figure 1(a)

This brings to light an issue regarding the stability of tel-econnection patterns Changing the domain and time periodof the analysis can introduce some differences Besides themissing PNA center there are other differences betweenthe seminal analysis of Wallace and Gutzler and Figure 1(a)The NAO-related pattern exhibits a noticeable eastward shiftas does the third PNA center Note data set differences areunlikely to be the origin as the large-scale rotational flow issimilarly represented in atmospheric data sets While notablethe teleconnectivity differences are viewed as modest consid-ering the nonoverlapping periods of the two analyses Thisassessment is supported by Figure 1(b) which shows theteleconnectivity map from a similar analysis of longer perioddata (1949ndash2009) one encompassing the previousnonoverlapping periods While the NAO centers and the thirdPNA center here are more in accord with the Wallace andGutzler analysis the fourth PNA center remains indiscernibleInterestingly the NPOWP centers of action are also notidentifiable in the longer period analysis Even so the overallstructure of the teleconnectivity map is similar in Figure 1(a)

The correlation analysis reveals the base points of a tele-connection pattern An index constructed using the magnitudeof fluctuations at the base points is used to track the patternamplitude For example the PNA index is defined as

aPNAethkTHORN frac14 13

f0Ak

SA f0

Bk

SBthorn f0

Ck

SC

where A B and C are the marked base points of the PNApattern (Figure 1(a)) Si in the above definition denotes thestandard deviation of height anomalies at the ith base pointLikewise an index can be constructed for the NAO- andNPOWP-related patterns using two base points in each case

Correlation analysis is a physically intuitive and objectivemethod for identifying climate teleconnections but the ob-tained patterns may not be independent especially if spatial

General Circulation of the Atmosphere j Teleconnections 91



structures overlap This for instance is the case in the Pacificwhere three prominent patterns ndash PNA WPNPO and theENSO response (shown later) ndash overlap to varying extent in theextratropical sector Not surprisingly the correlation methodproved unsuccessful in capturing the ENSO response in theextratropics as a teleconnection pattern Could this be a conse-quence of focusing on analysis of midtropospheric variabilityThe 500-hPa level ndash a level of near-zero divergence ndash is perhapsnot the level of choice for identifying tropicalndashextratropicalinteractions instigated by deep convective heating in theTropics Horizontal divergence associated with deep convectionsuch as during El Nintildeo winters is usually strongest in thetropical upper troposphere (200 hPa) The associatedmidlatitude response on the other hand is quasi-geostrophic incharacter and hence approximately nondivergent Tropicalndashextratropical interactions are thus best diagnosed at a pressurelevel that captures the divergent outflow in the Tropics and thatis near a nondivergent level in the extratropics The 200-hPalevel meets these criteria to a large extent

Repeating the correlation analysis of height fluctuations at200 hPa did not yield any new information on teleconnectivityIn particular no new base points were identified and the ENSOresponse in the extratropics remained unidentified as before

Principal Component Analysis

Principal component analysis (PCA) is an elegant and widelyused method for determining the structure of recurrent vari-ability A common name for it is empirical orthogonal function(EOF) analysis This method also analyzes the structure of thecorrelation matrix ndash the covariance matrix is preferred though ndash

but focuses on regions that account for a substantial portion ofthe temporal variance rather than just those which exhibitstrong negative correlations with distant points in the domainIn contrast with the previous method the technique yields boththe spatial patterns of recurrent variability and the extent towhich these are present or projected in the observed anomalyrecord The projection or amplitude is called the PC while thespatial pattern is referred to as the loading vector (LV) in thetechnical literature

Recurrent variability patterns identified from PCA arespatially and temporally independent or orthogonal Suchrelationship among patterns is often helpful in investigatingthe origin and governing mechanisms of variability but can berelaxed in one of the dimensions ndash space or time ndash if theorthogonality constraints prove restrictive For example it isconceivable for two temporally independent variabilitypatterns to have overlapping spatial structure Imposition oftemporal and spatial orthogonality constraints in this case maynot lead to a physically meaningful analysis

The EOF analysis results in a series of spatial patterns or LVseach one explaining successively smaller portion of the totaltemporal variance One can view these recurrent patterns asteleconnection patterns The pattern amplitude is given by thetime-dependent PC The anomaly record can be reconstructed bysumming the product of the spatial patterns (LVs) and their PCs

Rotated Principal Component Analysis

PCA provides an efficient and unique characterization ofrecurrent variability in terms of a small number of uncorrelatedspatial patterns The patterns are chosen so that each one

Figure 1 (a) 500-mb geopotential height teleconnectivity map for Northern Hemisphere winter (DJF) comparable to seminal Figure 7(b) in WallaceJM Gutzler DS 1981 Teleconnections in the geopotential height field during the northern hemisphere winter Mon Weather Rev 109 784ndash812Contourshading starts at a correlation of 05 and increases at 01 intervals Points A B Bb and C refer to the centers of actions of the PNA patternPoints B and Bb are generally associated with the Aleutian Low center of action (b) 500-mb geopotential height teleconnectivity map for NorthernHemisphere winter (DJF) comparable to seminal Figure 7(b) in Wallace JM Gutzler DS 1981 Teleconnections in the geopotential height fieldduring the northern hemisphere winter Mon Weather Rev 109 784ndash812 Figure 1 except for a prolonged analysis period from 1949 to 2009Contourshading starts at a correlation of 05 and increases at 01 intervals Points A B Bb and C refer to the centers of actions of the PNA patternPoints B and Bb are generally associated with the Aleutian Low center of action DJF December January February

92 General Circulation of the Atmosphere j Teleconnections



successively explains the maximal residual variance in theanomaly data set For instance the leading PC is obtained byrequiring that it maximizes the sum of the squared temporalcorrelation between itself and the anomaly time series at allspatial points in the domain The resulting PCs are temporallyorthogonal while the LVs are spatially orthogonal Spatialorthogonality can however be restrictive and in many casesundesirable as discussed earlier Although the leading LV is notdirectly impacted subsequent LVs are often constrained to havepredictable geometric relationships vis-a-vis the leading patterndomain geometry thus becomes an influencing factor itself

For these and other reasons a variant of PCA called therotated principal component analysis (RPCA) has becomepopular since it yields patterns that are no longer constrainedto be spatially orthogonal domain geometry is thus much lessinfluential rotated PCs continue to be temporally orthogonalthough The linear transformation (or solid rotation) of PCsthat is widely used in meteorology is called the lsquovarimaxrsquorotation It is determined by the requirement that the varianceof the squared correlations between each rotated PC and theoriginal time series be maximized Focusing on the variancerather than sum (as in unrotated analysis) of the squaredcorrelations increases spatial discrimination and facilitatesinterpretation of the obtained patterns

Typically only a subset of the leading PCs is rotatedAlthough several criteria exist to guide the choice of this subsetthe sensitivity of results to the rotated number offers goodpractical guidance In most meteorological applications 8ndash10of the leading PCs are rotated in most applications reportedhere the number is eight

A more technical discussion of RPCA can be found inBarnston and Livezy where RPCA is advanced as a preeminentmethod for defining andmonitoring of teleconnection patterns

Empirical Orthogonal Teleconnection

An alternate method for identifying the teleconnection patternsis empirical orthogonal teleconnection (EOT) analysis devel-oped by Van den Dool et al (2000) EOT analysis is conductedby finding a base point in the domain that explains the mostvariance at all the other points combined using linear regres-sions This leading spatial pattern is defined from the linearregression coefficients of this base point while its time series issimply the raw time series at the base point Subsequentpatterns are obtained by repeating this procedure on a reduceddata set from which the first pattern has been linearly removed

The analysis results in a series of spatial patterns or EOTs thatare not constrained to be spatially orthogonal but whose timeseries are independent In this way it is similar to RPCA Otherthan the inherent simplicity EOT analysis holds another advan-tage over RPCA InRPCA only a small subset of PCs are subject torotation no such truncation is required in EOT analysis

Northern Winter Teleconnections from RPCA

The leading patterns of recurrent height variability are extractedfrom RPCA and shown in Figure 2 The DJF anomalies during1979ndash2008 winters were analyzed in the 30 Sndash90 N domainAnalysis was conducted at the 200-hPa level in order to also

capture the tropicalndashextratropical interactions (eg ENSOrelated) that are prominently manifest in the upper tropo-sphere for reasons stated earlier Height anomalies weremultiplied by the square root of the cosine of latitude to ach-ieve grid-area parity which prevents polar regions with manymore points on a regular latitudendashlongitude grid from undulyinfluencing the analysis The covariance rather than correla-tion matrix was analyzed so that regions with large variancecan exert greater control on the analysis outcome The eightleading PCs were rotated using the varimax criterion The PCsare normalized and dimensionless and the spatial patterns(LVs) show nothing but the regression of the PCs onto the200-mb height field

The leading pattern explains 15 of the monthly variancein the domain with two centers of action in the Atlantic sectorNote the spatial similarity to the Atlantic Basin teleconnectivitystructure in Figure 1 sans its earlier noted eastward shift Thisleading pattern is clearly the NAO Its structure is consistent withits historical characterization as a northndashsouth dipole in SLPthat represents out-of-phase fluctuations of the Icelandic Lowand Azores High For example Hurrellrsquos monthly NAO index isconstructed from the difference of normalized SLP anomalies atPonta Delgada Azores and Stykkisholmur Iceland the NAOpattern depicted in Figure 2 is thus in its positive phase whichstrengthens the regional winter circulation features The identi-fication of NAO as the leading pattern in the 200hPa analysis isimportant both for explanation of regional variance and forshowing NAO variability to be temporally orthogonal to theother NH variability patterns The NAO pattern is manifest ontimescales ranging from subseasonal to decadal as seen inFigure 3 The NAO pattern is closely linked with meridionalexcursions of the Atlantic jet and related storm track displace-ments northward in the positive phase of the pattern

The second third and fourth leading patterns all havecenters of action in the Pacific sector attesting to RPCArsquos skill inseparating overlapping variability structures The secondleading pattern represents higher geopotential heights in allsectors of the northern subtropics but especially over thecentraleastern Pacific Higher upper-level heights are typicallyassociated with a warmer air column underneath since theatmosphere is in hydrostatic balance The pattern of variabilityis thus associated with a warming of the Tropics such as duringEl Nintildeo winters The presence of a subtropical ridge to thesoutheast of the Hawaiian Islands is indicative of linkage withENSO for deep convection in this tropical Pacific sector andthe related divergent outflow are linked with the developmentof an upper-level anticyclonic circulation in the subtropics Thesubtropical ridge also reflects the southeastward extension ofthe Asian-Pacific jet in El Nintildeo winters The ENSO pattern isalso characterized by an eastndashwest dipole near North Americawith a negative anomaly along and off the west coast anda positive anomaly near the south shore of Hudson Bay

The PC associated with this spatial pattern exhibits inter-annual variability and is correlated (r frac14 06) with the Nintildeo 34index which is the average sea surface temperature (SST)anomaly in the central-eastern equatorial Pacific (5 Sndash5 N170ndash120 W) and operationally used to define the state ofENSO While this pattern which explains 13 of themonthly winter variance is characterized as the ENSOresponse the related PC exhibits some long-term trend as well




It is remarkable that RPCA of upper-tropospheric heights canidentify the ENSO-related height pattern without any referenceto the underlying SST variability Note that teleconnectionanalysis of the 200-hPa height variability in the same periodwas unsuccessful in this regard

The third and fourth leading patterns of Northern winterheight variability each explain 11 of the monthly varianceThe third leading pattern in Figure 2 consists of a northndashsouthdipole in the North Pacific with a downstream height anomalyover eastern North America centered over Hudson Bay Thispattern is referred to as the NPOWP pattern as the related SLPpattern (shown later) closely resembles the NPO a recurrentsubseasonal variability pattern in SLP first identified by Walker

and Bliss in 1932 The lsquoWest Pacificrsquo part of the name is fromWallace and Gutzlerrsquos teleconnection analysis of 500-hPaheights The NPOWP pattern can be viewed as the NAOanalog in the Pacific Basin both patterns describe subseasonalheight fluctuations in the jet-exit regions of the climatologicaljets streams The fluctuations have a meridional dipole struc-ture with a stronger polar center in both cases and linked withmeridional displacements of the jets and related storm tracks Amultivariate description of the NPOWP pattern and its simi-larity with the NAO can be found in Linkin and Nigam

The fourth leading pattern is the PNA defined by thecoherent arcing pattern of height fluctuations of alternatingsigns beginning with the center over Hawaii and followed by

Figure 2 The four leading teleconnection patterns from a rotated EOF analysis conducted on monthly 200-mb wintertime height anomalies (a)NAO (b) ENSO related (c) NPOWP and (d) PNA The domain of the analysis is 30 Sndash90 N Contourshading interval is 10 m the zero contour issuppressed Percentage of explained variance (a) 146 (b) 125 (c) 114 (d) 111 NAO North Atlantic Oscillation ENSO El Nintildeo-Southern Oscilla-tion NPOWP North Pacific OscillationWest Pacific PNA PacificndashNorth America




centers over the tip of the Aleutians northwestern NorthAmerica and Southeast United States In the positive phase(depicted) the pattern consists of positive height anomaliesover northwestern North America In the Pacific sector theheight anomalies represent eastward displacement andmeridional focusing of the Asian-Pacific jet It is interesting tonote the position of the subtropical ridge in the PNA- andENSO-related patterns as this is helpful in pattern identifi-cation The ridge is over the Hawaiian Islands in the formerbut southeastward of Islands in the latter The variabilitytimescales of the patterns are also distinct subseasonal andinterannual respectively (Figure 3) The separation of thePNA- and ENSO-related patterns is important as PNA was

once viewed as the extratropical response of ENSO The PNAand NPOWP patterns are also distinct notwithstanding thebroadly similar structure over the North PacificndashNorthAmerican region closer inspection reveals the two patterns tobe in spatial quadrature in this sector for example theNPOWP trough tracks the North American coastline which isa nodal line in the PNA pattern

Surface Temperature and Precipitation Impacts

Teleconnection patterns are of considerable interest because oftheir impact on surface hydroclimate Significant climateanomalies on subseasonal-to-interannual timescales are often

Figure 3 The principal component time series for each pattern is shown (a) NAO (b) ENSO related (c) NPOWP (d) PNA Note that there are onlyvalues for the DJF season NAO North Atlantic Oscillation ENSO El Nintildeo-Southern Oscillation NPOWP North Pacific OscillationWest Pacific PNAPacificndashNorth America




attributable to a dominant phase of one or more tele-connection patterns Figure 4 shows the SLP surface temper-ature and precipitation footprints of the NAO Thetemperature and precipitation data sets used in this analysis arefrom the Climate Research Unit (CRU) of the University of EastAnglia namely the TS 31 climate database The SLP patternbears striking resemblance to the 200-hPa height pattern(Figure 2(a)) indicating an equivalent barotropic verticalstructure The strongest impacts of the NAO emerge overEurope and North Africa The NAO results in nearly all ofEurope experiencing above-normal temperatures while coolertemperatures prevail from Saharan Africa east-northeastward tothe eastern Mediterranean Middle East and the ArabianPeninsula The temperature pattern at the surface results largelyfrom advection of climatological temperature by the NAOcirculation (ie VNAO$Grad (TCLIM)) for example coolertemperatures over Africa arise from the cold advection by

northeasterlies tracking the eastern flank of the high SLP cellCorrelation coefficients are shown in Figure 4 to indicate thesignificance of the impact although they conceal the fact thattemperature anomalies over Europe are larger than those overAfrica The precipitation pattern likewise shows a wetndashdrydipole with northern Europe experiencing above-normal andthe Iberian Peninsula and southern Europe below-normalprecipitation This impact is not obvious from the geo-potential and SLP patterns unless the impact of NAO on the jetstream and thus storm tracks is considered In the depictedpositive phase NAO results in enhanced zonal wind north ofthe Azores center and reduced zonal wind to the south Thenorthward displaced jet stream and storm track account inpart for the observed precipitation dipole

The NAO influences the surface climate of North America aswell A northndashsouth dipole in temperature is again observedcold in northeastern North America warm to the south both

Figure 4 (a) Temperature (b) precipitation and (c) SLP footprints of the wintertime NAO pattern seen in Figure 2 Correlation between NAO prin-cipal component and CRU temperature and precipitation is contoured at 015 intervals the zero contour is suppressed For SLP the regression coeffi-cient between the principal component and the SLP field from the CFSR is contoured at 1-hPa intervals and the zero contour is suppressed NAONorth Atlantic Oscillation CFSR Climate Forecast System Reanalysis




consistent with the regional SLP signal The precipitationimpacts are less coherent and significant than in Europe exceptalong the west coast of the Labrador Sea and Davis Strait thedeficit here resulting from off-shore moisture advection by theNAO circulation Interestingly there is a significant dryanomaly on the central US West Coast that is associated witha weak ridge over the Gulf of Alaska (Figure 2(a)) This impactfeature is revisited when annular modes are discussed

The surface signatures of the ENSO-related pattern areshown in Figure 5 The SLP signature is notably weak andfocused in the Gulf of Alaska Unlike NAO where regionalthermal advection shaped surface temperature the significantENSO-related warming of the global northern Tropics (at bothlower and upper levels) results from the propagation ofatmospheric waves instigated by ENSO-related diabatic heatingchanges over the tropical Pacific There is also a tendency forwarmer winters across much of North America centered on theUnited StatesndashCanadian border along with northern Europeand East Asia Precipitation impacts are also notable in the

deep Tropics with drying in eastern South America andincreased precipitation over eastern Africa The ENSO-relatedpattern results in a southward shift of precipitation acrossNorth America with increased precipitation along the US WestCoast and southern tier states Equally noteworthy are thepositive precipitation anomalies over the northern Indiansubcontinent including the Himalayas It is important to notethat Figure 5 does not fully capture the canonical ENSO impacton winter hydroclimate (eg as defined from correlations ofthe Nintildeo 34 SST index) but comes close in most regions

The NPOWP pattern results in significant modulation ofthe North American surface climate (Figure 6) The SLP patternis robust and oriented much as the NAOrsquos reflecting similarvariability mechanisms The NPOWP circulation in the lowertroposphere leads to impressive temperature correlations(exceeding 05) over central-eastern North America fromanomalous thermal advection and related meteorologicalfeedbacks see Linkin and Nigam for more informationIntrusion of the sub-Arctic air into the continental interior in

Figure 5 Same as in Figure 3 except for ENSO pattern (a) temperature (b) precipitation (c) sea-level pressure ENSO El Nintildeo-Southern Oscillation




the negative NPOWP phase (weakened Aleutian Low) leads tobelow-normal temperatures especially to the east of theRockies for example The high correlations suggest thatbetween one-quarter and one-half of the monthly temperaturevariance in the core influence region is explainable by NPOWPvariability The precipitation impact is weaker and the spatialpattern less coherent but potentially important in the PacificNorthwest and southern Great Plains Like the NAO the NPOresults in meridional perturbation of the jet stream in its exitregion since this occurs over the wide Pacific Basin the relatedstorm track modification is not manifest in analysis of conti-nental precipitation

The PNA patternrsquos influence on North American surfacehydroclimate (Figure 7) is perhaps most impressive TheSLP footprint in the Bering Sea and Gulf of Alaska is6 hPa larger than of the other Pacific Basin patternsincluding ENSO The temperature impact is characterized bya strong northwestndashsoutheast dipole across North AmericaNot surprisingly the impact is focused on Northwest

Canada and the Southeast United States the third andfourth centers of action of the pattern respectively againwith anomalous thermal advection defining the large-scaleimpact structure if not its amplitude At first glance theprecipitation impact appears similar to the opposite phaseof the ENSO-related pattern (Figure 5(b)) Closer inspec-tion however reveals a subtle shift in the greatest precipi-tation correlation regions The PNA response is defined bylarge correlations over interior eastern North Americacentered on the Tennessee and Ohio Valleys and extendingacross the Great Lakes into Northern Canada A similarlysigned signal is evident in the Pacific Northwest and Alaskaas well The positive-phase PNA precipitation response(Figure 7(b)) thus consists of only deficits across theUnited States The deficit over the eastern United Statesresults from the southward displacement of the Atlantic jetand storm track that are linked to the circulation anomaliesassociated with the third and fourth centers of the PNAheight pattern

Figure 6 Same as in Figure 3 except for the WPNPO pattern (a) temperature (b) precipitation (c) sea-level pressure WPNPO West PacificNorthPacific Oscillation




Stratospheric Signature

Teleconnection patterns are obtained from the analysis ofconnectivity of tropospheric climate variability Understandingthe mechanisms generating the remote response ie tele-connectivity in some cases warrants analysis of the relatedstratospheric links since stratospheric circulation can be aninfluence pathway between distant tropospheric regions Theenabling role of the stratosphere in generating aspects oftropospheric teleconnectivity is getting increasing attention inobservational and modeling analyses Two of the four tropo-spheric teleconnection patterns ndash NAO and NPOWP ndash exhibitimpressive stratospheric links (Figure 8) Not surprisinglyboth involve meridional displacement of the subtropical jetsespecially in the jet-exit sector Such change in the troposphericjet structure (a poleward shift in the positive NAO and NPOWP phase) modulates the refractive index for the meridional-vertical propagation of planetary waves a poleward jet shiftfor instance reduces the orographically forced stationary waveresponse in both the troposphere and stratosphere in dynam-ical models Jet perturbations influence not only wave

propagation but also wave forcing the latter by modulatingflow over orography and regions of strong climatologicalvorticity gradients for instance The dynamical and thermo-dynamical mechanisms linking the troposphere and strato-sphere are of great interest in modern climate studies

Figure 8 shows the regressions of the NAO and NPOWPPCs on geopotential heights at 100 hPa (16 km) 50 hPa(20 km) and 10 hPa (30 km) the NAO ones are in the leftcolumn The NAO footprint in the stratosphere is significantlyannular ie anomalies of one sign in the polar region sur-rounded by opposite-signed anomalies in the midlatitudes Thetropospheric structure is less annular prompting some to viewNAO as the regional expression of the annular mode of SLP theannular modes are discussed in a later section The NAOstratospheric structure can be broadly characterized as consist-ing of out-of-phase variations of the polar vortex andsurrounding regions The NPOWP one on the other handreflects displacement of this vortex into North America Thisdifference between displacement of the vortex and the break-down or splitting of the vortex is analogous to the two distinct

Figure 7 Same as in Figure 3 except for the PNA pattern (a) temperature (b) precipitation (c) sea-level pressure PNA PacificndashNorth America




Figure 8 Stratospheric footprints of the two leading teleconnection patterns with substantial stratospheric signatures (a)ndash(c) show the regressionof the NAO PC onto 10-mb 50-mb and 100-mb normalized geopotential height anomalies respectively The contour interval is 02 standard devia-tions The zero contour is suppressed in all cases Panels (d)ndash(f) show the same except for the WPNPO pattern NAO North Atlantic Oscillation PCprincipal component WPNPO West PacificNorth Pacific Oscillation




types of sudden warming events that occur in the stratosphereIn the case of a split vortex or complete breakdown of the vortexan equivalent barotropic structure is observed as in the case ofthe NAO stratospheric signature The NPO-related vortexdisplacement tends to be more baroclinic in nature at least inthe stratosphere as evidenced by the westward tilt with height

SST Anomalies

The teleconnections described above represent standingmodes of atmospheric circulation variability on subseasonal-to-interannual timescales As such it is of interest to inquire

into the nature of related SST anomalies which in the PacificBasin can provide additional discrimination between theoverlapping height variability patterns Figure 9 shows the SSTcorrelation of the four PCs The NAO SST footprint revealsa characteristic tripole feature in the Atlantic Basin with coolertemperatures underneath strengthened westerlies in the NorthAtlantic SST anomalies associated with the NAO also appear inthe extratropical Pacific that are only slightly weaker than in theAtlantic It is noteworthy that NAO variability is not correlatedwith tropical Pacific SSTs at least contemporaneously duringNorthern winter months Herein lies a great dilemma thatserves as the basis for a significant portion of contemporary

Figure 9 Correlation between the leading wintertime teleconnection patterns (a) NAO (b) ENSO related (c) NPOWP (d) PNA and SSA anomaliesContourshading interval is 01 m while the zero contour is suppressed NAO North Atlantic Oscillation ENSO El Nintildeo-Southern Oscillation NPOWP North Pacific OscillationWest Pacific PNA PacificndashNorth America SSA sea surface temperature




subseasonal-to-interannual climate research ENSO which isslowly varying and more predictable on seasonal timescalesexplains relatively little hydroclimate variance in the Northernextratropics The NAO by contrast explains more surfaceclimate variance but is substantially less predictable

Not surprisingly the SST signature of the ENSO-relatedheight pattern consists of large tropical correlations thatresemble El Nintildeo SSTs in the equatorial Pacific Notablecorrelations are also present in the western Indian Ocean andthe tropical and extratropical Atlantic These include the well-documented ENSO links to warming of the adjacent basinsas well as capturing of more than just ENSO variability by thisPC including possibly the secular trend and multidecadalvariability The NPOWPrsquos SST correlations are not unlikeNAOrsquos Atlantic ones a weak tripolar horseshoe-like structurein the extratropical basin unlike NAO however it exhibitsweak links with other basins

The PNA SST correlations are broadly similar to theNPOrsquos ndash an SST tongue extending from Japan into themidlatitude Pacific surrounded by opposite-signed SSTsalong the North American coast from the Aleutians to BajaPeninsula ndash but somewhat stronger especially in the deepTropics A helpful distinction is the placement of the coastalanomalies the off-coastal maximum in NPOWP vis-agrave-vis thecoastal maximum in PNA The SST correlations of all threePacific height patterns likely contain elements of the ENSOand Pacific decadal SST variability patterns in view of thepotential aliasing of SST variability due to the short analysisperiod (1979ndash2008) and the atmospheric-only nature of thePCA An RPCA analysis of the combined variability of upper-tropospheric geopotential height and SST in a longer recordminimizes such aliasing see Nigam (2003) for moreinformation

The PNA SST correlations extend into the equatorial Pacific(Figure 9(d)) as noted above This pattern was once thoughtto be the extratropical atmospherersquos response to El Nintildeobecause of such correlations despite the dominant intra-seasonal timescales of its PC (or teleconnection index) and themodest correlation (03 04) of the latter with the Nintildeo 34SST index The potential aliasing of SST variability notedabove led to PNArsquos spurious linkage with tropical SSTs ThePNA PC from the combined analysis of 200-hPa geopotentialheight and SST variations is correlated with the Nintildeo 34 SSTindex at only 007

The contemporaneous SST correlations of the recurrent200-hPa height variability patterns raise the interesting ques-tion of cause and effect Are the height patterns forced by theunderlying SST patterns or vice versa The answer to thisquestion is unfortunately complex as it is dependent on boththe basin region and the variability pattern Generallyspeaking SST variations in the deep Tropics (eg ENSOrelated) are influential on the atmosphere on monthly time-scales whereas variations in the middlendashhigh latitude basinsare the response Consider the NAO In the displayed phase(positive) the SLP pattern (Figure 4(c)) strengthens both theIcelandic Low and Azores High along with a northwarddisplacement of the jet stream and storm track This will resultin anomalous westerly wind stress on the ocean northward ofthe climatological jet position leading to increased verticalmixing and latent heat flux and thus cooling of the surface

layer The opposite arguments apply in the region off the UScoast where SSTs are warmer The strengthened Azores Highis likewise associated with stronger trade winds in the tropicalAtlantic leading to more upwelling and mixing off the Africancoast and thus colder SSTs As noted earlier the influencedirection is dependent on the basin region and variabilitytimescales The extratropical oceans can be influential onlonger timescales and even on subseasonal ones through theirpreconditioning role Clearly this is an active research area inclimate science requiring the use of advanced modelingtechniques

Northern Summer Teleconnections from RPCA

The Northern summer stationary waves are considerablyweaker than the winter ones because of the weaker andnorthward displaced jet (and storm track) The summerregime thus consists of monsoonal circulations and onlymodest stationary Rossby wave propagation For thesereasons the winter anomalies have been the target of mostteleconnection analyses The structure of recurrent heightvariability in Northern summer is however still of interest Arotated EOF analysis (RPCA) of June July and August (JJA)200-hPa geopotential height anomalies is conducted ina fashion similar to the wintertime analysis except for theanalysis domain The domain used in the summer analysis isrestricted to 20ndash90 N This change is made to emphasize theextratropical modes inclusion of the Tropics leads to a noisieranalysis

The summer results are displayed in Figure 10 Theleading patterns are perhaps relatable to the winter patternsThe first one is labeled NPOWP due to the presence ofa northndashsouth dipole over the North Pacific near the date-line The second one is referred as the NAO on account of thedipole over the northeastern Atlantic An ENSO-relatedpattern or a PNA look-alike was not identified The thirdleading pattern exhibits higher geopotential heights in theglobal Tropics and a coherent ridge over the western sub-Arctic The first four modes explain less variance than inwinter indicating the more disorganized structure ofsummertime variability Notice that even the patternsresembling NPOWP and NAO are more regionally confinedconsistent with brief remarks on reduced Rossby waveactivity in Northern summer see Folland et al (2009) formore discussion of the summer NAO the most studiedsummer teleconnection

Southern Winter Teleconnections from RPCA

The very different distribution of continents and mountains inthe Southern Hemisphere (SH) results in muted seasonalityand stationary waves The SH winter climatology is domi-nated by a strong polar vortex and thus more zonallysymmetric flow The lack of continents and large mountainranges lead to reduced orographic forcing and landndashseacontrast both resulting in diminished stationary Rossby waveactivity An RPCA analysis of the SH winter height variabilityis briefly discussed here It is identical to the Northern winter




analysis the domain includes the Tropics (30 Nndash90 S) andthe eight leading patterns are rotated using the varimaxcriterion

The four leading patterns are shown in Figure 11 Thefirst one is analogous to the NAO but slightly more annularin nature The larger amount of explained variance(164) relative to the other patterns indicates that SHvariability is dominated by fluctuations of the polar vortexnot surprising given the zonal nature of the SH circulationPatterns 3ndash4 are more classic teleconnections The thirdpattern consists of a strong center in the Bellingshausen Seain the far South Pacific surrounded by three weaker centerswhile the fourth one is dominated by a wave-3 patternaround the pole Patterns 1 3 and 4 were identified insome form by Mo and White (1985) the first analysis of SH

winter variability see this study for more details on SHteleconnections

For reference Figures 12ndash14 show the surface temperatureand precipitation footprints of select winter patterns of SHheight variability in the interest of space only a few prom-inent features are noted below Because of its polar domainthe first pattern has weak temperature and precipitationfootprint over the inhabited continents Pattern 2 (Figure 12)is linked with wide-spread warming of the Tropics and moreprecipitation over Amazonia and southern South Africa anda precipitation dipole over Australia (dry to the east wet inthe north) some of these signals are well documented in theliterature as El Nintildeorsquos SH winter impact Pattern 3rsquos precipi-tation impact is weak but it is linked to a prominent northndashsouth temperature dipole over South America (Figure 13)

Figure 10 Four leading summertime (JJA) teleconnection patterns that emerge from a rotated EOF analysis of 200-mb geopotential height Thedomain of the analysis is 20ndash90 N Only the leading two modes (a) and (b) are clearly identifiable as the NPOWP and NAO respectively Contourshading interval is 8 m the zero contour is suppressed Percentage of explained variance (a) 119 (b) 85 (c) 80 (d) 74 JJA June July AugustEOF empirical orthogonal function NPOWP North Pacific OscillationWest Pacific NAO North Atlantic Oscillation




Although not expected from the height anomaly structuremode 4 exerts substantial impact over Australia (Figure 14)with wet anomalies everywhere except the southeasternregion

Annular Modes

Closely related to teleconnections are the annular modeswhich have received considerable attention in recent yearsWhile teleconnections are the climate links betweengeographically separated regions the annular modes representthe seesaw (ie out-of-phase links) between the poles and thesurrounding midlatitudes annular modes can be viewed asthe teleconnections of the polar region The annular mode inthe NH and SH is referred as the Arctic Oscillation and

Antarctic Oscillation (AO and AAO) respectively The annularpattern is obtained from EOF analysis of the monthly 1000-hPa geopotential height anomalies (700 hPa for SH) pole-ward of 20 for the entire year Key differences from theprevious RPCA analysis are the lack of rotation of the PCs andthe search for recurrent patterns among all calendar month(and not just winter) anomalies Note unrotated analysesoften generate more spatially extended patterns in order tomaximize the explanation of variance as seen in Figure 15which also shows the temperature and precipitation impactsof the annular patterns Note lsquoannular patternrsquo rather thanlsquoannular modersquo is perhaps more reasonable terminology asonly spatial (and not spatiotemporal) variability is analyzedfor their extraction

The annular modes bear close resemblance to the leadingmode extracted from RPCA of 200-hPa heights in each

Figure 11 The leading patterns that emerge from a rotated EOF analysis of 200-mb height in SH wintertime (JJA) The domain used is 30 Nndash90 S The contour interval is 10 m while the zero contour is suppressed Panels (a)ndash(d) represent patterns 1ndash4 Percentage of explained variance(a) 164 (b) 119 (c) 115 (d) 104 JJA June July August EOF empirical orthogonal function SH Southern Hemisphere




hemisphere The AO has a stronger coherent center of actionin the Aleutians in addition to the essential elements of theNAO pattern Not surprisingly AOrsquos year-round temperatureand precipitation impacts are very similar to the winterimpacts of the NAO (cf Figure 4) The close similarity of AOand NAO has led many to question whether the NAO ismerely a region manifestation of the larger-scale annularmode (AO) It is noteworthy that neither the AO nor AAO are

strikingly annular especially the AO The AO footprint in thelower stratosphere (not shown) however appears substan-tially more annular as does the NAO footprint (cfFigure 8(c))

Subseasonal Tropical Variability ndash MJO

The ENSO-related pattern reveals one way in which theTropics can influence the extratropics An interannual shiftin the preferred area of tropical convectionprecipitationleads to anomalous divergent motions which generatea Rossby wave source that in turn alters midlatitude stormtracks and extratropical climate Zonal variations in tropicalconvection occur on timescales other than interannual onesas well The MJO ndash representing convection variability inthe tropical Indian Ocean and Western Pacific basins on 30-to 50-day timescales ndash can like ENSO influence the climatein the extratropics albeit on subseasonal timescales

When the MJO is active convection is enhanced over theeastern Indian Ocean at the expense of the West Pacific Thisgenerates an anomalous divergent circulation that leads toretraction of the East Asian jet The jet retraction and relatedmodulation of storm tracks influence the circulation andhydroclimate over the North Pacific and North Americanregions for example weaker Aleutian Low and higher thannormal heights over eastern North America These MJOinfluences can implicitly influence the NAO and PNAstructure but only modestly as these teleconnectionpatterns are robustly manifest in the MJO-filtered dataas well

Figure 12 Same as in Figure 3 except for the SH pattern 2 This is likely related to ENSO ENSO El Nintildeo-Southern Oscillation SH SouthernHemisphere

Figure 13 Similar to Figure 12 except only showing the SouthAmerican temperature footprint for pattern 3




Subseasonal Teleconnection Analysis

Climate teleconnections discussed above were obtainedfrom the analysis of monthly averaged anomalies usingtechniques keyed to an optimal accounting of the temporalvariance These attributes steered the analysis towardextraction of the mature phase of the leading teleconnectionpatterns Monthly averaging filtered the large-scale synopticwaves in the troposphere and facilitated focus on thelsquostandingrsquo component of low-frequency variability A monthis not too long a period in context of large-scale oceanndashatmosphere interaction but it is an extended averagingperiod in context of the dynamical and thermodynamicalprocesses generating climate teleconnections Theseprocesses evolve rapidly ndash on pentad-to-weekly timescales ndash

and their resolution is critical for the description of thenascent-to-mature and decay phases of the teleconnectionpatterns and for advancing understanding of the excitationmechanisms Insights into the evolution of teleconnection

patterns can advance the subseasonal prediction of theclimate system

In the following the submonthly evolution of PNAvariability is briefly discussed to illustrate an interesting lineof contemporary teleconnection research The PNA patternis chosen as its evolution was described earlier froma simpler (and less accurate) analysis (Nigam 2003)facilitating assessment of the new analysis strategy TheCFSR used earlier at monthly resolution is analyzed atpentad (5-day average) resolution in this section Theanalysis period covers 30 extended winter (NovemberndashMarch) seasons from 1979 to 2008 and the 200-hPa geo-potential height anomalies are analyzed as before Toemphasize pattern development and decay a techniquecalled extended EOF analysis is used Extended EOF analysisis a powerful spatiotemporal analysis technique but onethat can be easily implemented using the standard EOFanalysis infrastructure by substituting a contiguous series(eg from T 2 to T thorn 2 time steps) of anomaly patterns isplace of the contemporaneous one (ie T frac14 0) The leadingvariability mode in this case is not a single spatial patternbut the most recurrent temporal series of spatial patterns orthe most recurrent spatiotemporal pattern In this illustra-tive example a 5-pentad window is chosen (T 2 to T thorn 2pentads) The seven leading modes are rotated using thevarimax criterion to allow for more spatial discrimination

The leading mode from this spatiotemporal analysisrepresents the development and decay of the NAO (notshown) The fourth leading mode is closely related to the PNAa 5-pentad evolution is shown in Figure 16 An interestingfinding here is that the development and decay of the PNA isassociated with westward progression of height anomaliesacross the high latitudes At T frac14 3 there is a notable heightcenter in the NAO region of the North Atlantic It is thus notsurprising that there is a small but significant lag correlationbetween modes one and four (r frac14 035) with the negativeNAO leading the positive-phase PNA pattern The retrogressionof the northern height anomalies is consistent with theadvection of planetary vorticity but cannot adequately explainthe development of the robust anomaly center south of theAleutians

To help set forth a plausible basis for the NAOndashPNAlink the impact of the NAO on the meridional wind isassessed meridional winds highlight the shorter zonalscales of the midlatitude quasi-geostrophic flow and thusthe zonal propagation of stationary Rossby wavesFigure 17(a) shows contemporaneous regressions of thePC associated with the positive NAO pattern onto the200-hPa meridional wind along with the climatologicalwinter position of the East Asian jet The NAO apparentlyforces a wave pattern in the subtropics between East Africaand southern Asia that culminates in the North PacificFigure 17(b) shows the 2-pentad lag regression of thesame NAO PC onto the 200-hPa zonal winds The resultis a subtle retraction of the East Asian jet consistent withthe negative-phase PNA pattern

The analysis presented above serves to illustrate the natureof contemporary research in climate teleconnections ratherthan definitive findings on PNA evolution durable andcorroborated findings on teleconnection evolution have yet to

Figure 14 Similar to Figure 12 except the temperature (a) andprecipitation (b) footprints are shown only over Australia for pattern 4




Figure 15 The two annular modes (a) the Arctic Oscillation and (b) the Antarctic Oscillation (AAO) are shown along with their temperature andprecipitation footprints Annular modes are calculated as the leading EOF of 1000-mb geopotential height (700-mb geopotential height for AAO) Notethat the entire year is used to calculate the pattern Contour interval for the height patterns is 5 m Hydroclimate footprints are shown as correlationsbetween the principal components and CRU temperature and precipitation contoured at 015 intervals For all panels the zero contour is suppressedEOF empirical orthogonal function




Figure 16 The fourth leading mode from a rotated extended EOF analysis on extended winter 200-hPa height at pentad resolution This modeclosely resembles the PNA The figure is presented as the leadlag regression (T frac14 4 to T frac14 0) of the PC onto the height field The contour intervalis 10 m and the zero contour is suppressed EOF empirical orthogonal function PNA PacificndashNorth America PC principal component




emerge The recent focus on submonthly climate predictionhas however spurred climate scientists to investigate thedevelopment and decay of the leading winter teleconnectionpatterns

See also Climate and Climate Change Climate VariabilitySeasonal and Interannual Variability Dynamical MeteorologyRossby Waves Stationary Waves (Orographic and ThermallyForced) StratosphereTroposphere Exchange and StructureGlobal Aspects Tropical Meteorology and ClimateMaddenndashJulian Oscillation Skeleton and Conceptual Models

Further Reading

Barnston AG Livezey RE 1987 Classification seasonality and persistence of low-frequency atmospheric circulation patterns Monthly Weather Review 1151083ndash1126

Folland CK Knight J Linderholm HW Fereday D Ineson S Hurrell JW2009 The summer North Atlantic Oscillation past present and future Journal ofClimate 22 1082ndash1103

Mo KC White GH 1985 Teleconnections in the southern hemisphere MonthlyWeather Review 113 22ndash37

Nigam S 2003 Teleconnections In Holton JR Pyle JA Curry JA (Eds)Encyclopedia of Atmospheric Sciences Academic Press Elsevier Science Londonpp 2243ndash2269

van den Dool HM Saha S Johansson Aring 2000 Empirical orthogonal tele-connections Journal of Climate 13 1421ndash1435

Wallace JM Gutzler DS 1981 Teleconnections in the geopotential height fieldduring the northern hemisphere winter Monthly Weather Review 109 784ndash812

Figure 17 (a) The contemporaneous regression of the leading PC (NAO) onto 200-hPa meridional wind is contouredshaded at 05 m s1 intervalsThe zero contour is suppressed Bold red contours show the winter climatology of the 200-hPa zonal wind at 10 m s1 intervals (b) The 2-pentadlag regression of the leading PC (NAO) onto the 200-hPa zonal wind field is contouredshaded at 1 m s1 intervals with the zero contour suppressedAs in (a) bold red contours represent the winter jet climatology NAO North Atlantic Oscillation PC principal component




TeleconnectionsS Nigam and S Baxter University of Maryland College Park MD USA

2015 Elsevier Ltd All rights reservedThis article is a revision of the previous edition article by S Nigam volume 6 pp 2243ndash2269 2003 Elsevier Ltd

Synopsis

Teleconnections refer to the climate variability links between non-contiguous geographic regions Teleconnection patterns areextracted from analysis of the sea-leveltropospheric pressure variations on monthly (and weekly) timescales The methodsused in extraction of teleconnection patterns are discussed and applied to recent period data Teleconnections are especiallywell-developed in Northern winter when they strongly influence subseasonal variability notably in surface temperature andprecipitation the stratospheric and SST links are also noted The patterns in boreal summer are shown as are SouthernHemisphere teleconnections Aspects of tropicalndashextratropical teleconnections are discussed along with the relationshipbetween annular modes and teleconnections Finally a teleconnection analysis with pentad (5-day averaged) data is pre-sented to initiate discussion of teleconnection evolution

Introduction

The term teleconnection is often used in atmospheric sciences todescribe the climate links between geographically separatedregions The remote region need not exhibit fluctuations of thesame sign in order to be lsquoteleconnectedrsquo In fact the interestingteleconnections often involve contemporaneous variations ofopposite signs Climate analysis is facilitated by the constructionof a teleconnection map which describes the linkage betweena region of interest (a base point) and all other points in thedomain that are farther than the decorrelation length scale of thevariable Teleconnection maps thus provide information aboutthe structure of recurrent climate variability especially itscorrelation-at-a-distance features The maps are useful becauseclimate variability is often manifest with such structure forexample winter variations in temperature and rainfall oversouthern Europe and the Iberian Peninsula are frequentlyopposite to those over northwestern Europe and Scandinavia

Teleconnection maps were first constructed for meteoro-logical parameters measured at the Earthrsquos surface such as theatmospheric pressure The selection of the base point is a criticalfirst step and was historically guided by the investigatorrsquosinsights and interests Today base points can be selected moreobjectively and the robustness of teleconnection maps can beascertained by independent analyses The statistical correlationof the fluctuations provides a measure of the teleconnectionstrength The structure and strength of the teleconnectionpatterns change with season altitude choice of variable andeven temporal averaging of data There are interesting differ-ences between the hemispheres too in part due to the presenceof extensive continents andmountainous zones in the NorthernHemisphere (NH) Teleconnectivity in the NH winter circula-tion has been extensively analyzed as teleconnections accountfor a significant portion of the winter variance in this region

Climate teleconnections are present in observations thathave been averaged in time over a period that is long enough tosuppress the day-to-day weather fluctuations but short enoughto retain the seasonal-to-interannual component of climatevariability Although teleconnection patterns often evolve onsubmonthly timescales (weeks) their spatial patterns arecharacterized in monthly and seasonal data Monthly averages

are typically used since a month is longer than the period ofmost large-scale synoptic waves in the troposphere Climateteleconnections thus highlight the lsquostandingrsquo component oflow-frequency variability ndash one with geographically fixed nodesand antinodes The connectivity of remote regions manifest inthe teleconnection map indicates the potential significance ofremote forcing in the generation of regional climate anomaliesA teleconnection map based on contemporaneous correlationshowever cannot by itself discriminate between the forcing andresponse regions

Although the structure of the prominent teleconnectionpatterns has been known for some time the reasons for theirorigin are not yet well understood For example themechanismsthat excite and sustain the North Atlantic Oscillation (NAO) ndashone of the notable and earliest discovered patterns ndash are stillbeing investigated In the context of such investigations it hasbeen questioned if the teleconnection patterns that are typicallyregional (eg NAO) robustly portray the spatial structure ofvariability from the viewpoint of elucidation of the underlyingdynamical processes

Such concerns are relevant since the canonical tele-connection patterns typically represent the mature phase ofvariability The mature-phase pattern however need notresemble the nascent-phase structure which may be morerevealing of the excitation mechanism Identification of theevolution process from analysis of the mature-phase structureis thus difficult The spatial imprint of variability captured bya teleconnection pattern can also be ineffective in revealing theunderlying mechanisms if the region in question is the locus oftwo temporally independent physical andor dynamicalprocesses While cautionary these remarks do not call fora radically new analysis paradigm Instead they point to theneed for more comprehensive analysis of variability particu-larly in the spatiotemporal domain to facilitate insights intothe evolutionary process While in-depth analysis of this kind isbeyond the scope of this article it is important to let the readerknow that contemporary teleconnection research is rooted inspatiotemporal analyses

Additionally while teleconnectivity has historically beenidentified from assessment of the correlation between somedefined base point and distant points in the domain not all

90 Encyclopedia of Atmospheric Sciences 2nd Edition Volume 3 httpdxdoiorg101016B978-0-12-382225-300400-X




Analysis Method


Correlation




Hij frac14PN

kfrac141f0ikf

0jkPN

kfrac141f02ik

1=2PNkfrac141f

02jk



Teleconnection Map








f0Ak

SA f0

Bk

SBthorn f0

Ck

SC






















































































SST Anomalies






























Annular Modes








































Further Reading












Analysis Method


Correlation




Hij frac14PN

kfrac141f0ikf

0jkPN

kfrac141f02ik

1=2PNkfrac141f

02jk



Teleconnection Map








f0Ak

SA f0

Bk

SBthorn f0

Ck

SC






















































































SST Anomalies






























Annular Modes








































Further Reading



























































































SST Anomalies






























Annular Modes








































Further Reading














































































SST Anomalies






























Annular Modes








































Further Reading






























































SST Anomalies






























Annular Modes








































Further Reading






















































SST Anomalies






























Annular Modes








































Further Reading














































SST Anomalies






























Annular Modes








































Further Reading







































SST Anomalies






























Annular Modes








































Further Reading































SST Anomalies






























Annular Modes








































Further Reading
























SST Anomalies






























Annular Modes








































Further Reading
















SST Anomalies






























Annular Modes








































Further Reading












SST Anomalies






























Annular Modes








































Further Reading


































Annular Modes








































Further Reading




















Annular Modes








































Further Reading












Annular Modes








































Further Reading











































Further Reading

































Further Reading





















Further Reading

















Further Reading













Further Reading











Documents

Author's personal copy - Atmospheric and Oceanic Sciencenigam/Nigam-Baxter... · Encyclopedia of Atmospheric Sciences, Second Edition, 2015, 90–109 teleconnection patterns are uncovered