Sea surface temperature in the north tropical Atlantic as a trigger for El Niño/Southern Oscillation events
Yoo-Geun Ham 1,2,Jong-Seong Kug 3*,Jong-Yeon Park 3and Fei-Fei Jin 4
El Niño events,the warm pha of the El Niño/Southern Oscillation (ENSO),are known to affect other tropical ocean basins through teleconnections.Converly,mounting evi-dence suggests that temperatu
re variability in the Atlantic Ocean may also influence ENSO variability 1–5.Here we u reanalysis data and general circulation models to show that a surface temperature anomalies in the north tropical At-lantic during the boreal spring can rve as a trigger for ENSO events.We identify a subtropical teleconnection in which spring warming in the north tropical Atlantic can induce a low-level cyclonic atmospheric flow over the eastern Pacific Ocean that in turn produces a low-level anticyclonic flow over the western Pacific during the following months.This flow generates easterly winds over the western equatorial Pacific that cool the equatorial Pacific and may trigger a La Niña event the following winter.In addition,El Niño events led by cold anomalies in the north tropical Atlantic tend to be warm-pool El Niño events,with a centre of action located in the central Pacific 6,7,rather than canonical El Niño events.We suggest that the identification of temperature anomalies in the north tropical Atlantic could help to forecast the development of different types of El Niño event.
The anomalous a surface warming in the tropical Pacific during the El Niño is known to produce robust teleconnections over the tropical ocean basins.It is generally agreed that the ENSO over the Pacific domain influences Atlantic climate variability with a few months’lag 8–11.In particular,this remote ENSO impact on the tropical Atlantic is robust in the north tropical Atlantic (NTA)over the northern subtropics.On the other hand,ENSO is not statistically relevant to the simultaneous equatorial Atlantic variability,becau the atmospheric bridge mechanism is cancelled by the opposite influence of oceanic process 12.
Significant a surface temperature (SST)anomalies over the NTA region are known to have broad impacts on climate variability over the US and Africa 13–15,and even interact remotely with North Atlantic climate 16,17.In addition,recent studies suggested that the Atlantic Ocean variability can influence the ENSO variability and its predictability 1–5.In particular,some studies argued that the equatorial Atlantic zonal mode (that is,Atlantic Niño)in boreal summer can affect ENSO development by modulating the Walker Circulation 4,5.Thus,it is also possible that the NTA SST may modulate the Pacific climate variability.In particular,the NTA SST can play a role in triggering ENSO events becau the NTA SST has strong variability during boreal spring when the ont of most ENSO events occurs.
1Global
Modeling and Assimilation Office,NASA/GSFC,Greenbelt,Maryland 20771,USA,2Goddard Earth Sciences T echnology and Rearch Studies and
Investigations,Universities Space Rearch Association,Columbia,Maryland 21044,USA,3Korea Institute of Ocean Science and T echnology,Ansan 426-744,Korea,4Department of Meteorology,University of Hawaii,Honolulu 96822,USA.*e-mail:jskug@kiost.ac
.
¬9¬7¬5¬33579
Pr (mm d ¬1)
40°a
b
c
ParReg ((NTA) – NINO3.4)
20°020°40°20°020°40°20°020°Figure 1|Regression with respect to NTA SST.a –c ,Lagged regressions between NTA SST (90◦W–20◦E,0◦–15◦N)averaged during February to April (the FMA ason)and SST,wind vector at 850hpa (vector)and precipitation during the MAM (a ),JJA (b )and SON (c )asons,after
excluding the impact of NINO3.4SST (170◦–120◦W,5◦S–5◦N)during the previous DJF ason.Only the values at the 95%confidence level or higher are shown.
To isolate the impact of NTA SST on the Pacific climate variability,lagged regressions are calculated with respect to the NTA SST (90◦W–20◦E,0◦–15◦N)during February–April (the FMA ason).As it is known that the NTA SST is influenced in part by
a
b
C a l e n d a r m o n t h
987
6C a l e n d a r m o n t h
543
6 2 ¬1Off-equatorial (Pr, Strm850, SST)Equatorial (gh250–gh850, Pr, SST)
¬9¬7¬5¬3
3579
Pr (mm d ¬1)魄力的近义词
团体游戏室内
Pr (mm d ¬1)
Figure 2|Two pathways for relaying Atlantic signals to the Pacific.a ,b ,Lagged regressions of 3-month-averaged subtropical precipitation,stream
function at 850hPa and SST (a ),and equatorial (10◦S–10◦N)thickness (that is,geopotential height at 250hpa minus geopotential height at 850hpa),precipitation and SST (b )from March to September.Note that the subtropical region of precipitation is 5◦–15◦N,whereas the region for other variables is 5◦–20◦N.Only the values with at least 95%confidence level are shown.
the ENSO variability 10,11,the ENSO effect is first removed using the linear regression with respect to the NINO3.4SST (170◦–120◦W,5◦S–5◦N)during the previous ason (December–February,DJF)and then the lagged regression is obtained.
During boreal spring (the MAM ason;Fig.1a),there is a significant positive SST over the NTA region.This SST anomaly enhances the convective activity,especially over the equatorial Atlantic where the Atlantic Intertropical Convergence Zone (ITCZ)is located.The associated diabatic heating gives ri to a low-level cyclonic flow over the subtropical eastern Pacific as a Gill-type Rossby-wave respon 18,which produces a northerly flow on its west flank.The northerly flow leads to surface cooling through the enhanced wind speed 19and cold/dry advection from higher latitudes 20,resulting in an anomalous sinking motion.The associated suppression in convection is clear over the off-equatorial region in the vicinity of the ITCZ,where climatological convective activity is strong.
During boreal summer (the JJA ason),the negative precip-itation anomaly over the subtropical Pacific is slightly shifted to the north owing to the northward migration of the ITCZ.At the same time,the negative precipitation anomalies induce a low-level anticyclonic flow anomaly,generating a pair of circulation anoma-lies along the Pacific ITCZ.This anticyclonic flow enhances the northerly flow at its eastern edge,which reinforces the negative precipitation anomaly.This strong coupling among negative pre-cipitation,the northerly flow and negative SST anomalies leads to a slow westward extension of the negative precipitation anomaly and the anticyclonic flow,which amplifies and relays the influence of the Atlantic to the Pacific.This subtropical anticyclonic flow is located over the western North Pacific,so it induces equatorial easterly anomalies over the western Pacific.This equatorial western Pacific easterly flow acts as an efficient trigger for a cold event over the equatorial Pacific by shoaling the equatorial thermocline and inducing westward zonal current anomalies 21.
The subtropical variability in the Pacific is known to be capable of exciting an equatorial ENSO signal 22,and our results demonstrate that the NTA SST is partly responsible for this subtropical signal.It is worth pointing out that there is a region of
anomalous westerly flow over the far-eastern Pacific,which is part of the NTA-forced cyclonic flow.It
prevents the eastern Pacific SST from being cooled by suppressing local upwelling,thus allowing the SST anomaly to develop mainly over the central Pacific.Once the negative SST is generated over the equatorial central Pacific,it is intensified through air–a coupled process and leads to strong cooling during the subquent ason (Fig.1c).During this ason,the robust equatorial cold SST signal locks the negative precipitation maximum over the Equator,whereas the subtropical signals are much weakened as the ITCZ migrates southwards.
So far,it has been shown that the NTA SST warming (cooling)can initiate equatorial Pacific cooling (warming)through a pair of low-level circulation respons along the Pacific ITCZ.Figure 2clearly illustrates the quential evolution of this atmospheric teleconnection.Once the positive NTA SST is developed in the FMA ason,the enhanced local convection generates a cyclonic flow over the eastern Pacific as a Gill-type Rossby-wave respon 18.Then,the anomalous convective activity,due to the anomalous northerly flow at the western part of this cyclonic flow,produces an anticyclonic flow over the western–central Pacific in the AMJ ason.As this anticyclonic flow is gradually intensified through air–a feedback,it leads to an equatorial easterly flow over the western Pacific that initiates surface cooling over the equatorial central Pacific.This westward relay of the Atlantic signal takes place along the Pacific ITCZ through the strong coupling with the wind-driven convection anomalies.
In addition to this subtropical teleconnection,the NTA SST can modulate western Pacific wind through eastward-propagating signals.During March,there is a quick respon of the atmospheric thickness anomaly from the Atlantic to the eastern Indian Ocean,which might be explained by the warm Kelvin-wave respon to the equatorial diabatic heating (Fig.2b).This positive thickness signal over the eastern Indian Ocean propagates eastwards slowly,possibly owing to a coupling with the convection.This warm thickness signal is linked to a low-level low-pressure anomaly,leading to an equatorial low-level easterly flow in the region of negative zonal pressure gradient.This easterly flow induces low-level convergence over the maritime continent and
MAM
SON ParReg((NTA)–NINO3.4)
Longitude Longitude红烧牛尾的做法
Longitude
D e p t h (m )
D e p t h (m )
D e p t h (m )
a
b
c
新西兰工签
Figure 3|Evolution of the oceanic variables with respect to NTA SST.
a –c ,Lagged regressions of equatorial (2◦S–2◦N)subsurface temperature,oceanic zonal currents and vertical velocities during the MAM (a ),JJA (
b )and SON (
c )asons,after excluding the impact of NINO3.4SST
(170◦–120◦W,5◦S–5◦N)during the previous DJF ason.Only the values at the 95%confidence level or higher are shown.The red and blue dots at the surface layer are marked when there are positive and negative 850-hPa zonal wind anomalies at or above the 95%confidence level,respectively.
a resultant positive precipitation anomaly in the AMJ ason.In turn,the positive precipitation anomaly enhances the anomalous equatorial easterly flow.
The anomalous easterlies,induced by NTA warming,give ri to significant oceanic respons along the Equator (Fig.3).During the MAM ason,there is a weak subsurface temperature anomaly in the central Pacific related to the equatorial easterly flow.The easterly forcing also induces anomalous westward currents with the upwelling over the central Pacific through the upwelling Kelvin waves.However,the upwelling Kelvin-wave signals are not extended to the eastern Pacific,and there are even weak downwelling signals due to the local westerlies (red dots in Fig.3a)as a part of the cyclonic flow (Fig.1a).Therefore,the subsurface temperature anomalies cannot develop further and conquently the surface cooling is suppresd in the eastern Pacific.
During boreal summer and autumn,the westward current anomaly is extended to the east with the extension of equatorial easterly forcing,thereby the surface cooling develops further over the central Pacific.On the other hand,in the eastern Pacific,the surface cooling is not significant owing to competing process between local westerly flow and remote easterly forcing.Therefore,the SST anomalies develop mainly over the central Pacific,rather than over the eastern Pacific.It is also noted that the temperature anomaly over the surface layer is co-located with the westward current anomaly,whereas the location of the surface anomaly is shifted to the west of the subsurface temperature anomaly,implying that the negative SST over the central Pacific is mainly driven by zonal advection 7.
It is shown that the NTA warming in boreal spring can trigger a cold event in the subquent winter by inducing western Pacific easterlies through both equatorial and off-equatorial pathways.As some ENSO events accompany the NTA SST events 10,11,our study also suggests that there is two-way feedback between ENSO and NTA SST,which leads to fast pha transition of ENSO (ref.1and Supplementary Fig.S9).That is,when the El Niño events accompany the NTA warming,the NTA warming can lead to fast termination of El Niño and possibly rapid development of La Niña in the subquent winter.This implies that the NTA SST plays a critical role in modulating ENSO evolution such as its ont,and pha transition.
Even though the obrvational analysis clearly emphasizes the role of NTA SST on the ENSO,it has limitations due to the
relatively
SST U (°C )
PDF (Exp CTRL , Exp NTA )
a
b
Figure 4|The impact of the NTA SST in the idealized experiment using a CGCM.a ,Enmble-mean difference in the 850-hPa zonal wind over the
western Pacific (120◦–160◦E,5◦S–5◦N)and NINO3.4SST from January to December between two CGCM simulations with and without NTA warming.Error bars indicate the 95%confidence intervals.b ,Probability distribution of NINO3.4SST in December using 60enmble simulations with the
climatological SST (black bar)or with the NTA warming (red bar).Note that the moderate (strong)ca is defined as the magnitude of NINO3.4SST being between one and two (two and three)standard deviations of the 60enmble members.
short records with a mixture of various climate variabilities. Therefore,this obrved role of the NTA SST is substantiated by idealized experiments using a coupled general circulation model (CGCM).To isolate the impact of NTA warming,two CGCM experiments are carried out with and without positive NTA SST anomalies(Supplementary Fig.S3).The differences between the two experiments are clearly due to the NTA SST,becau the only difference in the experimental designs is the existence of the NTA SST anomaly.
The model results confirm that the NTA warming leads to the development of La Niña during the subquent winter ason (Fig.4a).It is clear that the equatorial easterly anomalies are significant over the western Pacific during boreal spring,even though the equatorial Pacific SST signals are positive.This indicates that this anomalous easterly flow is induced by the NTA SST rather than by the Pacific SST.Owing to the equatorial anomalous easterlies,the La Niña signal begins to show up from boreal summer,and further develops during subquent asons.Note that the anomalous easterlies are weakened during boreal summer as the impact of the NTA warming is weakened,but they are reinforced during SON as the La Niña signal becomes stronger.
As in the process,the model tends to predict that the La Niña state is more probable in60enmble runs under the NTA warming(Fig.4b).The probability of La Niña is incread1.9times i
n the simulations with the NTA warming,and the probability of having a strong La Niña is even higher(×4).On the other hand,the probability of El Niño is reduced to52%with the NTA warming, which clearly shows that the NTA warming during spring time favours La Niña signals nine months later.
The key distinctive role of NTA SST is characterized by the opposite low-level wind respons between the western and eastern Pacific.In the ca of NTA cooling,there is a region of equatorial westerly flow over the western Pacific,whereas the equatorial easterly flow is over the eastern Pacific.Therefore,surface warming prevails in the central Pacific,whereas the eastern Pacific SST does not,owing to the opposite effect of the anomalous local easterly flow.This suggests that the NTA cooling favours the development of an El Niño event who action centre is over the central Pacific,rather than the eastern Pacific.In other words,the NTA cooling tends to be cloly related to the development of the warm-pool El Niño6,7,23,24(Supplementary Fig.S8for model results),which has occurred frequently in the recent decades25. Interestingly,the triggering role of NTA SST ems to be much stronger in the recent decades when the warm-pool El Niño events are frequently obrved.With the increasing scientific and socioeconomic interest in the warm-pool El Niño due to its distinctive global impacts6,25,a better understanding of NTA SST’s role may provide a unique,and more reliable predictor for the different types of El Niño event.
Methods
In Figs1–3,lagged regressions onto NTA SST during the FMA ason are applied, after removing the regresd fields,onto NINO3.4during the previous DJF ason. This is equivalent to the partial regression method26,which is often ud to calculate the linearly fitted equation between two variables after excluding the linear impact of the third variable.As it captures only a linear relationship,this regression analysis has a limitation,namely,it does not reflect the asymmetry between the NTA warming and cooling events.The significance test ud in
this study is a two-tailed t-test bad on the temporal standard deviation for which the distribution of the test statistics under the null hypothesis can be approximated by a t-distribution.
The monthly mean wind,geopotential height and precipitation from1980to 2010are obtained from the Modern-Era Retrospective Analysis for Rearch and Applications(MERRA,disc.sci.v/daac-bin/DataHoldings.pl)27. The obrved SST data during1980–2010are from the improved
Extended Reconstructed Sea Surface Temperature version2(ERSST V.2, aa.gov/oa/climate/rearch/sst/sst.html)from the National Climate Data Center of
the US(ref.28).In addition,the monthly mean subsurface temperature,oceanic current and vertical velocity from1980to2010are obtained from the NCEP Global Ocean Data Assimilation System(GODAS, aa.gov/psd/data/das.html)29.All of the data are de-trended first.
The GFDL CM2.1model is ud for the idealized CGCM experiment,which is developed by the Geophysical Fluid Dynamics Laboratory30.To isolate the impact of NTA warming,two CGCM experiments are carried out by prescribing different SST fields;one is using warm NTA SST anomalies added to the obrved climatological filed(Exp NTA,e Supplementary Fig.S3for the prescribed SST anomalies),and the other is the control experiment of prescribing the obrved climatological SST(Exp CTRL).The SST fields are prescribed only over the Atlantic Ocean(5◦S–25◦N)in both NTA and CTRL experiments,whereas the model is fully coupled over the other regions.For the prescribed SST of the NTA warming ca, the obrved SST anomalies from the NTA developing ason(that is,January)to the subquent winter ason(that is,December)are obtained using lag composite analysis when positive NTA SST during the FMA ason is larger than one standard deviation(Supplementary Fig.S1),and then the composited SST anomalies are added to the climatological SST.A total of60enmble members with different initial conditions are co
nducted,which are ud for both the Exp NTA and the Exp CTRL.The t-test to determine the significance of enmble-mean differences between two experiments is bad on the standard deviation of differences in two experiments of each enmble member.
Received8May2012;accepted27November2012;
published online6January2013;corrected online17January2013 References
1.Dommenget,D.,Semenov,V.&Latif,M.Impacts of the tropical Indian and
Atlantic Oceans on ENSO.Geophys.Res.Lett.33,L11701(2006).
2.Jann,M.F.,Dommenget,D.&Keenlyside,N.S.Tropical atmosphere–ocean
interactions in a conceptual framework.J.Clim.22,550–567(2009).
3.Frauen,C.&Dommenget,D.Influences of the tropical Indian and
Atlantic Oceans on the predictability of ENSO.Geophys.Res.Lett.39,
L02706(2012).
4. al.Are Atlantic Niños enhancing Pacific ENSO events
in recent decades?Geophys.Res.Lett.36,L20705.(2009).
5.Ding,H.,Keenlyside,N.S.&Latif,M.Impact of the Equatorial Atlantic on the
El Niño Southern oscillation.Clim.Dynam.38,1965–1972(2012).
6.Ashok,K.,Behera,S.K.,Rao,S.A.,Weng,H.&Yamagata,T.El Niño Modoki
and its possible teleconnection.J.Geophy.Res.112,C11007(2007).
三八节礼物7.Kug,J-S.,Jin,F-F.&An,S-I.Two types of El Niño events:Cold tongue El Niño
and warm pool El Niño.J.Clim.22,1499–1515(2009).
8.Enfield,D.B.&Mayer,D.A.Tropical Atlantic a surface temperature
variability and its relation to El Niño Southern Oscillation.J.Geophys.Res.102, 929–945(1997).
9.Saravanan,R.&Chang,P.Interaction between tropical Atlantic variability and
El Niño-Southern Oscillation.J.Clim.13,2177–2194(2000).
10.Alexander,M.A.&Scott,J.D.The influence of ENSO on air–a interaction
in the Atlantic.Geophys.Res.Lett.29,1701(2002).
11.Chiang,J.C.H.&Sobel,A.H.Tropical tropospheric temperature variations
caud by ENSO and their influence on the remote tropical climate.J.Clim.
15,2616–2631(2002).
12.Chang,P.,Fang,Y.,Saravanan,R.,Ji,L.&Seidel,H.The cau of the fragile
relationship between the Pacific El Niño and the Atlantic Niño.Nature443, 324–328(2006).
13.Enfield,D.B.Relationships of inter-American rainfall to tropical Atlantic and
Pacific SST variability.Geophys.Res.Lett.23,3305–3308(1996).
14.Chang,P.,Saravanan,R.,Ji,L.&Hegerl,G.C.The effect of local a surface
temperatures on atmospheric circulation over the tropical Atlantic ctor.
J.Clim.13,2195–2216(2000).
15.Wang,C.,Enfield,D.B.,Lee,S-K.&Landa,C.W.Influences of the Atlantic
warm pool on Western Hemisphere Summer Rainfall and Atlantic Hurricanes.
J.Clim.19,3011–3028(2006).
16.Watanabe,M.&Kimoto,M.Tropical-extratropical connection in the
拟把疏狂图一醉Atlantic atmosphere–ocean variability.Geophys.Res.Lett.26,
2247–2250(1999).
17.Smith, al.Skilful multi-year predictions of Atlantic hurricane
frequency.Nature Geosci.3,846–849(2009).
18.Gill,A.E.Some simple solutions for the heat induced tropical circulation.
Quart.J.Meterol.Soc.106,447–462(1980).
19.Xie,S.P.A dynamic ocean–atmosphere model of the tropical Atlantic decadal
variability.J.Clim.12,64–70(1999).
20.Ham,Y-G.,Kug,J-S.&Kang,I-S.Role of moist energy advection in formulating
anomalous Walker Circulation associated with ENSO.J.Geophy.Res.112, D24105(2007).
21.Kug,J-S.&Kang,I-S.Interactive feedback between the Indian Ocean and
ENSO.J.Clim.19,1784–1801(2006).
简单公主头扎法
22.Vimont,D.J.,Wallace,J.M.&Battisti,D.S.Seasonal footprinting in the
Pacific:Implications for ENSO.J.Clim.16,2668–2675(2003).
怎样提高英语23.Larkin,N.K.&Harrison,D.E.On the definition of El Niño and
associated asonal average US weather anomalies.Geophys.Res.Lett.32,
L13705(2005).
24.Kao,H-Y.&Yu,J-Y.Contrasting eastern-Pacific and central-Pacific types of
ENSO.J.Clim.22,615–632(2009).
25. al.El Niño in a changing climate.Nature461,511–514(2009).
26.Saji,N.H.&Yamagata,T.Structure of SST and surface wind variability
during Indian Ocean dipole mode events:COADS obrvations.J.Clim.16, 2735–2751(2003).
27.Rienecker, al.MERRA—NASA’s Modern-era retrospective analysis
for rearch and applications.J.Clim.24,3624–3648(2011).
28.Smith,T.&Reynolds,R.Improved extended reconstruction of SST
(1854–1997).J.Clim.17,2466–2477(2004).
29.Behringer,D.W.&Xue,Y. Integrated Obrving and Assimilation
Systems for Atmosphere,Oceans,and Land Surface,AMS84th Annual Mtg
11–15(Washington State Convention and Trade Center,2004).
30.Delworth, al.GFDL’s CM2global coupled climate models.
Part I:Formulation and simulation characteristics.J.Clim.
19,634–674(2006).Acknowledgements
This work was supported by the National Rearch Foundation of Korea(Grant
NRF-2009-C1AAA001-2009-0093042)funded by the Korean government(MEST),and KIOST(PE98801).F.F.J.was supported by US NSF grant ATM1034798,US Department of Energy grant DESC005110and US NOAA grant NA10OAR4310200.
Author contributions
Y-G.H.,J-S.K.and F-F.J.designed the rearch and wrote the paper.Y-G.H.and J-Y.P. performed the e
xperiments and analyd the data.All of the authors discusd the results and commented on the manuscript.
Additional information
Supplementary information is available in the online version of the paper.Reprints and permissions information is available online at /reprints.Correspondence and requests for materials should be addresd to J-S.K.
Competingfinancial interests
The authors declare no competing financial interests.
