Atlantic Climate Variability and Its Associated Atmospheric Circulation Cells
Total Page:16
File Type:pdf, Size:1020Kb
1516 JOURNAL OF CLIMATE VOLUME 15 Atlantic Climate Variability and Its Associated Atmospheric Circulation Cells CHUNZAI WANG Physical Oceanography Division, Atlantic Oceanographic and Meteorological Laboratory, NOAA, Miami, Florida (Manuscript received 1 May 2001, in ®nal form 19 October 2001) ABSTRACT Phenomena important for Atlantic climate variability include the Atlantic zonal equatorial mode, the tropical Atlantic meridional gradient mode, and the North Atlantic Oscillation (NAO). These climate phenomena and their associated atmospheric circulation cells are described and discussed using the NCEP±NCAR reanalysis ®eld and the NCEP sea surface temperature (SST) from January 1950 to December 1999. Atmospheric divergent wind and vertical motion are used for the identi®cation of atmospheric circulation cells. During the peak phase of the Atlantic equatorial mode, the Atlantic Walker circulation weakens and extends eastward, which results in surface westerly wind anomalies in the equatorial western Atlantic. These westerly wind anomalies are partly responsible for warming in the equatorial eastern Atlantic that occurs in the second half of the year. The Atlantic equatorial mode involves a positive ocean±atmosphere feedback associated with the Atlantic Walker circulation, similar to the Paci®c El NinÄo. The tropical Atlantic meridional gradient mode is characterized by a strong SST gradient between the tropical North Atlantic (TNA) and the tropical South Atlantic. Corresponding to the meridional gradient mode is an atmospheric meridional circulation cell in which the air rises over the warm SST anomaly region, ¯ows toward the cold SST anomaly region aloft, sinks in the cold SST anomaly region, then crosses the equator toward the warm SST region in the lower troposphere. The analysis presented here suggests that the Paci®c El NinÄo can affect the TNA through the Walker and Hadley circulations, favoring the TNA warming in the subsequent spring of the Paci®c El NinÄo year. The NAO, characterized by strong westerly air¯ow between the Icelandic low and the Azores high, is also related to an atmospheric meridional circulation. During the high NAO index, the atmospheric Ferrel and Hadley cells are strengthened, consistent with surface westerly and easterly wind anomalies in the North Atlantic and in the mid-to-tropical Atlantic, respectively. 1. Introduction phase. This interannual phenomenon is called the At- lantic zonal equatorial mode (or the Atlantic El NinÄo). Atlantic climate variability shows many important Early studies, using empirical orthogonal function phenomena on different timescales. Similar to the trop- (EOF) analysis or correlation of rainfall indices with ical Paci®c, the tropical Atlantic exhibits an equatorial SST ®elds, suggest that the tropical Atlantic also dis- cold tongue, and northeasterly and southeasterly trade plays a ``dipolelike'' variation for SST, wherein the trop- winds converging into an intertropical convergence ical regions to the north and south of the ITCZ are zone (ITCZ) north of the equator, all of which display antisymmetric (e.g., Weare 1977; Hastenrath 1978; a seasonal cycle. Unlike the tropical Paci®c, the seasonal Moura and Shukla 1981; Servain 1991; Nobre and cycle dominates the ocean±atmosphere signal in the Shukla 1996). Recent modeling studies demonstrate the tropical Atlantic. Superimposed on the seasonal cycle possibility that an ocean±atmosphere feedback mode is climate variability on interannual to interdecadal time- may produce such antisymmetries on the decadal time- scales. A phenomenon similar to but weaker than the scale (e.g., Chang et al. 1997; Xie 1999). The existence Paci®c El NinÄo also occurs in the Atlantic. The warm of such a dipole is controversial, however. For example, events reach their maximum strength in the second half of the year, with manifestations focused primarily near studies such as Houghton and Tourre (1992), En®eld the equator (e.g., Zebiak 1993; Carton and Huang 1994; and Mayer (1997), Mehta (1998), En®eld et al. (1999), Latif and Grotzner 2000). During a warm phase, trade and Dommenget and Latif (2000) suggest that the trop- winds in the equatorial western Atlantic are weak and ical North Atlantic (TNA) and tropical South Atlantic sea surface temperature (SST) is high in the equatorial (TSA) SST anomalies occur independently and dipole eastern Atlantic. The converse occurs during a cold con®gurations are not ubiquitous. Nevertheless, the in- dex formed by the SST anomaly difference between the TNA and TSA (Servain 1991), sometimes referred to as a dipole index, has proved to be an invaluable in- Corresponding author address: Dr. Chunzai Wang, NOAA/AOML/ PhOD, 4301 Rickenbacker Causeway, Miami, FL 33149. dicator of the meridional gradient in SST anomaly that E-mail: [email protected] is correlated with north±south displacements of the 1JULY 2002 WANG 1517 ITCZ and with strong climate anomalies over the sur- 7 document atmospheric circulation cells associated rounding land regions. It thus can be more usefully re- with the Atlantic zonal equatorial mode, the tropical garded as a meridional gradient index. Moreover, much Atlantic meridional gradient mode, the remote in¯uence evidence exists for an ocean±atmosphere couplingÐ of the Paci®c El NinÄo on the TNA, and the NAO, re- stronger in the TNA regionÐthat causes reinforcement spectively. Section 8 provides a discussion and sum- and persistence of meridional gradient anomalies mary. through positive feedback involving the trade winds and associated surface ¯ux anomalies (e.g., Carton et al. 2. Data 1996; Chang et al. 1997; Xie 1999). This meridional antisymmetry appears to work preferentially at the de- The major data source in this study is the NCEP± cadal timescale (Mehta and Delworth 1995; En®eld et NCAR reanalysis ®elds from January 1950 to December al. 1999), though it explains little of the total SST anom- 1999. The NCEP±NCAR reanalysis ®elds use a state- aly variance. In view of this and the possibly misleading of-the-art global data assimilation system on a 2.58 lon- nature of the term ``dipole,'' we prefer to refer to this gitude by 2.58 latitude grid [see Kalnay et al. (1996) for variability simply as the tropical Atlantic meridional details]. Variables used in this study are monthly sea gradient mode. level pressure (SLP) and monthly atmospheric horizon- Another important Atlantic climate phenomenon is tal wind velocity, vertical velocity, and velocity poten- the North Atlantic Oscillation (NAO) (e.g., Hurrell tial at levels of 1000, 925, 850, 700, 600, 500, 400, 1995, 1996) and its possible relationships and interac- 300, 250, 200, 150, and 100 mb. The vertical component tions with the tropical Atlantic. Much of the climate of wind ®eld in the reanalysis ®elds is pressure vertical variability over the North Atlantic and surrounding con- velocity. In our presentation, we multiply the pressure tinents has been correlated with changes in the intensity vertical velocity by 21, so positive values of the vertical of the NAO. The NAO is associated with the variations velocity indicate an upward movement of air parcels. of surface westerly wind in the midlatitude Atlantic, of Since we are interested in tropical atmospheric circu- the Icelandic low, and of the subtropical anticyclone lation and northern midlatitude atmospheric circulation centered near the Azores. Xie and Tanimoto (1998) sug- associated with Atlantic climate variability, our analyses gested that extratropical variability drives tropical var- in this paper are shown from 308Sto758N over the iability. Using an atmospheric general circulation mod- Atlantic and the eastern Paci®c. Note that our analyses el, Robertson et al. (2000) suggested an in¯uence of the were performed globally and results associated with Pa- southern Tropics on the NAO. This may not be consis- ci®c ENSO are reported in a companion paper (Wang tent, however, with the fact that the NAO index does 2002). not signi®cantly relate to the TSA SST anomalies. Horizontal wind velocity can be divided into a non- Deque and Servain (1989) emphasized the importance divergent (or rotational) part and a divergent (or irro- of the atmospheric circulation over the North Atlantic tational) part (e.g., Mancuso 1967; Krishnamurti 1971; midlatitudes in tropical Atlantic SST. Ruiz-Barrades et Krishnamurti et al. 1973): v 5 vc 1 vf 5 k 3 =c 1 al. (2000) recently concluded that the extratropical At- =f, where c is streamfunction and f is velocity po- lantic does not affect the tropical Atlantic, but that there tential. The ®rst part does not contribute to atmospheric are some impacts of the Tropics on the extratropics. divergent ®elds associated with atmospheric vertical As summarized above, all of these Atlantic climate motion (it is nondivergent). It is well known that the phenomena have been previously studied through either Walker and Hadley cells are thermally driven, associated data analyses or numerical models. However, variations with atmospheric convergence±divergence. Atmospher- of the atmospheric Walker, Hadley, and Ferrel circula- ic heating associated with convection induces atmo- tion cells associated with these Atlantic climate phe- spheric convergence±divergence that drives atmospher- nomena have not been documented and they have been ic vertical motion and circulation. What matters to at- little studied. The recent availability of the National mospheric cells associated with atmospheric conver- Centers for Environmental Prediction±National Center gence±divergence