VOLUME 122 MONTHLY WEATHER REVIEW SEPTEMBER 1994The South Pacific Convergence Zone (SPCZ): A Review DAYTON G. VINCENTDepartment of Earth and Atmospheric Sciences, Purdue University, West Lafayette, Indiana(Manuscript received 14 September 1993, in final form 21 December 1993) ABSTRACT The cimulation features associated with the South Pacific convergence zone (SPCZ) and its accompanyingcloud band are reviewed and discussed. The paper focuses on the following topics: location, structure, andcharacteristics of the SPCZ; theories and observations concerning its existence; the significance and scope ofthe SPCZ in global-scale circulation patterns; quasi-periodic changes in its location and strength; and synopficscale features within its regional influence (e.g., cyclones, subtropical jets). It concludes with some challengingproblems for the future.1. Introductory remarks The South Pacific convergence zone (SPCZ) contains one of the earth's most expansive and persistentconvective cloud bands, and is recognized as playing asignificant role in global-scale circulation patterns. Although first depicted in the surface analyses of Bergeron (1930), the vast extent of the SPCZ, both spatiallyand temporally, was not fully appreciated until satelliteimagery became available in the early 1960s (e.g., Hubert 1961). Since that time, numerous studies have beenconducted of this rather intriguing phenomenon. Theprimary purpose of this paper is to provide a convenientsummary of the knowledge gained from those SPCZstudies. In that context, the paper focuses on a reviewof past and exiting theories and observations about theSPCZ. It is important'to note that it is not the intent ofthe paper to choose between, or criticize, contrastingopinions where they may occhr. It should also be mentioned that, with regard to the observational results presented herein, it was frequently necessary to rely onthose obtained during the first special observing period(SOP-l) of the First GARP (Global Atmospheric Research Program) Global Experiment (FGGE); results,particularly above the surface, are otherwise quitesparse in the literature. For this reason, some of thefindings cited in this paper, especially those toward theend, are for January and February 1979. In these cases,it was difficult to draw any general conclusions. The organization of the paper is as follows. In section 2, the location, structure, and characteristics of theSPCZ are described, with attention devoted to seasonal Corresponding author address: Dr. Dayton G. Vincent, Department of Earth and Atmospheric Sciences, Purdue University, 1397Civil Engineering Building, West Lafayette, IN 47907.c 1994 American Meteorological Societychanges. This is followed by a summary of theories andobservations concerning its origin and maintenance(section 3). Next, the significance of the SPCZ and itsrole within global-scale circulation patterns are discussed (section 4). Section 5 presents a synopsis of thequasi-periodic behavior of the SPCZ. The focus is onintraseasonal and interannual oscillations, since seasonal changes are discussed in section 2. Section 6 discusses some of the synoptic-scale features associatedwith the SPCZ, such as cyclones, subtropical jetstreaks, and storm tracks. Here, the results discussedare only for a 2~week period in January 1979. Finally,in section 7 concluding remarks are given and somechallenging problems are offered. It will be seen thatsome of the problems arise because of the lack of observational evidence, especially outside the SOP-I ofFGGE. Before proceeding, the author would like tostate that he is indebted to many of his fellow scientistswhose solicited contributions and comments made thispaper possible. These colleagues are recognized in theacknowledgments section.2. Location, structure, and characteristics The name, SPCZ, is generally attributed to Trenberth(1976). In its mean annual position the axis of the SPCZstretches from New Guinea east-southeastward to about30-S, 120-W. In its northwestern sector the SPCZ becomes more zonally oriented and merges with the intertropical convergence zone (ITCZ), which extendswestward into the Indian Ocean. Figure 1, extractedfrom Trenberth (1991a), serves as a representative example of the main surface features and the cloud bandassociated with the average location of the SPCZ. Itshows, as Trenberth (1976, 1991a) and others (e.g.,Streten and Zillman 1984; Kiladis et al. 1989; Kodama1992, 1993) have noted, that the SPCZ lies in a region19491950 MONTHLY WEATHER REVIEW VOLUME 122I I I I.~~iiiiii', i~~iii'~ii iiiiiii'~i ii~?:i :~iill.,:4~:ii:. !i::::.::i:: ::::::i? r~~ ....~, ~,~,~, ,~,~: /..~i ~o- x '~ :~2 ::i;i ::i:'~ii'~ ;:~:i:i i~ ~ iiii?; _ u-_ 5 .. " b': '_~a' .~':~.'~:~ ::~::~;:~:::~::~'~:~:?~:~:~ :~:-~::~:~:~:?~:~:~ ~:~~::: ~.: . <,o ~. ~.... ,:.,.:.:.:.:. ~ z ::x :.: ::.: , .~ ~ ~ ~ iO~ 2 :?: ::?~ :~:~:: h ~ "~ ''~-~ ~?~ :~?~ ":~ ~ .~ ~ u ,.:~ .~ ~~:~'~ ... ~ ~.. ~ .'~ ..~ :~~:.~?? _~::~?? .?.~~ /: o ;~0 ~$?OE IZtO I-~ 180 IS'O I,,~ V ' ~1 FIG. 1. Schematic view of the main convergence zones, the ITCZand SPCZ, along with the annual mean sea level pressure contoursand surface wind streamlines. (Extracted from Trenberth 1991a.) 'of low-level moisture convergence, between the predominantly northeasterly flow west of the eastern Pacific subtropical high and the cooler predominantlysoutheasterly flow from higher latitudes. It is worth noting that Fig. 1 is a schematic representation of the meancirculation features that undergo changes on severaltemporal scales. For the present, only seasonal changesare discussed; longer- and shorter-term changes will beaddressed in sections 5 and 6, respectively. Figure 2 shows the mean sea level pressure (MSLP)patterns for the months of January and July, obtainedas 6-yr averages (1985-90) from the World ClimateResearch Programme/Tropical Oceans and Global Atmosphere (WCRP/TOGA) archive II European Centrefor Medium-Range Weather Forecasts (ECMWF) analyses. In January, the most prominent feature is thetrough of low pressure that extends eastward from themonsoonal low centered over northern Australia acrossthe Pacific to a location near the equator and 130-W.The western part of this trough is commensurate withthe zonal portion of the SPCZ. Also shown is the troughassociated with the diagonal portion of the SPCZ,which extends southeastward from 10-S, 170-E to30-S, 160-W. In July, the pattern has changed drastically. The trough of low pressure is now located alonga line that stretches from the low centered over Southeast Asia, associated with their summer monsoon, eastsoutheastward across the Pacific to just east of the dateline where it then remains close to the equator. In contrast to January, there is a high pressure center locatedover southern Australia and no obvious pressure troughover the South Pacific. Comparing the MSLP patternsin Fig. 2 to the one in Fig. 1 reveals that Trenberth's(1991a) schematic is approximately a blend of the January and July distributions. - The other pattern shown in Fig. 1 is surface streamlines. For comparison, Fig. 3 shows the 6-yr monthlydistributions of streamlines and isotachs at the surface(10 m) for January and July. In January, northeast tradewinds extend across the entire tropical and subtropicalNorth Pacific. In the South Pacific and over northernAustralia, there is strong confluence into the monsoonaltrough and SPCZ region. In July, southeast trades prevail over the South Pacific and there is an anticycloneover southern Australia. There is little evidence of theexistence of the SPCZ (i.e., no obvious confluent zone).As was seen in the pressure pattern, the streamline fieldin Fig. 1 appears to be an approximate blend betweenthe flow fields depicted for January and July in Fig. 3.It is worth noting that the pressure and wind fieldsshown in Figs. 2 and 3 are in good agreement withthose given by Sadler et al. (1987). The latter is anexcellent reference of the long-term mean monthly distributions of MSLP, surface wind and wind stress, andsea surface temperature in the Tropics and subtropics. As stated in the introduction, the SPCZ contains oneof the earth's most expansive and persistent cloudbands. Early evidence of this was given by Streten(1973) who used 5-day averages of satellite mosaicsover a 3-yr period (1968-71) in the Southern Hemisphere to construct seasonal means of percent cloudcover. He found that the SPCZ contained the highestfrequency of clouds in all seasons, with maximum annual values in excess of 40% stretching from NewGuinea to approximately 30-S, 120-W. A similar axisof maximum cloudiness was found by Atkinson andSadler (1970) and Gruber (1972). In later studies, however, it was determined that the SPCZ cloud band ismost intense in southern summer. For example, in astudy by Meehl (1987), the monthly mean values ofoutgoing longwave radiation (OLR) showed a more ex 140E ID 140W 100W FIG. 2. Mean sea level pressure (hPa), averaged for all Januarys1985-90 (upper) and all Julys 1985-90 (lower). Also shown are thelocations of the pressure troughs (dashed) associated with the equatorial trough zone and SPCZ.SEPTEMBER ] 994 V I N C E N T 195120N ?20S .... 100EC.2oNEq205120E 140E 160E 180 160W 140W100E 120E 140E 160E 180 160W 140Wbe EQ~,c.~ .~-~~~~~.,[ ......... ~~ ............... i ..........20S~ .... -----~ ......... I~E 120E 140E 160E 180 I~W 140Wd. ........... C' ............. ..... I~E 120E 140E 160E 180 160W 140WFiG. 3. Streamlines and isotachs (m s ~) at 10 m for January [(a) and (b)] and July [(c) and (d)], averaged for years 1985-90.tensive and convectively active cloud band in Januarythan in the other three midseason months (Fig. 4). There is additional evidence that indicates the SPCZis generally more active in southern summer than atother times during the year. As an example, considerFig. 5, which shows maps of vertical p velocity (w)at 500 hPa for the four 3-month seasons, JuneAugust, September-November, December-February,and March-May. As for Figs. 2 and 3, these resultswere obtained from the 6-yr averages (1985-90) ofWCRP/TOGA archive II ECMWF analyses. The figureclearly shows that midtropospheric upward motionalong the SPCZ, particularly in the diagonal portion, ismuch stronger and extensive in December-Februarythan in other seasons. A convenient summary of the seasonal changes thatoccur in the SPCZ region is provided by Meehl (1987).He illustrates the annual cycle of several variablesbased on island station reports across the Indian andPacific Oceans. His results clearly illustrate the southward and eastward progression of different measuresof convective activity from India in July to the SPCZregion in January. For example, his time series ofmonthly means show minimum surface pressure anomalies, maximum cloudiness, minimum OLR, and maximum precipitation reaching the zonal portion of theSPCZ, near northern Australia, by January. From January to March these variables propagate across thewestern Pacific Ocean to about the date line. In thediagonal portion of the SPCZ the respective minimaand maxima are not as distinct but show quasi-steadyvalues from December to March. An example ofMeeh!'s results is shown in Fig. 6, which depicts sealevel pressure anomalies and precipitation at stationsfrom India to the eastern Pacific. It is seen that thelowest surface pressures and highest precipitation ratesin the SPCZ region take place in the southern summerseason. Because all of the variables discussed thus farhave shown the SPCZ to be more intense in southernsummer, the emphasis for much of the remaining paperwill be on that season. Furthermore, as stated in sectionl, many of the detailed analyses of SPCZ features wereavailable only for the SOP-l, FGGE period (i.e., southern summer months of January and February). With this in mind, we shall continue our discussionof the location, structure, and characteristics of theSPCZ. Figure 7 shows maps of precipitation, sea surface temperature (SST), and surface wind convergencefor January, extracted from Kiladis et al. (1989). Theynote that, "the surface convergence maximum associated with the SPCZ lies to the south of the axis ofmaximum precipitation, typical of tropical convergencezones" [e.g., GARP (Global Atmospheric ResearchProgram) Atlantic Tropical Experiment (GATE) A/Bscale ship array]. They further note, as Fig. 7 showsthat, "maximum precipitation lies to the south of theaxis of maximum SST." The mean flow patterns at 200 and 850 hPa for January are shown in Fig. 8. These charts were obtainedfrom the same ECMWF dataset used to derive earlierfigures. Of primary interest here are the two SouthernHemisphere upper-level anticyclones that are approximately superimposed on (slightly east of) two low-levelcyclones. This pair of circulation systems, togetherwith the ridge (at 200 hPa) and trough (at 850 hPa)between them, lies along the axis of upper-level out1952 MONTHLY WEATHER REVIEW VOLUME 122a)~O*N$0'$ 45,1r ,JULY OUTGOING LONGWAVE~//.~iiiii::ii~!iiii!iiiiiiiiiiii!ii:l, ? , - / ~-~,o/ x\_ ~ ,o.,v~/~_.~.....~ :.~ ~ ~.o / ~~ ::-"" ":'::::~:::::::: ~ 260 4 / I/A~~~~~ ~:~o~ ~ ~ ~= o~=~~2/~'~~~~~'~~~~::~ ~ [~ ~ '. ~ leo :~::" 20 ~80 ~g ~: . . ~ , ~ ~~- ~~. _ ~..~ , ~ ~ ,~~'~ ,o.~ ~0' ~ ~80 ~0' Wb)~0' N, OCTOBER OUTGOING LONGWAVE?-----~'-~.o~ .---"-- ~ (\~:2.~ ~'----~o-~ "~-~._~/ ~J~ ~ ~,,~ ~ ~f ~ ~.~.~~~.~~~~ .~ ~ .... :~::~:~:%. t o ' . ~o ~O*S45'E 90'E 180 90'w~1 JANUARY OUTGOING LONGWAVE 30'N ~0~ x ~._.~f':"'~,o_.~-~ ..... ' ~ t4o-------"~. ~,-.~3o., .~,,~.~-260~ 240 \~% ~f' I 30'S 45-E 90- E 180 90' Wd) APRIL OUTGOING LONGWAVE"'" :~e..:..,_~~~t~'.:.'-'~:'~4o ~ ~o ~-- , ~ -~ ,o., . . ~ .... '""::~ii~i~i~i~ii~::. - ~_ ~ t~$O*S J '~ 46" I 30'$90' E 180 90'W FIG. 4. Eight-year means of OLR (W m-2) for the period June 1974-November 1983 (1978 missing) for(a) July; (b) October; (c) January; and (d) April. Areas of OLR <~ 220 W m-a are shaded (extracted fromMeehl 1987).flow and lower-level inflow associated with the zonalportion of the SPCZ. There is also an upper-level ridge(lower-level trough) that extends southeastward fromthe easternmost anticyclonic (cyclonic) center. Thisfeature is aligned along the approximate location of thediagonal portion of the SPCZ. The flow pattern at 200hPa is in good agreement with that presented in theatlas by Sadler (1975). The upper-tropospheric flow regime and its relationship to the SPCZ cloud band are now examined in moreSEPTEMBER 1994 V I N C E N T 1953ae 140E 160E ID 160W l~OWd. 14OE 160E ID 160W 140W FIG. 5. Seasonal averages for years 1985-90 of vertical p velocityw (Pa s)-~ at 500 hPa for (a) June-August, (b) September-November, (c) December-February, and (d) March-May. Upward motionless than or equal to 0.04 is hatched.detail. Figure 9 shows depictions of the mean state upper-level flow given by three different sets of authors[Sadler (1972) for January at 200 hPa, Madden andZipser (1970) for March-April 1967 at 250 hPa basedon the Line Islands Experiment, and Vincent (1982) for10-18 January 1979 at 200 hPa based on FGGE SOP-!data]. The latter two references also show the locationof the cloud band. Similar patterns are evident in eachdepiction and are also comparable to those in Fig. 8.An anticyclone is present near 10-S, just west of thedate line, with a ridge extending east-southeastwardfrom it to the central and, sometimes, eastern Pacific.The ridge is collocated with the cloud band, a positionthat is favorable for providing divergent outflow to sustain the convection. It could also be argued that thedeep convection within the SPCZ causes the ridge. Acol area is located just north of the equator near thedate line, with a prominent trough line extending fromit along an axis that parallels the ridge line. In the depiction by Vincent (1982), there is another trough lineto the southwest of the ridge line and cloud band. It isinteresting to note that Manabe et al. (1970) were ableto reproduce the trough-ridge pattern illustrated in Fig.9 with an early version of the Geophysical Fluid Dynamics Laboratory (GFDL) general circulation model(GCM), which contained a convective adjustmentscheme. Without the scheme (i.e., with their "dry"model), the flow across the subtropical South Pacificwas essentially zonal. Their results suggest, therefore,that SPCZ convection causes the upper-troposphericridge. The relationship between the surface pressuretrough, alluded to earlier, and the SPCZ cloud band areshown in Fig. 10 for a 9-day period during SOP-l,FGGE, 10-18 January 1979. The figure was createdfrom the results of Huang and Vincent (1983). Notethat the pressure trough lies along the southwesternedge of the cloud band and, thus, is poleward and westward of the upper-level ridge line discussed above. Ithas been observed or suggested by several authors thatthe trough-ridge system in the vicinity of the SPCZtilts poleward and westward with height (e.g., Trenberth 1976, 1991a; Vincent 1982; Kiladis et al. 1989;Kiladis and Weickmann 1992a). This tilt supports thesuggestion of many investigators, namely, that the diagonal portion of the SPCZ is highly baroclinic. The vertical structure of the zonal and diagonal portions of the SPCZ were depicted by Vincent (1982) forthe same period used to construct Fig. 10. Selected asan example is Fig. 11, which shows vertical cross sections of the 9-day averages (10-18 January 1979) ofseveral variables along a line that is perpendicular to,and crosses, the diagonal portion of the SPCZ cloudband. The center of the cloud band for this period was25-S, 142.5-W, and the cloud band width (convectivelyactive portion) was the equivalent distance of about 10-of latitude--that is, 5- each side of the center point.The figure shows that a deep layer of moist air occurs 1954 MONTHLY WEATHER REVIEW VOLUME 122~) SEA LEVEL PRESSUREb) PRECIPITATION I I ....'~ .... l' " ! '1 ~., i& $ ,- ~2 :~:~L .:x ~.:~.~Q~, , , ~x:~:x~.~, xx:~ '..~ '~ '~ ...~ .~ .... .k'. Z. t. "! ~ ?~??i 5. ~ ~;:::;J, -/~:~. ~4.'i.;:,, ~ ~'~ ,,:':C .,~ F ' x x. \ \,, x' xx \xx.,' x, \, x\,., x\:, v,\xx-. FIG. 6. Annual cycle of long-term mean station data of (a) sea level pressure and (b) precipitation, plotted as deviations of 3-month runningmeans from the annual mean. Stippling denotes highest pressures in (a) and greatest precipitation in (b), and vice versa for cross-hatching.Each tic mark on the horizontal axes indicates 5.0 mb in (a) and 100 mm month-~ in (b) (extracted from Meehl 1987).in conjunction with the narrow band of maxima in lowlevel convergence, upper-level divergence, and risingmotion. Outside the cloud band, drier air is descending.Near the center of the cloud band the relative vorticityis near zero at all levels; however, away from the cen 120*E I~~ I~~ 180' I~* I~~ 120 I00 ~ ~W Fro. 7. (a) Mean Janua~ precipitation (cm month-~) for the period1950-79. Also shown is the axis of highest precipitation (solid) andthe axis of maximum surface wind convergence for the pedod 1961 80 (dash); (b) mean Januau SST (-C) for the period 1950-79. Solidline depicts the axis of maximum SST, and the dashed line ~e axisof highest precipitation, taken from (a) (extracted from Kiladis et al.1989).ter, there is cyclonic vorticity to its southwest and anticyclonic vorticity to its northeast throughout the troposphere. Temperature anomaly and wind cross sections inFig. 11 offer the most revealing insight into the physical processes that occurred within the cloud band andits environment during the 9-day period. Anomalieswere calculated as a grid point's departure from theaverage of all grid points along each constant pressurelevel. It is seen that temperature anomalies in the middle and upper troposphere are appreciably higherwithin the band than they are outside the band, suggesting that latent heat dominates over adiabatic cooling in the strong ascending motion. Wind speeds showminimum vertical shear 'near the center of the cloudband and strong vertical shear toward the edges of theband, particularly to its southwest. This section closes with a comparison of the verticalprofiles of convective heating between the SPCZ andother tropical regions. Figure 12 shows such profilesderived from the Ql-budget technique of Yanai et al.(1973). The profile on the left is taken from Miller andVincent (1987) and represents the average convectiveheating for the SPCZ region bounded by 7.5--27.5-Sand 170--135-W, for the period 10-18 January 1979.The profiles on the right are those determined by otherinvestigators using similar methodology (Reed andReeker 1971; Nitta 1972; Yanai et al. 1973; ThompsonSEPTEMBER 1994 V I N C E N T 1955ao2o/~Co20N205100E 120E 140E 160E 180 160W 140WIOOE 120E 140E 160E 180 160W 140WFro. 8. Streamlines and isotachs (ms ~) at 200 hPa [(a) and (b)] b* 20S --t 100E 120E 140E 160E 180 160W 140W d. ~Si:~'~iiii'~i i i ~ ~?~Z'--,~"'"----a-~ ' ~ - ~/.~, ,~ i ~.*~?..L~.o~ i ' ' ~' ' -~-'- '~ ....... i ....... 20N 20~5-'~s> ~:~~~~ "~ ............. '~~ ................... -~,~ ~ i 120E 140E 160E 180 160W 140Wand 850 hPa [(c) and (d)] for Janu~y, averaged for years 1985-90.et al. 1979). In the first three studies, data gathered overthe Marshall Islands were used. Despite differences inthe large-scale conditions (i.e., the periods of studywere different), and analysis techniques, the three heating profiles are similar. They all show maximum heating in the 500-400-mb layer, comparable to the resultsof Miller and Vincent. Thompson et al. used data fromthe GATE B-scale ship array during phase III and obtained a somewhat different heating profile. Their profile shows maximum heating at a much lower level(~600 rob) than those obtained over the Marshall Islands area. Thompson et al. suggest that deep cloudsand convection frequently dominate the Marshall Islands area, while multiple cloud layers that are oftenobserved in the GATE region, yield a lower level ofmaximum heating. Finally, although it was not possibleto reconstruct a vertical profile of heating from thestudy by Song and Frank (1983), they showed for theA/B-scale area during phase 1 of GATE that maximumconvective heating, in cases when precipitation ratesexceeded 0.5 mm h-~, occurred near 550 mb. However,when rates were less than 0.5 mm h-t, the maximumoccurred below 800 rob. It would appear that if theirrainfall-rate categories were combined, the net profilethat would result would be similar to that of Thompsonet al. Thus, it seems that the results for the South Pacificare more in line with those for the western portion ofthe North Pacific Ocean than with those for the GATEarea. It is important to note that the profiles shown in Fig.12 represent mean values over areas that contain convective systems, as well as clear regtons. Miller andVincent (1987) examined the average vertical distribution of convective heating for a 12-h period on 12January 1979 near the center of a cyclone in the SPCZcloud band. They found that the maximum value wasseveral times greater than the maxima shown in Fig.12. Also, there was a much deeper layer of high valuesof convective heating that extended well into the uppertroposphere.3. Origin and maintenance As discussed and shown above, the summertimeSPCZ consists of a zonal portion, generally locatedover the western Pacific in the "warm pool" region,and a diagonal portion, oriented northwest-southeast.There have been several attempts to explain the location, orientation, and strength of each of these components of the SPCZ, but because the explanations aresomewhat interrelated (i.e., not mutually exclusive), itis difficult to separate them into distinct hypotheses.Nevertheless, an attempt is made here to sort out someof the ideas regarding the origin and maintenance ofthe SPCZ. One theory is that the SPCZ, particularly thezonal portion, owes its existence to warm underlyingSSTs or, more explicitly, to $ST gradients (e.g., vonStorch et al. 1988; Kiladis et al. 1989; Trenberth 1991a;Kiladis and Weickmann 1992a). The scenario in thisregard is as follows: SST gradients impose pressuregradients that in turn drive low-level winds that resultin moisture convergence. This argument is similar tothat proposed by Lindzen and Nigam (1987) for theexistehce of the ITCZ in the western and central Pacific. In brief, they suggest that SST gradients accountfor a significant portion of the forcing of low-level1956 MONTHLY WEATHER REVIEW ,VOLUME 1220 150E200 mb JANUARY ;~~_~ ~~ .... ~.7~m. . - " ~ I 180 I,~)OW 120W 90WL>ON~ON 0 lOS20S170E 180 170W 160 150 140W Tlme-/~-oged FIo~ 200rob 10-18 d~u~ry 1979ION .... ~ , ~ , ,40 170E 180 170W 160 150 140 130 120 II0W FIG. 9. Upper-tropospheric winds.and SPCZ cloud band based on(a) mean ]anuary at 200 hPa (wind speeds in knots) adapted fromSadler (1972); (b) a schematic of streamlines at 250 hPa during theLine Islands Experiment, extracted from Madden and Zipser (1970);(c) streamlines and isotachs (m s-~), extracted from Vincent (1982)'.Trough lines are dashed and ridge lines are zig-zag; shaded areas in(b) and (c) represent SPCZ cloud band; for reference, crosses indicatecommon latitude-longitude locations (10-S, 180- and 20-S, 150-W)on each diagram.winds (i.e., beneath the trade-wind inversion), resultingin mass convergence. As a consequence, high-0e airaccumulates near the rising branch of the ITCZ wherethe convergence (primarily of water vapor) helps toconcentrate the location of deep convection and the accompanying latent heat release that occurs in the middle and upper troposphere. They state that moistureconvergence tends to be a maximum where SST gradients are greatest. They further state that while latentheating undoubtedly drives the upper-level outflow,there is no compelling reason to Suppose it will impacton the low-level convergence. In numerical simulationsby Lindzen and,his collaborators, the best results wereobtained when the effects of both low-level forcing bySST gradients and upper-level forcing by latent heatrelease were coupled. However, they could not explicitly determine the role that latent heating plays in producing sufficient low-level convergence. Finally, theynote that in the region occupied by the zonal portion ofthe SPCZ, forcing appears to be due primarily to zonalgradients of SST, although they also found that zonalvariations in the meridional gradients could be important. Support for this theory can be seen in the map ofsurface temperatures for January 1986 (Fig. !3). Notethat the warmest SSTs (nearly 304 K) occur in the western Pacific along about 5--10-S, between 150- and170-E. Also note that the SSTs along this latitude decrease to about 302 K between 170-E and 170-W. Ithas been shown in several figures thus far that the zonalportion of the SPCZ frequently lies close to this latitudebelt between 150-E and 170-W. Another theory for the origin and maintenance of theSPCZ is that it is due to land-sea distributions. Therehave been a few studies, primarily using GCMs (e.g.;Kalnay et al. 1986; and Kiladis et al. 1989), that investigated the potential influence of the three mainlandmasses in the Southern Hemisphere on the locationof the SPCZ, particularly during the summer seasonwhen continental heating is a maximum and the SPCZis most intense. The net result of these studies suggeststhat the positioning of the land-ocean distribution haslittle influence on the location of the SPCZ; however,in the experiments by Kiladis et al. (1989), changes inthe heating and circulation patterns over Australia inJanuary did have an impact on the strength of' the2C2~ I I I ~70w ~65 160 ~55 150 145 ~0 ~35w FIG. ]0. Average sea level pressure and OLR ~ 225 W m-~(shaded) for 10-18 January ]979. Trough line indicated by dashedline, This figure was compiled from diagrams presented by Huangand Vincent (1983).SEFI'EMBER 1994 V I N C E N T 1957 T* (-K) Across Cloud Band9 Day Average 10-18 January 1979 t // ~ ~45 ~4o ~ ~wI00150200~50500 pOnb)~005O0350'000 RH (%) Across Cloud Band Day Average 10-18 January 1979 100 250 / --~ 300/ /// I \ \sso , ~, / 4 / , 3 , ~o15~ 150 145 I~ I~ I~W~ (lO-"mbs-~) Across Cloud Bend9 Day Average 10-18 January 1979 040'S 35 , 30 ~ 25 ~ 20 e d~w ~50 ~45 15 , IO'S145 135 150'W-11004150-1200,4250-.1300 p(mb) 400 500700~850~lOO0 101 (ms-~) Across Cloud Band9 Day Average 10-18 January 1979 - . ~ / , -~100 ~ - /'.---.-'~" % // 4,5o 15 2O0 b):D"S ~ 35 ~ 30 , 25 ~ 20 ~ 15 ~ lOS I000 ~55~W 150 145 140 I.'.'55 130"W ~ (lO'fs-~) Across Cloud Bond 9 Day Average 10-18 January 1979 - I00 - 150 Y$~ 35 ~ 30 ~5 ~ 20 ~ 15 lOS ,~'w ,~o ,~,5 ,40 ,35 ,~o'w %. ~ (10-%-') Across Cloud Band 9 Day Average 10-18 January 197940 %.~w35 :,0 FIG. I 1. Vertical cross sections of temperature deviations (K, upper left), wind speed (m s ~, upper right), relativehumidity (%, middle left), relative vorticity in 10_5 s ~ (middle right), vertical p velocity in 10-3 mb s-~ (lower left),and horizontal divergence in 10-6 s-' (lower right) across the diagonal portion of the cloud band.SPCZ. They found that the removal of the continent ofAustralia in their GCM run destroyed the southernmonsoon and weakened the zonal portion of the SPCZ. Still another theory, again primarily related to thezonal portion of the SPCZ, was suggested by Davidsonand Hendon (1989). They investigated the interactions1958 MONTHLY WEATHER REVIEW VObUME 122 P(rob) P(rob)80025O3004OO5007OO850 LATENT HEATING PROFILES -18 JAN 1979 ! I I I0 I 2 3 4 5 6 7 8 9 K day-~ FIG. 12. Vertical profiles of convective heating, Q~ - QR, determined as residuals in the dry static energy (Q0 budget in units ofkelvins per day for (a) SPCZ area by Miller and Vincent (1987) for10-18 January 1979, and (b) western North Pacific region by Reedand Recker (1971), Nitta (1972), and Yanai et al. (1973), and easternNorth Atlantic region by Thompson et al. (1979).among the three main components of the SouthernHemisphere monsoon (Indonesian monsoon, Australian monsoon, and SPCZ) during the period December1978-January 1979 using winds and relative vorticityat 850 hPa. They were able to trace the path of threeprogressive eastward-moving cyclonic (and anticyclonic) vortices that enhanced SPCZ convection andstrengthened the South Pacific vortex, back to northernAustralia in one case and to the Indonesian sector inthe other two cases. They hypothesized that these patterns of downstream development were consistent withthe eastward-propagating'energy dispersion process described by advocates' of linear barotropic dynamics(e.g., Matsuno 1966; Webster and Chang 1988). Davidson and Hendon did note, however, "that the triggering of the initial (upstream) disturbance still largelyremains a mystery." At present, the most appealing hypothesis regardingthe origin and maintenance of the diagonal portion ofthe SPCZ focuses around mechanisms involving tropical-midlatitude interactions and/or midlatitude influences. For example, Kiladis et al. (1989) conclude thatthe central and eastern parts of the SPCZ reflect thepreferred storm tracks that propagate from lower tohigher latitudes when the tropical convergence zoneinteracts with transient troughs in the midlatitude westerlies. Their results are compatible with those of otherauthors (e.g., Streten and Zillman 1984; Vincent 1985;Hurrell and Vincent 1990). Other investigators havestressed the importance of maintaining the diagonalportion of the SPCZ through disturbances and frontalactivity that propagate equatorward from higher latitudes, generally in the vicinity of New Zealand (e.g.,Streten 1973; Webster and Curtain 1975; Trenberth1976, 1991a; Kiladis and Weickmann 1992b). Thesesystems tend to strengthen the low-level convergencebetween themselves and the quasi-stationary high in theeastern Pacific. An example of some of the processes describedabove was illustrated in the GCM results of Hurrell andVincent (1992). They conducted a number of experi7ments in which SSTs (and thus SST gradients) werereduced by varying amounts in the Southern Hemi~sphere tropical western Pacific (i.e., zonal portion ofthe SPCZ). Within a few days, the upper-troposphericdivergent outflow, induced by the heating, was 'substantially reduced. In the same time frame the subtropical jet, located just southeast of the divergent outflowand in the vicinity of the diagonal portion of the SPCZ,decreased by a factor of 2. In addition, poleward- andeastward-propagating Rossby waves, present in theircontrol run, were essentially absent. Thus, it appearsthat the baroclinic structure of the diagonal portion of20Ngg~ 140E 160E ID 160W 140W FIG. 13. Surface temperatures over land and ocean (SSTs) in kelvins for January 1986, taken from WCRP/TOGA ECMWF archiveII dataset.SEPTEMBER 1994 V I N C E N T 1959the SPCZ was influenced by tropical heating (SST gradients, in particular). Hurrell and Vincent (1991, 1992)showed that both midlatitude wave activity and divergent outflow from a tropical heat source were importantin maintaining the subtropical jet. Figure 14 summarizes their results, which were obtained as 15-daymeans for the period 6-20 January 1979. The top panelin Fig. 14 shows the location of the subtropical westerlyjet in question; it is centered at about 25-S, 170-W, andhas a maximum strength in excess of 30 m s-L Theremaining panels depict contributions to the acceleration of the zonal wind by the two most important termsin the localized E-P flux equations of Trenberth(1986). These are the Coriolis force due to the diabatically driven divergent meridional wind (fv*) and theE-P flux divergence (V .E,), which is a measure of thenet mean forcing by eddies on the time mean flow. Itis seen that the tropical forcing (fv*) acts to acceleratethe jet in its entrance region, whereas the midlatitudeforcing (V' E,) acts to decelerate it. In the exit region,both terms act to accelerate the jet. The net effect appears to move the jet eastward during the 15-day period, as was observed. Moreover, the results verify (atleast for this case study) that both tropical and extratropical forcing are important in maintaining the jet,which is an integral part of the strength and location ofthe diagonal portion of the SPCZ.4. Significance in global-scale circulations In preceding sections, the persistence and expanse ofthe SPCZ was established. In this section the significance of the SPCZ with regard to global-scale circulation patterns is demonstrated. Krishnamurti et al.(1973), based on their analysis of upper-troposphericwinds, together with Atkinson and Sadler's (1970)nephanalysis, noted that there were three principalregions of tropical convection during northern winter.They stated that the convection centers must accountfor the three waves containing subtropical jet streamsin the Northern Hemisphere and the three midoceanictroughs in the Southern Hemisphere. They concludedthat the regions of convection provide a major link between the flows in the two hemispheres. Since one oftheir troughs is an integral part of the SPCZ, and theSPCZ is a dominant convection region in the SouthernHemisphere, it seems reasonable that the SPCZ mightplay an important part in the cross-equatorial flow overthe western and central Pacific. This feature is readilyidentifiable in Fig. 8, where outflow from the SPCZregion is seen to penetrate into the Northern Hemisphere between the longitudes of approximately 150-Eand 160-W in January. Within the confines of the Southern Hemisphere,there are several measures of the importance of theSPCZ. The interactions that occur between the ITCZand midlatitude systems that penetrate into the subtropics across the western and central Pacific have al ZONAL WIND jl~~o~ ~ ~ 40S160E 180 160W 14011t/,2SI020~0 E-P FLUX DIVERGENCE ~~__i~~ o ,-'--':)D , ,I~E180 ISOW 140W FV*; ~ 2S f"" ~ ~ ~ ~ -,5 ~ -.~ ~ \ -io ~ I0160E 180 160W2O 40S140W FIG. 14. Averages at 200 hPa of the zonal wind component (m s-~,upper), E-P flux divergence (ms-~ day-I, middle), andfv* (m s-sday t, lower), for the period 6-20 January 1979. Sources of zonalwind acceleration greater than or equal to 10 m s-~ day ~ are shaded.ready been alluded to as an important mechanism forcausing and maintaining the SPCZ. The SPCZ is alsoknown to be an integral part of the Southern Oscillation(SO) and the Walker circulation (e.g., Trenberth 1976;Streten and Zilhnan 1984; van Loon and $hea 1987).More will be said in section 5 about the role of theSPCZ with regard to the SO and other quasi-periodiccirculation patterns. Attention is now focused on results in which quantitative estimates of the significance of the SPCZ havebeen given. Hurrell and Vincent (1987) performed adiagnosis of eddy energy budget terms in the SPCZ and1960 MONTHLY WEATHER REVIEW VOLUME 122EDOY ENERGY CYCLELEGEND CA CKGE BAE - BKE I000 -I00 mbENERGY CONTENTS (lOSJn~') ALL OTHERS (Win'2) 10-24 JANUARY 1979 TROPICS (0-30S) 0.0 0.10.1 0.0 0.0 0.10.0 0.0 0.1 0.2SPCZ S.A.0.1 03 0.2 0.0 FIG. 15. Time-averaged eddy energy cycles for 0--30-S and thefour tropical regions defined by longitude in the text for the period10-24 January 1979. Values are vertically integrated and area averaged, with contents in 105 J m-2 and other terms in watts per squaremeter (extracted from Hurrell and Vincent 1987).other Southern Hemisphere tropical regions of similarareal extent for a 15-day period during SOP-l, FGGE.They computed vertical integrals that were area averaged from 0- to 30-S, for four subregions bounded bythe following longitudes [Africa (AFR), 25-W-65-E;Australia (AUST), 65-- 155-E; SPCZ, 155-E- 115-W;South America (SA), 115--25-W]. Their results aresummarized in Fig. 15, which shows 1) the energy cycle for all regions being dominated by a generation ofeddy potential energy by diabatic heating processes, aconversion of eddy potential to eddy kinetic energy,and a dissipation of eddy kinetic energy; and 2) theenergy cYCle for the SPCZ being dominant over thatfor the other three regions. It is also seen that the eddykinetic energy content is substantially greater in theSPCZ than in other regions. Thus, it is speculated thatthe SPCZ makes a significant contribution to the eddyenergy budget of' the Southern Hemisphere Tropics inJanuary. In an earlier paper, Huang and Vincent (1985) foundthat the standing eddies were responsible for nearly allof the conversion of potential to kinetic energy (CE) inthe SPCZ and that the conversion was a maximum inthe 300-250-hPa layer. In that paper and a later one(Huang and Vincent 1988) they found that wavenumber 4 dominated the CE conversion in the SouthernHemisphere Tropics in the 10-24 January 1979 periodwhen the SPCZ was quasi-stationary, very intense, andlocated east of its normal position. One of the fourwaves (the dominant one) corresponded to the SPCZ,while the other three were centered over the three continents of Africa, Australia, and South America. From24 to 27 January the SPCZ convection and its associated wave activity and subtropical jet all decayed. Subsequently, there was a realignment of the long wavesto a wavenumber 3 pattern that lasted until the end oftheir study period (13 February). The convection centers associated with the latter pattern were centeredover the western Indian Ocean, the western Pacific, andSouth America. It is not currently known if this trans~formation from wavenumber 4 to 3 (and presumablyback to 4) is a regular feature of the tropical SouthPacific. If it is, the meandering of the SPCZ betweenthe central and western Pacific may play a key role.Further reference to this feature is given in section 5. Another quantitative measure of the significance ofthe SPCZ is given in the precipitation estimates of Vincent et al. (1991a). They used the Q~-budget method ofYanai et al. (1973) to compute monthly mean rainfallrates (as the budget residual) for the regions shown inFig. 16 from June 1984 to May 1987. All regions, withthe possible exception of the SPCZ, exhibited a distinctannual cycle with about twice as much rain in midsummer as in midwinter; however, the SPCZ had more rainduring the 3-yr period than any of the other regions. Inthe last two years (i.e., June 1985-May 1987) theSPCZ had an annual average rate of 6.1 nun day-~. Bynormalizing the corresponding values for other regions,based on a value of 1.00 for the SPCZ, the followingresults were obtained: SPCZ--1.00, SAFR--0.77,AUST--0.73, SHT--0.65, INDO--0.69, NHT-0.77, SACZ--0.67, and OTHR--0.46, where OTHR,SHT, and NHT represent the other (remaining) Southern Hemisphere tropical areas not accounted for byspecified regions, the Southern Hemisphere total Tropics (0--30-S), and the Northern Hemisphere total Tropics (0--30-N). It is worth mentioning that the reliabilityof the Q~-budget estimates was tested by Vincent et al.(1991a). They compared their precipitation values tothose from island station and atoll reports, as well as Seteot, ed Sout, hor'n HemLepheee Re~lbOne0' 80' E t20' g 180' E 1~0' H S0' W FIG. 16, Regions for which area-averaged values of precipitationwere computed and presented in text (extracted from Vincent et al.1991a).SEPTEMBER 1994 V I N C E N T 1961to those presented in some rainfall atlases, and concluded that the Q~-budget technique is capable of providing mean monthly values over open oceanic areasthat are consistent with observational data. Furthermore, in an earlier study (Miller and Vincent 1987), itwas found that Ql-budget estimates of precipitation inthe SPCZ region were in good agreement with thosederived using an IR satellite algorithm.5. Quasi-periodic behavior The location and intensity of the SPCZ (particularlythe diagonal portion) have been observed to fluctuateon a variety of timescales ranging from synoptic to interannual [e.g., see Gruber (1972) for one of the earliestdocumentations]. In this rather wide spectra of temporal scales, the ones discussed in this paper, beginningwith the shortest are 1) synoptic (few to several days);2) intraseasonal (2-4 weeks and 30-60 days); 3) season/annual (several months); and 4) interannual [e.g.,E1 Nifio/Southern Oscillation (ENSO)]. A discussion ofthe seasonal/annual cycle was given in section 2, anda discussion of the synoptic-scale variability is the focus of section 6. In this section the intraseasonal andinterannual oscillations are discussed. As stated above,intraseasonal fluctuations in the SPCZ appear to occuron two preferred timescales, 2-4 weeks and 30-60days. The latter is often referred to as the MaddenJulian oscillation (MJO) after the pioneering work byMadden and Julian (1971, 1972). Kiladis and Weickmann (1992a,b) examined 5 years of filtered data in theSouthern Hemisphere Tropics and higher latitudes forthree distinct periods: 6-14, 14-30, 30-70 days. Theyfound that the zonal portion of the SPCZ contained apronounced summertime spectral peak in OLR withinthe MJO range at about 50 days, in agreement withseveral other authors (e.g., Lau and Chan 1985, 1988;Weickmann et al. 1985; Knutson and Weickmann1987; Vincent et al. 1991b). They also found that thediagonal portion of the SPCZ exhibited a significantpeak near 18 days, which was captured in the 14-30day band. More will be said concerning this band later;for the present, a discussion of the 30-60-day periodis continued. There is ample evidence that intense convective episodes in the SPCZ are out of phase with those in theIndian Ocean (e.g., Weickmann et al. 1985; Hurrell andVincent 1990; Pedigo and Vincent 1990), and it appears that this dipole is modulated on MJO timescales.For example, Weickmann et al. (1985) used filteredOLR data to show the evolution and propagation ofconvective cloudiness around the global Tropics. Figure 17, reproduced from their results, shows that OLRanomalies over the Indian Ocean tend to be out of phasewitl~ those over the western to central Pacific Ocean(i.e., SPCZ region) on 28-72-day timescales. The resuits of Hurrell and Vincent (1990) and Pedigo andVincent (1990) support the relationship shown in Fig.4O20N 0 40 SOS IOOE I4oE leo 14ow ioo 60 Low o ~ 6oE FIG. 17. Schematic of OLR evolution for 28-72-day timescales. A cycle of cloudiness goes from I to 2 to 3 to 4 to I. OLR anomalies at I and 3, and 2 and 4, tend to be out of phase (extracted from- Weickmann et al. 1985).17. These papers, however, examined only a portion ofthe 60-day SOP-1 period of FGGE; therefore, it wasnot possible for the authors to verify that their findingswere related to an MJO type of cycle. Nonetheless,their results did show a shift of convective activity andcorresponding circulation features from the SPCZ region in mid-January 1979 to the western Indian Oceanby early February. Figure 18, which shows the OLRpatterns for two 2-week periods in January-February1979, illustrates this shift. From 10 to 24 January, theSPCZ exhibited strong convection in the central Pacific, whereas in late January and early February, convective activity weakened as the SPCZ tried to reestablish itself nearer Australia. In contrast, the westernIndian Ocean was essentially free of deep convectionin the first period but became relatively active in thesecond. Hurrell and Vincent (1990) found that therewas a persistent area of subtropical wind maxima to thesouth and east of the OLR minimum in the SPCZ region during mid-January. When maximum convectiveactivity shifted to the Indian Ocean in late January, italso contained a corresponding subtropical jet to itssouth and east. In the other study alluded to above, Pedigo and Vincent (1990) derived area- and time-averaged precipitation rates for regions similar to those in Fig. 16. Theycompared values computed as residuals from the Q~and Q2-budget methods of Yanai et al. (1973) for two15-day periods, 10-24 January 1979 and 28 January11 February 1979. The first period is the same as, andthe second one is nearly the same as, those illustratedin Fig. 18. Their results are given in Fig. 19 and showthat the best agreement in precipitation rate estimatesbetween the two methods occurs in the SPCZ and Indian Ocean regions. Furthermore, it is seen that valuesdecreased (increased) substantially in the $PCZ (INDO.) regions from period 1 to period 2. Taken togetherwith other results cited above, this sequence of eventssuggests that the intraseasonal MJO may have had animpact on modulating the SPCZ.1962 MONTHLY WEATHER REVIEW OLR (Win'~) 10-24 JAN 79O"E 60 120 EISOeW 120 60VOLUME 122 OLR Wn~2). 26 JAN- 7 FEB 79 l O~E: 60 120 FtG. 18. Outgoing longwavc radiation (W m-2). Shaded regions axe values less than or equal to225 W m 2, indicative of dccD cumulus convection in the Tropics. The upper panel represents24 January 1979, and the lower pane! represents :26 January-? February 1979 (extracted fromPedigo and Vincent 1990). In still another study, von Storch and Xu (1990) useda statistical approach, known as principal oscillationpattern (POP) analysis, to investigate the influence ofthe MJO in the tropical atmosphere. They found that aPOP, corresponding to a 30-60-day wave, contained amarked annual cycle and accounted for 70% of the variance in the upper-troposphere velocity potential as ittraversed the SPCZ. In this context, Vincent et al.(1991b) found a high correlation in the MJO band between velocity potential at 200 hPa and OLR along thezonal portion of the SPCZ in the summer of 1984/85.In the following summer, however, the correlation wasconsiderably lower, showing that the velocity potentialis not always a good indicator of SPCZ convection.Q~ AND Q2BUDGET ESTIMATES OFAREA-AVERAGED PRECIPITATION RATES (mmh'l)FOR SELECTED REGIONS BETWEEN 2N-30S(DERIVED FROM GLA LEVEL 1Tl-b ANALYSES) PERIOD I: 10-24 JAN 79PERIOD2: 28 JAN-II FEB79 S.AFR IND 0. AUST SPCZ SACZ OTHER SH TROP NHTROP 5-40E 50-100E 110-180E 175-135W 70-25W GLOBAL GLOBAL I0.3 ~ 0.30.2 '--] 0.20.1 ~ 0.1o.o I--I-I I--l-1 o.o0.4 - ~'2 ''"~" I 0.40.3 -~_~ ~"-ll ~ 0.30.2- I I I 0.2o.o I ~ ~ o.~ 0.0 I 2 I 2 I 2 I 2 i 2 I Z I 2 I 2 FIG. 19. Histogram of precipitation rate estimates (ram h-~) for selected regions in the SouthernHemisphere Tropics. Also shown are the global values for the Northern and Southern HemisphereTropics (extracted from Pedigo and Vincent 1990).SEPTEMBER 1994 V I N C E N T 1963 In a study of the quasi periodicity of the motion ofthe SPCZ cloud band, $treten (1978) found a preferredperiodicity in the east-west movement of the easternportion of the band (between 20- and 40-S) of about25 days. He used 5-day averages of satellite imageryfrom 1968 to 1971, and noted that the cloud bandwould undergo several 5-day periods of eastwardmovement followed by a rather abrupt westward movement back to its original position. Figure 20 illustratesthese eastward propagations and recurring westwardreversals. Note that 30% of the reversals occur near30-S with a period of about 25 days. Streten (1973)previously suggested that the eastern sector of theSPCZ cloud band was related to midlatitude trough intrusions into the subtropical South Pacific. The sameconclusion was reached by several other authors (e.g.,Webster and Curtin 1975; Trenberth 1976, 1991a; Streten and Zillman 1984; Kiladis and Weickmann 1992b). We now turn our attention to interannual variabilityof the SPCZ. The primary circulation features of interest here are El Nifio, La Nifia, and the SO. The relationship between these phenomena and the SPCZ hasbeen discussed in detail by several authors (e.g., Trenberth 1976, 1991a; Rasmusson and Carpenter 1982;Trenberth and Shea 1987; van Loon and Shea 1987;Meehl 1987; Philander 1989). Although El Nifio andthe SO do not always occur in conjunction with oneanother (Trenberth and Shea 1987), they usually do andare referred to as ENSO. In general, during an E1 Nifio(warm) event, the SO is negative (i.e., surface pressuresare lower than normal over the South Pacific and higherthan normal over northern Australia). Figure 21, whichappears in Rasmusson (1985) and Trenberth and Shea(1987), shows a composite of the correlations of annualmean sea level pressures with those at Darwin. Notethat the average position of the diagonal portion of theSPCZ (dashed line) is about midway between the centers of positive and negative action. This implies thatthe SPCZ is the focal point for ENSO events. That is,during an E1 Nifio, the SPCZ is generally north and eastof its average position, and the pressures are lower overthe central South Pacific. Figure 22 shows the MSLPdistribution for January 1987, which was in the middleof an El Nifio period. By comparing this pattern withthe January climatology shown in Fig. 2, it is clearlyseen that the pressure during El Nifio is slightly lower(considerably higher) than normal over the tropicalcentral South Pacific (northern Australia) and that theSPCZ surface trough is northeast of its average position. During a La Nifia (non-E1 Nifio or cold) event,the SO is usually positive and the SPCZ lies south andwest of its average position. Figure 23 shows the MSLPmap for January 1989, which was in a La Nifia period.Note the higher (slightly lower) pressures over the central Pacific (northern Australia) and the more southwestern position of the SPCZ surface trough. Another example of the changing circulation features that can occur in the SPCZ region during E1 Nifio 2O -E t0LONG. t0 20 30 40 50 60 ?0 80 g0 100 I I I I I I I I I I I I I I I I I I.I I I I |NOV. 10 1968 HAR. 20 1969 DAYS 20 20 20 tO 20 20 10 ~ $ tO I~ 20 25 ~0 ~S ~0 OAYS F~G. 20. Motion of the South Pacific cloud band axis. Vectors showeastw~d or westward motion of the band in degrees of longitude at40-S. Displacements ~e relative to the band axis location for ~eprevious 5-day average imagery (upper). Percentage of recu~encesof westw~d reversals at indicated time intervals (days) for specifiedlatitudes (lower) (extracted from S~eten 1978).and La Nifia events is illustrated in Fig. 24. This showsdepartures of the zonal wind component at 200 hPafrom their climatological mean for the 3-month seasonsof March-May 1987 (an El Nifio) and December1988-February 1989 (a La Nifia). It is seen that ananomalously strong upper-level anticyclonic (cyclonic)circulation occurs over the central South Pacific duringEl Nifio (La Nifia). These patterns are compatible withthe observed fact that more (less) convective outflowtakes place over the South Pacific in an El Nifio (LaNifia). Figure 25, extracted from Meehl (1987), provides a convenient summary of the features associatedwith the two extremes of ENSO events describedabove. The strong annual cycle corresponds to a LaNifia, while the weak annual cycle is representative ofan El Nifio.1964 MONTHLY WEATHER REVIEW VOLUME 122 Correlltlone of Annual Mean Sea Level Preaaurea with Darwin20- W0*E 20' 40' 60' 60* 10<7 ~20- 140- 160' 180- 160- 140. 120' 100. e0* 60- 40-~0- ~0o 60-40* 40-20- 20*0- 0.20' 20-40' 40*60' ,60-20'W0-E 2(7' 40* 60* 60* 100. 120" 140' 160. 160- 160' 140- 120- 100- 60* 60- 40* FiG. 2 ]. Composite assessment of the correlations of annual mean sea level pressures with Darwin (extractedfrom Trenberth and Shea 1987). The average position of the axis of the diagonal portion of SPCZ cloud bandhas been added (heavy dash). Rasmusson and Carpenter (1982) examined thechanging characteristics of SST, MSLP, and precipitation anomalies during six E1 Nifio events. They gavea detailed description of the sequence of these variablesfor each of the five phases (antecedent, onset, peak,transition, and mature) of E1 Nifio, based on composited results. They noted that their analyses should beinterpreted with caution and that they may be biasedby the stronger of the six events. Nevertheless, theycomment that their results are consistent with a numberof statistical and case studies by other authors. Of importance to this paper is the fact that they found a transition in the pattern of variables and a shift in the position of the SPCZ during each of the E1 Nifios, whichare compatible with the results discussed above.6. Synoptic-scale features Because the SPCZ is influenced both by tropicalheating, moisture convergence, and extratropical waveactivity, it is a region where synoptic-scale circulationsystems frequently are spawned. The zonal portion ofthe SPCZ is where tropical cyclones are often observed,particularly during the Australian summer monsoonseason. In the diagonal portion.of the SPCZ, between20- and 40-S, a variety of synoptic-scale features arepresent and appear to be responsible for the maintenance of the cloud band. Thus, the focus here will beon those features that include cyclones, anticyclones,wave disturbances, fronts, upper-air troughs and ridges,and jet streaks, as well as some of the interactions thatoccur among these phenomena. Several early investigations, based on 1957-58 International GeophysicalYear data (Taljaard 1964, 1965, 1967; Taljaard and vanLoon 1962, 1963) and satellite imagery (Troup andStreten 1972; Streten and Troup 1973; Streten 1975),provided a climatology of cyclones and anticyclonesfor the Southern Hemisphere. These studies illustrated40S140E ID 140W FiG. 22. As in Fig. 2 except for January 1987.100W140E ID 140W FIG. 23. As in Fig. 2 except for January 1989.IOOWSEPTEMBER 1994 V I N C E N T 1965ao20N20Suary). Figure 26 shows averages of the MSLP and OLR~< 225 W m-2 for the 10-18, 19-24, and 25-27 January 1979 periods. The change in the pattern betweenthe first two periods occurred on synoptic timescalesand was due to a steady buildup of high pressure in theeastern Pacific. The central pressure rose from 1015hPa on 18 January to 1027 hPa on 23 January. A similar, but more dramatic change occurred between thelatter two periods in the central South Pacific (near20N'~~~~~' J : (J) STRONG ANNUAL CYCLE .. . _ . ., . ...... ,., ..... 100~ 120E 140E 160E 180 160W 140W :WA '" ~ ~ "' '~Fro. 24. Anomalous departures from 1985-90 average of the zonal ' '~, ' -'w indcomponent(ms-')at200hPa for(a) March-May1987 and(b) December 1988-February 1989.that a l~ge number of cyclones develop ~d propagate ~~"~~,~..~sou~eastw~d over the South Pacific Oce~ (e.g., S~e- . ......... /2~;.......~t~ ..........? ........... .~...ten and Troup 1973) in what is now ~own as ~e SPCZ.For the most p~, however, detailed ~alyses of the flowfeatures associated with ~ese cyclones have been lacking, phmmly due to sp~sity of obse~ational data. Oneexception to this occurred during SOP-l, FGGE, whenample data were available in the SPCZ region to providean analysis of cyclone activity and related synoptic-scalefeatures (Vincent 1982, 1985). The author and his colleagues authored several papers that were devoted todiagnosing these features during SOP-1. Their resultsare summarized herein. Before proceeding to a discussion of their papers, however, it is important to mentionthat Trenberth (1991b) wrote an excellent paper onstorm tracks in the Southern Hemisphere. His primaryfocus was on storms at higher latitudes and he offeredminimal reference to SPCZ storms. Furthermore, asstated by Streten and Troup (1973) and many other authors, cyclonic disturbances in the South Pacific have alarge variability both in time and space. Perhaps this iswhy the storms studied by Trenberth did not provide astrong signal in the SPCZ region. Vincent (1985) performed a diagnosis of the life cycles of three cyclones in the SPCZ that occurred from10 to 17 January 1979. During this period, the SPCZand its cloud band were quasi-stationary and conditionswere favorable for cyclone development (Huang andVincent 1983). Shortly afterward, however, majorchanges took place in the large-scale circulation patternacross the Pacific, and the SPCZ shifted westward ( 1924 January) and weakened considerably (25-27 Jan WEAK ANNUAL CYCLE - :."-' .... ............ o% 'q": ...... f k~ ............ .... .:.,,/- - FIG. 25. Schematic diagram illustrating processes that evolve during (a) a strong annual cycle--also can accompany an extreme coldevent or La Nifia, and (b) a weak annual cycle--also can accompanyan extreme warm event or El Nifio (extracted from Meehl 1987).1966 MONTHLY WEATHER REVIEW VOLUME 122ON 170E 180 170W 160 150 140 130 120 I lOWIONION0 .oE .owloft .---'~o";"'"' ' ~ ..... '(c)'Ps imbi - ION o 26 ~ 170E ~ 17~ ~ I~ ~ I~ 120 IIOW Fro. 26. Awragc s~a level pressure (rob) for (a) 10-18 Janu~y,(b) 19-24 Janua~, and (c) 24-27 Janu~y 1979, together with OLR~ 225 W m-~ (shaded). This figure was compiled from diagramspresumed by Huang and Vincent (1983).10S! 20 3040S160-W) and was due to a cold-air outbreak from midlatitudes. There, the pressure rose from 1017 hPa on 22January to a plateau value of almost 1030 hPa from 25to 27 January. These features are seen in Fig. 27, whichshows a time series of twice-daily central pressures ofthe two high pressure areas. Thus, as noted earlier,changes in the location and strength of the diagonalportion of the SPCZ may be related to variations in theeastern Pacific high, as well as to intrusions of polarair into the subtropics near the date line. Returning to -incent's (1985) study, the tracks ofthe three cyclones are shown in Fig. 28. It is seen thatone cyclone remained in the zonal portion of the SPCZ,while the other two propagated southeastward along thediagonal portion. The track of each cyclone seems tobe influenced by the upper-tropospheric flow, as indicated by the daily maps of 200-hPa streamlines andisotachs shown in Fig. 29. Note that L2, which deepened as it propagated toward midlatitudes (Fig. 28),was located in the exit region of a transient jet streak,whereas L3, which weakened considerably as it propagated poleward, was located in the entrance region ofthe same jet streak. Vincent also showed that L2 contained more convective activity and precipitation thanL3. It is worth noting that these three cyclonic disturbances were all active on 12 January, which was oneof the dates quoted by Davidson and Hendon (1989),in their paper referred regarding the origin of the SPCZ,that corresponded to a downstream strengthening of theSouth Pacific vortex. An important finding in Vincent's study was that L2and L3 exhibited reasonably strong frontal characteristics as far equat0rward as 27.5-S, even though it wasmidsummer (i.e., sun nearly overhead) and they werelocated over the open sea. Evidence of this is seen inFig. 30, which shows a time-height section of windsand vorticity at a grid point near the island of Rapa(station 958 in Fig. 28), as well as the surface reportsfrom Rapa. As the cyclones pass by this point, warm(cold-) air advection occurs ahead of (behind) them, asindicated by the backing (veering) of the 1000-700hPa winds. Both cyclones exhibit upper-troposphericanticyclonic (cyclonic) vorticity ahead of (behind) theirpassage near Rapa. There is also enhanced cyclonicvorticity in the lower troposphere as each cyclonepasses this location. The surface conditions at Rapashow increasing (decreasing) clouds and rain, and awind shift as each cyclone approaches (departs). Each of the features described above supports the general characteristic of the SPCZ alluded to in early sectionsof this paper, namely, that it is highly baroclinic. Robertson et al. (1989) examined the role of diabatic heatingand other processes in maintaining this baroclinic zoneand concluded that 1) a major balance existed betweenthe frontogenetical contribution by differential diabaticheating and the opposing diabatic tilting processes, and2) in contrast to midlatitude cyclones, the adiabatic contributions from the deformation and tilting terms were oflesser importance. Their findings were compatible withthe eddy energy budget results of Hurrell and Vincent(1987) discussed in section 4. The study by Robertson et al. (1989), together with those by Hurrell and Vincent (1990, 1991), address a very important issue with regard to forcing mechanisms of synoptic-scale systems that occur in the diagonal portion of the SPCZ. They state that it was difficult for them to say whether outflow from the cyclone's dia batic heating field caused an enhancement of the upper level winds (and, thus, established a baroclinic zone) or whether the jet streaks set up a baroclinic zone that was favorable for cyclogenesis and the attendant dia batic heating. It appears, therefore, that a complicated994 V I N C E N T 1967CENTER PRESSURE TRACES 10~2-- ,~ ~/! ~ , ~ ~ ~o2e J ~ / - '~'"x E,~STE.. ,~'~ ~, ~ ,/ - ,OE4~ '~ ,~ ~a'.. . ..... '~'.. / .~ 5 -. ,. ,, , ',, -- 2 ~ ~' ~. ~ i "x~ ~_.~i ~ ~ - ~ .,- -~ -~ '~ ~ .--/ ~ g__.(~) 1016~b --~~ /~--c ) ~ "~ [ /~'1~u~) ~-' ~ --- . ~ WESTERN HIGH ~ 1012 -- - i I I I .L .! t I I i I ~ , ~ ~ I I I I ~o~n ~ ~2 ~ ~ 15 ~ 17 m 19 ~ a~ ~a ~ ~4 ~ ~ 27 ~o~n ~ ~T ~G~T ~. ~. ~imc sc~cs o~ twice-daily mc~n sea ]cvc~ p~cssum (rob) ar the ccmc~s of ~hc quasi-stafo~y casrcm P~c~fic h~ (~)a~ migrating highs tn rhc ccntm~ to wcs~c~ Pacific (W) ~or rhc pc~o~ 1~-~ January 1979. ~hc~ ~incs indicate whc~c ~csrc~high was just oursi~c (somh) of rhc analysis area.- 10;52 028 !102~ O20 016 012two-way interaction between heating and wind fieldsexists in the SPCZ. Furthermore, as was suggested insection 3, the cause of the SPCZ jet streaks is not wellunderstood but is most likely due to a combination oftropical and midlatitude influences.7. Concluding remarks and challenging problems In this paper, an attempt has been made to reviewsome of the knowledge concerning the SPCZ. Topicsdiscussed were the structure and characteristics of theSPCZ; possible causes of its location, origin, and maintenance; its role in global-scale circulations, some observations concerning quasi-periodic fluctuations in itslocation and strength; and a few of the synoptic-scalefeatures that occur within its domain. In summary, itwas suggested that the processes that maintain andchange the zonal (tropical) portion of the SPCZ aremost likely different from those that are responsible forthe diagonal (subtropical and lower midlatitude) portion. In this regard, several theories and hypotheses15*S202530 *JANUARY 1979 35 I I tTO-W 165 160 I O~1 L I ~93~loo4~ )~ I~::).~,26,,~oo~ a~'~o~TM830 M 843 ~ 944954-~ 998~x~,,..,L :~II -58 1008 ~ I00~ , 150 145 140155135 FiG. 28. Cyclone tracks of L~, L2, and L3 during January 1979.Open circles depict locations at 1200 UTC, closed at 0000 UTC. Alsoshown are central pressures at 1200 UTC and locations of islandstations (crosses) in vicinity of tracks (extracted from Vincent 1985).were discussed. It was documented that the SPCZ is animportant part of, and often contributes to and interactswith, circulation features across the Pacific Ocean, bothin the Tropics and extratropics. Furthermore, it appearsto play a role in the cross-equatorial flow. It was alsoshown that interactions between the SPCZ and otherlocations occur on a variety of timescales from synopticto interannual. It appears, therefore, that a better understanding of the SPCZ would lead to improvementsin weather forecasting and climate prediction. There are a number of challenging problems, someof which, if solved, could enhance our knowledge ofthe SPCZ. For example, there is still much that needsto be learned about the relative importance of processesand mechanisms that are responsible for the locationand strength of the $PCZ and its circulation features.In the tropical (zonal) portion, the roles of warm SSTs,SST gradients and latent heat release need to be betterunderstood. In this regard, one of the primary issues isrelated to the following question: How does the SPCZrespond to, and interact with, low-frequency phenomena such as the intraseasonal oscillation and ENSOevents? One of the main contributors to this paper, Dr.Kevin Trenberth, suggested that a good climatology ofthe SPCZ and changes in ENSO would be useful. Asdiscussed in the text, the SPCZ generally lies north andeast of its "average" position during an El Nifio. Itwas pointed out by Vincent (1982), however, that from10 to 18 January 1979 the SPCZ was located north andeast of its normal position, even though this was not anEl Nifio period. On the other hand, Huang and Vincent(1988) and Hurrell and Vincent (1990) speculated thatthe SPCZ region was under the influence of the convectively active part of an intraseasonal oscillation during 10-18 January. Thus, as indicated above, a betterunderstanding of the relative roles of phenomena onlow-frequency timescales needs to be achieved. In the subtropical and lower midlatitude (diagonal)portion of the SPCZ, more knowledge is required concerning the importance of tropical versus higher-lati1968MONTHLY WEATHER REVIEW5'S I0 15 '~5303540200rnb FLOW 0000 GMT II JAN 79-,,'-c~',~ \ 'V ' /- n,~-,o-7 ./,~ ~, ,,5'S I0 15 20 25 30 3540200mb FLOW'O000 GMT 12 JAN 79:"-'"-~%-',~ 't ' l /'. .-~~. ~,__ ~j f ~', 180~'W5* $ I0 15 20 t)5 $0 3,5 40 170 160 150 140 130 125'W 180"W 170 160 150 140 130 12SEW200rnb FLOW 0000 GMT I$ JAN 79 200rnb FLOW 0000 GMT 14 JAN 795* $180-W 170 160 150 140 130 12~ 200mb FLOW 0000 GMT 15 JAN 79 5- S! I0 15 ZO 25 3,0 35 40180-W 170 160 150 140 130 125'~ 200mb FLOW 0000 GMT 16 JAN 79 '.-,o-~'~-;,'~-~ 'l L'x~I' - ~'\ ( %. \x., ' ~ ~ ~ ~ ~ ~ '~. ~~/'~ ~ ~3~ N~o~ ] ~ ~ ~ . ~zo~ x o' ~~~/:~ iX RX~.-~ ~-~/'./ ! 'Y '~ % .~~J ~~ ~ " ' ~,~. %~ _1 --- ,~ X ~'~ ~. ~.~q180-W 170 160 150 140 rso 125',w 180ow 170 160 150 140 r'Jo 125~t/ FIG. 29. Streamlines and isotachs (ms ~) at 200 hPa for six consecutive 0000 UTC times beginning on ] l January 1979. Locations of three cyclones also denoted (extracted from %/inccnt 1985).VOLUME 122tude influences. In this regard, most of the evidence provided in this paper came from analyses during SOP-l,FGGE. Further studies need to be conducted that addressthe following questions: 1) What are the respective rolesof tropical heat sources, the subtropical high over the eastern Pacific, and intrusions of cold polar air into the westem and central subtropical Pacific, with regard to changing the circulation features within the SPCZ? and 2) Isthere a relationship or teleconnection between the secondand third factors in question 1 ? It would also be useful tohave a better understanding of the interactions betweendiabatic (and adiabatic) processes in a convective disturbance and the upper-tropospheric jet that usually accompanies the disturbance. Another problem worthy of consideration is the linkbetween the circulation patterns over the Indian Oceanand the SPCZ. As mentioned in the text, there appearsto be an out-of-phase relation between these tworegions such that when one is convectively active theother is not. This relationship, which is generally attributed to intraseasonal oscillations, is but one of several that can alter the location and strength of the eastwest circulation across the Indian-Pacific Ocean tropical corridor. In this context, it appears that the keylocation to a better understanding resides in the warmpool region of the western Pacific. This region undoubtedly makes major contributions to, and interactswith, the SPCZ and other circulation features in its vicinity (e.g., Australian monsoon). It is believed, therefore, that results from recent field programs, such asthe Australian Monsoon Experiment and the TropicalOceans Global Atmosphere Coupled Ocean-AtmoSEPTEMBER ] 994 V I N C E N T 1969 2Z5- S, 145.0-W WINDS, I$OTACHS (miI) ,oo.~~~wx~ovC. & ~ ,oo ~10 I0 .so-~-~~5 ~ ~ /~ I000 I v /'~ m ~~ r '~ '~ ~ m~ ~' ~ ' I0/00 II/00 12/00 I~/00b 14/00 15/00 16/00 I~/00 I~ W~ ~1~ CA - I W~ I( CA~ L~ Ls ...... a., t "~-" ~ ~" ~;d~ ~ "~ "e~ ~ ~ ~ ~oO, so X ~ X~/[~,-~ l 200 250 ~00 P (rob) 400 500 ,oo 850 tO00 ~ i ~ I0/00 II/00 I~/00 13/00 14/00 15/00 16~0 17/00 FiG. 30. Time-height cross sections of winds (upper panel) ~drelative vo~icity (lower panel) at g~d point 27.5-S, 145.0-W. Isotachs ~c in meters per second, and a full b~b is 5 m s~. Also shown~c the passages of L2 and L~ and the pc~ods of implied lower-tropospheric w~-air (WA) and cold-air (CA) advection (extractedfrom Vincent 1985). In addition, the surface observations from islandstation Rapa (958 in Fig. 28) ~e given.RRPAsphere Response Experiment, will provide answers tosome of the questions and problems posed above. Acknowledgments. This paper could not have beenwritten without the assistance of several of the author'sscientific colleagues at other institutions. The authorsincerely thanks the following persons who answeredhis request for materials and information and, thus,made it possible for him to compile and present thisreview of the SPCZ: Drs. A. Gruber, G. Kiladis, R.Madden, N. Streten, K. Trenberth, and H. von Storch.He also thanks the two referees for their helpful suggestions that he feels greatly improved the manuscript-Messrs. Doug Miller and Jon Schrage for creating the figures, and Ms. Helen Henry for helping withcorrespondence and typing the paper. This researchwas supported primarily by the National Aeronauticsand Space Administration under Grant NAG8-836,with some assistance provided by the National ScienceFoundation under Grant ATM-9200534, both issued tothe author at Purdue University. REFERENCESAtkinson, G. D., and J. C. Sadler, 1970: Mean cloudiness and gra dient level wind charts over the tropics. Vol. II. Tech. Rep. 215, II, Air Weather Service, Scott Air Force Base, 1L, 48 pp.Bergeron, T., 1930: Richtlinien einer dynamischen Klimatologie. Meteor. Z, 47, 246-262.Davidson, N. E., and H. H. Hendon, 1989: Downstream development in the Southern Hemisphere monsoon during FGGEdWMO NEX. Mon. Wea. Rev., 117, 1458-1470.Gmber, A., 1972: Fluctuations in the position of the ITCZ in the Atlantic and Pacific Oceans. J. Atmos. Sci., 29, 193-197.Huang, H.-J., 1985: Significance of the South Pacific convergence zone in energy conversions of the Southern Hemisphere during FGGE, 10-27 January 1979. Mon. Wea. Rev., 113, 1359-137 I. , 1988: Active and inactive phases of the South Pacific conver gence zone and changes in global circulation patterns: A case study. Beitr. Phys. Atmos., 61, 123-134.--, and D. G. Vincent, 1983: Major changes in circulation features over the South Pacific during FGGE, 10-27 January 1979. Mon. Wea. Rev., 111, 1611-1618.Hubert, L. F., 1961: A subtropical convergence line of the South Pacific: A case study using meteorological satellite data. J. Geo phys. Res., 66, 797-812.Hurrell, J. W., and D. G. Vincent, 1987: Significance of the South Pacific convergence zone (SPCZ) in the energy budget of the Southern Hemisphere tropics. Mon. Wea. Rev., 115, 1797-1801.--, and --, 1990: Relationships between tropical heating and subtropical westerly maxima in the Southern Hemisphere during SOP-l, FGGE. J. Climate, 3, 751-768.--, and ,1991: On the maintenance of short-term subtropical wind maxima in the Southern Hemisphere during SOP-l- FGGE. J. Climate, 4, 1009-1022.--, and --., 1992: A GCM case study on the maintenance of short-term subtropical wind maxima in the summer hemisphere during SOP-l, FGGE. Quart. J. Roy. Meteor. Soc., 118, 51-70.Kalnay, E., K. C. Mo, and J. N. Paegle, 1986: Large-amplitude, short scale stationary Rossby waves in the Southern Hemisphere: Ob servations and mechanistic experiments to determine their ori gin. J. Atmos. Sci., 43, 252-275.Kiladis, G. N., and K. M. Weickmann, 1992a: Circulation anomalies associated with tropical convection during northern winter. Mon. Wea. Rev., 120, 1900-1923. , and ,1992b: Extratropical forcing of tropical Pacific con vection during northern winter. Mon. Wea. Rev., 120, 1924 1938. - H. yon Storch, and H. van Loon, 1989: Origin of the South Pacific convergence zone. J. Climate, 2, 1185-1195.Knutson, T. R., and K. M. Weickmann, 1987:30-60 day atmo spheric oscillations: Composite life cycles of convection and circulation anomalies. Mon. Wea. Rev., 115, 1407-1436.Kodama, Y.-M., 1992: Large-scale common features of subtropical precipitation zones (the Baiu frontal zone, the SPCZ, and the SACZ). Part I: Characteristics of subtropical frontal zones. J. Meteor. Soc. Japan, 70, 813-835.--, 1993: Large-scale common features of subtropical conver gence zones (the Baiu frontal zone, the SPCZ, and the SACZ). Part II: Conditions of the circulations for generating the STCZs. J. Meteor. Soc. Japan, 71, 581-610.Krishnamurti, T. N., M. Kanamitsu, W. J. Koss, and J. D. Lee, 1973: Tropical east-west circulations during the northern winter. J. Atmos. Sci., 30, 780-787.Lau, K.-M., and P. H. Chan, 1985: Aspects of the 40-50 day oscil lation during the northern winter as inferred from outgoing long wave radiation. Mon. Wea. Rev., 113, 1889-1909.1970 MONTHLY WEATHER REVIEW VOLUME 122--, and --., 1988: Intraseasonal and interannual variations of tropical convection: A possible link between the 40-50 day os cillation and ENSO? J. Atmos. Sci., 45, 506-521.Lindzen, R. S., and S. Nigam, 1987: On the role of sea surface tem perature gradients in forcing low level winds and convergence in the tropics, J. Atmos. Sci., 44, 2418-2436.Madden, R. A., and E. J. Zipser, 1970: Multi-layered structure of the wind over the equatorial Pacific during the Line Islands Exper iment. J. Atmos. Sci., 27, 336-342. , and P. R. Julian, 1971: Detection of a 40-50 day oscillation in the zonal wind in the tropical Pacific. J. Atmos. Sci., 28, 702 708. , and , 1972: Description of global-scale circulation cells in the tropics with a 40-50 day period. J. Atmos. Sci., 29, 1109 1123.Manabe, S., J. L. Holloway Jr., and H. M. Stone, 1970: Tropical circulation in a time-integration of a global model of the at mosphere. J. Atmos. Sci., 27, 580-613.Matsuno, T., 1966: Quasi-geostrophic motions in the equatorial area. J. Meteor. Soc. Japan, 44, 25-42.Meehl, G. A., 1987: The annual cycle and interannual variability in the tropical Pacific and Indian Ocean regions. Mort. Wea. Rev., 115, 27-50.Miller, B. L., and D. G. Vincent, 1987: Convective heating and pre cipitation estimates for the tropical South Pacific during FGGE, 10-18 January 1979. Quart. J. Roy. Meteor. Soc., 113, 189 212.Nitta, T., 1972: Energy budget of wave disturbances over the Mar shall Islands during the years 1956 and 1958. J. Meteor. Soc. Japan, 50, 71-84.Pedigo, C. B., and D. G. Vincent, 1990: Tropical precipitation rates during SOP-l, FGGE, estimated from heat and moisture budg ets. Mon. Wea. Rev., 118, 542-557.Philander, G., 1989: El Nifio and La Nifia. Amer. Sci., 77, 451-459.Rasmusson, E. M.i 1985:E1 Nifio and variations in climate. Amer.Sc-, 73, 168-177. - and T. H. Carpenter, 1982: Variations in tropical sea surface temperature and surface wind fields associated with the South em Oscillation/El Nifio. Mon. Wea. Rev., 110, 354-384.Reed- R. J., and E. E. Recker, 1971: Structure and properties of syn optic-scale wave disturbances in the equatorial western Pacific. J. Atmos. Sci., 28, 1117-1133.Robertson, F. R., D. G. Vincent- and D. M. Kann; 1989: The role of diabatic heating in maintaining the upper-tropospheric baro clinic zone in the South Pacific. Quart. J. Roy. Meteor. Soc., 115, 1253-1271.Sadler, J. C., 1972: The mean winds of the upper troposphere over the central and eastern Pacific. Tech. Rep. UHMET-72-04, Uni versity of Hawaii, Honolulu, HI, 35 pp. , 1975: The upper tropospheric circulation over the global trop ics. Tech. Rep. UHMET-75-05, Department of Meteorology, University of Hawaii, Honolulu, HI, 35 pp. - M. A. Lander, A. M. Hori, and L. K. Oda, 1987: Tropical marine climatic atlas: Vol. II, Pacific Ocean. Tech. Rep. UHMET-87-02, Department of Meteorology, University of Ha Waii, Honolulu, HI, 27 pp.Song, J.-L., and W. M. Frank, 1983: Relationship between deep con vection and large-scale processes during GATE. Mon. Wea. Rev., 111, 2145-2160. Streten, N. A., 1973: Some characteristics of satellite-observed bands of persistent cloudiness over the Southern Hemisphere. Mort. Wea. Rev., 101, 486-495. , 1975: Satellite-derived inferences to some characteristics of the South Pacific circulation associated with the E1 Nifio event of 1972-73. Mon. Wea. Rev., 103, 989-995. , 1978: A quasi-periodicity in the motion of the South Pacific cloud band. Mort. Wea. Rev., 106, 1211-1214. , and A. J. Troup, 1973: A synoptic climatology of satellite observed cloud vortices over the Southern Hemisphere. Quart. J. Roy. Meteor. Soc., 99, 56-72.--, and J. W. Zillman, 1984: Climate of the South Pacific Ocean. WorldSurvey of Climatology, H. van Loon, Ed., Elsevier, 263-375.Taljaard, J. J., 1964: Cyclogenesis, cyclones and anticyclones in the Southern Hemisphere during autumn. Notos, 13, 31-36.--., 1965: Cyclogenesis, cyclones and anticyclones in the Southern Hemisphere during the period June to December. Notos, 14, 73-84.---, 1967: Development, distribution and movement of cyclones and anticyclones in the Southern Hemisphere during the IGY. J. Appl. Meteor., 6, 973-987.--, and H. van Loon, 1962: Cyclogenesis, cyclones and anticy clones in the Southern Hemisphere during the winter and spring of 1957. Notos, 11, 3-20.--, and --, 1963: Cyclogenesis, cyclones and anticyclones in the Southern Hemisphere during the summer of 1957-1958. Notos, 12, 37-50.Thompson, R. M., Jr., J. W. Payne, E. E. Recker, and R. J. Reed, 1979: Structures and properties of synoptic-scale disturbances in the intertropical convergence zone of the eastern Atlantic. J. Atmos. Sci., 36, 53-72.Trenberth, K. E., 1976: Spatial and temporal variations of the south ern oscillation. Quart. J. Roy. Meteor. Soc., 102, 639-653.--, 1986: An assessment of the impact of transient eddies on the zonal flow during a blocking episode using localized Eliassen Palm flux diagnostics. J. Atmos. Sci., 43, 2070-2087J--, 1991a: General characteristics of El Ni~o-Southern Oscilla tion. Teleconnections Linking Worldwide Climate Anomalies. M. Glantz, R. W. Katz, and N. Nicholls, Eds., Cambridge Uni versity Press- 13-41.--., 1991b: Storm tracks in the Southern Hemisphere. J. Atmos. Sci., 48, 2159-2178.---, and D. J. Shea, 1987: On the evolution of the Southern Oscil lation. Mon. Wea. Rev., 115, 3078-3096.Troup, A. J., and N. A. Streten, 1972: Satellite-observed Southern Hemisphere cloud vortices in relation to conventional observa tions. J. Appl. Meteor., 11, 909-917.van Loon, H., and D. J. Shea, 1987: The Southern Oscillation. Part VI: Anomalies of sea level pressure on the Southern Hemisphere and of Pacific sea surface temperature during the development of a warm event. Mon. Wea. Rev., 115, 370-379.Vincent, D. G., 1982: Circulation features over the South Pacific during 10-18 January 1979. Mon. Wea. Rev., 110, 981-993.--, 1985: Cyclone development in the South Pacific convergence zone during FGGE, 10-17 January 1979. Quart. J. Roy. Meteor. Soc., 111, 155-172.--, K. H. North, R. A. Velasco, and P. G. Ramsey, 1991 a: Precip itation rates in the tropics based on the Qi-budget method: 1 June 1984-31 May 1987. J. Climate, 4, 1070-1086.--, T. Sperling, A. Fink, S. Zube, and P. Speth, 1991b: Intrasea sonal oscillation of convective activity in the tropical Southern Hemisphere: May 1984-April 1986. J. Climate, 4, 40-53. yon Storch, H., and J. Xu, 1990: Principal oscillation pattern analysis of the 30- to 60-day oscillation in the tropical troposphere. Part I: Definition of an index and its prediction. Climate Dyn., 4, 175-190. , H. van Loon, and G. N. Kiladis, 1988: The Southern Oscilla tion. Part VIII: Model sensitivity to SST anomalies in the trop ical and subtropical regions of the South Pacific convergence zone. J. Climate, 1, 325-331. Webster, P. J., and D. G. Curtin, 1975: Interpretation of the EOLE Experiment II. Spatial variation of transient and stationary modes. J. Atmos. Sci., 32, 1848-1863. --, and H.~R. Chang, 1988: Equatorial energy accumulation and emanation regions: Impacts on a zonally varying basic state. J. Atmos. Sci., 45, 803-829. Weickmann, K. M., G. R. Lussky, and J. E. Kutzbach, 1985: Intra seasonal (30-60 day) fluctuations of outgoing longwave radia tion and 250 mb streamfunction during northern winter. Mon. Wea. Rev., 113, 941-961. Yanai, M., S. Esbensen, and J. H. Chu, 1973: Determination of the bulk properties of tropical cloud clusters from large-scale heat and moisture budgets. J. Atmos. Sci., 30, 611-627.