1. Introduction
Recent theories and models of ENSO development have centered on equatorial ocean–atmosphere coupling as sole explanators of Southern Oscillation variability (Latif et al. 1998; Fedorov and Philander 2000). However, the failure of operational forecasting models to predict the rapid development, intensity, and abrupt termination of the intense 1997/98 El Niño (Anderson and Davey 1998; Trenberth 1998; Barnston et al. 1999; Landsea and Knaff 2000) challenged this assumption. Subsequent analysis of the event provided support for the two main theories of the Southern Oscillation.
First, the Southern Osillation can be explained as a weakly damped oscillator that needs to be triggered by a random disturbance. Westerly wind bursts in the western equatorial Pacific appear necessary at the onset of El Niño as they are associated with 1) an eastward shift of warm water and convection, and 2) the generation of oceanic Kelvin waves that warm the central and eastern Pacific through zonal advection and displacement of the thermocline. The timing and amplitude of 1997 warming was attributed to exceptionally strong westerly wind events/Madden–Julian oscillation (MJO) activity, strong downwelling kelvin waves (McPhaden and Yu 1999; Picaut et al. 2002; Lengaigne et al. 2003), and a buildup of upper-ocean heat early in the year (Meinen and McPhaden 2000; Sun 2003). MJO activity was also proposed as a “triggering mechanism” that accelerated the ending of the event (Takayabu et al. 1999; Straub et al. 2006).
However, the influence of the MJO on ENSO remains a controversial topic (McPhaden 2004) as periods of enhanced westerlies and MJO activity do not always lead to El Niño (Chen and Wu 2000; Bergman et al. 2001; Zhang and Gottschalck 2002; Jones et al. 2004; Levinson 2005). Of the 13 strongest MJO events since 1979, only the two events late in 1996 and early 1997 actually preceded an El Niño (Jones et al. 2004). Slingo et al. (1999) found that the statistical relationship between ENSO and MJO indices do not show that they are causally related, and Bergman et al. (2001) speculated that the MJO might be relevant to the timing and initial growth of El Niño rather than be responsible for the event.
Second, the Southern Oscillation can be viewed as a lower-frequency self-sustaining mode of oscillation in the the tropical Pacific (Chen et al. 2004). Wang and Weisberg (2000) found that off-equatorial SST and associated pressure anomalies in the subtropics played an important role in the development and decay of the 1997/98 El Niño. However, what went unnoticed was a major seesaw in pressure in the southern midlatitudes (van Loon and Shea 1985, 1987), and a large-scale propagating SLP signal noted in previous El Niño events (Barnett 1985; Krishnamurti et al. 1986).
When Bjerknes (1966, 1969) linked maximum oceanic warming in the eastern equatorial Pacific (El Niño) to the “Southern Oscillation” in pressure, the primary cause for these events was attributed to an anomalous weakening of the trade winds of the Southern Hemisphere and associated oceanic upwelling (Bjerknes 1969). Since the South Pacific high dominates the equatorial flow in the central and eastern Pacific to 10°N, Bjerknes (1966) speculated that reasons for this warming must be sought south of the equator.
Of critical importance to ENSO dynamics is the inference by van Loon (1984) that the quasi-permanent South Pacific trough in the surface westerlies weakens both the South Pacific high and associated trades during the important months between May, June, and July (MJJ, hereafter 3-month periods are denoted by the first letter of each respective month). At this time, the trough and subtropical jet stream reach their farthest northward extent as part of the semiannual oscillation (SAO) in trough positioning and intensity (van Loon 1967; van Loon 1972; Hurrell et al. 1998). Over the year leading into El Niño, the South Pacific trough changes from a weakened to an enhanced state, that is, takes on a more permanent feature (van Loon 1984). Associated with this is a west-to-east pressure anomaly reversal or “seesaw” in the southern midlatitudes (Berlage 1966; van Loon and Shea 1985, 1987) and other important wind and SST anomalies (Trenberth 1976; Rasmusson and Carpenter 1982; Harrison 1984; Harrison and Larkin 1996, 1998). Figure 1 sumarizes this process, which we shall call the “van Loon hypothesis.”
In austral winter/spring in the year before El Niño (year −1), the annual cycle is weak in the southern midlatitudes, with a strong trough pattern over the region of the Australian subtropical high, and a weak trough in the South Pacific (Fig. 1a). Northerly wind anomalies between these two regions warm the Coral Sea and the southwest edge of the South Pacific convergence zone (SPCZ) as it expands south in austral spring (−1). Increased convection and lower pressures over this warm water then coincide with anomalous westerlies on the equatorward side of the northern SPCZ (Rasmusson and Carpenter 1982; Von Storch et al. 1988). At the southern end of the SPCZ, negative pressure anomalies continue to form over warm water and this is associated with an equatorward repositioning of the surface trough into the South Pacific (van Loon and Shea 1985; Fig. 1b). This whole process is linked to the convective maximum in the annual cycle gradually moving from southeast Asia into the Australian monsoon region (austral summer), and onto the southern end of the SPCZ (austral autumn; Meehl 1987).
By austral winter in the El Niño year (year 0) the annual cycle is enhanced in the southern midlatitudes. Stronger highs occur over Australia, and negative SLP and upper-level height anomalies become established in the central South Pacific along with an equatorward movement in the subtropical jet (Karoly 1989; Kiladis and Mo 1998). The first result of this pattern is strong southerly wind anomalies to the east of an enhanced Australian high pressure. These southerlies converge with and intensify westerly wind anomalies in the equatorial western Pacific and appear to play a determining factor in the onset time of El Niño (Harrison 1984; Mitchum 1987; Chen and Wu 2000; Xu and Chan 2001). Second, transients within the South Pacific storm track shift equatorward with the axis of maximum baroclinicity (Solman and Menendez 2002). The resultant westerly wind changes are a vital part of a gradual progression of negative SLP anomalies into the wider Pacific basin, that is, weakening of the South Pacific high (van Loon et al. 2003). The enhanced southwesterly airflow in the southwest Pacific cools the SST along the SPCZ (Fig. 1b), which then preceeds a strengthening of the South Pacific ridge (trades) and the demise of El Niño.
More recently, the focus has been on atmospheric variability and SST anomalies in the northern midlatitudes as a coupled component of the stochastic forcing of ENSO (Vimont and Battisti 2003; Anderson 2003, 2004; Thompson and Lorenz 2004; Anderson and Maloney 2006). The implication of these studies is that the Southern Oscillation and southern midlatitudes have a minimal role in ENSO, but is this the case?
We propose that the van Loon hypothesis becomes more pronounced in stronger El Niño events and that air–sea interactions between the southern midlatitudes and the equator are necessary for such events to develop. If the Southern Oscillation is principally a seesaw (or standing wave) in atmospheric mass exchange between the Western and Eastern Hemispheres (Berlage 1966; Trenberth 2000), then strong negative pressure anomalies observed in the Australian region in year −1 should be observed in the South Pacific as El Niño approaches. We tested this hypothesis by comparing the composite sequence of sea level pressure (SLP), vector wind, and SST anomalies leading into stronger and weaker El Niño events.
2. Data
The present study utilizes four types of data: 1) SLP, SST, surface winds, and 850-hPa winds. Maps showing the spatial evolution of SLP and surface wind anomalies associated with El Niño development were generated from the National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) reanalysis data from 1950 to 2003 (Kalnay et al. 1996). This has seasonal-mean surface fields of sea level pressure and low-level winds represented at a 2.5° resolution in both the meridional and zonal direction. The reanalysis data should be considered a blend of observational and model-simulated values.
To compliment the spatial analysis, monthly SLP data were obtained for individual stations across the Pacific Ocean and Australia. SST anomalies over the Niño-3 region in the eastern equatorial Pacific (5°S–5°N, 150°–90°W) and area-averaged 850-hPa trade wind index values for the western Pacific (5°N–5°S, 135°E–180°) were extracted from the Climate Prediction Center Web site (www.cpc.noaa.gov/data/indices). Area-averaged SST data over the central South Pacific (20°–30°S, 160°–120°W) from 1950 to 1998 were extracted from the Hadley Centre (Met Office) Global Sea Ice and Sea Surface Temperature dataset (GISST3), which is available on a 1° grid. Subsequently, the Reynolds SST dataset (RSST; 1950–99) was updated with the extended reconstructed SST (ERSST; Smith and Reynolds 2003) to map spatial composites of SST anomalies at a 2.5° resolution between 1950 and 2003.
The El Niño events chosen in this study followed the 11 year (0) events defined by Harrison and Larkin (1998), though we used a slightly less stringent criteria (Niño-3 seasonal SST anomalies exceeding +0.5 standard deviations for at least five months) to define 14 events up to 2003. When mapping spatial composites, we split these events into strong and weak categories on the basis of the Bjerknes ENSO index (BEI), a composite measure of eastern equatorial SLP, zonal wind, and meridional wind anomalies (Harrison and Larkin 1998). We assigned the six events with the highest BEI3 intensity (maximum intensity of event) as strong, that is, 1997, 1982, 1972, 1991, 1965, and 1957. The six events in 1951, 1953, 1963, 1976, 1994, and 2002 were classified as weak (note: we did not have a BEI3 value for 2002 but noted that other ENSO indices were weak). It was decided to seperately map the the late-developing 1968/69 and 1986/87 events, as the weak warming that developed late in the first year was associated with a different sequence of pressure anomalies compared with other events.
When examining the composite of SST anomalies, it must be noted that these were affected by warming beginning along the South American coast and spreading westward prior to 1979, whereas later events had warming progressing eastward from the western Pacific (Trenberth et al. 2002), that is, 3 (2) of the 6 (6) strong (weak) events had eastward warming. Since there is a small sample size, and differences between individual events, we will only comment on major differences between the composites.
3. Southeast Australian pressure
Acording to van Loon and Shea (1985), low pressure over southeastern Australia in year −1 appears to be a useful predictor of El Niño events 12–18 months later. As an independent test of this proposition, we correlated seasonal SLP averaged over the region 25°–35°S, 135°–145°E from JAS with Niño-3 SST the equivalent period a year later. Between 1985 and 2002 the correlation was –0.71, significant at the 99% level. This concurs with Anderson (2003), who found that a pressure index in this region was significantly related (r = −0.56) with the JFM Niño-3.4 index 18 months later.
In Fig. 2, SLP anomalies from Mildura (34.2°S, 142.1°E) in southeastern Australia are compared to those at Darwin closer to the equator (12.4°S, 130.9°E). As expected, the variability in SLP is greater at the higher-latitude station, but this is still dominated by ENSO. The largest negative values typically occur in the winter–spring in the year prior to El Niño. Over the following year, there is a dramatic rise in SLP (highlighted with straight dashed line) as the normal strengthening of the subtropical high over the continent is gradually enhanced. What is most striking is the pronounced pressure reversal leading into stronger El Niño events, especially in 1957, 1965, 1972, 1982, and 1997.
At Darwin, there is not such a close association with El Niño strength and preceding pressure anomalies. The pressure reversal from year −1 to year 0 is also less pronounced and less consistent. The correlation between SLP anomalies at Darwin, between July and September, and El Niño strength in the following year measured by BEI intensity was only r = 0.18, compared to r = 0.69 at Mildura. The question that results is as follows: does this low pressure anomaly centered over southeastern Australia play a role in the formation of ENSO events, and are larger low pressure anomalies associated with more pronounced tropical–extratropical interactions leading into stronger El Niño events?
4. Spatial aspects of El Niño development
We first compare the evolution of SST, SLP, and winds associated with the development of strong and weak El Niño events. In Fig. 3 spatial composites of developing SST anomalies were complimented with equivalent composites of SLP and vector wind anomalies in Fig. 4. Composites were plotted as a sequence of seasonal averages from MJJ (–1) through to FMA (+1). We plot actual anomalies because wind tends to flow downgradient along the equator, and roughly along isobars farther poleward, that is, the actual impact of large-scale pressure anomalies can be highlighted visually with wind stress forcing arrows. Also, ENSO monitoring bulletins (e.g., CPC 1997) plot data in this way as a standard.
To address the issue of field significance, SST and SLP anomalies were standardized and a Student’s two-tailed t test was applied. Regions that were significantly different at the 90% level are highlighted with a solid dashed line. When considering this approach, Harrison and Larkin (1998) noted that the Student’s t calculations and bootstrap method gave very similar results to the normal-z statistic in nearly every location. They also found the larger and more intense the composite signal, the more robust the signal tends to be.
a. Sea surface temperatures
For both stronger and weaker El Niños we found the same underlying sequence and pattern of SST anomalies from MJJ (−1) to FMA (+1) as those found by Harrison and Larkin (1998). Warm SST anomalies to the north and east of Australia gradually extend eastward and become most pronounced in the eastern equatorial Pacific as El Niño events begin to develop (Fig. 3). As the warming progresses eastward into the Pacific, SST anomalies become cooler in the Australian region, and in the characteristic horseshoe-shaped region in the western Pacific. SST anomalies along the SPCZ noticeably change sign from warm to cold between FMA (0) and MJJ (0).
When comparing SST anomalies between stronger and weaker events, the following features are apparent. First, the spatial extent and magnitude of SST anomalies are larger for strong events, although critically, the magnitude of Pacific warming is not indicated in year −1. Second, SST anomalies in strong events bifurcate around the equator in NDJ (0), before intensifying along the equator (160°E–140°W) in FMA (0), while for weaker events warm SST anomalies do not extend toward Hawaii and are cooler near the warm pool between NDJ (0) and FMA (0). Third, in the important months of MJJ (0) warming is weaker and centered farther east along the equator for weaker events.
It should be pointed out that warm SST anomalies along the equator or SPCZ do not always lead to El Niño, but require ocean–atmosphere interactions, which we will now elucidate.
b. Mean sea level pressure and wind anomalies
For both strong and weak El Niño events Fig. 4 confirms the large-scale eastward progression of low SLP anomalies from the Eastern Hemisphere to the Western Hemisphere from year −1 to year 0 identified by Barnett (1985). This generally coincides with the eastward progression in SST anomalies along the equator and subtropics (Fig. 3), and the movement in the convective maximum in the annual cycle (Meehl 1987). Similar to SST anomalies, the magnitude and spatial extent of pressure and wind anomalies is greater for the strong events, and this pattern begins in Austral winter in year −1.
Figure 4 also confirms the sequence of events proposed in the van Loon hypothesis and illustrates that the seesaw in low pressure between Australia and the south central Pacific is most pronounced for the strong events (Figure 4a,b cf. Figure 4i,j). A distinct feature of the seesaw in strong El Niño events is the large area of significant low pressure anomalies in the winter hemisphere as the sequence develops. That is, with significant low pressure over Australia between MJJ and ASO (−1), the central North Pacific between NDJ (0) and FMA (0), and the central South Pacific between MJJ and ASO (0).
We now examine this sequence more closely by approximately following the first five phases of Harrison and Larkin’s (1998) schematic ENSO composite.
1) Phase 1: pre—May (−1) to October (−1)
In this buildup stage, the magnitude of Pacific warming in the following year is indicated by the magnitude of the low pressure anomaly in the Australian region. Concurrent higher-than-normal pressure over the Pacific means that the Southern Oscillation is acting like a tilted seesaw before a major change. In terms of wind anomalies, the increased northerly flow to the east of Australia assists warmer SST anomalies along the SPCZ for stronger events. Along the equatorial Pacific the southeast trades are enhanced, and combined with stronger westerly wind anomalies at the western end of the Maritime Continent, provide a favorable scenario for the buildup of sea level and heat in the western Pacific. Thus, the “heat pump” picture of ENSO proposed by Jin (1997), and the buildup of sea level (Wyrtki 1975, 1985), is seen here as a wider enhancement of the zonal Walker Circulation.
2) Phase 2: ante—November (−1) to April (0)
Antecedent conditions occur when the low pressure anomalies progress east of Indo-Australia in a complex way. In strong events low pressure over Australia bifurcates about the equator [NDJ (0)] and gradually extends over a larger area to the northwest and southwest Pacific. This bifurcation in pressure appears to be coupled to warm SST anomalies, which takes on a similar spatial pattern with northwesterly flow along the SPCZ, maintaining warm SST in the south, and southerly flow warming the northwest Pacific (Figure 4e–h). In strong events, significant lower pressures in the north-central Pacific in boreal winter appear to be an enhancement of planetary Rossby waves and the precipitation maximum in the annual cycle. Likewise in FMA (0), significant low pressure over the southern end of the SPCZ occurs in the humid wet season coincident with the precipitation maximum in the annual cycle.
We confirm that westerly wind anomalies migrate from west of Sumatra into the western equatorial Pacific at this time (Clarke and Van Gorder 2003), but the magnitude of these anomalies only partially relates to the strength of the following El Niño. Two important lead time features are evident: 1) low pressure changes in the Pacific midlatitudes lead pressure changes along the equator that become established in MJJ (0); and 2) the southeast/northeast trade winds weaken in the midlatitudes before equatorial trades weaken and the east–west SST gradient along the equator is reduced.
3) Phase 3: onset—May (0) to July (0)
In the critical months of MJJ (0) there is a widespread fall in pressure across the Pacific and a rise in pressure in the Indo-Australian region for both event types. The van Loon hypothesis is most pronounced in stronger events, while weak events have significant low pressure anomalies farther east in the southeastern Pacific that line up longitudinally with equatorial warming centered closer to South America (Figs. 3i,j and 4i,j). These results confirm two features of Rossby waves identified by Colls and Whitaker (2001). First, as the zonal flow of air becomes more disturbed, the amplitude of the upper waves becomes greater, with the corresponding reduction in the wavelength indicating baroclinic growth. Second, the net result of these changes in pressure is a greater meridional movement of air: increased equatorward movement of cooler air into the western equatorial and subtropical Pacific 40°S–10°N, 160°E–170°W; and a greater poleward flow in the southeastern Pacific in the region of the South Pacific high (10°–35°S, 120°–80°W). In the latter case, there is an increase of mass transport to the region south of 40°S, coincident with greater subsidence and blocking over the Bellingshausen Sea (Renwick and Revell 1999).
The strength of the El Niño events is clearly a function of westerly wind anomalies in the western equatorial Pacific and the increased SLP gradient between Australia and the South Pacific. The North Pacific high is weaker in strong events, but for weak events remains near normal strength. Weaker El Niño events therefore form from the Southern Oscillation acting over regions to the south of 10°N, further emphasizing the requirement for a weaker South Pacific high in El Niño development (Bjerknes 1966).
4) Phase 4: peak—August (0) to January (+1)
At the peak stage, significant low pressure anomalies cover most of the eastern Pacific in strong events; however for weaker events significant negative anomalies cover less of the equator and are more noticeable across a band from 30° to 40°S in the southeastern Pacific. The increased zonal extent and duration of westerly wind anomalies in strong events (Philander 1981) is clearly seen in Figs. 4m–p. The dynamics of El Niño decay proposed by Chan and Xu (2000) can be seen to begin in the midlatitudes in NDJ (+1) as the high pressure anomaly over Indo-Australia begins to extend toward Hawaii and the central South Pacific over two broad bands of cool SST anomalies. Significant high pressures close to Hawaii in weak and strong events would contribute to the relationship between Hawaii SLP anomalies and subsequent Niño-3.4 SST anomalies found by Anderson (2003).
5) Phase 5: decay—February (+1) to April (+1)
As we move to FMA (+1), the SLP pattern differs more between individual events and this depends on whether a transition to La Niña or neutral conditions is occurring. In stronger events the significant low pressure anomalies weaken along the equator with stronger ridges continuing to gradually extend toward the east (Fig. 4o). Low pressures persist near Tahiti, which causes the Southern Oscillation index (SOI) to remain negative and lag many other indicators that are suggesting a breakdown of El Niño conditions. Associated with this persistent low pressure is a southward shift in equatorial westerly winds (Fig. 4o), which reduces the direct wind forcing along the equator and contributes to the decay process (Harrison and Vechi 1999). By MJJ (+1) (not shown), there is a return to stronger equatorial easterlies and a stronger South Pacific high in both types of events.
The importance of the southern midlatitudes was also found for the late developing El Niño events of 1968/69 and 1986/87 (not shown). Warming developed in NDJ (0), then reamplified in MJJ (0) as a strong South Pacific trough linked up with low pressure anomalies along the equatorial Pacific.
5. Temporal time series aspects of El Niño development
Given the sparse network of weather stations in the South Pacific, normalized time series composites of pressure and SST anomalies are presented in Fig. 5 for individual locations illustrated in Fig. 1. For the 11 El Niño events defined by Harrison and Larkin (1998) monthly normalized SST anomalies are averaged and plotted, while normalized pressure anomalies were further converted into 3-month running means to minimize the Southern Oscillation “signal” being dominated by monthly “noise” (Trenberth 1976).
In the western region around Australia (Fig. 5a), normalized pressure anomalies generally coincide with normalized Niño-3 SST anomalies during the transition into El Niño, emphasizing the broad-scale coupling of the ocean and atmosphere. Between June and October (−1), the largest pressure anomalies occur in winter over southeastern Australia and Darwin. Negative pressure anomalies in the northwestern SPCZ become slightly more significant during NDJ (0) as negative SLP anomalies move east at this time (Fig. 4). SLP rises and falls in the Australian region as Niño-3 warms and cools, respectively, although SLP returns to normal values more quickly over southeastern Australia.
In regions east of the date line (Fig. 5b), the largest anomaly from late in year −1 through to March (0) is warm SST in the south-central Pacific (20°–30°S, 120°–160°W) surrounding Rapa Island. Northerly wind anomalies that flow down the SPCZ at this time (Figure 4c,e,g) maintain warm SST anomalies and enhance the normal gradual decrease in pressure observed in the annual cycle at Rapa Island. Negative pressure anomalies first appear at Rapa in October (−1) and become most significant in the critical months of February–June (0). Significantly, these negative pressure anomalies lead, both a fall in pressure farther north of the SPCZ at Tahiti and a gradual erosion of strength of the South Pacific high centered farther east near Easter Island. The fact that SST anomalies in Niño-3 warm the most between April and July (0) (Fig. 5b), when van Loon (1984) speculated that the South Pacific trough most influences Pacific trade winds (MJJ), is an important consideration. Negative SLP at Rapa that coincides with rapid Niño-3 warming in April implies an earlier northward influence of the trough on the trades.
The underlying mechanism for this process appears to be a standing wave of pressure anomalies between southeastern Australia and the South Pacific (Fig. 6). Three evenly spaced locations between Australia and the central South Pacific were chosen at 30°S: southeastern Australia appears to represent a western antinode, the date line (Raoul, Kermadec Islands, 29°S, 178°W) a stationary node (until November 0), and Rapa Island the eastern antinode. The points of maximum amplitude (antinodes) oscillate from over southeastern Australia in year −1, are equally found at both antinodes in year 0, and finally end in the South Pacific in year +1 when El Niño decays. Near the date line, pressure anomalies are part of broad-scale negative anomalies in the Australian region before El Niño, and broad-scale positive anomalies in the South Pacific after El Niño (Figs. 4 and 6).
To investigate the strength of the anomalies in relation to the strength of the event, we divided the 11 warm events defined by Harrison and Larkin (1998) into weak, moderate, and strong based on the BEI3 power: weak—1951, 1953, and 1969 (BEI3 Power 0–3); moderate—1957, 1965, 1976, 1987, and 1991 as moderate (BEI3 power 6–12); and, strong—1972, 1982, and 1997 (BEI3 Power 20–70). When the pressure anomalies were averaged for these categories, the magnitude of negative anomalies over southeastern Australia in winter (−1) is reversed by winter (0) (Fig. 7a). These anomalies match 1) the magnitude of negative anomalies at Rapa Island in winter (0) (Fig. 7b); and 2) the rate of SST warming in the Niño-3 region through April–July (0) (Fig. 7c). Thus, consistent with a standing wave, the magnitudes of pressure anomalies in winter (0) over southeastern Australia (Figure 4a,b) are later observed in the central South Pacific in year 0 (Figure 4i,j) and these coincide with the rate of SST increase farther north along the equator for the critical period of trough enhancement between April and July (0). The timing of the trough strengthening in the South Pacific also appears important in determining the strength of the El Niño events. Negative SLPs appear earlier at Rapa at the end of year −1 and early in year 0 for weaker and moderate events, but are more significant in austral winter (0) when strong events are developing.
At Hawaii, very negative SLP anomalies are clearly indicated between April and July (0) for strong El Niño events (Fig. 7d). This supports the assumption by Wang and Weisberg (2000) that a dipole of complementary negative pressure anomalies are needed in the northern and southern midlatitudes to force the strong westerly wind anomalies needed at the onset of extreme events. For weaker events there are weaker SLP anomalies between April and July (0) at Hawaii, implying that the North Pacific has a more minor role in the formation of these events (Figs. 4j,l). Figure 7d also supports the leading role that Hawaii SLP in NDJ (0) has at indicating a subsequent warming of Niño-3.4 SST (Anderson 2003) but suggests that this is probably dominated by the few strong El Niño events.
6. The relationship between pressure and the Pacific trade winds
SLP anomalies at Rapa Island in the South Pacific appear to play a vital role in equatorial SST changes farther north in the Niño-3 region (Fig. 8a). The correlation between the two curves is −0.42, significant at the 99% level. Warm and cold events are clearly identified with divergence between the two time series, with the largest Rapa pressure anomalies leading the largest SST anomalies. All El Niño events are associated with a stronger South Pacific trough, and all La Niña events are associated with a stronger South Pacific ridge. The critical role of the South Pacific trough is also emphasized by the shift to more consistent negative pressure anomalies at Rapa Island between 1976 and 1997, and the increased occurrence of El Niño (Fig. 8a).
To look at the influence of the South Pacific trough on equatorial trades, Rapa SLP anomalies are plotted against the anomalies in the area-averaged 850-hPa trade wind index for the western Pacific (Fig. 8b). Pressure anomalies at Rapa lead anomalies in the trades, but more importantly, coincide in anomaly strength at the beginning of El Niño (1982, 1986/87, 1991, and 1997) and La Niña (1988, 1995, and 1998). When the trade wind and SLP curves have opposite anomalies, ENSO events do not develop, even if the trade winds are very strong (1984) or weak (early 1990).
The unprecedented rapid rise and fall in SST in the 1997 El Niño and the 1998 La Niña appear to be related to the most pronounced weakening and strengthening of the South Pacific ridge between April and July since 1950 (Figs. 8a,b). In the decay of the event, the South Pacific ridge (trades) are clearly strengthening, and Niño-3 SST is cooling, long before the “small-scale” MJO event reaches the eastern Pacific in May 1998 (Takayabu et al. 1999; Straub et al. 2006). This supports the conclusion by Wang and Weisberg (2000) and Boulanger et al. (2004) that strong off-equatorial low (high) SLP and westerly (easterly) wind anomalies in the western Pacific in early 1997 (1998) were the main drivers of the accelerated development (termination) of this event.
Figure 9a shows that in El Niño (La Niña) years the enhanced (weakened) annual cycle at 30°S causes a greater (weaker) pressure difference in austral winter between Australia and the South Pacific. In 1997, there was an unprecedented eight consecutive months between March and October when SLP was lower at Rapa Island than over southeastern Australia, that is, southwesterly wind flow directed into the southwest Pacific (Fig. 9b). There is a fundamentally different SLP gradient between these two regions when we compare 1997 to 1998 (Figs. 9b,c). This difference begins in April and continues through to October.
To quantify the relative importance of the Pacific subtropical ridges on ENSO, SLP anomalies at Rapa Island and Hawaii were correlated with the western Pacific 850-hPa trade wind index (Fig. 10). In spite of the fact that Rapa Island is approximately 4880 km to the southeast of the western equatorial Pacific, and 1450 km farther from this region than Hawaii, the trade winds are significantly correlated to the strength of the South Pacific ridge between March and November and only between May and August for the North Pacific ridge. Weaker correlations over austral summer at Rapa would be expected as the subtropical high is to the south of Rapa Island and the surface trough in the westerlies is farthest south in the annual cycle. High correlations between trade winds and Rapa SLP in February–March and March–April occur in the wettest months in the year at Rapa when the SPCZ is active and extends southward, that is, El Niño is linked to enhancement in the annual cycle of precipitation.
Figure 11a shows that Rapa pressure is lower than normal in March–April for all the El Niño years and higher than normal for most of the La Niña years. Thus, the pattern of SLP anomalies found in the South Pacific appears to play a significant role in setting up a favorable pattern for ENSO development or transitions (r = 0.62). In contrast, Fig. 11b shows that for the same March–April period there is an insignificant relationship between Hawaii pressure and western Pacific trade winds (r = 0.26). Very low pressures are seen in the strong El Niño years 1982 and 1997, but also in neutral and La Niña years. Given the stronger negative SLP anomalies in strong El Niño years, but the low correlation with trade winds for all years (Fig. 11b), it appears that the North Pacific pressures 1) respond to broad-scale pressure changes from year −1, 2) contribute to the formation of strong El Niño events (Figs. 4e,g,i), and 3) adjust to trade wind strength that develops along the equator between May and August (0).
In recent years there has been speculation that westerly wind anomalies in the western equatorial Pacific play an important role in triggering El Niño and weakening the trade winds across the Pacific. If this was the case, then the Pacific subtropical highs and ridges would weaken after the westerly anomalies appeared. To test this, we correlated South Pacific SLP anomalies with trade wind strength with lead time, and then reversed the lead/lag relationship (Figs. 11c,d). Rapa Island SLP anomalies in March–April explained half the variance of the following west Pacific trade winds in May–June (r = 0.70, significant at the 99% level), while the reverse relationship was weaker and insignificant (r = 0.34). Shorter-term periods of weaker trades associated with the passage of MJO events are obviously related to oceanic Kelvin wave development and subsequent east Pacific warming, however Figs. 4 and 5 and Figs. 7 –11 suggest that El Niño events will not develop unless there is support from the broad-scale SLP pattern, especially in the South Pacific.
7. Discussion and conclusions
By comparing El Niño development on the basis of strength this study has clarified a number of previous findings. A broad-scale dynamical framework for ENSO development is possible if the interactions between the midlatitudes and equator are properly accounted for.
First, we find that the development of stronger El Niño events is dominated by the classical Southern Oscillation pressure signal observed on the Southern Hemisphere by van Loon and Shea (1985, 1987). We confirm that wave fluctuations in the southern subtropics are closely coupled to the Southern Oscillation (Kidson 1975; Trenberth 1980; Karoly 1989). The greatest Pacific warming (April–July of year 0) occurs when Rossby waves in the southern midlatitudes have a high amplitude and short wavelength. Energy, or momentum, in the Rossby waves appears to be tranferred from Australia (year −1) to the South Pacific (year 0) via a low-frequency standing wave of SLP, and a fundamental reversal in the strength of the annual cycle close to 30°S. These changes must occur in sequence with global-scale pressure changes (Barnett 1985; Krishnamurti et al. 1986) coincident with a transition from an enhanced Walker Circulation (buildup of heat in Indo–Australia), to a weaker Walker circulation (movement of heat heat away from Indo–Australia; Meinen and McPhaden 2000; Sun 2003). Such changes are also linked to the annual cycle of clouds and precipitation, and the movement of the convective maximum from the Indo–Australian region into the southwest Pacific (Meehl 1987).
Since the Southern Oscillation does act like a seesaw, or tilted pendulum, the van Loon hypothesis appears to be critical for the development of strong events, but must also incorporate the complimentary role of the North Pacific Rossby waves in strong events (Wang and Weisberg 2000). The less significant pressure signals found prior to weak events suggest that they are more variable in their development, and that air–sea interations along the equator could be more significant in their formation.
The wave, or seesaw features, related to the van Loon hypothesis occurred in a dramatic way leading into the intense 1997 El Niño. In the southern midlatitudes there was an unprecedented 1) low SLP anomaly over southeastern Australia in JAS 1996 (which Fig. 2 partly illustrates), 2) east–west pressure reversal between Australia and the central South Pacific (Figs. 2 and 8), 3) warm SPCZ in March 1997 (CPC 1997), 4) strong South Pacific trough (March–July; Fig. 8), and 5) eight consecutive months when SLP at Rapa Island was lower than that recorded over southeastern Australia (Fig. 9). Wang and Weisberg (2000) only showed weak southerly wind anomalies off the east coast of Australia in the 1997 El Niño development phase, but in doing so did not map anomalies between April and July, when equatorial warming was greatest.
Second, we show that the standing wave of SLP in the southern midlatitudes is part of a low-frequency eastward propogation of 1) SLP at a large scale (Barnett 1985; Krishnamurti et al. 1986; Meehl 1987; Kiladis and van Loon 1988), and 2) surface wind anomalies along the equator (Gutzler and Harrison 1987; Chen and Wu 2000; Clarke and Van Gorder 2003). Gill and Rasmusson (1983), Barnett (1985), and Barnett et al. (1991) show that low SLP anomalies on the equator propagate eastward through an ocean–atmosphere feedback such that larger-than-normal absolute ocean temperatures are continually created to the east of the central low pressure anomaly and the release of latent heat in the middle troposphere. This phase shift between the forcing and response effectively “pulls” the center of convection and low pressure eastward toward the warmer SST, with westerly wind anomalies and advective cooling occurring to the west of the convective zone (Barnett et al. 1991). We show that this process extends into higher latitudes as meridional wind anomalies associated with large-scale pressure anomalies flow from opposite directions in each hemisphere and support SST changes in the subtropics.
Third, a key finding of this study is that the Pacific troughs strengthen long before changes are observed on the equator (Figs. 4, 5, 7, 8 and 11c,d). Thus, we confirm the lead role that SLP changes play in the South Pacific (Quinn 1974; Trenberth 1976, 1980; Chen 1982; Trenberth and Shea 1987; Wright et al. 1988) and the North Pacific (Anderson 2003, 2004). We do not see evidence for a reduced east–west SST gradient along the equator reducing the trade winds in the central eastern Pacific (Lengaigne et al. 2003), but rather see the trades weakening as a result of broad-scale pressure changes that first appear in the midlatitudes. Peixoto and Oort (1992) found that changes in the SOI lead changes in zonal wind stress in the central equatorial Pacific by 2 months, and eastern Pacific SST by 4.5 months. Our analysis suggests that pressure changes in the midlatitudes could give an even longer lead warning of Pacific warming. Significant high (low) pressure anomalies near Tahiti at the beginning (end) of weaker (stronger) events cause the SOI to lag other indicators in some ENSO transitions.
It could be argued that mid-Pacific surface troughs are enhanced by anomalous convection near the date line in MJJ (0) (Kiladis and van Loon 1988) and, through the mass circulation, force the extension of the subtropical jet and associated South Pacific storm track. However, the fact that the SLP changes in the midlatitudes lead changes along the equator is a strong counterargument to this proposition. Nevertheless, positive feedbacks between weaker trades, surface warming, and eastward shifts in convection (McPhaden 2004) would benefit from lower pressure in the subtropics and vice versa. Once tropical convection becomes anomolously strong and extends into the central Pacific, it would act as a positive feedback between the ocean and atmospheric circulation that would favor the maintenance and continued strengthening of the event. It is interesting to note that the South Pacific troughs are strengthened farther east in weaker events, which means any feedback between the Tropics and subtropics would be weaker in the western Pacific.
Fourth, we confirm that the strength of winter hemisphere planetary Rossby waves has an important role in ENSO development and decay (van Loon hypothesis; Anderson 2003, 2004; Vimont et al. 2003; Anderson and Maloney 2006). In particular, an enhanced South Pacific trough causes wind stress forcing that adds to the westerly momentum change occurring farther north along the equator, both in the western and the eastern Pacific. Such changes in momentum should be a major contributor to ENSO being phase locked to the annual cycle, and to the seasonal dependence in MJO strengthening in MJJ (0) in El Niño years (Teng and Wang 2003).
Harrison and Larkin (1998) found that the wind anomaly patterns in their composites mainly occurred within 10° of the equator and have “little in common with the structure of the northeast or southeast trades.” This study, along with Wang and Weisberg (2000) and Anderson (2003), contradicts this view. By grouping weak, strong, and late-developing events together, Harrison and Larkin (1996, 1998) could have obscured the role of the midlatitudes in their analysis. We also consider that the role of the South Pacific trough would have been diminished in the study of Anderson (2003) because a detrending process would have removed the trend to lower pressure in the South Pacific, which coincided with more eastern Pacific warming (Fig. 8).
Our composites are unsuitable for testing the contasting views of ENSO dynamics although they do support Chen et al. (2004), who found that stronger El Niño events appear as low-frequency, self-sustaining oscillations that set up in the large-scale state, especially SST. However, we would concur with Guilyardi et al. (2004) that the atmosphere plays the dominant role in setting El Niño amplitude by allowing air–sea interactions on the equator to occur. The link between the two main theories of ENSO dynamics and our results could be provided by Eisenman et al. (2005), who found that the low-frequency component of westerly wind bursts, which force ENSO, are modulated by ENSO through the large-scale equatorial SST field in the western Pacific. We find a notable difference in the SST field between stronger and weaker events near the Pacific warm pool between NDJ (0) and FMA (0) (Figure 4e–h).
Finally, there are three implications for ENSO monitoring that follow from this study. First, stronger events should be predictable at a longer lead time if SLP changes over southeastern Australia are monitored. Second, a broad-scale index of the Southern Oscillation is needed to properly account for pressure changes mapped in Fig. 4. Third, the greater spatial extent of significant pressure and SST anomalies in strong warm events provides a plausible explanation to the spatial extent of drought in the global scale (Ropelewski and Halpert 1987; Allan et al. 1996; Lyon 2004).
In summary, equatorial Pacific warm events are clearly not only a result of reciprocal action between air and sea, but also reciprocal interactions between equatorial and midlatitudinal weather systems. We find that midlatitudinal and equatorial pressure anomalies combine to determine the strength and persistence of westerly wind anomalies in the western/central Pacific and the magnitude of Pacific warming. The reversal of trough strength in the southern midlatitudes before ENSO extremes, combined with a buildup and weakening of trades (heat), offers a consistent dynamical framework for the connection between the Southern Oscillation and the magnitude of eastern Pacific warming.
Acknowledgments
We thank George Kiladis (NOAA/Earth Systems Research Laboratory), Neville Nicholls (Australian BMRC), Ian Foster (DAFWA), and two anonymous referees for many useful comments on this paper. Ian Smith (CSIRO) kindly provided GISST3 South Pacific SST data and Jim Renwick (NIWA) NCEP–NCAR reanalysis data. Patrick Varney (Meteo-France), Jim Salinger (NIWA), and Denis Shea (NCAR) all provided SLP data for stations in the Pacific Ocean. The Australian Grains Research and Development Corporation (GRDC) provided funding for the project.
REFERENCES
Allan, R., J. Lindesay, and D. Parker, 1996: El Niño—Southern Oscillation and Climatic Variability. CSIRO Publishing, 416 pp.
Anderson, B. T., 2003: Tropical Pacific sea-surface temperatures and preceding sea level pressure anomalies in the subtropical North Pacific. J. Geophys. Res., 108 .4732, doi:10.1029/2003JD003805.
Anderson, B. T., 2004: Investigation of a large scale mode of ocean–atmosphere variability and its relation to tropical Pacific sea surface temperature anomalies. J. Climate, 17 , 4089–4098.
Anderson, B. T., and E. Maloney, 2006: Interannual tropical Pacific sea surface temperatures and their relation to preceding sea level pressures in the NCAR CCSM2. J. Climate, 19 , 998–1012.
Anderson, D. L. T., and M. K. Davey, 1998: Predicting the El Niño of 1997/98. Weather, 53 , 303–310.
Barnett, T. P., 1985: Variations in near-global sea level pressure. J. Atmos. Sci., 42 , 478–501.
Barnett, T. P., M. Latif, E. Kirk, and R. Roeckner, 1991: On ENSO physics. J. Climate, 4 , 487–515.
Barnston, A. G., M. H. Glantz, and Y. He, 1999: Predictive skill of statistical and dynamical climate models in SST forecasts during the 1997–98 El Niño episode and the 1998 La Niña onset. Bull. Amer. Meteor. Soc., 80 , 217–243.
Bergman, J. W., H. H. Hendon, and K. M. Weickmann, 2001: Intraseasonal air–sea interactions at the onset of El Niño. J. Climate, 14 , 1702–1719.
Berlage, H. P., 1966: The Southern Oscillation and world weather. Mededelingen en Verhandelingen 88, Koninklijk Nederlands Meteorologisch Instituut, 152 pp.
Bjerknes, J. A., 1966: Possible response of the atmospheric Hadley Circulation to equatorial anomalies of ocean temperature. Tellus, 97 , 820–829.
Bjerknes, J. A., 1969: Atmospheric teleconnections from the equatorial Pacific. Mon. Wea. Rev., 97 , 163–172.
Boulanger, J. P., C. Menkes, and M. Lengaigne, 2004: Role of high- and low-frequency winds and wave reflection in the onset, growth and termination of the 1997–1998 El Niño. Climate Dyn., 22 , 267–280.
Chan, J. L., and J. Xu, 2000: Physical mechanisms responsible for the transition from a warm to a cold state of the El Niño–Southern Oscillation. J. Climate, 13 , 2056–2071.
Chen, D., M. A. Cane, A. Kaplan, S. E. Zebiak, and D. Huang, 2004: Predictability of El Niño over the past 148 years. Nature, 428 , 733–735.
Chen, L., and R. Wu, 2000: The role of the Asian/Australian monsoons and the Southern/Northern Oscillation in the ENSO cycle. Theor. Appl. Climatol., 65 , 37–47.
Chen, W. Y., 1982: Assessment of Southern Oscillation sea level pressure indices. Mon. Wea. Rev., 110 , 800–807.
Clarke, A. J., and S. Van Gorder, 2003: Improving El Niño prediction using a space-time integration of Indo-Pacific winds and equatorial Pacific upper ocean heat content. Geophys. Res. Lett., 30 .1399, doi:10.1029/2002GL016673.
CPC, 1997: Climate Diagnostic Bulletin 97(3), NOAA, 78 pp.
Colls, K., and R. Whitaker, 2001: The Weather Book. Reed New Holland, 180 pp.
Eisenman, I., L. Yu, and E. Tziperman, 2005: Westerly wind bursts: ENSO’s tail rather than the dog? J. Climate, 18 , 5224–5238.
Fedorov, A. V., and S. G. Philander, 2000: Is El Niño changing? Science, 288 , 1997–2002.
Gill, A. E., and E. M. Rasmusson, 1983: The 1982–83 climate anomaly in the equatorial Pacific. Nature, 306 , 229–234.
Guilyardi, E., and Coauthors, 2004: Representing El Niño in coupled ocean–atmosphere GCMs: The dominant role of the atmospheric component. J. Climate, 17 , 4623–4629.
Gutzler, D. S., and D. E. Harrison, 1987: The structure and evolution of seasonal wind anomalies over the near-equatorial eastern Indian and western Pacific Oceans. Mon. Wea. Rev., 115 , 169–192.
Harrison, D. E., 1984: The appearance of sustained equatorial surface westerlies during the 1982 Pacific warm event. Science, 224 , 1099–1102.
Harrison, D. E., and N. K. Larkin, 1996: The COADS sea level pressure signal: A near-global El Niño composite and time series view, 1946–1993. J. Climate, 9 , 3025–3055.
Harrison, D. E., and N. K. Larkin, 1998: El Niño–Southern Oscillation sea surface temperature and wind anomalies, 1946–1993. Rev. Geophys., 36 , 353–399.
Harrison, D. E., and G. A. Vechi, 1999: On the termination of El Niño. Geophys. Res. Lett., 26 , 1593–1596.
Hurrell, J. W., H. van Loon, and D. J. Shea, 1998: The mean state of the troposphere. Meteorology of the Southern Hemisphere, Meteor. Monogr., No. 49, Amer. Meteor. Soc., 1–46.
Jin, F-F., 1997: An equatorial ocean recharge paradigm for ENSO. Part I: Conceptual model. J. Atmos. Sci., 54 , 811–829.
Jones, C., L. M. V. Carvalho, R. W. Higgins, D. E. Waliser, and J. K. E. Schemm, 2004: Climatology of tropical intraseasonal convective anomalies: 1979–2002. J. Climate, 17 , 523–539.
Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77 , 437–471.
Karoly, D. J., 1989: Southern Hemisphere circulation features associated with El Niño–Southern Oscillation events. J. Climate, 2 , 1239–1252.
Kidson, J. W., 1975: Tropical eigenvector analysis and the Southern Oscillation. Mon. Wea. Rev., 103 , 187–196.
Kiladis, G. N., and H. van Loon, 1988: The Southern Oscillation. Part VII: Meteorological anomalies over the Indian and Pacific sectors associated with the extremes of the oscillation. Mon. Wea. Rev., 116 , 120–136.
Kiladis, G. N., and K. C. Mo, 1998: Interannual and intraseasonal variability in the Southern Hemisphere. Meteorology of the Southern Hemisphere, Meteor. Monogr., No. 49, Amer. Meteor. Soc., 307–336.
Krishnamurti, T. N., S. H. Chu, and W. Iglesias, 1986: On sea level pressure of the Southern Oscillation. Arch. Meteor. Geophys. Bioklimatol., 34A , 385–425.
Landsea, C. W., and J. A. Knaff, 2000: How much skill was there in forecasting the very strong 1997–98 El Niño? Bull. Amer. Meteor. Soc., 81 , 2107–2119.
Latif, M., and Coauthors, 1998: A review of the predictability and prediction of ENSO. J. Geophys. Res., 103 , C7. 14375–14393.
Lengaigne, M., J. Boulanger, C. Menkes, G. Madec, P. Delecluse, E. Guildyardi, and J. Slingo, 2003: The March 1997 westerly wind event and the onset of the 1997/98 El Niño: Understanding the role of the atmospheric response. J. Climate, 16 , 3330–3343.
Levinson, D. H., 2005: State of the climate in 2004. Bull. Amer. Meteor. Soc., 86 , 1–86.
Lyon, B., 2004: The strength of El Niño and the spatial extent of tropical drought. Geophys. Res. Lett., 31 .L21204, doi:10.1029/2004GL020901.
McPhaden, M. J., 2004: Evolution of the 2002/03 El Niño. Bull. Amer. Meteor. Soc., 85 , 677–695.
McPhaden, M. J., and X. Yu, 1999: Equatorial waves and the 1997–98 El Niño. Geophys. Res. Lett., 26 , 2961–2964.
Meehl, G. A., 1987: The annual cycle and interannual variability in the tropical Pacific and Indian Ocean regions. Mon. Wea. Rev., 115 , 27–50.
Meinen, C. S., and M. J. McPhaden, 2000: Observations of warm water volume changes in the equatorial Pacific and their relationship to El Niño and La Niña. J. Climate, 13 , 3551–3559.
Mitchum, G. T., 1987: Trade wind fluctuations associated with El Niño-Southern Oscillation events. J. Geophys. Res., 92 , 9464–9468.
Peixoto, J. P., and A. H. Oort, 1992: Physics of Climate. American Institute of Physics, 520 pp.
Philander, S. G. H., 1981: The response of equatorial oceans to a relaxation of the trade winds. J. Phys. Oceanogr., 11 , 176–189.
Picaut, J., E. Hackert, A. J. Busalacchi, R. Murtugudde, and G. S. E. Lagerleof, 2002: Mechanisms of the 1997–1998 El Niño–La Niña, as inferred from space-based observations. J. Geophys. Res., 107 .3037, doi:10.1029/2001JC000850.
Quinn, W. H., 1974: Monitoring and predicting El Niño invasions. J. Appl. Meteor., 13 , 825–830.
Rasmusson, E. M., and T. H. Carpenter, 1982: Variations in tropical sea surface temperature and surface wind fields associated with the Southern Oscillation/El Niño. Mon. Wea. Rev., 110 , 354–384.
Renwick, J. A., and M. J. Revell, 1999: Blocking over the South Pacific and Rossby wave propogation. Mon. Wea. Rev., 127 , 2233–2247.
Ropelewski, C. F., and M. S. Halpert, 1987: Global and regional scale precipitation patterns associated with the El Niño/Southern Oscillation. Mon. Wea. Rev., 115 , 1606–1626.
Slingo, J. M., D. P. Powell, K. R. Sperber, and F. Nortley, 1999: On the predictability of the interannual behaviour of the Madden-Julian oscillation and its relation to ENSO. Quart. J. Roy. Meteor. Soc., 125 , 583–609.
Smith, T. M., and R. W. Reynolds, 2003: Extended reconstruction of global sea surface temperatures based on COADS data (1854–1997). J. Climate, 16 , 1495–1510.
Solman, S. A., and C. G. Menendez, 2002: ENSO-related variability of the Southern Hemisphere winter storm track over the eastern Pacific–Atlantic sector. J. Atmos. Sci., 59 , 2128–2140.
Straub, K. H., G. N. Kiladis, and P. E. Ciesielski, 2006: The role of equatorial waves in the onset of the 1998 South China Sea summer monsoon. Dyn. Atmos. Oceans, 42 , 216–238.
Sun, D-Z., 2003: A possible effect of an increase in the Pacific warm-pool SST on the magnitude of El Niño warming. J. Climate, 16 , 185–205.
Takayabu, Y. N., T. Iguchi, M. Kachi, A. Shibati, and H. Kanzawa, 1999: Abrupt termination of the 1997–98 El Niño in response to a Madden-Julian oscillation. Nature, 402 , 279–282.
Teng, H. Y., and B. Wang, 2003: Interannual variations of the boreal summer intraseasonal oscillation in the Asian–Pacific region. J. Climate, 16 , 3572–3584.
Thompson, D. W. J., and D. J. Lorenz, 2004: The signature of the annular modes in the tropical troposphere. J. Climate, 17 , 4330–4342.
Trenberth, K. E., 1976: Spatial and temporal variations of the Southern Oscillation. Quart. J. Roy. Meteor. Soc., 102 , 639–653.
Trenberth, K. E., 1980: Planetary waves at 500 mb in the Southern Hemisphere. Mon. Wea. Rev., 108 , 1378–1389.
Trenberth, K. E., 1998: Development and forecasts of the 1997/98 El Niño: CLIVAR scientific issues. CLIVAR Exchanges, No. 3, International CLIVAR Project Office, Southampton, United Kingdom, 4–14.
Trenberth, K. E., 2000: The Southern Oscillation revisited: Sea level pressures, surface temperatures, and precipitation. J. Climate, 13 , 4358–4365.
Trenberth, K. E., and D. J. Shea, 1987: On the evolution of the Southern Oscillation. Mon. Wea. Rev., 115 , 3078–3096.
Trenberth, K. E., J. M. Caron, D. P. Stepaniak, and S. Worley, 2002: Evolution of the El Niño–Southern Oscillation and global atmospheric surface temperatures. J. Geophys. Res., 107 .4065, doi:10.1029/2000JD000298.
van Loon, H., 1967: The half-yearly oscillation in the middle and high southern latitudes and the coreless winter. J. Atmos. Sci., 24 , 472–486.
van Loon, H., 1972: Temperature, pressure, wind, cloudiness and precipitation in the Southern Hemisphere. Meteorology of the Southern Hemisphere, Meteor. Monogr., No. 49, Amer. Meteor. Soc., 25–111.
van Loon, H., 1984: The Southern Oscillation. Part III: Associations with the trades and with the trough in the westerlies of the South Pacific Ocean. Mon. Wea. Rev., 112 , 947–954.
van Loon, H., and D. J. Shea, 1985: The Southern Oscillation. Part IV: The precursors south of 15°S to the extremes of the oscillation. Mon. Wea. Rev., 113 , 2063–2074.
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.
van Loon, H., G. A. Meehl, and R. F. Milliff, 2003: The Southern Oscillation in the early 1990s. Geophys. Res. Lett., 30 .1478, doi:10.1029/2002GL016307.
Vimont, D. J., J. M. Wallace, and D. S. Battisti, 2003: The seasonal footprinting mechanism in the Pacific: Implications for ENSO. J. Climate, 16 , 2668–2675.
Von Storch, H., H. van Loon, and G. N. Kiladis, 1988: The Southern Oscillation. Part VIII: Model sensitivity to SST anomalies in the tropical and subtropical regions of the South Pacific convergence zone. J. Climate, 1 , 325–331.
Wang, C., and R. H. Weisberg, 2000: The 1997–98 El Niño evolution relative to previous El Niño events. J. Climate, 13 , 488–501.
Wright, P. B., J. M. Wallace, T. P. Mitchell, and C. Deser, 1988: Correlation structure of the El Niño/Southern Oscillation phenomenon. J. Climate, 1 , 609–625.
Wyrtki, K., 1975: El Niño—The dynamic response of the equatorial Pacific Ocean to atmospheric forcing. J. Phys. Oceanogr., 5 , 572–584.
Wyrtki, K., 1985: Water displacements in the Pacific and the genesis of El Niño cycles. J. Geophys. Res., 90 , 7129–7132.
Xu, J., and J. C. L. Chan, 2001: The role of the Asian–Australian monsoon system in the onset time of El Niño events. J. Climate, 14 , 418–433.
Zhang, C. D., and J. Gottschalck, 2002: SST anomalies of ENSO and the Madden–Julian oscillation in the equatorial Pacific. J. Climate, 15 , 2429–2445.
Schematic diagram showing major weather systems and anomalies in the Southern Hemisphere in winter–spring in (a) the year before El Niño [year (−1)] and (b) during El Niño [year (0)]. The Walker Circulation (white) and westerly surface troughs (red curve) sit above surface anomalies measured at key stations represented by crosses [Alice Springs (A), Darwin (D), Easter Island (E), Mildura (M), Noumea (N), Rapa Island (R), Tahiti (T), Willis Island (W)]. The SPCZ is represented by parallel gray lines. Warm (red) and cold (green) SST anomalies are indicated, as are wind anomalies (black arrows). High and low SLP are indicated by “H” and “L” and anomalous pressures in the midlatitudes are dashed.
Citation: Journal of Climate 20, 10; 10.1175/JCLI4131.1
The 3-month running mean of SLP anomalies at (a) Darwin, northern Australia, and (b) Mildura, southeastern Australia. Solid dashed straight lines are superimposed on the curves for the 12 months leading into El Niño development.
Citation: Journal of Climate 20, 10; 10.1175/JCLI4131.1
The evolution of SST anomalies for strong and weak El Niño composites: from (a) MJJ (−1) through to (p) FMA (+1). Anomaly contours are 0.1°C with positive anomalies shaded as indicated by the legend on the next page. Superimposed on SST anomalies are dashed thick lines showing regions of significant anomalies at the 90% level based on the normal-z statistic, and the Student’s t test.
Citation: Journal of Climate 20, 10; 10.1175/JCLI4131.1
The evolution of SLP and surface wind anomalies for strong and weak El Niño composites: from (a) MJJ (−1) through to (p) FMA (+1). SLP anomaly contoured every 0.5 hPa with negative anomalies shaded. Superimposed on the pressure anomalies are dashed thick lines showing regions of significant SLP anomalies at the 90% level based on the normal-z statistic, and the Student’s t test. The vector wind anomaly is measured by the scale arrow indicated on the next page.
Citation: Journal of Climate 20, 10; 10.1175/JCLI4131.1
Composite time series of averaged normalized SST anomalies in the Niño-3 region (solid line, diamonds) for 11 El Niño events compared to equivalent anomalies at (a) western region, SLP at Darwin (solid line, solid triangle), northwest SPCZ (average Noumea and Willis Island; dashed line, open squares), and southeastern Australia (average of Alice-Springs and Mildura; thin line, open triangles); (b) eastern region, central South Pacific SST (solid line, open squares), and SLP anomalies at Rapa Island (solid line, solid circles), Easter Island (thin line, open circles), and Tahiti (dashed line, solid triangles). All SLP anomalies have been smoothed by a 3-month running mean.
Citation: Journal of Climate 20, 10; 10.1175/JCLI4131.1
Composite time series of averaged SLP anomalies for 11 El Niño events for Raoul (Kermadec Islands) near the date line (open circles); Rapa, west of the date line (solid circles); and southeastern Australia, east of the date line (solid triangles). All SLP anomalies have been smoothed by a 3-month running mean.
Citation: Journal of Climate 20, 10; 10.1175/JCLI4131.1
Composite time series of averaged normalized anomalies for three weak El Niño events (faint line), five moderate events (dark thin line), and three strong events (bold thick line) for (a) SLP in southeastern Australia (average of Alice Springs and Mildura), (b) SLP on Rapa Island, (c) Niño-3 SST, and (d) SLP in Hawaii. All pressure data have been smoothed by a 3-month running mean.
Citation: Journal of Climate 20, 10; 10.1175/JCLI4131.1
(a) Plot of normalized Rapa Island SLP anomalies since 1951 [black line, left (y) axis] vs actual Niño-3 SST anomalies [red and blue color fill, right (y) axis]. (b) Plot of normalized Niño-3 SST anomalies (red curve) vs normalized Rapa Island SLP anomalies (blue curve) and normalized 850-hPa west Pacific trade winds (purple curve) (1979–2004). All curves have been smoothed using a 5-month running mean.
Citation: Journal of Climate 20, 10; 10.1175/JCLI4131.1
Comparison between SLP at Rapa Island and southeastern Australia (average of Alice Springs and Mildura: (a) plot of 5-month running mean in 1994–98, Rapa (dashed line) and southeast Australia (solid line); and (b) monthly values in 1997, Rapa (solid triangle) and southeast Australia (solid dot). (c) Same as in (b), but for 1998.
Citation: Journal of Climate 20, 10; 10.1175/JCLI4131.1
Correlations between bimonthly standardized 850-hPa trade wind anomalies in the western equatorial Pacific and bimonthly standardized pressure anomalies at Rapa Island in the South Pacific and Hawaii in the North Pacific (1979–2001). Dashed lines represent the 99% and 95% significance levels.
Citation: Journal of Climate 20, 10; 10.1175/JCLI4131.1
(a) Simultaneous correlations between average March–April standardized SLP anomalies at Rapa Island in the South Pacific and standardized 850-hPa trade wind anomalies in the western equatorial Pacific; (b) simultaneous correlations between average March–April standardized SLP anomalies at Honolulu in the North Pacific and standardized 850-hPa trade wind anomalies in the western equatorial Pacific; (c) correlations between average March–April standardized SLP anomalies at Rapa Island and standardized 850-hPa trade wind anomalies in the western equatorial Pacific in May–June; and (d) correlations between average March–April standardized 850-hPa trade wind anomalies in the western equatorial Pacific and average standardized SLP anomalies at Rapa Island in May–June.
Citation: Journal of Climate 20, 10; 10.1175/JCLI4131.1
