1. Introduction
The worldwide impact of the El Niño–Southern Oscillation (ENSO) phenomenon (Trenberth 1997) has been extensively documented for the last decades (e.g., Ropelewski and Halpert 1987; Kiladis and Diaz 1989; Klein et al. 1999; Lau and Nath 2001). A better understanding of how ENSO is generated would significantly help to improve its prediction (Battisti and Sarachik 1995) and thus help to anticipate associated positive and negative impacts in many regions of the world (CPC 1997; Changnon 1999; Pascual et al. 2000; Glantz 2001; Pezzi and Cavalcanti 2001). However, our knowledge of what causes El Niño (EN) events is limited and, as a result, the accuracy of predictions is still far from the inherent limits to predictability (Chen and Cane 2008), especially for long-lead times (Chang et al. 2009). At present, purely statistical (Graham et al. 1987; Tang et al. 1997; Xue et al. 2000), hybrid (Barnett et al. 1993), and fully physical (Cane et al. 1986; Kirtman et al. 1997) coupled models show similar skills, which seem to have reached a plateau at the moderate predictability level (Chen and Cane 2008). The main problems arise from uncertainties in initialization schemes relevant for ENSO prediction (Clarke 2008) and in the modest advances to date in the understanding of ENSO physics (Wang and Picaut 2004). Long-range forecasting research is, therefore, relevant to determine if we are facing a fundamental barrier in the characterization of ENSO physics (Luo et al. 2005; Deng et al. 2009), and whether this limit is inherent to the characterization of the ENSO phenomenon (Latif et al. 1994, 1998; McPhaden et al. 1998).
Since the eighties, our understanding of the ENSO phenomenon has improved significantly (Wang and Fiedler 2006), but its nature is still an open issue. ENSO has been described as a self-sustained natural oscillatory mode of the coupled ocean–atmosphere system, with EN and La Niña (LN) as two particular phases of the oscillation (Wang and Picaut 2004) but also as a damped oscillation sustained by stochastic forcing (Moore and Kleeman 1999; Thompson and Battisti 2001; Zavala-Garay et al. 2003). Some other authors have a mixed point of view and consider ENSO as a weakly damped mode sustained by random disturbances (Philander and Fedorov 2003). This discussion has profound practical implications far beyond the theoretical sphere (Chen et al. 2004). The predictability of ENSO, when considered as a self-sustained oscillatory mode, would only be constrained by the growth of initial errors (Zebiak 1989; Goswami and Shukla 1991; Xue et al. 1997). Instead, if EN is interpreted as a temporal departure from normal conditions triggered by random disturbances, then predictability would be largely limited by natural noise, such as the westerly wind bursts (Penland and Sardeshmukh 1995; Perigaud and Cassou 2000; Boulanger et al. 2001).
Mechanistically speaking, at the seasonal time scale, the western tropical Pacific (WPAC) is a known key area for the activation and deactivation of ENSO (Wang et al. 1999; Boulanger and Menkes 2001; Boulanger et al. 2003). Off-equatorial ocean–atmosphere anomalies inducing equatorial westerly wind anomalies therein have been connected with the onset of most EN events since the regime shift around 1977 (Zhang et al. 1997; Weisberg and Wang 1997). These coupled ocean–atmosphere anomalies appear to propagate eastward via equatorial downwelling Kelvin waves and are connected with a positive feedback between suppressed equatorial easterlies and upwelling, enhanced zonal SST advection, and decreased Walker circulation (Bjerknes feedback; Bjerknes 1969; Wang and Fiedler 2006). Although ENSO is generally considered as a tropical ocean–atmosphere coupled mode of variability in the Pacific Ocean (Philander 1985), dynamical processes have also been described in the Southern and Northern Hemisphere subtropics and extratropics during the development of ENSO events (van Loon and Shea 1985; Lysne et al. 1997; Chan and Xu 2000; Anderson 2007).
In the present work, we introduce a new high-latitude tracer in the southern Pacific Ocean that is followed by EN events in the recent observational record around 9 months later on average. Lead–lag event composites of this new feature have been explored to describe the transition from an initial perturbation taking place in WPAC at long leads up to the mature phase of EN. These composites were also compared to the results of a complex empirical orthogonal function (CEOF) analysis capturing the rotational component of ENSO variability.
2. Data and methods
Monthly observations of satellite-derived sea surface temperature (SST) and sea ice concentration for the 1981–2008 period were used in this study [the National Oceanic and Atmospheric Administration (NOAA) Optimum Interpolation v2; Reynolds et al. (2002)]. The presatellite period was excluded from the analysis because SST estimations in the southern extratropical oceans are not fully reliable. The sea ice edge was identified by means of a concentration criterion of 50% (White et al. 2004). Only grid points free of sea ice for the whole time period were used in the analysis of SST (see the dashed thick line in Fig. 1e). Sea level pressure (SLP) and tropospheric winds were derived from the National Centers for Environmental Prediction (NCEP)–National Center for Atmospheric Research (NCAR) reanalysis (Kalnay et al. 1996; Kistler et al. 2001). Variables were standardized by subtracting the long-term mean and then dividing the difference by the standard deviation.
An eigendecomposition analysis (M = 40) was used to reconstruct variability at frequencies lower than the annual cycle for both the Ross–Bellingshausen (RB; difference between mean SST in [180°–160°W] × [65°–50°S] minus [100°–80°W] × [65°–50°S]) and Niño-3.4 (N34, mean SST in [170°–120°W] × [5S°–5°N]) indices (Figs. 2a,b; Dettinger et al. 1995). For comparison, a data-tapering window procedure to minimize spectral leakage was applied to each series to ensure no component was lost in the reconstruction. These reconstructed indices were used for the identification of the major RB peaks surpassing the +1 standard deviation threshold criterion (Fig. 3). This way, RB peaks were selected to occur in October 1986, March 1991, February 1997, January 2002, and October 2004 (Fig. 2a). RB peaks were then composited as a function of time lag in Figs. 2c, 4, 5, 6a–c, and 7a–c. Despite finding the same results for the 1982/83 RB and EN peaks (not shown), they were not included in the lead–lag composites because the SST dataset was only available since November 1981. In the composites, low-pass-filtered anomalies of spatiotemporal SST and SLP data were computed by a recursive Butterworth procedure (Moron and Plaut 2003), which was effective at removing high-frequency variability for periods lower than 18 months, therefore filtering out the annual cycle.
Principal component (PC) and CEOF analyses were applied to area-weighted low-pass-filtered SST anomalies in the southern high-latitude Pacific ([120°E–30°W] × [65°–50°S]; Fig. 1). CEOF is an analysis technique that decomposes variability into real and imaginary spatial maps that are amplified by real and imaginary time-varying coefficients, respectively (White and Annis 2004). These patterns together characterize the main modes of variability in the original spatiotemporal dataset as a function of the phase (φ) in periodic spatial coefficients (xyCEOFi) and periodic temporal scores (tCEOFi); that is, xyCEOFi(x, y, φ) = −xyCEOFi(x, y, φ + 180°) = xyCEOFi(x, y, φ + 360°) and tCEOFi(t, φ) = −tCEOFi(t, φ + 180°) = tCEOFi(t, φ + 360°) (von Storch and Zwiers 1999). The temporal scores at different phases were here correlated with the RB and N34 indices (Figs. 2a,b) and with low-pass-filtered spatiotemporal data (Figs. 6d–f and 7d–f).
Statistical significance (p < 0.05, unless specified otherwise) was computed by means of 1000 white-noise time series, which were then tested against an autoregressive process of order 1 having the same mean, standard deviation, and temporal one-lag autocorrelation value as the interannual RB index.
3. Results
In the search of a possible southern high-latitude tracer of ENSO activity far from its generation domain, a PC analysis of SST anomalies in the southern high-latitude Pacific was computed (Figs. 1a,b). The first two components define orthogonal phases of oceanic variability in the region. The first mode (PC1, 40%) describes a large-scale SST anomaly centered around 120°W, an oceanic feature that has been described in the literature as a local impact of the Pacific–South American (PSA; Ghil and Mo 1991) mode during and a few months after the peak of an EN event (e.g., Garreaud and Battisti 1999; see also Figs. 6c and 7c below). Instead, the second component (PC2, 21%), traditionally overlooked, depicts a zonal dipole near the Ross and Bellingshausen Seas, whose role within the ENSO mode of variability as a leading tracer of EN events is explored in the present work. More interestingly, a CEOF analysis of SST anomalies in the same region reveals that both PC components are indeed orthogonal states (around phases 105° and 195°) of the first CEOF mode (Figs. 1c–i). This rotating mode explains around half of the variability in the region, and it is basically characterized by an eastward displacement of the dipole, with strong amplification of SST anomalies when they reach longitude 120°W (i.e., PC1).
The RB index is here derived from these analyses and defined as the difference between SST anomalies near the Ross and Bellingshausen Seas (red and blue boxes in Fig. 1e, respectively). These areas are chosen as approximating the central regions for poles in PC2, where their difference approaches a maximum. The dark-blue curve in Fig. 2a depicts the recent interannual variability of the RB index. Simple temporal correlation analysis shows that the RB index can be alternatively expressed as the temporal score of the CEOF mode at phase 102° (light-blue line in Fig. 2a; r = 0.79, p < 0.001). Similarly, the interannual component of the N34 index (red curve in Fig. 2b) can also be represented through a particular phase (φ = 169°) of the CEOF mode (pink line in Fig. 2b; r = 0.78, p < 0.001).
The dipole is thus a feature of special interest, even though the magnitude of spatial coefficients in the CEOF mode is minimum during the dipole phases (e.g., magnitudes in Fig. 1e are much smaller than in Fig. 1h). ENSO is the main source of interannual variability, and the tropical Pacific is the main teleconnected region worldwide, therefore variability induced by EN in the southern high-latitude Pacific during its mature phase is expected to be, in general, larger compared to the amount of variability occurring there several months before the peak of EN. As a leading feature, the RB dipole must be therefore a pattern of lower amplitude. The difference in magnitude between both phases (RB dipole at 102° versus ENSO impact between 169° and 195°) might thus give a reasonable explanation to why this oceanic feature has been traditionally overlooked.
The main positive peaks of the RB index (see methods) were composited to explore the lead–lag relationship between the SST dipole and variability occurring in the tropical Pacific. Figure 2c depicts regional averages of lead–lag SST anomalies for a window size of more than 6 yr around RB peaks. These results show that positive RB peaks in the recent observational record were followed by EN events around 9 months later on average, but in turn, RB maxima were anticipated by positive SST anomalies occurring around a year in advance in the WPAC region (here defined as [140°–160°E] × [5°S–5°N]; gray boxes in Figs. 3a–e). This triple relationship highlights the status of the RB dipole as an effective and outstanding intermediate tracer of transient tropical Pacific dynamics, between the warm anomalies in the WPAC region and the subsequent onset of an EN event in the central and eastern tropical Pacific.
Despite the statistically significant relationship between the three features (WPAC, RB, and EN), individual events are subject to some minor differences with regard to the average picture. Figure 3 shows the configuration of SST anomalies in the tropical Pacific before and after each individual RB peak. Note that lagged warm SST anomalies in the N34 region (Figs. 3f–j) were indeed identified as EN events by the Climate Prediction Center (July 1987, Decemeber 1991, November 1997, October 2002, and February 2005, respectively; available online at http://www.cpc.noaa.gov/products/analysis_monitoring/ensostuff/ensoyears.shtml). Generally speaking, warmer than normal SST anomalies are observed a year in advance in WPAC and 9 months later in the N34 region. In some cases, however, the warm SST area in WPAC appears slightly shifted (e.g., to the west in Fig. 3b). In addition, the range of magnitudes between EN events is quite large, suggesting that the intensity of RB events and the subsequent maximum in the N34 index are not subject to a direct linear relationship. These nonlinearities are also expressed in the shorter-than-normal time–lag relationship between 2004/05 RB and N34 maxima (Fig. 3j).
Referring to the average composite picture, Figs. 4 and 5 describe the dynamical link between the WPAC region and the RB dipole. In these figures, SST, SLP, and wind anomalies were composited for the year prior to the peak of the RB dipole (monthly time lags −12, −09, −06, and −03). The warm SST anomaly area peaking in the WPAC region a year before RB maxima (Fig. 4a) is seen to initially enhance the local convection and the upper-troposphere divergence (Fig. 4c). Accordingly, the Walker circulation also appears to be strengthened, with upper- (lower-) troposphere westerly (easterly) wind anomalies across the tropical Pacific and downward motion in the central Pacific. Nevertheless, convection in the equatorial band is typically observed in the climatology to the west of the date line (e.g., Wang 2005), and so the ascending motion in the composites is weakened to the east of 155°E (Fig. 4c). The Hadley circulation is thus active in this area of anomalous subsidence, driving lower- (upper-) troposphere poleward (equatorward) wind anomalies in both hemispheres (Figs. 4b,d). The response is, however, clearly asymmetric, and so the southern Hadley cell is especially active, with poleward wind anomalies reaching much higher latitudes (around 50°S). Instead, the anomalous circulation in the northern counterpart is essentially confined to the tropics (until 20°N). Three months later, the warm SST area in WPAC starts to extend slightly to the northeast (Figs. 4e,g), close to the area of poleward low-level wind anomalies (Fig. 4h). Meanwhile, in the Southern Hemisphere, poleward surface winds appear to be embedded in a significant low pressure anomaly area arising around 35°S, with cyclonic wind anomalies becoming increasingly active in the western subtropical Pacific (Fig. 4f).
Half a year before the peak of the RB dipole, the low pressure area and cyclonic wind anomalies in the austral hemisphere become more active and propagate farther south to 40°S (Fig. 5b). An equivalent barotropic (upper-troposphere anomalies not shown here) anomalous wave train extending eastward and poleward thus results from the quasi-stationary Rossby wave response to the equatorial warming, convection, and upper-troposphere divergence (Trenberth et al. 1998). Thus, a secondary high pressure area (not significant yet) and counterclockwise wind anomalies appear around 155°W, 50°S (Fig. 5b). Poleward low-level geostrophic wind anomalies enhanced by the wave train lead to the anomalous warming of the ocean surface near the Ross Sea (Fig. 5a). Similarly, a significant low pressure area arises in the boreal hemisphere, with cyclonic wind anomalies around 20°N (Fig. 5b), and thus the ocean surface also warms in the northern tropical sector of the central Pacific (Fig. 5a). The latitudinally asymmetric response of the Hadley circulation, with subsequent low pressure areas arising at 40°S and 20°N, appears to be the key factor explaining why a northern high-latitude tracer, either expressed as a dipole or not, does not appear in the boreal hemisphere. Both warm areas in the northern tropical region and in the southern extratropics become increasingly strong compared to the decaying equatorial anomaly in the WPAC region. Three months before the RB maxima, the equatorial SST in the WPAC region finally disappears (Fig. 5c). In parallel with the off-equatorial warm anomaly near the date line, and pushed by equatorial westerly wind anomalies in WPAC, convection moves to the central Pacific, and as a consequence, the Walker circulation starts to weaken. The anomalous wave train in the austral hemisphere strengthens, as well as it does the cyclonic (anticyclonic) circulation around the low (high) pressure area in the southern high latitudes (Fig. 5d) and thus the wind-induced warm region near the Ross Sea (Fig. 5c).
The triple relationship arising among the WPAC, RB, and N34 regions is outlined in Figs. 6a–c and 7a–c, and then compared to the CEOF mode in Figs. 6d–f and 7d–f. For such a comparison, the rotating mode was generalized farther away from this oceanic region, and associated large-scale patterns were displayed by means of Pearson correlations between the temporal score of the CEOF mode at a particular phase and spatiotemporal SST anomalies. Note that this procedure is indeed a generalization of the rotating mode for other regions because it keeps the same SST configuration within the southern high-latitude Pacific (cf. Figs. 1c,e,g with Figs. 6d–f).
These results show that variability occurring in the Pacific Ocean during the onset of EN is ultimately linked to the WPAC region (Figs. 6a and 7a). However, the warm anomaly developing in WPAC can appear even when positive, neutral, or nonsignificant negative SST anomalies exist in the central and eastern tropical Pacific, though in some cases it can equally occur following a typical LN-like warm horseshoe pattern (see discussion below). Indeed, N34 SST anomalies around lag −12 are shown to be weak and do not stand out as significant (Fig. 2c), showing that the origin of the RB dipole is not necessarily linked to a previous LN event. Note that the equivalent barotropic anomalous wave train in the austral hemisphere defines an atmospheric tripole configuration that resembles one of the PSA-like modes shown by Mo 2000 (Fig. 7b). Thus, simultaneously to the warm region appearing near the Ross Sea, a cold SST anomaly is generated near the Bellingshausen Sea, making up the so-called RB dipole (Fig. 6b), whose intensity is maximum a year after the SST peak in the WPAC area (Fig. 2c).
As part of this overall ENSO cycle, the warm SST region in the central tropical Pacific keeps moving to the east in parallel with the strengthening of the Bjerknes feedback (Figs. 6b and 7b). This ocean–atmosphere coupling exponentially grows on the initial warming, and when the enhanced westerlies reach the central Pacific, an EN event rapidly grows in the eastern Pacific (Figs. 6c and 7c). The atmospheric configuration appearing in the southern Pacific during the mature phase of EN corresponds to the PSA mode (Fig. 7c; Lau et al. 1994), which is a teleconnection pattern linking tropical variability with the Southern Hemisphere (Mo and Higgins 1998). This mode of variability has been traditionally associated with the impact of ENSO events in the interannual band in the southern Pacific (Mo and Paegle 2001). The particular impact of the PSA mode in the southern high-latitude Pacific is basically characterized by a warm SST region forced by local winds that reaches its maximum intensity just after the mature phase of EN (Figs. 1g,h and 6c). This warm SST region indeed corresponds to the anomaly appearing near the Ross Sea 9 months before, which moved eastward after the RB peak in parallel with the tropical warming and convection and the southern wave train (cf. Figs. 1e,g with Figs. 6b,c and 7b,c).
The CEOF mode strikingly mimics the evolution shown in the lead–lag composite analysis. Here, we show values for SST and SLP, but analogous results are found for other variables (e.g., geopotential height and winds). An initial warming takes place in the WPAC region, although the warm SST region is embedded in a LN-like warm horseshoe pattern owing to the characteristics of the technique (phase 45° in Figs. 6d and 7d; see discussion below). Again, the equivalent barotropic atmospheric wave train in the austral hemisphere evolves in parallel with the thermal dipole configuration of the underlying ocean surface (phase 105° in Figs. 6e and 7e). Meanwhile, the warm SST anomaly in the tropical region arrives in the central Pacific, it weakens the Walker circulation, and favors the positive phase of ENSO (phase 165° in Figs. 6f and 7f).
4. Discussion
In the present work, a CEOF analysis was applied to SST data in the southern high-latitude Pacific, in the search of a proper description of the onset of EN events. This methodology was compared to lead–lag peak composites of a new high-latitude tracer of tropical Pacific dynamics, the RB dipole. Both methodologies appeared to successfully trace the transition among the WPAC, RB, and N34 regions during the onset of EN. However, the CEOF mode also displays some differences when compared to the lead–lag composites of RB peaks, which need to be discussed.
As a rotating mode, spatial coefficients of the CEOF component are cyclic by construction [i.e., SST(x, y, φ) = −SST(x, y, φ + 180°) = SST(x, y, φ + 360°)]. The phase evolution of the CEOF is thus defined by positive values in the WPAC, RB, and N34 regions (WPAC+, RB+, and EN), and then by their negative counterparts (WPAC−, RB−, and LN). Figure 8 summarizes this cyclic component of ENSO variability occurring in the Pacific Ocean. Note that the largest amount of variability within this rotating mode is observed during the peak of EN and LN. The cyclic nature of the CEOF, which optimizes the extraction of features involving a large amount of variance, masks the incipient warm spot in the WPAC region, which occurs in the same tropical area and at similar phases than the preceding LN warm horseshoe; and therefore, it highlights this much wider feature occurring in the tropical Pacific.
Hence, as an inherent feature of the circularity of the methodology, the initial warm SST area in the WPAC region is forced to appear circumscribed within a clearly established LN event, together with its subsequent oceanic warm horseshoe (Figs. 6d and 7d). Nevertheless, according to the alternate analyses, the only oceanic LN-like feature that seems to be a necessary prerequisite for the appearance of the RB dipole is the (much more local) equatorial warming in the WPAC region (Figs. 2c and 6a). Although many different processes might be interacting during the generation of the initial WPAC warm spot in each particular individual event (e.g., a moderate LN event was observed the year before the February 1997 positive RB peak but not before the March 1991 RB event; Figs. 2a,b), lead–lag composites tend to highlight those common features appearing among the averaged individual cases (i.e., the local warming in WPAC).
ENSO variability is, however, far more complex than the simplification provided by a cyclic component. To begin with, EN and LN events exhibit some degree of asymmetry (Wang and Fiedler 2006). In addition, the cyclic approximation particularly fails to reproduce the observed interactions between the cold phases of anomalies in the WPAC, RB, and N34 regions (WPAC−, RB−, and LN in Fig. 8, respectively). For example, the November 1988 LN event was followed by a negative RB peak 3 months later (Figs. 2a,b). The maximum resemblance between the CEOF mode and the lead–lag composites is thus observed during the positive phases of anomalies occurring in the RB and N34 regions (RB+ and EN in Fig. 8).
Despite all these differences, the CEOF mode described in the present work provides a new way to interpret low-frequency variability in the tropical Pacific. In addition, it further stresses the role of the southern high-latitude Pacific as a privileged nearly 90° out-of-phase extratropical observer of tropical dynamics in the Pacific Ocean. The ability to trace tropical dynamics particularly shows up both when the largest peaks of the dipole are composited and when a CEOF analysis of the region is computed. Regarding the role of the RB dipole, there is no evidence that might suggest that it actively forces or modulates the occurrence of EN events. In that way, the RB dipole might be seen as a leading extratropical surrogate signature of an eastward-propagating EN event in the tropical Pacific. As such, this new tracer does not stand in contradiction to current science, in the sense that tropical SST anomalies have the largest impact on atmospheric variability (e.g., Sterl and Hazeleger 2005).
This new high-latitude tracer cannot be considered, however, to lead the whole ENSO variability, but instead only EN, because it is intimately linked to tropical variability occurring in the WPAC region during the onset of eastward propagating EN events. Thus, while the RB dipole is a high-latitude tracer that leads the mature phase of EN, it is at the same time composed within the overall envelope of ENSO variability, and therefore it is anticipated by other earlier stages of the recurring ENSO phenomenon. The CEOF analysis thus further stresses this interpretation of ENSO variability as an ordered sequence of features appearing in the WPAC, RB, and N34 regions (e.g., Jin and Kirtman 2009).
An example of extratropical ENSO-forced pattern leading to EN peaks was shown by Karoly (1989) for the 1972/83 period, who described a weak upper-troposphere anomaly wave train extending poleward and eastward in the southern Pacific during the developing stage of an EN event. According to Karoly (1989), this feature typically occurs in summer, two seasons in advance of the seasonal winter peak of EN. Note that this feature also appears in the lead–lag composites shown in the present work (i.e., time lag +03 after the RB peak; see also Kidson and Renwick 2002). Moreover, Vimont et al. (2001, 2003) showed an elongated warm SST anomaly in the northern tropical and subtropical Pacific, imparted by midlatitude intrinsic atmospheric variability (seasonal-footprinting mechanism). This anomaly typically occurs in winter, a year before the seasonal winter peak of EN. As expected, this feature also appears in our lead–lag composites, this time at time lag −03 before the RB peak (Figs. 5c,d).
Our results showed that the oceanic RB dipole is coupled with an equivalent barotropic atmospheric tripole (Figs. 6b and 7b). The ocean–atmosphere interaction giving rise to the warm (cold) area near the Ross (Bellingshausen) Sea is explained by poleward (equatorward) low-level geostrophic wind anomalies driven by low (high) SLP anomalies to the west and high (low) values to the east (White and Chen 2002). This mechanism ultimately originates the out-of-phase juxtaposition of SLP−/SST+/SLP+/SST−/SLP− anomalies along the southern extratropical Pacific. On the other hand, warm SST anomalies are also associated with latent heat flux-induced lower- (upper-) level diabatic cooling (warming), which is balanced by deep convection and poleward wind throughout the whole tropospheric column above the surface ocean warming (deep diabatic heating scenario; find further details in White et al. 2004). Both mechanisms define a positive feedback between ocean temperatures and atmospheric circulation in the extratropical southwestern Pacific, which might be initially stimulated by the quasi-stationary Rossby wave response to the equatorial warming and the active southern Hadley circulation (e.g., see Figs. 4d,h for local northerlies and ascending motion in the southern extratropics). Once the feedback has been activated, the coupled ocean dipole–atmosphere tripole system can persist for a long time after the WPAC warming has decayed.
Many authors have used alternative techniques for the exploration of features giving rise to EN events. For example, this is the case for optimal perturbations of SST, which isolate the most rapidly growing perturbations in a system where dynamics is assumed to be linear (Farrell 1982). This technique has been applied to a wide range of coupled model simulations, including those from purely dynamical (Chen et al. 1997) or hybrid (Thompson and Battisti 2001) models and leads to very different results in each case (Moore and Kleeman 2001). This kind of analysis has been historically restricted to tropical latitudes in the Pacific Ocean, and model simulations usually run for time periods not longer than half a year (Penland and Sardeshmukh 1995; Thompson 1998; Moore and Kleeman 2001), due to the computationally intensive nature of these calculations. These methodological constraints might give a reasonable explanation to why the high-latitude RB dipole and the long-lead precursor in the WPAC region have been traditionally overlooked in these studies. In addition, most of the leading positive SST anomalies isolated by means of this technique are restricted to the eastern tropical Pacific (Chen et al. 1997; Kug et al. 2010), suggesting that climate models in these studies mostly reproduce westward-propagating EN events. Under all these circumstances, these studies and our results are not easily comparable.
Different theories have been invoked to explain how EN events are generated, but a clear description of why and how EN is activated is still an issue open to debate (Wang and Fiedler 2006). The two methodologies described in the present work, analyzed together, might suggest new clues for a better understanding of the nature and origin of EN events, which might be accommodated into a common unified framework. In that search, our CEOF analysis has shown that a large amount of low-frequency ENSO variability (e.g., Fig. 2b) can be fairly well approximated by a cyclic mode of variability (Fig. 8). Although this periodic mode (LN–EN–LN) might be valid for some of the analyzed individual EN events, lead–lag composites additionally suggest that the cyclic approach is not generally applicable and that some of the EN events are initially preceded by only a much more local warm SST area in the WPAC region. The origin of this initial perturbation might be, for instance, related to high-frequency stochastic forcing and the subsequent collapse of the trade winds in the WPAC region. Alternatively, it might be linked to slow eastward-propagating ocean–atmosphere waves in the southern extratropics, which traveled equatorward across the Indian Ocean and up to the WPAC region after the 1977 regime shift (cf. Figs. 6a and 6b in White and Annis 2004). This issue is far beyond the scope of this study, and it will require further research. One way or the other, it seems clear that this mechanism is in principle only valid for eastward-propagating EN events. These kind of events have become dominant during the time period with available satellite observations, and therefore it is impossible to determine whether other tracers in the southern high-latitude Pacific were observed in previous time periods where other types of EN predominated.
5. Conclusions
The role of the RB dipole as a new tracer for the onset of recent eastward EN events has been established in the present work. This precursory area in the southern high-latitude Pacific was followed by EN events around, on average, 9 months later. The RB dipole is also found to occur a year after the development of a warm SST anomaly in the WPAC region, which generates an anomalous wave train extending eastward and poleward in the Southern Hemisphere. Changes in the circulation lead to warm SST anomalies in the central tropical Pacific, which are then strengthened by suppressed equatorial easterlies. Convection thus starts to move to the central Pacific. This process is associated to a weakening of the Walker circulation, setting up the positive Bjerknes feedback that exponentially grows on top of the incipient warming. The tropical warm anomaly then propagates eastward heading to the central Pacific, where an EN event rapidly starts to develop.
Acknowledgments
JB was in receipt of a fellowship from the Catalan Ministry of Innovation and Science. XR was in receipt of funds from a NOAA/NSF grant and from the Spanish Ministry of Innovation and Science (MICINN) through the PANDORA Project (CGL2007-63053). Authors want to expressly thank the very useful comments provided by Dan Cayan, Dave Pierce, and three anonymous reviewers.
REFERENCES
Anderson, B. T., 2007: Intraseasonal atmospheric variability in the extratropics and its relation to the onset of tropical Pacific sea surface temperature anomalies. J. Climate, 20 , 926–936.
Barnett, T. P., M. Latif, N. E. Graham, M. Flugel, S. Pazan, and W. White, 1993: ENSO and ENSO-related predictability. I: Prediction of equatorial Pacific sea surface temperature with a hybrid coupled ocean–atmosphere model. J. Climate, 6 , 1545–1566.
Battisti, D. S., and E. S. Sarachik, 1995: Understanding and predicting ENSO. Rev. Geophys., 33 , 1367–1376.
Bjerknes, J., 1969: Atmospheric teleconnections from the equatorial Pacific. Mon. Wea. Rev., 97 , 163–172.
Boulanger, J. P., and C. Menkes, 2001: The trident Pacific model. Part 2: Role of long equatorial wave reflection on sea surface temperature anomalies during the 1993–1998 TOPEX/POSEIDON period. Climate Dyn., 17 , 175–186.
Boulanger, J. P., and Coauthors, 2001: Role of non-linear oceanic processes in the response to westerly wind events: New implications for the 1997 El Niño onset. Geophys. Res. Lett., 28 , 1603–1606.
Boulanger, J. P., S. Cravatte, and C. Menkes, 2003: Reflected and locally wind forced interannual Kelvin waves in the western Pacific Ocean. J. Geophys. Res., 108 , 3311. doi:10.1029/2002JC001760.
Cane, M. A., S. E. Zebiak, and S. C. Dolan, 1986: Experimental forecasts of El Niño. Nature, 321 , 827–832.
Chan, J. C. 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.
Chang, R., Q. Y. Zhang, and R. F. Li, 2009: North Pacific premonitory sign of the ENSO event. Geophys. Res. Lett., 36 , L03818. doi:10.1029/2008GL036597.
Changnon, S. A., 1999: Impacts of 1997/98 El Niño generated weather in the United States. Bull. Amer. Meteor. Soc., 80 , 1819–1828.
Chen, D., and M. A. Cane, 2008: El Niño prediction and predictability. J. Comput. Phys., 227 , 3625–3640.
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 , 733–736.
Chen, Y. Q., D. S. Battisti, T. N. Palmer, J. Barsugli, and E. S. Sarachik, 1997: A study of the predictability of tropical Pacific SST in a coupled atmosphere–ocean model using singular vector analysis: The role of the annual cycle and the ENSO cycle. Mon. Wea. Rev., 125 , 831–845.
Clarke, A. J., 2008: An Introduction to the Dynamics of El Niño and the Southern Oscillation. Elsevier, 308 pp.
CPC, 1997: Special Climate Summary, 97/3. NWS, NOAA, 14 pp.
Deng, Z., Y. Tang, and X. Zhou, 2009: The retrospective prediction of El Niño–Southern Oscillation from 1881 to 2000 by a hybrid coupled model. I: Sea surface temperature assimilation with ensemble Kalman filter. Climate Dyn., 32 , 397–413.
Dettinger, M. D., C. M. Strong, W. Weibel, M. Ghil, and P. Yiou, 1995: Software for singular spectrum analysis of noisy time series. Eos, Trans. Amer. Geophys. Union, 76 , 12–14.
Farrell, B. F., 1982: The initial growth of disturbances in baroclinic flow. J. Atmos. Sci., 39 , 1663–1686.
Garreaud, R. D., and D. S. Battisti, 1999: Interannual (ENSO) and interdecadal (ENSO-like) variability in the Southern Hemisphere tropospheric circulation. J. Climate, 12 , 2113–2123.
Ghil, M., and K. C. Mo, 1991: Intraseasonal oscillations in the global atmosphere. Part II: Southern Hemisphere. J. Atmos. Sci., 48 , 780–790.
Glantz, M. H., 2001: Currents of Change: Impacts of El Niño and La Niña on Climate and Society. Cambridge University Press, 252 pp.
Goswami, B. N., and J. Shukla, 1991: Predictability of a coupled ocean–atmosphere model. J. Climate, 4 , 3–22.
Graham, N. E., J. Michaelsen, and T. P. Barnett, 1987: An investigation of the El Niño–Southern Oscillation cycle with statistical models. 2. Model results. J. Geophys. Res., 92 , 14271–14289.
Jin, D., and B. P. Kirtman, 2009: Why the Southern Hemisphere ENSO responses lead ENSO. J. Geophys. Res., 114 , D23101. doi:10.1029/2009JD012657.
Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77 , 437.
Karoly, D. J., 1989: Southern Hemisphere circulation features associated with El Niño–Southern Oscillation. J. Climate, 2 , 1239–1252.
Kidson, J. W., and J. A. Renwick, 2002: The Southern Hemisphere evolution of ENSO during 1981–99. J. Climate, 15 , 847–863.
Kiladis, G. N., and H. F. Diaz, 1989: Global climatic anomalies associated with extremes in the Southern Oscillation. J. Climate, 2 , 1069–1090.
Kirtman, B. P., J. Shukla, B. Huang, Z. Zhu, and E. K. Schneider, 1997: Multiseasonal predictions with a coupled tropical ocean global atmosphere system. Mon. Wea. Rev., 125 , 789–808.
Kistler, R., and Coauthors, 2001: The NCEP/NCAR 50 Year Reanalysis: Monthly means CD-ROM and documentation. Bull. Amer. Meteor. Soc., 82 , 247.
Klein, S. A., B. J. Soden, and N. C. Lau, 1999: Remote sea surface temperature variations during ENSO: Evidence for a tropical atmospheric bridge. J. Climate, 12 , 917–932.
Kug, J. S., Y. G. Ham, M. Kimoto, F. F. Jin, and I. S. Kang, 2010: New approach for optimal perturbation method in ensemble climate prediction with empirical singular vector. Climate Dyn., 35 , 331–340.
Latif, M., T. P. Barnett, M. A. Cane, M. Flügel, N. E. Graham, H. Storch, J. S. Xu, and S. E. Zebiak, 1994: A review of ENSO prediction studies. Climate Dyn., 9 , 167–179.
Latif, M., and Coauthors, 1998: A review of the predictability and prediction of ENSO. J. Geophys. Res., 103 , 14375–14393.
Lau, K. M., P. J. Sheu, and I. S. Kang, 1994: Multiscale low-frequency circulation modes in the global atmosphere. J. Atmos. Sci., 51 , 1169–1193.
Lau, N. C., and M. J. Nath, 2001: Impact of ENSO on SST variability in the North Pacific and North Atlantic: Seasonal dependence and role of extratropical sea–air coupling. J. Climate, 14 , 2846–2866.
Luo, J. J., S. Masson, S. Behera, S. Shingu, and T. Yamagata, 2005: Seasonal climate predictability in a coupled OAGCM using a different approach for ensemble forecasts. J. Climate, 18 , 4474–4497.
Lysne, J., P. Chang, and B. Giese, 1997: Impact of the extratropical Pacific on equatorial variability. Geophys. Res. Lett., 24 , 2589–2592.
McPhaden, M. J., and Coauthors, 1998: The Tropical Ocean–Global Atmosphere observing system: A decade of progress. J. Geophys. Res., 103 , 14169–14240.
Mo, K. C., 2000: Relationships between low-frequency variability in the Southern Hemisphere and sea surface temperature anomalies. J. Climate, 13 , 3599–3610.
Mo, K. C., and R. W. Higgins, 1998: The Pacific–South American modes and tropical convection during the Southern Hemisphere winter. Mon. Wea. Rev., 126 , 1581–1596.
Mo, K. C., and J. N. Paegle, 2001: The Pacific–South American modes and their downstream effects. Int. J. Climatol., 21 , 1211–1229.
Moore, A. M., and R. Kleeman, 1999: Stochastic forcing of ENSO by the intraseasonal oscillation. J. Climate, 12 , 1199–1220.
Moore, A. M., and R. Kleeman, 2001: The differences between the optimal perturbations of coupled models of ENSO. J. Climate, 14 , 138–163.
Moron, V., and G. Plaut, 2003: The impact of El Niño–Southern Oscillation upon weather regimes over Europe and the North Atlantic during boreal winter. Int. J. Climatol., 23 , 363.
Pascual, M., X. Rodó, S. P. Ellner, R. Colwell, and M. J. Bouma, 2000: Cholera dynamics and El Niño–Southern Oscillation. Science, 289 , 1766–1769.
Penland, C., and P. D. Sardeshmukh, 1995: The optimal growth of tropical sea surface temperature anomalies. J. Climate, 8 , 1999–2024.
Perigaud, C. M., and C. Cassou, 2000: Importance of oceanic decadal trends and westerly wind bursts for forecasting El Niño. Geophys. Res. Lett., 27 , 389–392.
Pezzi, L. P., and I. F. A. Cavalcanti, 2001: The relative importance of ENSO and tropical Atlantic sea surface temperature anomalies for seasonal precipitation over South America: A numerical study. Climate Dyn., 17 , 205–212.
Philander, S. G., 1985: El Niño and La Niña. J. Atmos. Sci., 42 , 2652–2662.
Philander, S. G., and A. Fedorov, 2003: Is El Niño sporadic or cyclic? Annu. Rev. Earth Planet. Sci., 31 , 579–594.
Reynolds, R. W., N. A. Rayner, T. M. Smith, D. C. Stokes, and W. Wang, 2002: An improved in situ and satellite SST analysis for climate. J. Climate, 15 , 1609–1625.
Ropelewski, C. F., and M. S. Halpert, 1987: Global- and regional-scale precipitation patterns associated with the El Nino–Southern Oscillation. Mon. Wea. Rev., 115 , 1606–1626.
Sterl, A., and W. Hazeleger, 2005: The relative roles of tropical and extratropical forcing on atmospheric variability. Geophys. Res. Lett., 32 , L18716. doi:10.1029/2005GL023757.
Tang, F. T., W. W. Hsieh, and B. Tang, 1997: Forecasting the equatorial Pacific sea surface temperature by neural network models. Climate Dyn., 13 , 135–147.
Thompson, C. J., 1998: Initial conditions for optimal growth in a coupled ocean–atmosphere model of ENSO. J. Atmos. Sci., 55 , 537–557.
Thompson, C. J., and D. S. Battisti, 2001: A linear stochastic dynamical model of ENSO. Part II: Analysis. J. Climate, 14 , 445–466.
Trenberth, K. E., 1997: The definition of El Niño. Bull. Amer. Meteor. Soc., 78 , 2771–2777.
Trenberth, K. E., G. Branstator, D. Karoly, A. Kumar, N. C. Lau, and C. Ropelewski, 1998: Progress during TOGA in understanding and modeling global teleconnections associated with tropical sea surface temperatures. J. Geophys. Res., 103 , 14291–14324.
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.
Vimont, D. J., D. S. Battisti, and A. C. Hirst, 2001: Footprinting: A seasonal connection between the tropics and midlatitudes. Geophys. Res. Lett., 28 , 3923–3926.
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., and F. W. Zwiers, 1999: Statistical Analysis in Climate Research. Cambridge University Press, 499 pp.
Wang, C., 2005: ENSO, Atlantic climate variability, and the Walker and Hadley circulations. The Hadley Circulation: Present, Past and Future, H. F. Diaz and R. S. Bradley, Eds., Springer 173–202.
Wang, C., and P. C. Fiedler, 2006: ENSO variability and the eastern tropical Pacific: A review. Prog. Oceanogr., 69 , 239–266.
Wang, C., and J. Picaut, 2004: Understanding ENSO physics, A review. Earth’s Climate: The Ocean–Atmosphere Interaction. Geophys. Monogr. Vol. 147, Amer. Geophys. Union, 21–48.
Wang, C., R. H. Weisberg, and J. I. Virmani, 1999: Western Pacific interannual variability associated with the El Niño–Southern Oscillation. J. Geophys. Res., 104 , 5131–5149.
Weisberg, R. H., and C. Wang, 1997: A western Pacific oscillator paradigm for the El Niño–Southern Oscillation. Geophys. Res. Lett., 24 , 779–782.
White, W. B., and S. S. Chen, 2002: Thermodynamic mechanisms responsible for the tropospheric response to sst anomalies in the antarctic circumpolar wave. J. Climate, 15 , 2577–2596.
White, W. B., and J. Annis, 2004: Influence of the Antarctic circumpolar wave on El Niño and its multidecadal changes from 1950 to 2001. J. Geophys. Res., 109 , C06019. doi:10.1029/2002JC001666.
White, W. B., P. Gloersen, and I. Simmonds, 2004: Tropospheric response in the Antarctic circumpolar wave along the sea ice edge around Antarctica. J. Climate, 17 , 2765–2779.
Xue, Y., M. A. Cane, and S. E. Zebiak, 1997: Predictability of a coupled model of ENSO using singular vector analysis. Part I: Optimal growth in seasonal background and ENSO cycles. Mon. Wea. Rev., 125 , 2043–2056.
Xue, Y., A. Leetmaa, and M. Ji, 2000: ENSO prediction with Markov models: The impact of sea level. J. Climate, 13 , 849–871.
Zavala-Garay, J., A. M. Moore, C. L. Perez, and R. Kleeman, 2003: The response of a coupled model of ENSO to observed estimates of stochastic forcing. J. Climate, 16 , 2827–2842.
Zebiak, S. E., 1989: On the 30–60 day oscillation and the prediction of El Niño. J. Climate, 2 , 1381–1387.
Zhang, Y., J. M. Wallace, and D. S. Battisti, 1997: ENSO-like interdecadal variability: 1900–93. J. Climate, 10 , 1004–1020.
Leading modes of area-weighted low-pass-filtered SST anomalies in the southern high-latitude Pacific Ocean. (a),(b) Spatial coefficients of the first two PCs. Spatial coefficients of the first CEOF at phases (c) 45°, (d) 75°, (e) 105°, (f) 135°, (g) 165°, (h) 195°, and (i) 225°. Note that the spatial configuration at phase φ is by definition the opposite to phase φ + 180° [e.g., (c) and (i) have equal magnitude but opposite sign]. The red and blue rectangles in (e) depict the areas for the computation of the RB dipole.
Citation: Journal of Climate 24, 5; 10.1175/2010JCLI3619.1
Leading modes of area-weighted low-pass-filtered SST anomalies in the southern high-latitude Pacific Ocean. (a),(b) Spatial coefficients of the first two PCs. Spatial coefficients of the first CEOF at phases (c) 45°, (d) 75°, (e) 105°, (f) 135°, (g) 165°, (h) 195°, and (i) 225°. Note that the spatial configuration at phase φ is by definition the opposite to phase φ + 180° [e.g., (c) and (i) have equal magnitude but opposite sign]. The red and blue rectangles in (e) depict the areas for the computation of the RB dipole.
Citation: Journal of Climate 24, 5; 10.1175/2010JCLI3619.1
Leading modes of area-weighted low-pass-filtered SST anomalies in the southern high-latitude Pacific Ocean. (a),(b) Spatial coefficients of the first two PCs. Spatial coefficients of the first CEOF at phases (c) 45°, (d) 75°, (e) 105°, (f) 135°, (g) 165°, (h) 195°, and (i) 225°. Note that the spatial configuration at phase φ is by definition the opposite to phase φ + 180° [e.g., (c) and (i) have equal magnitude but opposite sign]. The red and blue rectangles in (e) depict the areas for the computation of the RB dipole.
Citation: Journal of Climate 24, 5; 10.1175/2010JCLI3619.1
(Dark blue and red lines) Standardized (unitless) interannual component of monthly SST in the (a) RB and (b) N34 regions, respectively. (Light blue and pink lines) Standardized (unitless) temporal scores of the CEOF at phases (a) 102° and (b) 169°, respectively, at which the largest Pearson correlation with the RB and N34 indices are found. (c) Lead–lag standardized (unitless) low-pass-filtered SST anomalies for the WPAC area (gray), the Ross (red) and Bellingshausen (blue) Seas, and the N34 region (black). Anomalies were averaged for the 5 RB peaks reaching the +1 standard deviation criterion (see Fig. 3). Circles in (c) indicate significant anomalies.
Citation: Journal of Climate 24, 5; 10.1175/2010JCLI3619.1
(Dark blue and red lines) Standardized (unitless) interannual component of monthly SST in the (a) RB and (b) N34 regions, respectively. (Light blue and pink lines) Standardized (unitless) temporal scores of the CEOF at phases (a) 102° and (b) 169°, respectively, at which the largest Pearson correlation with the RB and N34 indices are found. (c) Lead–lag standardized (unitless) low-pass-filtered SST anomalies for the WPAC area (gray), the Ross (red) and Bellingshausen (blue) Seas, and the N34 region (black). Anomalies were averaged for the 5 RB peaks reaching the +1 standard deviation criterion (see Fig. 3). Circles in (c) indicate significant anomalies.
Citation: Journal of Climate 24, 5; 10.1175/2010JCLI3619.1
(Dark blue and red lines) Standardized (unitless) interannual component of monthly SST in the (a) RB and (b) N34 regions, respectively. (Light blue and pink lines) Standardized (unitless) temporal scores of the CEOF at phases (a) 102° and (b) 169°, respectively, at which the largest Pearson correlation with the RB and N34 indices are found. (c) Lead–lag standardized (unitless) low-pass-filtered SST anomalies for the WPAC area (gray), the Ross (red) and Bellingshausen (blue) Seas, and the N34 region (black). Anomalies were averaged for the 5 RB peaks reaching the +1 standard deviation criterion (see Fig. 3). Circles in (c) indicate significant anomalies.
Citation: Journal of Climate 24, 5; 10.1175/2010JCLI3619.1
Lead–lag standardized (unitless) low-pass-filtered SST anomalies for the 5 RB peaks reaching the +1 standard deviation criterion. Anomalies are shown for monthly time lags (a)–(e) −12, (f)–(i) +09, and (j) +04 before (negative lags) or after (positive lags) the RB peaks in (a),(f) October 1986, (b),(g) March 1991, (c),(h) February 1997, (d),(i) January 2002, and (e),(j) October 2004. Gray and red rectangles depict the WPAC and N34 regions, respectively.
Citation: Journal of Climate 24, 5; 10.1175/2010JCLI3619.1
Lead–lag standardized (unitless) low-pass-filtered SST anomalies for the 5 RB peaks reaching the +1 standard deviation criterion. Anomalies are shown for monthly time lags (a)–(e) −12, (f)–(i) +09, and (j) +04 before (negative lags) or after (positive lags) the RB peaks in (a),(f) October 1986, (b),(g) March 1991, (c),(h) February 1997, (d),(i) January 2002, and (e),(j) October 2004. Gray and red rectangles depict the WPAC and N34 regions, respectively.
Citation: Journal of Climate 24, 5; 10.1175/2010JCLI3619.1
Lead–lag standardized (unitless) low-pass-filtered SST anomalies for the 5 RB peaks reaching the +1 standard deviation criterion. Anomalies are shown for monthly time lags (a)–(e) −12, (f)–(i) +09, and (j) +04 before (negative lags) or after (positive lags) the RB peaks in (a),(f) October 1986, (b),(g) March 1991, (c),(h) February 1997, (d),(i) January 2002, and (e),(j) October 2004. Gray and red rectangles depict the WPAC and N34 regions, respectively.
Citation: Journal of Climate 24, 5; 10.1175/2010JCLI3619.1
Lead–lag standardized (unitless) low-pass-filtered SST, SLP, and wind anomalies averaged for the 5 RB peaks reaching the +1 standard deviation criterion (see Fig. 3). Anomalies are shown for monthly time lags (a)–(d) −12 and (e)–(h) −09 before the RB peaks. (a),(b),(e),(f) SST, SLP (shaded areas), and horizontal surface wind (arrows) anomalies. (c),(g) Zonal, vertical (arrows), and meridional (shaded areas, positive means northward) wind components within the brown rectangle shown in the maps. (d),(h) Meridional, vertical (arrows), and zonal (shaded areas, positive means eastward) wind components within the green rectangle shown in the maps. Note that only significant anomalies are shown in the shading.
Citation: Journal of Climate 24, 5; 10.1175/2010JCLI3619.1
Lead–lag standardized (unitless) low-pass-filtered SST, SLP, and wind anomalies averaged for the 5 RB peaks reaching the +1 standard deviation criterion (see Fig. 3). Anomalies are shown for monthly time lags (a)–(d) −12 and (e)–(h) −09 before the RB peaks. (a),(b),(e),(f) SST, SLP (shaded areas), and horizontal surface wind (arrows) anomalies. (c),(g) Zonal, vertical (arrows), and meridional (shaded areas, positive means northward) wind components within the brown rectangle shown in the maps. (d),(h) Meridional, vertical (arrows), and zonal (shaded areas, positive means eastward) wind components within the green rectangle shown in the maps. Note that only significant anomalies are shown in the shading.
Citation: Journal of Climate 24, 5; 10.1175/2010JCLI3619.1
Lead–lag standardized (unitless) low-pass-filtered SST, SLP, and wind anomalies averaged for the 5 RB peaks reaching the +1 standard deviation criterion (see Fig. 3). Anomalies are shown for monthly time lags (a)–(d) −12 and (e)–(h) −09 before the RB peaks. (a),(b),(e),(f) SST, SLP (shaded areas), and horizontal surface wind (arrows) anomalies. (c),(g) Zonal, vertical (arrows), and meridional (shaded areas, positive means northward) wind components within the brown rectangle shown in the maps. (d),(h) Meridional, vertical (arrows), and zonal (shaded areas, positive means eastward) wind components within the green rectangle shown in the maps. Note that only significant anomalies are shown in the shading.
Citation: Journal of Climate 24, 5; 10.1175/2010JCLI3619.1
As in Figs. 4a,b,e,f, but for monthly time lags (a),(b) −06 and (c),(d) −03 before the RB peaks.
Citation: Journal of Climate 24, 5; 10.1175/2010JCLI3619.1
As in Figs. 4a,b,e,f, but for monthly time lags (a),(b) −06 and (c),(d) −03 before the RB peaks.
Citation: Journal of Climate 24, 5; 10.1175/2010JCLI3619.1
As in Figs. 4a,b,e,f, but for monthly time lags (a),(b) −06 and (c),(d) −03 before the RB peaks.
Citation: Journal of Climate 24, 5; 10.1175/2010JCLI3619.1
(a)–(c) Lead–lag standardized (unitless) low-pass-filtered SST anomalies averaged for the 5 RB peaks reaching the +1 standard deviation criterion (see Fig. 3). Anomalies are shown for monthly time lags −09, +00, and +09 before (negative lags) or after (positive lags) the RB peaks. (d)–(f) Pearson correlations between the temporal scores of the CEOF shown in Figs. 1c,e,g and low-pass-filtered anomalies of global spatiotemporal SST. Panels correspond to CEOF phases 45°, 105°, and 165°. Shaded areas indicate significant anomalies.
Citation: Journal of Climate 24, 5; 10.1175/2010JCLI3619.1
(a)–(c) Lead–lag standardized (unitless) low-pass-filtered SST anomalies averaged for the 5 RB peaks reaching the +1 standard deviation criterion (see Fig. 3). Anomalies are shown for monthly time lags −09, +00, and +09 before (negative lags) or after (positive lags) the RB peaks. (d)–(f) Pearson correlations between the temporal scores of the CEOF shown in Figs. 1c,e,g and low-pass-filtered anomalies of global spatiotemporal SST. Panels correspond to CEOF phases 45°, 105°, and 165°. Shaded areas indicate significant anomalies.
Citation: Journal of Climate 24, 5; 10.1175/2010JCLI3619.1
(a)–(c) Lead–lag standardized (unitless) low-pass-filtered SST anomalies averaged for the 5 RB peaks reaching the +1 standard deviation criterion (see Fig. 3). Anomalies are shown for monthly time lags −09, +00, and +09 before (negative lags) or after (positive lags) the RB peaks. (d)–(f) Pearson correlations between the temporal scores of the CEOF shown in Figs. 1c,e,g and low-pass-filtered anomalies of global spatiotemporal SST. Panels correspond to CEOF phases 45°, 105°, and 165°. Shaded areas indicate significant anomalies.
Citation: Journal of Climate 24, 5; 10.1175/2010JCLI3619.1
Schematic representation of the relationship between lead–lag positive RB peak composites (left panels in Figs. 6 and 7) and the CEOF shown in Figs. 1c–i (right panels in Figs. 6, 7). The distance between the lead–lag composites (noncircular curve) and the CEOF (central circle) at a particular lag/phase represents a schematic measure of similarity between both techniques.
Citation: Journal of Climate 24, 5; 10.1175/2010JCLI3619.1
Schematic representation of the relationship between lead–lag positive RB peak composites (left panels in Figs. 6 and 7) and the CEOF shown in Figs. 1c–i (right panels in Figs. 6, 7). The distance between the lead–lag composites (noncircular curve) and the CEOF (central circle) at a particular lag/phase represents a schematic measure of similarity between both techniques.
Citation: Journal of Climate 24, 5; 10.1175/2010JCLI3619.1
Schematic representation of the relationship between lead–lag positive RB peak composites (left panels in Figs. 6 and 7) and the CEOF shown in Figs. 1c–i (right panels in Figs. 6, 7). The distance between the lead–lag composites (noncircular curve) and the CEOF (central circle) at a particular lag/phase represents a schematic measure of similarity between both techniques.
Citation: Journal of Climate 24, 5; 10.1175/2010JCLI3619.1