Change of El Niño onset location around 1970

Change of El Niño onset location around 1970 | npj Climate and Atmospheric Science

Abstract

A shift of El Niño onset location from eastern Pacific (EP) to western Pacific (WP) occurred around 1970. It was accompanied by a faster mean sea surface temperature (SST) warming in WP and a change of precursory SST and wind anomaly patterns. The eigenvalue analysis of a simple coupled model shows that an SST anomaly (SSTA) in WP may grow under the post-1970 mean condition but cannot under the pre-1970 mean condition. As a result, a warm SSTA appeared in WP accompanying to a preceding La Niña condition in EP after 1970, whereas such a warming was rarely seen before 1970. The preceding SSTA patterns led to distinctive zonal wind responses in EP, favoring El Niño onset in EP prior to 1970. For the post-1970 El Niño onset, an initial warming in WP was induced by anomalous downward solar radiation in association with atmospheric meridional overturning circulation or anomalous horizontal advection associated with thermocline induced eastward geostrophic currents.

Introduction

El Niño-Southern Oscillation (ENSO), the largest year-to-year variability in the Earth’s climate system, exerts great impacts on weather and climate around the globe 1, 2, 3, 4, 5, 6. El Niño exhibits great diversities in spatial structure 7, 8, amplitude 9, 10 and temporal evolution 11, 12. For instance, El Niños were divided into eastern Pacific (EP) and central Pacific (CP) types based on the longitudinal location of maximum sea surface temperature (SST) anomalies during their mature phase 13, 14, 15, 16, 17, 18. A marked asymmetry of temporal evolutions was found between El Niño and its counterpart, La Niña 11.

An outstanding unresolved issue of ENSO dynamics is what causes the diversity of El Niño onset location. The classic El Niño composite was characterized by an initial warming in the coast of South America and subsequent westward propagation 19. This is in great contrast to El Niño initiation during the recent decades when an initial warming started to develop in the western Pacific (WP) 20, 21, 22.

Several theories were proposed to understand El Niño initiation processes 23, 24, 25, 26, 27, 28, 29, 30, 31, 32. Whereas the “delayed oscillator” theory emphasized the effect of delayed ocean waves and associated thermocline changes 24, 25, the key of the “recharge oscillator” theory lied on zonally mean thermocline fluctuations 26, 27. The “seasonal footprint” theory 28, 29, which typically plays a role in the influence of “Pacific Meridional Mode” (PMM) on ENSO 30, 31, 32, 33, 34, along with the “trade wind charging” mechanism 35, 36, 37, on the other hand, involved subtropical and mid-latitude processes. The former emphasized the thermodynamic-coupling process, i.e., wind-evaporation-SST (WES) feedback, while the latter focused on the wind stress curl induced subsurface heat content. Besides, factors like the Australian winter monsoon 38, early-spring Aleutian Low 39, and the position of easterly/westerly wind convergence zone 40 can also play a role in triggering El Niño. In addition, cross-basin interaction may also affect the El Niño onset 41, 42, 43, 44, 45, 46, 47, 48. For example, an SST anomaly (SSTA) in the tropical Atlantic may determine the occurrence of a CP or EP type of El Niño in the subsequent months 43, 44. A positive Indian Ocean Dipole (IOD) or high-frequency westerly wind burst (WWB) might trigger an El Niño event 10, 42, 47, 48, 49, 50, 51.

The objective of the current study is two folds. Firstly, we demonstrate that there was indeed an interdecadal change of El Niño onset location around 1970s from a 41-year running window Empirical Orthogonal Function (EOF) analysis of the observed SST during the past 110 years. Secondly, we intend to reveal physical mechanisms behind the observed change. Unraveling the El Niño onset characteristics and change is critical in advancing our current understanding of ENSO dynamics and improving the seasonal prediction, which would benefit the society for the prevention of natural disasters and for human adaptation and mitigation.

Results

A shift of El Niño onset location around 1970s

In the following, the mean state is referred to a slowly varying interdecadal mean climate state (including the warming trend). The anomalies are obtained by the initial fields minus the climatology (including removing the linear warming trend). An EOF analysis of the equatorial SSTA for two periods (1910–1970 and 1970–2020) shows that the El Niño onset location has an interdecadal shift (Fig. 1a, b). The leading EOF mode during 1910–1970 is characterized by an initial warming in EP; after the initial warming, the maximum SSTA moves westward and reaches a peak in December (Fig. 1a). In contrast, the leading EOF mode during 1970–2020 shows an initial warming in WP, followed by eastward propagation (Fig. 1b). For both the periods, the variance of the leading EOF mode explains about 50% of the total variance, indicating the dominance of the leading mode compared to other higher-order modes.

The leading EOF mode of time-longitude sector of the SSTA averaged over 2°S–2°N during a 1910–1970 and b 1970–2020; c–f, g–j, k–n the composite evolutions of the SSTA (shading, unit: °C), anomalous 850 hPa wind (vector, unit: \(\rm m\,\space s^{-1}\)) and anomalous precipitation (slash line, unit: mm day−1) fields for the MEP, SBW and MCP composite from DJF(0) to SON(0). Linearly detrended data were used here. “0” and “1” represent the El Niño developing year and decaying year. Green lines and arrows denote the propagation of maximum SSTA centers. Black dots in c–n denote the SSTA that exceeds a 95% confidence level, and slash green (brown) lines denote positive (negative) precipitation anomalies that exceed a 95% confidence level. The anomalous precipitation exceeding a 95% confidence level is shown. Black dots in c–n denote the SSTA that exceeds a 95% confidence level. The green (brown) slashes denote the regions of significant positive (negative) precipitation anomalies. Wind vector with speed less than 0.3 m s−1 is omitted.

The interdecadal shift of El Niño onset location was validated by a 41-year moving window EOF analyses in which the time-longitude sections of the equatorial SSTA during the period of 1910–2020, with a central reference time at 1930, 1940, 1950, 1960, 1970, 1980, 1990, and 2000, were analyzed respectively. The result (Supplementary Fig. 1) confirms the interdecadal shift of initial El Niño warming location around 1970. Accompanied by the distinctive onset location shift are opposite zonal propagation directions of the maximum SSTA, as indicated by green lines in Supplementary Fig. 2, points out a robust shift of westward to eastward speed around 1970. A further sensitivity test with a slight change in the separation time from 1970 to 1980 (Supplementary Fig. 3c, d) confirms the interdecadal change characteristic.

The shift characteristic is also supported by ref. 20, who examined the diversity of El Niño onset type and found a transition of the onset type around 1970, from a moderate eastern Pacific (MEP) type prior to 1970 to strong basin-wide (SBW) and moderate central Pacific (MCP) types after 1970 (Supplementary Fig. 4). The classification of the three types of El Niño was based on a nonlinear K-means cluster analysis of equatorial SSTA evolution characteristics from El Niño pre-onset to its mature phases. The MEP originated from EP and propagates westward. The SBW initiated in WP and propagates eastward, forming a basin-wide SSTA at its mature phase. The MCP initiated in WP, and then shifted slightly eastward and matured in CP. The evolving horizontal patterns of anomalous SST and surface wind fields associated with the three types of El Niño onsets are illustrated in Fig. 1c–n and Supplementary Fig. 5. Note that an initial warming appears in the coast of South America in boreal spring [MAM(0)] in MEP. In contrast, warming has already developed in WP prior to DJF(0) in SBW and MCP.

Because MEP is the dominant onset type prior to 1970 whereas SBW and MCP are dominant types after 1970, in the following we will focus on their distinctive SSTA patterns during El Niño preceding winter. While a La Niña-like precursory SSTA pattern occurs in DJF(0) for all the three types, a marked difference appears in WP, that is, an accompanying warm SSTA occurs in SBW/MCP while no apparent SSTA signals are found in WP in MEP (Fig. 1c, g, k). This implies that the post-1970 mean state favors the development of an SSTA in WP whereas the pre-1970 mean state does not. In the subsequent section we will investigate the possible role of the mean state in affecting the SSTA development. Accompanied by the single-pole SSTA pattern prior to 1970 and the dipole SSTA patterns after 1970 in DJF(0) (Fig.

It is interesting to note that a PMM-like pattern appears during the preceding winter DJF(0) and spring MAM(0) in SBW and MCP (Fig. 1g, h, k, l). Given that the SBW and MCP types are derived based on El Niño onset evolution 20, it is not surprising to see the accompanied PMM pattern in DJF(0) and MAM(0). Such a connection of PMM to El Niño, however, is unclear in MEP prior to 1970 (Fig. 1c, d). This points out a possible link between El Niño initiation location and PMM formation.

Impact of the mean state on SSTA growth in WP: an eigenvalue analysis

Previous studies suggested that the interdecadal change of background mean state influenced ENSO behavior 52, 53, 54, 55, 56. Here we hypothesize that the post-1970 mean state has a great impact on El Niño onset in WP. To test this hypothesis, a simple ocean-atmosphere coupled model was constructed (see Methods for detailed explanations and derivations). The key parameters in the model are listed in Table 1.

The mean state differences of SST and 850 hPa wind and specific humidity fields between the two periods (1910–1970 and 1970–2020) are illustrated in Fig. 2a, b. Note that a more rapid mean SST warming appears in WP than in the central and eastern Pacific (CEP). As a result, both the trade wind and the zonal SST gradient strengthen along the equatorial Pacific (Fig. 2a). The larger WP warming also leads to a greater increase of the mean specific humidity in situ (Fig. 2b). Physically, we argue that an SSTA perturbation can grow faster under a greater climatological mean moisture condition because the same SSTA perturbation can stimulate stronger diabatic heating and thus circulation responses.

a Changes (post-1970 minus pre-1970) of the mean SST (shading, units: °C) and 850 hPa wind (vector, unit: m s−1) fields; b Changes of specific humidity vertical profiles averaged over 130°–170°E along latitudinal bands at 0°, 5°S, 10°S, 5°N and 10°N respectively (units: \({\rm kg}\,\space{{\rm kg}}^{-1}\)); c Growth rates (unit: ×10−7 s−1) derived from CTRL, SEN1 and SEN2 experiments under the pre- and post- 1970 mean-state conditions; d Contributions of individual terms in Eq. ( 10) to the growth rate (units: ×10−7 s−1) in CTRL. Black dots in a represents the interdecadal mean SST change that exceeds a 95% confidence level.

By solving the eigenvalue problem of the equations, the growth rate (σi), which represents how fast an SSTA disturbance grows (its value corresponds to a reversed e-folding time scale), can be derived [Eqs. ( 10)–( 14) in Methods]. It depends on four terms in association with the following four air-sea interaction processes: zonal advective feedback, upwelling feedback, wind-evaporation-SST (WES) feedback and cloud-radiation-SST feedback. To test the sensitivity of the model solution to zonal wavelength, we calculated the growth rates with the following three wavelengths, 20° longitudes in a control experiment (CTRL) and 40° and 10° longitudes in sensitivity experiments (SEN1 and SEN2) respectively. Figure 2c shows the calculated growth rates for the two periods. It turns out that the model solution is not sensitive to the different wavelengths. The growth rate is weak and negative under the pre-1970 mean condition but becomes strong and positive under the post-1970 mean condition. Thus, the eigenvalue analysis confirms the hypothesis that the post-1970 mean state favors the development of an SSTA perturbation in WP whereas the pre-1970 mean state does not.

The relative contributions of the individual feedback terms at the right-hand side of Eq. ( 10) are further examined. Figure 2d shows that the first three terms contribute positively, as expected, while the last term is a damping term and contributes negatively. The largest contributor to the growth rate increase arises from the WES feedback, followed by the upwelling and zonal advective feedbacks. Among all the mean state variables, the change of the mean moisture appears most critical in affecting the growth rate, because it not only has a direct impact on the WES feedback through the modulation of sea-air specific humidity difference [see Eqs. ( 4) and ( 13) in Methods] but also exerts an indirect impact on anomalous upwelling and zonal current through enhanced atmospheric diabatic heating and wind responses.

A sensitivity test with a different meridional profile \(\it \varPhi \left(y\right)={{\varPhi }_{2}{H}_{2}\left(y\right)e}^{-\frac{\beta }{2{C}_{0}}{y}^{2}}={\left(4\frac{\beta }{{C}_{0}}{y}^{2}-2\right)\varPhi }_{2}{e}^{-\frac{\beta }{2{C}_{0}}{y}^{2}}\) was further carried out. Here \({H}_{2}\left(y\right)=(4\frac{\beta }{{C}_{0}}{y}^{2}-2)\) denotes the second order Hermite function. The growth rate is derived as Supplementary Eqs. ( 28)–( 31). As seen from Supplementary Fig. 6, the result is quite similar compared to Fig. 2c, d, which points out that the eigenvalue analysis result is robust.

To further validate the conclusion that the change of the interdecadal background mean moisture is most critical in affecting the growth rate, additional two sets of sensitivity experiments were conducted. In the first experiment (SEN_q_post1970), all parameters keep the same original values except that the mean moisture (\(\bar{q}\)) is set to 0.018 kg kg−1 for both the periods. In the second experiment (SEN_q_pre1970), all parameter values are kept same except that the mean moisture (\(\bar{q}\)) is set to 0.016 kg kg−1 for both the periods. The results show that in SEN_q_post1970, the pre-1970 growth rate changes from a nearly zero negative value to a large positive value, very close to the post-1970 growth rate (Supplementary Fig. 7). In SEN_q_pre1970, the post-1970 growth rate decreased by ~90% (Supplementary Fig. 8). The two additional sensitivity experiments indicate that the interdecadal change of the mean-state moisture plays a crucial role in affecting the growth rate of the SSTA in WCP.

One may wonder what causes the change of mean state SST. By conducting an EOF analysis of the 8-yr low-pass filtered observed monthly SST field during 1910–2020, it turns out that the first mode is related to long-term trend, and its structure resembles the one shown in Fig. 2a, with a greater warming in the WP and less warming in the CEP. The second and third modes are related to tropical Pacific decadal variability (TPDV) and Atlantic Multidecadal Oscillation (AMO). After reconstructing the SST using the spatial modes with their corresponding principal components, the change of SST due to climate change, TPDV and AMO can be obtained. Besides, the randomness is obtained by subtracting the sum of SST changes due to the above three modes from the total SST change. The result indicated that the change of mean state SST is mainly attributed to the climate change (figure not shown).

Impact of the single-pole and dipole SSTA patterns on wind responses

The eigenvalue analysis above helps us understand why an anomalous WP warming is triggered and maintained accompanying to a La Niña-like SSTA during the preceding year of El Niño development in SBW/MCP but not in MEP (Supplementary Fig. 5). As a result, a single-pole SSTA pattern formed in tropical Pacific in DJF(0) before 1970, and it led the development of a traditional EP El Niño as shown in Rasmusson and Carpenter (1982) 19, whereas a dipole SSTA pattern occurred after 1970, and it led to the development of either a CP or a super EP El Niño event. In this section we further investigate how the single-pole and dipole SSTA patterns affected zonal wind anomalies at the equator (Fig. 1c, g, k).

The zonal distributions of anomalous SST and vertical overturning circulation along the equator are plotted in Fig. 3. Note that a single descent branch with negative precipitation anomalies occurs in the equatorial central Pacific in MEP. In response to the negative heating forcing, low-level easterly anomalies appear in CP, while weak westerly anomalies appear in EP as a consequence of boundary-layer divergence (Fig. 3a). In contrast, in response to the dipole SSTA pattern in SBW/MCP, anomalous vertical motion and precipitation fields also exhibit a dipole pattern (Fig. 1g, k; Fig. 3b, c), with anomalous ascent near 150°E and descent to its east. The pronounced low-level easterly anomalies over the entire central and eastern Pacific are a direct Kelvin wave response to the positive heating near 150°E.

a–c Anomalous zonal wind (black vector, unit: m s−1) and vertical velocity (shading, unit: ×10−2 Pa s−1) fields averaged over 5°S-5°N in DJF(0) during MEP prior to 1970 and SBW and MCP after 1970. The units on the y-axis is hPa; d–f SSTA zonal profiles averaged over 5°S–5°N in DJF(0) during MEP, SBW and MCP types (units: °C). Green dots and arrows denote the zonal wind and vertical velocity anomalies exceeding a 95% confidence level. Right circles in a–c highlight the zonal wind anomaly in EP. “0” denotes the El Niño developing year.

To validate whether the distinctive wind patterns result from the underlying SSTA forcing, we conducted atmospheric general circulation model (GCM) experiments using ECHAM4.6 (see Methods). In the control experiment, the climatological monthly mean SST is specified, whereas in the sensitivity experiments, the SSTA patterns associated with the MEP, SBW and MCP in DJF(0) are superposed onto the climatological SST field. Supplementary Table 1 lists the details of the GCM experiments. The result (see Supplementary Fig. 9) indicates that the SSTA patterns do affect the anomalous wind responses. For instance, the single-pole SSTA pattern in MEP does stimulate easterly anomalies in CP and weak westerly anomalies in EP (Supplementary Fig. 9a), resembling the observed (Fig. 3a). The dipole SSTA patterns associated with SBW and MCP do stimulate anomalous easterlies over the entire CEP (Supplementary Fig. 9b, c).

Therefore, the consequence of a single-pole SSTA pattern is the westward shift of easterly anomalies at the equator, which allows initial warming in far EP. The dipole-pole SSTA pattern, on the other hand, leads to easterly anomalies across the eastern and central equatorial Pacific, prohibiting the initial warming in EP.

Cause of initial EP warming prior to 1970

To understand El Niño onset processes prior to 1970, we conducted a mixed-layer heat budget analysis 10, 11(see Methods section for detail) in EP during MEP El Niño onset phase. Figure 4 shows the diagnosis result during MAM(0) averaged over (5°S–5°N, 100°–80°W). Note that the MLT anomaly (MLTA) tendency is primarily attributed to the following ocean advection terms: upwelling feedback \(\left(-{w}^{{\prime} }\frac{\partial \bar{T}}{\partial z}\right)\), nonlinear meridional advection \(\left(-{v}^{{\prime} }\frac{\partial {T}^{{\prime} }}{\partial y}\right)\) and zonal advective feedback \(\left(-{u}^{{\prime} }\frac{\partial \bar{T}}{\partial x}\right)\). The warming arises from positive thermocline depth anomalies induced by anomalous anticyclonic wind stress forcing in association with the La Niña-like SSTA pattern in DJF(0) (Fig. 1c). As seen from Fig. 4b, positive thermocline depth anomalies are accumulated in WP in DJF(0). In subsequent seasons they propagate eastward as a downwelling Kelvin wave (Fig. 4c, d). The positive thermocline depth anomalies, on the one hand, induce eastward geostrophic current anomalies and a positive zonal advection anomaly \(\left(-{u}^{{\prime} }\frac{\partial \bar{T}}{\partial x}\right)\), initiating warming in EP, and, on the other hand, induce anomalous downwelling in EP, leading to a positive vertical advection anomaly \(\left(-{w}^{{\prime} }\frac{\partial \bar{T}}{\partial z}\right)\).

a Mixed-layer heat budget diagnosis results (unit: °C mon−1) in MAM(0) averaged over (5°S–5°N, 100°–80°W) during MEP; b–d Composite pattern evolution of anomalous thermocline depth (shading, unit: m) and ocean current (vector, unit: m s−1) fields during MEP in DJF(0), MAM(0) and JJA(0). The thermocline depth is defined as the depth of 20°C isotherm. Linearly detrended data was used. “0” represents the El Niño developing year. The green dots in b–d represent the thermocline depth anomaly that exceeds a 95% confidence level. The ocean current velocity less than 0.01 m s−1 is omitted. The x-axis in a from left to right represents: MLTA tendency \(\left(\frac{\partial {T}^{{\prime} }}{\partial t}\right),\) the sum of advection and heat flux terms, the sum of advection terms, the sum of heat flux terms, and individual advection and heat flux terms:\(-\bar{u}\frac{\partial {T}^{{\prime} }}{\partial x}, -{u}^{{\prime} }\frac{\partial \bar{T}}{\partial x}\),\(-{u}^{{\prime} }\frac{\partial {T}^{{\prime} }}{\partial x}\), \(-\bar{v}\frac{\partial {T}^{{\prime} }}{\partial y}\), \(-{v}^{{\prime} }\frac{\partial \bar{T}}{\partial y}\), \(-{v}^{{\prime} }\frac{\partial {T}^{{\prime} }}{\partial y}\), \(-\bar{w}\frac{\partial {T}^{{\prime} }}{\partial z}\), \(-{w}^{{\prime} }\frac{\partial \bar{T}}{\partial z}\), \(-{w}^{{\prime} }\frac{\partial {T}^{{\prime} }}{\partial z}\), net longwave radiation, net solar radiation, surface latent heat flux and surface sensible heat flux, respectively.

A positive latent heat flux (LHF) anomaly also contributes to the initial warming in EP. This is attributed to the local westerly anomalies as these anomalous winds are against the climatological mean easterly. The decrease in the surface wind speed leads to a decrease in surface evaporation. Such a LHF effect is opposite in the post-1970 groups, as demonstrated by an MLT budget analysis in SBW/MCP. As shown in Supplementary Fig. 10, the cooling effect of the LHF anomaly is clearly presented in SBW/MCP. The cooling is caused by increased wind speed as the anomalous easterlies strengthen the mean easterly in EP. The difference between SBW and MCP lies in the ocean dynamic heating term. In MCP the ocean dynamic impact is negative, whereas in SBW it is positive, owing to much stronger positive thermocline depth anomalies at the equator associated with this type (Supplementary Fig. 10c, e).

The spring onset prior to 1970 is closely related to the discharge of anticyclonic wind stress-induced positive thermocline depth anomalies in the preceding season from the western Pacific. Such a positive thermocline depth anomaly, once moving to the equatorial EP at a form of equatorial Kelvin waves, triggers a warming in the America coast where the mean thermocline is shallow. Such a process is suppressed after 1970 to a certain extent due to the presence of the anomalous easterly in EP. Meanwhile, an anomalous warming is easily triggered in WP under the post-1970 mean state, as demonstrated by the instability analysis of a theoretical model. Such a warming occurs in preceding fall and winter in accompanying to a La Niña-like SSTA pattern in EP (Supplementary Fig. 5b, c).

Cause of El Niño onset in WP after 1970

Because the dominant El Niño onset types after 1970 are SBW and MCP 20, we will focus on initial triggering processes associate with the two modes. As seen from Supplementary Fig. 5b, c, the initial warming for both the modes starts in WP around JJA prior to El Niño developing year [JJA(–1)]. To reveal physical processes responsible for the initiation of the WP warming, we conducted a mixed-layer heat budget analysis for both the modes in JJA(−1).

The diagnosis result shows that the main contributor of the warming in SBW is anomalous downward shortwave radiative flux, whereas it arises from anomalous ocean advection in MCP (Fig. 5a, b). In SBW, the increased downward shortwave radiation is associated with anomalous descent in the equatorial region, which is the part of local Hadley Cell, as seen in Fig. 5c. The anomalous local vertical overturning circulation is driven by positive SST and precipitation anomalies in subtropics near 20°N and 20°S. The so-induced descent at the equator causes less cloud cover and more downward shortwave radiation in situ, leading to a warm SSTA.

Mixed-layer heat budget diagnosis results (unit: °C mon−1) in JJA(–1) averaged over a (5°S–5°N, 135°–155°E) during SBW and b (5°S–5°N, 160°E–180°) during MCP; c Anomalous meridional (unit: m s−1) and vertical (shading, unit: ×\({10}^{-2}\space\,\rm{Pa}\) \(\rm {s}^{-1}\)) wind fields (black vector), SSTA (red curve, unit: °C) and anomalous precipitation (green curve, unit: mm day−1) fields averaged over 135°–155°E in JJA(−1) during SBW; d Anomalous thermocline depth (shading, units: m) and ocean current (vector, units: m s−1) fields in JJA(–1) during MCP. “−1” represents the preceding year of El Niño development. Green dots in c represent wind anomalies that exceed a 95% confidence level. Green dots in d represent thermocline depth anomalies that exceed a 95% confidence level. Ocean current velocity less than 0.01 m s−1 is omitted.

In contrast, the warming in MCP arises from the ocean dynamic heating term while the net heat flux term plays a negative role. Among the dynamic heating terms, anomalous zonal and meridional advective processes are most important. It is found that eastward zonal current anomalies (caused by thermocline-induced geostrophic current component) are critical for the initial warming. As shown in Fig. 5d, negative (positive) thermocline depth anomalies appear off (at) the equator. This meridional structure favors eastward geostrophic current anomalies 57. The eastward current anomalies advect higher mean SST eastward, leading to a warming tendency. Once the warming occurs, climatological mean meridional currents further advect the warm signal poleward, strengthening the positive MLTA tendency in situ.

Discussion

In this work, we demonstrate that there is an interdecadal shift of El Niño onset location around 1970, that is, an EP onset before 1970 and a WP onset after 1970, and reveal physical mechanisms responsible for the shift. Firstly, through the eigenvalue analysis of a simple coupled atmosphere-ocean model, we demonstrate that the post-1970 mean state favors the growth of an SSTA perturbation in WP whereas the pre-1970 mean state does not. The specific air-sea interaction processes responsible for the faster growth include the wind-evaporation-SST feedback, the upwelling feedback, and the zonal advective feedback. The mean-state moisture change is critical in strengthening the aforementioned feedbacks. The result suggests that an SSTA perturbation can be easily triggered and maintained in WP under the post-1970 mean condition.

Because of the mean state impact, distinctive preceding anomalous SST patterns occurred between the pre- and post- 1970 El Niño composites. A single-pole La Niña-like SSTA pattern appeared in DJF(0) prior to 1970, whereas a dipole SSTA pattern with a positive (negative) SSTA in WP (EP) occurred after 1970. The different SSTA patterns led to distinctive anomalous zonal wind responses in EP, as confirmed by atmospheric GCM experiments. For the pre-1970 El Niño composite (MEP), weak westerly anomalies appeared in EP in DJF(0), which worked together with the ocean dynamic process and triggered El Niño onset in EP. In contrast, pronounced easterly anomalies appeared in EP in the post-1970 El Niño composites (SBW and MCP), which induced a cooling tendency through enhanced surface evaporation.

While the post-1970 mean state favors the growth and maintenance of the WP SSTA, two specific processes triggered El Niño onset in WP after 1970. One is through downward solar radiation forcing in association with anomalous atmospheric meridional overturning circulation. The source to generate the anomalous overturning circulation is from SSTA forcing in the subtropics. Another is via anomalous zonal advection associated with thermocline-induced eastward geostrophic currents.

While the current study identifies both pre- and post-1970 El Niño onset processes, the following questions remain unresolved. For post-1970 El Niño events, what is the origin of the subtropical SSTA that drives the anomalous vertical overturning circulation? What determines the meridional distribution of the thermocline depth anomaly so that eastward geostrophic current anomalies are generated? Also, what contributes to a more rapid mean SST warming in WP during the past century? To what extent does the cross-basin interaction contribute to El Niño onset location shift? All these questions above require further in-depth observational, theoretical and modeling studies.

The current study reveals, from both dynamical and thermodynamical air-sea coupling perspectives, the role of the background mean state change in causing the El Niño onset location shift. Through a simplified theoretical model, we demonstrate the important role of the background mean state change, in particular the mean moisture change, in affecting the SSTA growth in WP. Through the diagnosis of the mixed-layer heat budget, we identify key processes that trigger El Niño initial warming prior to and after 1970. Through idealized atmospheric GCM experiments, we confirm the role of the precursory dipole and single-pole SSTA patterns in causing distinctive wind responses. The current study also sheds lights in understanding the future projection of El Niño. It is anticipated that more frequent El Niño onset will happen in WP as long as the greater WP warming trend persists.

Methods

Observational datasets

Ensemble mean monthly SST datasets from the Extended Reconstructed SST analyses version 5 (ERSSTv5) 58 with a 2° × 2° horizontal resolution and the Hadley Centre Sea Ice and Sea Surface Temperature reanalysis version 1.1 (HadISSTv1.1) 59 at a 1° × 1° horizontal resolution for the period from 1910 to 2020 are used in the current study. Merged monthly atmospheric specific humidity and wind fields are derived from the following reanalysis products: NOAA Twentieth Century Reanalysis version 2c (20CRv2c) 60 with a horizontal resolution of 2° × 2° from 1910 to 1978, the National Centers for Environmental Prediction/Department of Energy (NCEP/DOE) Reanalysis II (NCEP2) 61 at a 2.5° × 2.5° horizontal resolution from 1979 to 2020, ECMWF reanalysis of the 20th Century (ERA20C) 62 from 1910 to 1957, ECMWF Reanalysis ERA40 63 from 1958 to 2001, ECMWF reanalysis Interim (ERA-Interim) 64 from 2002 to 2018, and ECMWF reanalysis 5 (ERA5) 65 from 2002 to 2020. To ensure temporal consistency, the mean state of ERA-Interim and ERA5 datasets is calibrated to ERA40 based on their overlapping period 1979–2001. To reduce the uncertainty of the datasets, the ensemble mean of the NOAA and ECMWF datasets are applied. The climatological mean states mentioned here are referred to as the mean states during the two interdecadal periods, 1910–1969 and 1970–2020.

The precipitation datasets are the ensemble average of the following datasets, 20CRv2c (1910–2014), ERA20C (1910–2010), ERA-Interim (1979–2018), NCEP2 (1979–2020), NCEP/National Center for Atmospheric Research Reanalysis I (NCEP1) 66(1948–2020), and the Global Precipitation Climatology Project Version 2.3 Combined Precipitation Dataset (GPCPv2.3) 67(1979–2020). The purpose to use the ensemble average is to reduce the uncertainty of the products.

Monthly mixed-layer ocean temperature and three-dimensional ocean currents are derived from the ensemble mean of the following oceanic reanalysis datasets, the Simple Ocean Data Assimilation version 2.2.4 (SODAv2.2.4) 68 at a 0.5° × 0.5° horizontal resolution and 40 vertical levels and the NCEP Global Ocean Data Assimilation System (GODAS) 69 at a 1° zonal resolution, \(\frac{1}{3}\)° meridional resolution, and 40 vertical levels (with a 10-m resolution in the upper 105 m). The continuity equation was applied to calculate vertical ocean currents. SODAv2.2.4 covers 1910–2008 while GODAS covers 2009–2020. Oceanic thermocline depth (20°C isotherm depth) field is obtained from SODAv2.2.4 and GODAS. The climatological mean of GODAS is calibrated to SODAv2.2.4 based on their overlapping period 1980–2008.

The surface heat flux fields are also taken from the ensemble mean of the following six datasets: 20CRv2c (1910–2014), NCEP2 (1979–2020), ERA20C (1910–2010), ERA40 (1958–2001), ERA-Interim (1979–2018) and Woods Hole Oceanographic Institute Objectively Analyzed Air-Sea Fluxes (OAFlux) 70(1984–2009), to reduce the uncertainty of the datasets. All the above datasets are interpolated into 1° × 1° horizontal resolution and the merged datasets cover the period of 1910–2020.

Definition of El Niño

El Niño years are defined when the ONI index [(SSTA averaged over the Niño3.4 region (5°S–5°N, 170°–120°W)] averaged for October–November–December–January–February is greater than 0.5 °C. After removing the linear trend of the SST, 25 El Niños are identified. Based on their onset characteristics 20, these El Niño events are separated into SBW El Niño onset type (1972, 1982, 1997, 2015), MEP El Niño onset type (1911, 1913, 1918, 1923, 1925, 1939, 1951, 1957, 1963, 1965, 1968, 1976) and MCP El Niño onset type (1986, 1991, 1994, 2002, 2004, 2006, 2009, 2014, 2018).

Ocean mixed-layer heat budget analysis

The ocean mixed-layer heat budget equation is written as follows:

where T denotes MLT; u, v, and w represent 3D mixed-layer current velocity. All mixed-layer variables are calculated based on an averaged over the upper 50 m. A prime denotes the anomaly field and a bar represents the climatological mean annual cycle field. \({Q}_{n{\rm{e}}t}^{{\prime} }\) denotes the net surface heat flux anomaly that includes anomalous net longwave radiation, net shortwave radiation, latent heat flux and sensible heat flux; ρ = 1015 kg m−3 denotes the density of water; \({Cp}=4000\) \(\rm J\,\space{{kg}}^{-1}\,\space{K}^{-1}\) is the specific heat of water. In this study a constant mixed-layer depth (H = 50 m) was used, based on the zonal average of the climatological mean mixed-layer depth along the equatorial Pacific. The choice of this value is also based on an intermediate coupled model of Zebiak and Cane 71, in which a constant mixed layer depth of 50 m was used. We did a sensitivity test of the mixed-layer heat budget with a 20% increase or decrease of the mixed-layer depth, and the result is essentially the same. R is a residual term representing processes other than heat fluxes and temperature advection 72, which may be related to sub-grid ocean mixing process, diffusion, high-frequency advective rectification processes 73, 74, 75. As seen from Fig. 4a and Fig. 5a, b, the mixed-layer heat budget analysis is not perfectly closed, and there are some errors. But the warming tendencies during the El Niño onset phases are well captured, which can be seen from the summation of all dynamic and thermodynamic terms. Therefore, the budget analyses identify key physical processes relevant to the initial warming during the El Niño onset phases. The actual value of the residual term R may be inferred from the difference of the first two terms in the figures (i.e., the tendency term and the sum term).

Simple ocean-atmosphere coupled model

Similar to Cane-Zebiak coupled model 71, the atmospheric component is the Gill Model 76 on an equatorial beta plane with a heating anomaly being related to the underlying SSTA [Eqs. ( 1)–( 3)]. The Gill model is the simplest atmospheric model in the tropics that describes the first baroclinic mode atmospheric response to prescribed heating. It is in particular suitable for studying the atmospheric response in the western Pacific to an underlying SSTA forcing 76. Equation ( 4) describes the oceanic mixed-layer temperature (MLT) equation, with the MLT tendency depending on anomalous zonal and vertical advection, anomalous wind-induced surface latent heat flux and a thermal damping associated with the cloud-shortwave radiation-SST feedback 77. In the presence of a deep mean thermocline in WP, anomalous thermocline feedback is weak and can be neglected. Following Zebiak and Cane 71 and Chang and Philander 78, the ocean mixed-layer currents are determined by anomalous surface wind stress forcing [Eqs. ( 5)–( 7)]. The governing equations can be written as:

where ε and εo represent the damping coefficients at free atmosphere and ocean respectively; u′, \({v}^{{\prime} }\), \({\phi }^{{\prime} }\), \({T}^{{\prime} }\), \({u}_{o}^{{\prime} }\), \({v}_{o}^{{\prime} }\), \({w}_{o}^{{\prime} }\) are the atmospheric anomalous zonal wind, meridional wind, geopotential height, mixed-layer temperature anomaly (MLTA), zonal ocean current, meridional ocean current and vertical ocean current; \(\bar{u}\), \(\bar{q}\) and Va represents mean state surface zonal wind, specific humidity and wind speed. For simplicity, let \({V}_{a}=|\bar{u}|\). \({\tau }_{x}^{{\prime} }={\rho }_{a}{c}_{D}{V}_{a}{u}^{{\prime} }\) is the anomalous zonal wind stress at the surface; \(\bar{T}\) is the mean state MLT. \(\beta =2.3\times{10}^{-11}\,\space\rm{rad}\) \(\rm {m}^{-1}\) \(\rm {s}^{-1}\) represents the planetary vorticity; \({C}_{0}=50\,\space \rm m\) \(\rm {s}^{-1}\) denotes gravity wave speed for the first baroclinic mode; I represents the heating coefficient due to free-troposphere wave convergence, and it depends on the mean specific humidity at the surface; γ denotes the SSTA induced atmospheric condensational heating and thus is the function of the mean surface specific humidity; \({\rho }_{a}=1.2\,\space\rm{kg}\) \(\rm {m}^{-3}\) represents the density of atmosphere; \({\rho }_{o}={10}^{3}\,\space \rm{kg}\) \(\rm {m}^{-3}\) represents the density of ocean; \({{c}_{E}=1.5\times{10}^{-3}}\) denotes the moisture exchange coefficient; \({c}_{w}=4.27\times{10}^{3}\) denotes the ocean specific heat; \({L}_{v}=2.5\times{10}^{6}\,\space \rm J\) \(\rm {{kg}}^{-1}\) represents the latent heat of evaporation; H1 = 50 m denotes the depth of mean oceanic mixed layer; RH = 0.85 represents the mean relative humidity at the surface; α denotes the cloud-radiation-SST feedback coefficient following Li et al. 77; \({c}_{D}=1.5\times{10}^{-3}\) denotes the drag coefficient. In all the equations above, a bar represents the mean state variable whereas a prime represents the perturbation.

Assume (\({u}^{{\prime} }\), \({v}^{{\prime} }\) \({\phi }^{{\prime} }\), \({T}^{{\prime} }\), \({u}_{o}^{{\prime} }\), \({v}_{o}^{{\prime} }\), \({w}_{o}^{{\prime} }\)) \(=\) [\(U\left(y\right)\), \(V\left(y\right)\), \(\varPhi \left(y\right)\), \(T\left(y\right)\), \({U}_{o}\left(y\right)\), \({V}_{o}\left(y\right)\), \({W}_{o}\left(y\right)\)] \({e}^{i\left({kx}-\sigma t\right)}\), where k denotes the zonal wavenumber; σ = σr + iσi is a complex parameter, where σr denotes the frequency and σi denotes the growth rate. Eliminating \(U,V,T,{U}_{o},{V}_{o},{W}_{o}\), one may derive a single equation that involves only Φ (for detailed derivation, readers are referred to the Supplementary Notes).

where A1, A2, A3, A1, B2, B3 are the functions of y and k (see Supplementary Notes). Equation ( 8) can be written as

Assuming the perturbation has a meridional profile \(\it {\Phi }(y)={{\Phi }}_{0}{H}_{0}(y){e}^{-\frac{\beta }{2{C}_{0}}{y}^{2}}={{\Phi }}_{0}{e}^{-\frac{\beta }{2{C}_{0}}{y}^{2}}\), where \({H}_{0}\left(y\right)=1\) is the zero order Hermite function, one may derive the following dispersion equation for the growth rate (σi) at the equator:

where C1, C2, C3 and C4 represent the effects of the zonal advective feedback, the upwelling feedback, the WES feedback and the cloud-radiation-SST feedback on the growth rate, respectively.

The eigenvalue analysis performed in this study involves a series of simplification of the governing equations and the calculation of the growth rate based on the observational values prior to and after 1970. For specific derivations of these equations, the reader is referred to the Supplementary Notes.

ECHAM4.6 Model description and sensitivity experiments

An atmospheric general circulation model, the ECHAM version 4.6 (ECHAM4.6), developed by Max-Plank Institute for Meteorology 79, is used. The model has a 2.8° × 2.8° (T42) horizontal resolution with 19 vertical levels in a hybrid sigma pressure coordinate system (from surface to 10 hPa). This model was used to study various tropical climate dynamics problems from ENSO to MJO 80, 81, 82. Here we employed this model to investigate the atmospheric wind response to a specified SSTA forcing.

In the control experiment (EXP_CTRL) the model is driven by climatological monthly SST. In the first sensitivity experiment (EXP_MEP), the model is forced by an observed SSTA pattern for MEP over the tropical Pacific (10°S–10°N, 150°E–80°W) in DJF(0) that is superposed onto the climatological monthly SST field. The second (EXP_SBW) and third (EXP_MCP) experiments resemble the first one except that a dipole SSTA pattern over tropical Pacific (10°S–10°N, 140°E–80°W) in DJF(0) for SBW and MCP is specified respectively. The difference between the sensitivity experiments and control experiments is considered as the anomalous atmospheric response to the SSTA forcing. All the experiments are integrating for 30 years and the ensemble mean fields during the last 20 years are shown.

Data availability

All data needed to evaluate the conclusions in the paper are presented in the paper. The ERSSTv5 dataset is available at https://psl.noaa.gov/data/gridded/data.noaa.ersst.v5.html. The HadISSTv1.1 dataset is available at https://www.metoffice.gov.uk/hadobs/hadisst/. The NOAA 20CRv2c dataset is available at https://psl.noaa.gov/data/gridded/data.20thC_ReanV2c.html. The NCEP2 dataset is available at https://psl.noaa.gov/data/gridded/data.ncep.reanalysis2.html. The NCEP1 dataset is available at https://psl.noaa.gov/data/gridded/data.ncep.reanalysis.html. The GPCP dataset is available at https://psl.noaa.gov/data/gridded/data.gpcp.html. The SODAv2.2.4 dataset is available at http://apdrc.soest.hawaii.edu/erddap/search/index.html?page=1&itemsPerP…. The GODAS dataset is available at https://psl.noaa.gov/data/gridded/data.godas.html. The OAFlux dataset is available at http://apdrc.soest.hawaii.edu/erddap/search/index.html?page=1&itemsPerP…. The ERA5 dataset is available at https://cds.climate.copernicus.eu/cdsapp#!/dataset/reanalysis-era5-sing… and https://cds.climate.copernicus.eu/cdsapp#!/dataset/reanalysis-era5-pres…. Codes used in this study are available from the corresponding author on request.

1.
Ropelewski, C. F. & Halpert, M. S. Global and regional scale precipitation patterns associated with the El Niño/Southern Oscillation. Mon. Weather Rev. 115, 1606–1626 (1987).
2.
Huang, R. & Wu, Y. The influence of ENSO on the summer climate change in China and its mechanism. Adv. Atmos. Sci. 6, 21–32 (1989).
3.
Bove, M. C., Eisner, J. B., Lansea, C. W., Niu, X. F. & O’Brien, J. J. Effect of El Niño on U.S. landfalling hurricanes, Revisited. Bull. Am. Meteorol. Soc. 79, 2477–2482 (1998).
4.
Changnon, S. A. Impacts of 1997-98 El Niño generated weather in the United States. Bull. Am. Meteorol. Soc. 80, 1819–1827 (1999).
5.
Mcphaden, M. J., Zebiak, S. E. & Glantz, M. H. ENSO as an integrating concept in Earth Science. Science 314, 1740–1745 (2006).
6.
Li, T. & Hsu, P. C. ENSO Dynamics. Fundamentals of Tropical Climate Dynamics, 236 (Springer International Publishing, 2017).
7.
Capotondi, A. et al. Understanding ENSO diversity. Bull. Am. Meteorol. Soc. 96, 921–938 (2015).
8.
Timmermann, A. et al. El Niño–Southern Oscillation complexity. Nature 559, 535–545 (2018).
9.
Chen, L., Li, T., Behera, S. K. & Doi, T. Distinctive precursory air-sea signals between regular and super El Niños. Adv. Atmos. Sci. 33, 996–1004 (2016).
10.
Chen, L., Li, T., Wang, B. & Wang, L. Formation mechanism for 2015/16 super El Niño. Sci. Rep. 7, 2975 (2017).
11.
Chen, M., Li, T., Shen, X. & Wu, B. Relative roles of dynamic and thermodynamic processes in causing evolution asymmetry between El Niño and La Niña. J. Clim. 29, 2201–2220 (2016).
12.
An, S. I., Tziperman, E., Okumura, Y. M. & Li, T. ENSO irregularity and asymmetry. El Niño Southern Oscillation in a Changing Climate, Geophys. Monogr. Vol. 253, 153–172 (John Wiley & Sons, Inc., 2020).
13.
Larkin, N. K. & Harrison, D. E. ENSO warm (El Niño) and cold (La Niña) event life cycles: Ocean surface anomaly patterns, their symmetries, asymmetries, and implications. J. Clim. 15, 1118–1140 (2005).
14.
Ashok, K. et al. El Niño Modoki and its possible teleconnection. J. Geophys. Res. 112, C11007 (2007).
15.
Kao, H. Y. & Yu, J. Y. Contrasting eastern-Pacific and central-Pacific types of ENSO. J. Clim. 22, 615–632 (2009).
16.
Kug, J. S., Jin, F. F. & An, S. I. Two types of El Niño events: cold tongue El Niño and warm pool El Niño. J. Clim. 22, 1499–1515 (2009).
17.
Yeh, S. W. et al. El Niño in a changing climate. Nature 461, 511–515 (2009).
18.
Yu, J. Y. & Kim, S. T. Reversed spatial asymmetries between El Niño and La Niña and their linkage to decadal ENSO modulation in CMIP3 models. J. Clim. 24, 5423–5434 (2011).
19.
Rasmusson, E. M. & Carpenter, T. H. Variations in tropical sea surface temperature and surface wind fields associated with the Southern Oscillation/El Niño. Mon. Weather Rev. 110, 354–384 (1982).
20.
Wang, B. et al. Historical change of El Niño properties sheds light on future changes of extreme El Niño. Proc. Natl Acad. Sci. 116, 22512–22517 (2019).
21.
Philander, S. G. H., Yamagata, T. & Pacanowski, R. C. Unstable air-sea interactions in the tropics. J. Atmos. Sci. 41, 604–613 (1984).
22.
Philander, S. G. El Niño, La Niña, and the Southern Oscillation. 289 (Academic Press, 1990).
23.
Bjerknes, J. Atmospheric teleconnections from the equatorial Pacific. Mon. Weather Rev. 97, 163–172 (1969).
24.
Suarez, M. J. & Schopf, P. S. A delayed action oscillator for ENSO. J. Atmos. Sci. 45, 3283–3287 (1988).
25.
Battisti, D. S. & Hirst, A. C. Interannual variability in a tropical atmosphere-ocean model: Influence of the basic state, ocean geometry and nonlinearity. J. Atmos. Sci. 46, 1687–1712 (1989).
26.
Jin, F. F. An equatorial ocean recharge paradigm for ENSO. Part I: conceptual model. J. Atmos. Sci. 54, 811–829 (1997).
27.
Li, T. Phase transition of the El Niño-Southern oscillation: a stationary SST mode. J. Atmos. Sci. 54, 2872–2887 (1997).
28.
Vimont, D. J., Wallace, J. M. & Battisti, D. S. The seasonal footprinting mechanism in the Pacific: Implications for ENSO. J. Clim. 16, 2668–2675 (2003).
29.
Alexander, M. A., Vimont, D. J., Chang, P. & Scott, J. D. The impact of extratropical atmospheric variability on ENSO: testing the seasonal footprinting mechanism using coupled model experiments. J. Clim. 23, 2885–2901 (2010).
30.
Chang, P., Zhang, L. & Saravanan, R. Pacific meridional mode and El Niño–Southern Oscillation. Geophys. Res. Lett. 34, L16608 (2007).
31.
Larson, S. M. & Kirtman, B. P. The Pacific Meridional Mode as a trigger for ENSO in a high-resolution coupled model. Geophys. Res. Lett. 40, 3189–3194 (2013).
32.
Larson, S. M. & Kirtman, B. P. The Pacific Meridional Mode as an ENSO precursor and predictor in the North American Multimodel Ensemble. J. Clim. 27, 7018–7032 (2014).
33.
Geng, T. et al. Increased occurrences of consecutive La Niña events under global warming. Nature 619, 774–781 (2023).
34.
Amaya, D. J. The Pacific Meridional mode and ENSO: a review. Curr. Clim. Change Rep. 5, 296–307 (2019).
35.
Anderson, B. T. Intraseasonal atmospheric variability in the extratropics and its relation to the onset of tropical Pacific sea surface temperature anomalies. J. Clim. 20, 926–936 (2007).
36.
Anderson, B. T., Perez, R. C. & Karspeck, A. Triggering of El Niño onset through trade wind-induced charging of the equatorial Pacific. Geophys. Res. Lett. 40, 1212–1216 (2013).
37.
Anderson, B. T. & Perez, R. C. ENSO and non-ENSO induced charging and discharging of the equatorial Pacific. Clim. Dyn. 45, 2309–2327 (2015).
38.
Chen, M., Yu, J. Y., Wang, X. & Chen, S. Distinct onset mechanisms of two subtypes of CP El Niño and their changes in future warming. Geophys. Res. Lett. 48, e2021GL093707 (2021).
39.
Chen, S. F. et al. Enhanced impact of the Aleutian Low on increasing the Central Pacific ENSO in recent decades. npj Clim. Atmos. Sci. 6, 29 (2023).
40.
Eusebi Borzelli, G. L. & Carniel, S. Where the winds clash: what is really triggering El Niño initiation? npj Clim. Atmos. Sci. 6, 119 (2023).
41.
Cai, W. et al. Pantropical climate interactions. Science 363, eaav4236 (2019).
42.
Wang, C. Three-ocean interactions and climate variability: a review and perspective. Clim. Dyn. 53, 5119–5136 (2019).
43.
Ham, Y. G., Kug, J. S., Park, J. Y. & Jin, F. F. Sea surface temperature in the north tropical Atlantic as a trigger for El Niño/Southern Oscillation events. Nat. Geosci. 6, 112–116 (2013).
44.
Jiang, L. S. & Li, T. Impacts of tropical North Atlantic and equatorial Atlantic SST anomalies on ENSO. J. Clim. 34, 5635–5655 (2021).
45.
Yu, J. Y. et al. Linking emergence of the Central Pacific El Niño to the Atlantic multi-decadal oscillation. J. Clim. 28, 651–662 (2015).
46.
Chao, J. H., Fan, G. Z. & Ding, R. Q. Influence of the North American Dipole on ENSO onset as simulated by a coupled ocean–atmosphere model. Atmos. Ocean. Sci. Lett. 14, 100058 (2021).
47.
Jo, H. S. et al. Southern Indian Ocean Dipole as a trigger for Central Pacific El Niño since the 2000s. Nat. Commun. 13, 6965 (2022).
48.
Kug, J. S., An, S. I., Jin, F. F. & Kang, I. S. Preconditions for El Niño and La Niña onsets and their relation to the Indian Ocean. Geophys. Res. Lett. 32, L05706 (2005).
49.
Lengaigne, M. et al. Triggering of El Niño by westerly wind events in a coupled general circulation model. Clim. Dyn. 23, 601–620 (2004).
50.
Chen, D. et al. Strong influence of westerly wind bursts on El Niño diversity. Nat. Geosci. 8, 339–345 (2015).
51.
Fedorov, A. V., Hu, S., Lengaigne, M. & Guilyardi, E. The impact of westerly wind bursts and ocean initial state on the development, and diversity of El Niño events. Clim. Dyn. 44, 1381–1401 (2015).
52.
Wang, B. Interdecadal changes in El Niño onset in the last four decades. J. Clim. 8, 267–285 (1995).
53.
Gu, D. & Philander, S. G. H. Interdecadal climate fluctuations that depend on exchanges between the tropics and extratropics. Science 275, 805–807 (1997).
54.
Fedorov, A. V. & Philander, S. G. Is El Niño changing? Science 288, 1997–2002 (2000).
55.
Choi, J., An, S. I., Kug, J. S. & Yeh, S. W. The role of mean state on changes in El Niño’s flavor. Clim. Dyn. 37, 1205–1215 (2011).
56.
Xiang, B., Wang, B., Yu, W. & Xu, S. How can anomalous western North Pacific subtropical high intensify in late summer? Geophys. Res. Lett. 40, 2349–2354 (2013).
57.
Su, J. Z. et al. Causes of the El Niño and La Niña amplitude asymmetry in the equatorial eastern Pacific. J. Clim. 23, 605–617 (2010).
58.
Huang, B. et al. Extended reconstructed sea surface 5 temperature, version 5 (ERSSTv5): upgrades, validations, and intercomparisons. J. Clim. 30, 8179–8205 (2017).
59.
Rayner, N. A. et al. Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century. J. Geophys. Res.: Atmos. 108, 4407 (2003).
60.
Compo, G. P. et al. The twentieth-century reanalysis project. Q. J. R. Meteorol. Soc. 137, 1–28 (2011).
61.
Kanamitsu, M. et al. NCEP–DOE AMIP-II reanalysis (R-2). Bull. Am. Meteorol. Soc. 83, 1631–1644 (2002).
62.
Poli, P. et al. ERA-20C: an atmospheric reanalysis of the twentieth century. J. Clim. 29, 4083–4097 (2016).
63.
Uppala, S. M. et al. The ERA‐40 re‐analysis. Q. J. R. Meteorol. Soc. 131, 2961–3012 (2005).
64.
Dee, D. P. et al. The ERA‐interim reanalysis: configuration and performance of the data assimilation system. Q. J. R. Meteorol. Soc. 137, 553–597 (2011).
65.
Hersbach, H. et al. Global reanalysis: goodbye ERA-Interim, hello ERA5. ECMWF Newsl. 159, 17–24 (2019).
66.
Kalnay, E. et al. The NCEP/NCAR 40-year reanalysis project. Bull. Am. Meteor. Soc. 77, 437–472 (1996).
67.
Adler, R. F. et al. The version-2 global precipitation climatology project (GPCP) monthly precipitation analysis (1979-present). J. Hydrometeorol. 4, 1147–1167 (2003).
68.
Carton, J. A. & Giese, B. S. A reanalysis of ocean climate using Simple Ocean Data Assimilation (SODA). Mon. Weather Rev. 136, 2999–3017 (2008).
69.
Saha, S. et al. The NCEP climate forecast system. J. Clim. 19, 3483–3517 (2006).
70.
Sun, B., Yu, L. & Weller, R. A. Comparisons of surface meteorology and turbulent heat fluxes over the Atlantic: NWP model analyses versus moored buoy observations. J. Clim. 16, 679–695 (2003).
71.
Zebiak, S. E. & Cane, M. A. A Model El Niño-Southern Oscillation. Mon. Weather Rev. 97, 163–172 (1987).
72.
Li, T., Zhang, Y., Lu, E. & Wang, D. Relative role of dynamic and thermodynamic processes in the development of the Indian Ocean dipole: An OGCM diagnosis. Geophys. Res. Lett. 29, 2110 (2002).
73.
An, S. I. & Jin, F. F. Nonlinearity and asymmetry of ENSO. J. Clim. 17, 2851–2865 (2004).
74.
Hsu, P. C. & Li, T. Interactions between boreal summer intraseasonal oscillations and synoptic-scale disturbances over the western North Pacific. Part II: Apparent heat and moisture sources and eddy momentum transport. J. Clim. 24, 942–961 (2011).
75.
Xue, M. & Li, T. To what extent does ENSO rectify the tropical Pacific mean state? Clim. Dyn. 61, 3875–3891 (2023).
76.
Gill, A. E. Some simple solutions for heat-induced tropical circulation. Q. J. R. Meteorol. Soc. 106, 447–462 (1980).
77.
Li, T., Wang, B., Chang, C. P. & Zhang, Y. S. A theory for the Indian Ocean dipole-zonal mode. J. Atmos. Sci. 60, 2119–2135 (2003).
78.
Chang, P. & Philander, S. G. A coupled ocean-atmosphere instability of relevance to the seasonal cycle. J. Atmos. Sci. 51, 3627 (1994).
79.
Roeckner, E. et al. The atmospheric general circulation model ECHAM-4: model description and simulation of present-day climate. MPI report 218, (Max-Planck-Institut für Meteorologie, 1996).
80.
Zhu, Z. & Li, T. A new paradigm for the continental United States summer rainfall variability: Asia–North America teleconnection. J. Clim. 29, 7313–7327 (2016).
81.
Wang, L., Li, T., Maloney, E. & Wang, B. Fundamental causes of propagating and nonpropagating MJOs in MJOTF/GASS models. J. Clim. 30, 3743–3769 (2017).
82.
Jiang, L. & Li, T. Relative roles of El Niño-induced extratropical and tropical forcing in generating Tropical North Atlantic (TNA) SST anomaly. Clim. Dyn. 53, 3791–3804 (2019).

Acknowledgements

This study was supported by NSFC project 42088101 and Jiangsu Province Postgraduate Scientific Research Innovation Project KYCX21_0946.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

About this article

Cite this article

Pan, X., Li, T. & Yu, J. Change of El Niño onset location around 1970. npj Clim Atmos Sci 7, 163 (2024). https://doi.org/10.1038/s41612-024-00709-y

-
Received:
-
Accepted:
-
Published:
-
DOI: https://doi.org/10.1038/s41612-024-00709-y