Understanding the Development of the 2018/19 Central Pacific El Niño

El Niño Southern Oscillation (ENSO) is the dominant interannual variation of sea surface temperature (SST) in the tropical Pacific region. The event-to-event diversity of individual ENSO events can lead to different climate impacts (Ashok et al., 2007; Wang and Wang, 2013, 2014; Capotondi et al., 2015; Yu et al., 2017). El Niño events are often classified as Eastern Pacific (EP) or Central Pacific (CP) El Niño events (Yu and Kao, 2007; Kao and Yu, 2009; Yu and Kim, 2011). An EP El Niño first exhibits surface warming in the cold-tongue region in the eastern Pacific, while a CP El Niño first exhibits surface warming in the central tropical Pacific. The latter is also referred to as an El Niño Modoki (Ashok et al., 2007) or a dateline El Niño (Larkin and Harrison, 2005).

Recent studies have revealed that different types of El Niño events are dominated by different physical processes (Kug et al., 2009; Yu et al., 2010; Chen et al., 2015; Lai et al., 2015; Wang et al., 2019b). Kug et al. (2009) demonstrated that a CP El Niño arises mainly from zonal advective feedback, while warming in the eastern tropical Pacific is suppressed by enhanced upwelling and evaporation caused by equatorial easterly anomalies. Yu et al. (2010) pointed out that during the development of a CP El Niño, SST anomalies first appear in the northeastern subtropical Pacific and extrend towards the central equatorial Pacific through wind–evaporation–SST (WES) feedback (Xie and Philander, 1994). There are also studies suggesting that the forcings from the northern subtropical Pacific could trigger the development of a CP El Niño via the seasonal footprinting mechanism (Vimont et al., 2001; Yeh et al., 2015; Yu and Fang, 2018; Fang and Yu, 2020). Chen et al. (2015) and Lai et al. (2015) emphasized that the diversity of El Niño events is a combined effect of both zonal wind anomalies and subsurface temperature anomalies in the equatorial Pacific.

Wang and Wang (2013) investigated the diversity in SST evolution and climate impact of CP El Niño events and further classified them into two subtypes, Modoki-I and Modoki-II (referred to as CP-I and CP-II hereafter). Despite the similarity of both CP types warming in the central tropical Pacific, they have distinctly different spatial patterns and regional climatic effects (Liu et al., 2014; Wang and Wang, 2014; Tan et al., 2016; Liu et al., 2017; Chen et al., 2019; Wang et al., 2020b; Kim et al., 2021). A CP-I (CP-II) El Niño tends to induce positive (negative) SST anomalies in the South China Sea during developing autumn, mainly by affecting the latent heat flux (Tan et al., 2016). Associated with a weaker Walker circulation in the Indo–Pacific region, a CP-I (CP-II) El Niño favors a positive (negative) Indian Ocean Dipole via the Bjerknes feedback (Wang and Wang, 2014). As for the western Pacific subtropical high during the decaying summer, a CP-I (CP-II) El Niño imposes little (strong positive) impact (Chen et al., 2019, 2021a). Wang et al. (2018) came up with an index to identify the two subtypes of CP El Niño events and showed that CP-II events occurred the most often after 1990. Wang et al. (2019c) investigated CP El Niño events occurring from 1900 to 2010 by analyzing the heat budget of the mixed layer water in the Niño-4 region (5°S–5°N, 160°E–150°W) and discovered that zonal advective (Ekman pumping) feedback is the leading contributor to CP-I (CP-II) El Niño events. Chen (2021b) concluded that a CP-I El Niño is triggered by the weakening of the Australian winter monsoon, while a CP-II El Niño is mainly forced by the Pacific Meridional Mode (PMM) via WES feedback.

In 2018/19, an El Niño event occurred, with the warming center located at the Date Line. The SST anomalies first appeared in the northern subtropical Pacific and then extended towards the central tropical Pacific. The surface warming was stronger in the Niño-4 region (5°S–5°N, 160°E–150°W) than in the Niño-3 region (5°S–5°N, 150°–90°W; Fig. 1). Based on observations and model forecasts, Liu et al. (2020) argued that the central tropical Pacific warming in the 2018/19 El Niño, along with tropical Atlantic warming and interdecadal variation, is one of the major factors leading to the extremely wet winter of 2018/19 in the lower reach of the Yangtze River. Wang et al. (2020a) suggested that the 2018/19 El Niño induced a remote teleconnection pattern with pronounced low-level southerly anomalies over East China, which transported moisture from oceans to the continent and caused persistent rainy days in the 2018/19 winter in Shanghai, China.

Figure 1.  Evolution of SST (shading, °C) and wind stress (vector, $ \mathrm{N}\;{\mathrm{m}}^{-2} $) anomalies in the tropical Pacific during the development of the 2018/19 El Niño. (a) MAM, (b) JJA, and (c) SON indicate the periods of March–April–May, June–July–August, and September–October–November in 2018, respectively. (d) DJF indicates the period from December 2018 to February 2019. Black vectors indicate wind stress anomalies exceeding ±1.5 standard deviations. The dotted area indicates where SST anomalies exceed ±1.5 standard deviations.

The 2018/19 El Niño event presents an opportunity to examine the genesis mechanisms of a CP El Niño. In this work, the physical processes associated with the 2018/19 El Niño are analyzed. It is demonstrated that the off-equatorial forcings in the northern subtropical Pacific are the primary cause of the 2018/19 El Niño. The contributions of the equatorial forcings, however, can be opposing during different stages.

Monthly SSTs from the Extended Reconstructed Sea Surface Temperature, version 5 (ERSST v5) dataset with a resolution of 2.0° × 2.0° (Huang et al., 2017), monthly sea level pressures, surface winds and heat fluxes from the National Centers for Environmental Prediction-National Center for Atmospheric Research (NCEP/NCAR) Reanalysis 1 with a resolution of 2.5° × 2.5° (Kalnay et al., 1996), and monthly wind stress, ocean current, and temperature data from the Global Ocean Data Assimilation System (GODAS) (Behringer and Xue, 2004) with a horizontal resolution of 0.333° latitude × 1.0° longitude are used to analyze the evolution of the 2018/19 El Niño. Ocean current and temperature data at 40 vertical levels from GODAS are used. These monthly variables cover the period from 1950 to January 2019. Monthly oceanic current and temperature data from the Simple Ocean Data Assimilation (SODA) 2.2.4 (Carton and Giese, 2008; Giese and Ray, 2011) and monthly sea surface wind data from the Twentieth Century Global Reanalysis Version 2 (20CR v2) (Compo et al., 2011), with a longer time span covering 1900–2010, are used for the composite analyses of CP-I and CP-II El Niño events. The 20CR v2 data are on a global T62 Gaussian grid. Pentad wind stress data from GODAS are also analyzed for the development of the 2018/19 El Niño. The precipitation rate data is from the Global Precipitation Climatology Centre (GPCC) Monitoring Product Version 6 (2014–present; Schneider et al., 2011) and Version 7 (1901–2013; Schneider et al., 2016), with a resolution of 1.0° × 1.0°.

The classification of EP or CP El Niño is based on the El Niño Modoki index (EMI) proposed by Ashok et al. (2007):

where the subscripts C, W, and E indicate the SST anomalies averaged in the central (165°E–140°W, 10°S–10°N), western (125°–145°E, 10°S–20°N), and eastern (110°–70°W, 15°S–5°N) Pacific, respectively. The El Niño event is considered a “typical” CP event when the EMI is equal to or greater than 0.7$ \sigma $, where $ \sigma $ is the climatological standard deviation.

CP El Niño events, as mentioned above, can be further classified as CP-I or CP-II El Niño events, with reference to their opposite influences on the precipitation in southern China (Wang and Wang, 2013). Additionally, Wang et al. (2018) developed a new index (CP-II index in this paper) to identify the subtypes of CP El Niño events. The CP-II index is defined as the leading principal component of multivariate empirical orthogonal function analysis of the normalized EMI, Niño-4 index, and 850-hPa relative vorticity anomalies averaged near the Philippine Sea (115°–145°E, 10°–25°N) during boreal autumn. CP El Niño events with a CP-II index larger than one standard deviation are designated as a CP-II El Niño.

Following Wang and Wang (2013), seven CP-I El Niño events (1914/15, 1940/41, 1941/42, 1963/64, 1987/88, 1990/91, and 2002/03) and six CP-II El Niño events (1968/69, 1979/80, 1991/92, 1992/93, 2004/05, and 2009/10) are selected as the historical collections. Monthly data are utilized to conduct the heat budget analysis of the ocean mixed-layer temperature based on the equation from Huang et al. (2010):

where the primes indicate temporal anomalies. T_t represents the temperature tendency. Q_u, Q_v, and Q_w represent the zonal, meridional, and vertical advective feedback. Q_q represents the net surface flux. R represents the residuals. The ocean mixed-layer depths, defined as the depth where the buoyancy difference with respect to the surface level is equal to 0.03 cm s^–2, are from the GODAS dataset.

The advection terms on the right side of Eq. (2) are further decomposed as

where T is the ocean mixed-layer temperature. u, v, and w represent the zonal, meridional, and vertical current, respectively. ρ is the seawater density. c_p is the specific heat capacity of seawater under constant pressure. h is the depth of the mixed layer. The overbars indicate temporal average.

Firstly, the EMI in 2018 is calculated, which exceeds 0.7$ \sigma $ from June to November [Fig. S1 in the electronic supplementary material (ESM)]. Thus, the 2018/19 El Niño is identified as a CP El Niño. Secondly, there are two ways to separate CP-I and CP-II El Niño events. The southern China rainfall during the boreal autumn of 2018 was significantly positive, the same as that of a CP-I El Niño (Fig. S2 in the ESM). The CP-II index of 2018 is 0.87$ \sigma $ (less than one standard deviation). So, the 2018/19 El Niño should be categorized as a CP-I El Niño event under either classification approach. The 2018/19 El Niño exhibited equatorial westerly anomalies during the developing February and March (Fig. 2c). The equatorial westerly anomalies generated downwelling Kelvin waves that propagated eastward and caused positive zonal advective feedback (Fig. 3b), leading to positive ocean heat content (OHC) anomalies in the central Pacific during the developing February to March (Fig. 2c). All of these features are the precursors of a CP-I El Niño, and the zonal advective feedback is supposed to support the warming in the central equatorial Pacific until the peak. The subsequent development process, however, tells a different story.

Figure 2.  Evolutions of ocean heat content (OHC) anomalies in the top 300 m (shading; units: ${10}^{9}\;\mathrm{J}\;{\mathrm{m}}^{{-}2}$) and surface zonal wind anomalies (vector; units: m s⁻¹) in the equatorial Pacific (averaged between 5°S–5°N) during the (a) CP-I El Niño, (b) CP-II El Niño, and (c) 2018/19 El Niño. In (a) and (b), (0) and (+1) on the y-axis indicate the developing and decaying year, respectively. Black dots in (a) and (b) denote areas exceeding the 90% confidence level. Black dots in (c) denote areas exceeding ±1 standard deviation.

Figure 3.  (a) Ocean mixed-layer heat budget analysis of the Niño-4 region during 2018 and (b) accumulative contributions of each term. Anomalies of horizontal current (vector) in the upper 150 m and its divergence (shading) in (c) January–February–March (JFM) and (d) June–July–August (JJA) 2018. In (a) and (b), T_t represents the temperature trend anomaly. uT, vT and wT represent anomalies of zonal, meridional and vertical advection, respectively. Qnet represents the anomaly of the net surface heat flux. R represents the residuals. Black dots in (c) and (d) denote areas exceeding ±1.5 standard deviations.

Easterly anomalies spread across the equatorial Pacific in late boreal spring and summer (April to August), induced negative zonal advective feedback (Fig. 3b), and suppressed the increase of oceanic subsurface temperature in the central equatorial Pacific. Yet, the oceanic subsurface temperature in the central equatorial Pacific still increased notably in late boreal summer without significant westerly anomalies in the western equatorial Pacific (Figs. 2b, c). It will be demonstrated below that this notable warming in the central equatorial Pacific is mainly ascribed to the Ekman feedback induced by the subtropical westerly anomalies.

During the spring of 2018, westerly anomalies were mainly located in the central subtropical Pacific (5°–10°N, 170°E–130°W), and positive SST anomalies (exceeding 1°C) were in the northeastern subtropical Pacific (Fig. 1a). Under the influence of WES feedback, the surface warming extended from the northeastern Pacific to the central equatorial Pacific. When the central Pacific SST anomalies were adequately positive, westerly anomalies arose in the western equatorial Pacific (Fig. 1c). The westerly anomalies and the surface warming in the central Pacific enhanced themselves through Bjerknes feedback (Bjerknes, 1969) and peaked in the boreal winter of 2018/19. Particularly, the developing patterns of SST anomalies and wind anomalies resembled those of a CP-II El Niño (Figs. S3e–h in the ESM).

Evolutions of composite OHC anomalies in the top 300 m and surface zonal wind anomalies in the equatorial Pacific during CP-I and CP-II El Niño events are illustrated in Fig. 2a and Fig. 2b, respectively. For the CP-I El Niño events, the significantly positive OHC anomalies in the eastern-central Pacific and westerly anomalies in the western-central equatorial Pacific persist during the whole developing year, indicating that the equatorial ocean and the atmosphere are well coupled. For the CP-II El Niño events, OHC anomalies in the central equatorial Pacific begin to increase remarkably in developing summer, even though strong westerly anomalies have not taken place in the equatorial Pacific yet. According to Wang et al. (2019c), the warming in the central Pacific is mainly caused by the westerly anomalies in the northern subtropical Pacific via Ekman feedback. As a critical feature of a CP-II El Niño, this process recurred in 2018. During the summer (June–August) of 2018, while strong westerly anomalies were still absent in the western-central equatorial Pacific, the OHC anomalies increased sharply (Fig. 2c). After the increase of OHC in the central equatorial Pacific, westerly anomalies prevailed in the western equatorial Pacific until the spring of 2019 (Fig. 2c).

Characteristic features of the CP-I, CP-II, and 2018/19 El Niño mentioned above include the evolutions of equatorial SST anomalies (Fig. S4 in the ESM) and thermocline anomalies (Fig. S5 in the ESM). In particular, the multi-occurrence of downwelling Kelvin waves, which is attributed to the evolution characteristics of the wind anomalies, during the boreal spring, autumn, and winter of 2018 is clearly illustrated in Fig. S1c and Fig. S4c.

To isolate the warming mechanism, the ocean mixed-layer heat budget for the Niño-4 region is analyzed for the CP-I, CP-II, and 2018/19 El Niño. For the CP-I El Niño, the leading contributors to the mixed-layer temperature warming are the zonal and meridional advective feedbacks (Fig. S6a in the ESM). During the whole developing phase of the CP-I El Niño, the contribution from vertical advective feedback remains small. In comparison, vertical advective feedback is the leading accumulative contributor in boreal spring and summer during the CP-II El Niño (Fig. S6c). Due to the absence of significant surface zonal wind anomalies in the western equatorial Pacific (Fig. 2b), the contributions of zonal advective feedback and meridional advective feedback are rather weak in boreal spring (Fig. S6d).

It is evident that vertical advective feedback plays the leading role in the development of the 2018/19 El Niño, especially in boreal spring and autumn (Figs. 3a, b). Thus, the positive ocean temperature anomalies in the central equatorial Pacific during the CP-II and 2018/19 El Niño can reach a deeper depth than those during the CP-I El Niño (Fig. S7 in the ESM). The only characteristic in line with the CP-I El Niño is the early boost of the zonal advective feedback. The contribution from zonal advective feedback reached its peak before April of 2018 (Fig. 3b), when westerly anomalies prevailed in the western equatorial Pacific and generated downwelling Kelvin waves that propagated eastward to the central Pacific (Fig. 2c). Associated with eastward current anomalies in the upper equatorial Pacific (Fig. 3c), the resultant positive zonal advective feedback made a considerable contribution to the warming of the central Pacific. Thus, the upper-ocean temperature warms much earlier in the 2018/19 El Niño than it does in the CP-II El Niño (Fig. S7i), consistent with the results of the OHC evolution (Fig. 2c). In July, easterly anomalies took the place of the westerly anomalies in the western Pacific (Fig. 2c), generating upwelling Kelvin waves, and the equatorial current anomalies turned to be westward (Fig. 3d). Thus, the zonal advective feedback reversed to be negative, and the Niño-4 temperature trend anomaly dropped sharply. While the zonal current anomalies changed their direction, the central equatorial Pacific (Niño-4 region) maintained anomalous convergence in the upper ocean the entire time (Figs. 3c, d). In this way, the climatological upwelling in the central equatorial Pacific was hindered, and the Ekman feedback kept the Niño-4 temperature anomalies increasing.

The upper-layer divergence in the central equatorial Pacific during 2018 summer is further decomposed into two parts (Fig. 4). One part is led by the zonal current anomalies, and the other is led by the meridional current anomalies. While the zonal-led divergence anomalies (du'/dx) are mainly positive in the Niño-4 region (averaged 1.87 × 10^–8 s^–1; Fig. 4b), the meridional-led divergence anomalies (dv'/dy) are mostly negative (averaged –2.31 × 10^–8 s^–1; Fig. 4d). Therefore, the total divergence in the Niño-4 region adds up to be negative (i.e., an anomalous convergence) and is dominated by the meridional-led part.

Figure 4.  (a) The oceanic convergence in the upper 150 m solely led by the zonal current anomalies ($\mathrm{d}{u}^{{'}}/\mathrm{d}x$) in June–July–August (JJA) 2018. (b) The oceanic convergence in (a) averaged between 5°S and 5°N (units: 10^–8 s^–1). (c) and (d) are the same as (a) and (b), except that the convergence is solely led by the meridional current anomalies ($\mathrm{d}{v}^{{'}}/\mathrm{d}y$). Dashed lines in (b) and (d) indicate the zonal boundaries of the Niño-4 region.

In order to understand the influence of subtropical westerly wind forcing on the anomalous convergence through Ekman transport, the meridional Ekman flow is calculated by:

where V_E is the meridional Ekman current (positive for northward flow, negative for southward flow), τ_x is the zonal wind stress, ρ is the seawater density (1025 kg m^–3), and f is the Coriolis parameter (f = 2Ωsinθ, with Ω and θ equal to Earth’s angular velocity and latitude, respectively). Zonal wind stress averaged within (6°–10°N, 160°E–150°W) is used to represent the westerly wind forcing in the subtropical North Pacific. The resultant V_E anomaly in 2018 summer is –0.94 m s^–1 (see Fig. S8 in the ESM for details), while the anomaly of the mean meridional current in the mixed layer across the section (5°N, 160°E–150°W) is –0.32 m s^–1. This demonstrates the vital role of the subtropical westerly wind forcing on the central equatorial Pacific.

To sum up, the anomalous convergence in the central equatorial Pacific was induced by the off-equatorial westerly anomalies (Figs. 1a, b). It favored the warming in the central equatorial Pacific by suppressing the climatological upwelling and inducing positive vertical advective feedback (Wang et al., 2019c). When the Niño-4 region got warm enough in autumn, the Bjerknes feedback became effective, bringing about westerly anomalies and downwelling Kelvin waves again in the equatorial Pacific. As a result, the zonal advective feedback became positive and dominant. In the wake of the propagation of downwelling Kelvin waves, the whole equatorial Pacific warmed up and the El Niño conditions formed.

In this work, the evolution of the 2018/19 El Niño was investigated. The 2018/19 El Niño is classified as a CP-I El Niño according to existing classification methods. However, heat budget analysis demonstrated that the development of the 2018/19 El Niño is mainly attributed to subtropical processes, which is the key feature of a CP-II El Niño. This apparent paradox results from the tortuous developing process of the 2018/19 El Niño, since both equatorial and off-equatorial forcings were extensively involved. In early 2018, both the equatorial and subtropical conditions were favorable for the development of an El Niño event. While the favorable equatorial conditions faded and even reversed in boreal summer, the positive Ekman feedback caused by subtropical westerly anomalies kept the western-central Pacific SST increasing. Consequently, the subtropical factor took control of the development process and finally materialized a CP-II El Niño. It is illustrated that forcing factors from the equatorial and subtropical Pacific have the potential to promote an El Niño together, like the widely discussed 2015/16 El Niño (Huo and Xiao, 2016; Palmeiro et al., 2017). Although the influences of equatorial and off-equatorial forcing elements on the development of El Niño events have been discussed in previous studies (Yu et al., 2010; Chen et al., 2018; Yeh et al., 2018), their joint effect is seldom addressed. The findings of this study demonstrate that various driving forces may coexist in a particular El Niño event and play different roles in different stages. This is the reason for the inconsistency between the classification result and the evolution features of the 2018/19 El Niño. Such a hybrid situation is not considered in the existing classification approaches. To improve the classification, a more sophisticated method is needed. Perhaps the components of the CP-II index can be subdivided. Or in a more direct and mechanistic way, the contributions of the zonal and vertical advective feedbacks could be used; an understanding of their combined effect will help us better understand the complexity of ENSO diversity and improve ENSO prediction skill.

The importance of wind anomalies in the northern subtropical Pacific is revealed in this work. Yet the source of these wind anomalies has not been clear. Previous studies have suggested that the North Pacific Oscillation (NPO, Rogers, 1981) and Pacific Meridional mode (PMM, Chiang et al., 2004) in the subtropical North Pacific play an important role in the onset of the CP El Niño (Yu and Kim, 2011; Yeh et al., 2015; Wang et al., 2019b). The westerly anomalies in the subtropical northern Pacific, which are demonstrated to be a crucial forcing for the 2018/19 El Niño in this study, are closely related to both the NPO and PMM. The impacts of the NPO and PMM on the generations of other CP El Niño events will be explored in a follow-up study. Another crucial element that affected the development of the 2018/19 El Niño is the reversal of the central equatorial wind anomalies from westerly to easterly in the boreal summer of 2018. This equatorial forcing element is likely to be a result of the large zonal SST anomaly (SSTA) gradient in the equatorial Pacific, which is associated with positive (negative) SSTA in the western (eastern) Pacific (Fig. S4, January–June). It is also suggested that the mean state of the Pacific (negative phase of the Pacific Decadal Oscillation /Interdecadal Pacific Oscillation) is favorable for the formation of the easterly anomalies in the tropical Pacific (Min et al., 2015; Hu and Fedorov, 2016). The interrelationship between these intraseasonal and decadal stimuli deserve further analysis. In addition, a coupled model will be needed to further quantify the effects of equatorial and off-equatorial forcings.

Acknowledgements. This work is supported by the National Natural Science Foundation of China (Grant Nos. 41925024 and 41876021), Strategic Priority Research Program of Chinese Academy of Sciences (Grant No. XDB42000000), Key Special Project for Introduced Talents Team of Southern Marine Science and Engineering Guangdong Laboratory (Guangzhou) (GML2019ZD0306), Innovation Academy of South China Sea Ecology and Environmental Engineering, Chinese Academy of Sciences (ISEE2021ZD01), and Natural Science Foundation of Shandong Province, China (Grant No. ZR2020QD065).

Electronic supplementary material: Supplementary material is available in the online version of this article at https://doi.org/10.1007/s00376-022-1410-1.