1 Introduction

The El Niño, which occurs in the equatorial Pacific Ocean, is a significant natural variability in atmosphere–ocean interaction systems (Chen et al. 2016; Yang et al. 2018). El Niño has a distinct effect not only on the equatorial Pacific but also on the global circulation and various climate factors in the atmosphere and oceans (Cane 2005; Cheon 2020). Half a century ago, Bjerkens proposed a mechanism for the El Niño—anomalous sea surface temperature (SST) warming in the eastern Pacific. The weakening of trade winds due to unspecified reasons has been attributed to the movement of warm water from the western Pacific to the eastern Pacific based on low-resolution oceanic data (Bjerkens 1969). Since 1990, two extreme (1997 and 2015) and three moderate (1994, 2003, and 2012) El Niño events have occurred (Li et al. 2020). The 1997 and 2015 El Niño events are considered the two most extreme El Niños in the historical recorded (Kao and Yu 2009; Iskandar et al. 2018). Typically, El Niño events are believed to be triggered by the weakening of trade winds due to multiple westerly wind bursts (WWBs) (Fedorov 2002; Gebbie and Tziperman 2009). As a strong WWB occurred in the spring of 2014 (Menkes et al. 2014), many research institutes predicted that a strong El Niño event would occur at the end of that year (Tollefson 2014; Levine and McPhaden 2016). Contrary to those predictions, the strongest El Niño occurred in 2015. The 2015 El Niño was characterized by strong and continuous WWBs during the spring and summer months (Chen et al. 2017; Hong et al. 2017; Hu and Fedorov 2017), leading to the accumulation and subsequent expansion of heat in the western Pacific (Su et al. 2017). The 1997 El Niño was also thought to have grown rapidly after WWBs during the spring months (Lengaigne and Vecchi 2010).

Recent studies have classified each El Niño as Eastern Pacific El Niño (EP El Niño) or Central Pacific El Niño (CP El Niño) events according to the spatial distribution of their SST (Ashok et al. 2007; Kug et al. 2009; Karnauskas 2013). The traditional EP El Niño explains the rise of SST in the eastern Pacific with thermocline variation and upwelling, and it is widely accepted and used in various El Niño conceptual models (Jin 1997; Kao and Yu 2009; Webb 2018). However, the CP El Niño, which is caused by zonal advection and wind-driven forcing and has predominantly occurred since 1990, has significantly different characteristics from the traditional El Niño in its initiation stage and impact on society (Ham and Kug 2012; Yeh et al. 2014; Fang and Mu 2018). El Niño evolution mechanisms exhibit significant irregularities in intensity, duration, and spatiotemporal structure (Vimont et al. 2016). General circulation models reflect the characteristics of the traditional EP El Niño but do not adequately simulate the CP El Niño that has been dominant since the 2000s (Ham and Kug 2012). The diversity and complexity of El Niño have recently highlighted the significance of these mechanisms for El Niño evolution, yet the argument between EP and CP El Niño continues (McPhaden et al. 2006; Ren and Jin 2013; Guan and McPhaden 2016).

Due to the influence of trade winds, the major ocean currents near the equator in the Pacific Ocean are the North Equatorial Current (NEC) and the South Equatorial Current (SEC), with the North Equatorial Countercurrent (NECC) flowing toward the east. The NEC caused by easterly trade winds in the tropical Pacific Ocean is a westward-flowing current between latitudes 10° N and 25° N, forming a clockwise-rotating subtropical gyre in the Northern Hemisphere. As the NEC does not connect to the equator, it separates from the SEC, and the SEC rotates counterclockwise. The westward trade winds and eastward trade winds converge to form the Intertropical Convergence Zone (ITCZ) between the SEC and the NEC, between latitudes 5° N and 10° N. Ocean currents in the equatorial region change their location and intensity according to seasonal shifts in the ITCZ (Schneider et al. 2014; Xie et al. 2018) and interannual fluctuations in the strength of the trade winds associated with El Niño(Zhao et al. 2013; Webb 2018; Wijaya and Hisaki 2021).

Ocean surface currents (OSC) in the tropical Pacific play an important role in heat transfer, and it is important to monitor OSC to study El Niño evolution (Picaut et al. 1996; Bonjean and Lagerloef 2002; Jo et al. 2014; Azminuddin et al. 2019; Chang 2021). OSC also varies with latitude and at the equator, where the El Niño occurs (Dohan 2017). A major component of the equatorial current system, the NECC, flows eastward across the Pacific Ocean between latitudes 3° N and 10° N against the prevailing trade winds in the tropical Pacific (Johnson et al. 2002). Several studies have shown that the NECC plays a role in regulating the tropical Pacific climate by transporting significant amounts of heat from the western Pacific warm pool to the eastern Pacific (Donguy and Meyers 1996; Masunaga and L’Ecuyer 2011; Zhou et al. 2021). Most classical theories focus on the weakening of trade winds throughout the tropical Pacific and wind-driven forcing, such as equatorial Kelvin and Rossby waves (Kessler et al. 2003; Milliff et al. 2004; Zhao et al. 2013). Simple waves, such as equatorial Kelvin waves, are capable of transferring energy and momentum, but they have difficulty in transferring individual particles or particle properties such as temperature and salinity (Webb 2018). Hsin and Qiu (2012) suggest that only EP El Niño has a significant effect on the NECC, whereas the effect of CP El Niño is minimal. However, Tan and Zhou (2018) argue that both EP and CP El Niños have significant effects on the NECC, with EP El Niño exerting a greater influence on the eastern basin and CP El Niño exerting a greater influence on the western basin, because the amplitudes of their wind forcing anomalies are comparable but their spatial distributions are distinct. Numerical models cannot fully explain the dynamic complexity of the NECC, and single satellite and local observations cannot explain the overall changes in the Pacific Ocean. Therefore, the role of the NECC in moving warm waters is still unclear.

The background mean state (BMS) of the ocean in the tropical Pacific is important, because it affects the location and intensity of the El Niño (Zhao et al. 2016; Wang and Hendon 2017). The Niño region, the standard El Niño monitoring area, is an important area for studying and predicting the evolution of El Niño (Chen et al. 2023). In this study, we use satellite-derived observational data sets to conduct a comprehensive analysis of the spatiotemporal patterns of key surface variables (zonal wind stress (ZWS), OSC, and SST) within the El Niño monitoring region. In addition, we identify the variability in the intensity of these parameters according to latitude and longitude. This approach enables us to discern and quantify the distinctive impacts exerted by various El Niño types on these crucial climatological factors.

2 Data and Methods

We used high-quality time series satellite data to identify the relationship between winds, currents, and SST in the equatorial Pacific. Long-term wind products are estimated using surface wind data from a variety of scatterometers and radiometers on board ocean-observing satellites. Satellite-based scatterometers detect microwave backscatter from centimeter-scale waves at the ocean surface, which are influenced by atmosphere–ocean stress rather than airspeed. Wind stress calculated from scatterometer data is a natural way to account for ocean motion. The scatterometers and radiometers used to estimate long time series have some interesting characteristics from 1991 to the present (Table 1). These wind data are used for 6-hourly (00 h:00, 06 h:00, 12 h:00, and 18 h:00 UTC) global surface wind analysis using a 0.25° latitudinal and longitudinal grid based on the kriging technique combined with external drift (Bentamy and Fillon 2012). The wind data set was obtained from the Copernicus Marine Environment Monitoring Service (CMEMS) for Global Ocean Wind Products (WIND_GLO_WIND_L4_REP_OBSERVATIONS_012_006) and the new hourly L4 multi-year (MY) product is released (WIND_GLO_PHY_L4_MY_012_006) https://doi.org/10.48670/moi-00185. We used the downward surface eastward stress data from the wind field data. The surface stress refers to the shear stress applied by wind at the atmospheric lower boundary. Specifically, “downward surface eastward wind stress” describes the component of this stress that is both downward and eastward. This stress is a downward momentum flux moving eastward, effectively accelerating the surface layer (like ocean currents) toward the east while simultaneously imparting a westward momentum to the layer above. This concept is crucial in understanding wind-driven surface currents over large water bodies, where wind stress acts as a primary force influencing their direction and strength. In this study, we used the downward surface eastward stress as the zonal wind stress.

Table 1 Satellite-based observation instruments for ocean surface wind and current
Full size table

OSC can be estimated from satellite data using a simplified physical model of an ocean surface layer. Currents are averaged over the first 30 m of the solution. Ocean surface current analyses real-time (OSCAR), funded by NASA Earth Science, is an ongoing research project aimed at improving the accuracy of both surface current calculations based on various satellite data sets (Table 1) and the methods used to validate these calculations (Santiago-Mandujano and Firing 1990; Lagerloef et al. 1999; Bonjean and Lagerloef 2002). The velocity product from OSCAR is calculated using altimetry, scatterometer winds, sea surface temperature, and in situ instruments. Additional surface buoyancy gradient terms are added to the model to account for geostrophic effects. The ocean surface velocity includes zonal and meridional velocities. The near-surface velocity is directly derived from sea surface height, wind stress, and SST. The gradient of the sea surface height fields is used to calculate the geostrophic term, and the thermal wind coupling also contributes to the vertical shear via gradients of the SST. The OSCAR source data sets are obtained as follows: sea surface height data: SSALTO/DUACS delayed-time absolute dynamic topography (ADT) global gridded product distributed by the Copernicus Marine Environment Monitoring Service (CMEMS) and produced by the DUACS multi-mission altimeter data processing system, wind data: ERA5 10 m reanalysis vector winds produced by ECMWF (European Centre for Medium-Range Weather Forecasts), SST data (1993 to 2015): Canada Meteorological Center, SST data: (2016 to present) (Brasnett 2008; Meissner et al. 2016). This data set has been extensively calibrated and validated using both 15-m drifters and fixed-position observations from the global TAO OceanSITES (Ocean Sustained Interdisciplinary Timeseries Environment Observation System) project (Johnson et al. 2007; Dohan and Maximenko 2010). OSCAR currents are computed at 5-day intervals or the half cycle of Jason-1 or TOPEX/Poseidon. The data coverage is from − 80° to 80° latitude and from 0° to 360° longitude, and the resolution is a one-third degree grid (Johnson et al. 2007). The OSCAR products were produced using Earth Space Research (ESR) https://www.esr.org/research/oscar/oscar-surface-currents/. The data set used in this work was obtained from the OSCAR third-degree resolution ocean surface currents (OSCAR_L4_OC_third-deg); https://doi.org/10.5067/OSCAR-10D01.

The National Oceanic and Atmospheric Administration (NOAA) derives SST data from a series of NOAA satellites (Banzon et al. 2016) using high-resolution optimum interpolation (OI) from the advanced very-high-resolution radiometer (AVHRR) infrared satellite. SST analysis was performed using OI with a spatial grid resolution of 0.25° and a temporal resolution of 1 day (Reynolds et al. 2007; Huang et al. 2021). We used the NOAA OI SST V2.1 product provided by the Physical Sciences Division of the NOAA Earth System Research Laboratory (ESRL). The NOAA OI SST V2.1 product was provided by the Physical Sciences Division of the NOAA Earth System Research Laboratory (ESRL) (https://psl.noaa.gov/data/gridded/data.noaa.oisst.v2.highres.html).

We used an in-situ global ocean temperature data set called Coriolis Ocean database for ReAnalysis 5 (CORA5) (Cabanes et al. 2013; Szekely et al. 2019). The latest version, CORA 5.2 has been released: https://doi.org/10.17882/46219 (Szekely et al. 2019). The CORA5.2 data set provides in-situ ocean temperature and salinity measurements collected from various sources (mainly Argo floats, XBT, CTD and XCTD, moorings, sea mammal data, and some drifting buoys). The CORA5.2 measurements are carefully validated by the Coriolis data center and distributed by the Copernicus Marine Service. This product covers the period 1990–2021 and from the surface to a depth of 2000 m. In this paper, we calculated the ocean heat contents in the equatorial Pacific Ocean as follows:

$$ H_{{\text{cell}}} = \rho \times {C_{\text{p}}} \times T \times V $$

where ρ is the density of seawater (approximately 1025 kg/m3), Cp is the specific heat capacity of seawater at constant pressure [approximately 3993 J/(kg·K)], and T is the seawater temperature. We calculated the volume of each grid cell with a 1-degree latitude and longitude interval and a depth of 10 m.

3 Results

3.1 Background Mean State (BMS) (27-Year Period) of ZWS, OSC, and SST

Figures 1 and 2 show the BMS of the Pacific Ocean during the 27-year satellite period. The ZWS indicates that northeast and southeast winds were generated north and south of the equator, respectively (Fig. 1). The strong westward winds were observed, which dominated across the Pacific Ocean. The south and north equatorial currents flowed from east to west owing to northeast and southeast trade winds. The NECC, which is biased to the north in the opposite direction to the northeast trade wind, flows from west to east. A warm (27 °C or higher) water region was observed in the western Pacific, with a warm water band extending to the eastern Pacific between latitudes 0°–10° N. The ZWS flowed westward from the equator with comparable strength in the Niño 3 and Niño 4 regions. However, the OSC intensities varied with respect to the equator.

Fig. 1
figure 1

Mean spatial pattern of 27-year satellite data recorded in the tropical Pacific. The mean vector was obtained using data from 1993 to 2019. a Near-surface wind climatology for wind speed vector (black arrow) and zonal wind stress (ZWS) (color contours). The arrow corresponds to 6 m/s. The color of the contours shows the intensity of the east–west ocean surface zonal current (OSC). b Climatology for OSC and intensity (color contours). The arrow corresponds to 1 m/s. The color of the contours represents the OSC intensity in the east–west directions (c) sea surface temperature (SST) climatology. Color shading depicts the monthly mean SST above 18 °C in the tropical Pacific

Full size image
Fig. 2
figure 2

Seasonal mean spatial pattern using 27-year data (1993–2019) for near-surface wind (NSW), ocean surface current (OSC), and sea surface temperature (SST). a–d NSW for wind speed vector (black arrows) and zonal wind stress (ZWS) (color contours) and e–h OSC climatology on MAM, JJA, SON, and DJF. The color of the contours represents the ZWS and OSC intensities in the east and west directions. i–l SST climatology in MAM, JJA, SON, and DJF. The 3-month mean values are presented for MAM (March, April, and May), JJA (June, July, and August), SON (September, October, and November), and DJF (December, January, and February)

Full size image

The ZWS spatial pattern displayed strong westward winds throughout the tropical Pacific, with weak winds across the western and eastern Pacific continents (Fig. 2). In the western Pacific, there were weak eastward winds at latitudes 0°–10° S in March, April, and May (MAM), moving north of 0°–10° N in June, July, and August (JJA) and September, October, and November (SON), and stronger eastward winds in December, January, and February (DJF), extending to the dateline and descending to 0°–10° N. In the 5° N region in JJA and SON, westward winds decreased owing to the influence of southeast trade winds. The NECC intensity increased in JJA and moved south, then north in SON, and thereafter decreased in DJF. The distribution of the high SST above 29 °C extending into the eastern Pacific was the widest in SON and the narrowest in JJA. In JJA and SON, the SST zone above 27 °C extended to latitudes above 20° N. In JJA and SON, the position of the warm water band shifted to the north. It appears spatially different from the location and intensity of the NECC.

Between the ZWS zonal mean (mean longitude of 160° E–90° W) and latitude 20° S–20° N, there was only a westward wind (Fig. 3). At the 27-year background mean, the ZWS shows seasonally varying intensities and is dominated by westward winds. The westward wind had a maximum value (− 0.1 N/m2) at latitude 10° S in DJF and a minimum value (− 0.01 N/m2) at latitude 7° S in DJF. The interval between the maximum and minimum values was 0.09 N/m2 in the Northern Hemisphere and 0.06 m/s in the Southern Hemisphere, with the range of ZWS variation being larger in the Northern Hemisphere. The OSC is dominated by westward currents, with the exception of the NECC. The NECC intensity peaks were located at 5° N (0.03 m/s), 6° N (0.25 m/s), 7° N (0.3 m/s), and 6° N (0.19 m/s) in MAM, JJA, SON and DJF, respectively. The NECC intensity was the strongest in SON and weakest in MAM. In SON and JJA, the OSC and ZWS had almost identical peaks at latitudes 6° N and 7° N. However, in MAM and DJF, the ZWS moved westward, and the eastward current was weaker than that in other periods. In the off-equatorial region, winds were stronger in MMA and JJA with a westward wind intensity, while currents were strongest in the equatorial region. Near latitude 20° N, the seasonal patterns in OSC and ZWS were similar, but between the equator and latitude 10° N, the strength and direction of the seasonal trends differed. This suggested that the two parameters may be related but were influenced by different factors or had different response times to seasonal changes.

Fig. 3
figure 3

East–west intensity of seasonal climatology (1993–2019) for zonal wind stress (ZWS) and ocean surface current (OSC) for latitudes 20° S–20° N in the tropical Pacific Ocean. The mean longitude is in the range of 160° E–90° W. At each latitude, the mean ZWS and OSC intensity in the east–west direction is expressed in a 1-degree grid. East–west a ZWS and b OSC intensities in MAM, JJA, SON, and DJF. The 3-month mean values are presented for MAM (March, April, and May), JJA (June, July, and August), SON (September, October, and November), and DJF (December, January, and February)

Full size image

The intensity of the ZWS with longitude (mean latitude 3° S and 10° N) had peak values of a westward wind intensity between longitudes 180° W and 130° W (Fig. 4). The variation in ZWS with longitude was dominantly westward across the tropical Pacific, although there were variations in intensity. The variability in intensity and location of each period was nearly identical. The OSC eastward intensity was strongest in the western Pacific during JJA, SON, and DJF, weakened toward longitude 120° W across the central Pacific, and then increased weakly again in the eastern Pacific. The change in movement shown by the OSC was larger than that shown by the ZWS, and all periods except for MAM showed noticeable trends. Owing to the low intensity of the eastward movement from the MAM, this trend was not well observed. The shift and intensity of ZWS and OSC at specific longitudes were not consistent with longitude. The OSC does not coincide with the ZWS in time and sometimes shows opposite trends. We infer from Figs. 3 and 4 that ZWS can have an indirect effect on surface currents but it is not strong enough to have a direct effect. OSC can be driven not only by wind but also by thermodynamic processes within the ocean, the slope of the sea surface, and factors such as the Coriolis effect.

Fig. 4
figure 4

East–west intensity of seasonal climatology (1993–2019) for zonal wind stress (ZWS) and ocean surface current (OSC) for longitudes 160° E–80° W in the tropical Pacific Ocean. The mean latitude is in the range of 3° S–10° N. At each longitude, the mean ZWS and OSC intensity in the east–west direction was expressed in a 1-degree grid. East–west a ZWS and b OSC intensities in MAM, JJA, SON, and DJF. The 3-month mean values are present for MAM (March, April, and May), JJA (June, July, and August), SON (September, October, and November), and DJF (December, January, and February)

Full size image

3.2 Variation in ZWS, OSC, and SST During the Extreme El Niño Events

The El Niño onset phase was defined as the period in which positive SST anomalies first appeared in the eastern or central Pacific, and the main phase was defined as the period in which continuous positive SST anomalies appeared in the eastern Pacific. The decay phase was defined as the period during which the positive SST anomaly decreased after reaching its maximum phase. In 2014, the positive SST anomaly almost vanished from the central Pacific in JJA, although it slightly reappeared in 2014 SON and 2014 DJF. This event continued to grow in a mixed form of EP and CP El Niño through 2014 SON, 2014 DJF, and 2015 MAM; peaked in 2015 SON and 2015 DJF; and then gradually declined through 2016 MAM (Fig. 5). We considered the following: 2014 JJA to be the El Niño collapse period; 2014 SON, 2014 DJF, and 2015 MAM to be the onset phases; 2015 JJA, 2015 SON, and 2015 DJF to be the main phases; and 2016 MAM to be the decay phase.

Fig. 5
figure 5

Maps of vectors (black arrows) and intensities (color contours) for zonal wind stress (ZWS) and ocean surface current (OSC) and the sea surface temperature (SST) anomaly in the tropical Pacific Ocean during the onset, main, and decay period of the 2015 extreme El Niño event. The left-hand-side column shows the ZWS spatial pattern, middle column the OSC spatial pattern, and right-hand-side column the spatial pattern of the SST anomaly. SST anomalies were calculated with respect to the 1993–2019 climatology. The dark green rectangles indicate the Niño 3 and Niño 4 regions

Full size image

The 2014 ZWS was somewhat stronger than the BMS in 2014 JJA, 2014 SON (latitude 0°–10° N), and 2014 DJF (0°–10° S) in the western Pacific but had a comparable pattern (Fig. 5). An eastward wind occurred between latitudes 5° S–5° N in the 2015 MAM in the western Pacific Ocean. The region of occurrence moved northward from the BMS and extended toward the dateline. In 2015 JJA and SON, the western Pacific eastward wind weakened and moved northward to latitude 10° N. At this time, the equatorial trade winds weakened in an S-shape, lying from the central to the eastern Pacific. At latitudes 0°–10° S, the eastward wind from 2015 DJF to the western Pacific declined as the northwest wind from the North Pacific intensified. It strengthened up to longitude 170° W, and the trade winds in the central and eastern Pacific regained their intensity. During the El Niño main phase, the eastward wind intensity in the western Pacific strengthened and expanded with an increase in the SST anomaly in the central and eastern Pacific, but at separate locations. Under the BMS, the southeast and northeast trade winds weakened in JJA and SON, and the southeast trade wind tilted toward the eastern Pacific. The ZWS of the 2015 El Niño was characterized by stronger eastward winds than the BMS and weaker equatorial trade winds. The eastward wind of the ZWS was isolated to a region between latitudes 0°–10° N and longitudes 150°–180° E during the main phase; although the westward wind had weakened in the central and eastern Pacific, it remained dominant. From 2014 to 2015 SON, the westward wind was weak across the equator, with 2015 JJA and 2015 SON being the weakest periods owing to the southeast trade winds. At this time, the NECC intensity was strong in 2014 JJA and 2014 SON (onset) and 2015 JJA and 2015 SON (main). The NECC was weakened in 2014 DJF (onset), 2015 DJF, and 2016 MAM (dissipation). The NECC in the Niño region flowed outward in the onset and main phases and disappeared in the decay phase, except for the weak movement of 2015 JJA and 2015 MAM. Except for the 2015 MAM, the NECC followed the same pattern as the BMS.

The spatial pattern of the 1997 positive SST anomaly began to develop in the eastern Pacific in 1997 MAM (onset) and rapidly reached its main phase in 1997 JJA and SON (Fig. 6). It spread north and south from the equator in the eastern and central Pacific in 1997 DJF, and then declined in 1998 MAM (dissipation). In 1997 MMA, an eastward wind occurred in the western Pacific, and the trade wind intensity weakened across the equator during the main phase. The latitude with the weakened trade wind shifted to the north owing to the effect of the southeast trade winds tilting toward the eastern Pacific in 1997 JJA and 1997 SON. In 1997 DJF, the eastward wind shifted south from the western Pacific to latitudes 0°–10° S, and the trade wind intensity across the equator once again strengthened. The eastward wind formed between latitudes 0° and 10° S with the strengthening of the northeast trade wind in 1998 MAM, and a strong eastward wind formed between longitudes 160° W and 150° W. In 1998 JJA, the eastward wind completely dissipated in the western Pacific and occurred at the equator in the eastern Pacific.

Fig. 6
figure 6

Maps of vectors (black arrows) and intensities (color contours) for zonal wind stress (ZWS) and ocean surface current (OSC) and the sea surface temperature (SST) anomaly in the tropical Pacific Ocean during the onset, main, and decay period of the 1997 extreme El Niño event. The left-hand-side column shows the ZWS spatial pattern, middle column the OSC spatial pattern, and right-hand-side column the spatial pattern of the SST anomaly. SST anomalies were calculated with respect to the 1993–2019 climatology. The dark green rectangles indicate the Niño 3 and Niño 4 regions

Full size image

The eastward current was the weakest during the MAM period in BMS, and in 1997 MAM and SON, the eastward current moved from the western Pacific to longitude 150° W within the Niño region, whereafter it moved out of this region. In 1997 DJF, the NECC moved weakly out of the Niño region, and the westward current intensity in the Niño region increased. In 1998 MAM, the westward current flowed into the Niño region, and the NECC dissipated in 1998 JJA.

For the 1997 El Niño, the positive SST anomaly was strongest in 1997 SON and 1997 DJF. In 1997 JJA, 1997 SON, and 1998 MAM, the eastward wind appeared strongly throughout the equatorial region, and at this time, NECC movement occurred strongly in the order of 1997 SON, 1997 JJA, and 1997 MAM. In the 1997 El Niño period, the eastward current and wind were strong during the main phase, but their intensities and directions differed during the onset and decay phases. In 1998 MAM, the wind intensity was weakened throughout the equatorial region (especially at longitude 150° W), but the westward current moved strongly inside the Niño region in the eastern Pacific. The eastward wind was similar to that of the BMS, whereas the NECC was different. This result for the NECC could be explained by its mechanism, which was different from that of wind-driven forcing.

During the 2015 El Niño period, the eastward wind and current differed in intensity; however, their positions were similar to those during the BMS. As it takes approximately 2–3 months for the eastward current to transport warm water in the western Pacific (Lumpkin and Dohan 2016; Dohan 2017), the NECC must move strongly into the Niño region during the onset phase prior to the main phase. However, in the 1997 El Niño, the eastward current was strong during the main phase. The eastward current did not move into the Niño region, except in 2015 JJA. The westward current was dominant inside the Niño region, and the position and intensity of this current pushed the NECC out of the Niño region. However, in the 1997 El Niño, a strong eastward current was observed during the main phase. The eastward current did not move into the Niño region during the 2015 El Niño period, except in the 2015 JJA. Within the Niño region, the westward current was dominant, and its position and intensity pushed the NECC out of the Niño region.

Figure 7 depicts the mean ZWS, OSC, and SST anomalies from the central to the eastern Pacific between 180° W and 80° W. The wind intensity in the 2015 El Niño was stronger than that in the BMS, but the peak positions were comparable. During El Niño evolution, the westward wind dominated in the central and eastern Pacific between latitudes 20° S and 20° N, although its intensity varied during each phase. At latitude 7° N in 2015 SON, there was only a weak eastward wind (main).

Fig. 7
figure 7

East–west intensities for zonal wind stress (ZWS) and ocean surface current (OSC) for latitudes 20° S–20° N during the 2015 extreme El Niño in the tropical Pacific Ocean. The mean longitude is in the range of 160°–90° W. At each latitude, the mean ZWS and OSC intensity in the east–west direction is expressed in a 1-degree grid. West–east a ZWS and b OSC intensity in MAM (March, April, and May), JJA (June, July, and August), SON (September, October, and November), and DJF (December, January, and February)

Full size image

In decreasing order of eastward current intensity, the NECC was 2015 SON (peak 5° N), 2015 JJA (peak 5° N), 2014 JJA (peak 6° N), 2014 SON (peak 7° N), 2015 DJF (peak 6° N), 2014 DJF (peak 6° N), 2015 MAM (peak 5° N), and no eastward current in 2016 MAM. The SST anomaly peaked at the equator (main) and at latitudes 1° N (onset) and 5° S (dissipation). The peaks in currents and winds occurred at different latitudes, and the peaks in SST occurred at different times and latitudes, suggesting that these three factors had independent dynamics. This shows that the interaction between the ocean and atmosphere during El Niño is more complex than a simple cause-and-effect relationship and requires a detailed analysis of how each factor interacts during El Niño events. During the 1997 El Niño period, the ZWS differed from the BMS with latitude. However, as with 2015 El Niño, there was a weak eastward wind at 1997 SON latitude 4° N, but the westward wind continued to predominate. The peak of the positive SST anomaly was located at the equator (0°). Similar to the 2015 El Niño, the peak positions of the ZWS, NECC, and SST anomalies differed.

Figure 8 shows the changes in intensity and location of the ZWS, OSC, and SST anomalies with longitude in the Niño region. There was an eastward wind from the eastern Pacific to the dateline location in 2015 MAM and 2015 JJA, and it extended to longitude 170° W in 2015 SON, where a westward wind dominated after that. The westward wind weakened after longitude 140° W. There was a weak eastward current from the eastern to the central Pacific in 2015 MAM, 2015 JJA, and 2015 SON, but thereafter, with the exception of 2015 MAM, the westward current was predominant. During the main phase, the positive SST anomaly reached its maximum value in the central Pacific and became saturated in the eastern Pacific. The 1997 ZWS data differed significantly between the periods. The eastward wind extended to the dateline in 1997 MAM, up to 170° W in 1997 JJA, and up to longitude 150° W in 1997 SON, but the westward wind was dominant. In 1997 MAM, 1997 JJA, and 1997 SON, the OSC had a weak eastward current from the western to the eastern Pacific. The positive SST anomaly reached its maximum value at longitude 100° W during the main phase, decreased toward the central Pacific, and became negative as it crossed the dateline. The ZWS of the 2015 El Niño was weaker than that of the 1997 El Niño, and the expansion of the 2015 El Niño was limited to the western Pacific. The OSC intensity and location of the two extreme events extended to the eastern Pacific Ocean, with the 1997 El Niño being more intense and reaching further east than the 2015 El Niño. In the Niño region, it was validated that the 1997 El Niño was more effective for the east–west flow of wind and currents than the 2015 El Niño, in which the CP and EP El Niño were mixed.

Fig. 8
figure 8

East–west intensities for zonal wind stress (ZWS) and ocean surface current (OSC) for longitudes 160° E–90° W during the extreme El Niño events in the equatorial Pacific Ocean. The mean latitude is in the range of 5° S–5° N. At each longitude, the mean ZWS and OSC intensity in the east–west direction is expressed in a 1-degree grid. The graphs depict the 2015 a ZWS intensity, b OSC intensity, and c sea surface temperature (SST) anomaly, as well as the 1997, d ZWS intensity, e OSC intensity, and f SST anomaly

Full size image

4 Summary and Discussion

The development of high-quality satellite products has revealed a mechanism for El Niño formation and development that has remain unknown since Bjerkens’ theory was proposed in 1960. In this theory, the development of El Niño by wind forcing (weakening of the trade winds, westerly wind bursts, and Kelvin waves) remains an important factor in the interaction between the atmosphere and the ocean and is the subject of ongoing research. Most El Niño events since the 2000s have been CP El Niños. However, despite the vast amount of observational data and improved model performance, the ability of general circulation models to accurately distinguish CP El Niños is still limited and may not fully reflect their complexity. Although various hypotheses have been proposed, the thermal origin of the sudden SST warming in the eastern Pacific remains unclear.

During an extreme El Niño event, the direction and intensity of the surface flow of the NECC is stronger than the background mean state, but strong eastward flow within the Niño region is rarely observed. In the 2015 El Niño, which had a long onset and duration, the NECC moved outside the Niño region, while westward flow was dominant. In the 1997 El Niño, which began relatively quickly, the NECC moved into the El Niño region only during the main phase of the El Niño. During the 1997 El Niño (EP El Niño), which developed rapidly and can be explained by the classical mechanism, the NECC shifted inward into the Niño region. Thus, the NECC is influenced by the speed and type of El Niño development.

Figures S1, S2, and S3 show the 3-month mean spatiotemporal pattern of a moderate El Niño (CP El Niño). Previous studies have shown that the zonal effect of CP El Niño events is an important factor in the development of El Niño. During the three El Niño periods, both the NSW and the NECC flowed out of the Niño region. During the main phase, the NECC was strong in the 2002 and 2009 El Niños but weak in the 1994 El Niño. The changes in ZWS and NECC strength during the main phase occurred simultaneously, suggesting that there are additional factors that alter ocean currents and winds. For moderate CP El Niño events, as for extreme CP El Niño events, a continuous flow of the NECC extending into the eastern Pacific is rarely observed in El Niño monitoring area.

A limitation of this study is that the equatorial climate system is very complex during El Niño events, so it is difficult to assess the overall mechanisms using only the indicators NSW, NECC, or SST peak location. The NECC can contribute to the transport of warm water from the western to the eastern Pacific over a large area above the equator. The differences between the NSW, NECC, and SST peaks during extreme El Niño events in this study are indicative of the complexity of ocean–atmosphere interactions associated with El Niño events. These differences may raise questions about the accuracy of El Niño prediction models. If models do not fully reflect the influence of ocean currents such as the NECC, then their predictions of the onset and progression of El Niño events may be less accurate. From the study of the spatiotemporal variability of the NSW, NECC during the 1997 and 2015 El Niño using satellite observations can be interpreted as an important signal that calls for a reassessment of existing El Niño theories and improvements in models. Furthermore, the complexity of the El Niño evolution with the dominance of the CP El Niño since the 2000s requires additional factors to be considered based on existing theories. Since SSTs peak at the equator despite the continuous flow of the NECC during El Niño, we propose the indirect effects of the horizontal movement of the NECC and the importance of the Equatorial Undercurrent in transporting heat within the ocean contributors to the equatorial temperature increase.

Figure 9a represents the climatology for the period from 1993 to 2019, where the heat content is evenly distributed across various temperature ranges. During the 2015 El Niño period, a significant increase in heat content at temperatures above 28 °C was observed, with the 30 °C mark also showing a notable rise. This indicates an increase in seawater temperatures during the El Niño event, suggesting that water at higher temperatures can contain more heat. The augmented heat content at these elevated temperatures during El Niño events reflects the substantial thermal impact of this climatic phenomenon on the ocean's temperature distribution. In the classical model of El Niño, the weakened trade wind allows the eastward ocean surface current, which transports the warm water normally in the western Pacific to the east. Without the additional heat, regardless of the physical distribution movement of the water at each temperature, the volume distribution of the heat should be the same within the seasonal variation. Therefore, we propose a hypothesis that the additional heat transported below the surface contributes more to the development of El Niño than the heat transported by the winds or the NECC at the surface. In future work, we aim to investigate the source of this excess heat and its impact on El Niño of the transport of the Equatorial Undercurrent, which moves rapidly from the western Pacific to the eastern Pacific in the subsurface. Since the Equatorial Undercurrent is expected to play a significant role in the redistribution of internal heat in the equatorial Pacific, we propose an international program of intensive in-situ monitoring of this current in addition to the current ARGO project.

Fig. 9
figure 9

Monthly distribution of ocean heat content and temperature above 20 °C from latitudes 5° S to 5° N latitude and longitudes 140° E to 70° W during the 2015 El Niño Event. Each bar corresponds to a specific month from February to December and illustrates the variation in heat content across a temperature range of 20 °C to 30 °C. a Climatology based on the period from 1993 to 2019, b 2015 El Niño, c 2015 anomaly, calculated as the difference between the 2015 and the 27-year climatology

Full size image