Originally, El Niño was the name of a southward-flowing current off the coast of Peru which affected fisheries . However, the studies reported in and connected it to the very large changes in surface temperature and thermocline depth observed in the nearby Equatorial Pacific Cold Pool (also referred to as “Cold Pool” in the following).

was the first to make the further connection between this oceanographic El Niño and the strength of the NECC in the western Pacific. He also suggested that the NECC might trigger an El Niño by transporting West Pacific Warm Pool water to the east.

then showed that the NECC transport increased by between 25 % and 50 % while the strong 1982–1983 El Niño was developing. They also found that the total transport of warm water (> 28 ∘C) by the NECC was consistent with their estimate of the loss of warm water from the West Pacific Warm Pool.

Although Wyrtki's hypothesis was not developed further, his work complemented the study by which showed a correlation between water temperatures in the equatorial Pacific and the Southern Oscillation . This led to further uses of the term “El Niño” to describe both its worldwide influence and the air–sea interaction (or mechanism) behind the event .

also argued that the east–west temperature gradient along the Equator was involved, and this is usually taken to imply that the temperature of the Equatorial Pacific Cold Pool is a key part of the El Niño mechanism. However, doubts this, because the east Pacific winds are only weakly affected by an El Niño. If he is correct, temperature changes in the Cold Pool are a consequence of and not a cause of El Niños.

Further studies showed that Bjerknes' correlation was strongest when using ocean temperatures from the central Pacific. As a result, most modern studies of the Southern Oscillation use the Nino 3.4 index to measure the strength of an El Niño . This is based on the mean sea surface temperature anomaly in the central Pacific between 170 and 120∘ W, and between 5∘ S and 5∘ N.

As well as affecting surface winds in the western and central Pacific, the atmospheric El Niño requires the centre of deep atmospheric convection to move from the maritime continent and the western Pacific into the central Pacific. , following and , confirmed that this requires sea surface temperatures of 28 ∘C and above during wet seasons and even higher temperatures during dry periods.

In the central Pacific, such temperatures are found on the Equator during an El Niño, but eastern Pacific equatorial temperatures never reach this value Fig. 2. Instead, the eastward limit of warm water on the Equator is closely correlated with the Southern Oscillation Index .

In the eastern Pacific, temperatures sufficient to trigger deep atmospheric convection are often found further north in the ocean off Central America and along the path of the NECC close to the ITCZ.

Meteorological studies also showed that the start of an El Niño was often connected with one or more Madden–Julian Oscillations (MJOs) . In his review, describes MJOs as individual events with lifetimes of between 30 and 100 d, and with the property that there is never more than one major event in existence at a time. Each one consists of an area, spanning the Equator, in which there is an increased density of short-lived convective cloud groups.

The events progress eastwards at speeds around 5 ms-1, starting in the western Indian Ocean and dying out after passing the western Pacific. At sea level, they are associated with inflowing winds from the east and west, but in the Pacific the westerly winds may extend past the centre of the cloud groups.

Following Wyrtki and Bjerknes, most theoretical studies of the ocean's contribution to El Niño have focused on the Cold Pool. This is a narrow band of low sea surface temperatures, centred on the Equator in the eastern Pacific. During a normal year, this results from the wind-generated upwelling of cold water from depths of around 200 m, but during an El Niño the surface thermocline becomes thicker and less cold water is upwelled.

Early theoretical studies, using analytic models or numerical models with a few layers , showed that the changes in the Cold Pool could be generated by wind variations in the western and central Pacific. This occurred via the propagation of baroclinic Kelvin waves in the equatorial wave guide. The studies also showed that only winds close to the Equator were responsible and that off-equatorial currents, such as the NECC, were not connected with the El Niño changes to the Cold Pool.

With the development of computer power, it was eventually possible to run ocean models with much more detailed physics and with enough horizontal and vertical resolution to resolve the key currents and structure of the ocean . The versions developed by , and have been widely used in stand-alone mode and, coupled to an atmospheric model, for studies of climate.

Unfortunately, the NECC can be poorly represented in such models . For ocean-only models, found that the poor results were largely due to errors in the atmospheric forcing datasets.

concentrated on the forcing of the NECC and showed that a weak NECC resulted from the atmospheric reanalysis datasets underestimating the curl of the wind field near the latitudes of the ITCZ and overestimating the strength of easterly winds on the Equator.

The study also found that, of the datasets studied, the ECMWF reanalysis was best at generating a realistic NECC. This may be because the ECMWF model was using a spherical harmonic expansion which requires less smoothing than normal finite difference schemes to overcome non-linear instabilities.

There is a related problem with coupled ocean–atmosphere models which fail to generate realistic El Niños . This again may be due to the resolution of the atmospheric component being insufficient to accurately represent features such as the ITCZ.

Further studies using observational data have indicated that there are two types of El Niño . The most common are the central Pacific (CP) or Modoki events, in which the centre of atmospheric convection moves to the central Pacific. More rarely, the centre of convection moves even further east to give the strong eastern Pacific (EP) El Niños . These behave like the classic El Niños , which proposed were triggered by the NECC.

Although neglected for many years, Wyrtki's hypothesis is supported by more recent studies using altimeter data .

These confirm the increased transport by the NECC during EP El Niños, showing that increases in the transport of the NECC in the western Pacific occur around 3 months before the corresponding change in the Nino 3.4 index. Also in the western Pacific, show that there is a delay of between 40 and 90 d between changes in the depth of the North Equatorial Trough, north of the NECC, and changes in sea level near the Equator.

The results thus leave open the possibility that the increased transport has a role in triggering subsequent EP El Niños, rather than being part of the ocean's response to the event. The studies also indicate similar changes in the NECC during CP El Niños, indicating that both types of El Niño may be affected by the NECC in the western Pacific.

Further support for the hypothesis comes from , who used results from a high-resolution global version of the NEMO ocean model to study the strong EP El Niños of 1982–1983 and 1997–1998.

For this study, NEMO was forced by the DRAKKAR dataset , based on the ECMWF reanalysis fields, and produced sea level and sea surface temperature fields in the equatorial Pacific in close agreement with satellite observations . The results indicate that the modified ECMWF reanalysis fields accurately represented the wind stress curl and so generated a realistic NECC.

Unlike previous studies, concentrated on water with temperatures above 28 ∘C, which are sufficient to trigger deep atmospheric convection. The study showed that, while the strong El Niños were developing, the NECC carried warm water into the eastern Pacific far in advance of water with similar temperatures on the Equator.

The study also helped explain why the NECC does not carry Warm Pool water into the eastern Pacific every year, by showing that in normal years tropical instability eddies mix in cold upwelled water from near the Equator and that this rapidly reduces temperatures in the NECC See:. In El Niño years, the eddies are weaker so their effect on the NECC is reduced.

It also found that the start time of strong El Niño events resulted in warm water reaching the central and eastern Pacific at a time of year when the annual Rossby wave meant that the NECC was strongest there. This resulted in the warm water arriving off South America around the end of the year, thus providing a physical mechanism to explain the name originally given to classic EP El Niños.

Finally, the study showed that both the strong 1982–1983 and 1997–1998 El Niño events started when the depth of the western section of the North Equatorial Trough was unusually low. As a result, the NECC transport, and the flux of Warm Pool water eastwards, were both much stronger than usual.

suggested that the low sea level in the trough was due to an unusually strong annual Rossby wave prior to each El Niño. However, further investigation showed that this is unlikely, as in each case the wave is not unusually large as it crosses the central Pacific.

Another possible explanation is that the stratification in the Pacific, prior to an El Niño, is sufficient to focus the Rossby wave and so cause lower-than-normal sea levels. Alternatively, the low sea level may not be due to the Rossby wave but instead due to some other process associated with the wind field. This could be a local feature occurring at or just before the period when sea level drops. It could also be a response to forcing elsewhere in the ocean, which later propagates into the western Pacific.

To help clarify the cause of the drop in sea level, this paper starts by reporting on series of short ocean model runs which focus on changes in the western Pacific prior to the 1982–1983 El Niño. As in , this allows the development of hypotheses, which are then tested by treating them as predictions for the strong El Niños of 1997–1998 and 2015–2016 for which better observations are available.

Occam, the global ocean model used for the tests, is an updated version of the OCCAM project ocean model . In each run, the model is initialised from one of the archive datasets from the original run of the NEMO model and forced with ocean surface stresses calculated during the same run.

The results imply that the low sea levels, at the western end of the North Equatorial Trough, are due to local Ekman pumping acting over a period of many months.

This hypothesis is tested against the data from the original run of the NEMO 1/12∘ global ocean model, initially during the growth of the 1982–1983 El Niño and then for the strong EP El Niños of 1997–1998 and 2015–2016. A similar comparison covering the whole period from 1976 to 2015 is also carried out.

In the first part of the paper, Sect. 2 describes the model being used and how the model fields and forcing were converted from the 1/12∘ NEMO grid to the 1/4∘ Occam grid.

Section 3 reports on tests carried out to validate the lower-resolution model. In these, the model was started from the NEMO archive datasets from early January 1981 and 1982 and then run for a year using the surface wind stresses from the same year.

The following sections report tests with different wind forcing and initial conditions. This includes runs where the model was again started from early January 1981 and 1982 but with the wind forcing from the opposite year. The results show that the main changes in sea level depend primarily on the wind field and, to first order, are independent of the stratification at the start of the run.

The tests are repeated but starting later in the year when the annual Rossby wave has developed and is starting to cross the central Pacific. This is to check whether it is responsible for the sea level drop in the western Pacific. The results are similar to the earlier test, indicating that the wind-generated Rossby wave is not responsible.

A final test is designed to investigate whether the winds causing the drop in sea level are local to the western equatorial Pacific or propagate in from other parts of the ocean. The results indicate that it is the local winds that are responsible.

The most likely way in that the local winds could affect sea level in the North Equatorial Trough is through Ekman pumping and the resulting rise in density surfaces within the ocean. In the second part of the paper, this hypothesis is tested.

Section 7 compares the expected uplift due to Ekman pumping during 1982 with the rise in density surface in the NEMO 1/12∘ model. The results are promising so in Sect. 8, similar tests are also carried out for the 1997–1998 and 2015–2016 EP El Niños.

Finally, a similar survey of the whole period 1976 to 2015 is carried out to see if strong Ekman pumping in the western section of the North Equatorial Trough during the early part of the year is common and whether it might also be a sufficient condition for the development strong EP El Niños.

The model runs reported here were initialised by averaging the archived data from the high-resolution NEMO model onto the Occam grid. Variable values within each Occam ocean cell were calculated by averaging over the intersection of the Occam and NEMO cells. In the case of vector quantities, vectors were rotated to the Occam grid before averaging.

The runs were also carried out with zero flux of heat and salt across the ocean surface. Each run is for only a few months and, although the surface temperatures and salinities are affected, this approximation allows the analysis to concentrate on the effects of wind stress, and of advection and diffusion within the ocean.

Each NEMO archive dataset contains the ocean and forcing fields averaged over the previous 5 d. Thus, when initialising the Occam model and when specifying the wind forcing, the time of each NEMO dataset is set to the central time of the averaging period. However, when specifying particular NEMO archive files, the original archive date is used.

The use of 5 d averages filters out higher-frequency components of the wind field and also means that the ocean state used to initialise the Occam model may be unusually smooth. The resulting lack of high-frequency oscillations may affect vertical mixing in the ocean but should otherwise have little effect on lower-frequency variations in sea level and current velocity, which are the focus of this study.

In Fig. , the top two pairs of figures show sea level from the NEMO and Occam models on 6 June and 9 September 1981. 6 June is chosen because the corresponding date in 1982 can be used to check the early deepening of the North Equatorial Trough in the western Pacific that occurs prior to the 1982–1983 El Niño. Similarly, 9 September is chosen as a check on how well Occam reproduces the increased depth of the trough, and corresponding increased strength of the NECC, across the entire Pacific, during the autumn months.

At the largest scales, the Occam model is in close agreement with NEMO, the main differences arising from slightly more extreme maxima and maxima. Thus, in September, maximum sea level in the South Pacific Gyre is slightly higher in Occam. Similarly, the minimum sea levels in the Cold Pool on the Equator and within the North Equatorial Trough are slightly lower.

This paper is primarily concerned with the behaviour of the NECC and the associated sea levels in the North Equatorial Trough. Here, Occam shows a similar westward propagation of the trough during the year and a similar growth of the ridge on the southern side of the NECC. The September figures also show a similar development in Occam of the short wavelength features on the ridge, associated with the growth of tropical instability eddies.

In June, both NEMO and Occam show that the trough is particularly weak in the central Pacific, resulting in an almost non-existent NECC. However, in September, with a greater contrast between ridge and trough, both models show a stronger NECC.

At the western ends of the trough and ridge, Occam shows a similar development of the high- and low-sea-level regions associated with both the Halmahera Eddy and Mindanao Eddy and the initial meanders of the NECC. In September, these decay in amplitude, in contrast to the growth of the trough and ridge features seen in the central and eastern Pacific.

The development of sea level during the year at 6∘ N in both NEMO and Occam is shown in Fig. a and c.

The average slope in sea level in Occam matches that of NEMO, and in the central and eastern Pacific, Occam shows a similar development of the ridges due to the growth and propagation of tropical instability eddies. The figure also shows the low sea levels due to the propagation on the annual Rossby wave at 6∘ N. This starts in the eastern Pacific early in the year and reaches the dateline in July and August.

The corresponding sea levels on the Equator are shown in Fig. a and c. Again, the overall east–west slope is similar and Occam reproduces the eastward-propagating equatorial waves that can be generated by westerly wind bursts in the western equatorial Pacific. There are differences between the two model runs but in both models sea level is a maximum either close to the western boundary or near 150∘ E.

Sea surface temperatures, in June and September 1981, are shown in the top pairs of panels in Fig. . As Occam includes no surface warming or cooling, it is expected to show no seasonal changes. This is seen to be the case, with Occam showing warmer temperatures in the Southern Hemisphere in both June and September and cooler temperatures in the north.

The differences are largest off Central America, where Occam fails to reproduce the East Pacific Warm Pool, and in the western section of the South Pacific, where there is little reduction of Occam temperatures between June and September. The regions of cold water, with temperatures at or below 20 ∘C, also behave very differently in Occam and NEMO.

Along the line of the NECC at 6∘ N, Occam fails to show the warming during the first half of the year, but by September it does show a narrow band of warm water being advected into the eastern Pacific by the NECC. It also shows it being eroded by tropical instability eddies.

Corresponding sets of panels for 1982, during the development of the 1982–1983 El Niño, are shown in Figs. –.

In Fig. , Occam shows, in agreement with NEMO, a much deeper North Equatorial Trough in early June together with a well-developed NECC. This is emphasised further in September with a deep trough in the west which extends almost as far as the Mindanao Eddy.

The Hovmöller diagrams of Fig.  show that in 1982 the annual Rossby wave, in both Occam and NEMO, links with the region of low sea level that develops at the western end of the trough. argued that it was this drop which generated the increase the NECC strength which triggered the development of a classic oceanographic El Niño.

At the Equator (Fig. ), Occam shows, in agreement with NEMO, the movement of the maximum sea level away from the western boundary to a region near 200∘ E. In Figs.  and , this is seen to correspond to the region of high sea level on the Equator where the warmest equatorial temperatures are also found.

The results show that the Occam model is capturing the key features in the sea level field which affect the strength of the NECC, the strength of tropical instability eddies and processes occurring on the Equator. It is less effective at capturing changes to sea surface temperature field. This may partly be due to the lack of surface heat fluxes in this version of the model. It may also be partly due to the much reduced horizontal resolution of the model.

However, bearing these strengths and weaknesses into account, there is no reason why it cannot be used to study the effect of the winds and the initial state of the ocean on the development of the 1982–1983 El Niño, as is done here.

In this test, the model was initialised from the NEMO archive dataset dated 5 January 1982 but then forced with winds from 1981. Figure e shows how sea level developed at 6∘ N. During the first few months of the year, the annual Rossby wave develops and propagates westwards as normal but in the middle of the ocean the amplitude declines, so by the time the wave reaches 180∘ E the signal is weak.

On the Equator (Fig. e), sea level starts high in the western Pacific, as in the earlier run with 1982 winds forcing the 1982 ocean (Fig. d), but as the year develops it stays in the west and there is no movement into the central Pacific.

Figures c and c show the sea level and temperature fields for June and September. Previous runs with 1982 winds show that by June the depth of the trough is starting to increase more than in a normal year but here this is not happening and instead the response is closer to that of NEMO in 1981. Similarly, in September, the trough fails to develop further and the response again lies closer to that of NEMO in 1981.

The surface temperature field is also closer to the NEMO results from 1981, with the bulk of the Warm Pool water remaining in the west. However, there is a thin core of warmer NECC water extending into the eastern Pacific, which is not present in the run starting in January 1981 with 1981 winds. This indicates that there was some change in the structure of the ocean between January 1981 and 1982 which aided the transport of warm water eastwards by the NECC.

Despite this, the main conclusion from this run is that the ocean state at the start of 1982 was not sufficient to trigger an El Niño in the middle of the year. Although a reasonable annual Rossby wave was generated early in the year, this was not focused and did not generate or contribute to the lowering of sea level in the western Pacific.

In the second test, the ocean is initialised from the NEMO-archived dataset dated 5 January 1981 but then forced with 1982 winds. Figure f shows the sea level during the year at 6∘ N. The annual Rossby starts as before, but this time it continues past 180∘ E and links up with a region of low sea level in the western Pacific – as might be expected at the start of a strong El Niño.

On the Equator (Fig. f), the ocean again starts with high sea levels in the west. Mid-year, these move into the central Pacific – again as might be expected at the start of a strong El Niño.

Figures c and c show that in the middle of the year the distribution of sea surface temperature is similar to the other Occam runs, but sea level shows a well-developed North Equatorial Trough with a strong NECC developing on its southern slope.

By September, the trough has developed further and the temperature plot shows warm surface water being advected rapidly into the eastern Pacific. The area of warm water involved is not as large as in the original NEMO run in 1982, but as with Fig. b, this is probably a result of setting the surface flux of heat to zero in these test runs.

One conclusion from this test is that the ocean will produce an El Niño-like response when forced by the winds from an El Niño year. This has been reported before but based primarily on the large sea level and temperature anomalies that develop in the Equatorial Pacific Cold Pool towards the end of the first year of a classic oceanographic El Niño.

The present study is different in that it shows the winds having a significant effect during the first half of the year, well before any changes in the Cold Pool region are noticeable. The results also show that by the middle of the year the 1982 winds have produced significant changes in the western section of the North Equatorial Trough. What is not clear is whether the increased depth of the trough is due to the mean winds during the first part of the year or due to one or more isolated events.

It is also not clear whether it is due to a local change in the winds along the line of the trough, or whether the key wind forcing occurs elsewhere and the signal propagates into the region lowering the level of the trough. The remaining model runs reported here attempt to obtain a clearer answer to these questions.

Figure  shows the changes in sea level from the different runs, averaged over the region 140 to 170∘ E and 5 to 7∘ N, as outlined in Fig. . The Occam runs start with a slight offset from the corresponding runs of the NEMO model, due to the initial adjustment to the new coast and topography.

In this figure, colour is used to distinguish the different winds forcing the ocean: blue for 1981 and red for 1982. Even with the different starting conditions, the runs soon split into two distinct groups: sea levels with the 1981 winds staying roughly constant, whereas the 1982 winds generate a 20 cm drop between March and August. As pressures in the deep ocean remain roughly constant, this drop in sea level must correspond to a significant rise in the density surfaces within the ocean.

To quantify the differences between the groups, sea levels from the end of April and August are given in Table , split into runs forced by 1981 winds and those forced either by 1982 winds or by the combined wind fields. A t test shows that the probability of both sets belonging to the same group at the end of April is around 0.5, as might be expected. By the end of August the probability drops to around 0.00002. A test using the sea level change over the period gives a similar value.

Periods of positive El Niño index are often thought to be triggered by periods of westerly winds in the western Pacific associated with the Madden–Julian Oscillation.

One possible mechanism, connecting the two, requires westerly wind bursts to generate equatorial Kelvin waves which eventually result in a warmer Equatorial Pacific Cold Pool. However, there are doubts about this mechanism because the Equatorial Pacific Cold Pool region is never warm enough to trigger deep atmospheric convection.

However, the westerly winds can also reverse the direction of the Equatorial Current in the western Pacific, and so they transport Warm Pool water into the central Pacific. The temperature of the advected water is sufficient to trigger deep atmospheric convection and its effect on the position of the warm water front in the central Pacific may explain much of the normal year-to-year variation in the El Niño index.

These mechanisms focus on waves and currents near the Equator, but, as discussed earlier, observations show that water warm enough to cause deep atmospheric convection is carried even further eastwards by the North Equatorial Counter Current.

investigated the cause of the strong 1982–1983 and 1997–1998 El Niños and argued they were due to the NECC advecting a larger than normal volume of Warm Pool water into the central and eastern Pacific as a result of low sea levels in the North Equatorial Trough early in the year.

The study reported in this paper shows that these low sea levels are the result of local winds upwelling water into the surface layer. Although it is affected by the topography of the adjacent ocean continent, the northward extent of MJOs and other similar wind features in the western Pacific is close to the scale of the equatorial Rossby radius in the atmosphere and so includes the latitudes of the NECC.

As a result, it is very possible that the mismatch between the equatorial Rossby radii in the ocean and atmosphere means that MJO westerlies may trigger strong EP Pacific El Niños as a result of Ekman-driven upwelling, at the latitude of the North Equatorial Trough, in addition to their impact on currents near the Equator.