By Andy May
The Oceanic Niño Index or ONI is NOAA’s primarily indicator for monitoring the sea surface temperature (SST) anomaly in the critical Niño 3.4 region. It is a 3-month running mean of ERSST.v5 SST anomalies in the Niño 3.4 region, defined as 5°N-5°S and 120°W-170°W. Figure 1 shows the ONI as computed from the NOAA ERSST dataset. ERSST is a two-degree gridded dataset, so the region averaged for figure 1 is 6°N-6°S and 120°W-170°W.
Per convention a three-month moving average has been applied to the raw ONI data in figure 1. Sometimes you will see the ONI detrended, but the curve in figure 1 is not detrended and has an upward slope of one-half degree per century. The 3-month moving average has to exceed 0.5°C for five consecutive months to define an El Niño, so the chart is colored red above 0.5°C. The same is true for La Niña, but in reverse. The white area between -0.5 and +0.5 is ENSO neutral.
The current ENSO state, as of July 2025, is ENSO neutral, with an average ONI of about zero. NOAA prefers to use a base period for their ONI anomalies of 1991-2020, but we use 1961-1990 to be consistent with the other posts in this series and with HadCRUT5. There is a visual trend over the past 175 years, Niños are more common now and stronger than in previous years. Climate models have a very hard time duplicating ENSO over both short and long periods of time (IPCC, 2021, p. 115). The Niño 3.4 region is shown in figure 2 in red.
Strong Niños expel a lot of ocean heat into the atmosphere, and this warms Earth’s surface for a few years, but in the longer term, Niños signal future cooling, just as in the longer term, frequent Niñas signal warming or a stable global temperature. However, this is qualitative, the correlation between ENSO or ONI and HadCRUT5 is poor, as shown in figure 3.
As figure 3 shows, the correlation between ONI and HadCRUT5 is especially poor since 1990. Since then, they trend in different directions, the ONI trends down at 0.4°C per century and HadCRUT5 trends upward at 2.3°C per century.
On climate.gov, NOAA’s Michelle L’Heureux writes:
“ENSO is one of the most important climate phenomena on Earth due to its ability to change the global atmospheric circulation, which in turn, influences temperature and precipitation across the globe.”
Indeed, this is what nearly everyone believes. The current phase of ENSO does have a planetwide influence on weather, thus it is odd that the current measure of climate change, global mean surface temperature or GMST from HadCRUT5, does not trend with ENSO or even in the same direction over the past 35 years. This 35-year climatic period is when we have the best SST data.
History
The term El Niño has evolved in its meaning considerably over the years (Trenberth, 1997). It originally named a weak warm coastal current that runs southward along the coast of Ecuador around Christmas. However, it is now used to describe extreme warming events along the tropical western South American coast every ~2.5 to ~6 years (Rasmusson & Carpenter, 1982) and (Ghil, et al., 2002) that are associated with changes in Pacific equatorial currents and atmospheric circulation. Rasmussen & Carpenter give a detailed chronology of the study of the El Niño phenomenon in their 1982 paper, and they discuss the early work on this complex and important oscillation from before 1920, even then it was recognized as an oscillation with a global reach.
The first effort to standardize the definitions of El Niño and La Niña episodes was by Kevin Trenberth (Trenberth, 1997). He defined an El Niño event as when a 5-month running mean of sea surface temperature (SST) anomalies in the Niño 3.4 region (5°N-5°S, 120°-170°W) exceeded 0.4°C for 6 months or more. Since his original paper was published this definition has been modified as given above, such that the six months was reduced to five, the 0.4°C raised to 0.5°C, and the 3-month running mean rather than Trenberth’s 5-month running mean was used. The same definition applies to Niñas, but for a decrease in Niño 3.4 SST of 0.5°C.
Regardless of the definition used, the ENSO phase changes involve huge changes in the ocean and air circulation over the entire equatorial Pacific, and they affect the entire world. I found the NOAA illustration shown in figure 4 to be very helpful.
The Niño 3.4 region is roughly located on the figure, also see figure 2. The Niño 3.4 region is warm during Niños, cool during Niñas, and in between during neutral periods.
Discussion
The global mean surface temperature, the global mean ocean surface temperature and the AMO, in the 20th century, follow a 64-year period (Wyatt, et al., 2012a). ENSO and the ONI do not follow this pattern, the only statistically significant oscillation periods in the various measures of ENSO are from an analysis of SOI, Southern Oscillation Index, which is very similar to the ONI. The periods that are significant at the 99% level are 5.5 years and 2.4 years (Ghil, et al., 2002). Ghil’s analysis is illustrated in figure 5.
ENSO has a worldwide impact but does not appear to contain the global 64-year oscillation described by Marcia Wyatt or the ~60-year oscillation described by Nicola Scafetta and many other writers and researchers. In fact, the only oscillations in this series that contain a significant 60-70-year period component are the global mean surface temperature, the global mean sea surface temperature and the AMO (see here).
In the next post we will look at the global ocean temperature and reflect back on the oscillations discussed in this series.
Download the bibliography here.
Previous posts in this series:
Climate Oscillations 1: The Regression
Climate Oscillations 2: The Western Hemisphere Warm Pool (WHWP)
Climate Oscillations 3: Northern Hemisphere Sea Ice Area
Climate Oscillations 4: The Length of Day (LOD)
Climate Oscillations 6: Atlantic Meridional Model
Climate Oscillations 7: The Pacific mean SST
Climate Oscillations 8: The NPI and PDO
Climate Oscillations 9: Arctic & North Atlantic Oscillations
Climate Oscillations 10: Aleutian Low – Beaufort Sea Anticyclone (ALBSA)
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
That’s what my oscilloscope looks like when the V/cm control needs some contact cleaner.
“The global mean surface temperature, the global mean ocean surface temperature…”
Are complete fantasies.
I prefer to say they can be estimated, but with regard to climate or climate change they are meaningless.
And tolerably pointless. Satisfying to measurebators, but no-one else.
Explain nothing, predict nothing – but are a continuing source of enjoyment to those with nothing better to do.
I rather like NOAA’s naïveté, and complete disregard for reality and the laws of physics, when they apparently state –
Complete nonsense. Warm water floats on colder water, being less dense. And, unfortunately, water finds its own level with respect to the geoid, being a liquid. The sorts of things that GHE believers refuse to believe, being suckers for brightly coloured cartoon pictures delisting fantasies and fairytales.
Climate is the statistics of weather observations, as are all “oscillations”, “waves”, “cycles”, and all the rest.
Curiosities of no practical value.
Using any temperature record that does not have a verifiable repeatable method of measurements over its history is a waste of time for detailed analysis.
The only valid temperature data for the Nino34 region is the satellite data that has been calibrated against a consistent surface measurement. The one I know of is the Reynolds OI SST data set.
I did the attached Fourier analysis on that data and it had a distinct peak at 10.6 years. That will no doubt be tied to Jupiter’s orbit and its influence on the Sun and possibly Earth’s orbit as well.
I am not yet convinced that the Sun’s weird orbit is valid because reducing such a large mass to a point source may not be a valid method of determining its motion when it dominates the gravitational field. Based on the daily acceleration data for the Sun for this century I find a solution of a circular orbit around the barycentre more convincing than the dogs breakfast that NASA JPL arrives at. Irrespective, Jupiter plays a large role in the Sun’s orbit and that has a large influence on the solar radiation Earth experiences.
What causes the ENSO and 60 year climate cycle?
The best answer is I don’t know. We know the Neutral/La Nina part of ENSO is related to solar activity, but the El Nino part is seemingly random. See figure 2.4 here:
https://andymaypetrophysicist.com/2022/08/08/the-sun-climate-effect-the-winter-gatekeeper-hypothesis-ii-solar-activity-unexplained-ignored-effects-on-climate/
The 60-64-year cycle is probably related to Jupiter and Saturn in some way, but I don’t know how. Also the 60 or 64 year cycles are influenced by planetary motions, as well as volcanic eruptions and greenhouse gas emissions. It is likely that the true natural cycle is 60 years and modulation by volcanism and GHGs causes it to look like 64. More in the next post, also see Nicola Scafetta’s paper here:
https://www.mdpi.com/2073-4433/12/2/147
Andy,
Let me try to give you some suggestions:
ENSO is a natural oscillation between excessive ocean cooling and inadequate ocean cooling.
When one looks at the heat balance in the tropical oceans, the most important cooling mechanism, offsetting solar heating, is water evaporation.
At tropical ocean temperatures greater than 25C, the most important factor influencing the rate of water evaporation is the wind speed, and the most important factor influencing wind speed is the rate of water evaporation, i.e. a self-reinforcing feed-back loop.
In the warming phase, when the winds are low, there is not enough cooling by water evaporation to offset solar heating and the water temperature rises. As the temperature rises the rate of water evaporation increases, and at water temperatures higher than approx 28C the rate of water evaporation becomes high enough to begin to drive the winds, setting off the feed-back loop described above. With high wind speeds the rate of evaporative cooling becomes excessive and the system enters the cooling phase.
Because of the tremendous masses involved in winding up and unwinding the winds, the temperatures will always overshoot in either direction, thus the oscillation.
2. 64 year Ocean Cycle:
This is likely the time it takes to complete one circuit of the Global Conveyor Belt. One must remember that in any cyclical process the most important factor affecting the amplitude is usually the amplitude of the previous cycle. It is imprudent to assume that the effects of external forcings (for example variations in solar energy inputs that are absorbed deep into the oceans) decay in a time that is less than periodicity of this ocean cycle, and that there is no ‘memory’ associated with the cycle. Thus the amplitude of the 64 year ocean cycle is likely to be the accumulation of external forcings over previous cycles (i.e. many centuries) that have not yet fully decayed, plus the effects of the external forcings accumulated during the current cycle. The periodicity will be the time required to complete one circuit.
3. No correlation between ENSO and the 64 year ocean cycle.
Try using a 120 month (10 year) moving average of ENSO and you are sure to find the correlation.
The global conveyor belt, if it exists at all, takes much longer than 64 years.
As for the lack of correlation between ENSO and GMST, if you want more documentation see the references in this post, plus a few more in my next post which will be out in a few days.
Regarding the 64 year cycle.
Regarding the relationship between ENSO and ocean temperature cycles.
McLean, J.D., de Freitas, C.R. and Carter R.M. “Comment on the Influence of the Southern Oscillation on tropospheric temperature”; Journal of Geophysical Research, DOI:10.1029, 13 August, 2009.
Shakun, J. D., and J. Shaman (2009), Tropical origins of North and South Pacific decadal variability, Geophys. Res. Lett., 36, L19711, doi:10.1029/2009GL040313.
Tisdale, Bob; Regression Analyses Do Not Capture The Multiyear After-effects of Significant El Nino Events, Climate Observations, July 27, 2009.
More on this in the next post. You can find a statistical correlation between any two time series, particularly if they both trend up, that does not mean it is significant. As you will see in the next post, there is no 60-64-year statistically significant period in ENSO. The ultimate source is:
https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2000RG000092
But, also see Scafetta, 2010, Empirical evidence for a celestial origin of the climate oscillations and its implications.
Thanks, Andy. The state of El Nino, Neutral, or La Nina, is immediately obvious when we vacation on the coast of Chile a couple of times in December and January. In El Nino conditions large flocks of Pelicans, and other diving sea birds, along with many seals and sea lions, are eating fish near-shore. In La Nina only a few scattered birds and seals, eating only a few fish (Neutral is irregular and averages between the two effects). How great it is to have an Index that is easily and directly observable! Tipping Points?…..forget about it!
“However, this is qualitative, the correlation between ENSO or ONI and HadCRUT5 is poor”
That’s a weird take. The effects are obvious when you look at global temperature spikes and El Niños. There is a lag of around half a year which makes it clearer. But nobody is suggesting that ENSO causes long term warming – that’s why you need to look at de-trended data.
Here’s my graph showing a combination of CO2 and a lagged ONI value.
The R^2 is 0.93, compared with 0.91 for just CO2. And it gives a p-value for ONI of 5.95e-10.
Bellman,
You are seeing the effects of autocorrelation. R^2 is not very meaningful between two monotonically increasing time series. See here:
https://andymaypetrophysicist.com/2021/11/13/autocorrelation-in-co2-and-temperature-time-series/
Perform an autocorrelation exercise, then you will see what I mean. To a trained eye, the graph you made shows that the correlation is very poor in reality.
“You are seeing the effects of autocorrelation.”
I disagree. I’m not sure how autocorrelation explains the relationship between ENSO and temperature spikes.
“R^2 is not very meaningful between two monotonically increasing time series.”
Temperature is not monotonically increasing. If you mean CO2, I’m only using it here to illustrate the correlation between ENSO and temperatures when you detrend temperatures. I could have just as easily used a loess smothing.
Here is a more thorough look at CO2 and solar impacts regresses agains GMST.
https://andymaypetrophysicist.com/2023/11/25/climate-co2-and-the-sun/
The regression study also includes Nino 3.4, which was a minor addition to the regression, but it is interchangeable with the solar cycles, which I found interesting.
I was arguing about that a couple of years ago. I argued it’s just over fitting multiple sine waves. You need to show that you can predict temperatures outside the fitted period.