Sea surface temperature is only half of El Niño. The Southern Oscillation Index is the atmospheric half, and as of July 20 the 30-day value sits at -27.87. Here is what that number does, and does not, tell you about how strong 2026's event will get.
For six weeks the story of the 2026 El Niño has been told almost entirely in sea surface temperature. The models point toward an unusually strong event, with NOAA assigning an 81 percent chance of a very strong October to December peak. The Niño 3.4 region is warm and getting warmer. But sea surface temperature is only half of the El Niño Southern Oscillation, and it is the half that describes the ocean. The other half lives in the atmosphere, and its clearest single gauge is the Southern Oscillation Index. As of July 20, that index has gone deeply negative: the 30-day value sits at -27.87. This post is about what that number means, what it does not mean, and the one test that will tell us whether the atmosphere is now locked in step with the ocean.
The Southern Oscillation Index is a standardized measure of one thing: the difference in surface air pressure between Tahiti, out in the central Pacific, and Darwin, in northern Australia. Both the Bureau of Meteorology and Queensland's LongPaddock calculate versions of the Troup SOI from Tahiti and Darwin pressure data, but they use different climatological base periods. LongPaddock publishes a daily-updated series.
During El Niño, pressure tends to rise over the western Pacific near Darwin and fall over the central Pacific near Tahiti. That pushes the Tahiti-minus-Darwin difference negative. Physically, that pressure pattern is the fingerprint of a weakened Walker circulation and slackened trade winds, which is precisely the atmospheric response that reinforces a warm ocean rather than fighting it. The Bureau's own rule of thumb: sustained values below -7 indicate El Niño is active, and sustained values above +7 are typical of La Niña. The sign and the persistence are what matter, not any single day.
As of the July 20 update from LongPaddock, the 30-day SOI is -27.87 and the 90-day average is -19.97. The monthly figures have been tightening steadily all season: April came in at -9.88, May at -13.22, and June at -22.69. Each month has been more negative than the one before it.
NOAA is saying the same thing from the other side of the Pacific. The Climate Prediction Center's July 9 ENSO Diagnostic Discussion noted that the "traditional and equatorial Southern Oscillation indices were significantly negative," alongside a weekly Niño 3.4 value of +1.2°C and an 81 percent chance of a very strong El Niño by October to December. Put plainly: a 30-day reading of -27.87 is far below the Bureau's -7 El Niño rule-of-thumb threshold and represents a strongly negative atmospheric signal. On the atmospheric gauge, this is not a marginal event.
Last month the open question on this site was whether the atmosphere would actually couple to the exceptionally strong ocean warming, or lag behind it. It is a real distinction. A warm ocean sitting under an unresponsive atmosphere produces a smaller real-world footprint than the raw sea surface temperature number implies. That gap is part of why NOAA adopted the Relative Oceanic Niño Index (RONI), which subtracts the average tropical SST anomaly from Niño 3.4 and therefore better distinguishes ENSO-relative warming from broad tropical-ocean warming.
The SOI provides one of the clearest atmospheric indicators of coupling, because it measures the atmosphere itself. Read alongside the westerly wind anomalies, shifted convection and ocean warming reported by NOAA, a deeply negative index that keeps deepening strongly supports the conclusion that the atmosphere is now participating in the event rather than lagging it. That mutual reinforcement, the Bjerknes feedback, is what turns a warm patch of water into a full event with global reach. As of mid-July, the atmosphere is engaged.
The SOI is noisy, and the Bureau says so explicitly: single-day and even weekly values are dominated by ordinary weather, not climate, and should not be used on their own. The wild daily swings in early July, contributions near -45 on the 7th and 8th and around -40 on the 15th, are partly day-to-day weather and passing tropical pulses, not pure ENSO signal. That is exactly why the 30-day and 90-day averages are the figures worth quoting.
Base periods differ too. LongPaddock's primary SOI product uses an 1887 to 1989 baseline, the Bureau's own Troup SOI uses a 1933 to 1992 baseline, and NOAA uses a different standardized method entirely, so the absolute numbers will not match source to source. Direction and persistence travel across methods; the exact decimal does not. And a 30-day window is not a season. El Niño is phase-locked to the calendar and typically peaks between November and January. A deeply negative SOI in July tells you the atmosphere is engaged now. It does not fix the magnitude of a peak that is still five months out.
Four concrete things will settle whether this signal holds:
The sea surface temperature forecast will keep making the headlines, and fairly so. The SOI is the quieter number that tells you whether the atmosphere is actually along for the ride. Right now, it is. You can follow it, and the other indices, on our teleconnections dashboard.