In the week centered July 22, Niño 1+2 reached +3.0°C and Niño 3 reached +1.8°C, while the benchmark Niño 3.4 held at +1.4°C for a second straight week and Niño 4 slipped to +0.3°C, its lowest since April. The east-west contrast is widening. Here is what the shape of an El Niño tells you, what this particular shape does not yet settle, and the atmospheric and subsurface evidence that belongs alongside it.
Almost all of the attention on this event has gone to a single number: a forecast peak of 3.6°C. That figure comes from Zeke Hausfather's July synthesis of 667 ensemble members across 14 seasonal forecast systems, initialized in July 2026, and it is the multi-model median of each member's peak monthly detrended Niño 3.4 anomaly. Note that "detrended" is not the same thing as NOAA's relative index: the RONI subtracts the tropical-mean SST anomaly from Niño 3.4 and then rescales the variance to match the original index, which is a more specific operation than removing a trend. In RONI terms Hausfather's own ensemble gives a lower record probability, 77 percent versus 91 percent.
Either way, that is a forecast of magnitude. It says nothing about structure. And over the past six weeks the structure has been doing something worth watching, visible in the weekly CPC data and almost entirely absent from the coverage.
ENSO is monitored through four regions strung along the equatorial Pacific. Niño 4 sits furthest west (5N-5S, 160E-150W) and normally has less variance than the others. Niño 3.4 straddles the central Pacific (5N-5S, 170W-120W) and is the benchmark, adopted because the key coupled ocean-atmosphere interactions sit there rather than further east. Niño 3 covers the eastern basin (5N-5S, 150W-90W). Niño 1+2 (0-10S, 90W-80W) is the smallest and easternmost, hugging the coasts of Ecuador and Peru, and it is where El Niño was named in the first place.
Reported together, those numbers describe not just how strong an event is but where its center of gravity sits. That is the part being missed right now.
| Week centered | 1+2 | 3 | 3.4 | 4 | 1+2 minus 3.4 | 1+2 minus 4 |
|---|---|---|---|---|---|---|
| 10 June | +2.1 | +1.0 | +0.9 | +0.8 | 1.2 | 1.3 |
| 17 June | +2.4 | +1.3 | +1.1 | +0.8 | 1.3 | 1.6 |
| 24 June | +2.5 | +1.5 | +1.2 | +0.6 | 1.3 | 1.9 |
| 1 July | +2.7 | +1.5 | +1.2 | +0.5 | 1.5 | 2.2 |
| 8 July | +2.6 | +1.6 | +1.3 | +0.5 | 1.3 | 2.1 |
| 15 July | +2.8 | +1.6 | +1.4 | +0.4 | 1.4 | 2.4 |
| 22 July | +3.0 | +1.8 | +1.4 | +0.3 | 1.6 | 2.7 |
Over those six weeks Niño 1+2 gained 0.9°C and Niño 3 gained 0.8°C, while Niño 3.4 gained 0.5°C and Niño 4 lost 0.5°C. Niño 4 has declined from +0.8°C in mid-June to +0.3°C, with four weekly drops and two unchanged readings. The eastern boxes warmed nearly twice as much as the benchmark, and seven weekly readings are not a stable rate estimate, so treat that as a direction rather than a velocity.
The inclusion of Niño 3 matters. If the signal appeared only in Niño 1+2, the tiny and most volatile box, it would be easy to dismiss as noise. Niño 3 is a large region and it moved almost as much, which means the eastward emphasis is not an artifact of one twitchy index.
There is a formal index for a closely related east-west contrast. Trenberth and Stepaniak proposed the Trans-Niño Index in 2001 to capture the gradient between the eastern and central Pacific, defined as the difference in normalized anomalies between Niño 1+2 and Niño 4, smoothed with a five-month running mean. The last column above is an unnormalized weekly difference between those same two regions, which is the closer analogue of the two contrasts shown, but it is still not the TNI and should not be quoted as one.
SST boxes on their own would be a thin basis for any claim about an event's character. Fortunately there is more. NOAA's July 9 diagnostic discussion also reports low-level westerly wind anomalies and upper-level easterly anomalies over the western and central equatorial Pacific, significantly negative traditional and equatorial Southern Oscillation indices, enhanced convection over the central and east-central equatorial Pacific with suppressed convection over Indonesia, and a recent downwelling Kelvin wave that deepened the thermocline and raised temperatures in the eastern Pacific.
Those atmospheric and subsurface signals confirm a strengthening coupled El Niño, which is the important part. They do not yet settle where the mature event's center of action will lie. Proper classification looks at the full SST anomaly field, the longitude of maximum convection and atmospheric heating, low-level wind anomalies, thermocline depth and subsurface heat, outgoing longwave radiation patterns, and seasonal rather than weekly structure. This post is reading one slice of that.
Researchers commonly distinguish eastern-Pacific and central-Pacific expressions of El Niño. The traditional, or canonical, pattern is generally east-weighted and sets up a single anomalous Walker circulation cell across the basin. A central Pacific event, often called El Niño Modoki, puts the warm anomaly near the dateline with cooler water flanking it on both sides, producing two anomalous cells with rising motion over the central Pacific, wetter conditions in the middle and drier conditions to the east and west. Different classification methods do not always place individual events in exactly the same category, and whether Modoki is a genuinely separate phenomenon or one end of a continuum remains an open question in the literature.
Two benchmark extreme events, 1982-83 and 1997-98, developed strongly east-weighted structures. The 2015-16 event reached comparable Niño 3.4 magnitude but was substantially warmer farther west, which illustrates exactly why a single strength ranking can conceal important structural differences. "Strongest since" tells you less than most headlines imply.
The practical stakes are highest for South America. Peru's ENFEN commission defines its own coastal El Niño on the eastern Pacific specifically, because that region tracks coastal rainfall and fishery impacts more closely than Niño 3.4 does. A +3.0°C weekly Niño 1+2 anomaly is highly relevant to the Peruvian coast, but ENFEN's formal classification relies on the ICEN and on sustained monthly or seasonal conditions rather than any single weekly value. For context on what has already occurred there this year, see our earlier report on Peru's coastal flooding, Lima heat and anchovy closure.
And the flat Niño 3.4 reading is not evidence of a stall. ENSO events typically mature during boreal late autumn or winter, so most of the intensification in a developing event normally arrives from late summer through autumn. Two identical weekly values in July mean very little in either direction.
One more number, because it is routinely mangled. NOAA's unrounded RONI estimate for April to June 2026 is +0.47°C, effectively at the +0.5°C historical threshold but slightly below it at full precision, and NOAA's public table rounds values to one decimal place. Three months earlier, for February to April, the value was -0.45°C, so the relative index has climbed nearly a full degree in two seasons, which is genuinely fast.
Two caveats belong with that. A historical El Niño episode requires the RONI to hold at or above +0.5°C across five consecutive overlapping three-month seasons, so a single season cannot establish the eventual episode record either way. And NOAA explicitly warns that because of the filtering applied to the ERSSTv5 data, recent RONI values may change for up to two months after first posting and should be treated as estimates. None of this makes the event unofficial: an El Niño Advisory is in effect and CPC states plainly that El Niño conditions are present. For why the relative index exists at all, see our explainer on why NOAA changed how El Niño is measured.
The forecast peak is a number about intensity. The gradient is a number about consequences. Both are worth tracking, and only one of them is getting written about. You can follow all four Niño regions as they update on our teleconnections dashboard, and see where we first flagged the eastern signal in June's look at the Niño 1+2 gap.