Key findings

1. Why this paper exists

Three weeks ago we published a long paper asking whether the 2026 El Niño is a rerun of the catastrophic 1877 event. The verdict was split. The ocean signal was real, the 1877 analogue was weak, and the season would be decided by the Indian Ocean Dipole. Since then the story has moved fast, and the headlines have moved faster. "Strongest in 75 years." "Super El Niño." "Race to save Australia." The reader reflex, and it is a healthy one, is to file it all under hype.

So this paper does one job. It takes the two articles that landed in our inbox and checks every checkable claim in them against the primary sources. Not against other news coverage. Against the CPC weekly data file, the Bureau of Meteorology SOI archive, the published model forecasts, the peer reviewed papers being quoted, and the rain gauge on our own hill. Where the articles are right we say so. Where they dress the data up we show exactly where the dressing was added.

The honest summary up front: we went in expecting to write a debunk. The data did not fully cooperate. The packaging is hype. The ocean is not.

2. What changed in the three weeks since our June paper

Our June paper was written on data to 14 June. In the begin of July the event shifted gear, and every core indicator moved the same direction at once. The table below puts the June snapshot next to the current one.

Sources: Bureau of Meteorology southern hemisphere monitoring (14 June and 14 July 2026 issues); NOAA CPC weekly Niño indices and ENSO diagnostic discussion (9 July 2026); Bureau SOI archive.

Two of those rows deserve a sentence each. The Bureau's relative index gained more than half a degree in four weeks, and the Bureau, which retired its cautious Watch and Alert dial back in December 2024, now writes plainly that "El Niño is underway" with ocean and atmosphere coupled. And the SOI did not drift down. It fell through the floor. A 30 day value of −25.8 means the pressure pattern that drives the trade winds has collapsed in a way the monthly record has not shown in over twenty years. We checked the Bureau's full monthly SOI table back through the record: the last month lower than June's −24.9 was February 2005 at −28.6. Even the 2015/16 super event never pushed a month below −22.

That last point matters for the hype question. The SOI line in the ZeroHedge piece, "levels not seen since 2005", read like classic headline exaggeration to us. We verified it. It is simply true.

3. The race: 2026 against 1997, 2015 and 2023 at the same week

Strength claims about a developing El Niño only mean something against the same date in past events, so we pulled the weekly file straight from CPC archive and lined up the four big contenders at the same week of July. These are raw weekly Niño3.4 anomalies from the same satellite dataset with the same 1991 to 2020 climatology, so the comparison is clean.

Weekly anomalies: NOAA CPC weekly Niño region SST file (OISST v2.1). Peaks: NOAA CPC Oceanic Niño Index (ERSST v5). Compiled by Dayboro Weather Research, 15 July 2026.

Read that middle column twice. The two monsters of the modern record, 1997 and 2015, both sat at +1.3°C in the second week of July. The current event is at +2.0°C, roughly 0.7°C ahead of both, and it crossed the +2.0°C "very strong" line on 8 July. We searched the whole weekly file for the first week each past event managed that: 3 September for 1997, 16 September for 2015, 6 October for 1982, 22 November for 2023. This event got there eight weeks before the fastest of them. In a weekly satellite record that runs back to September 1981, no event has been anywhere near this strong this early. That is a measured fact, not a forecast, and no amount of scepticism about headlines changes it.

The eastern Pacific is more extreme again. The Niño1+2 region off South America, the classic El Niño coastal signature, read +3.4°C in the same week. When the ZeroHedge piece said parts of the Pacific were running 2.7°C above normal, it was if anything behind the data.

4. What the forecasters said, next to what the headline said

The headline claim is "strongest in more than 75 years", presented as the CPC's warning. Well, that is not what the discussion says. We read the 9 July ENSO diagnostic discussion and the strength outlook behind it. The CPC's actual statements are these. First, an 81 per cent chance of a very strong El Niño in October to December, where very strong means a peak of at least +2.0°C. Second, that such an event would "rank among the largest" in the record since 1950. Third, in their own words, a 1 in 5 chance the event stays weaker than very strong. And fourth, the standard caveat the headline never quotes: even the strongest events do not deliver typical impacts everywhere.

"Among the largest since 1950" is a claim about membership of the top tier, which historically means the five events that peaked above +2.0°C: 1972, 1982, 1997, 2015 and 2023. "Strongest in 75 years" is a claim about beating every one of them, including 2015/16 at +2.75°C. The CPC never made that second claim. The headline writer made it, and the article then mixed in a 63 per cent "super" probability that was actually the previous month's number for a different season window. That is the mechanics of how a careful forecast becomes a loose headline: take the top of the range, drop the caveats, and let two different months' probabilities sit side by side as if they were independent confirmations.

Here is the twist, though. While the headline overstates the CPC, the raw model consensus now sits closer to the headline than to the CPC's careful wording. Zeke Hausfather, a climate scientist nobody would file under doom media, published an analysis on 13 July of 667 ensemble runs across 14 seasonal forecast models. The median peak sits at +3.6°C. Around 91 per cent of runs finish above the 2015/16 record. His framing is blunt: the gap between the strongest and fifth strongest El Niño of the past 150 years is only about half a degree, and the median forecast now sits nearly a full degree above the record. If the models are right, the 75 year claim is not hype. It is conservative, because 75 years understates it.

If. The same analysis carries the caveat that matters most, and we will not bury it: no seasonal model ensemble has ever forecast a +3.6°C event and then been verified against one, because no such event exists in the observational record. Models are calibrated on the events they have seen. Outside that envelope, the error bars are honest guesses. The forecast deserves respect, not worship.

5. Measuring a record in an ocean that is already warmer

There is a measurement question buried under all of this, and skipping it would flatter the doom case. The Pacific of 2026 starts warmer than the Pacific of 1997 because the whole tropical ocean has warmed. Part of any raw anomaly is background warming, not El Niño. The forecasting community's answer is the relative index, which subtracts the tropical mean and measures how much the Niño3.4 region stands out against the rest of the tropics (L'Heureux and colleagues, 2024). That is the index the Bureau now quotes, and it is why the Bureau says +1.47°C while NOAA's raw weekly file says +2.0°C for the same ocean in the same week. Neither number is wrong. They answer different questions.

Does the record chase survive the switch to the relative measure? That is not so simple, and the honest answer is: mostly. On the relative index the current event reads weaker, about +0.5°C for the April to June season, so some of the raw spectacle really is background warmth. But Hausfather ran the record odds both ways, and even in relative terms about 77 per cent of model runs still produce the strongest event on record. So the measurement nuance shaves the odds from 9 in 10 to about 3 in 4. It does not rescue the hype verdict.

Our read: quote both numbers, always. Any article that gives you only the raw anomaly is selling the event a little stronger than it is. Any sceptic who quotes only the relative number is selling it a little weaker. The two together are the honest picture, and on both, this event is exceptional.

6. The 2023 test: the last time these words were used

The strongest card in the sceptic's hand is 2023. That year the CPC gave a 63 per cent chance of a very strong event that would "rank among the largest El Niño events in the historical record", near enough word for word the language of the current discussion, and the media cycle produced the same super El Niño headlines we are reading now. The event peaked at +2.06°C, fifth on the list, behind even 1972. Sicne then, "super El Niño" headlines deserve a discount, and we applied one going in.

But apply the discount fairly. In mid July 2023 the weekly Niño3.4 anomaly was +1.1°C, and the June 2023 monthly SOI printed +0.2, dead neutral. The atmosphere spent months refusing to couple to the warm ocean. This time the same week reads +2.0°C with the SOI at −25.8 and the Bureau explicitly confirming ocean atmosphere coupling. The 2023 lesson is real: forecast language can outrun the event. It just is not this event's situation. The 2026 numbers are close to double the 2023 numbers on the same calendar date, on both sides of the ocean and atmosphere ledger.

What 2023 does still teach is about impacts, and here the sceptics have their best point, covered next with our own data.

7. What our own gauge says

Global indices are one thing. Here is what the event looks like from Lyndhurst Hill. Our archive runs from December 2005, which gives a twenty year monthly baseline (with some gap years, which we note for honesty). The 2026 year to date against that baseline:

Dayboro Weather Research station archive, December 2005 to July 2026. Averages computed from 2006 to 2025.

January to June totals 868.7mm against a 1,024mm average, about 15 per cent down, and the deficit is all in the dry start to the year. May ran two thirds above average, and June's 95.9mm made it the wettest June our gauge has logged since 2016, right in the middle of a declared El Niño. We wrote that up at the time. The media line that southern Queensland has been drying since January is true for the region in general and not particularly true on our hill. That is exactly the kind of gap between an index and a rain gauge that this site exists to document. July has produced 1.8mm in two weeks, which fits the El Niño script, and one dry fortnight in July proves nothing either way.

Now the sharper question: when a big El Niño does peak, what does it actually do here? We have two test cases in our own archive, and they could not disagree more.

Monthly rainfall in mm at our gauge through the two most recent major El Niño peaks. Dayboro Weather Research station archive.

The 2015/16 super event was a proper local drought: every single month from September to February came in under half the average, with January 2016 at a quarter of it. The 2023/24 event opened the same way, with a bone dry spring and an October of 17.5mm, then flipped, and January 2024 dumped 469.8mm on us, double the January average, in the middle of the fifth strongest El Niño on record. Both rows are from the same gauge on the same hill. Anyone who tells you they know what a +3°C El Niño does to Dayboro's summer is guessing. The spring lean, though, is consistent in both rows and in the wider record: nine of the ten driest winter to spring runs for eastern Australia happened in El Niño years (IPCC, 2021; Bureau of Meteorology). Dry spring is the bet. The summer is the open question, and the Indian Ocean Dipole, still neutral at −0.06°C on 12 July, remains the tell we flagged in June.

8. The money angle: trillions that were never a bill

The supply chain half of the ZeroHedge piece leans on a Goldman Sachs note: the top three exporters control 60 to 90 per cent of trade in soybeans, corn, rice, sugar and palm oil, so a weather shock in the wrong place can trigger export bans that amplify it. As a structural observation that is fair, and it is the same fragility argument our June paper made about just in time food systems, so we will not argue it all over again here. The June paper's answer stands: the cheap hedge is the one in your backyard.

The number that needs a haircut is the damage figure. The news.com.au piece prices the 1997 El Niño at 8.2 trillion Australian dollars of "global damage". The source behind that figure is a real and clever study, Callahan and Mankin in Science (2023), which estimated global income losses of 4.1 trillion and 5.7 trillion US dollars in the five years following the 1982 and 1997 events. Note what that is: a modelled estimate of growth that did not happen relative to a world without the event, integrated over half a decade, from a statistical model of country growth rates. It is a legitimate research finding at the top end of the literature, and it is also not money anyone paid, lost from an account, or could have insured. Converting it to Australian dollars and calling it "damage" turns an opportunity cost into a wreckage photo. The distinction between a modelled income loss and a bill is exactly the distinction between research and headline.

9. The cloud brightening story: reading past the rescue plot

The news.com.au piece is the more entertaining of the two, because it is structured as a thriller: a "race to save Australia", scientists "desperately working", dominoes to be stopped. We controlled every number in that story against the published paper behind it, which is Wan and colleagues, released in Science Advances on 8 July 2026, one week before the article.

What the study actually did is elegant. The 2019/20 Australian bushfires pumped smoke over the southeast Pacific, brightened the cloud deck, cooled the surface, and appear to have helped kick the ocean toward the triple La Niña that followed. That was an accidental natural experiment. The researchers reproduced it in a climate model, then asked whether doing it on purpose, spraying seawater to brighten clouds over the same patch of ocean, could weaken an El Niño already forming. In their simulations of the 1997 and 2015 events, nine months of targeted brightening cut roughly 1°C off a 2°C warm anomaly and brought the following La Niña forward. In the model, "we were able to turn what was an extreme or super El Niño into a neutral event", as the lead author put it.

What the study is not: a plan, a program, a trial, or a race. It is a modelling result from two replayed historical events, published days ago, with side effects the authors flag themselves, including one simulation where the strengthened La Niña disrupted rainfall as far away as the Horn of Africa. Nobody is spraying anything over the Pacific this year, no agency has proposed to, and the paper would not support doing so if they had. Turning that into a "race to save Australia" required no false facts, every number in the article is sourced. It required a movie trailer's sense of framing. The geoengineering debate deserves better, and plainly reported model results like these are how it should be had.

10. Verdict: how much of this is hype?

Score the two articles the way we scored the 1877 story in June, claim by claim.

Verified and real: the CPC probabilities, the SST anomalies, the SOI at a 21 year extreme, the event's record pace against 1997 and 2015, the concentration of crop export markets, and the existence and findings of the cloud brightening study. Every hard number in the ZeroHedge piece survived checking, and two of them were understatements.

Inflated in packaging: "strongest in 75 years" attributed to forecasters who said "among the largest since 1950" with a 1 in 5 chance of less. Last month's 63 per cent stacked beside this month's 81 per cent as if independent. A modelled five year income loss dressed as trillion dollar damage. A one week old simulation dressed as a rescue mission. July heatwaves in Europe and floods in China attributed to an El Niño whose main impacts arrive months from now, through a northern summer when its reach is weakest.

Left out entirely: the relative versus raw index question, the unverified amplitude caveat on the +3.6°C median, the 2023 precedent of similar language underdelivering, the Bureau's own warning that a strong Niño3.4 signal "does not necessarily mean strong impacts on Australia's climate", and the still neutral Indian Ocean Dipole that will do more to set Dayboro's spring than the Niño3.4 number will.

So the reader instinct that sent us these articles was half right. The hype is real, but it is hype about attribution, framing and rescue plots. It is not hype about the ocean. The Pacific is doing something the weekly record has not seen, the models say record, the honest caveats say the models are outside their experience, and our own gauge says the local story is still wide open. We would rather hand you that split verdict than a tidy one.

What to watch from here, in order of usefulness for a Dayboro reader. The weekly Indian Ocean Dipole value, neutral today, decisive if it turns strongly positive by September. The Bureau's relative Niño3.4, currently +1.47°C and climbing about 0.1°C a week. Our own spring rainfall against the table in section 7. And the preparation advice from the June paper has not changed: fuel, tanks, pumps, garden beds. Both futures start with the same weekend jobs.

Related Reading on Dayboro.au

References