The 2026–27 El Niño: state of the event and expected regional effects (2026-09-04)

Created
· Expires 2026-12-03 15:35 · 260904-yw

The 2026–27 El Niño: state of the event and expected regional effects

A technical briefing. Current to 4 September 2026.

This document takes off from Mandy Freund's 4 September 2026 piece in The Conversation, "A record-shattering El Niño is almost here. Here's what that means for the world's weather," which is written for a general audience and centred on Australia. It keeps her question, what a record event means for the world's weather, and answers it in drier terms, region by region, from the agency outlooks current on the day of writing. Where her figures did not survive checking against primary sources, that is noted rather than repeated. Section 11 explains how the document was made and how to reproduce it.

Season abbreviations follow meteorological convention: SON is September–November, OND October–December, DJF December–February, JFM January–March, MAM March–May. "Anomaly" means departure from a 1991–2020 climatology unless an agency states otherwise.

1. State of the event

Declarations. The World Meteorological Organization (WMO) announced El Niño onset on 2 June 2026, the US National Oceanic and Atmospheric Administration (NOAA) on 11 June, Australia's Bureau of Meteorology (BoM) on 16 June. Onset in boreal late spring is early; the events that reached very strong intensity (1982, 1997, 2015) all had an established warm anomaly by mid-year.

Ocean and atmosphere. NOAA's Climate Prediction Center (CPC) discussion of 13 August reported July anomalies of +1.4 °C in Niño-3.4 (5°N–5°S, 170°W–120°W), +1.7 °C in Niño-3, and +2.9 °C in Niño-1+2, the far-eastern region off Peru. Warmest in the east marks this as an eastern-Pacific (EP) type event in the mould of 1982–83 and 1997–98, not a central-Pacific type like 2009–10. The thermocline was deeper than average across 180°–100°W, the reservoir the surface draws on through autumn. Growth continued: BoM's 1 September update gave the relative Niño-3.4 index for the week ending 30 August as +2.45 °C against a threshold of +0.80 °C. The atmosphere has coupled: Southern Oscillation Index values near two standard deviations negative in July, −14.6 on BoM's 30-day index to 30 August (threshold −7), westerly wind anomalies from the western to the east-central Pacific, convection enhanced in the central and eastern basin and suppressed over Indonesia. The Bjerknes feedback loop is closed.

Forecast. CPC gives a greater-than-90% chance of a very strong event through boreal autumn and winter, and a 69% chance that the OND three-month relative index reaches +2.5 °C or more, which "would exceed the strength of previous El Niño events dating back to 1950." BoM forecasts values "above the warmest observed since reliable records began in 1950" and persistence into austral autumn 2027. WMO's August update gave a nearly 100% probability of persistence through DJF, the highest confidence it has ever expressed. Peak: OND to NDJ.

The "relative" index, and the source article's 2.3 °C. In 2026 NOAA moved its classification to the Relative Oceanic Niño Index (RONI), which subtracts the tropical-mean (20°S–20°N) sea-surface temperature anomaly from the Niño-3.4 anomaly and rescales the result. The atmosphere responds to the gradient between the central Pacific and the rest of the tropics, not to the absolute value, and as the whole tropical ocean warms the absolute anomaly overstates that gradient. A record on the relative index is a record in the quantity that drives teleconnections, and a smaller number than the absolute anomaly. The Conversation's "2.3 °C above average" for August was not found in any NOAA, BoM or WMO product; the nearest published figure is BoM's +2.45 °C weekly value. A monthly mean near 2.3 is plausible arithmetic but is not a published number.

Concurrent conditions. The Indian Ocean Dipole (IOD) index was +0.22 °C for the week ending 30 August, back in neutral after three weeks above the +0.4 °C positive threshold; models still lean toward a weak positive IOD in austral spring. A positive IOD reinforces the drying over Australia and Indonesia and the wetting over East Africa. Global sea-surface temperature set a daily record of 21.0 °C on 21 June (Copernicus Marine), with about 82% of the ocean under marine-heatwave conditions at the end of June, and BoM reported record daily global SSTs through August. Peru's coast is far ahead of the basin index: ENFEN's 28 August communiqué held a Coastal El Niño alert with anomalies reported up to +6 °C, and 893 districts were placed under a state of emergency (per Infobae; not confirmed against ENFEN's text).

Already attributed. Three tropical cyclones were active simultaneously in the eastern Pacific on 2 September (Hurricane Lowell, briefly Category 5; Hurricane Karina, Category 4; Tropical Storm Marie). The Conversation's "four active tropical cyclones" appears to conflate this with the 2015 case that the same reporting cites as precedent.

2. Mechanism, briefly

The tropical Pacific normally holds its warmest water in the west, under the rising branch of the Walker circulation. In El Niño the warm pool spreads east and the deep convection that drives the tropical atmosphere relocates by thousands of kilometres. Three channels carry that to the rest of the world. Direct displacement of tropical rain: less over Indonesia, northern Australia and the Amazon, more over the central and eastern Pacific and the Peruvian coast; the most reliable channel. Rossby wave trains into the extratropics: in the north, the Pacific–North American pattern (deep Aleutian Low, western-Canada ridge, southeastern-US trough, a strengthened and southward-shifted subtropical jet across the southern US and into the Atlantic); in the south, the Pacific–South American pattern reaching the Amundsen Sea. These are strongest when the winter jet is strongest, so they are DJF signals in the north and winter-to-spring signals in the south. The atmospheric bridge: changed winds and cloud warm the Indian and tropical Atlantic oceans with a lag of a season or two, which is why global temperature peaks the year after the winter peak, and why 2027, not 2026, is the candidate record year.

Two consequences for reading what follows. Every teleconnection is a shift in the probability distribution of a seasonal mean, not a deterministic forecast, and in mid-latitudes the atmosphere's own variability is usually larger. And teleconnections are phase-locked to the annual cycle, so a regional effect without a season attached is not saying much.

3. "Expected" versus "potential"

Expected effects satisfy two conditions: the composite of strong El Niño events shows a consistent sign, and the responsible agency's current outlook for this event shows the same sign. Potential effects fail one: the composite is inconsistent, the outcome hinges on a second driver (IOD, Southern Annular Mode, North Atlantic Oscillation) not yet settled, or no agency outlook exists. Australia's rainfall is the canonical potential: most major Australian droughts coincide with El Niño, but not every El Niño produces one, and strength does not scale the effect. A reader who treats every potential as expected will over-prepare for some outcomes and then discount the expected list when the potentials fail to arrive.

4. North and Central America, with the Caribbean

Expected, DJF (United States). CPC's 20 August outlook carries the classic strong-El Niño winter: above-normal temperature across the northern tier, with probabilities exceeding 50% from the Pacific Northwest to the Northeast; above-normal precipitation across the Southeast, exceeding 60% probability by NDJ and persisting through JFM; and above-normal precipitation for California and the Southwest from DJF through FMA, peaking in JFM, on the ground that "recent EP-based strong El Niños have produced copious precipitation, especially in late winter." The extended subtropical jet steers Pacific storms into the southern US: a wet, stormy Gulf Coast and Florida with elevated cool-season tornado risk, a milder and drier Ohio Valley and northern Rockies.

Potential, California. The composite says wet; the last two strong events split. 1997–98 delivered; 2015–16 under-delivered for Southern California while the Sierra did well. Snowpack matters more than rain for the Colorado and the Sierra reservoirs, and a warm winter raises the snow line. "One wet year can make a major difference. It cannot resolve the river's larger structural problems" (Water Education Foundation, 20 August). Debris flows on Southern California burn scars are the flood risk to watch if the late-winter track verifies.

Expected, Atlantic hurricanes. NOAA's 6 August update cut the season to 7–13 named storms, 2–6 hurricanes, 0–2 majors, with a 75% chance of a below-normal season. Through early September: five named storms, no hurricanes, and July Caribbean shear the highest in the record back to 1979. The eastern Pacific is the mirror: 15 named storms, 6 hurricanes, 4 majors including two Category 5s. Quiet is not zero; late-season Caribbean development is possible when shear relaxes.

Canada. Environment and Climate Change Canada's winter outlook is due in late November. The composite is a mild winter with reduced snow across the Prairies and southern Ontario: expected on the composite, unconfirmed by outlook.

Expected and under way, Central America. Honduras extended a drought red alert to 103 of 298 municipalities on 26 August, citing deficits of 100–200 mm and 1.8 million people in food crisis; El Salvador and Panama declared national emergencies on 25 August. An Oxfam and Action Against Hunger assessment of 58 Dry Corridor municipalities, published 31 August, found 78.6% of corn-growing households had lost 75–100% of production. Panama's Pacific side was 60% below normal and the Canal cut transits to 32 vessels a day. The World Food Programme (WFP) projects the largest proportional increase in acute food insecurity anywhere, more than 16 million people, in Central America and the Caribbean, with effects running through 2027 because they land on future harvests.

Expected, Caribbean. The Caribbean Institute for Meteorology and Hydrology forecasts below-normal rainfall from the Guianas across the Antilles through October, persisting into March 2027. Fewer hurricanes and a dry wet season are one phenomenon: the shear that suppresses storms suppresses rain.

Mexico. Abnormally dry conditions covered 22.5% of the territory to 15 August, up from 10.8% in January, but the drought monitor did not name El Niño and Mexico City's Cutzamala system closed August at 81%. The winter composite is wetter over northern Mexico. The concern Mexican researchers flag is drought in 2027 as the event decays: potential, not expected.

5. South America

Expected and under way, Peru and Ecuador. With Niño-1+2 at +2.9 °C in July and coastal anomalies up to +6 °C, ENFEN assigns 62% to the "extraordinarily warm" category over September–January and forecasts above-normal rain on the north and central coast and northwest highlands from November. The precedents (1982–83, 1997–98, the 2017 coastal event) are torrential rain on a desert coast, Piura and Tumbes flooding, western-Andes landslides, and dengue and leptospirosis afterward. The 893-district emergency is the state acting on that history in advance.

Already realised, the anchovy fishery. Peru's first north-central season opened with a quota of 1.91 million tonnes, the lowest in a decade, and was suspended after about a quarter was caught because of juvenile presence; the industry association reported industrial catch down more than 80%. Peru supplies a large share of global fishmeal, so the effect on aquaculture feed prices is global and fast.

Expected, northern and northeastern Brazil. INMET, CPTEC and INPE forecast reduced rainfall and more frequent dry spells across the Nordeste and eastern Amazon, with Amazon drought "similar to or even higher than 2015," when fire counts ran 36% above the 12-year average. NASA Earth Observatory reported the Rio Negro at Manaus falling 77 cm in 26 days from a 2025 low that was the lowest in 122 years (observation date not confirmed). Fire risk in Mato Grosso and the North Region runs through October; river transport and northern hydropower are the second-order effects.

Expected, the southeastern quadrant. Southern Brazil, Uruguay and northeastern Argentina sit under the strengthened subtropical jet and its stalled fronts. INMET projects up to 100 mm above normal in Rio Grande do Sul and Santa Catarina; INUMET gives northern Uruguay a 60% chance of above-normal rain; Argentina's SMN favours above-normal SON rainfall in the northern Litoral, with September anomalies in Misiones possibly exceeding 150–200 mm on a normal of 120–150. Paraná and Uruguay river flooding, as in 1998 and 2016, is the expected form. Central Chile tends wet in El Niño winters and the last DMC bulletin (May–July) projected normal-to-above rainfall.

Expected, the north coast. Colombia's IDEAM confirmed El Niño three months early and gives 63% to very strong intensity for NDJ; Venezuela expects a 20–25% rainfall reduction. Colombia's grid is heavily hydroelectric and its El Niño years (1992, 2016) are its rationing years.

6. Europe and the Mediterranean

Europe has the weakest and least reliable El Niño signal of any region, and the most useful thing this section can do is say so precisely.

What the outlooks say. The Copernicus Climate Change Service (C3S) seasonal page of 10 August: "Over Europe, there are signals for a wetter-than-average autumn in southern and western parts, but here forecast uncertainty is higher and forecast skill is lower than in tropical regions." Temperature is "more likely than not to be above the 80th percentile" over all but the far east and north, which is trend and warm surrounding seas, not El Niño. No 2026 agency or peer-reviewed statement attributes a European seasonal signal to this event. World Weather Attribution's 23 July analysis of the western-European heat and drought does not mention El Niño; it attributes the evaporative drought to anthropogenic warming, at about 80 times more likely.

What the literature supports. The documented pathway runs through the stratosphere: El Niño winters favour a weakened polar vortex, a negative North Atlantic Oscillation tendency in February–March, a southward-displaced storm track, wetter Iberia and Mediterranean, colder late winter in the north. The effect is modest, strongest for strong EP events, and non-stationary: it varies with the Pacific Decadal Oscillation, and seasonal models capture the early-winter predictability but not the non-stationarity. Potential, and label it so.

What Europe gets with certainty is the global-temperature channel. A very strong El Niño peaking in DJF raises the odds that 2027 sets a global record, and Europe warms at roughly twice the global rate. The Mediterranean was already in marine-heatwave conditions in mid-2026, with anomalies near 6 °C in the Gulf of Lion. The European reading of this event is not "expect an El Niño winter" but "plan for the summer of 2027."

7. Africa

Expected, East Africa and the Horn: a wet OND. The IGAD Climate Prediction and Applications Centre (ICPAC) outlook of 18 August gives a 90% chance of enhanced short rains over southern Ethiopia, central-to-southern Somalia and northeastern Kenya, and a high chance of totals above 400 mm in central-to-southern Somalia, central Kenya, the Lake Victoria basin, Burundi, western Rwanda and western Tanzania. ICPAC's analogues are 1997 and 2023, both flood years: 1997–98 brought the worst floods in Kenya and Somalia in decades; OND 2023 flooded the Juba and Shabelle basins and displaced more than a million people in Somalia. The mechanism is El Niño and a positive IOD together, so with the IOD back in neutral the sign is more certain than the magnitude. OND supplies up to 70% of annual rainfall in parts of Ethiopia, Kenya and Somalia, so the season that threatens floods also refills pastoral water. Rift Valley fever, cholera and displacement are the standard sequels. ICPAC also warned of a dry July–September in the northern Horn first: drought, then flood.

Expected, southern Africa: a dry, hot 2026–27 season. The SADC Climate Services Centre outlook of 27 August favours below-normal OND rainfall across much of Angola, southern Zambia, Zimbabwe, Mozambique, Namibia, Botswana, most of South Africa, Eswatini and Lesotho, continued below-normal JFM rainfall across Angola, Namibia, Botswana, Zimbabwe, Zambia and parts of South Africa, and above-average temperatures nearly everywhere. Precedents: 2015–16, the worst regional drought in 35 years, and 2023–24, when Zambia, Zimbabwe and Malawi declared national disasters after maize failure. FEWS NET projects delayed onset and below-average rain for October–May, southern Madagascar first. WFP counts more than 18 million additional food-insecure people across eastern and southern Africa. One buffer: Lake Kariba had recovered to about 42% of usable storage by mid-2026, so Zambian and Zimbabwean hydropower starts this event with margin it lacked in 2024.

Potential only, the Sahel and West Africa. The composite is weakly dry in El Niño summers, but the 2026 season is largely over and ran near the April forecast of average-to-below totals with above-normal pockets in Chad, Niger, northern Nigeria and Burkina Faso. AGRHYMET's July update said explicitly that "the presence of strong El Niño does not automatically lead to a poor agricultural season in the sub-region," and Nigeria's most damaging 2026 event was May flooding. WFP still counts nearly 6 million additional food-insecure people across West and Central Africa, driven as much by conflict and prices as by weather. North Africa has no robust signal.

8. Asia

Expected and largely realised, India. The India Meteorological Department (IMD) lowered its seasonal forecast to about 90–92% of the long-period average as El Niño developed (two figures in different reports, not reconciled), and the season has tracked it: a large June deficit, an above-normal July, August 16% below normal, cumulative deficit about 14% at the end of August (reported; IMD's table was not retrievable). IMD projects below-normal September rainfall and above-normal temperatures; reservoirs entered the season at 34% of capacity; about 200 districts were flagged for contingency planning. A deficient monsoon in a strong El Niño year is the continent's most robust teleconnection, and it has arrived. Potential: the OND northeast monsoon over Tamil Nadu tends to be enhanced in El Niño years, especially with a positive IOD, but no 2026 IMD statement on it was found.

Expected, the rest of South Asia. The regional forum in May forecast below-normal monsoon rain and above-normal temperatures; Pakistan's PMD said the same through September; Bangladesh expected less rain and up to ten heatwaves; Nepal put the chance of below-average rain at 55–65% across its southern provinces. Sri Lanka's department, in a release that reads as pre-dating the declaration, described ENSO-neutral conditions: the one outlier.

East Asia: wet summer realised, warm winter expected, 2027 to watch. China's National Climate Center described two major rain belts, heavier rain over southern China and the lower Yangtze, more and stronger typhoons (three Category 5s by 1 August), and repeated heavy-rain events in the north. Japan's July NINO.3 anomaly of +2.5 °C tied July 1997 as the highest since 1949. The composite winter is mild across China, Korea and Japan. The larger risk is the decay year: the summers after strong El Niño winters, 1998 and 2016, gave the Yangtze basin its worst floods of their decades as the lagged Indian Ocean warming strengthened the western Pacific subtropical high. Potential for summer 2027, worth planning now.

Expected and under way, the Maritime Continent and Southeast Asia. Suppressed convection over Indonesia is the most direct teleconnection there is. Indonesia recorded 15,652 fire hotspots in Kalimantan and Sumatra from January to 27 August, Kalimantan rising from 1,339 in July to 9,849 in August. The ASEAN Specialised Meteorological Centre raised its haze alert to Level 3, the highest, on 27 August, with Borneo air quality unhealthy to hazardous and smoke thick enough to obscure hotspot counts; the haze reached Metro Manila on 1–2 September and closed schools. BMKG forecasts low rainfall over 77% of Indonesia in September and dry palm-oil provinces through October. The Philippines' PAGASA forecasts below-normal rain spreading from northern Luzon in September to most of the country in November; Angat Dam hit its lowest level since 1967 on 25 July. For the Mekong, low flows, delta salinity intrusion and hydropower shortfall are the expected form, though no 2026 Mekong River Commission bulletin was retrieved. Palm-oil and rice yield effects lag by six to twelve months, placing them in 2027. WFP counts 8.2 million additional food-insecure people across Asia and the Pacific.

9. Australia, New Zealand and the Pacific Islands

Australia: the "textbook response," with the textbook's footnotes. BoM's spring outlook favours drier conditions for parts of the south and east, wetter for much of Western Australia, above-average day and night temperatures almost everywhere, the dry signal strongest in September with a "wet signal" emerging from October. The fire authorities' council (AFAC) outlook of 19 August finds elevated fire risk across seven states and territories, citing "a very strong El Niño in the tropical Pacific Ocean, a climate driver that typically promotes abnormally warm and dry weather over large areas of Australia during spring," on top of high fuel loads from wet preceding years. Murray–Darling storage was reported near 46% in June, Hume at 23% (reported; the MDBA page was not retrievable).

The footnotes are the ones Freund gives, and they are right. 1982–83 put the eastern two-thirds of the continent into the driest decile and ended with Ash Wednesday. 1997–98, the strongest event on the absolute index until now, confined its rainfall impact to coastal southeastern Australia and Tasmania. 2015–16 was dry and hot in the east. 2023–24 was dry through spring and then wetter than average in summer. BoM's own line is that "the strength of an El Niño is not directly proportional to the rainfall impacts." The modulators are the IOD, now neutral and forecast only weakly positive, and the Southern Annular Mode, strongly positive in July, which favours a wetter east while it lasts. Expected for spring: heat, a dry September, elevated eastern fire danger. Potential for summer: continental drought, with 1997–98 and 2023–24 as the two counter-examples in the last four strong events.

Great Barrier Reef. AIMS's 2025–26 survey found hard-coral cover rising in the northern (30.0% to 35.1%) and central (28.6% to 31.6%) regions, "the start of recovery from the impacts of the 2024 mass bleaching event." Summer 2026–27 is a higher-risk season. The fourth global bleaching event, which touched 84% of the world's reef area over 2023–25, was judged ended only in mid-2025; the gap before the next may be the shortest yet.

New Zealand: expected, the classic pattern. NIWA's August–October outlook: a prevailing westerly regime, below-normal rain for the north and east of the North Island (50% probability) and the east of the South Island, above-normal for the west of the South Island, warm nearly everywhere, elevated wildfire risk in Canterbury, Otago and Marlborough.

Pacific Islands: expected, and the most severe per capita. The South Pacific Convergence Zone shifts toward the equator. SPREP declared a regional El Niño phase in June, warning of drought for twenty states including Fiji, Papua New Guinea, the Solomon Islands and Vanuatu, while NIWA's Island Climate Update forecasts a pronounced wet shift on the equator for Nauru, most of Kiribati, Tuvalu and Tokelau. Papua New Guinea is already the worst hit: rainfall below average for nearly a year, Highlands frost destroying gardens in Chimbu, Southern Highlands and Hela, Oxfam's estimate of up to 3 million people affected, the Porgera mine suspended for lack of water. That is the 1997–98 PNG scenario recurring. Cyclone genesis shifts east, raising risk for the Cook Islands and French Polynesia. Sea level in the western Pacific falls as the warm pool migrates: May anomalies were −5 to −25 cm across Micronesia, the Marshalls, Guam and Palau, and in 1982–83 and 1997–98 Samoa experienced taimasa, the "foul-smelling tide," when reef flats stood exposed for months and their coral died.

10. What a record event changes, and what it doesn't

What scales with strength. The direct-displacement effects: Indonesian and Amazon drought, Peruvian coastal rain, the East African short rains, Central American drought, the eastern-Pacific cyclone season and the Atlantic's suppression. So does the global-temperature response: the Berkeley Earth and Texas A&M estimate of a temporary +0.3 °C on top of trend puts 2027 near 1.7–1.8 °C above pre-industrial, a record by a wide margin, with coral heat stress and marine heatwaves scaling with it.

What does not. The mid-latitude wave trains saturate near "strong"; a very strong event makes the PNA winter more likely, not proportionally larger. Australia's rainfall does not scale. Europe's signal does not scale into reliability. The Sahel is governed by other drivers.

What the flavour changes. An EP event, with the largest anomalies off South America, produces the 1982 and 1997 signatures: strongest coastal-Peru and California late-winter responses, strongest Atlantic suppression. The July gradient says EP, but events evolve; CPC's next discussion is due 10 September.

What "record" means here. The record being forecast is on the relative index, the gradient in a tropical ocean that is itself at record temperature. That is the physically relevant record. It is not the same as the absolute Niño-3.4 anomaly exceeding 1997's, though the models suggest it may.

The decay. Strong El Niños are usually followed by La Niña within a year (1983, 1998, 2016), and the transition brings some of the largest impacts: the 1998 Yangtze floods, the Atlantic hurricane rebound, the recovery monsoon. The event ends in austral autumn 2027; the consequences do not.

The reference-class caveat. 2023–24 was also forecast as a possible record, peaked near +2.0 °C on the traditional index, and gave Australia a wet summer. August forecasts of the peak have skill and have been wrong in both directions. Everything above is conditional on the peak arriving as forecast.

11. How this document was made: method, biases, replication

Pipeline. The author is Saga, an AI agent running on Claude Fable 5.1, working for Peter Kaminski. The work was done in one session on 4 September 2026 in about twenty minutes of wall-clock time from launch to first full draft (08:04 to 08:21 Pacific), plus a trimming pass and this review:

  1. The Conversation article was fetched and distilled to a list of factual claims.
  2. Four scoping questions went to the principal (audience, sourcing policy, regional grouping, length); the answers fixed the structure: a technically literate generalist as reader (terms defined once, indices named); live agency outlooks plus canonical teleconnections, cited, with "expected" and "potential" kept separate; six merged regional sections rather than nine; about 3,000 words of body, with this section as additional budget.
  3. Four research agents (Claude Sonnet 5, with web search and page fetch) ran in parallel with written briefs: the state of the event and forecast agencies; the Americas and Europe; Africa and Asia; Australia, the Pacific, the ocean and cryosphere. Each brief demanded a URL and issue date per fact, "not found" over filling a gap from memory, and exact quotation of numbers. The briefs are preserved verbatim in the working files.
  4. The AI author read the four reports, re-fetched the primary pages behind the highest-stakes numbers (the CPC discussion PDF, the BoM ENSO page, the WFP release, the ICPAC and SADC outlooks, the phys.org cyclone report) and confirmed them, then wrote the document. The Sri Lanka outlier and the two Conversation discrepancies were kept, not smoothed.

Where the pipeline is weak.

  • The author's training data ends in June 2026. Everything about this event is from fetched pages. Mechanism and precedent (1982, 1997, 2015, 2023) are from training, well-represented and stable material, but not re-verified here.
  • Agency sites block automated fetching. BoM's outlook pages and the Murray–Darling Basin Authority returned HTTP 403 to the agents, so several Australian figures come from search snippets and secondary reporting, marked "reported" in the text. NOAA's press-release server and IMD's rainfall tables were likewise unretrievable.
  • Anglophone and institutional bias. Sources are overwhelmingly English-language agencies and humanitarian bodies. Lusophone Africa, Francophone West Africa and Latin America lean on wire reports; the Peruvian figures come from Infobae's reporting of ENFEN, not ENFEN. A region with a strong English-language climate service (India, Australia, the US) is covered better than a region of equal exposure without one (Mozambique, the Sahel, the Mekong).
  • Humanitarian framing. Impact numbers come from WFP, FEWS NET and Oxfam, whose purpose is to raise resources; their projections are conditional and stated as "at least" or "could," and are reproduced with those hedges intact.
  • Uneven research depth. Two agents made 36 and 41 tool calls; the other two made 5 and 4, delegating to sub-searches whose transcripts were not reviewed. The Africa/Asia and Americas/Europe sections therefore rest on a shallower verification trail, which is why the author re-fetched ICPAC, SADC and WFP directly.
  • The expected/potential classification is the author's judgment, applied consistently but not derived from any single published composite. A reader with the CPC, BoM or ECMWF composite atlases can check each call.
  • Selection. The document reports what agencies chose to forecast. The Arctic, Middle East, Central Asia and the Antarctic sea-ice response are absent because no 2026 statement was found, not because the literature is silent.

How to replicate. Fetch, in this order, and compare against Sections 1 and 4–9:

  1. NOAA CPC ENSO Diagnostic Discussion, cpc.ncep.noaa.gov/products/analysis_monitoring/enso_advisory/ensodisc.pdf, and the RONI outlook and probability tables linked from it. A new issue is due 10 September 2026 and supersedes Section 1.
  2. BoM, bom.gov.au/climate/enso/ (fetch with a browser user-agent; the outlook subpages refuse scripts).
  3. WMO El Niño/La Niña Update and Global Seasonal Climate Update, wmo.int.
  4. IRI ENSO forecast plume, iri.columbia.edu/our-expertise/climate/forecasts/enso/current/ (graphic only).
  5. Regional: CPC seasonal outlooks, NOAA hurricane outlook, CIMH (cimh.edu.bb), ENFEN (gob.pe/enfen), INMET/CPTEC, SMN Argentina, C3S seasonal (climate.copernicus.eu/seasonal-forecasts), ICPAC (icpac.net), SADC CSC, AGRHYMET (agrhymet.cilss.int), IMD (mausam.imd.gov.in), JMA TCC (ds.data.jma.go.jp/tcc), BMKG, PAGASA, ASMC (asmc.asean.org), AFAC, AIMS, NIWA (niwa.co.nz), SPREP.
  6. Impact: WFP release of 5 August 2026; FEWS NET country outlooks; Copernicus Marine press releases; NOAA Coral Reef Watch.

Queries that reproduced the agents' results were of the form "<agency>" "El Niño" 2026 outlook <season> and "<region>" "El Niño" 2026 site:<agency domain>. Anything a re-run cannot find at the primary source should be treated as unverified regardless of what this document says.

Sources

By section, with issue date. † marks items the author confirmed against the primary page; the rest are as reported by the research agents with the URL they gave.

Event state

North and Central America, Caribbean

South America

Europe

Africa

Asia

  • ICRA, Mid-Monsoon Update, Aug 2026 (PDF); Business Standard, 4 Sep 2026 (headline only); Down To Earth, reservoir storage, 18 May 2026, and district contingency planning, 2026.
  • Express Tribune / Dawn, SASCOF-34 and PMD, May–Jun 2026; Daily Star (Bangladesh), Jun 2026; Nepal DHM via Nepal News.
  • Global Times / CMA, 18 Aug 2026; China Daily, 10 Jul 2026; Insurance Journal, 11 Aug 2026.
  • JMA Tokyo Climate Center, El Niño outlook, 10 Aug 2026. https://ds.data.jma.go.jp/tcc
  • The Star (Malaysia), hotspot counts, 29–30 Aug 2026; ASMC, Level 3 alert, 27 Aug 2026 (https://asmc.asean.org/home/); Manila Times, 2 Sep 2026; Palm Oil Magazine citing BMKG, 26 Jul 2026; PAGASA advisory, 9 Jun 2026 (https://pagasa.dost.gov.ph/press-release/216) and 26 Aug 2026 (secondary).

Australia, New Zealand, Pacific

Source article

About this document

Released under the Creative Commons Attribution 4.0 International licence (CC BY 4.0). Attribute to Peter Kaminski. Standalone; not part of any website or series.

Drafted by Saga, an AI agent running on Anthropic's Claude Fable 5.1, under Peter Kaminski's direction, with research assistance from four Claude Sonnet 5 agents as described in Section 11. Peter Kaminski directed the work, set its scope, and is responsible for it. The source article by Mandy Freund is separately licensed by The Conversation under CC BY-ND and is cited, not reproduced.

Figures describe the situation as of 4 September 2026 and are superseded by the NOAA CPC discussion of 10 September 2026 and by each agency's next update.