Across 76 winters since 1951, El Niño strength tracks neither UK winter temperature nor UK winter rainfall. But of the nine strongest El Niño winters, six ended February with the blocked, colder circulation that steers Atlantic storms away from Britain — against about a third of all other winters.
El Niño is a periodic warming of the surface of the tropical Pacific. It moves where the biggest tropical rainstorms sit, and that in turn nudges wind patterns far beyond the tropics. Its cool opposite is La Niña; the cycle swings back and forth every two to seven years.
NOAA's Climate Prediction Center, the body that formally declares the phase, issued an El Niño Advisory on 13 August. Sea surface temperatures in its central Pacific monitoring box averaged 1.4C above normal through July, and 2.9C above normal in the eastern box nearer South America.1
What comes next is forecast rather than observed, and is treated here as such. The Center puts a greater than 90 per cent chance on a very strong event through the northern autumn and winter, and a 69 per cent chance that October to December exceeds every event in its records back to 1950.1 The Met Office's long-range model shows Pacific values above 3C in the months ahead; its head of long-range forecasting, Professor Adam Scaife, said he had never seen an El Niño signal this intense in our forecasts.
2
The question for readers in Britain is what any of that implies for the weather outside their window. The honest answer has three parts, and only the third is interesting.
The obvious test is to line up every winter since 1951 against how strong El Niño was at the time — NOAA's relative Niño index for December to February5 against the Met Office's UK average temperature for the same three months.3
The result is a cloud, not a line. Nine winters since 1951 qualify as strong events, meaning the index reached +1.5C or above. Their average UK temperature was 3.68C. The average across the other sixty-seven winters was 3.67C — a difference of one hundredth of a degree, against year-to-year swings with a standard deviation of 1.15C.
Temperature is only one way a winter can go wrong. Britain's costly winters are usually wet ones, so the more useful test is rainfall — and the Met Office publishes a UK winter total on the same basis, back to 1836.4
If anything the rainfall cloud is flatter than the temperature one. Strong El Niño winters averaged 328.2mm; all other winters averaged 323.0mm. The gap is 5.2mm — under two per cent of a typical winter's rain, against a standard deviation of 77.8mm. Within the nine, the range runs from 251.1mm in 1992 to 504.9mm in 2016, the second wettest winter in the 76.
Two national averages, two flat results. On the face of it, that settles the matter. It does not, and the reason is worth following.
Wind is harder to test than rain. The Met Office's public national series covers temperature, rainfall, sunshine, rain days and air frost, but not a long homogeneous record of winter windiness, so there is no equivalent chart to draw. What forecasters use instead is the North Atlantic Oscillation, or NAO — an index of the pressure difference between Iceland and the Azores that governs where the Atlantic jet stream, and with it the storm track, actually goes.6
In plain terms: when the index is positive, the jet is strong and aimed at Britain, and winters are mild, wet and windy. When it is negative, the pattern is blocked, the jet is pushed south or buckles, and winters are calmer, drier and colder. This is a proxy for storminess rather than a direct count of storms, and it is used here as one.
Published research had already flagged something the seasonal averages hide: the El Niño signal over the North Atlantic and Europe changes sign partway through winter, tending positive early and negative late.7 That is a testable claim, so we tested it — correlating El Niño strength against the NAO index month by month across all 76 winters.
The two lines cross. In strong El Niño winters the circulation starts the season tilted towards the Atlantic — milder, wetter, stormier — and ends it tilted towards blocking. In every other winter the tilt runs the other way. Averaged over December to February, the two halves largely cancel. That is the mechanism by which a real signal produces a flat seasonal number, and it is why the first two exhibits should not be read as an all-clear.
Averages across nine cases can conceal as much as they reveal, so here are the nine individually, each drawn from its December circulation value to its February one.
Six of the nine ended February blocked. Among the other 67 winters, 23 did — roughly two in three against roughly one in three. December runs the other way: only two of the nine strong winters had a blocked December, against 28 of the other 67. Three winters, 1973, 1992 and 2016, ended February firmly on the Atlantic side of neutral, which is the plainest possible demonstration that this is a tendency and not a rule.
The abstraction becomes concrete in the two extreme cases already visible in Exhibit 01. Winter 2015/16 followed a very strong El Niño and arrived with the circulation firmly positive: December 2015 carried an NAO index of +2.24, among the highest December values in the record.6 Storms Desmond, Eva and Frank ran in over a fortnight. Cumbria recorded new UK 24-hour and 48-hour rainfall records, 341mm at Honister Pass and 405mm at Thirlmere, and flood defences in England protected some 20,000 properties during December alone.10 That winter was the wettest on record for the island of Ireland in a series running back to 1850.9
Winter 2009/10 also followed a strong El Niño and did the opposite. Its December index was −1.93 and its February −1.98 — blocked throughout. The UK average temperature was 1.62C, the third coldest winter since 1951, and rainfall was 254.6mm, in the driest fifth of the record.34 The costs that winter were snow, ice and burst pipes rather than floods.
Both are consistent with a weak, sign-changing El Niño influence. Neither was caused by El Niño in any sense the data here can establish. That distinction matters: an association across 76 winters is not a causal account of any one of them.
The Met Office expects this event to raise the chance of wetter, stormier conditions across north-west Europe, including the UK, through the coming autumn and early winter.2 That is the early-winter half of the pattern above, and it points the same way as the November and December figures in Exhibit 03.
For late winter, the textbook expectation is the reverse — a raised chance of blocking, and with it colder, calmer, drier spells. It is a shift in the odds, not a forecast, and it has failed before. In the strong El Niño winter of 2023/24 the expected pattern did not appear at all; Met Office researchers reviewing that season found North Atlantic circulation close to neutral, with the seasonal forecast itself weakly positive rather than negative.8
The practical reading, then. A wetter and windier than usual autumn and early winter is the better-supported half of the expectation, and flood risk is the thing worth planning around. A cold, blocked late winter is a raised possibility rather than an expectation, and on this record it fails about one time in three. Anyone offering a confident month-by-month UK forecast on the strength of a Pacific sea temperature is going beyond what the record supports.
Nine strong events in 76 years is a thin sample, and it cannot separate El Niño from everything else happening in those particular winters — the stratospheric polar vortex, the quasi-biennial oscillation and Atlantic sea temperatures all act on the same seasons. Wind is tested here only through a circulation index, not through a direct record of UK wind speeds or gusts, because no long homogeneous public series of the latter exists in the same form. Floods are illustrated by two cases, not counted systematically; a proper flood test would need river-flow and insurance-loss records this piece does not use. NOAA notes that its recent Niño index values are estimates, revisable for up to two months.5 And nothing here establishes cause in any single winter.
Every strong event since 1951, with the four measures used above. Winters are dated by the year in which February fell.
| Winter | Index | Temp (C) | Rain (mm) | Dec NAO | Feb NAO |
|---|---|---|---|---|---|
| 1957/58 | +2.0 | 3.26 | 344.1 | +0.12 | −1.06 |
| 1965/66 | +1.7 | 3.12 | 354.2 | +1.37 | −1.39 |
| 1972/73 | +2.0 | 4.05 | 257.0 | +0.19 | +0.85 |
| 1982/83 | +2.5 | 3.41 | 337.4 | +1.78 | −0.53 |
| 1986/87 | +1.6 | 2.76 | 301.7 | +0.99 | −0.73 |
| 1991/92 | +2.3 | 4.11 | 251.1 | +0.46 | +1.07 |
| 1997/98 | +2.2 | 5.35 | 349.2 | −0.96 | −0.11 |
| 2009/10 | +1.5 | 1.62 | 254.6 | −1.93 | −1.98 |
| 2015/16 | +2.2 | 5.46 | 504.9 | +2.24 | +1.58 |