Route 03 · The Copernicus systems on ENSO
El Niño, system by system
Most seasonal outlooks are downstream of one question: what is the tropical Pacific going to do? Here are the ten Copernicus systems' answers, side by side — each measured against its own reforecast so that its own drift cancels, combined with one vote per system, and scored against what the ocean actually did. The consensus is the headline; the ten lines behind it are the argument.
Niño 3.4 now
+2.98 °C
ERA5, September 2026
Consensus peak
+4.08 °C
December 2026, one vote per system
Ten systems at the peak
+3.4 … +4.5 °C
lowest to highest system mean that month
Relative Niño 3.4 peak
+2.83 °C
December 2026 — the warm basin taken out
What the systems say
10 systems · started October 2026The ocean the ten systems start from is already well into El Niño: ERA5 put Niño 3.4 at +2.98 °C in September 2026. Their consensus — one vote per system — carries it to +4.08 °C in December 2026, and not one of the ten lets it fall below +1.8 °C at any month through March 2027. For scale, the strongest month in ERA5's record since 1940 on this same baseline is +2.88 °C (December 1997), then +2.85 (January 1983). The consensus peak would top it by +1.20 °C, and even the lowest system at that month (NCEP, +3.39) would top it too.
They agree on the peak and part company on the decline. The spread of the ten system means is 0.22 °C in the first month and 0.71 °C by the last. At the peak month they run from +3.4 (NCEP) to +4.5 (DWD); by March 2027 from +1.8 (NCEP) to +4.2 (DWD). Individually they peak 3 in November, 5 in December, 2 in January.
Part of that number is not El Niño. The relative index subtracts the whole tropical band, which has warmed since 1993–2016, and it peaks at +2.83 °C in December 2026. At the consensus peak about 31 % of the raw Niño 3.4 anomaly is the basin being warmer, not the event — the part to take away before comparing with events of the past. On the relative index the strongest month on record is +2.59 °C (January 1983).
In the Indian Ocean the consensus puts the dipole positive, +0.90 °C in October 2026, and back to −0.23 by March 2027. Its reforecast correlation has fallen to 0.58 by January 2027, so the later months of that row say little.
How much to trust it. Over 1993–2016 the consensus Niño 3.4 correlates with ERA5 at 0.99 in the first month and 0.93 by the sixth, with an RMSE of 0.11 rising to 0.35 °C. Correlation rewards getting the rhythm of the years right; the RMSE is the size of a typical miss, and at the sixth month it is a sizeable fraction of a degree. Individually the best at month six is ECMWF (0.95), the weakest NCEP (0.82); the coastal Niño 1+2 box is the hardest of the six, 0.42 for the consensus.
Index
Raise a system
The consensus
ten systems, one vote each, six months aheadNiño 3.4 5°N–5°S, 170°W–120°W
ERA5 observedConsensus — mean of the ten system means10–90 % of all members, each system weighted equallyEach system's meanThe system you pick±0.5 °C threshold
Relative Niño 3.4 Niño 3.4 minus the 20°N–20°S mean (not rescaled)
ERA5 observedConsensus — mean of the ten system means10–90 % of all members, each system weighted equallyEach system's meanThe system you pick±0.5 °C threshold
Niño 3 5°N–5°S, 150°W–90°W
ERA5 observedConsensus — mean of the ten system means10–90 % of all members, each system weighted equallyEach system's meanThe system you pick±0.5 °C threshold
Niño 4 5°N–5°S, 160°E–150°W
ERA5 observedConsensus — mean of the ten system means10–90 % of all members, each system weighted equallyEach system's meanThe system you pick±0.5 °C threshold
Niño 1+2 0–10°S, 90°W–80°W
ERA5 observedConsensus — mean of the ten system means10–90 % of all members, each system weighted equallyEach system's meanThe system you pick±0.5 °C threshold
Indian Ocean Dipole DMI: 10°N–10°S 50–70°E minus 0–10°S 90–110°E
ERA5 observedConsensus — mean of the ten system means10–90 % of all members, each system weighted equallyEach system's meanThe system you pick±0.4 °C threshold
System by system
one shared axis · tap a panel to raise it abovethat system's 10–90 % of membersits meanconsensus, for referenceERA5
Where they disagree
every system's mean, coloured against the consensusColour: each system against the consensus that month, °C0.6+ cooler0.35–0.60.15–0.35within ±0.150.15–0.350.35–0.60.6+ warmer
| System, sorted by Dec ’26 | Oct ’26* | Nov ’26 | Dec ’26 | Jan ’27 | Feb ’27 | Mar ’27 |
|---|---|---|---|---|---|---|
| DWDGCFS2.1 · 50 | +3.9 | +4.2 | +4.5 | +4.6 | +4.5 | +4.2 |
| ECCC 5CanESM5.1 / GEM5.2-NEMO · 20 | +3.8 | +4.4 | +4.4 | +4.1 | +3.6 | +3.1 |
| BoMACCESS-S2 · 110 | +3.8 | +4.2 | +4.3 | +4.0 | +3.4 | +2.8 |
| Météo-FranceSystem 9 · 51 | +3.8 | +4.1 | +4.2 | +3.8 | +3.0 | +2.0 |
| CMCCSPSv4 · 50 | +3.5 | +3.9 | +4.2 | +4.3 | +4.0 | +3.5 |
| ECMWFSEAS5.1 · 51 | +3.6 | +4.0 | +4.2 | +4.1 | +3.8 | +3.5 |
| ECCC 4CanESM5.1 / GEM5.2-NEMO · 20 | +3.5 | +3.9 | +4.1 | +3.8 | +3.2 | +2.5 |
| JMACPS3/CPS4 · 150 | +3.4 | +3.8 | +4.0 | +3.9 | +3.7 | +3.2 |
| Met OfficeGloSea6 · 60 | +3.4 | +3.7 | +3.6 | +3.4 | +2.9 | +2.2 |
| NCEPCFSv2 · 120 | +3.2 | +3.5 | +3.4 | +2.8 | +2.4 | +1.8 |
| Consensusone vote per system | +3.59 | +3.98 | +4.08 | +3.89 | +3.46 | +2.86 |
| Spreadsd of the ten means | 0.22 | 0.26 | 0.33 | 0.48 | 0.58 | 0.71 |
Colour: each system against the consensus that month, °C0.6+ cooler0.35–0.60.15–0.35within ±0.150.15–0.350.35–0.60.6+ warmer
| System, sorted by Dec ’26 | Oct ’26* | Nov ’26 | Dec ’26 | Jan ’27 | Feb ’27 | Mar ’27 |
|---|---|---|---|---|---|---|
| DWDGCFS2.1 · 50 | +3.2 | +3.3 | +3.6 | +3.7 | +3.5 | +3.2 |
| ECCC 5CanESM5.1 / GEM5.2-NEMO · 20 | +2.8 | +3.3 | +3.2 | +2.9 | +2.5 | +2.0 |
| Météo-FranceSystem 9 · 51 | +2.8 | +3.0 | +3.0 | +2.7 | +1.9 | +1.0 |
| ECCC 4CanESM5.1 / GEM5.2-NEMO · 20 | +2.5 | +2.8 | +2.9 | +2.6 | +2.1 | +1.4 |
| ECMWFSEAS5.1 · 51 | +2.5 | +2.7 | +2.8 | +2.7 | +2.4 | +2.1 |
| CMCCSPSv4 · 50 | +2.4 | +2.7 | +2.8 | +2.8 | +2.5 | +2.1 |
| BoMACCESS-S2 · 110 | +2.5 | +2.8 | +2.7 | +2.4 | +1.9 | +1.4 |
| JMACPS3/CPS4 · 150 | +2.3 | +2.6 | +2.7 | +2.6 | +2.4 | +1.9 |
| Met OfficeGloSea6 · 60 | +2.4 | +2.5 | +2.4 | +2.2 | +1.7 | +1.1 |
| NCEPCFSv2 · 120 | +2.3 | +2.5 | +2.3 | +1.7 | +1.3 | +0.9 |
| Consensusone vote per system | +2.57 | +2.81 | +2.83 | +2.62 | +2.23 | +1.70 |
| Spreadsd of the ten means | 0.28 | 0.29 | 0.37 | 0.48 | 0.56 | 0.67 |
Colour: each system against the consensus that month, °C0.6+ cooler0.35–0.60.15–0.35within ±0.150.15–0.350.35–0.60.6+ warmer
| System, sorted by Dec ’26 | Oct ’26* | Nov ’26 | Dec ’26 | Jan ’27 | Feb ’27 | Mar ’27 |
|---|---|---|---|---|---|---|
| BoMACCESS-S2 · 110 | +4.5 | +4.9 | +5.1 | +4.7 | +3.8 | +3.2 |
| ECMWFSEAS5.1 · 51 | +4.4 | +4.6 | +4.8 | +4.6 | +4.2 | +3.8 |
| ECCC 5CanESM5.1 / GEM5.2-NEMO · 20 | +4.4 | +4.7 | +4.7 | +4.3 | +3.7 | +3.2 |
| CMCCSPSv4 · 50 | +4.1 | +4.5 | +4.7 | +4.6 | +4.1 | +3.8 |
| JMACPS3/CPS4 · 150 | +4.1 | +4.4 | +4.7 | +4.7 | +4.4 | +3.9 |
| Météo-FranceSystem 9 · 51 | +4.2 | +4.5 | +4.6 | +4.1 | +3.3 | +2.3 |
| ECCC 4CanESM5.1 / GEM5.2-NEMO · 20 | +4.0 | +4.3 | +4.4 | +4.0 | +3.4 | +2.7 |
| Met OfficeGloSea6 · 60 | +4.1 | +4.4 | +4.4 | +4.0 | +3.2 | +2.7 |
| NCEPCFSv2 · 120 | +3.6 | +4.1 | +4.2 | +3.9 | +3.3 | +2.5 |
| DWDGCFS2.1 · 50 | +3.6 | +3.7 | +3.9 | +3.8 | +3.8 | +3.7 |
| Consensusone vote per system | +4.11 | +4.41 | +4.55 | +4.27 | +3.71 | +3.19 |
| Spreadsd of the ten means | 0.27 | 0.33 | 0.33 | 0.33 | 0.39 | 0.56 |
Colour: each system against the consensus that month, °C0.6+ cooler0.35–0.60.15–0.35within ±0.150.15–0.350.35–0.60.6+ warmer
| System, sorted by Nov ’26 | Oct ’26* | Nov ’26 | Dec ’26 | Jan ’27 | Feb ’27 | Mar ’27 |
|---|---|---|---|---|---|---|
| CMCCSPSv4 · 50 | +1.9 | +2.4 | +2.7 | +2.9 | +2.9 | +2.5 |
| ECCC 5CanESM5.1 / GEM5.2-NEMO · 20 | +1.8 | +2.3 | +2.2 | +2.0 | +1.9 | +1.8 |
| DWDGCFS2.1 · 50 | +1.9 | +2.2 | +2.2 | +2.2 | +2.0 | +1.8 |
| ECMWFSEAS5.1 · 51 | +1.7 | +2.2 | +2.3 | +2.3 | +2.2 | +2.1 |
| Météo-FranceSystem 9 · 51 | +1.8 | +2.1 | +2.0 | +1.6 | +1.4 | +1.2 |
| BoMACCESS-S2 · 110 | +1.9 | +2.1 | +2.0 | +1.9 | +2.0 | +1.8 |
| JMACPS3/CPS4 · 150 | +1.7 | +2.0 | +2.0 | +2.0 | +1.8 | +1.6 |
| Met OfficeGloSea6 · 60 | +1.6 | +1.7 | +1.6 | +1.6 | +1.6 | +1.3 |
| ECCC 4CanESM5.1 / GEM5.2-NEMO · 20 | +1.5 | +1.6 | +1.7 | +1.7 | +1.7 | +1.4 |
| NCEPCFSv2 · 120 | +1.6 | +1.3 | +1.2 | +0.7 | +0.5 | +0.4 |
| Consensusone vote per system | +1.73 | +1.99 | +1.98 | +1.89 | +1.80 | +1.57 |
| Spreadsd of the ten means | 0.14 | 0.33 | 0.40 | 0.53 | 0.58 | 0.54 |
Colour: each system against the consensus that month, °C0.6+ cooler0.35–0.60.15–0.35within ±0.150.15–0.350.35–0.60.6+ warmer
| System, sorted by Oct ’26 | Oct ’26* | Nov ’26 | Dec ’26 | Jan ’27 | Feb ’27 | Mar ’27 |
|---|---|---|---|---|---|---|
| CMCCSPSv4 · 50 | +5.3 | +5.2 | +4.9 | +4.4 | +3.8 | +4.1 |
| BoMACCESS-S2 · 110 | +5.3 | +5.4 | +5.4 | +4.6 | +3.3 | +3.0 |
| ECMWFSEAS5.1 · 51 | +5.2 | +5.1 | +4.8 | +4.3 | +3.7 | +3.7 |
| JMACPS3/CPS4 · 150 | +5.1 | +5.0 | +4.9 | +4.5 | +3.8 | +3.4 |
| Met OfficeGloSea6 · 60 | +4.9 | +4.7 | +4.4 | +3.8 | +2.8 | +2.8 |
| ECCC 5CanESM5.1 / GEM5.2-NEMO · 20 | +4.8 | +4.2 | +4.1 | +3.7 | +2.8 | +2.4 |
| Météo-FranceSystem 9 · 51 | +4.2 | +3.9 | +3.7 | +3.5 | +3.0 | +3.0 |
| ECCC 4CanESM5.1 / GEM5.2-NEMO · 20 | +4.1 | +3.4 | +3.1 | +2.9 | +2.4 | +2.0 |
| NCEPCFSv2 · 120 | +3.9 | +3.7 | +3.8 | +3.5 | +2.9 | +2.3 |
| DWDGCFS2.1 · 50 | +2.4 | +2.4 | +2.4 | +2.1 | +1.5 | +1.4 |
| Consensusone vote per system | +4.52 | +4.29 | +4.16 | +3.71 | +3.00 | +2.80 |
| Spreadsd of the ten means | 0.86 | 0.90 | 0.87 | 0.75 | 0.67 | 0.76 |
Colour: each system against the consensus that month, °C0.3+ cooler0.18–0.30.08–0.18within ±0.080.08–0.180.18–0.30.3+ warmer
| System, sorted by Oct ’26 | Oct ’26* | Nov ’26 | Dec ’26 | Jan ’27 | Feb ’27 | Mar ’27 |
|---|---|---|---|---|---|---|
| ECCC 5CanESM5.1 / GEM5.2-NEMO · 20 | +1.3 | +1.2 | +0.8 | +0.4 | +0.1 | −0.1 |
| CMCCSPSv4 · 50 | +1.0 | +1.1 | +0.8 | +0.2 | −0.4 | −0.5 |
| ECCC 4CanESM5.1 / GEM5.2-NEMO · 20 | +1.0 | +0.9 | +0.6 | +0.4 | +0.1 | +0.1 |
| BoMACCESS-S2 · 110 | +1.0 | +1.0 | +0.7 | +0.2 | −0.1 | −0.2 |
| JMACPS3/CPS4 · 150 | +0.9 | +0.9 | +0.5 | +0.1 | −0.3 | −0.4 |
| Met OfficeGloSea6 · 60 | +0.9 | +0.7 | +0.6 | +0.2 | −0.1 | −0.2 |
| NCEPCFSv2 · 120 | +0.9 | +0.7 | +0.4 | +0.2 | 0.0 | −0.1 |
| ECMWFSEAS5.1 · 51 | +0.8 | +0.5 | +0.2 | 0.0 | −0.3 | −0.4 |
| DWDGCFS2.1 · 50 | +0.6 | +0.2 | +0.2 | 0.0 | −0.1 | −0.2 |
| Météo-FranceSystem 9 · 51 | +0.6 | +0.4 | +0.2 | 0.0 | −0.2 | −0.2 |
| Consensusone vote per system | +0.90 | +0.78 | +0.51 | +0.16 | −0.13 | −0.23 |
| Spreadsd of the ten means | 0.17 | 0.30 | 0.24 | 0.15 | 0.16 | 0.17 |
How much to trust it
reforecast 1993–2016 vs ERA5Score
| Correlation with ERA5, 1993–2016 | Lead 1 Oct ’26* | Lead 2 Nov ’26 | Lead 3 Dec ’26 | Lead 4 Jan ’27 | Lead 5 Feb ’27 | Lead 6 Mar ’27 |
|---|---|---|---|---|---|---|
| Consensusmean of system means | 0.99 | 0.99 | 0.98 | 0.97 | 0.94 | 0.93 |
| ECMWFSEAS5.1 | 0.99 | 0.99 | 0.98 | 0.97 | 0.96 | 0.95 |
| ECCC 5CanESM5.1 / GEM5.2-NEMO | 0.99 | 0.99 | 0.97 | 0.97 | 0.94 | 0.93 |
| DWDGCFS2.1 | 0.97 | 0.96 | 0.93 | 0.93 | 0.91 | 0.91 |
| JMACPS3/CPS4 | 0.99 | 0.97 | 0.96 | 0.96 | 0.93 | 0.91 |
| Météo-FranceSystem 9 | 0.99 | 0.98 | 0.97 | 0.96 | 0.94 | 0.90 |
| BoMACCESS-S2 | 0.99 | 0.98 | 0.97 | 0.96 | 0.93 | 0.90 |
| CMCCSPSv4 | 0.99 | 0.98 | 0.97 | 0.96 | 0.93 | 0.90 |
| Met OfficeGloSea6 | 0.99 | 0.99 | 0.98 | 0.97 | 0.92 | 0.88 |
| ECCC 4CanESM5.1 / GEM5.2-NEMO | 0.99 | 0.98 | 0.94 | 0.91 | 0.87 | 0.85 |
| NCEPCFSv2 | 0.94 | 0.94 | 0.92 | 0.88 | 0.82 | 0.82 |
| Root-mean-square error (°C), 1993–2016 | Lead 1 Oct ’26* | Lead 2 Nov ’26 | Lead 3 Dec ’26 | Lead 4 Jan ’27 | Lead 5 Feb ’27 | Lead 6 Mar ’27 |
|---|---|---|---|---|---|---|
| Consensusmean of system means | 0.11 | 0.20 | 0.29 | 0.30 | 0.36 | 0.35 |
| ECMWFSEAS5.1 | 0.13 | 0.22 | 0.28 | 0.29 | 0.34 | 0.35 |
| ECCC 5CanESM5.1 / GEM5.2-NEMO | 0.28 | 0.43 | 0.42 | 0.38 | 0.42 | 0.41 |
| DWDGCFS2.1 | 0.27 | 0.34 | 0.47 | 0.43 | 0.45 | 0.43 |
| JMACPS3/CPS4 | 0.16 | 0.32 | 0.38 | 0.40 | 0.49 | 0.57 |
| Météo-FranceSystem 9 | 0.17 | 0.26 | 0.32 | 0.36 | 0.39 | 0.39 |
| BoMACCESS-S2 | 0.15 | 0.25 | 0.31 | 0.34 | 0.41 | 0.43 |
| CMCCSPSv4 | 0.12 | 0.24 | 0.35 | 0.42 | 0.45 | 0.49 |
| Met OfficeGloSea6 | 0.12 | 0.20 | 0.28 | 0.32 | 0.43 | 0.43 |
| ECCC 4CanESM5.1 / GEM5.2-NEMO | 0.16 | 0.29 | 0.44 | 0.51 | 0.50 | 0.41 |
| NCEPCFSv2 | 0.38 | 0.46 | 0.54 | 0.59 | 0.61 | 0.48 |
| Correlation with ERA5, 1993–2016 | Lead 1 Oct ’26* | Lead 2 Nov ’26 | Lead 3 Dec ’26 | Lead 4 Jan ’27 | Lead 5 Feb ’27 | Lead 6 Mar ’27 |
|---|---|---|---|---|---|---|
| Consensusmean of system means | 0.99 | 0.98 | 0.97 | 0.96 | 0.93 | 0.91 |
| ECMWFSEAS5.1 | 0.99 | 0.98 | 0.97 | 0.97 | 0.95 | 0.93 |
| ECCC 5CanESM5.1 / GEM5.2-NEMO | 0.99 | 0.98 | 0.96 | 0.96 | 0.92 | 0.90 |
| DWDGCFS2.1 | 0.96 | 0.95 | 0.91 | 0.92 | 0.88 | 0.89 |
| BoMACCESS-S2 | 0.99 | 0.97 | 0.96 | 0.96 | 0.91 | 0.88 |
| Météo-FranceSystem 9 | 0.99 | 0.98 | 0.97 | 0.96 | 0.92 | 0.88 |
| CMCCSPSv4 | 0.99 | 0.97 | 0.95 | 0.94 | 0.88 | 0.85 |
| JMACPS3/CPS4 | 0.98 | 0.96 | 0.94 | 0.94 | 0.90 | 0.85 |
| Met OfficeGloSea6 | 0.99 | 0.98 | 0.97 | 0.96 | 0.90 | 0.85 |
| ECCC 4CanESM5.1 / GEM5.2-NEMO | 0.99 | 0.97 | 0.93 | 0.88 | 0.85 | 0.81 |
| NCEPCFSv2 | 0.95 | 0.94 | 0.92 | 0.89 | 0.81 | 0.80 |
| Root-mean-square error (°C), 1993–2016 | Lead 1 Oct ’26* | Lead 2 Nov ’26 | Lead 3 Dec ’26 | Lead 4 Jan ’27 | Lead 5 Feb ’27 | Lead 6 Mar ’27 |
|---|---|---|---|---|---|---|
| Consensusmean of system means | 0.12 | 0.21 | 0.28 | 0.27 | 0.32 | 0.30 |
| ECMWFSEAS5.1 | 0.12 | 0.21 | 0.26 | 0.25 | 0.28 | 0.28 |
| ECCC 5CanESM5.1 / GEM5.2-NEMO | 0.25 | 0.39 | 0.39 | 0.32 | 0.36 | 0.37 |
| DWDGCFS2.1 | 0.28 | 0.34 | 0.45 | 0.40 | 0.41 | 0.41 |
| BoMACCESS-S2 | 0.12 | 0.24 | 0.29 | 0.29 | 0.34 | 0.34 |
| Météo-FranceSystem 9 | 0.14 | 0.24 | 0.30 | 0.30 | 0.32 | 0.30 |
| CMCCSPSv4 | 0.14 | 0.26 | 0.34 | 0.37 | 0.41 | 0.41 |
| JMACPS3/CPS4 | 0.18 | 0.31 | 0.37 | 0.36 | 0.41 | 0.47 |
| Met OfficeGloSea6 | 0.13 | 0.21 | 0.27 | 0.29 | 0.38 | 0.37 |
| ECCC 4CanESM5.1 / GEM5.2-NEMO | 0.17 | 0.29 | 0.42 | 0.46 | 0.42 | 0.33 |
| NCEPCFSv2 | 0.31 | 0.38 | 0.46 | 0.47 | 0.49 | 0.37 |
| Correlation with ERA5, 1993–2016 | Lead 1 Oct ’26* | Lead 2 Nov ’26 | Lead 3 Dec ’26 | Lead 4 Jan ’27 | Lead 5 Feb ’27 | Lead 6 Mar ’27 |
|---|---|---|---|---|---|---|
| Consensusmean of system means | 0.99 | 0.99 | 0.98 | 0.97 | 0.90 | 0.82 |
| ECMWFSEAS5.1 | 1.00 | 0.99 | 0.98 | 0.97 | 0.92 | 0.87 |
| ECCC 5CanESM5.1 / GEM5.2-NEMO | 0.99 | 0.99 | 0.98 | 0.96 | 0.89 | 0.86 |
| ECCC 4CanESM5.1 / GEM5.2-NEMO | 0.98 | 0.97 | 0.95 | 0.94 | 0.88 | 0.82 |
| JMACPS3/CPS4 | 0.99 | 0.97 | 0.97 | 0.96 | 0.90 | 0.82 |
| DWDGCFS2.1 | 0.96 | 0.95 | 0.91 | 0.91 | 0.84 | 0.80 |
| CMCCSPSv4 | 0.99 | 0.99 | 0.97 | 0.96 | 0.91 | 0.77 |
| Met OfficeGloSea6 | 1.00 | 0.99 | 0.98 | 0.97 | 0.87 | 0.77 |
| BoMACCESS-S2 | 0.99 | 0.98 | 0.97 | 0.97 | 0.89 | 0.75 |
| Météo-FranceSystem 9 | 0.99 | 0.97 | 0.97 | 0.97 | 0.89 | 0.74 |
| NCEPCFSv2 | 0.93 | 0.92 | 0.91 | 0.87 | 0.77 | 0.70 |
| Root-mean-square error (°C), 1993–2016 | Lead 1 Oct ’26* | Lead 2 Nov ’26 | Lead 3 Dec ’26 | Lead 4 Jan ’27 | Lead 5 Feb ’27 | Lead 6 Mar ’27 |
|---|---|---|---|---|---|---|
| Consensusmean of system means | 0.14 | 0.23 | 0.29 | 0.27 | 0.43 | 0.50 |
| ECMWFSEAS5.1 | 0.11 | 0.21 | 0.28 | 0.31 | 0.47 | 0.48 |
| ECCC 5CanESM5.1 / GEM5.2-NEMO | 0.18 | 0.24 | 0.33 | 0.37 | 0.48 | 0.45 |
| ECCC 4CanESM5.1 / GEM5.2-NEMO | 0.24 | 0.37 | 0.47 | 0.40 | 0.42 | 0.40 |
| JMACPS3/CPS4 | 0.18 | 0.31 | 0.35 | 0.43 | 0.64 | 0.77 |
| DWDGCFS2.1 | 0.52 | 0.68 | 0.76 | 0.59 | 0.47 | 0.43 |
| CMCCSPSv4 | 0.12 | 0.22 | 0.33 | 0.34 | 0.42 | 0.64 |
| Met OfficeGloSea6 | 0.10 | 0.19 | 0.29 | 0.31 | 0.49 | 0.59 |
| BoMACCESS-S2 | 0.17 | 0.27 | 0.33 | 0.30 | 0.48 | 0.65 |
| Météo-FranceSystem 9 | 0.18 | 0.30 | 0.36 | 0.32 | 0.49 | 0.59 |
| NCEPCFSv2 | 0.42 | 0.51 | 0.59 | 0.64 | 0.69 | 0.65 |
| Correlation with ERA5, 1993–2016 | Lead 1 Oct ’26* | Lead 2 Nov ’26 | Lead 3 Dec ’26 | Lead 4 Jan ’27 | Lead 5 Feb ’27 | Lead 6 Mar ’27 |
|---|---|---|---|---|---|---|
| Consensusmean of system means | 0.99 | 0.98 | 0.97 | 0.94 | 0.92 | 0.93 |
| ECMWFSEAS5.1 | 0.98 | 0.98 | 0.96 | 0.93 | 0.92 | 0.93 |
| Met OfficeGloSea6 | 0.99 | 0.99 | 0.98 | 0.94 | 0.92 | 0.92 |
| ECCC 5CanESM5.1 / GEM5.2-NEMO | 0.99 | 0.98 | 0.96 | 0.93 | 0.92 | 0.91 |
| BoMACCESS-S2 | 0.98 | 0.98 | 0.97 | 0.93 | 0.91 | 0.91 |
| Météo-FranceSystem 9 | 0.97 | 0.98 | 0.97 | 0.94 | 0.93 | 0.91 |
| JMACPS3/CPS4 | 0.98 | 0.97 | 0.95 | 0.92 | 0.88 | 0.90 |
| CMCCSPSv4 | 0.98 | 0.97 | 0.94 | 0.91 | 0.90 | 0.90 |
| DWDGCFS2.1 | 0.95 | 0.96 | 0.91 | 0.88 | 0.87 | 0.85 |
| ECCC 4CanESM5.1 / GEM5.2-NEMO | 0.98 | 0.97 | 0.95 | 0.89 | 0.84 | 0.81 |
| NCEPCFSv2 | 0.95 | 0.93 | 0.89 | 0.81 | 0.73 | 0.70 |
| Root-mean-square error (°C), 1993–2016 | Lead 1 Oct ’26* | Lead 2 Nov ’26 | Lead 3 Dec ’26 | Lead 4 Jan ’27 | Lead 5 Feb ’27 | Lead 6 Mar ’27 |
|---|---|---|---|---|---|---|
| Consensusmean of system means | 0.16 | 0.17 | 0.24 | 0.33 | 0.35 | 0.29 |
| ECMWFSEAS5.1 | 0.17 | 0.21 | 0.29 | 0.38 | 0.37 | 0.30 |
| Met OfficeGloSea6 | 0.13 | 0.15 | 0.19 | 0.31 | 0.35 | 0.31 |
| ECCC 5CanESM5.1 / GEM5.2-NEMO | 0.23 | 0.35 | 0.35 | 0.38 | 0.39 | 0.35 |
| BoMACCESS-S2 | 0.14 | 0.18 | 0.23 | 0.34 | 0.38 | 0.33 |
| Météo-FranceSystem 9 | 0.25 | 0.25 | 0.30 | 0.35 | 0.34 | 0.32 |
| JMACPS3/CPS4 | 0.14 | 0.22 | 0.28 | 0.38 | 0.43 | 0.36 |
| CMCCSPSv4 | 0.20 | 0.27 | 0.40 | 0.49 | 0.47 | 0.42 |
| DWDGCFS2.1 | 0.39 | 0.36 | 0.51 | 0.53 | 0.50 | 0.45 |
| ECCC 4CanESM5.1 / GEM5.2-NEMO | 0.15 | 0.20 | 0.27 | 0.43 | 0.50 | 0.45 |
| NCEPCFSv2 | 0.25 | 0.34 | 0.46 | 0.59 | 0.64 | 0.55 |
| Correlation with ERA5, 1993–2016 | Lead 1 Oct ’26* | Lead 2 Nov ’26 | Lead 3 Dec ’26 | Lead 4 Jan ’27 | Lead 5 Feb ’27 | Lead 6 Mar ’27 |
|---|---|---|---|---|---|---|
| Consensusmean of system means | 0.98 | 0.97 | 0.95 | 0.85 | 0.62 | 0.42 |
| ECCC 5CanESM5.1 / GEM5.2-NEMO | 0.98 | 0.96 | 0.95 | 0.86 | 0.71 | 0.54 |
| ECMWFSEAS5.1 | 0.98 | 0.98 | 0.97 | 0.84 | 0.64 | 0.52 |
| ECCC 4CanESM5.1 / GEM5.2-NEMO | 0.95 | 0.90 | 0.90 | 0.82 | 0.62 | 0.48 |
| BoMACCESS-S2 | 0.96 | 0.96 | 0.95 | 0.88 | 0.69 | 0.48 |
| Met OfficeGloSea6 | 0.98 | 0.98 | 0.96 | 0.83 | 0.65 | 0.44 |
| CMCCSPSv4 | 0.98 | 0.97 | 0.93 | 0.82 | 0.66 | 0.43 |
| JMACPS3/CPS4 | 0.97 | 0.95 | 0.95 | 0.84 | 0.62 | 0.43 |
| Météo-FranceSystem 9 | 0.97 | 0.95 | 0.89 | 0.76 | 0.44 | 0.22 |
| NCEPCFSv2 | 0.87 | 0.80 | 0.82 | 0.71 | 0.36 | 0.20 |
| DWDGCFS2.1 | 0.77 | 0.77 | 0.77 | 0.62 | 0.46 | 0.18 |
| Root-mean-square error (°C), 1993–2016 | Lead 1 Oct ’26* | Lead 2 Nov ’26 | Lead 3 Dec ’26 | Lead 4 Jan ’27 | Lead 5 Feb ’27 | Lead 6 Mar ’27 |
|---|---|---|---|---|---|---|
| Consensusmean of system means | 0.36 | 0.41 | 0.40 | 0.51 | 0.60 | 0.81 |
| ECCC 5CanESM5.1 / GEM5.2-NEMO | 0.27 | 0.41 | 0.40 | 0.50 | 0.53 | 0.73 |
| ECMWFSEAS5.1 | 0.24 | 0.26 | 0.32 | 0.55 | 0.66 | 0.80 |
| ECCC 4CanESM5.1 / GEM5.2-NEMO | 0.47 | 0.68 | 0.67 | 0.59 | 0.58 | 0.76 |
| BoMACCESS-S2 | 0.36 | 0.35 | 0.35 | 0.46 | 0.55 | 0.78 |
| Met OfficeGloSea6 | 0.25 | 0.29 | 0.34 | 0.56 | 0.58 | 0.81 |
| CMCCSPSv4 | 0.29 | 0.31 | 0.45 | 0.57 | 0.60 | 0.85 |
| JMACPS3/CPS4 | 0.34 | 0.41 | 0.36 | 0.59 | 0.74 | 0.92 |
| Météo-FranceSystem 9 | 0.41 | 0.44 | 0.56 | 0.76 | 0.94 | 1.14 |
| NCEPCFSv2 | 0.69 | 0.77 | 0.68 | 0.70 | 0.86 | 0.99 |
| DWDGCFS2.1 | 0.88 | 0.89 | 0.86 | 0.77 | 0.66 | 0.86 |
| Correlation with ERA5, 1993–2016 | Lead 1 Oct ’26* | Lead 2 Nov ’26 | Lead 3 Dec ’26 | Lead 4 Jan ’27 | Lead 5 Feb ’27 | Lead 6 Mar ’27 |
|---|---|---|---|---|---|---|
| Consensusmean of system means | 0.97 | 0.90 | 0.80 | 0.58 | 0.34 | 0.31 |
| Met OfficeGloSea6 | 0.95 | 0.87 | 0.78 | 0.57 | 0.47 | 0.44 |
| BoMACCESS-S2 | 0.91 | 0.78 | 0.65 | 0.36 | 0.12 | 0.40 |
| JMACPS3/CPS4 | 0.95 | 0.87 | 0.78 | 0.35 | 0.03 | 0.32 |
| CMCCSPSv4 | 0.97 | 0.91 | 0.74 | 0.57 | 0.33 | 0.30 |
| ECCC 5CanESM5.1 / GEM5.2-NEMO | 0.98 | 0.91 | 0.81 | 0.64 | 0.36 | 0.27 |
| Météo-FranceSystem 9 | 0.94 | 0.88 | 0.66 | 0.16 | 0.24 | 0.16 |
| NCEPCFSv2 | 0.93 | 0.90 | 0.64 | 0.16 | -0.06 | 0.16 |
| ECMWFSEAS5.1 | 0.97 | 0.91 | 0.79 | 0.57 | 0.21 | 0.14 |
| DWDGCFS2.1 | 0.91 | 0.80 | 0.24 | -0.14 | 0.15 | 0.06 |
| ECCC 4CanESM5.1 / GEM5.2-NEMO | 0.96 | 0.89 | 0.77 | 0.54 | 0.26 | -0.03 |
| Root-mean-square error (°C), 1993–2016 | Lead 1 Oct ’26* | Lead 2 Nov ’26 | Lead 3 Dec ’26 | Lead 4 Jan ’27 | Lead 5 Feb ’27 | Lead 6 Mar ’27 |
|---|---|---|---|---|---|---|
| Consensusmean of system means | 0.22 | 0.30 | 0.25 | 0.25 | 0.36 | 0.32 |
| Met OfficeGloSea6 | 0.27 | 0.36 | 0.32 | 0.26 | 0.34 | 0.31 |
| BoMACCESS-S2 | 0.37 | 0.43 | 0.34 | 0.29 | 0.38 | 0.31 |
| JMACPS3/CPS4 | 0.29 | 0.33 | 0.26 | 0.29 | 0.40 | 0.33 |
| CMCCSPSv4 | 0.22 | 0.37 | 0.42 | 0.26 | 0.36 | 0.33 |
| ECCC 5CanESM5.1 / GEM5.2-NEMO | 0.17 | 0.29 | 0.26 | 0.24 | 0.35 | 0.32 |
| Météo-FranceSystem 9 | 0.30 | 0.35 | 0.32 | 0.30 | 0.37 | 0.33 |
| NCEPCFSv2 | 0.33 | 0.34 | 0.34 | 0.31 | 0.41 | 0.33 |
| ECMWFSEAS5.1 | 0.20 | 0.28 | 0.26 | 0.25 | 0.37 | 0.34 |
| DWDGCFS2.1 | 0.42 | 0.42 | 0.41 | 0.35 | 0.38 | 0.35 |
| ECCC 4CanESM5.1 / GEM5.2-NEMO | 0.26 | 0.32 | 0.26 | 0.26 | 0.37 | 0.35 |
Darker is better in both tables: a higher correlation, a smaller error. Correlation asks whether a system placed each of the 24 reforecast years on the right side of normal and by roughly the right amount relative to the others; RMSE is the typical size of the miss in °C. The consensus is scored as the mean of the system means, the same way it is drawn.
All six at a glance
the consensus for every index| Index — consensus °C, with its reforecast correlation | Oct ’26* | Nov ’26 | Dec ’26 | Jan ’27 | Feb ’27 | Mar ’27 |
|---|---|---|---|---|---|---|
| Niño 3.45°N–5°S, 170°W–120°W | +3.6r 0.99 | +4.0r 0.99 | +4.1r 0.98 | +3.9r 0.97 | +3.5r 0.94 | +2.9r 0.93 |
| Relative Niño 3.4Niño 3.4 minus the 20°N–20°S mean (not rescaled) | +2.6r 0.99 | +2.8r 0.98 | +2.8r 0.97 | +2.6r 0.96 | +2.2r 0.93 | +1.7r 0.91 |
| Niño 35°N–5°S, 150°W–90°W | +4.1r 0.99 | +4.4r 0.99 | +4.6r 0.98 | +4.3r 0.97 | +3.7r 0.90 | +3.2r 0.82 |
| Niño 45°N–5°S, 160°E–150°W | +1.7r 0.99 | +2.0r 0.98 | +2.0r 0.97 | +1.9r 0.94 | +1.8r 0.92 | +1.6r 0.93 |
| Niño 1+20–10°S, 90°W–80°W | +4.5r 0.98 | +4.3r 0.97 | +4.2r 0.95 | +3.7r 0.85 | +3.0r 0.62 | +2.8r 0.42 |
| Indian Ocean DipoleDMI: 10°N–10°S 50–70°E minus 0–10°S 90–110°E | +0.9r 0.97 | +0.8r 0.90 | +0.5r 0.80 | +0.2r 0.58 | −0.1r 0.34 | −0.2r 0.31 |
The Indian Ocean Dipole runs smaller than the Niño boxes, so its row is coloured on its own scale — ±0.4, ±0.8, ±1.2 °C rather than ±0.5, ±1.5, ±3.
How to read this
four things it is notBaseline
Each system against itself
Every anomaly is taken against that system's own 1993–2016 reforecast at the same lead, so a model that runs its Pacific warm or cold does not carry that into the forecast. ERA5 uses the same years. A different baseline — CPC's 1991–2020 ONI, for one — gives different numbers for the same ocean.
Weighting
One system, one vote
JMA flies 155 members and each ECCC system 20. Pooling members would hand the answer to whichever centre runs the most; the consensus instead averages the ten system means, and the shaded band weights every system equally.
Relative index
Not rescaled
Relative Niño 3.4 here is Niño 3.4 minus the 20°N–20°S mean, as a plain difference. CPC's RONI then rescales it to the traditional index's variance; this page does not, so its numbers run smaller than RONI's and it says so rather than borrowing the name.
Reach
Six months, not twelve
The Copernicus systems stop at six months. For what comes after — how the event decays and whether the Pacific swings cool — NOAA's SFS runs twelve, with its skill measured against persistence, on route 15.