Prolepsia forecastYes / no
Will 2027 be the warmest calendar year on record for global mean surface air temperature in the Copernicus ERA5 dataset?
As of 17 November 2026, Prolepsia puts the chance of yes at 95%.
95%chance of yes
How Prolepsia reasons
The report Prolepsia wrote with this forecast.
How forecasts are made · Track record
TL;DR
The forecast is 95.5% Yes and 4.5% No that 2027 will be the single warmest calendar year in ERA5. Strong late-2026 warmth, El Niño’s delayed global effect and continued background warming favor a record; the main uncertainty is how much warmth persists across the full year.
Key evidence
- The benchmark is high, but already close. The existing record is 2024, at approximately 15.10°C, or 0.72°C above 1991–2020 (Copernicus). The following year ranked third, 0.13°C cooler (2025 assessment). Copernicus’s November assessment gives 2026 about a 20% chance of matching or exceeding 2024, so 2027 must beat whichever year finishes higher (2026 outlook).
- Warmth is arriving at the right time for 2027. October’s ERA5 anomaly reached 0.904°C above 1991–2020, and November 1–15 averaged approximately 1.067°C, with the final day preliminary (monthly bulletin; daily series). These are short-period observations, not annual forecasts, but they establish a warm starting point.
- The Pacific signal extends into next year. NOAA assigns a 79% probability to strong or very strong El Niño in January–March 2027 (strength outlook). The global temperature response typically peaks after El Niño develops, favoring the following calendar year (WMO).
Uncertainties
A strong winter does not guarantee a record annual mean. NOAA gives only a 6% probability of El Niño in June–August, while GFDL also anticipates weakening toward neutral conditions in late spring or summer (NOAA; GFDL). Rapid cooling, a weaker global response, or cooler conditions outside the Pacific could reduce the annual result.
The event’s unusual strength also pushes temperature relationships beyond familiar historical conditions. Published projections for a six-dataset average support the warming direction but are not interchangeable with an ERA5-specific forecast (Carbon Brief). An exceptional cooling shock remains another failure route. Resolution requires a strictly higher annual ERA5 value: neither a monthly record nor a genuine annual tie is enough.
Read the full report
TL;DR
The forecast is 95.5% Yes and 4.5% No that 2027 will be the single warmest calendar year in ERA5. Exceptionally warm late-2026 conditions, El Niño’s delayed global effect and continued background warming make a new record the clear expectation. The decisive risk is weaker full-year warming, not whether individual months set records.
Context
As of November 17, 2026, the completed-year record belongs to 2024. Its global mean surface air temperature was approximately 15.10°C, equivalent to 0.72°C above 1991–2020 and 1.60°C above the estimated pre-industrial baseline (Copernicus). The next year ranked third, 0.13°C below that record (2025 assessment). The unfinished 2026 is most likely to rank second, but it remains a possible new benchmark (November assessment).
The question concerns the January–December global, area-weighted mean of near-surface air temperature over land and ocean. It does not concern sea-surface temperature alone, a monthly maximum or a rolling annual peak. The first complete official ERA5 assessment for 2027 determines the result, using the highest reasonably available precision. A genuine tie fails; a positive record margin qualifies even when rounded published values look identical.
Evidence
The historical backbone is the tendency for global warmth to peak in the year after a strong El Niño develops. The prominent decay years 1998, 2016 and 2024 all set global temperature records. Copernicus discusses the earlier pattern, and its subsequent assessment confirms that 2024 exceeded 2023 by 0.12°C (2023 highlights; 2024 highlights). These episodes support the timing argument. They are a small, selected reference class, not a universal rule.
The counterexample matters too. The strong El Niño decay year 1983 did not establish a new record in the historical comparisons examined here. It followed the El Chichón eruption, which complicates its use as an ordinary ENSO analogue (volcanic-response study). The lesson is that El Niño raises global temperature without controlling every other influence. Historical record counts also depend on the temperature dataset and whether a near-tie is judged from rounded or underlying values.
The physical timing is well supported. WMO describes El Niño as typically adding approximately 0.1–0.2°C to annual global temperature, with its largest effect in the year after development (WMO assessment). A study covering 47 June–May years associates a 1°C December Niño-3.4 anomaly with approximately 0.03°C of global warming in the preceding June–August, but 0.11°C in the following February (ENSO-conditioned temperature study). This supports early-2027 warmth. It does not justify applying February’s response to every month.
Continued human-induced warming strengthens that starting position. The Indicators of Global Climate Change assessment estimates warming during 2016–2025 at 0.27°C per decade, with an assessed range of 0.2–0.4°C per decade (assessment). That is a background trend, not an ERA5 forecast for a particular year. Its relevance is that the climate has continued warming since the existing record, so the Pacific does not have to supply the entire difference.
The observed temperature path supplies the strongest immediate evidence. The full January–October 2026 monthly sequence shows relatively moderate warmth early in the year, followed by a sharp rise. Values below are ERA5 global surface-air-temperature anomalies relative to 1991–2020 (Copernicus monthly series).
| Month in 2026 | Anomaly, °C |
|---|---|
| January | 0.511 |
| February | 0.531 |
| March | 0.530 |
| April | 0.523 |
| May | 0.550 |
| June | 0.559 |
| July | 0.671 |
| August | 0.846 |
| September | 0.824 |
| October | 0.904 |
October was the warmest October on record (Copernicus bulletin). November 1–15 then averaged approximately 1.067°C above the daily climatology; November 15 remained preliminary (Climate Pulse daily series). These observations show that warmth is already present, rather than resting entirely on a distant Pacific forecast. They do not establish the rest of November, December or any part of the 2027 annual mean.
The benchmark remains uncertain. Copernicus’s November 11 assessment, based on January–October observations, assigns approximately a 20% probability to 2026 matching or exceeding 2024. Its 95% prediction interval is 0.68–0.73°C above 1991–2020 (official outlook). Later warmth pushes in two directions: it supports persistence into 2027, but it also raises the chance that 2026 sets a higher hurdle. The correct comparison is with both years, not just a fixed 2024 value.
The winter ENSO outlook adds a strong forward signal. NOAA assigns a 79% probability to strong or very strong El Niño in January–March 2027. By June–August, however, its probabilities are 6% El Niño, 61.0% neutral and 33.0% La Niña (NOAA strength outlook). GFDL’s November 16 outlook also anticipates weakening toward neutral conditions in late spring or summer (GFDL update). The evidence favors a powerful early-year influence, not year-round El Niño.
There is also a meaningful downgrade behind the strong-event headlines. The probability of an October–December relative index of at least +2.5°C fell from 83% in October to approximately 60.0% in November (October discussion; November discussion). This does not erase the warming signal. It does weaken the case for treating the most extreme earlier projections as settled.
Other ocean basins provide support and caution. Mercator reports above-average sea-surface temperatures across 75% of the extra-polar ocean in October. Its North Atlantic mean was nevertheless below October 2023 and slightly below October 2024 (ocean bulletin). Warmth is widespread, but not every contributor to the existing record is stronger now. Ocean temperatures inform the outlook; they are not the air-temperature measure used for resolution.
Published assessments broadly support the warming direction. The Met Office describes 2027 as very likely to replace 2024, while Carbon Brief’s analysis also favors a new record (Met Office; Carbon Brief). But the latter targets a six-dataset composite and explicitly extrapolates beyond historical ENSO strength. Agreement on the direction is useful; it is not proof that the exact ERA5 outcome is assured.
What's non-obvious
El Niño does not need to last all year for 2027 to set a record. The global response lags the Pacific, and warm early months count toward the annual mean after the tropical Pacific begins cooling. This is why a spring decline is compatible with the forecast. The opposing mistake is to take a strong winter index and assume its peak global effect persists unchanged through December.
The other subtlety is that the current record already includes El Niño warmth and exceptional conditions elsewhere. A stronger Pacific event cannot simply be added in full to the 2024 annual value. Dataset choice matters too: the Carbon Brief analysis identifies ERA5’s unusually warm recent benchmark as harder to beat than some other temperature records (analysis). The case for YES rests on the combination of timing, observed late-year warmth and background warming—not on a single “record-strength El Niño” headline.
Uncertainties
The main uncertainty is the size and duration of the global response. An unusually strong Pacific event requires extrapolation beyond familiar historical conditions. Faster decay, an earlier or stronger La Niña, or weaker heat transfer into global air temperatures could leave the annual result below the record. Further winter observations and spring ENSO forecasts would narrow this gap.
Conditions outside the Pacific are another open question. Atlantic and Indian Ocean variability, clouds, aerosols and circulation contributed to recent warmth, but their persistence is not settled (Copernicus assessment). A large sulfur-rich volcanic eruption is a separate cooling mechanism. Historical Pinatubo evidence demonstrates substantial cooling, but does not establish the probability or annual effect of a future eruption (volcanic evidence).
Forecast accuracy also has limits. A review of annual global temperature forecasts reports errors around 0.09–0.13°C across several forecast categories and periods, with recent exceptional warmth exposing larger misses (forecast review). A verified ERA5-specific calibration for an event this strong, including first-assessment ranking differences, remains unavailable. That gap argues against treating a strong physical signal as certainty.
Finally, the completed 2026 annual value and all 2027 observations remain unknown. The final assessment must distinguish a strict record from a tie at underlying precision. Replacement by another reanalysis product would not itself imply NO: this question still requires ERA5, and insufficient official data would trigger annulment under its rules.
Forecast history
| Date | Forecast |
|---|---|
| 95% chance of yes |
Sources Prolepsia read
- Global Climate Highlights 2024 | Copernicusclimate.copernicus.eu
- ACP - Surface temperature response to the major volcanic eruptions in multiple reanalysis data setsacp.copernicus.org
- State of the climate in 2026 gives new meaning to phrase “off the charts”wmo.int
- ENSO-Conditioned Evolution of Global Mean Surface Temperatureagupubs.onlinelibrary.wiley.com
- ESSD - Indicators of Global Climate Change 2025: annual update of key indicators of the state of the climate system and human influenceessd.copernicus.org
- Climate Pulse daily seriessites.ecmwf.int
- Climate Prediction Center - Official NOAA CPC ENSO Strength Probabilitiescpc.ncep.noaa.gov
- November 2026 El Niño Predictions – Geophysical Fluid Dynamics Laboratorywww.gfdl.noaa.gov
- Ocean temperature bulletin - October 2026 - Mercator Ocean Internationalwww.mercator-ocean.eu
- An ‘unprecedented’ El Niño and its implications for the weather forecast - Met Officewww.metoffice.gov.uk
The question
This question asks whether the global mean surface air temperature for the calendar year 2027 (January 1 through December 31, 2027) will be the highest annual value recorded in the Copernicus Climate Change Service (C3S) ERA5 global atmospheric reanalysis dataset, considering all calendar years from 1940 through 2027, inclusive. As of November 17, 2026, the warmest calendar year in ERA5 is 2024, when global mean surface air temperature reached approximately 15.10°C, or 1.60°C above the estimated 1850–1900 pre-industrial average. The year 2025 ranked third warmest, approximately 0.13°C below 2024. According to Copernicus's November 11, 2026 assessment, 2026 is most likely to become the second-warmest year on record, with approximately a 20% probability of matching or exceeding the 2024 record. The forecast for 2026 reflects exceptionally warm conditions, including a developing strong El Niño. Consequently, the record that 2027 must surpass may be established in either 2024 or 2026, depending on the final 2026 annual temperature. The question concerns the global, area-weighted annual mean of near-surface (2-metre) air temperature over both land and ocean, as represented by ERA5. It does not concern ocean-only sea surface temperatures, individual monthly or daily temperature records, or temperature estimates from other datasets. Background sources: Copernicus Global Climate Highlights 2025 (https://climate.copernicus.eu/copernicus-2025-was-third-hottest-year-record) and Copernicus November 11, 2026 climate bulletin (https://climate.copernicus.eu/2026-and-4-year-global-temperature-average-track-exceed-15degc-above-pre-industrial-levels).
How it resolves
The question resolves YES if the annual global mean surface air temperature for January 1–December 31, 2027, as reported by the Copernicus Climate Change Service (C3S) using the ERA5 reanalysis dataset, is strictly greater than the annual global mean surface air temperature of every preceding calendar year in the ERA5 record, beginning in 1940 and ending in 2026. The question resolves NO if any preceding calendar year has an annual global mean surface air temperature equal to or greater than that of 2027. The primary resolution source will be the Copernicus Global Climate Highlights 2027 report, expected to be published in January 2028, or the equivalent official C3S annual global temperature assessment. If that report explicitly identifies 2027 as the single warmest calendar year in ERA5, the question resolves YES. If it identifies another year as warmer, or identifies 2027 as tied for the warmest year, the question resolves NO. If the report's ranking is ambiguous because temperatures are rounded, the underlying ERA5 annual global mean temperature values, using the same methodology and dataset version employed by C3S for its annual assessment, will determine the outcome. Comparisons will use the highest precision reasonably available rather than rounded published values. If the Global Climate Highlights 2027 report is not published, resolution will rely on another official C3S or European Centre for Medium-Range Weather Forecasts (ECMWF) publication that provides the relevant ERA5 annual global surface air temperature rankings or sufficient data to calculate them. The outcome will be determined using the first official, complete annual ERA5 assessment for 2027, without waiting for subsequent routine revisions. If no official source or sufficiently complete ERA5 data permitting a reliable comparison is available by December 31, 2028, the question will be annulled rather than resolved YES or NO.
Fine print
The ERA5 record begins in January 1940, so 'on record' refers specifically to calendar years covered by ERA5, not the full instrumental temperature record extending into the nineteenth century. Annual temperatures will be compared using a consistent ERA5 methodology across all years. Temperature anomalies relative to a common reference period, such as 1991–2020, may be used instead of absolute annual temperatures because subtracting the same reference climatology does not change annual rankings. The estimated 1850–1900 pre-industrial baseline is provided for context only and is not itself part of the ERA5 historical record. A tie at the precision of the underlying data does not qualify as a new record. A year that exceeds the previous record by any positive amount at the available underlying precision qualifies, even if the difference is smaller than 0.01°C. The question does not require 2027 to exceed any particular temperature anomaly, such as 1.5°C above pre-industrial levels; it requires only that 2027 be strictly warmer than every preceding year in ERA5. Temperatures recorded in 2028 or later are irrelevant to the determination. If ERA5 is superseded by a newer reanalysis product before resolution, the original ERA5 dataset remains the required source, provided the necessary values remain accessible. Later revisions to the underlying dataset or historical temperature rankings after the question has resolved will not change the outcome.
More public forecasts
What will Copernicus report as the 2027 global mean surface air temperature above the 1850-1900 pre-industrial level, in degrees Celsius (ERA5)?
1.74median
Prolepsia forecast
Will the United States carry out air or missile strikes on Iranian territory between 1 January and 31 December 2027?
88%chance of yes
Prolepsia forecast
Will the FAO Food Price Index average for calendar year 2027 be higher than its average for 2026?
76%chance of yes
Prolepsia forecast
Will the US FTC or DOJ sue to block more than five mergers in calendar year 2027?
23%chance of yes
Prolepsia forecast
Will US securities class action filings exceed 230 in calendar year 2027?
49%chance of yes
Prolepsia forecast
Will Brent crude average above $75 a barrel in calendar year 2027?
66%chance of yes
Prolepsia forecast