Will This Be the Warmest Year on Record?
Will 2026 be the warmest year on record? After 10 months the outlook gives it less than a 1% chance across this site's stations. The chance for each country and station, updated daily and tested on past years.
Data as of . Last revised (2026-11-17).
Key Facts
- From January to October 2026 the mean over this site's stations was +0.75 C (+1.35 F) against the 1991 to 2020 normal, the 4th warmest January to October of 76 years since 1951.
- The projected mean of the whole of 2026 is +0.76 C (+1.37 F), with a 5 to 95 percent range of +0.66 to +0.87 C (+1.19 to +1.57 F). That value would rank 2nd of 76 years, between 2nd and 5th across the range. The outlook gives 2026 less than a 1% chance of ending as the warmest year on record in this site's station series and more than a 99% chance of ending among the five warmest. The record is 2024 at +0.93 C (+1.67 F).
- 22 countries have at least 3 complete months of 2026 with 5 or more stations. Countries where a record is out of reach or did not happen: Thailand and South Africa.
- 4,471 stations in 45 countries have an outlook for 2026: each has 30 complete years among the 45 before it and 3 to 10 complete months of 2026. Of the 4,471 stations, 7 are already certain of a record, 1,580 have a chance of 50% or more and 172 are out of reach.
- The chances for 2026 add up to about 1,588 stations with a record year (36% of them). Averaged per 5 degree grid cell and weighted by area, over 250 cells, the mean chance is 21%.
- Of the 72,755 station outlooks of 2000 to 2025 that stated a 10 to 20 percent chance of a record year, 8,634 (12%) ended with one. In 2000 to 2025 the mean error of the projected yearly mean fell from 0.44 C (0.80 F) with 3 months complete to 0.17 C (0.31 F) with 11.
2026 So Far and Where It Is Heading
January to October completeFrom January to October 2026 the mean over this site's stations was +0.75 C (+1.35 F) against the 1991 to 2020 normal, the 4th warmest January to October of 76 years since 1951. The projected mean of the whole of 2026 is +0.76 C (+1.37 F), with a 5 to 95 percent range of +0.66 to +0.87 C (+1.19 to +1.57 F). That value would rank 2nd of 76 years, between 2nd and 5th across the range.
The outlook gives 2026 less than a 1% chance of ending as the warmest year on record in this site's station series and more than a 99% chance of ending among the five warmest. The record is 2024 at +0.93 C (+1.67 F).
These values are the mean over the weather stations of this site, each measured against its own 1991 to 2020 normal, averaged in 5 degree grid cells and weighted by area. They cover the land where the stations are and no ocean, so they are not a global mean temperature and should not be compared with one. The newest month used, October 2026, rests on 4,285 stations in 296 grid cells, 51% of the weight of all cells that ever reported that month. That month is provisional: late reports can still change it.
| Mean from January through | 2026 so far | Projected | 5th percentile | 95th percentile | Record (2024), whole year |
|---|---|---|---|---|---|
| January | +0.23 | n/a | n/a | n/a | +0.93 |
| February | +0.52 | n/a | n/a | n/a | +0.93 |
| March | +0.48 | n/a | n/a | n/a | +0.93 |
| April | +0.52 | n/a | n/a | n/a | +0.93 |
| May | +0.51 | n/a | n/a | n/a | +0.93 |
| June | +0.56 | n/a | n/a | n/a | +0.93 |
| July | +0.59 | n/a | n/a | n/a | +0.93 |
| August | +0.65 | n/a | n/a | n/a | +0.93 |
| September | +0.72 | n/a | n/a | n/a | +0.93 |
| October | +0.75 | +0.75 | +0.75 | +0.75 | +0.93 |
| November | n/a | +0.75 | +0.70 | +0.81 | +0.93 |
| December | n/a | +0.76 | +0.66 | +0.87 | +0.93 |
| Year | Yearly value | Projected 2026 | 5th percentile | 95th percentile | Record (2024) |
|---|---|---|---|---|---|
| 1951 | -0.86 | n/a | n/a | n/a | +0.93 |
| 1952 | -0.97 | n/a | n/a | n/a | +0.93 |
| 1953 | -0.55 | n/a | n/a | n/a | +0.93 |
| 1954 | -0.92 | n/a | n/a | n/a | +0.93 |
| 1955 | -0.99 | n/a | n/a | n/a | +0.93 |
| 1956 | -1.33 | n/a | n/a | n/a | +0.93 |
| 1957 | -0.85 | n/a | n/a | n/a | +0.93 |
| 1958 | -0.82 | n/a | n/a | n/a | +0.93 |
| 1959 | -0.71 | n/a | n/a | n/a | +0.93 |
| 1960 | -0.96 | n/a | n/a | n/a | +0.93 |
| 1961 | -0.65 | n/a | n/a | n/a | +0.93 |
| 1962 | -0.71 | n/a | n/a | n/a | +0.93 |
| 1963 | -0.75 | n/a | n/a | n/a | +0.93 |
| 1964 | -1.16 | n/a | n/a | n/a | +0.93 |
| 1965 | -1.09 | n/a | n/a | n/a | +0.93 |
| 1966 | -1.06 | n/a | n/a | n/a | +0.93 |
| 1967 | -0.82 | n/a | n/a | n/a | +0.93 |
| 1968 | -1.02 | n/a | n/a | n/a | +0.93 |
| 1969 | -1.25 | n/a | n/a | n/a | +0.93 |
| 1970 | -0.99 | n/a | n/a | n/a | +0.93 |
| 1971 | -0.94 | n/a | n/a | n/a | +0.93 |
| 1972 | -1.15 | n/a | n/a | n/a | +0.93 |
| 1973 | -0.64 | n/a | n/a | n/a | +0.93 |
| 1974 | -1.06 | n/a | n/a | n/a | +0.93 |
| 1975 | -0.69 | n/a | n/a | n/a | +0.93 |
| 1976 | -1.24 | n/a | n/a | n/a | +0.93 |
| 1977 | -0.74 | n/a | n/a | n/a | +0.93 |
| 1978 | -0.92 | n/a | n/a | n/a | +0.93 |
| 1979 | -0.81 | n/a | n/a | n/a | +0.93 |
| 1980 | -0.69 | n/a | n/a | n/a | +0.93 |
| 1981 | -0.32 | n/a | n/a | n/a | +0.93 |
| 1982 | -0.81 | n/a | n/a | n/a | +0.93 |
| 1983 | -0.43 | n/a | n/a | n/a | +0.93 |
| 1984 | -0.88 | n/a | n/a | n/a | +0.93 |
| 1985 | -0.92 | n/a | n/a | n/a | +0.93 |
| 1986 | -0.63 | n/a | n/a | n/a | +0.93 |
| 1987 | -0.64 | n/a | n/a | n/a | +0.93 |
| 1988 | -0.32 | n/a | n/a | n/a | +0.93 |
| 1989 | -0.42 | n/a | n/a | n/a | +0.93 |
| 1990 | -0.15 | n/a | n/a | n/a | +0.93 |
| 1991 | -0.32 | n/a | n/a | n/a | +0.93 |
| 1992 | -0.68 | n/a | n/a | n/a | +0.93 |
| 1993 | -0.59 | n/a | n/a | n/a | +0.93 |
| 1994 | -0.40 | n/a | n/a | n/a | +0.93 |
| 1995 | -0.09 | n/a | n/a | n/a | +0.93 |
| 1996 | -0.59 | n/a | n/a | n/a | +0.93 |
| 1997 | -0.26 | n/a | n/a | n/a | +0.93 |
| 1998 | +0.06 | n/a | n/a | n/a | +0.93 |
| 1999 | -0.13 | n/a | n/a | n/a | +0.93 |
| 2000 | -0.22 | n/a | n/a | n/a | +0.93 |
| 2001 | -0.07 | n/a | n/a | n/a | +0.93 |
| 2002 | +0.03 | n/a | n/a | n/a | +0.93 |
| 2003 | +0.02 | n/a | n/a | n/a | +0.93 |
| 2004 | -0.12 | n/a | n/a | n/a | +0.93 |
| 2005 | +0.18 | n/a | n/a | n/a | +0.93 |
| 2006 | +0.12 | n/a | n/a | n/a | +0.93 |
| 2007 | +0.27 | n/a | n/a | n/a | +0.93 |
| 2008 | -0.03 | n/a | n/a | n/a | +0.93 |
| 2009 | -0.07 | n/a | n/a | n/a | +0.93 |
| 2010 | +0.15 | n/a | n/a | n/a | +0.93 |
| 2011 | 0.00 | n/a | n/a | n/a | +0.93 |
| 2012 | +0.07 | n/a | n/a | n/a | +0.93 |
| 2013 | +0.13 | n/a | n/a | n/a | +0.93 |
| 2014 | +0.16 | n/a | n/a | n/a | +0.93 |
| 2015 | +0.51 | n/a | n/a | n/a | +0.93 |
| 2016 | +0.66 | n/a | n/a | n/a | +0.93 |
| 2017 | +0.52 | n/a | n/a | n/a | +0.93 |
| 2018 | +0.29 | n/a | n/a | n/a | +0.93 |
| 2019 | +0.47 | n/a | n/a | n/a | +0.93 |
| 2020 | +0.52 | n/a | n/a | n/a | +0.93 |
| 2021 | +0.35 | n/a | n/a | n/a | +0.93 |
| 2022 | +0.33 | n/a | n/a | n/a | +0.93 |
| 2023 | +0.75 | n/a | n/a | n/a | +0.93 |
| 2024 | +0.93 | n/a | n/a | n/a | +0.93 |
| 2025 | +0.76 | +0.76 | +0.76 | +0.76 | +0.93 |
| 2026 | n/a | +0.76 | +0.66 | +0.87 | +0.93 |
| Last complete month | Projected yearly value | 5th percentile | 95th percentile | Record (2024) |
|---|---|---|---|---|
| March | +0.62 | +0.48 | +0.88 | +0.93 |
| April | +0.61 | +0.48 | +0.86 | +0.93 |
| May | +0.60 | +0.49 | +0.85 | +0.93 |
| June | +0.63 | +0.49 | +0.85 | +0.93 |
| July | +0.65 | +0.51 | +0.85 | +0.93 |
| August | +0.70 | +0.55 | +0.86 | +0.93 |
| September | +0.74 | +0.59 | +0.87 | +0.93 |
| October | +0.76 | +0.66 | +0.87 | +0.93 |
After March 2026 the method gave it a 1% chance of a record; after October it gave it less than a 1% chance of a record. In the 30 reference years of this series, for each degree the months so far were above their trend, the rest of the year was 0.55 degrees above its own.
| Month | Against normal | Stations | Grid cells | Coverage | State |
|---|---|---|---|---|---|
| January 2026 | +0.23 C (+0.41 F) | 6,154 | 335 | 61% | Used |
| February 2026 | +0.81 C (+1.46 F) | 6,071 | 320 | 60% | Used |
| March 2026 | +0.41 C (+0.74 F) | 6,260 | 344 | 60% | Used |
| April 2026 | +0.64 C (+1.15 F) | 6,231 | 344 | 61% | Used |
| May 2026 | +0.44 C (+0.79 F) | 6,268 | 350 | 61% | Used |
| June 2026 | +0.82 C (+1.48 F) | 6,245 | 342 | 61% | Used |
| July 2026 | +0.78 C (+1.40 F) | 6,106 | 319 | 54% | Used |
| August 2026 | +1.06 C (+1.91 F) | 6,037 | 322 | 54% | Used |
| September 2026 | +1.24 C (+2.23 F) | 5,902 | 318 | 55% | Used |
| October 2026 | +1.06 C (+1.91 F) | 4,285 | 296 | 51% | Used, provisional |
Countries
5 or more stations22 countries have at least 3 complete months of 2026 with 5 or more stations. Countries where a record is out of reach or did not happen: Thailand and South Africa. Each country's value is the area mean of its own stations, and each has its own last complete month, because some networks report later than others. Values are degrees C against the 1991 to 2020 normal; multiply by 1.8 for degrees F.
| Country | Months complete | Year so far against normal, C | Projected year (5 to 95%), C | Chance of the warmest year | Likely rank | Record year (C) | Stations |
|---|---|---|---|---|---|---|---|
| France | 9 (to September) | +2.43 | +2.33 (+1.92 to +2.76) | over 99% | 1st of 76 | 2022 (+1.42) | 42 |
| Spain | 9 (to September) | +1.49 | +1.43 (+1.15 to +1.79) | 95% | 1st (1st to 2nd) of 76 | 2023 (+1.20) | 81 |
| United Kingdom | 9 (to September) | +1.32 | +1.27 (+0.83 to +1.68) | 95% | 1st (1st to 3rd) of 76 | 2025 (+0.97) | 54 |
| Austria | 9 (to September) | +1.61 | +1.52 (+1.10 to +2.09) | 73% | 1st (1st to 4th) of 74 | 2023 (+1.41) | 5 |
| Netherlands | 9 (to September) | +1.30 | +1.34 (+0.84 to +1.93) | 56% | 1st (1st to 7th) of 76 | 2024 (+1.30) | 30 |
| Switzerland | 6 (to June) | +1.69 | +1.51 (+0.72 to +2.22) | 53% | 1st (1st to 8th) of 70 | 2022 (+1.51) | 8 |
| Kazakhstan | 10 (to October) | +2.04 | +2.18 (+1.22 to +2.67) | 44% | 2nd (1st to 3rd) of 61 | 2025 (+2.20) | 24 |
| United States | 10 (to October) | +0.92 | +0.97 (+0.74 to +1.16) | 18% | 3rd (1st to 3rd) of 76 | 2016 (+1.06) | 3,306 |
| China | 6 (to June) | +0.78 | +0.84 (+0.46 to +1.16) | 14% | 4th (1st to 9th) of 70 | 2024 (+1.09) | 101 |
| Germany | 9 (to September) | +1.14 | +1.18 (+0.71 to +1.77) | 11% | 3rd (1st to 9th) of 76 | 2024 (+1.48) | 317 |
| Romania | 7 (to July) | +0.16 | +0.93 (+0.35 to +1.55) | 5% | 6th (2nd to 18th) of 75 | 2023 (+1.58) | 20 |
| Australia | 10 (to October) | +0.54 | +0.57 (+0.31 to +0.70) | 2% | 4th (2nd to 12th) of 76 | 2024 (+0.76) | 378 |
| Norway | 9 (to September) | +0.55 | +0.57 (-0.32 to +1.05) | 1% | 12th (3rd to 35th) of 76 | 2025 (+1.35) | 73 |
| Finland | 9 (to September) | +0.15 | +0.40 (-0.96 to +1.06) | under 1% | 13th (4th to 43rd) of 68 | 2020 (+1.77) | 116 |
| Sweden | 9 (to September) | +0.14 | +0.21 (-1.00 to +0.90) | under 1% | 21st (4th to 50th) of 76 | 2020 (+1.60) | 160 |
| Estonia | 8 (to August) | -0.38 | +0.33 (-0.81 to +1.07) | under 1% | 17th (5th to 43rd) of 76 | 2020 (+1.89) | 17 |
| Puerto Rico [United States] | 10 (to October) | +0.67 | +0.65 (+0.49 to +0.75) | under 1% | 2nd (2nd to 4th) of 74 | 2024 (+0.94) | 8 |
| Greenland [Denmark] | 9 (to September) | +0.69 | +0.83 (+0.37 to +1.36) | under 1% | 7th (3rd to 13th) of 61 | 2010 (+2.46) | 6 |
| Thailand | 10 (to October) | +0.42 | +0.44 (+0.27 to +0.63) | Out of reach | 7th (3rd to 9th) of 74 | 2024 (+1.06) | 12 |
| Canada | 10 (to October) | +0.42 | +0.41 (-0.05 to +0.78) | under 1% | 15th (7th to 21st) of 76 | 2010 (+1.59) | 320 |
| Ukraine | 9 (to September) | -0.56 | +0.08 (-0.57 to +0.66) | under 1% | 18th (4th to 33rd) of 68 | 2019 (+1.54) | 18 |
| South Africa | 10 (to October) | -0.10 | -0.12 (-0.23 to +0.11) | Out of reach | 45th (27th to 53rd) of 65 | 2024 (+0.67) | 7 |
Stations
4,471 stations4,471 stations in 45 countries have an outlook for 2026: each has 30 complete years among the 45 before it and 3 to 10 complete months of 2026.
Of the 4,471 stations, 7 are already certain of a record, 1,580 have a chance of 50% or more and 172 are out of reach.
The chances for 2026 add up to about 1,588 stations with a record year (36% of them). Averaged per 5 degree grid cell and weighted by area, over 250 cells, the mean chance is 21%.
The count of stations follows the network, not the climate. The country with the most of the 4,471 stations is United States with 3,132 (70%), followed by Germany (299), Australia (241), Canada (128), Sweden (114). That is why the area weighted share is given next to the count.
| Where the station stands | Stations | Share |
|---|---|---|
| Already certain of a record | 7 | 0% |
| Chance of 50% or more | 1,580 | 35% |
| Chance of 10 to 50% | 750 | 17% |
| Chance under 10% | 1,962 | 44% |
| Out of reach | 172 | 4% |
| Station | Latitude | Longitude | Chance of the warmest year |
|---|---|---|---|
| USC00021314 | 32.99 | -111.54 | over 99% |
| USC00023150 | 33.01 | -109.99 | over 99% |
| USC00028396 | 32.72 | -113.92 | over 99% |
| USC00053005 | 40.58 | -105.09 | over 99% |
| USC00053553 | 40.40 | -104.70 | over 99% |
| USC00053867 | 40.49 | -107.25 | over 99% |
| USC00058793 | 37.38 | -102.30 | over 99% |
| USC00265846 | 36.55 | -114.46 | over 99% |
| USC00293530 | 33.20 | -108.21 | over 99% |
| USC00295150 | 34.77 | -106.76 | over 99% |
| USC00411000 | 35.53 | -102.26 | over 99% |
| USC00416135 | 34.22 | -102.73 | over 99% |
| USC00486440 | 43.86 | -110.59 | over 99% |
| USC00487260 | 42.88 | -109.86 | over 99% |
| USS0005G05S | 42.28 | -105.58 | over 99% |
| USW00023158 | 33.62 | -114.71 | over 99% |
| USW00023169 | 36.07 | -115.16 | over 99% |
| USW00024018 | 41.15 | -104.81 | over 99% |
| USW00093026 | 31.46 | -109.61 | over 99% |
| USW00093042 | 36.03 | -102.55 | over 99% |
| ASN00070005 | -36.91 | 149.24 | over 99% |
| FRE00104052 | 48.45 | 0.11 | over 99% |
| FRE00104937 | 49.73 | -1.94 | over 99% |
| FRM00007207 | 47.29 | -3.22 | over 99% |
| UK000003808 | 50.22 | -5.32 | over 99% |
| USC00021050 | 35.14 | -114.57 | over 99% |
| USC00026796 | 34.57 | -112.43 | over 99% |
| USC00028162 | 34.13 | -109.28 | over 99% |
| USC00029359 | 35.24 | -112.19 | over 99% |
| USC00048606 | 39.53 | -121.16 | over 99% |
| USC00256200 | 41.13 | -101.72 | over 99% |
| USC00294426 | 32.62 | -106.74 | over 99% |
| USC00349404 | 36.58 | -98.88 | over 99% |
| USC00416210 | 31.96 | -96.69 | over 99% |
| USC00416776 | 35.55 | -100.97 | over 99% |
| USC00418566 | 33.48 | -100.88 | over 99% |
| USR0000NCUR | 38.76 | -115.41 | over 99% |
| USR0000NCYW | 38.28 | -114.76 | over 99% |
| USW00023023 | 31.94 | -102.19 | over 99% |
| USW00023050 | 35.04 | -106.62 | over 99% |
Stations already certain of their warmest year in 2026
| Station | Country | Months complete | Projected year (5 to 95%), C | Record so far, C | Chance of the warmest year | Likely rank |
|---|---|---|---|---|---|---|
| Idaho Falls 16 Se | US | 10 | 9.47 (8.75 to 9.94) | 7.53 in 2015 | Already certain | 1st of 51 |
| Northglenn | US | 10 | 14.06 (13.42 to 14.46) | 12.25 in 2025 | Already certain | 1st of 35 |
| Strontia Springs Dam | US | 10 | 11.34 (10.91 to 11.87) | 9.74 in 2012 | Already certain | 1st of 40 |
| Potter Valley Powerhouse | US | 10 | 17.64 (17.25 to 18.14) | 16.22 in 2015 | Already certain | 1st of 65 |
| Arches Np Hqs | US | 10 | 16.61 (16.26 to 16.98) | 15.39 in 2012 | Already certain | 1st of 38 |
| Havasu Arizona | US | 10 | 23.53 (23.19 to 23.90) | 22.35 in 2001 | Already certain | 1st of 31 |
| Verde Arizona | US | 10 | 18.95 (18.63 to 19.26) | 18.01 in 2025 | Already certain | 1st of 32 |
Highest chances among the stations still open
| Station | Country | Months complete | Projected year (5 to 95%), C | Record so far, C | Chance of the warmest year | Likely rank |
|---|---|---|---|---|---|---|
| Grand Lake 1 Nw | US | 9 | 6.22 (5.33 to 6.69) | 4.62 in 2024 | over 99% | 1st of 79 |
| Guthrie | US | 10 | 20.43 (19.68 to 20.86) | 18.88 in 2024 | over 99% | 1st of 50 |
| Trinidad Lake | US | 10 | 12.58 (12.10 to 13.09) | 11.04 in 2012 | over 99% | 1st of 35 |
| Ft Thomas 2 Sw | US | 8 | 20.23 (19.58 to 20.58) | 18.69 in 1989 | over 99% | 1st of 38 |
| Moran 5wnw | US | 10 | 5.32 (4.86 to 5.91) | 3.80 in 1934 | over 99% | 1st of 106 |
| Cheesman | US | 10 | 11.47 (10.80 to 12.04) | 9.96 in 1934 | over 99% | 1st of 113 |
| Cavern City Ap | US | 10 | 20.42 (19.94 to 21.08) | 18.93 in 2024 | over 99% | 1st of 76 |
| Canyonlands-The Neck | US | 10 | 14.88 (14.34 to 15.29) | 13.42 in 2017 | over 99% | 1st of 51 |
| Dimmitt | US | 10 | 17.28 (16.59 to 17.93) | 15.83 in 2024 | over 99% | 1st of 57 |
| Wheat Ridge 2 | US | 10 | 12.70 (12.05 to 13.38) | 11.25 in 2012 | over 99% | 1st of 39 |
| Willis Creek Colorado | US | 10 | 8.61 (8.14 to 9.13) | 7.22 in 2025 | over 99% | 1st of 37 |
| Copeland Lake | US | 10 | 7.58 (7.11 to 8.04) | 6.20 in 2025 | over 99% | 1st of 31 |
| Turkey | US | 10 | 19.59 (18.87 to 20.07) | 18.21 in 2012 | over 99% | 1st of 48 |
| Brighton 3 Se | US | 10 | 13.08 (12.51 to 13.61) | 11.75 in 2012 | over 99% | 1st of 48 |
| Walsh 1 W | US | 10 | 14.97 (14.38 to 15.66) | 13.65 in 2024 | over 99% | 1st of 51 |
| Ft Collins | US | 10 | 13.24 (12.72 to 13.83) | 11.94 in 2012 | over 99% | 1st of 128 |
| Copper Gulch Colorado | US | 10 | 10.96 (10.33 to 11.69) | 9.66 in 2025 | over 99% | 1st of 35 |
| Cheeseman Colorado | US | 10 | 11.53 (11.06 to 12.08) | 10.24 in 2012 | over 99% | 1st of 37 |
| Evergreen | US | 10 | 9.33 (8.65 to 9.81) | 8.05 in 2012 | over 99% | 1st of 55 |
| Casa Grande Nm | US | 9 | 23.58 (23.18 to 23.94) | 22.30 in 1981 | over 99% | 1st of 88 |
How the Method Did on Past Years
2000 to 2025The same computation was replayed on every complete year from 2000: for each year and each number of complete months from 3 to 11, only the years before it and its first months were read, and the result was compared with what the year became. Of the 72,755 station outlooks of 2000 to 2025 that stated a 10 to 20 percent chance of a record year, 8,634 (12%) ended with one. 156,926 station outlooks of 2000 to 2025 were called already certain or out of reach; none of them was wrong.
In 2000 to 2025 the mean error of the projected yearly mean fell from 0.44 C (0.80 F) with 3 months complete to 0.17 C (0.31 F) with 11. The 5 to 95 percent range held the final value in 92% of the 1,098,384 cases; 90% is the aim.
| Chance stated | Outlooks | Mean chance stated | Came true | Share that came true |
|---|---|---|---|---|
| Ruled out | 156,730 | 0.0% | 0 | 0.0% |
| 0 to 10% | 719,130 | 1.4% | 8,790 | 1.2% |
| 10 to 20% | 72,755 | 14.0% | 8,634 | 11.9% |
| 20 to 30% | 40,245 | 23.9% | 7,854 | 19.5% |
| 30 to 40% | 27,321 | 33.7% | 7,737 | 28.3% |
| 40 to 50% | 20,058 | 43.4% | 7,366 | 36.7% |
| 50 to 60% | 20,258 | 54.5% | 9,906 | 48.9% |
| 60 to 70% | 12,217 | 66.0% | 7,427 | 60.8% |
| 70 to 80% | 10,273 | 75.7% | 7,374 | 71.8% |
| 80 to 90% | 8,879 | 85.4% | 7,119 | 80.2% |
| 90 to 100% | 10,322 | 95.9% | 9,676 | 93.7% |
| Already certain | 196 | 100.0% | 196 | 100.0% |
A stated chance is reliable when about that share of the cases came true. The cases are not independent: in a warm year many neighbouring stations set records together, so the 26 years, not the number of outlooks, are the real size of the test. The next table shows it year by year: the records expected from the chances stated after June and after September, and the records that happened.
| Year | Outlooks | Records expected after June | Records expected after September | Records that happened |
|---|---|---|---|---|
| 2000 | 5,584 | 486.7 | 453.5 | 447 |
| 2001 | 5,581 | 191.6 | 184.8 | 91 |
| 2002 | 5,635 | 583.8 | 647.6 | 216 |
| 2003 | 5,626 | 263.3 | 383.1 | 301 |
| 2004 | 5,553 | 226.5 | 161.3 | 136 |
| 2005 | 5,346 | 193.2 | 195.4 | 190 |
| 2006 | 5,423 | 444.2 | 406.8 | 501 |
| 2007 | 5,279 | 679.8 | 549.5 | 459 |
| 2008 | 5,123 | 352.2 | 180.3 | 147 |
| 2009 | 5,016 | 204.9 | 197.8 | 147 |
| 2010 | 5,221 | 287.2 | 314.6 | 231 |
| 2011 | 5,039 | 236.4 | 251.0 | 351 |
| 2012 | 4,671 | 938.2 | 915.2 | 828 |
| 2013 | 4,211 | 152.1 | 142.3 | 82 |
| 2014 | 4,239 | 523.0 | 482.4 | 749 |
| 2015 | 4,150 | 538.2 | 429.5 | 484 |
| 2016 | 4,130 | 579.3 | 524.8 | 327 |
| 2017 | 4,123 | 596.2 | 359.2 | 251 |
| 2018 | 4,093 | 247.6 | 384.1 | 276 |
| 2019 | 3,701 | 269.9 | 263.2 | 229 |
| 2020 | 3,938 | 494.6 | 470.4 | 437 |
| 2021 | 4,019 | 164.5 | 152.8 | 151 |
| 2022 | 4,069 | 273.5 | 315.4 | 304 |
| 2023 | 4,150 | 388.4 | 459.2 | 499 |
| 2024 | 4,115 | 753.8 | 749.7 | 904 |
| 2025 | 4,012 | 319.9 | 250.1 | 384 |
| Months complete | Stations: mean error | Stations: inside the range | Countries: mean error | Countries: inside the range | World: mean error |
|---|---|---|---|---|---|
| 3 (to March) | 0.44 C (0.80 F) | 92% | 0.33 C (0.59 F) | 89% | 0.11 C (0.19 F) |
| 4 (to April) | 0.40 C (0.73 F) | 92% | 0.31 C (0.56 F) | 92% | 0.10 C (0.18 F) |
| 5 (to May) | 0.37 C (0.67 F) | 93% | 0.28 C (0.51 F) | 93% | 0.09 C (0.17 F) |
| 6 (to June) | 0.34 C (0.62 F) | 93% | 0.27 C (0.48 F) | 93% | 0.09 C (0.16 F) |
| 7 (to July) | 0.32 C (0.57 F) | 92% | 0.25 C (0.45 F) | 91% | 0.09 C (0.15 F) |
| 8 (to August) | 0.30 C (0.54 F) | 92% | 0.23 C (0.42 F) | 90% | 0.08 C (0.14 F) |
| 9 (to September) | 0.27 C (0.49 F) | 92% | 0.21 C (0.37 F) | 89% | 0.07 C (0.12 F) |
| 10 (to October) | 0.24 C (0.43 F) | 91% | 0.17 C (0.30 F) | 90% | 0.05 C (0.09 F) |
| 11 (to November) | 0.17 C (0.31 F) | 91% | 0.11 C (0.20 F) | 92% | 0.04 C (0.06 F) |
The test uses the data as they stand today. On the day itself the newest months had fewer stations than they have now, so the real outlooks of those years, had they been published, would have been somewhat less sure than the test suggests. For the world series the test has only 234 cases, one series over 26 years.
How the Outlook Is Made
A station month counts when it has at least 20 days with a maximum and 20 with a minimum reading; its value is the mean of the two monthly means. A year's value is the mean of its twelve months, so only a year with all twelve can be ranked. A station's month is complete once the station has reported a later month, which is the station's own evidence and does not depend on the day this page was built. An outlook starts with 3 complete months from January and needs 30 complete years among the 45 before the year, the reference years.
For each reference year the page takes the sum of the same months so far and the sum of the months that remain. The remainder is regressed on the year (the trend) and on the months so far, over the reference years. The slope on the months so far is the persistence: how much of a warm or cool start carried into the rest of the year in the past. A negative estimate is set to zero. The expected remainder of 2026 is the trend plus the persistence times this year's departure from its own trend.
Each reference year then gives one member of an ensemble: the months of 2026 so far, plus the expected remainder, plus the amount by which that year's remainder missed the regression, widened as a prediction error must be. The projected value is the median of the 30 members and the range runs from their 5th to their 95th percentile. The chance of a record is the share of members at or above the warmest earlier year, counted as (members + 0.5) / (30 + 1). When no member or every member reaches the record, the members cannot tell a near miss from a hopeless case, and the chance comes from the bell shaped error curve of the regression instead. A tie with the record counts for the later year, as on the warmest years page.
"Already certain" and "out of reach" are stricter than a chance of 100 or 0 percent. A year is already certain when it stays the warmest even with a remainder colder than that of any earlier year on record, colder than every member and 6 standard deviations below the expected one. Out of reach is the mirror image. The test above counts how often such a call was wrong.
The world and country values use the rules of the month in review: station anomalies against 1991 to 2020 (at least 20 years), station months more than 15 C from normal dropped as errors, 5 degree cells weighted by the cosine of latitude, and a month only when its cells hold at least 50 percent of the weight of all cells that ever reported that calendar month. A month is published 7 days after it ends. While it is younger than 45 days it is used only once its coverage reaches 90 percent of the mean coverage of the 3 months before it. A country needs 5 stations in a month. The series start in 1951.
What the Outlook Cannot Know
The method knows the past of each series and nothing else. It uses no weather forecast, no ocean state and no knowledge of El Nino or La Nina, so a forecast centre can know more about the coming months than this page does. It assumes that the next months relate to the year so far as they did in the reference years and that the trend of those years continues in a straight line.
Reports from many networks reach the source days to months after the day they describe. The newest months therefore rest on fewer stations than they will later, and a value can change after it is first shown; provisional months are marked in the table of months. A station with late data keeps its own, earlier, last complete month, and the count of stations with an outlook is dominated by the networks that report quickly.
The stations are not spread evenly over the land, and the world value is the mean over the land where this site has stations with a 1991 to 2020 normal. The country with the most of the 4,471 stations is United States with 3,132 (70%), followed by Germany (299), Australia (241), Canada (128), Sweden (114). The warming page builds its yearly series from long-record stations and complete station years, so its values and its ranking of years can differ from the ones here. The coverage column of the table of months shows how much of that area each month of 2026 holds so far.
A station record is used as measured. A station move, a new instrument or a change in observing time can shift a series by more than the margin between two warm years, and such a step can make a year look like a record or hide one (see station changes). A station whose year so far is more than 8 C from its reference years is left out as a data error; smaller steps are not detected here.
A chance is a statement about many cases, not about one. Of the outlooks that state 20 percent, about one in five should come true, and the table of past years shows how close the method came. This page describes what the station data imply for the yearly mean and does not claim a cause for a warm or a cool year.
Download the Data
Source: NOAA National Centers for Environmental Information, Global Historical Climatology Network Daily (GHCN-Daily). Records are computed by World Extreme Weather Monitor and are not official NOAA products. (US Government open data (NOAA NCEI)) Data as of 2026-11-17. Year-to-date means, projections, chances, ranks and the hindcast computed by this site from GHCN-Daily daily maximum and minimum temperatures; a statistical outlook from past years, not a forecast and not an official NOAA product.