Wholestory

← Top Stories

Climate Change

°C above 1850–1900Observed to Sep 2026

Observed records are set against 1850–1900 from their own data. Each assessment is drawn as it stated itself: a rate wedge where it gave a near-term rate (FAR, SAR, TAR, AR4), anchored at the observed five-year mean ending in its baseline year and run to its stated horizon; a level path where it gave levels against 1850–1900 (AR5, AR6). A range stated without a central estimate (TAR, AR5) is a band only.

Hansen 1988Hansen 1988: 0.28 (0.12–0.3) per decade, Scenario B (“perhaps the most plausible of the three”), with A (continued exponential trace-gas growth) and C (emissions curtailed, forcing flat after 2000) as the range: each scenario’s global temperature trend over 1988–2017, the period Hausfather et al. 2020 evaluate. Stated Jun 23, 1988, to 2017.IPCC FAR 1990IPCC FAR 1990: 0.3 (0.2–0.5) per decade, business-as-usual (Scenario A). Stated May 25, 1990, to 2100.IPCC SAR 1995IPCC SAR 1995: 0.182 (0.091–0.318) per decade, IS92a: about 2°C by 2100 from 1990 (range 1.0–3.5°C as the TAR restated it), expressed per decade over 110 years. Stated Dec 15, 1995, to 2100.IPCC TAR 2001IPCC TAR 2001: 0.127–0.527 per decade (range only), 35 SRES scenarios: 1.4–5.8°C over 1990–2100, expressed per decade; no central estimate stated. Stated Jan 20, 2001, to 2100.IPCC AR4 2007IPCC AR4 2007: 0.2 per decade, about 0.2°C per decade for the next two decades, range of SRES scenarios; no range stated. Stated Feb 2, 2007, to 2027.IPCC AR5 2013IPCC AR5 2013: 2016–2035 likely 0.3–0.7°C above 1986–2005, which the SPM puts 0.61 [0.55–0.67]°C above 1850–1900; no central estimate stated (range only). Stated Sep 27, 2013, to 2035.IPCC AR6 2021IPCC AR6 2021: SSP2-4.5, Table SPM.1: best estimate and very likely range for 2021–2040 and 2041–2060, from the assessed 2011–2020 observed 1.09 [0.95–1.20]°C. Stated Aug 9, 2021, to 2060.NASA GISTEMPHadCRUT5Berkeley Earth (high-resolution)1Jun 23, 1988 — The earth is warmer in 1988 than at any time in the history of instrumental meas2May 25, 1990 — Under the IPCC Business-as-Usual (Scenario A) emissions... a rate of increase of3Dec 15, 1995 — For the mid-range IPCC emission scenario, IS92a, the model-projected global mean4Jan 20, 2001 — The globally averaged surface temperature is projected to increase by 1.4 to 5.85Feb 2, 2007 — For the next two decades, a warming of about 0.2°C per decade is projected for a6Sep 27, 2013 — The global mean surface temperature change for the period 2016–2035 relative to 7Dec 12, 2015 — The Paris Agreement is adopted8Oct 8, 2018 — The IPCC prices the difference between 1.5°C and 2°C9Aug 9, 2021 — Global surface temperature will continue to increase until at least mid-century

Observed to Sep 2026 · latest
NASA GISTEMPHadCRUT5Berkeley Earth (high-resolution)Stated range, central estimate dashed
  1. 1Hansen tells the Senate warming has begun
  2. 2IPCC FAR projects 0.3°C a decade
  3. 3IPCC SAR projects about 2°C by 2100
  4. 4IPCC TAR projects 1.4–5.8°C by 2100
  5. 5IPCC AR4 projects 0.2°C a decade
  6. 6IPCC AR5 projects 0.3–0.7°C by 2016–2035
  7. 7Paris Agreement adopted
  8. 8IPCC puts 1.5°C on the table
  9. 9IPCC AR6: 1.5°C by 2021–2040
Global surface temperature as the major records measure it, against what each assessment projected from the year it was published. Drag the bar to any vintage to see what was known and what was forecast at that moment. Numbered dots are the moments the record marks.

billion tonnes CO₂ per year

China
United States
EU27
India
Rest of world
Territorial fossil CO₂ including cement and excluding the cement carbonation sink, annual, as Our World in Data relays the Global Carbon Budget; international aviation and shipping sit in the world total and so in “rest of world”
  1. 1Kyoto sets binding targets
  2. 2Copenhagen ends in an accord
  3. 3Paris Agreement adopted
  4. 4Endangerment finding rescinded
  5. 5China’s emissions fall
  6. 6EPA repeals power-plant rules
Fossil carbon dioxide by country and region since 1960, stacked. Numbered dots are the moments the record marks on each country’s line.

billion tonnes CO₂-equivalent per year

Electricity and heat
Transport
Manufacturing and construction
Agriculture
Buildings
Fugitive emissions
Industry
Waste
Aviation and shipping
Other fuel combustion
Land-use change and forestry
All greenhouse gases (100-year CO₂-equivalent), annual, to 2023; the inventory runs two years behind the fossil CO₂ data
  1. 1Kyoto sets binding targets
  2. 2Copenhagen ends in an accord
  3. 3Paris Agreement adopted
  4. 4Endangerment finding rescinded
  5. 5China’s emissions fall
  6. 6EPA repeals power-plant rules
All greenhouse gases by the sector that emits them. The sector inventories run a year or two behind the country data.

GW installed, cumulativeObserved to 2025Illustrative data — pending sourcing

Each World Energy Outlook’s main-scenario solar PV capacity path, from its vintage to the horizon it stated, against installed capacity as reported. Paths are stated levels, not rates, and are drawn from the forecast’s own figures.

IEA WEO 2006IEA WEO 2006: main scenario. Stated Nov 7, 2006, to 2030.IEA WEO 2010IEA WEO 2010: main scenario. Stated Nov 9, 2010, to 2030.IEA WEO 2014IEA WEO 2014: main scenario. Stated Nov 12, 2014, to 2030.IEA WEO 2018IEA WEO 2018: main scenario. Stated Nov 13, 2018, to 2030.IEA WEO 2022IEA WEO 2022: main scenario. Stated Oct 27, 2022, to 2030.Installed capacity (IRENA)1Dec 31, 2025 — Renewables overtake coal worldwide for the first time

Observed to 2025 · latest
Installed capacity (IRENA)Stated range, central estimate dashed
  1. 1Renewables overtake coal
Solar capacity as installed, against what each World Energy Outlook forecast from the year it was published.

Carbon dioxide in the atmosphere averaged 426.8 parts per million in 2025, 153% of the pre-industrial level, the World Meteorological Organization reported on 23 October. Methane and nitrous oxide also set records. The year’s rise, 2.8 ppm, was slower than the one before, as land carbon sinks recovered after the last El Niño.

The plans meant to bend that curve fall short. The UN climate secretariat’s synthesis of 127 new national plans, published on 27 October, finds that delivered in full they would bring emissions to about 6% below 2019 levels by 2035, or 9% with their conditional pledges. It calls that “substantially off track” for 1.5°C. A week earlier, UNEP put the world on course for 2.6°C even if every plan is kept.

The Lancet Countdown linked heat to a record 710,000 deaths in 2023, more than 70% of them attributable to warming. In Japan, warming added 9% to the moisture behind Typhoon Dujuan’s deadly rains, though its effect on the rainfall itself could not be pinned down.

The IPCC will reconvene online on 1–3 December to try again on its assessment timetable. In the US, the EPA has proposed loosening its methane rules for oil and gas.

New

9

in all one topics

The whole story1 chapters
The state of it — as of October 29, 2026
1.4
above pre-industrial, latest full year (NASA GISTEMP)
−0.1 since 2024
2nd warmest
January–September 2026 on record, 0.07°C behind 2024
3.11°C
daily Niño 3.4 reading, 21 September — the highest El Niño measured

In this story

US Federal Climate Policy

The EPA has proposed cutting back the 2024 methane standards for oil and gas. It would scrap the super-emitter program, which lets certified third parties flag large leaks, let oil wells flare their gas instead of piping it to sale, and lighten the rules for low-producing wells. President Trump announced it on 27 October as a termination, but it is a proposal: the standards stay in force while comments run to 16 December. The agency projects 60 million more short tons of methane over 2027–2050, and oil and gas prices moving by a fraction of a percent. It rests on the Clean Air Act, not on February’s rescission of the endangerment finding.

The repeal of most power-plant carbon standards takes effect on 16 November, and comments on scrapping the rest close on 2 November. Challenges to both the repeal and the rescission sit undecided in the appeals court in Washington.

New York has appealed to the Second Circuit the two rulings that struck down its climate superfund law. In the Supreme Court, where the federal government argued on 5 October that federal law bars state climate suits, a ruling in Suncor v. Boulder County is expected by next summer.

Continue: US Federal Climate Policy →

billion tonnes CO₂ per year

Electric power
Transportation
Industrial
Residential
Commercial
Energy-related CO₂ by sector, annual, as the Energy Information Administration reports it (Monthly Energy Review, tables 11.2–11.6)
1.5
US power-sector CO₂, latest full year
+0.1 since 2024
4
challenges to the endangerment rescission pending in the D.C. Circuit, none decided
26
petitioners led by New York challenging the power-plant repeal

Today

Typhoon Dujuan: Warming Loaded the Air With Water, but Its Rain Defied a Number

Typhoon Dujuan skirted Chiba's Boso Peninsula on 21 September and drenched the Kantō region. Seventeen people died, most of them in landslides. A World Weather Attribution study found that human-caused warming added 9% to the water vapour in the air over the region, and that the seas the storm crossed were 2–3°C warmer than in a pre-industrial climate. It could not put a figure on warming's effect on the rainfall itself, because the models disagreed. Simulated look-alike storms rained 20–40% more where rain was heaviest, and El Niño played no part. A separate Japanese government study, published on 23 October, found warming may have raised the rain by about 20%.

Read

World Weather Attributionworldweatherattribution.org85% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.Japan Ministry of Education, Culture, Sports, Science and Technologymext.go.jp85% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.Red Cross Red Crescent Climate Centreclimatecentre.org80% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.

Heat Was Linked to a Record 710,000 Deaths in 2023, Most of Them Attributable to Warming

The tenth Lancet Countdown on health and climate change was published on 28 October. It is led by University College London in collaboration with the World Health Organization. It linked heat to a record 710,000 deaths worldwide in 2023, more than 70% of them attributable to human-caused climate change. In the years since the Paris Agreement, heat deaths have averaged nearly 36% more a year than in the decade before. In 2025, heat cost 573 billion potential work hours, and drought touched 51.8% of the world's land for at least a month.

Read

Lancet Countdown on Health and Climate Changelancetcountdown.org85% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.United Nationsnews.un.org85% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.World Health Organizationwho.int85% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.

New National Climate Plans Would Cut Emissions About 6% Below 2019 by 2035

The UN climate secretariat's 2026 synthesis of national climate plans was published on 27 October. It covers 127 new plans from 154 parties, about 71% of global emissions. Delivered in full, they would bring emissions to a peak and then to an average of 6% below 2019 levels by 2035, within a range of 0–12%. With their conditional parts, the cut is about 9%. If countries' long-term net-zero targets are also met, the secretariat's best estimate is 1.7–2.3°C of warming by 2100. It calls progress "substantially off track" for 1.5°C. The plans' renewable targets add up to 3 terawatts by 2030, against 5.2 TW already installed in 2025.

Read

United Nations Framework Convention on Climate Change (UNFCCC)unfccc.int90% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.United Nations Framework Convention on Climate Change (UNFCCC)unfccc.int90% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.United Nations Framework Convention on Climate Change (UNFCCC)unfccc.int90% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.

Earlier this month

13
New

Greenhouse Gases Set New Records in 2025, With Carbon Dioxide at 426.8 ppm

The World Meteorological Organization's annual Greenhouse Gas Bulletin, released on 23 October, put the 2025 global average concentration of carbon dioxide at 426.8 parts per million, 153% of the pre-industrial level. Methane, at 1,948 parts per billion, and nitrous oxide, at 339.1, also set records. CO₂ rose 2.8 ppm over the year. That is faster than the 2011–2020 average of 2.4 but slower than the 2023–24 surge, which WMO puts down to land carbon sinks recovering after the last El Niño.

Read

World Meteorological Organization (WMO)wmo.int90% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.United Nationsnews.un.org85% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.

The Gap Widens: 2.8°C on Current Policies, 2.6°C if Every Pledge Is Kept

The UN Environment Programme's Emissions Gap Report, published 21 October, projects 2.8°C of warming this century if current policies continue, and 2.6°C if every national plan is delivered. UNEP says that is worse than last year's report, which put full delivery at 2.3–2.5°C, and blames the United States' withdrawal from the Paris Agreement and revised estimates of past emissions; newer plans barely moved it. Delivered in full, the plans cut emissions 5–8% below 2025 by 2030 and 11–15% by 2035. Its lowest-overshoot path needs 18% and 45%, peaks at 1.7–1.8°C, and returns below 1.5°C by 2100 only with even odds.

Read

United Nations Environment Programmeunep.org88% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.United Nations Environment Programmeunep.org88% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.United Nationsnews.un.org85% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.United Nations Environment Programmeunep.org88% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.
Our latest Global Seasonal Climate Update forecasts that sea surface temperatures in the tropical Pacific – the epicentre of El Niño - may rise to 3.8°C above average during November 2026 to January 2027.

— World Meteorological Organization (WMO)

Not yet checked

Not yet checkable: it settles on the November 2026–January 2027 Niño 3.4 average, early in 2027. 'May rise to 3.8°C' is scored as reaching 3.5°C or more; the seasonal record is +2.6°C (2015–16). WMO named neither the index nor its baseline.

Read

World Meteorological Organization (WMO)wmo.int90% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.United Nationsnews.un.org85% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.
New

The IPCC Breaks Off Without a Date for Its Next Assessment

The Intergovernmental Panel on Climate Change suspended its 65th session in Addis Ababa on the night of 16 October without agreeing when the three working-group reports of its seventh assessment are due. It was the sixth session running to fail on this. Small-island states and developed countries want all three reports before the second global stocktake, in 2028; a group led by Saudi Arabia, India and Russia wants the impacts and mitigation reports after it. Those three also blocked the budget; at current contributions the panel's fund runs dry by the end of 2028. On 26 October the secretariat asked delegates to hold 1–3 December for a virtual resumed session on the work plan and budget.

Read

Intergovernmental Panel on Climate Change (IPCC)ipcc.ch88% credibilityEarned from 1 checked claim.Earth Negotiations Bulletin (IISD)enb.iisd.org90% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.Agence France-Presse (AFP)afp.com85% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.Intergovernmental Panel on Climate Change (IPCC)ipcc.ch88% credibilityEarned from 1 checked claim.Carbon Briefcarbonbrief.org85% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.The Guardiantheguardian.com90% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.Intergovernmental Panel on Climate Change (IPCC)apps.ipcc.ch88% credibilityEarned from 1 checked claim.Third World Networktwn.my65% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.

Hurricane Polo: Warming Outweighed a Record El Niño, but Its Rain Could Not Be Attributed

Polo went from depression to Category 5 within two days from 20 September, held it a Pacific-record two and a half days and killed five people in Mexico. A rapid study by World Weather Attribution and Climate Central, released 15 October and not peer-reviewed, found human-caused warming made the ocean conditions that powered it about 23 times more likely, and El Niño about 14 times. Peak winds were about 2.7 m/s stronger. It could not quantify warming's role in the rainfall, because models fail to reproduce the region's observed trends. Three outside scientists told AP it likely understates warming's share.

Read

World Weather Attributionworldweatherattribution.org85% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.Associated Pressapnews.com95% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.
NCEI’s statistical analysis indicates a greater than 99% probability that 2026 will rank among the five warmest years on record, with the highest likelihood favoring the second-warmest year on record.

— NOAA

Not yet checked

Not yet checkable. It settles on where 2026 ranks in NOAA’s own annual record, due in January 2027. In August the same analysis favoured third or fourth.

Read

NOAAncei.noaa.gov95% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.NOAAncei.noaa.gov95% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.

September Is a Second Straight Month Past 1.5°C, and the Ocean Sets Another Record

Europe's ERA5 record puts September 2026 at 1.63°C above the 1850–1900 average and 0.82°C above 1991–2020, second only to September 2023 — and, after August, the first pair of consecutive months above 1.5°C since November 2025. Sea surface temperature between 60°S and 60°N averaged 21.05°C, a record for September, beating 2023's 20.92°C. Western Europe tied its warmest September, and parts of the Rhine, Danube, Vistula and Garonne ran at record-low September flows. One month is not the long-term average the Paris limit is measured against.

Read

Copernicus Climate Change Service (C3S)climate.copernicus.eu95% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.
We are in a new run of months above 1.5°C, and with a strong El Niño still intensifying it is likely to last into 2027.

— Copernicus Climate Change Service (C3S)

Not yet checked

Not yet checkable. It settles on Copernicus's monthly ERA5 figures for October 2026 through January 2027; read charitably, every one must stay above 1.5°C. August (1.65°C) and September (1.63°C) began the run.

Read

Copernicus Climate Change Service (C3S)climate.copernicus.eu95% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.
For October-November-December, the seasonal average Niño 3.4 sea surface temperature anomaly is forecast to reach approximately +3.7°C, peaking around December.

— World Meteorological Organization (WMO)

Not yet checked

Not yet checkable: it settles on the October–December 2026 Niño 3.4 average, early in 2027. The previous seasonal record was +2.6°C (2015–16). WMO did not state the anomaly's baseline.

Read

World Meteorological Organization (WMO)wmo.int90% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.

El Niño Passes the Very-Strong Mark Week by Week, Not Yet Season by Season

NOAA’s weekly reading for the central equatorial Pacific stood 2.3°C above average in its 5 October update, past the 2.0°C line used for a very strong El Niño. The three-month measure by which an event’s strength is judged is lower: 1.7°C for July to September. Its September outlook gave over 90 percent odds of a very strong event through the northern winter, and 75 percent odds that October to December would exceed every El Niño since 1950; its October discussion, as reported by Yale Climate Connections, raised that to 83 percent. It expects El Niño to last through March–May 2027.

Read

NOAAcpc.ncep.noaa.gov95% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.Yale Climate Connectionsyaleclimateconnections.org80% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.

Europe’s Summer in Full: Record Heat Stress and the Lowest Rivers in Three Decades

Copernicus’s review of summer 2026 puts western Europe at 21.7°C, 2.5°C above the 1991–2020 average, past the 2003 record; Europe overall ranked third. A record 52 percent of the continent was exposed at some point to “very strong” heat stress, a feels-like 38°C or more, and western Europe averaged 41 days of strong heat stress, though some places saw none. Simulated river flows were the lowest for any summer in a record that starts in 1992, the Rhine, Danube, Po and Loire low at once. EU burned area ran almost twice the 2006–2025 average by early September.

Read

Copernicus Climate Change Service (C3S)climate.copernicus.eu95% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.Copernicus Climate Change Service (C3S)climate.copernicus.eu95% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.
2026 as a whole is now on track to end up roughly tied with 2024 as the warmest year on record in ERA5, at 1.6C above preindustrial. And 2027 waiting in the wings to be quite a bit warmer on the back of a record-setting El Nino event.

— Zeke Hausfather

Not yet checked

Not yet checkable. It settles on the 2026 annual figure in the European ERA5 record, due in January 2027. NOAA, reading its own dataset, calls second place likeliest.

Read

Yale Climate Connectionsyaleclimateconnections.org80% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.

September Runs Second-Warmest, and the Oceans Set an All-Time High

September 2026 was the second-warmest September on record, behind 2023, in both NOAA’s and NASA’s analyses: 1.33°C above the twentieth-century average in NOAA’s, 1.38°C above 1951–1980 in NASA’s. The two figures use different baselines and are not interchangeable. The global ocean surface ran 1.10°C above its twentieth-century average — a fourth straight monthly record, and the largest departure NOAA has measured for any month. Record warmth covered 16.7 percent of the planet’s surface, the widest footprint in data from 1951. January to September stands second-warmest, 0.07°C behind the same months of 2024. NOAA now gives better than 99 percent odds of a top-five year and calls second place the likeliest finish; in August it favoured third or fourth.

Read

NOAAncei.noaa.gov95% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.NASAdata.giss.nasa.gov95% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.Yale Climate Connectionsyaleclimateconnections.org80% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.NOAAncei.noaa.gov95% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.

This quarter

22

El Niño Breaks Its All-Time Record, and the Forecast Goes Higher

The tropical Pacific is warmer than in any El Niño measured. On 21 September the daily Niño 3.4 reading hit 3.11°C above normal, past the 3.08°C of November 2015, by Carbon Brief's count; Copernicus put September's average surface temperature there at 29.70°C, above the peaks of 1982–83, 1997–98 and 2015–16. On 6 October the World Meteorological Organization forecast an October–December average near +3.7°C, peaking around December, against a previous seasonal record of +2.6°C. Measures differ: NOAA's 8 October discussion gave Niño 3.4 at +2.1°C and still calls a strong-to-very-strong event likely through early 2027, and an index that adjusts for the warming ocean has not yet broken its record.

Read

Carbon Briefcarbonbrief.org85% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.Copernicus Climate Change Service (C3S)climate.copernicus.eu95% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.World Meteorological Organization (WMO)wmo.int90% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.NOAAcpc.ncep.noaa.gov95% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.

The Arctic Minimum Lands Tenth-Lowest, After a Summer That Began Near the Record

Arctic sea ice reached its 2026 minimum on 12 September at 4.60 million km², tied with 2008, 2010 and 2025 for tenth-lowest in the satellite record. Extent ran near record lows through May; from June a cool, stormy pattern over the central Arctic slowed the melt. September averaged 4.81 million km², thirteenth-lowest and 1.6 million below the 1981–2010 average. The decline runs at 11.6 percent a decade; the twenty lowest minimums are the last twenty years. Antarctic ice provisionally peaked on 14 September at 17.59 million km², its third-lowest maximum, and by early October sat at or near record lows.

Read

National Snow and Ice Data Center (NSIDC)nsidc.org90% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.NASAscience.nasa.gov95% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.CIRES (University of Colorado Boulder)cires.colorado.edu90% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.

El Niño Is Being 'Supersized,' and the Forecast Is Near-Certain It Lasts Into 2027

The World Meteorological Organization issued its most unequivocal El Niño forecast yet: the event is firmly established and will strengthen to 'very strong,' peaking towards the end of 2026, with a likelihood its models put near 100 percent that it persists through February 2027. The tell is in the tropical Pacific, where the Niño 3.4 index climbed from about 1.5°C above normal in May–July to between 2.2 and 2.6°C by mid-August, and subsurface water ran more than 8°C above average in places. Above-normal temperatures are expected across almost all land through autumn. El Niño is a natural cycle, not a product of warming — but it is landing on an already record-warm ocean, and its heat usually shows up in global temperatures most strongly the year after it develops, pointing the record toward 2027. UN chief António Guterres said the planet is 'in uncharted waters.'

Read

World Meteorological Organization (WMO)wmo.int90% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.Reutersreuters.com95% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.United Nationsnews.un.org85% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.
El Niño will strengthen to a very strong event peaking towards the end of 2026, and forecasts put at nearly 100 percent the likelihood it persists through February 2027.

— World Meteorological Organization (WMO)

Read

World Meteorological Organization (WMO)wmo.int90% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.

The Arctic Ice Record Comes Back Online, and the Summer Tracked 2012

The National Snow and Ice Data Center resumed monthly sea-ice analysis in September, eleven months after funding lapsed and suspended much of its Sea Ice Today service. Philanthropic support has restored the monthly posts and the announcements of the annual maximum and minimum; mid-month updates, the ice-volume series and several analysis tools remain shut. Its first post back reports that Arctic extent this summer tracked 2012 — the record-low September year — until August. The 2026 minimum, normally reached in the second half of September, has not yet been called.

Read

National Snow and Ice Data Center (NSIDC)nsidc.org90% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.National Snow and Ice Data Center (NSIDC)nsidc.org90% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.

Britain's Warmest Summer Was Made 130 Times More Likely by Warming

Summer 2026 was provisionally the UK's warmest on record at a 16.5°C mean, beating the mark set the previous year by 0.4°C, with England, Wales and Northern Ireland each setting national records. A Met Office rapid attribution study found that a summer this hot has become roughly 130 times more likely in today's climate than it would have been without human-caused warming. It was also strikingly dry — drought declarations, hosepipe bans and wildfires followed, and Wisley went more than 60 days without measurable rain. The figure is a probability ratio, not a claim the season was impossible before.

Read

UK Met Officemetoffice.gov.uk90% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.

China's Emissions Fall on Oil, Not Coal, for the First Time

China's carbon dioxide emissions fell about 1 percent year-on-year in the second quarter of 2026 — the first time a drop in oil use, rather than coal, has been enough to turn the world's largest emitter down. Oil consumption fell roughly 9 percent overall and 16 percent in transport after supply disruption through the Strait of Hormuz, while coal burnt for power rose 2.4 percent as wind and solar output was curtailed. Electric vehicles displaced more oil in the first half of the year than the United Kingdom consumed over the same period. The analysis, made by the Centre for Research on Energy and Clean Air from national statistics, customs and power-industry data, cautions that about 60 percent of the 32 percent fall in oil imports reflects stockpile drawdown rather than lower consumption. Emissions remain marginally up across the half-year, and flat since peaking in March 2024.

Read

Carbon Briefcarbonbrief.org85% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.National Bureau of Statistics of Chinastats.gov.cn75% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.
China is on track to add enough wind, solar, nuclear and hydropower this year to cover electricity demand growth, despite a slowdown in new capacity.

— Centre for Research on Energy and Clean Air (CREA)

Read

Carbon Briefcarbonbrief.org85% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.

UNEP Says the Crossing Is Coming, and Names the Peak

A UNEP report, Limiting Overshoot, concludes that global temperature rise is set to cross 1.5°C above pre-industrial levels — the long-term average, not a single month — likely within the next few years. Its most optimistic scenario peaks at 1.8°C; most others peak higher, some above 2°C. It argues that an overshoot-peak-decline pathway is now the best remaining route back below 1.5°C, and that the exceedance should be made as small and as short as possible. Read alongside August’s single month at 1.65°C, it is the difference between weather inside a trend and the trend itself.

Read

United Nations Environment Programmeunep.org88% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.

August Was the Warmest Month Ever Measured, and the First Above 1.5°C Since November

August 2026 averaged 16.96°C globally — 0.85°C above the 1991–2020 August average, the warmest August in the ERA5 record, and the joint-warmest calendar month ever measured, level with July 2023. Against the 1850–1900 pre-industrial baseline it stood 1.65°C higher, the first single month above 1.5°C since November 2025. That is a month, not the long-term average the Paris target is defined against. June to August was the joint-warmest boreal summer on record with 2024; Western Europe had its warmest summer ever, passing 2003. Extra-polar sea surface temperature averaged 21.07°C, the highest for any August and joint-highest of any month, with a daily record of 21.11°C — though Copernicus notes some other datasets do not show an August SST record. Arctic sea ice ranked 12th lowest for the month, Antarctic fourth lowest.

Read

Copernicus Climate Change Service (C3S)climate.copernicus.eu95% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.

A Land Convention Meets for Twelve Days and Cannot Agree on Drought

UNCCD COP17 closed in Ulaanbaatar having adopted 34 decisions — on rangelands and pastoralists, which cover 54 percent of the world's land and support nearly 500 million people, on sand and dust storms, land tenure, finance and the science-policy interface — and having failed on the one it was called to settle. Parties reached no substantive agreement on drought, nor on migration, aridity or knowledge sharing, and IISD's negotiations bulletin records an unprecedented number of agenda items blocked from consideration outright, with gender, civil society participation and inter-convention synergies pushed to COP18. On drought the obstacle was named: a small group led most vocally by the United States would accept no international instrument, binding or voluntary, holding drought to be domestic, bilateral or regional. The Africa Group argued the opposite. The same disagreement stopped the question at Riyadh in 2024.

Read

Earth Negotiations Bulletin (IISD)enb.iisd.org90% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.Down To Earthdowntoearth.org.in75% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.

Attribution Looks at the Assam Floods and Finds No Signal

The floods that hit Upper Assam and neighbouring states between late June and early August killed at least 99 people in Assam, affected more than 500,000 and submerged over 45,000 hectares of cropland. World Weather Attribution found the rainfall behind them was not exceptional: a return period of under two years, no trend in the observations, and no material change in likelihood or intensity from 1.4°C of warming in the models. It points instead at exposure — unplanned urbanisation, deforestation and embankment-led flood control. Sparse rain gauges, the team cautions, leave the trend uncertain rather than proven absent.

Read

World Weather Attributionworldweatherattribution.org85% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.

The Caribbean Drought Is Attributed — Except Where the Data Runs Out

World Weather Attribution assessed the drought gripping Puerto Rico, Jamaica and the Lesser Antilles since May. The May-to-July rainfall shortfall was a roughly 50-year event in Puerto Rico and 30 years in Jamaica; measured as effective precipitation — rainfall minus the evaporation warmth drives — 15 years in Puerto Rico, 30 in Jamaica. Combining observations with models, the study puts about 6 percent of Jamaica's rainfall decline and 9 percent of Puerto Rico's down to human-caused warming, and finds a shortfall this severe now about 40 percent more likely in Jamaica and 30 percent more likely in Puerto Rico than in a climate 1.4°C cooler. For the Lesser Antilles it declines to attribute: small islands, sparse stations, observations and models pointing opposite ways. On the ground: water rationing in Puerto Rico, more than 300 brushfires there in July, restrictions across several Jamaican parishes.

Read

World Weather Attributionworldweatherattribution.org85% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.Carbon Briefcarbonbrief.org85% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.

A Fire Season That Would Have Been a Fifty-Year Event

World Weather Attribution assessed the weather behind Canada's 2026 fire season, in which more than 3 million hectares burned — about 94 percent of it in the five weeks from late June to late July — and smoke carried air-quality alerts to more than 120 million people from Minnesota to New Hampshire. In the Northwest Territories the week-long extreme in fire severity now recurs roughly every two years; in a climate 1.4°C cooler it would have been a fifty-year event, and the month-long measure a hundred-year one. Ontario's week-long extreme was about 3.5 times more likely. Climate models show a much weaker effect than the observations do, so the study's combined estimates are lower: around five-fold for the Northwest Territories week, about two-fold elsewhere. It declines to attribute the two years of dryness that preceded the fires.

Read

World Weather Attributionworldweatherattribution.org85% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.

For the First Time, Britain's Heat Deaths Outnumber Its Cold Deaths

New Office for National Statistics data put more than 40,000 excess deaths on England's hottest days between 1988 and 2025, against 139,000 on its coldest — and, for the first time in the series, heat overtook cold in the latest year, with 2,805 heat-related deaths in England and Wales exceeding 2,244 cold-related ones. Nearly every English region and Wales saw the risk of dying rise about 50 percent once temperatures passed 24°C. The count stops before summer 2026, already estimated to be more lethal still: the UK health agency attributed some 2,877 deaths to the May and June heatwaves alone.

Read

UK Government (gov.uk)ons.gov.uk90% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.The Guardiantheguardian.com90% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.

A Mountainside Falls in Nepal, and Attribution Stops at the Temperature

On 26 August a rock wall on Langtang Lirung, on the Nepal–China border, collapsed and took part of the glacier above it down with it. The debris flood reached the Rasuwagadhi border crossing, 22 kilometres downstream, within seven minutes. By mid-September Nepal had confirmed more than 1,300 deaths, with over 5,000 people still missing. World Weather Attribution found about 1.5°C of the region’s July–August warming attributable to human influence, and a freezing line climbing roughly 100 metres a decade. It did not assess whether the avalanche itself would have happened without that warming, and describes climate change as a destabilising factor acting on a slope already predisposed to fail — possibly by the 2015 earthquake — rather than the cause.

Read

World Weather Attributionworldweatherattribution.org85% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.

The Ocean Sets Its Warmest Day in August, Not April

The daily average sea surface temperature across the extra-polar ocean, 60°S to 60°N, reached 21.1°C on 22 August, the highest in the ERA5 record. It beat the March 2024 mark by a hundredth of a degree — inside the dataset's precision, and not the interesting part. Ocean temperatures normally peak in March and April after the southern summer; this was August, and it beat the previous August high of 20.98°C by a wider margin. It follows a run of monthly records through 2026 and arrives as a strong El Niño develops. Copernicus notes daily records must be read in context.

Read

Copernicus Climate Change Service (C3S)climate.copernicus.eu95% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.

Europe's Seas Warmed Faster Than the Models Said They Would

World Weather Attribution assessed July's record sea surface temperatures around Europe across four regions. The warmest 14-day period of the year has warmed by 1.2°C in Celtic waters, 1.5°C off Biscay and Iberia, 2.6°C in the eastern Mediterranean and 3.1°C in the western — the last two well above the rate of global warming. In the Celtic region the probability ratio is about 30, making this a roughly one-in-a-hundred-year event in a climate 1.4°C cooler; in the other three, July values like these would have been virtually impossible before industrialisation. Marine heatwaves covered nearly 70 percent of the eastern Mediterranean against 9 percent in that cooler climate. One finding cuts against the instruments: in three of the four regions the observed trend is much stronger than the models simulate, which is why WWA calls its attributable figures — about 2°C in both Mediterranean regions — a conservative estimate.

Read

World Weather Attributionworldweatherattribution.org85% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.

EU emissions tick up on the quarter, down on the year

Eurostat’s first estimate for 2026 put EU greenhouse gas emissions at 837 million tonnes of carbon dioxide equivalent in the first quarter, 0.3% above the previous quarter while economic output was flat. Against the same quarter of 2025 the picture reverses: emissions down 1.2% while GDP grew 0.8% — the decoupling on which the bloc’s target of a 55% net cut by 2030 depends. Electricity, gas and heat supply drove the quarterly rise, at +4.8%, while households, manufacturing and transport all fell. Twenty member states emitted more than in the previous quarter and seven less. Eurostat counts on a residence basis, which differs from the accounting the EU submits under the Paris Agreement.

Read

European Commissionec.europa.eu90% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.

Our reading

Two forecasters, two different finishes for 2026

NOAA’s annual outlook, published on 10 August, says it is very likely that 2026 will rank among the five warmest years, “with the highest probabilities favoring it as the third- or fourth-warmest year on record”. Carbon Brief’s projection of 24 July put the year second-warmest, at around 1.51°C above pre-industrial levels, with a 35% chance of beating 2024 outright. Much of the gap is a question of which thermometer is being read: Carbon Brief’s own table of record odds runs from 66% in Berkeley Earth and 65% in NASA GISTEMP down to 24% in NOAA’s dataset and 13% in the European reanalysis. NOAA’s record is therefore the second most conservative of the six, and its outlook reads accordingly.

Compared

NOAAncei.noaa.gov95% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.Carbon Briefcarbonbrief.org85% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.

July ties the warmest on record and the oceans set a new high

Two independent records agree that July 2026 was as hot as any July measured, and differ only on the ranking. NOAA put it 1.18°C above its twentieth-century average — tied with 2024 as the warmest July in the record — with North America, Africa and Asia each having their warmest July and Northern Hemisphere land setting an all-time July high. The European reanalysis put it at 1.47°C above pre-industrial levels and called it joint-second, behind 2023 and level with 2024. They agree on the ocean: sea surface temperatures were the highest ever recorded for a July, and in NOAA’s accounting this was the first July to pass 1.0°C above average, a departure matched only by months inside previous El Niño events. Western Europe had its hottest June–July on record, with drought deep enough to push the Seine, Rhine and Danube to exceptionally low flows. The past four Julys are the four warmest measured.

Read

NOAAncei.noaa.gov95% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.Copernicus Climate Change Service (C3S)climate.copernicus.eu95% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.

Attribution study ties record France and Spain wildfires to warming

World Weather Attribution found human-caused climate change sharply raised the odds of the fire-weather conditions behind the July 2026 wildfires in France and Spain. Using the seven-day fire-severity index against a 1.4°C-cooler counterfactual, it judged the extreme at least 20 times more likely — and about 42% more intense — in central Spain, and at least twice as likely (about 46% more intense) in southwest France. Spain's Ávila fire burned roughly 44,000 hectares, its largest on record; France logged a record ~115,000 hectares burned year-to-date, with about 220,000 people evacuated near Bordeaux. The study points to 'weather whiplash' — a wet winter growing fuel, then a hot, dry summer drying it out — as an increasingly important driver.

Read

World Weather Attributionworldweatherattribution.org85% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.

Earlier this year

14

Renewables overtake coal worldwide for the first time

Full-year 2025 data showed renewables supplying 33.8% of global electricity (10,730 TWh) and overtaking coal (33.0%, 10,476 TWh) for the first time in the modern power system — with coal falling below a third of generation for the first time in history. Record solar growth (+636 TWh to 2,778 TWh, up 30%) meant clean power met all of the year's demand growth, so global fossil-fuel generation did not rise for the first time since 2020. Fossil generation fell in both China (−0.9%) and India (−3.3%), the historic engines of its growth, and solar overtook wind for the first time.

Read

Emberember-energy.org85% credibility · not yet testedAn opening estimate from its institutional history — none of its claims checked yet.

Earlier

25
Covered since Jun 1988Read it from the start →