Level 1 — Absolute Beginner

When we burn coal, oil or gas, we make a gas called carbon dioxide. Scientists write it as CO2. This gas goes into the air. It stays there for a long time. It makes the planet warmer.

Every year, scientists count how much CO2 the world makes. More than 140 scientists work on this report. They showed the new numbers on 10 November at a big climate meeting in Turkey.

This year there is some good news. The amount of CO2 from fossil fuels will go down a little. It will fall by 0.4 percent. The total is 38.2 billion tonnes. This is the first fall in several years.

Why is it falling? Oil is very expensive now because of the war with Iran. So people drive less and fly less. Also, there are more solar panels, more wind turbines and more electric cars. But the scientists are not happy yet. They say a war is not a good way to help the planet.

coal

A hard black rock from the ground that people burn for energy.

gas

A substance like air that has no fixed shape.

carbon dioxide

The gas we make when we burn coal, oil or gas; written as CO2.

planet

A very large round object in space, such as the Earth.

scientist

A person who studies the world carefully to learn how it works.

report

A written document that gives information about a subject.

solar panel

A flat board that turns sunlight into electricity.

tonne

A unit of weight equal to one thousand kilograms.

Level 2 — Elementary

Once a year, an international team measures how much carbon dioxide humanity has added to the atmosphere. The project is called the Global Carbon Budget, and the 2026 edition is its twenty-first. More than 140 scientists from over 90 institutions contributed, led by Professor Pierre Friedlingstein of the University of Exeter. The results were presented on 10 November at the COP31 climate conference in Antalya, Turkey.

The headline finding is a small decline. Emissions from fossil fuels and cement are projected to fall 0.4 percent in 2026, to 38.2 billion tonnes. That matters because such falls are rare. The authors are careful to add that the projection has a range, from a drop of 1.5 percent to a rise of 0.6 percent, so a small increase cannot be ruled out entirely.

Oil did most of the work. Oil emissions are down 1.9 percent, a reduction of 243 million tonnes of CO2, which is one of the largest annual falls ever recorded outside the pandemic year of 2020. The cause was price: petrol and diesel cost between 10 and 50 percent more than in 2025 in the United States, China and the European Union, after the conflict with Iran and the closure of the Strait of Hormuz disrupted supply. Coal emissions actually rose 0.6 percent, because strong growth in India offset sharp falls in Europe and America, and gas fell 0.3 percent.

Clean energy helped, but not enough on its own. Roughly two thirds of the growth in global energy demand was met by solar power, wind power and electric vehicles, and electric cars alone account for about a fifth of the drop in oil emissions. Yet total emissions including land use still reached a record 43.3 billion tonnes, up from 42.8 billion, because land-use emissions climbed from 4.4 to 5.1 billion tonnes after peat fires and forest damage in Southeast Asia. Carbon dioxide in the atmosphere now stands at 427.7 parts per million, 54 percent above pre-industrial levels.

atmosphere

The layer of gases surrounding the Earth.

emissions

Gases released into the air by human activity.

projection

An estimate of what a figure will be, based on current data.

institution

An established organisation such as a university or research centre.

range

The span between the lowest and highest likely values.

to offset

To cancel out part or all of an opposite effect.

land use

How humans use land, including clearing forests for farming.

parts per million

A way of measuring how concentrated a gas is in the air.

Level 3 — Intermediate

The Global Carbon Budget is the closest thing climate science has to an annual audit, and its twenty-first edition landed on 10 November with a finding that invites easy misreading. Produced by more than 140 researchers across more than 90 institutions, led by Professor Pierre Friedlingstein of the University of Exeter and published as a preprint in Earth System Science Data, the report projects that carbon dioxide from fossil fuels and cement will decline by 0.4 percent in 2026, to 38.2 billion tonnes. It was launched at a press conference at COP31 in Antalya, Turkey, where the temptation to call this a turning point was considerable.

The arithmetic underneath the headline resists that reading. The decline is dominated by oil, where emissions fell 1.9 percent, or 243 million tonnes, one of the largest single-year reductions in the entire record apart from 2020, 2009 and 1992. The cause was not policy but price. Petrol and diesel cost 10 to 50 percent more than in 2025 across the United States, China and the European Union, following the energy shock from the Iran conflict and the closure of the Strait of Hormuz. Coal, meanwhile, rose 0.6 percent, with Indian growth outweighing sharp reductions in the European Union and the United States, and gas slipped 0.3 percent. Coal still supplies roughly 41 percent of fossil emissions, oil 33 percent, gas 21 percent and cement about 2 percent.

Structural change is visible, but it is not yet decisive. Around two thirds of the increase in global energy demand was absorbed by solar power, wind power and electric vehicles, equivalent to 0.54 of the 0.78 billion tonnes of demand-driven growth, and electric vehicles alone delivered roughly a fifth of the fall in oil emissions. International aviation and shipping each declined 1 percent, aviation partly because of reduced activity in the Middle East. Working in the other direction, data centre expansion driven largely by artificial intelligence added about 0.03 billion tonnes in 2025. The regional picture is uneven: China, responsible for 32 percent of global fossil emissions, is down 0.9 percent; the United States down 1.3 percent; the European Union down 1.5 percent; Japan down 1.0 percent; and India up 4.4 percent, driven by a 5.9 percent rise in coal.

The most sobering numbers concern totals and time. Land-use emissions rose from 4.4 to 5.1 billion tonnes, pushed up by peat fires and forest degradation in Southeast Asia associated with El Nino conditions, so total carbon dioxide emissions including land use reached a record 43.3 billion tonnes, up from 42.8 billion. Atmospheric concentration is projected at 427.7 parts per million, 54 percent above pre-industrial levels, rising 2.2 parts per million this year compared with 2.1 last year and a record 3.8 in 2024. The remaining budget for a 50 percent chance of limiting warming to 1.5 degrees now stands at 150 billion tonnes counted from the start of 2027, which is under four years at current rates. Friedlingstein's own verdict is blunt: a reduction driven partly by geopolitical shocks is not a sustainable climate solution. Glen Peters of CICERO put the same point differently, noting that a one-year drop is not a peak unless structural change is behind it.

Level 4 — Advanced

The annual Global Carbon Budget functions as the discipline's balance sheet, and its twenty-first edition, released on 10 November as a preprint in Earth System Science Data and launched at a COP31 press conference in Antalya, carries a headline almost designed to be misunderstood. Compiled by more than 140 researchers at over 90 institutions under the direction of Professor Pierre Friedlingstein of the University of Exeter, the report projects a 0.4 percent decline in carbon dioxide from fossil fuels and cement during 2026, to 38.2 billion tonnes. For a delegation hall full of negotiators searching for encouragement, this is the first fall in years and an obvious candidate for a turning point. The authors have gone to some trouble to foreclose that interpretation, and the uncertainty band alone, spanning a 1.5 percent decrease to a 0.6 percent increase, means the direction of travel is not strictly settled.

Decomposing the decline explains the authors' caution. Oil accounts for almost all of it, falling 1.9 percent, or 243 million tonnes, which places 2026 among the largest annual contractions on record alongside 2020, 2009 and the Soviet dissolution in 1992. The driver was neither regulation nor technology but a price shock: petrol and diesel ran 10 to 50 percent above 2025 levels across the United States, China and the European Union in the aftermath of the Iran conflict and the closure of the Strait of Hormuz. Coal moved the other way, up 0.6 percent, as Indian growth of 5.9 percent overwhelmed contractions of 6.7 percent in American coal and 2.9 percent in European, while gas edged down 0.3 percent. The composition of the remaining 38.2 billion tonnes is roughly 41 percent coal, 33 percent oil, 21 percent gas and 2 percent cement after the carbonation sink is netted off. An emissions reduction purchased by making fuel unaffordable is not a policy achievement; it is a recession in disguise, and it reverses the moment supply normalises.

Beneath the shock, genuine structural progress is measurable without yet being sufficient. Clean deployment absorbed roughly 0.54 of the 0.78 billion tonnes of demand-driven growth, around 69 percent, with solar, wind and electric vehicles meeting about two thirds of the increase in energy demand, and electric vehicles alone contributing approximately a fifth of the oil decline. International aviation and shipping each fell 1 percent, aviation partly reflecting suppressed Middle Eastern activity. Pulling against this, data centre growth attributed largely to artificial intelligence added about 0.03 billion tonnes in 2025. Regionally, China, at 32 percent of global fossil emissions, declined 0.9 percent on a 4.6 percent fall in oil despite coal rising 0.3 percent; the United States fell 1.3 percent, the European Union 1.5 percent and Japan 1.0 percent, against India's 4.4 percent increase. Per-capita divergence remains stark, at 2.2 tonnes in India against 14.5 in the United States in 2025. A companion paper in Nature Climate Change, led by Professor Corinne Le Quere of the University of East Anglia, concludes that a sustained peak and decline could be a few years away, which is encouragement contingent on acceleration rather than a forecast of it.