A study published in Geophysical Research Letters in March 2026 argues that the rate of global warming has risen sharply. The paper, by Grant Foster and Stefan Rahmstorf of the Potsdam Institute for Climate Impact Research, puts warming at about 0.35°C per decade since 2015, against an average of around 0.2°C per decade since the 1970s.

That is roughly three-quarters higher, and the authors describe it as the fastest decadal warming since the instrumental record began in 1880. If that pace were to hold, they argue, the world would pass 1.5°C of long-term warming, the central figure in the Paris Agreement, around 2030.

This is one study, not a settled consensus, and a respected line of argument holds that the underlying warming rate has not changed at all.

What the paper actually measured

The raw temperature record is noisy. El Niño and La Niña, volcanic eruptions, and the roughly eleven-year solar cycle all push global temperatures up and down by amounts that have nothing to do with the long-term trend. The authors removed estimates of those three influences from five global temperature datasets, following a method Foster and Rahmstorf first set out in 2011. What was left, they report, shows a statistically significant acceleration beginning around 2015.

That last point is the substance of the claim. Earlier attempts to detect acceleration in global temperatures had not reached the 95 per cent confidence threshold scientists usually treat as significant, which is why the question stayed open for years. The new paper says it clears that bar once the short-term noise is stripped out. The full paper, Foster and Rahmstorf, 2026, in Geophysical Research Letters, sets out the method.

Whether the threshold has truly been crossed is part of what other researchers dispute.

Why not everyone is convinced

Michael Mann, a climate scientist at the University of Pennsylvania, has pushed back. In comments reported by Carbon Brief and CNN, he said there is no evidence of acceleration over the past ten years, and that the recent run of record years reflects something more specific: a drop in industrial aerosol pollution, particularly sulphur, that had been masking part of the warming greenhouse gases were already producing. On his account the planet is warming at a roughly constant rate until carbon emissions reach zero, which he is quick to add is bad enough on its own.

The mechanism he points to is well documented. Sulphate aerosols reflect sunlight directly and help seed bright, reflective low clouds, so cutting them removes a cooling effect rather than adding a fresh source of heat.

Shipping is the clearest case. In 2020 the International Maritime Organization sharply lowered the permitted sulphur content of marine fuel, a public-health measure aimed at cutting air pollution. Sulphur dioxide emissions from large ships fell by around 80 to 85 per cent, the reflective “ship tracks” over busy ocean lanes thinned, and the cooling they had provided faded with them. Estimates of how much warming this unmasks vary widely, from a few hundredths of a degree to around 0.1°C this decade depending on the assumptions, with the effect concentrated in the Northern Hemisphere where most shipping runs. Carbon Brief has a useful analysis of the low-sulphur shipping rules and their climate effect.

The 2023 spike, and what it leaves unexplained

Much of the disagreement traces back to a single awkward year. 2023 came in at about 1.45°C above pre-industrial levels, far enough above the previous record that El Niño and steady greenhouse warming together struggle to account for the size of the jump.

A study by Goessling and colleagues in Science pointed to a record-low planetary albedo, the share of incoming sunlight Earth reflects back to space, driven mainly by a decline in low cloud cover over parts of the tropics and northern mid-latitudes. Whether that cloud loss is a passing fluctuation, a consequence of the aerosol cuts, or a feedback from warming itself is not resolved. The paper appears in Science.

Berkeley Earth, in its report on 2025, took a careful middle position. It placed 2025 as the third warmest year on record and kept the long-run trend since 1980 at about 0.2°C per decade. It treated the 2023 to 2025 spike as a likely sign of faster warming, cautioning that the past rate may no longer be a reliable predictor of the next decade, with natural and human causes tangled together in it. It also made a point that is easy to miss: in some ocean regions the sulphur cuts have largely run their course, which limits how much additional warming they can still unmask.

What “bringing forward deadlines” does and does not mean

It is worth being precise about the Paris Agreement, because single-year figures are routinely misread. 2024 was the first calendar year to average more than 1.5°C above pre-industrial temperatures. That is not the same as breaching the Paris target, which is defined over a multi-decade average rather than any one hot year.

The acceleration claim, if it holds, brings that multi-decade threshold closer. The Rahmstorf study puts it near 2030. A separate 2025 assessment, the Indicators of Global Climate Change report led by Piers Forster, estimated current warming at about 0.27°C per decade, roughly half again the rate of the 1990s and 2000s, and judged that the 1.5°C goal is now very likely to be missed while 2°C remains within reach if emissions fall.

On the question that matters most for policy, the two camps agree. Warming continues until emissions reach zero, and how fast it runs after that depends on how quickly that happens.

What to watch

The acceleration argument rests on a small number of recent years, which is exactly why it is contested. A run of cooler years would weaken it. Continued record warmth would strengthen it. The 2026 and 2027 figures, now that the 2023–24 El Niño has faded, will say more than any single paper. So will further work pinning down how much of the recent heat the shipping aerosol cuts and the cloud changes are actually responsible for.

Until that settles, the most defensible summary is a modest one. Warming is running at least as fast as expected, may be running faster, and shows no sign of slowing.