Earth’s Warming Rate Has Nearly Doubled Since 2015

The numbers had been there all along, buried in five separate global temperature records stretching back to 1880. What kept scientists from seeing the pattern wasn’t a lack of data. It was noise.

Every time the planet lurches through an El Niño cycle, global temperatures jump. Every time a volcano punches a column of sulphur dioxide into the stratosphere, they dip. Solar output waxes and wanes on an eleven-year cycle, adding its own gentle wobble. These aren’t small perturbations; they’re big enough to swamp a real underlying signal, to make a genuine acceleration in warming look like ordinary fluctuation if you don’t know where to look. For years, that’s roughly what happened: researchers suspected warming might be speeding up since around 2015, but nobody could demonstrate it to the 95% statistical confidence that science typically demands. The signal was there. The noise was louder.

Grant Foster, a US statistics expert, and Stefan Rahmstorf at the Potsdam Institute for Climate Impact Research decided to do something about that. Their approach was, in retrospect, almost straightforward: subtract the noise. Using established models for El Niño variability, volcanic aerosol loading, and solar irradiance, they estimated the temperature contribution of each natural factor in each month going back over a century, then removed those contributions from all five major global temperature records (NASA, NOAA, HadCRUT, Berkeley Earth, and the ERA5 reanalysis). What remained was a cleaner look at the underlying human-driven trend.

The result was unambiguous. “The adjusted data show an acceleration of global warming since 2015 with a statistical certainty of over 98 percent, consistent across all data sets examined and independent of the analysis method chosen,” says Rahmstorf. In all five datasets, the acceleration begins to emerge in 2013 or 2014. Not a blip, not an artefact. A persistent shift in the rate at which the planet is heating.

How big a shift? From 1970 to 2015, global temperatures rose at roughly 0.2°C per decade, a pace that was itself already alarming. Over the past ten years, the adjusted data suggest the rate has nearly doubled, to about 0.35°C per decade, depending on the dataset. “We can now demonstrate a strong and statistically significant acceleration of global warming since around 2015,” Foster says. That warming rate, if you’re looking for a comparison, is higher than in any decade recorded since instrumental measurements began in 1880.

Even 2023 and 2024, which generated considerable debate (some attributed their record temperatures almost entirely to an unusually strong El Niño), remain the two warmest years on record even after El Niño’s contribution is stripped out. The El Niño story was real. It just wasn’t the whole story.

What the paper does not do is explain why warming has accelerated. That question, which is frankly the more unsettling one, is left open. Climate models do predict that warming rates should increase as atmospheric CO₂ concentrations rise; the feedback loops baked into the physics of the climate system mean faster warming begets more warming. So an acceleration is not, in itself, unexpected. Whether the specific post-2015 step-change has a particular cause beyond the steady accumulation of greenhouse gases is not something this study can say. “We filter out known natural influences in the observational data, so that the ‘noise’ is reduced, making the underlying long-term warming signal more clearly visible,” Foster explains, describing the methodology rather than its implications. The causes are a problem for another paper.

The implications, though, follow fairly directly from the numbers. The Paris Agreement set a goal of limiting long-term warming to 1.5°C above pre-industrial levels. Climate scientists have been tracking how close we’re getting; global mean temperature is now reckoned at roughly 1.3°C above that baseline. At the pace of 0.35°C per decade that the adjusted data show, the 1.5°C threshold would become a permanent fixture rather than a target before 2030. “If the warming rate of the past 10 years continues, it would lead to a long-term exceedance of the 1.5°C limit of the Paris Agreement before 2030,” Rahmstorf says bluntly.

The study doesn’t stop at that observation either. “How quickly the Earth continues to warm ultimately depends on how rapidly we reduce global CO₂ emissions from fossil fuels to zero.” That’s perhaps the plainest statement of where things stand. The statistical analysis has done its job: removed the noise, confirmed the signal, quantified the rate. What happens to that rate is, from here, a policy question rather than a scientific one.

There’s something almost odd about how the result was obtained. You might expect a paper about accelerating climate change to open with satellite data, or ice cores, or some atmospheric measurement campaign. Instead it starts with subtraction; with the quiet, slightly painstaking work of estimating what the planet’s temperature would have been in a world without El Niño, without Pinatubo, without the solar maximum of the early 2000s. The signal Foster and Rahmstorf were looking for isn’t dramatic in itself. It’s the difference between 0.2°C per decade and 0.35°C per decade. A number that took 140 years of thermometer readings and a careful statistical scrub to extract cleanly.

Whether that acceleration continues, plateaus, or (in more optimistic scenarios) reverses as renewable energy deployment scales up will be, in some sense, the central empirical question of the next decade. The thermometers are already in place. The noise-removal methodology now exists. We’ll know the answer, one way or another.

Study link: https://agupubs.onlinelibrary.wiley.com/doi/abs/10.1029/2025GL118804