Scientists have long tracked rising global temperatures, but a new analysis argues the pace has quickened in a measurable way. The study reports an accelerated rate of global warming over the past decade, with recent warming nearly doubling the late-20th-century trend.
Researchers focused on the human-driven signal in the temperature record and worked to separate it from short-term natural swings. Those swings include El Niño and La Niña cycles, volcanic aerosols, and changes in solar output, which can temporarily boost or dampen surface temperatures.
Across five major global temperature datasets, the authors estimate that long-term warming ran at roughly 0.2°C per decade from about 1970 to 2015. They then find the rate increased to roughly 0.34-0.42°C per decade after an apparent shift around February 2013 to February 2014, depending on the dataset.
How Researchers Filtered Out “Noise” From Natural Variability
The analysis relies on a statistical approach that removes the estimated influence of three major natural drivers: ENSO, volcanic activity, and solar variation. The goal is not to deny those effects, but to avoid mistaking a temporary spike for a lasting change in the underlying trend.
To test robustness, the team applied the method to five independent global temperature records, including datasets from NASA, NOAA, the HadCRUT5 record, Berkeley Earth, and Copernicus ERA5. The adjusted series still shows 2023 and 2024 among the warmest years on record, even after the method trims part of the El Niño-related boost.
The authors also used two statistical checks for acceleration. One fits a curved trend through time; the other looks for a change point—the moment when the warming rate shifts. Both approaches point to acceleration at high confidence after the study filters out natural variability.
That finding addresses a key scientific question that has grown louder in the wake of recent record heat: did the planet simply experience an unusually hot stretch, or did the baseline trend itself steepen? The new work argues that the baseline changed.
What It Could Mean for the 1.5°C Threshold Before 2030
The Paris Agreement’s 1.5°C goal refers to long-term warming, often interpreted over multi-decade averages rather than a single hot year. Even so, a sustained increase in the warming rate can move the world closer to a durable exceedance sooner than expected.
Based on their acceleration estimates, the authors project that the long-term breach of 1.5°C could occur between 2026 and 2029, depending on the temperature dataset used. In the Copernicus ERA5-based calculation, the timing lands earlier than in several other records, largely because ERA5 shows a higher post-shift warming rate in its analysis.
Independent scientists who reviewed the work in media coverage describe the result as concerning, while also stressing uncertainty about whether the higher rate will persist. A temporary acceleration has precedent, and researchers note that continued monitoring will determine whether the shift reflects a lasting change or a shorter-lived episode amplified by variability.
The “Mystery” Behind the Speed-Up and Why It Matters
The study sits inside a broader debate about what drives the recent run of exceptional heat. Scientists agree that greenhouse gas emissions drive the long-term rise, but they continue to test how much recent extra warmth comes from internal variability versus shifts in the climate system’s response.
Some research has questioned whether the temperature record yet shows a clearly detectable “surge” without adjusting for natural variability, arguing that statistics can struggle to confirm acceleration in noisy time series. The new analysis emphasizes that the adjustment step clarifies the signal and makes acceleration easier to detect across multiple datasets.
The practical implication is straightforward: if the accelerated global warming rate holds, it compresses decision timelines. Countries will have less time to cut emissions enough to slow the trend, and planners will need to prepare for heat extremes that build on a higher baseline.