A research preprint by Stefan Rahmstorf and Grant Foster reports evidence that the pace of global warming accelerated after 2015 when three major sources of natural temperature variability were removed from the record. The paper had not been peer reviewed by a journal when presented on Research Square, an important limitation on how its conclusion should be interpreted.
The study addresses a statistical question raised by a sequence of record-hot years: whether the underlying warming trend has accelerated, rather than temperatures simply being pushed temporarily higher by natural fluctuations. According to the authors’ abstract, previous analysis had not found acceleration at the conventional 95% confidence threshold because natural variability makes a short-term change in trend difficult to distinguish from noise.
Rahmstorf and Foster adjusted the temperature data for three influences: El Niño, volcanic activity and changes in solar output. Each can move global temperatures independently of the longer-term warming signal. El Niño can produce unusually warm conditions over a limited period, volcanic material can cool the planet by reducing incoming sunlight, and solar variation can add another smaller oscillation. The preprint’s stated aim was to reduce the effect of those factors before testing the recent rate of change.
After making those adjustments, the authors found that global temperature after 2015 increased significantly faster than in any previous 10-year period beginning in 1945 or later. The abstract describes the result as statistically significant, contrasting it with earlier work that could not cross the 95% threshold. The supplied page does not provide a numerical warming rate or quantify how much faster the recent period was, so no such estimate can be inferred from the headline alone.
The distinction between an observed run of hot years and a change in the underlying trend matters for climate assessment. A record can be driven by both long-term warming and short-lived natural conditions; separating those components affects whether recent temperatures are interpreted as an exceptional cluster or evidence of a steeper trajectory. The authors’ approach argues for the latter after adjustment, but its validity depends on the data choices, treatment of natural influences and statistical model described in the full study.
Research Square identifies the work as Version 1 and explicitly states that it is a preprint. The platform allows findings to be shared before journal review, meaning specialists have not yet completed the formal process of evaluating the methods and conclusions for publication. The page lists no competing interest and makes the paper available under a Creative Commons Attribution 4.0 licence.
The result is therefore a research claim, not a settled reassessment by a scientific body. It strengthens the case for examining recent acceleration while leaving peer review, replication and comparison with other temperature analyses as necessary next steps.
