Forty kilometers above Earth’s surface, the air has been quietly losing heat for decades. A study published in Nature Geoscience in early 2026 puts a number on that loss: roughly 2 degrees Celsius since the mid-1980s, a cooling rate the researchers estimate is more than 10 times what natural processes alone would produce. The finding offers one of the sharpest physical signatures of human-driven climate change ever documented, written not in surface thermometers or melting ice but in the thinning energy of the upper atmosphere itself.
The research, led by Ben Santer and colleagues at the Lawrence Livermore National Laboratory, was distributed through the AAAS platform EurekAlert and detailed in the Nature Geoscience paper. The team developed a quantitative theory connecting rising CO₂ concentrations directly to falling stratospheric temperatures. The mechanism is straightforward in principle: as greenhouse gases accumulate in the lower atmosphere, they trap outgoing infrared radiation closer to the surface, reducing the energy that reaches the stratosphere. Meanwhile, CO₂ molecules in the thin air above 20 kilometers radiate heat to space more efficiently than they absorb it from below. The net result is a lower atmosphere that warms and an upper atmosphere that cools, a vertical contrast that acts like a fingerprint unique to greenhouse gas forcing.
Multiple instruments tell the same story
The 2°C figure does not rest on a single dataset. A synthesis of upper-air observations spanning 1979 through 2018, drawing on satellite microwave sounders, weather balloon radiosondes, and GPS radio occultation measurements, found stratospheric cooling of roughly 1 to 3 Kelvin across that period. Each instrument type carries its own biases and sampling quirks, so when all three converge on the same signal, confidence in the trend’s reality rises sharply. The 2°C value cited in the new study, focused on the period since the mid-1980s, sits squarely within that range.
Government records reinforce the picture. NOAA’s National Centers for Environmental Information maintains a lower-stratosphere temperature record built from MSU and AMSU satellite instruments and calibrated with GPS radio occultation data, covering 1986 through 2019. That record shows a clear downward slope in global-mean temperature, punctuated by short-lived spikes from volcanic eruptions but never reversing course over the long term.
The 2022 Scientific Assessment of Ozone Depletion, conducted under the Montreal Protocol, confirmed continued cooling in the middle and upper stratosphere and attributed it primarily to well-mixed greenhouse gases, especially CO₂, along with evolving ozone concentrations. An earlier 2006 assessment had documented a decrease of approximately 0.5 Kelvin per decade in global-mean lower-stratospheric temperature over 1979 through 2004, while noting a slowdown in that cooling after the late 1990s.
Separating CO₂ from the ozone signal
That slowdown is important context. During the 1980s and early 1990s, the destruction of stratospheric ozone, particularly over Antarctica, amplified cooling in the upper atmosphere. Ozone absorbs ultraviolet radiation and warms its surroundings; remove it, and temperatures drop further. After the Montreal Protocol curbed ozone-depleting chemicals and the ozone layer began stabilizing, that extra cooling influence faded, and the overall rate of stratospheric temperature decline moderated.
What makes the new Nature Geoscience study significant is its effort to disentangle the CO₂ contribution from the ozone effect. By combining radiative-transfer theory with observed temperature trends, Santer’s team isolated the portion of cooling attributable to carbon dioxide alone. That separation is what supports the claim that CO₂-driven cooling exceeds the natural baseline by a factor of more than 10. Earlier analyses that lumped greenhouse gases and ozone changes together could not make that distinction with the same precision.
The vertical structure of the trend also rules out a competing explanation. If increased solar output were driving recent warming, both the troposphere and the stratosphere would warm together, because a brighter Sun adds energy throughout the atmospheric column. Instead, the observed pattern shows pronounced warming near the surface and sustained cooling above. That contrast is difficult to explain without invoking greenhouse gas accumulation.
Gaps in the record and open questions
Several uncertainties temper how far these results can be pushed. The most recent comprehensive government dataset for lower-stratosphere temperatures covers through 2019, and the multi-dataset synthesis examines data through 2018. That leaves a gap of several years during which transient events could have temporarily altered the cooling trajectory. The January 2022 eruption of Hunga Tonga-Hunga Ha’apai, which injected an unprecedented volume of water vapor into the stratosphere, is one such event. Major eruptions can temporarily warm the stratosphere by introducing aerosols that absorb and scatter sunlight, and the recovery can take years. Without a peer-reviewed update that accounts for these disturbances, the precise cooling rate over the most recent period remains an open question.
Regional variation adds another layer of complexity. The verified data describe global averages, but the stratosphere does not cool uniformly. Polar regions, where ozone depletion was most severe, followed different temperature paths than tropical latitudes. Shifts in the polar vortices, changes in planetary wave activity, and evolving circulation patterns all shape how radiative forcing translates into local temperature change. The available primary sources do not break out regional cooling rates in enough detail to confirm whether the 10-times multiplier applies everywhere or primarily in certain latitude bands.
Projections carry their own uncertainty. The Nature Geoscience study builds a theoretical link between CO₂ and stratospheric temperature, but the sources reviewed here do not offer specific numerical forecasts for how much additional cooling to expect under different emissions pathways. Generating those projections would require coupling the new theory with climate model simulations that account for evolving greenhouse gas concentrations, ozone recovery, volcanic activity, and circulation changes.
Why a colder stratosphere matters on the ground
Stratospheric cooling is not an abstraction confined to the upper atmosphere. Temperature changes at those altitudes influence the formation of polar stratospheric clouds, which play a direct role in ozone chemistry. A colder stratosphere can promote the formation of these clouds even as ozone-depleting substances decline, potentially slowing the recovery of the ozone layer over polar regions.
The effects ripple downward. The temperature gradient between the stratosphere and the troposphere helps govern the strength and position of the jet stream, the river of fast-moving air that steers weather systems across the mid-latitudes. Changes in that gradient can alter the frequency of sudden stratospheric warming events, episodes in which the polar vortex weakens or collapses and sends blasts of Arctic air toward lower latitudes. Researchers have linked some of the most disruptive winter cold outbreaks in North America and Europe to these stratospheric disruptions.
The sources reviewed here acknowledge these connections but do not quantify how much additional stratospheric cooling might shift the odds of specific weather outcomes. Those links remain an active area of research. What the evidence does establish, with high confidence across multiple independent measurement systems, is that the cooling trend is real, that it matches what physics predicts from rising CO₂, and that it constitutes one of the most unambiguous markers of a planetary energy balance being reshaped by human activity. The signal is not subtle. It is large, persistent, and visible from space.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
