2026-02-23 マサチューセッツ工科大学（MIT）
Researchers detected statistically significant changes in global atmospheric temperatures in response to three major natural events: the eruption of Mount Pinatubo in 1991 (inset on left), the Australian wildfires in 2019-2020 (center), and the eruption of the underwater volcano Hunga Tonga in the South Pacific in 2022 (bottom right).Credit: MIT News; Richard P. Hoblitt, USGS (Pinatubo); iStock (wildfire); NASA GOES-West (Hunga Tonga); NASA’s Goddard Space Flight Center (background)
＜関連情報＞
- https://news.mit.edu/2026/study-reveals-climatic-fingerprints-wildfires-and-volcanic-eruptions-0223
- https://www.pnas.org/doi/10.1073/pnas.2525500123
山火事や火山噴火に対する地球規模の気温反応の検出 Detectable global temperature responses to wildfires and volcanic eruptions
Yaowei Li, Benjamin D. Santer, Susan Solomon, +1 , and Qiang Fu
Proceedings of the National Academy of Sciences Published:February 23, 2026
Significance
Major volcanic eruptions and wildfires provide natural experiments for testing how Earth’s atmosphere responds to sudden disturbances. While the tropospheric cooling and stratospheric warming effects of sulfate particles from the 1991 Pinatubo eruption are well established, the climate impacts of wildfire smoke and volcanic water vapor have been far less clear. Here, we show that the 2019-2020 Australian wildfires and the 2022 Hunga Tonga eruption also produced statistically detectable global atmospheric temperature signals. Using a robust detection framework, we evaluate whether, when, and for how long these signals can be distinguished from internal climate variability (“noise”). Our results highlight emerging forcings that must be included in climate models to better interpret past variability and anticipate impacts of future perturbations.
Abstract
Large volcanic eruptions and intense wildfires perturb Earth’s atmospheric temperature. Understanding the climate response to such natural forcings is essential for obtaining reliable estimates of its response to anthropogenic greenhouse gas emissions. While the climate impacts of volcanic sulfate aerosols are well documented, other natural forcings—including wildfire smoke reaching the stratosphere and water vapor injections from a submarine eruption—pose new challenges for detecting and attributing their atmospheric temperature impacts. Here, we demonstrate robust detection of statistically significant temperature anomalies in the troposphere and stratosphere using multidecadal satellite observations and internal variability estimates from a climate model ensemble and from observations. We analyze three landmark events: the 1991 Pinatubo eruption, the 2019-2020 Australian wildfires, and the 2022 Hunga Tonga eruption. Each leaves a fingerprint with distinct altitudinal, geographical, and temporal structure. The global-mean stratospheric signal from Australian wildfires is detectable even in time averages extending beyond 10 mo, despite injecting only ~5% of Pinatubo’s aerosol mass. For Hunga Tonga, we detect significant and prolonged stratospheric cooling, but no robust tropospheric signal in the first 2 y. These findings show that both sulfate and nonsulfate stratospheric perturbations produce distinct, statistically identifiable global temperature signals. Accounting for such forcings in climate model simulations is therefore essential for improving comparisons of simulated and observed variability.
