In a groundbreaking study published in Nature, researchers have extended our understanding of Earth’s atmospheric history by revealing detailed snapshots of greenhouse gas concentrations spanning over three million years. Utilizing ice samples recovered from the Allan Hills Blue Ice Area (BIA) in Antarctica, this study offers an unprecedented glimpse into the stability and variability of carbon dioxide (CO₂) and methane (CH₄) levels long before the advent of the well-documented glacial-interglacial cycles of the late Pleistocene epoch.
Antarctic ice cores have long served as invaluable archives of past atmospheric conditions, reliably recording fluctuations in greenhouse gases over the past 800,000 years. These cores have illuminated the rhythmic oscillations of glacial and interglacial periods, revealing strong correlations between greenhouse gas concentrations and global climate shifts. However, extending this valuable record beyond the 800,000-year threshold has remained challenging, due to the absence of continuous ice deposits dating further back in time.
The current investigation breaks new ground by drawing on shallow ice cores extracted from blue ice areas—unique sites where ancient ice is exposed and preserved near the surface through natural wind and sublimation processes. The Allan Hills BIA, situated within the East Antarctic Ice Sheet, provides a natural archive of ancient ice layers spanning as far back as 3.1 million years. By analyzing discrete ice samples within this span, the researchers sought to reconstruct the atmospheric CO₂ and CH₄ levels during significant climate epochs predating the ice core records traditionally relied upon.
A key outcome from this novel dataset is the overarching stability of atmospheric methane throughout this entire period. Contrary to some expectations, no significant trend in methane concentrations was observed from 3.1 to 0.5 million years ago. This finding suggests that methane-emitting processes, potentially including wetlands, permafrost thawing, and biomass burning, remained remarkably consistent during the late Pliocene and early Pleistocene climatic transitions.
Conversely, the study identifies a modest decline in atmospheric CO₂ concentrations, approximately 20 parts per million (ppm), between 2.9 and 1.2 million years ago. Post this interval, CO₂ levels plateaued, demonstrating relative stability with fluctuations constrained within ±10 ppm across the transformational mid-Pleistocene Transition. This transition reflects a pivotal period when Earth’s glacial cycles shifted from a 40,000-year periodicity to longer, 100,000-year cycles of ice volume change.
One of the largest challenges facing this research was accounting for postdepositional processes affecting the ice in blue ice areas, such as potential gas alteration due to biological activity or physical diffusion. To mitigate these complexities, the team applied sophisticated corrections based on stable carbon isotope ratios of trapped CO₂ (δ¹³C) to reconstruct more accurate atmospheric signals. Specifically, samples aged between 2.8 and 3.1 million years required adjustments to compensate for respiration effects that could otherwise skew CO₂ concentration estimates.
Intriguingly, after these isotope-based corrections, the mean atmospheric CO₂ concentrations during the late Pliocene emerge as indistinguishable from early Pleistocene values, averaging around 250 ± 10 ppm. This challenges prior assumptions that CO₂ concentrations were significantly higher in this period of global cooling and sea level decline and denotes a more complex connection between greenhouse gas levels and climate state than previously envisaged.
The research highlights the critical role of accumulation rates and climate-dependent biases in the interpretation of blue ice area records. Given that the recovered ice is not a continuous deposit, but rather consists of discrete, spatially separated samples, the resulting atmospheric reconstructions likely represent weighted averages over multiple glacial cycles. This weighting reflects differences in snow accumulation under varying climate conditions, which the study assumes to be relatively constant for approximation purposes.
By pushing ice core greenhouse gas measurements into the late Pliocene epoch, this study fundamentally expands the temporal range of direct atmospheric data. It provides climate scientists with more extensive empirical constraints on models of Earth’s climate system at a time when global temperatures were in secular decline and sea levels were falling dramatically. Such data are essential to elucidate the mechanisms underpinning critical climate transitions and the role of greenhouse gases therein.
The broader scientific community has long debated the interplay between atmospheric greenhouse gases and glacial cycles, especially surrounding the mid-Pleistocene Transition—a profound shift in Earth’s climatic rhythms. By revealing persistent stability in greenhouse gases over millions of years, the study calls for new perspectives on the contributions of orbital forcing, ice sheet dynamics, and feedback mechanisms to Earth’s ancient climate regimes.
Moreover, the Allan Hills BIA ice core record introduces a powerful proxy for interrogating ancient climate variability through direct measurement, complementing indirect methods such as marine sediment analysis and terrestrial proxies. Incorporating isotopic and concentration data from vast geological timescales will refine climate reconstructions and enhance predictive capabilities for future climate scenarios.
This work also underscores the adaptability and potential of blue ice area sampling techniques. Traditionally, ice core drilling has relied on deep, continuous cores that limit the temporal reach of greenhouse gas reconstructions. By contrast, blue ice areas offer unique windows into ancient atmospheres—albeit with complexities that require careful analytical corrections and modeling to unravel.
Overall, this research marks a transformative step in paleoclimatology, providing a robust, geochemically supported atmospheric greenhouse gas record extending over three million years. It crystallizes a vision of Earth’s climate system where greenhouse gas concentrations were broadly stable across major climatic transitions, a finding that challenges and refines current understanding of the drivers of long-term climate change.
As climate change continues to demand urgent attention, unraveling Earth’s ancient atmospheric history serves as a critical cornerstone for contextualizing modern and future greenhouse gas dynamics, offering vital lessons from epochs long past but geopolitically urgent for our planet’s future.
Subject of Research:
Greenhouse gas reconstructions from Antarctic blue ice areas informing long-term atmospheric variability and climate transitions over the past 3 million years.
Article Title:
Broadly stable atmospheric CO₂ and CH₄ levels over the past 3 million years.
Article References:
Marks-Peterson, J., Shackleton, S., Higgins, J. et al. Broadly stable atmospheric CO₂ and CH₄ levels over the past 3 million years. Nature 651, 647–652 (2026). https://doi.org/10.1038/s41586-025-10032-y
DOI:
https://doi.org/10.1038/s41586-025-10032-y (Published 19 March 2026)
Tags: Allan Hills Blue Ice Area researchancient ice core analysisAntarctic ice core climate recordsatmospheric greenhouse gas historycarbon dioxide variability in deep pastEast Antarctic Ice Sheet paleoenvironmentglacial-interglacial cycle precursorslong-term climate change proxiesmethane concentration in prehistoric atmospheremethane stability in ancient epochspre-Pleistocene greenhouse gas datastable CO2 levels over millions of years
