Somewhere in the lush, subtropical hills of South China, scientists pulled secrets from ancient stone. In layers of sediment compacted over eons, they found evidence of something shocking—millions of years ago, Earth’s natural systems belched enormous amounts of carbon dioxide into the air. The planet responded in kind: oxygen levels in the oceans plummeted, and life felt the squeeze.
Now, researchers say this ancient phenomenon may be repeating itself—except this time, the carbon isn’t rising from deep within the Earth, but from human hands.
A new study, led by a team from the University of California, Davis, the Chinese Academy of Sciences, and Texas A&M University, reveals that Earth experienced five major drops in oceanic oxygen levels between 310 and 290 million years ago, each following large-scale increases in atmospheric CO₂. The study, published in the journal Proceedings of the National Academy of Sciences, doesn’t just revisit ancient environmental changes—it draws a line straight from prehistory to the climate crises of today.
“This is our only analog for big changes in carbon dioxide at levels comparable to what we’re living in today,” said Dr. Isabel P. Montañez, senior author and Distinguished Professor at UC Davis. “And the kicker is: we’re doing it faster—hundreds of times faster.”
Carbon Dioxide and the Silent Drowning of the Seas
When Earth burped carbon dioxide 300 million years ago, it wasn’t just a gentle release. These were massive pulses—likely triggered by geological upheaval such as widespread volcanic activity. And each time it happened, it had a profound effect on the oceans.
Oxygen levels in the ocean dropped by 4% to 12%, based on new measurements. These might sound like small numbers, but the consequences for marine life were anything but. Marine biodiversity suffered—not in great extinction waves, but in stuttering pauses. The fossil record shows telltale slowdowns in evolution and diversification.
“We do see these pauses in biodiversity each time these burps happen,” Montañez said. “It had an impact. Most likely, coastal regions were impacted the most.”
These anoxic, or oxygen-depleted, events didn’t last for a moment—they stretched on for 100,000 to 200,000 years each. And while the Earth’s atmosphere back then held 40% to 50% more oxygen than today, it wasn’t enough to stop the oxygen drain from the seas. If it could happen then, Montañez warns, it can happen again—perhaps more easily this time.
A Tale Written in Isotopes
To unlock this buried story, the research team turned to the Naqing formation in South China—a fossil-rich, ancient marine deposit that holds an uninterrupted geological record from the late Carboniferous to early Permian periods.
Using carbonate uranium isotopes extracted from deep-sea sediment cores, scientists reconstructed the chemistry of long-vanished oceans. These isotopes act like time-stamped fingerprints, revealing how oxygen moved—or stopped moving—through marine systems over vast geological spans.
“Through that analysis, we see these ‘burps’ not just in carbon dioxide but in the ocean’s uranium isotope signature too,” said Montañez. “They’re totally aligned.”
To further test their interpretations, the team used advanced climate models, fed with all known geochemical data and run hundreds of thousands of times using supercomputers. The goal: simulate Earth’s ancient climate with as much realism as the data allows. What emerged from the modeling was clear—each spike in carbon dioxide was followed by a predictable drop in ocean oxygen.
A Faster, Human-Driven Future
Here’s where ancient geology collides with modern reality: the scale of those ancient CO₂ releases is comparable to what we’re seeing today. But while the past burps of carbon unfolded over millennia, the modern one is happening two to three orders of magnitude faster—in decades.
“We’re creating a burp now,” said Montañez, “and at a rate two, maybe three, orders of magnitude faster than in the past.”
That acceleration is a critical difference. While Earth’s natural systems once had time to absorb and adapt to rising CO₂, today’s rapid increases offer no such grace. With ocean oxygen already falling in some regions, and dead zones—areas devoid of marine life—expanding in coastal waters, many scientists see echoes of ancient trends in today’s headlines.
Coastal Consequences Looming
The study’s findings suggest that future ocean deoxygenation may hit coastal ecosystems hardest—the very areas that host the most productive fisheries, rich coral reefs, and some of the greatest biodiversity on the planet.
These regions are already vulnerable to human impacts: pollution, warming waters, and habitat loss. Add rapidly decreasing oxygen to the mix, and you have the potential for widespread ecological disruption, particularly in regions where human communities rely on marine resources for survival and economic stability.
If ancient burps caused local biodiversity to pause, what happens when the entire Earth’s system is thrown into fast-forward?
Lessons From a Hotter, Stranger Past
The late Carboniferous and early Permian periods were a transitional phase for Earth—similar to what we’re experiencing today. Glaciations were ending, ecosystems were reshaping, and carbon dioxide was on the move. The similarities make this study not just a retrospective, but a mirror.
“This is a huge discovery,” said Montañez. “Because how do you take an ocean sitting under an atmosphere with much more oxygen than today and permit this? The message for us is, don’t be so sure that we can’t do this again.”
Understanding the patterns of Earth’s deep past may offer the best clues for how the planet could respond to today’s changes. And while Earth will endure, the future for ecosystems—and for humanity—depends on whether we heed the geological warnings that the stones have carried for hundreds of millions of years.
The Earth has spoken before. The question is, are we listening now?
Reference: Chen, Jitao et al, Repeated occurrences of marine anoxia under high atmospheric O2 and icehouse conditions, Proceedings of the National Academy of Sciences (2025). DOI: 10.1073/pnas.2420505122. doi.org/10.1073/pnas.2420505122
