Over roughly the past 60 million years, Earth moved from the intense warmth of the early Cenozoic toward a cooler climate with permanent polar ice. Atmospheric carbon dioxide fell substantially along the way. Yet reconstructing where that carbon went has remained one of the harder accounting problems in Earth science.

A 2026 study points to an overlooked connection among shorelines, phosphate and microscopic ocean life. As sea level rose and fell, it changed the area of shallow continental shelves. Those shelves determined how much phosphate was trapped in coastal sediment before the nutrient could reach the open ocean. Phosphate availability then influenced marine productivity and the fraction of organic carbon that survived long enough to be buried.

The researchers describe the result as a feedback in Earth’s long-term carbon cycle. Calling it a planetary thermostat captures the self-regulating part of the idea, but not a fixed temperature setting. The mechanism could first strengthen carbon burial as seas fell, then weaken once sea level became too low for oxygen-poor water to interact with the right shelf sediments.

Every cell requires phosphorus. In the ocean, biologically available phosphate helps determine how much new organic matter plankton can make. It is not the only limiting nutrient, and adding phosphate alone would not make every patch of sea productive. Light, nitrogen, iron, temperature and ecology all matter. Across geological timescales, however, the size of the marine phosphate reservoir can place a powerful constraint on biological production.

Photosynthetic plankton take carbon from dissolved carbon dioxide and build it into cells. Most of that organic carbon is quickly recycled. Organisms respire it, animals eat it, and microbes decompose it, returning carbon dioxide to the water and eventually the atmosphere. Only a small share sinks to the seafloor and avoids complete decay.

That small surviving share is important. Once organic carbon is sealed into sediment, it leaves the active ocean-atmosphere system for a long time. Sustained changes in burial can therefore alter atmospheric carbon dioxide over millions of years. An earlier ScienceBlog report on ocean stability, productivity and carbon burial describes the same broad biological pump in the modern ocean, although the new work asks a much deeper-time question.

The study in Proceedings of the National Academy of Sciences was led by Rosalind Rickaby at the University of Oxford, working with Thomas Wood, Zunli Lu and Christian Bjerrum. Rather than treating one sediment core as a global history, the team combined a newly compiled dataset with several different geochemical reconstructions.

First, carbon-isotope mass balance was used to estimate the proportion of carbon buried as organic matter rather than carbonate. Second, phosphorus accumulation in deep-sea sediments supplied evidence about how efficiently the nutrient was being removed from seawater. Third, iodine-to-calcium ratios measured in fossil foraminifera acted as a proxy for oxygen conditions in the ancient water column.

The Oxford record for the paper summarizes the pattern: organic carbon burial was suppressed during the Eocene hothouse, when phosphate in the open ocean was scarce and the water column was comparatively oxygenated. Burial generally increased toward the present as Earth cooled.

No one proxy measures all three processes directly. Carbon-isotope calculations require assumptions about sources and sinks. Sediment coverage is uneven across ocean basins, and oxygen proxies must be interpreted within their chemical context. The case comes from the agreement among records, not from a single perfect thermometer or carbon meter.

During the Eocene, high sea level flooded broad areas along continental margins. These shallow shelf seas accumulated sediment rapidly. Phosphate delivered by weathering on land and carried through rivers could be captured in coastal mud before circulating through the wider ocean.

This creates a counterintuitive possibility. More flooded ocean area did not necessarily support more marine production. By expanding the zone where phosphate was efficiently buried, high sea level could starve the open ocean of a nutrient needed for life. Less phosphate meant less new organic matter, less material sinking to the bottom and less organic carbon ultimately preserved.

The University of Oxford’s account of the research uses the image of a grime ring left as water drains from a bathtub. When the waterline moves down, the active zone of coastal sedimentation also moves downslope and occupies a smaller area. In the proposed Cenozoic feedback, shrinking that zone allowed a greater proportion of phosphate to escape into open water.

Saying that high seas allowed more carbon dioxide to remain in the atmosphere does not mean sea level single-handedly set atmospheric CO₂. It means that, all else being equal, weaker burial removed carbon from the active system more slowly. Volcanic emissions, chemical weathering and other carbon fluxes continued at the same time.

As Earth cooled and sea level declined, the broad shelf trap contracted. More phosphate remained available to ocean life. Greater productivity created more sinking organic material, and microbes consumed dissolved oxygen as they decomposed it. Oxygen-minimum zones became more extensive.

Low oxygen can alter the fate of phosphorus in sediment. Under well-oxygenated conditions, reactive iron minerals help retain phosphate. When bottom waters and pore waters lose oxygen, those minerals can dissolve or change chemically, releasing phosphate back into seawater. Recycled nutrient can support another round of plankton growth.

The result is an amplifying loop. More phosphate permits more production. More organic matter sinks and drives greater oxygen consumption. Lower oxygen promotes phosphate recycling. Recycled phosphate sustains still more production, increasing the amount of organic carbon available for burial and gradually drawing carbon dioxide out of the ocean-atmosphere system.

This broad framework fits modern understanding that marine phosphorus is governed by interacting biological, mineral and redox processes, as reviewed in Nature Reviews Earth & Environment. The new study’s contribution is to connect those processes to changing shelf geometry across most of the Cenozoic.

If the loop only grew stronger as seas fell, it could appear to drive unlimited cooling. The reconstruction instead suggests a natural limit. The authors estimated the greatest organic carbon burial when sea level stood roughly 20 to 40 metres above the modern level.

At that intermediate position, oxygen-minimum zones could overlap carbon-rich sediment on the continental shelf. Oxygen concentrations below about 90 micromoles per kilogram favoured phosphate recycling, keeping the nutrient available and sustaining elevated burial for more than a million years.

When sea level fell further, the top of the oxygen-minimum zone could sit below the shelf break. Oxygen-poor water then lost contact with the shelf sediment best placed to return phosphate to the water. Recycling weakened, and so did the carbon burial feedback. In other words, the thermostat worked through a sweet spot rather than a simple rule that lower water always meant more cooling.

The idea was not invented after seeing the new data. A 2006 ocean chemistry modelling study predicted that sea-level changes could modify the marine phosphate inventory, biological productivity and organic carbon burial. The 2026 paper argues that carbon, phosphorus and oxygen records now reveal the relationship expected from that earlier model.

Sea level was both a participant in the feedback and an outcome of a changing climate. Cooling can grow ice sheets and lower the ocean, while tectonic changes can alter the volume of ocean basins over longer intervals. The study does not reduce 60 million years of climate history to one external sea-level dial.

Nor does it argue that phosphate feedback was the only route by which the Cenozoic atmosphere lost carbon dioxide. Volcanic outgassing, mountain building, silicate rock weathering, ocean circulation, ecosystem evolution and the burial of carbonate all contributed. The authors’ narrower claim is that marine organic carbon burial may have played a larger role in long-term cooling than previously recognized.

The timescale is equally important. The proposed feedback unfolded over hundreds of thousands to millions of years. It is not a mechanism capable of cancelling the rapid modern rise in greenhouse gases. NASA’s summary of climate evidence attributes the current warming trend to human activity and documents a pace far faster than this sedimentary regulator operates.

So the word “thermostat” deserves humility. Earth did not remain at one temperature, and its feedbacks did not prevent enormous climate shifts. What the study offers is a plausible self-limiting connection: coastlines changed the burial of a nutrient, the nutrient changed ocean life and oxygen, and a fraction of dead life changed the amount of carbon left in the atmosphere.

No single step looks planetary on its own. Linked together and repeated across 60 million years, they may help explain how Earth moved from a greenhouse world toward an icehouse without continuing all the way into permanent global glaciation.