Why is there so much oxygen?

Oxygen-producing photosynthesis evolved in cyanobacteria, and the isotopic evidence puts the rise in atmospheric oxygen between about 2.4 and 2.3 billion years ago, the Great Oxidation Event.

The marker is subtle: before the rise, sulphur isotopes in sediments fractionate in a way that only happens under ultraviolet light in an oxygen-free atmosphere. The disappearance of that signature is what dates the transition.

Oxygen then rose and fell again before settling low for more than a billion years, at a level still argued over: a 2014 review suggests perhaps less than a tenth of one per cent of today’s, and a 2021 study of how ancient organic matter weathered puts the minimum at 2 to 24 per cent.

It rose again from the Neoproterozoic, though on a 2023 consensus curve it reached 15 to 20 per cent of the atmosphere only in the Devonian.

A 2016 study dated the transition itself rather than bracketing it, using sulphur isotopes through a continuous sequence in three South African drill cores.

It places the disappearance of the oxygen-free signature at about 2.33 billion years ago and constrains the transition to less than ten million years, which is fast for a change of this kind.

A transition that fast fits models in which the atmosphere flips between two stable states rather than drifting between them. Whether the flip came from a rise in production or a fall in the sinks is left open.

Later work on the same South African succession reads what follows differently: oxygen went on crossing back and forth over the threshold after that date, over about two hundred million years in all, and stayed up only from about 2.22 billion years ago.

A 2023 reanalysis of those cores finds the returns to an oxygen-free atmosphere brief, perhaps tens of thousands of years each, and reads them as an unsteady in-between state rather than a clean flip.

Photosynthesis and respiration nearly cancel: what one produces, the other consumes. Oxygen only accumulates when organic carbon, and sulphur as pyrite, is buried before it can be oxidised, which takes it out of the cycle.

So the long-term oxygen level is set by burial and by the supply of reducing material from volcanoes and weathering, rather than simply by how much photosynthesis is happening.

Evidence for oxygen production hundreds of millions of years before the atmosphere changed implies that something consumed it for a very long time, though a minority of researchers argue that the metabolism appeared only at the Great Oxidation Event, leaving no delay to explain. Candidate sinks include reduced volcanic gases and dissolved iron in the ocean, whose supply may have waned as the continents grew, and the relative sizes of these are not agreed.

The gap between a metabolism appearing and its waste product accumulating has neither a settled length nor a settled explanation: estimates for when the metabolism appeared span well over a billion years.

The Neoproterozoic increase coincides with the snowball glaciations and with the appearance of animals, and each has been proposed as a cause of another: oxygen for animals, animals for oxygen, glaciation for oxygen and oxygen for glaciation.

The geochemical proxies for oxygen in this interval seem to contradict one another, which is the honest reason the story keeps changing.

Atmospheric oxygen has stayed within about a factor of two of today’s level for the last four hundred million years. The floor is read from charcoal, which needs roughly fifteen per cent oxygen to form and is in the record from the mid-Silurian on, with gaps whose meaning is uncertain.

Feedbacks involving phosphorus, fire, the slowing of plant growth at high oxygen and the burial of organic carbon are proposed to hold it there. Which of them does the work is not agreed, and the stability remains better observed than explained.