Earth’s atmosphere has not always carried the roughly 21 percent oxygen we breathe today. For most of the planet’s history, oxygen levels were far lower, and during one remarkable stretch about 300 million years ago they climbed significantly higher, possibly reaching around 30 percent. The question in the title actually gets the story backwards in an interesting way: the atmosphere spent its first two billion years with almost no free oxygen at all, then spiked, dipped, spiked again, and eventually settled into the range we know now. Understanding why oxygen rose and fell means following a trail of microbes, forests, volcanoes, and the slow churn of tectonic plates.
The First Two Billion Years Were Nearly Oxygen-Free
Earth formed about 4.5 billion years ago, and for roughly the first half of its existence the atmosphere contained virtually no free oxygen. The early air was dominated by nitrogen, carbon dioxide, methane, and water vapor. Any oxygen molecules that appeared were immediately gobbled up by chemical reactions with iron, sulfur, and volcanic gases. This was not a minor imbalance; the sinks for oxygen were so large that the gas simply could not accumulate.
The organisms that would eventually change everything, cyanobacteria, evolved the ability to split water molecules and release oxygen as a byproduct of photosynthesis. Genetic analyses suggest cyanobacteria originated deep in the Archean eon, perhaps around 3.6 billion years ago, well before oxygen started building up in the atmosphere.1PubMed Central. Timing the Evolutionary Advent of Cyanobacteria and the Later Great Oxidation Event Using Gene Phylogenies of a Sunscreen Molecular clock studies also support an Archean origin for cyanobacteria and indicate they had already developed multicellular forms before oxygen finally rose.2PubMed Central. Cyanobacteria and the Great Oxidation Event: evidence from genes and fossils So photosynthetic oxygen was being produced for possibly a billion years before it could gain a lasting foothold in the air. That enormous gap tells us something important: producing oxygen and accumulating oxygen are two very different problems.
Oxygen Oases Before the Big Rise
During that long gap, oxygen did not simply vanish the instant cyanobacteria released it. In certain shallow, nutrient-rich coastal waters, small pockets of dissolved oxygen built up locally even while the rest of the ocean and atmosphere remained oxygen-free. Researchers call these “oxygen oases,” and geochemical evidence from rocks in places like Steep Rock Lake in Canada shows that shallow-water limestones about 2.8 billion years old formed in oxygenated seawater, while deeper sediments nearby were laid down in water containing no oxygen at all.3Precambrian Research. Identification of an Archean marine oxygen oasis
Modeling work suggests these oases were modest, with typical dissolved oxygen concentrations of only about one to three micromolar, compared to over 200 micromolar in well-mixed modern surface waters.4Chemical Geology. Quantifying the areal extent and dissolved oxygen concentrations of Archean oxygen oases Recent work has traced these episodes to pulses of phosphorus recycling in the ocean: when bottom waters became more sulfide-rich, sediments released phosphorus, which fertilized photosynthetic life and temporarily boosted local oxygen production. These intervals predate the first global atmospheric oxygen accumulation by about 500 million years.5Nature Geoscience. Archaean oxygen oases driven by pulses of enhanced phosphorus recycling in the ocean The oases were spatially restricted and temporary, but they were the proving grounds where oxygen-dependent chemistry and biology could begin to develop.
The Great Oxidation Event
About 2.4 billion years ago, atmospheric oxygen finally crossed a threshold and began accumulating permanently. This transition, known as the Great Oxidation Event, was not a single afternoon of drama. It unfolded over tens to hundreds of millions of years, but the result was unmistakable in the rock record: certain sulfur isotope signatures that can only form in an oxygen-free atmosphere disappeared and never came back.
Why did it take so long? A major part of the answer lies underground. Earth’s mantle was more chemically reducing in the Archean, meaning volcanic gases released at the surface were rich in hydrogen, methane, and hydrogen sulfide, all of which react with and consume oxygen. Quantitative modeling of mantle data shows that these reducing volcanic emissions could have prevented atmospheric oxygen from accumulating until roughly 2.5 billion years ago with very high probability.6PubMed Central. Mantle data imply a decline of oxidizable volcanic gases could have triggered the Great Oxidation Over time, the mantle itself became more oxidized. As deeply buried oxidized material mixed upward, volcanic gases shifted toward more oxidized species like carbon dioxide and sulfur dioxide rather than methane and hydrogen sulfide. That shift removed a massive oxygen sink and allowed the gas to accumulate in the atmosphere.7PubMed Central. Destabilization of deep oxidized mantle drove the Great Oxidation Event
Even after the Great Oxidation Event, oxygen levels were still far below what we enjoy today. Estimates for the mid-Proterozoic, sometimes called the “boring billion” because of its apparent biological and chemical stability, have been surprisingly low. Fluid inclusions trapped in 1.4-billion-year-old salt crystals suggest atmospheric oxygen around 7.84 millibar, which works out to a few percent of the atmosphere at best, well below today’s roughly 210 millibar.8PubMed Central. Breathing life into the boring billion: Direct constraints from 1.4 Ga fluid inclusions reveal a fair climate and oxygenated atmosphere Similar halite-based measurements from 815-million-year-old rocks in Australia indicate oxygen around 10 percent of the atmosphere by volume.9Gondwana Research. Atmospheric gas in modern and ancient halite fluid inclusions: A screening protocol In other words, for more than a billion years after the Great Oxidation Event, oxygen hovered at levels that would leave a modern human gasping.
A Second Oxygen Surge and the Rise of Animals
A second major oxygenation pulse occurred in the Neoproterozoic, roughly 800 to 540 million years ago, bringing oxygen closer to modern levels. The triggers were partly tectonic: the breakup of the supercontinent Rodinia, combined with massive volcanic eruptions that deposited phosphorus-rich rocks in the tropics, may have flushed an unprecedented amount of bioavailable phosphorus into the ocean. That fertilizer would have driven photosynthesis and organic carbon burial on a scale the Earth had not seen before.10Geochemistry, Geophysics, Geosystems. Did phosphorus derived from the weathering of large igneous provinces fertilize the Neoproterozoic ocean?
The rise of oxygen during the Neoproterozoic dovetailed with the appearance of complex animal life. Atmospheric oxygen levels have broadly tracked biodiversity and body size through Earth’s history, enabling organisms to move from water to land and eventually to the air.11PubMed Central. Evolution of air breathing: oxygen homeostasis and the transitions from water to land and sky This was also when Earth’s nitrogen cycle was reshaped: the appearance of widespread nitrate in oxygenated oceans opened new metabolic pathways and may have helped cyanobacteria and early eukaryotic algae diversify rapidly.12Nature. Onset of the aerobic nitrogen cycle during the Great Oxidation Event While the precise oxygen thresholds that enabled animal body plans are still debated, the correlation between rising oxygen and the Cambrian explosion of animal diversity is hard to ignore.
The Carboniferous Peak and Why It Happened
If any single period answers the title question most dramatically, it is the late Carboniferous and early Permian, roughly 350 to 280 million years ago. Oxygen may have reached around 30 percent of the atmosphere, compared to today’s 21 percent. The principal driver was the rise of large vascular land plants, which massively increased the global burial of organic carbon.13Annual Review of Earth and Planetary Sciences. Phanerozoic Atmospheric Oxygen Even earlier land plants, simple non-vascular forms that colonized land surfaces from about 470 million years ago onward, had already begun pushing oxygen higher by boosting organic carbon burial.14PubMed Central. Earliest land plants created modern levels of atmospheric oxygen
Here is the key mechanism in plain terms. When a plant grows, it pulls carbon dioxide from the air and builds that carbon into wood, leaves, and roots. If the dead plant material decays, the carbon returns to the atmosphere and no net oxygen is gained. But if the dead material gets buried in sediment before it can decompose, say in a waterlogged swamp, the carbon stays locked away and the oxygen that was released during photosynthesis stays in the air. During the Carboniferous, vast tropical swamp forests covered low-lying continental interiors, creating ideal conditions for organic matter to pile up and form the coal deposits that give the period its name.
Another contribution came from pyrite burial. When sulfate-reducing bacteria produce sulfide in ocean sediments, that sulfide can react with iron to form pyrite. Burying pyrite removes sulfur and iron from the active cycle, and the net chemical effect is equivalent to adding oxygen to the atmosphere.15PubMed Central. Sulfur isotopes in coal constrain the evolution of the Phanerozoic sulfur cycle During periods when both organic carbon burial and pyrite burial were high, oxygen accumulated rapidly.
The Fungal Lag Debate
A popular explanation you may have encountered claims that Carboniferous coal accumulated because fungi had not yet evolved the enzymes needed to break down lignin, the tough structural polymer in wood. Under this idea, dead trees piled up for millions of years because nothing could digest them, and once white-rot fungi finally evolved ligninolytic enzymes, decomposition caught up and coal formation dropped. The timing is suggestive: a 2012 genomic study of 31 fungal genomes found that the origin of enzymatic lignin decomposition roughly coincided with the decline in organic carbon burial near the end of the Carboniferous.16PubMed. The Paleozoic origin of enzymatic lignin decomposition reconstructed from 31 fungal genomes
However, a later assessment challenged this narrative, arguing on the basis of phylogenomic, geochemical, paleontological, and stratigraphic evidence that a delayed evolution of lignin-degrading fungi did not actually cause the Paleozoic peak in coal production.17PubMed Central. Delayed fungal evolution did not cause the Paleozoic peak in coal production The alternative view points to tectonic and climatic factors: the arrangement of continents, the extent of tropical wetlands, and sea level changes determined where and how much organic material could be buried, regardless of how quickly fungi could decompose it. The debate is far from settled, but the simple version of the story, “nothing could eat wood,” is probably too neat.
The Post-Carboniferous Crash
Oxygen did not stay at 30 percent. Over a roughly 20-million-year stretch spanning the late Permian and into the early-to-middle Triassic, atmospheric oxygen dropped dramatically. One widely cited modeling study calculates a continuous decline from around 30 percent to about 13 percent.18Geochimica et Cosmochimica Acta. The carbon and sulfur cycles and atmospheric oxygen from middle Permian to middle Triassic That is a staggering change, equivalent to breathing at high altitude compared to sea level.
Several factors converged. High-biomass forests were gradually replaced by lower-biomass herbaceous vegetation, reducing the total amount of organic debris entering sediments. Increasing aridity dried out the swampy lowlands where coal had formed. Sea levels fell, shrinking the coastal wetlands that had served as burial grounds for organic carbon. Meanwhile, the consolidation of the supercontinent Pangea deformed and uplifted older coal-bearing basins, exposing vast peat deposits to the air where they oxidized, releasing their stored carbon back as carbon dioxide.19Geology. Late Permian global coal hiatus linked to 13C-depleted CO2 flux into the atmosphere during the final consolidation of Pangea This erosion of old organic carbon was essentially undoing millions of years of burial, pulling oxygen out of the atmosphere by reversing the reactions that had put it there.
Giant Insects and the Biological Consequences of High Oxygen
The Carboniferous oxygen peak left biological fingerprints that fascinate paleontologists. Dragonfly relatives with wingspans approaching 70 centimeters, millipede-like creatures over two meters long, and other oversized arthropods populated late Paleozoic ecosystems. Because insects deliver oxygen to their tissues through a network of passive air tubes called tracheae rather than using lungs and blood, their maximum body size is strongly tied to how much oxygen is available in the ambient air. When oxygen is higher, the tracheal system can supply tissues over a longer distance, allowing larger bodies.20PubMed Central. Increase in tracheal investment with beetle size supports hypothesis of oxygen limitation on insect gigantism When atmospheric oxygen was hyperoxic, giant insects thrived; the correlation is strong enough that researchers consider oxygen a central factor in setting the relatively small size of modern insects.21PubMed Central. Atmospheric oxygen level and the evolution of insect body size
The effects extended beyond bugs. A denser, more oxygen-rich atmosphere would have enhanced aerodynamic force production for early flying insects and may have boosted the metabolic capacity of early tetrapods venturing onto land. Multiple independent origins of vertebrate flight in the fossil record also line up with geological periods of elevated oxygen concentration and atmospheric density.22PubMed. Atmospheric oxygen, giant Paleozoic insects and the evolution of aerial locomotor performance
Why Oxygen Did Not Just Keep Climbing
If burying organic carbon raises oxygen, what stopped levels from shooting past 30 percent or 40 percent? Fire is one answer. Higher oxygen makes vegetation easier to ignite and fires harder to extinguish. At some point, wildfire frequency and intensity would have been high enough to burn through forests and organic matter before it could be buried, creating a negative feedback loop. The question is where that ceiling sits. Some researchers have argued forests could not survive above 25 or 30 percent oxygen, but experimental work and modeling suggest the real limit may be considerably higher, above 35 percent, because moisture in living vegetation can counteract the combustibility boost from extra oxygen.23Nature Communications. Increased fire activity under high atmospheric oxygen concentrations is compatible with the presence of forests The presence of charcoal in the fossil record throughout the Carboniferous confirms that fires were common, yet forests persisted. Fire acted as a brake on oxygen accumulation, but not a hard wall.
How Scientists Reconstruct Ancient Oxygen Levels
You might wonder how anyone claims to know the oxygen concentration 300 million or a billion years ago. There is no single method; researchers triangulate using several independent lines of evidence. Charcoal abundance in sedimentary rocks constrains when and where fires burned, which in turn sets minimum oxygen thresholds (fire cannot sustain itself below about 15 to 16 percent oxygen). Sulfur isotope anomalies in ancient rocks disappear after the Great Oxidation Event, providing a sharp marker for when oxygen first became persistent. Carbon isotope ratios in marine sediments track how much organic carbon was being buried versus recycled, which is directly linked to oxygen production.
One of the more direct methods involves ancient halite, or rock salt. When salt crystals form at the surface, they can trap tiny bubbles of the ambient air. Carefully extracting and analyzing the gas in those fluid inclusions gives a snapshot of atmospheric composition at the time the salt formed. Measurements from 1.4-billion-year-old halite, for example, revealed oxygen partial pressures consistent with a few percent of the atmosphere.8PubMed Central. Breathing life into the boring billion: Direct constraints from 1.4 Ga fluid inclusions reveal a fair climate and oxygenated atmosphere And 815-million-year-old halite from Australia yielded an estimate of roughly 10 percent oxygen.9Gondwana Research. Atmospheric gas in modern and ancient halite fluid inclusions: A screening protocol These direct measurements are rare because suitable, well-preserved samples are scarce, but when they exist they provide a valuable cross-check on the model-based estimates.
Oxygen in the Modern Atmosphere and Where It Is Headed
Today’s atmosphere sits at about 20.946 percent oxygen. Fossil fuel combustion is the largest draw on that supply in the current era, consuming an estimated 38.2 gigatonnes of oxygen per year as of 2015, up from about 2 gigatonnes per year in 1900. Under a high-emissions scenario, around 100 gigatonnes per year could be consumed through the end of this century, which would nudge atmospheric oxygen down to about 20.825 percent by 2100.24PubMed Central. The origin of Cretaceous black shales: a change in the surface ocean ecosystem and its triggers That decline, while real and measurable with sensitive instruments, is tiny in terms of what you would notice breathing. The atmosphere holds such an enormous reservoir of oxygen, roughly 1.2 million gigatonnes, that even aggressive fossil fuel burning barely dents the total. The much more pressing atmospheric concern from burning fossil fuels is the carbon dioxide being added, not the oxygen being removed.
Over geological timescales, though, the long-term future of atmospheric oxygen depends on the same factors that have always controlled it: the balance between organic carbon burial and organic carbon weathering, the redox state of volcanic emissions, the extent of plant-covered land, and the vigor of ocean circulation that delivers nutrients to photosynthetic organisms. If anything, Earth’s oxygen story teaches us that stable atmospheric conditions are a temporary arrangement. The chemistry of the air has swung wildly in both directions over the past four billion years, and the balance we enjoy now is the product of a specific, and by no means permanent, alignment of biological and geological processes.
Seafloor Oxygen Swings and Early Animal Stress
One often-overlooked dimension of ancient oxygen is what conditions were like right at the seafloor, where early animals actually lived. Even when average atmospheric oxygen was high enough to support animal metabolism, benthic environments experienced dramatic daily swings. Modeling work suggests that in warm shallow seas, photosynthesis during daylight hours could push bottom waters to full oxygenation, while nighttime respiration would drive the same waters to complete anoxia in less than 20 minutes.25Nature Communications. Benthic diel oxygen variability and stress as potential drivers for animal diversification in the Neoproterozoic-Palaeozoic In cooler conditions the fluctuations were gentler, but hypoxic nights were still common. This means early animals did not simply enjoy a steady climb in oxygen availability. They lived in a world of daily oxygen boom and bust, which may have been a powerful selective pressure driving metabolic innovation and body-plan diversification. The atmospheric average is only part of the story; the lived experience at the sediment-water interface was far more volatile.