Great Dying: How Earth Nearly Ran Out of Life 251 Million Years Ago
History

Great Dying: How Earth Nearly Ran Out of Life 251 Million Years Ago

BookOfWorldHistory June 2, 2026 14 min · 2,715 words
Reading settings
18px
Share

Forget the asteroid that killed the dinosaurs. The extinction that almost ended everything happened 185 million years earlier — and it had nothing to do with a rock from space. Ninety-six percent of marine species. Seventy percent of land vertebrates. The largest insect die-off in Earth's history. This is the story of the Great Dying, the worst thing that has ever happened to life on this planet, and why the cause should make you uncomfortable about the world right now.

There is a number that stops people when they hear it for the first time: 96. Ninety-six percent of all marine species gone. Not reduced. Not pushed to the edge. Gone, as in no longer present in the fossil record after a specific horizon in geological time. Add to that 70% of terrestrial vertebrate species. Seventy percent of land animals with backbones, erased in — geologically speaking — almost no time at all. This happened around 251.9 million years ago, at the boundary between the Permian and Triassic periods. Scientists call it the Permian-Triassic mass extinction event, or PTME. Most people who study it call it something older and more honest: the Great Dying. The extinction that killed the non-avian dinosaurs 66 million years ago is the famous one. The one that gets the movies, the museum exhibits, the endless popular science coverage. But the K-Pg event — for all the devastation it caused — was not the worst thing that ever happened to life on Earth. The Great Dying was. It wasn't even close.

Artistic reconstruction of the Siberian Traps volcanic eruptions that caused the Permian-Triassic mass extinction 251 million years ago.

The Siberian Traps — a large igneous province covering around 2 million square kilometers — erupted in pulses over hundreds of thousands of years, releasing carbon dioxide, sulfur dioxide, and toxic mercury into the atmosphere and oceans in quantities that destabilized Earth's entire climate system.

What the Permian World Looked Like Before It Collapsed

The late Permian was not a world in obvious decline. It had diversity, structure, established ecological relationships that had been in place for tens of millions of years. In the oceans, the Paleozoic evolutionary fauna — dominated by brachiopods, crinoids, tabulate and rugose corals, bryozoans, fusulinid foraminifera, and trilobites — had been stable since the Ordovician. These weren't struggling holdovers from some earlier age. They were successful, numerous organisms that had survived every extinction that came before them, including several that were themselves significant. On land, synapsids — the group that includes the ancestors of all mammals — were the dominant large vertebrates, particularly the dicynodonts (stocky, beak-mouthed herbivores) and the gorgonopsians (fast predators with large canine teeth). Insects were at their most diverse, with forms flying and crawling through Permian forests that included no equivalent anywhere in the modern world — some with wingspans pushing 70 centimeters. All of that was about to come apart. And the mechanism was not a sudden catastrophe from outside. It built, in pulses, from below.

The Siberian Traps: What It Means When a Continent Erupts

The scientific consensus on the cause of the Great Dying points, overwhelmingly, to flood basalt volcanism — specifically the eruptions that produced the Siberian Traps, a large igneous province that covered around 2 million square kilometers of what is now Siberia with lava. To understand what this actually means, it helps to push past the word "eruption," which carries connotations of localized, dramatic, relatively contained events — Krakatoa, Pinatubo, Vesuvius. The Siberian Traps were none of those things. This was volcanic activity on a continental scale, running in pulses over hundreds of thousands of years, with individual pulses enormous enough to dwarf anything the modern world has experienced in recorded history. The Siberian Traps sat over an unusually dangerous piece of lithosphere. The rocks beneath the eruption zone were rich in halogens — extremely destructive to the ozone layer — as well as thick sequences of carbonate rocks, evaporites, and coal-bearing sediments. When the magma intruded through all of this, it didn't just release the gases produced by melting rock. It cooked the surrounding material. It ignited coal deposits. It baked organic-rich sediments, releasing carbon that had been sequestered for hundreds of millions of years. The result was a cascade of greenhouse gas emissions substantially larger than what the volcanic activity alone would have produced. Atmospheric carbon dioxide, estimated at around 400 parts per million before the extinction, rose to somewhere between 2,500 and 8,000 ppm — depending on the model and the site — during the event. Global average surface temperatures, sitting around 18°C before the crisis, climbed toward 35°C. At high southern latitudes, warming of 10 to 14°C has been recorded in the geological evidence. In what is now Iran, tropical sea surface temperatures that had been between 27 and 33°C jumped above 35°C. On top of the warming, the Siberian Traps released sulfur dioxide in volumes large enough to cause short, sharp volcanic winters — brief but severe cooling events that punctuated the longer warming trend. Mercury poured into the atmosphere and ocean in quantities that left anomalies visible in sediments on every continent. Around 18 teratonnes of hydrochloric acid were emitted. Ozone production dropped by an estimated 60 to 70%, letting ultraviolet radiation reach the surface at levels far above anything organisms had been exposed to in the preceding millions of years.

Siberian Traps flood basalt lava plains, the volcanic province responsible for the Permian-Triassic mass extinction.

The Siberian Traps erupted over rock unusually rich in halogens, coal, and organic sediments — meaning the magma didn't just release volcanic gases, it cooked surrounding material and ignited buried carbon stores, multiplying the greenhouse effect far beyond what the eruptions alone would have produced.

What the Oceans Became

Carbon dioxide dissolves in seawater. When you pour enough of it into the atmosphere, you acidify the oceans — and ocean pH dropped by as much as 0.7 units during the Great Dying. For organisms that built shells and skeletons from calcium carbonate, this was chemistry working directly against their ability to exist. But acidification was only one of several things happening to the oceans simultaneously, and it may not have been the primary killer. Warming reduced the solubility of oxygen in seawater. Increased nutrient runoff from soil erosion — itself a consequence of warming and the death of land plants — fed algal blooms that consumed oxygen. Stagnating ocean circulation in a hotter world cut off the delivery of oxygenated water to depth. The oceans, in large parts, went anoxic: starved of oxygen, filled instead with hydrogen sulfide produced by sulfate-reducing bacteria thriving in the dead zones. This condition — called euxinia, oxygen-absent and sulfide-rich — left its signature in the sediment record across the Tethys and Panthalassic Oceans. Biomarkers for green sulfur bacteria, which require both sunlight and hydrogen sulfide, appear in the Permian-Triassic boundary layers at site after site around the world. These bacteria live in the photic zone, the shallow, sunlit surface waters. Their presence means euxinia had pushed all the way up to the surface in shallow seas. The combination — acidification, anoxia, euxinia, heat, mercury poisoning — hit organisms with calcium carbonate skeletons hardest. Corals (tabulate and rugose) lost 96% of genera. Brachiopods lost 96%. Bryozoans lost 79%. Crinoids came within a thread of complete extinction — 98% of genera gone. Foraminifera, 97%. Radiolarians, 99%. Ammonites, 97%, and that was after 30 million years of decline leading into the event. Trilobites, which had been declining since the Devonian and were already reduced to just 5 genera by the late Permian, finally disappeared entirely. They had survived four previous mass extinctions. They did not survive this one. Conodonts — those eel-like vertebrates whose mineralized tooth elements are one of the most useful biostratigraphic tools in geology — were severely affected, though they would survive into the Triassic before going extinct at the Triassic-Jurassic boundary later.

On Land: 70% of Vertebrates, Gone

The terrestrial extinction is harder to pin down than the marine one, partly because the continental fossil record is patchier and partly because the timing on land appears to have varied from region to region — some evidence suggests the terrestrial crisis preceded the marine one, other evidence suggests they were simultaneous or that the land extinction came slightly later in some places. What is clear is that it was severe. Over two thirds of terrestrial labyrinthodont amphibians, reptiles, and therapsid taxa disappeared. All Permian anapsid reptiles except the procolophonids died out. Gorgonopsians — the dominant large predators of the late Permian — went extinct. All dinocephalian genera had already been lost in an earlier extinction event at the end of the Guadalupian, the epoch that preceded the Lopingian. The kill mechanism on land was primarily aridification driven by extreme global warming. The supercontinent Pangaea's interior, already prone to aridity simply by virtue of being far from any ocean, got dramatically drier as temperatures spiked. The Glossopteris flora — the seed fern-dominated vegetation that covered high-latitude Gondwana — collapsed in Australia around 370,000 years before the Permian-Triassic boundary. River systems shifted from meandering to braided patterns, which in fluvial geology is a signature of the widespread death of rooted vegetation that normally stabilizes riverbanks. Insects, which had been more diverse in the Permian than at any point before or since, suffered the largest mass extinction of insects in Earth's history. Eight or nine insect orders became completely extinct. Ten more were heavily reduced. The giant winged forms of the Paleozoic — some of the largest arthropods to ever fly — were gone. The one animal that came through with apparent ease was Lystrosaurus, a pig-sized herbivorous dicynodont therapsid that went on to constitute as much as 90% of some earliest Triassic land vertebrate faunas. The Triassic, in its opening stages, belonged almost entirely to this one genus. The diversity of the Permian world had been replaced by something resembling a monoculture.

Reconstruction of Lystrosaurus, the dicynodont therapsid that dominated early Triassic land vertebrate faunas after the Permian-Triassic extinction.

Lystrosaurus, a stocky, beak-mouthed herbivore about the size of a pig, made up as much as 90% of some earliest Triassic land vertebrate communities — a level of dominance by a single genus that has no parallel elsewhere in the vertebrate fossil record.

How Long Did It Actually Take?

The main pulse of marine extinction has been constrained by uranium-lead zircon dating of volcanic ash beds at the Global Stratotype Section and Point for the Permian-Triassic boundary at Meishan, China. The extinction occurred between 251.941 and 251.880 million years ago — a duration of 60,000 years, give or take 48,000 years. On geological timescales, that is effectively instantaneous. But the full picture is more complicated. There are credible arguments for multiple extinction pulses rather than a single event. Studies of different sections have found evidence for two distinct extinction waves with different causes, different timing, and different victim groups. Ostracod and brachiopod extinctions, for example, appear to have been separated by somewhere between 670,000 and 1.17 million years. The declining diversity of the Permian is also partly entangled with an earlier, separate mass extinction at the end of the Guadalupian epoch — often called the Capitanian extinction — that occurred roughly 9.4 million years before the Permian-Triassic boundary and was itself a significant event. Under some interpretations, what we call the Great Dying is really a protracted crisis spanning multiple events rather than one clean catastrophe. That doesn't make it less severe. It may, in some ways, make it more — the world being hit repeatedly over millions of years rather than having one bad interval followed by genuine recovery.

The Recovery That Took 30 Million Years

Some extinctions bounce back relatively quickly. The Triassic-Jurassic extinction, bad as it was, showed marine recovery beginning almost immediately in some basins. The Great Dying did not work that way. The immediate aftermath of the extinction was dominated by disaster taxa — organisms that thrive specifically in degraded, low-diversity conditions and that would not normally be able to compete in a functioning ecosystem. In the oceans, bivalves like Claraia, Unionites, and Eumorphotis blanketed seafloors. Microbial mats — the kind of biological community more typical of the Precambrian than the Paleozoic — took over shallow marine environments. Siliceous sponges and keratinous sponges became the structural organisms of early Triassic reef-like communities, filling a role that corals wouldn't reclaim until the end of the Triassic. On land, the recovery was also slow and interrupted. Gymnosperms returned within a few thousand years after the boundary, but around 500,000 years later they were replaced by lycophytes during another extinction pulse. This oscillation — gymnosperms, lycophytes, gymnosperms again, lycophytes again — repeated several times during the Early Triassic, reflecting an environment that kept destabilizing as the Siberian Traps continued to erupt in additional pulses well into the Triassic period. A zone around the equator, between roughly 30°N and 40°S, may have been effectively uninhabitable for land vertebrates during the hottest phases of the Early Triassic — too hot for most tetrapods to survive. Recovery in terrestrial vertebrates, already moving slowly, had to work around a latitudinal dead zone. Global marine diversity didn't reach pre-extinction values until the Middle Jurassic — approximately 75 million years after the event. Terrestrial vertebrate fauna, by paleontologist Michael Benton's estimate, wasn't fully recovered until the Late Triassic, 30 million years after the boundary. Some terrestrial vertebrate guilds in Russia were still absent 15 million years into the recovery. What eventually emerged on the other side of all this was a fundamentally different world. The Paleozoic evolutionary fauna — brachiopods, crinoids, rugose corals, trilobites — was gone or reduced to shadows of its former self. The Modern evolutionary fauna — bivalves, snails, sea urchins, bony fish, eventually marine reptiles — had taken over the oceans. On land, the survivors of the therapsid groups that hadn't died gave rise to mammals. The archosaurs — initially rarer than the therapsids but gradually displacing them through the Triassic — gave rise to dinosaurs and crocodilians. The world the Permian had built was finished. The Mesozoic was something new.

Early Triassic seafloor reconstruction showing microbial mats and disaster taxa bivalves that dominated oceans in the aftermath of the Permian-Triassic extinction.

The immediate post-extinction marine environment was controlled by disaster taxa — bivalves, inarticulate brachiopods, and microbial communities that thrived specifically in degraded conditions. The diverse, structured ecosystems of the Paleozoic wouldn't return to these environments for millions of years.

Other Explanations — And Why Most of Them Don't Hold Up

The Siberian Traps explanation has been tested against alternatives, and the alternatives mostly fail on specifics. Asteroid impact was proposed partly because the success of the K-Pg impact hypothesis encouraged researchers to look for similar causes elsewhere. Reported evidence from the Permian-Triassic boundary included rare grains of shocked quartz in Australia and Antarctica, fullerenes supposedly trapping extraterrestrial noble gases, and meteorite fragments in Antarctica. Most of these claims have been challenged or refuted on reexamination. The quartz grains, on closer inspection, showed features consistent with tectonic rather than impact origin. Iridium levels at most Permian-Triassic boundary sites are not anomalous. There is no known impact crater of sufficient size dated to coincide with the boundary. The isotopic signatures of the extinction don't match what impact events produce. Methane clathrate release — the "clathrate gun" hypothesis — was proposed as a supplement or alternative mechanism, and there may be some role for it. But as a primary driver, it runs into problems with the isotopic patterns. The carbon isotope excursions of the Early Triassic don't fit a straightforward methane release scenario, and models show that the volcanic carbon dioxide signal accounts for the bulk of what's observed without needing clathrate input to explain it. A 2014 paper proposed something genuinely strange: that a genus of methanogenic archaea called Methanosarcina, having acquired a new metabolic pathway through gene transfer, underwent exponential population growth and consumed vast deposits of organic carbon in marine sediments, releasing methane and carbon dioxide on a massive scale. Nickel from the Siberian Traps eruptions, which is a cofactor for methanogen enzymes, would have fertilized this bloom. It's a creative hypothesis and it found some support in chemostratigraphic analysis, but the timing of nickel concentration increases relative to the carbon isotope shifts at the canonical Meishan section doesn't quite line up. The volcanic explanation remains where the evidence points, and it has accumulated further confirmation with every new study of mercury anomalies, carbon isotopes, and ocean chemistry from Permian-Triassic boundary sections around the world.

The Part That Researchers Keep Coming Back To

The Great Dying was caused by carbon dioxide. By warming. By ocean acidification, anoxia, and the collapse of ecosystems that couldn't tolerate the speed of change happening around them. The rate of carbon release during the Siberian Traps eruptions is poorly constrained, but it was almost certainly pulsed rather than continuous — concentrated into shorter intervals that were likely comparable in pace to modern anthropogenic emissions during their peaks. One study estimated the carbon dioxide emission rate of the first major Siberian Traps pulse as around half the current rate of human emissions. The total amount of carbon added to the ocean-atmosphere system during the PTME has been estimated between 3,900 and 12,000 gigatonnes. Currently, human activity adds roughly 50 gigatonnes of carbon dioxide per year to the atmosphere. The current atmospheric CO₂ concentration is about 426 parts per million — already above the estimated pre-extinction Permian baseline of 400 ppm, and rising. Modern oceans are already showing measurable drops in pH and declining oxygen levels. Coral reefs are already under documented pressure from water temperatures and acidification. Researchers who study the Great Dying are not in the habit of alarmism. But they are also people who have spent careers looking at what happens to marine ecosystems when seawater chemistry changes at speed, what happens to terrestrial food chains when aridity spreads faster than species can adapt, and what a 30-million-year recovery looks like in the rock record. When they draw comparisons between the PTME and the present, they do so because the parallels are there in the data, not because they're reaching for a dramatic ending. The Great Dying is not just old news about something that happened to unfamiliar organisms in an unimaginably distant past. It is the most thoroughly documented example on Earth of what rapid, large-scale carbon cycle disruption actually does to life — over timescales shorter than a species typically exists, in an ocean that runs by the same chemistry as the one we have now.