Siberian Traps Erupted for 300,000 Years and Nothing Died
The worst extinction in Earth's record is dated to at most sixty thousand years, by uranium clocks inside crystals in five thin ash beds at a quarry in eastern China. The usual explanation is that the Siberian Traps erupted. But the Traps were dated too, and the lava had been running for three hundred thousand years, with two thirds of its volume already on the ground, while the fossil record barely moved.
Chapters
- 0:00 Five thin bands of clay
- 0:22 The lava had been running for 300,000 years
- 0:46 252 million years ago
- 1:02 The clock and the mechanism
- 1:28 Meishan, section D, bed by bed
- 1:52 How a crystal keeps time
- 2:22 60,000 years, give or take 48,000
- 2:48 That number is a ceiling
- 3:15 The boundary sits inside the extinction
- 3:48 The carbon moved first
- 4:24 Light carbon has a history
- 4:48 The usual suspect
- 5:10 The Traps were dated too
- 5:38 Not how much magma. Where it went
- 6:08 What a sill does to a basin
- 6:38 The carbon the planet had buried
- 6:57 Four to nine times more
- 7:16 Metal that should not be in basalt
- 7:52 Ten degrees of ocean warming
- 8:17 96 percent, or 81
- 8:50 The volcano was not the weapon
- 9:13 A rate, not a volume
- 9:35 Next: the oldest thing on Earth
Transcript
Five thin bands of clay 0:00
In a roadside quarry in eastern China, a few metres of grey limestone sit in the open air. Running through it are five thin bands of clay. Each one was volcanic ash that settled out of the sky and turned to mud. Below one of those bands is a full, crowded marine world. Above it, almost nothing.
The lava had been running for 300,000 years 0:22
You have probably heard the explanation for that gap. The Siberian Traps erupted, and almost everything died. The rock says something stranger. By the time the dying started, that eruption had been running for three hundred thousand years. Two thirds of all its lava was already on the ground, and the fossil record had barely flinched.
252 million years ago 0:46
This sits two hundred and fifty-two million years ago, at the end of the Permian. It is the boundary between the Palaeozoic and the Mesozoic, and the worst extinction the rock record keeps. Everything after it, including us, is what grew back.
The clock and the mechanism 1:02
Two places carry this story, twelve thousand kilometres apart. The first is Meishan, a quarry section near Changxing, in Zhejiang Province, eastern China. It is the global reference section for the Permian-Triassic boundary. The golden spike. The second is the Tunguska Basin, in central Siberia. That one holds the mechanism.
Meishan, section D, bed by bed 1:28
Meishan is read as numbered beds, counted upward. The Changhsing Formation below, the Yinkeng Formation above. Bed twenty-five is a claystone a few centimetres thick. Below it, the late Permian sea is full: brachiopods, corals, crinoids, foraminifera. At bed twenty-five, that list collapses.
How a crystal keeps time 1:52
The ash beds are what make this section a clock. Ash carries zircon, a crystal that traps uranium when it forms and rejects lead. Every lead atom inside one, afterwards, arrived by radioactive decay. Researchers dated single grains, one at a time: sixteen from bed twenty-five, thirteen from bed twenty-eight. The result is a rock two hundred and fifty-two million years old, pinned to about thirty-seven thousand.
60,000 years, give or take 48,000 2:22
Here is what those crystals give. The extinction begins at two hundred fifty-one point nine four one million years ago. It ends at two hundred fifty-one point eight eight zero. The interval between them is sixty thousand years, give or take forty-eight thousand. For the largest extinction in the history of animal life, that is not an era. It is an incident.
That number is a ceiling 2:48
And that number is a ceiling, not a measurement. Meishan is a condensed section: very little sediment piled up per thousand years. In sections that accumulated faster, like Penglaitan in China, or the Gartnerkofel core in the Alps, the extinction looks sharper. So sixty thousand years does not say how fast it was. It says how slow it could possibly have been.
The boundary sits inside the extinction 3:15
There is a second oddity in the dating, and it is about the boundary itself. The Permian-Triassic boundary is not defined by the extinction. It is defined by the first appearance of a conodont: Hindeodus parvus, a tooth-shaped microfossil, in bed twenty-seven c. That first appearance dates to two hundred fifty-one point nine zero two. Thirty-nine thousand years after the dying started, and twenty-two thousand before it stopped. The line named after the event sits inside it.
The carbon moved first 3:48
The same beds carry a chemical record, in the carbon isotopes of the limestone. Through the late Permian it sits at a stable plus three to four parts per thousand. Then, just above the base of bed twenty-three, it starts sliding down. That is about sixty thousand years before the first species disappear. In the top six centimetres of bed twenty-four e it falls off a cliff: from plus two to about minus four. Depending on how fast that mud piled up, the whole plunge took between two thousand and nineteen thousand years.
Light carbon has a history 4:24
A drop in that direction means light carbon, and light carbon has a history. Living things prefer the lighter isotope, so anything that has passed through biology carries that signature. Carbon dioxide straight from the mantle does not. Whatever was loading the ocean and the air, it was not fresh from the deep Earth. It was an old stockpile.
The usual suspect 4:48
Which brings us to the usual suspect. The Siberian Traps are the largest continental flood basalt eruption of the last five hundred million years. More than a million cubic kilometres of lava, poured across northern Asia. The timing is roughly right, the scale is enormous, and the story writes itself.
The Traps were dated too 5:10
Except the Traps were dated too, with the same technique, on the same kind of crystal. The lava phase starts at two hundred fifty-two point two four million years ago. That is about three hundred thousand years before the first species disappears at Meishan. Across that whole interval, roughly two thirds of the province's lava erupts. And the carbon record and the fossil record show little sign of anything going wrong.
Not how much magma. Where it went 5:38
So the question is not what started. It is what changed. At two hundred fifty-one point nine zero seven, the province changes how it delivers magma. Instead of piling up as flood lava on the surface, it begins spreading sideways underground. Sheets of magma, forced between existing layers of rock. Sills. The sill complex covers more than one and a half million square kilometres.
What a sill does to a basin 6:08
What matters is what those sills were spreading into. The Tunguska Basin is a stack of sedimentary rock: salt and other evaporites, carbonate, and rock rich in organic carbon. A sill arriving at eleven hundred degrees does not simply sit there. It bakes a halo of country rock around itself, and baked organic rock gives off gas. Pressure builds in that halo until it vents, through pipes of shattered rock that reach the surface.
The carbon the planet had buried 6:38
This channel has spent an episode on why that organic rock exists at all. A wet tropical belt sitting over basins that kept sinking, during the assembly of Pangaea. Coal is carbon the planet took tens of millions of years to bury. In Siberia, magma found it.
Four to nine times more 6:57
The size of that effect has been estimated, and it is not a rounding error. Contact metamorphism around these sills releases four to nine times more carbon dioxide than the magma would on its own. A single square metre of that baked halo can yield fifty to eighty tonnes of it.
Metal that should not be in basalt 7:16
And there is a specimen that carries this mechanism as an object. Basalt does not normally contain metallic iron. Iron in lava is bound up with oxygen. But in a handful of Siberian Traps intrusions it is there: grey metal, grown inside the rock. Those bodies sit directly on Carboniferous and Permian coal-bearing sediment. Carbon from that coal stripped the oxygen off the iron and left the metal behind. The magma did not merely pass the carbon. It reacted with it, and the rock kept the receipt.
Ten degrees of ocean warming 7:52
At Meishan, the consequences are measurable in the same rock. Sea surface temperature rises about ten degrees Celsius across the extinction interval, from roughly twenty-three to thirty-three. Calcium isotopes shift the way they shift when seawater acidifies. Spread over sixty thousand years, that warming works out to about one degree every six thousand years.
96 percent, or 81 8:17
One more number deserves the same scrutiny as the rest. You will usually see this event described as killing ninety to ninety-six percent of marine species. That figure comes from a nineteen seventy-nine estimate, calculated from the loss of genera, not counted from species. A twenty sixteen reanalysis, correcting for background extinction and for how losses cluster, puts it nearer eighty-one percent. Still the worst in the record. Still an inference, not a body count.
The volcano was not the weapon 8:50
Put together, the record says something narrower than the headline. An eruption of that size is not automatically lethal. This one proved it for three hundred thousand years. What turned it lethal was where the magma went, and what happened to be buried there. The volcano was not the weapon. What it cooked was.
A rate, not a volume 9:13
Which makes the useful number here a rate, not a volume. A degree of ocean warming every six thousand years, sustained, was enough to do this. That is the pace that ended more than four out of five marine species. It took five to ten million years before the ocean carried anything resembling a full ecosystem again.
Next: the oldest thing on Earth 9:35
Next time, this channel follows the same mineral to the far end of the record. A single grain of zircon from the Jack Hills, in Western Australia. Four billion, four hundred and four million years old. Older than any rock on Earth. And the oxygen locked inside it says that the eon we named after hell already had liquid water.
Description and sources
At Meishan, in Zhejiang Province, five volcanic ash beds run through the few metres of limestone that record the end of the Permian. U-Pb dating of single zircon grains puts the marine extinction at 251.941 and 251.880 million years ago: 60,000 years, plus or minus 48,000. Meishan is a condensed section, so that is a ceiling: how slow the event could have been, not how fast it was.
Two things in that dating do not fit the tidy story. The Permian-Triassic boundary is defined by the first appearance of a conodont, Hindeodus parvus, dated to 251.902 — 39,000 years after the dying started and 22,000 before it stopped. The line named after the event sits inside it. And the carbon moves first, plunging from +2 to about -4 per mil inside the top six centimetres of one bed.
The Siberian Traps are the usual suspect, and the same technique dates them too. What coincides with the extinction is not the eruption but a change in plumbing: at 251.907 the magma stops piling up on the surface and spreads sideways underground, as sills covering more than 1.5 million square kilometres of the Tunguska Basin — salt, carbonate and organic-rich rock. Baking it releases four to nine times more carbon dioxide than the magma would degas alone, and a specimen from the Putorana Plateau carries the reaction as an object: native metallic iron grown inside basalt.
The eruption was not the weapon. What it cooked was.
PRINT-READY, FROM THIS CHANNEL
The Deep Time Record, Vol. 01 — the cores, the curves and the clocks
https://therepository.gumroad.com/l/deep-time-record
PRIMARY SOURCES
- Burgess, S. D., Bowring, S., & Shen, S. — High-precision timeline for Earth's most severe extinction. PNAS 111(9), 3316-3321, 2014. https://www.pnas.org/doi/10.1073/pnas.1317692111
- Burgess, S. D., Muirhead, J. D., & Bowring, S. A. — Initial pulse of Siberian Traps sills as the trigger of the end-Permian mass extinction. Nature Communications 8, 164, 2017. https://www.nature.com/articles/s41467-017-00083-9
- Svensen, H. H., et al. — Sills and gas generation in the Siberian Traps. Philosophical Transactions of the Royal Society A 376, 20170080, 2018. https://royalsocietypublishing.org/doi/10.1098/rsta.2017.0080
- Stanley, S. M. — Estimates of the magnitudes of major marine mass extinctions in earth history. PNAS 113(42), E6325-E6334, 2016. https://www.pnas.org/doi/10.1073/pnas.1613094113
- Pernet-Fisher, J. F., et al. — Atmospheric outgassing and native-iron formation during carbonaceous sediment-basalt melt interactions. Earth and Planetary Science Letters 460, 2017. https://www.sciencedirect.com/science/article/abs/pii/S0012821X1630735X
ARCHIVE PLATES — public-domain documents, with author, date, archive and licence.
... and 5 more plates, each credited on screen in the video and listed in this project's manifest.
The 5 generated plates were made for this channel with FLUX.1-schnell. No person is depicted in any of them.
Deep Time Archives rebuilds the pre-human Earth from the physical record — outcrops, drill cores, golden spikes, catalogued specimens. Every episode starts from an object you can go and check, and the sources are listed above.
About the images. Five photographs are archive images, credited on screen: a catalogued zircon crystal, an Upper Permian organic-rich shale, a limestone-over-coal outcrop, a coal seam in section, and native iron in Siberian Traps basalt from the Putorana Plateau. The quarry face, the basalt terraces, the sill and the empty sea are AI-generated reconstructions, because nobody photographed a 252-million-year-old world. This video is labelled as containing altered or synthetic content. The diagrams are drawn in code from the published numbers, not generated.