End-Permian extinction hit North China’s land and sea in step, study says

A 606-metre stack of river rocks in Shanxi puts North China’s carbon-cycle upheaval at 252.1 to 252.0 million years ago, in line with the marine die-off, but each age carries an uncertainty of about 220,000 years.

Peer reviewed · Rock-record dating study · One river section · Full paper paywalled; abstract, supplement and data read

The Great Dying, the most severe mass extinction of the past 540 million years, may have reached the land and the oceans at the same time, at least in North China. A team led from Peking University by Hongliang Lu and Baochun Huang dated a thick run of river-laid rocks at Sigou, Shanxi Province, and report in PNAS that its carbon-isotope upheaval matches the marine extinction. That matters because records from Australia and South Africa have been read as showing land ecosystems failing first. The catch: the Sigou ages are uncertain by about 0.22 million years.

Evidence at a glance

Is this a big deal?

A solid step, not a verdict. One more well-measured land record now has a timescale, and it lines up with the sea. A lag of a couple of hundred thousand years would still fit inside its error bars.

What scientists found

The study tracks two chemical signals through the Sunjiagou Formation. Mercury, which volcanoes release, peaks at about 252.4 Ma. The highest of the 48 published values is 27 nanograms per gram, roughly five times the median of 5.2. The authors align that peak with the main eruptive phase of the Siberian Traps, the vast lava outpourings in what is now Russia (for the basics, see what triggers volcanic eruptions).

About 0.3 million years later, the ratio of carbon-13 to carbon-12 in the rocks’ organic matter starts to fall, at 252.1 Ma, and reaches its low at 252.0 Ma. In North China, earlier work cited by the team found such carbon shifts at the same level as plant extinctions. The authors read the Sigou shift as the land extinction’s peak and find it synchronous with the Meishan marine boundary section.

The team also places the Induan–Olenekian boundary below a normal-polarity zone called SG4n, which makes the first stage of the Triassic last 1.6 to 1.8 million years.

How they found it

The supplement describes no dated volcanic ash at Sigou; its zircon crystals were washed in by rivers, so they give only a maximum age. The team built a clock from the rocks themselves.

First, they drilled 2.5-centimetre cores at 289 levels through 606 metres of rock and measured the ancient magnetic direction locked in each. Earth’s field flips now and then, so a section carries a barcode of normal and reversed stripes; Sigou has 14 such zones. Matching barcodes lets a land section be lined up with a marine one. The analogy has a limit: one stripe looks like any other, so only the whole pattern can be matched.

Second, they took 935 gamma-ray readings. Mudstones give off more natural gamma rays than sandstones, and the alternation tracks wet and dry spells that the authors tie to a 405,000-year cycle in Earth’s orbit. Counting those cycles gave a floating timescale. Pinning one level to a dated magnetic zone in the Shangsi marine section, at 251.89 ± 0.1 Ma, fixed it in time.

The supplement is candid about river rocks. Floods, shifting channels and erosion can blur short cycles, and the authors write that “no single analysis is presented as proving astronomical causation.” Instead, four statistical tests point the same way. The best-fit pile-up rate is about 13.6 centimetres per thousand years, a little over a tenth of a millimetre a year, and it varies from about 6 to 20.

Why it matters to you

This won’t change your day. It speaks to a deep-time question, whether life on land and in the sea collapsed together when the carbon cycle was pushed hard. For a related ancient upheaval, see how extreme warmth drove plant turnovers in fossil pollen.

What the evidence does not show

It does not date the extinction with the precision of the marine record. Meishan’s die-off is pinned to between 251.941 and 251.880 Ma. Sigou’s carbon-shift onset at 252.1 Ma is about 0.16 million years older than that, which falls inside its ±0.22 million-year uncertainty. “Synchronous” here means indistinguishable at this resolution.

In the parts SciQuest could read, the extinction level is inferred from the carbon signal; the supplement reports no fossil counts from Sigou. The anchor is also a correlation with a marine section, so the land clock leans on marine time. One more caution: an earlier study reported a gap of about 20 million years at the base of the Sunjiagou Formation; the authors set it aside because it compared zircons from sections 150 kilometres apart.

How this fits what we already knew

The marine timeline is tightly dated. Seth Burgess and colleagues dated five ash beds at Meishan and put the extinction within 60 ± 48 thousand years, in 2014. On land, the picture splits. In Australia’s Sydney Basin, Christopher Fielding’s team found that the Glossopteris flora collapsed before 252.3 Ma, about 370,000 years before the marine extinction. In South Africa’s Karoo, Robert Gastaldo’s group dated an ash above the traditional turnover level to 252.24 ± 0.11 Ma, while Pia Viglietti and colleagues argue for a protracted, roughly one-million-year extinction there.

In the tropics of southwest China, Qiong Wu and colleagues found the rainforest collapse began after the marine onset, and proposed that land extinction moved from high latitudes to the tropics. The Sigou result fits between those ends, which is how the authors frame it.

Independent perspective

SciQuest found no comment on this paper from researchers outside the author team as of 28 October 2026, 09:20 UTC, and has not interviewed one. Published independent work points both ways. Daoliang Chu and colleagues reported in Geology in 2024 that North China’s land crisis began at least 300,000 years before South China’s, and called the land extinction a set of regional losses rather than one global event.

How other headlines framed it

We found no news coverage to compare as of 28 October 2026, 09:20 UTC.

What happens next

The clearest test would be a dated volcanic ash at Sigou or a nearby North China section, which could tighten the error bars sharply; the Meishan ash dates carry uncertainties of a few tens of thousands of years. Fossil plant and pollen counts tied to the same metre levels would also show whether the carbon shift and the die-off coincide there. For another rock-record clock from northern China, see the Cambrian SPICE study near Beijing.

In one sentence

River rocks in North China suggest the Great Dying hit land and sea there within the same few hundred thousand years, but the dating is not yet fine enough to say they struck together.

Read next

• Cambrian SPICE event: sea-level falls preceded seafloor species losses: another rock-record clock from northern China.
• Fossil pollen ties plant life’s two biggest turnovers to extreme warmth: how plants respond to ancient heat.
• What triggers volcanic eruptions? The basics behind the Siberian Traps.

Sources and disclosure

1. Lu H, Huang B, Li M, Zhao Q, Li X, Liao Y, Pei Z, Xue Y, Parés JM, Zhang H, Jiang Q. “Synchronous end-Permian mass extinction zenith in marine and low–middle latitude terrestrial North China.” PNAS, published online 27 October 2026. doi:10.1073/pnas.2538105123.
2. Lu H et al. Supporting Information and Dataset S1. Zenodo, 2026. doi:10.5281/zenodo.22998231.
3. Burgess SD, Bowring S, Shen S-Z. “High-precision timeline for Earth’s most severe extinction.” PNAS 111: 3316–3321 (2014). doi:10.1073/pnas.1317692111.
4. Burgess SD, Bowring SA. “High-precision geochronology confirms voluminous magmatism before, during, and after Earth’s most severe extinction.” Science Advances 1: e1500470 (2015). doi:10.1126/sciadv.1500470.
5. Fielding CR et al. “Age and pattern of the southern high-latitude continental end-Permian extinction constrained by multiproxy analysis.” Nature Communications 10: 385 (2019). doi:10.1038/s41467-018-07934-z.
6. Gastaldo RA et al. “The base of the Lystrosaurus Assemblage Zone, Karoo Basin, predates the end-Permian marine extinction.” Nature Communications 11: 1428 (2020). doi:10.1038/s41467-020-15243-7.
7. Viglietti PA et al. “Evidence from South Africa for a protracted end-Permian extinction on land.” PNAS 118: e2017045118 (2021). doi:10.1073/pnas.2017045118.
8. Wu Q et al. “The terrestrial end-Permian mass extinction in the paleotropics postdates the marine extinction.” Science Advances 10: eadi7284 (2024). doi:10.1126/sciadv.adi7284.
9. Chu D et al. “Diachronous end-Permian terrestrial crises in North and South China.” Geology 53: 55–60 (2024). doi:10.1130/G52655.1.

Disclosure: This article was drafted with AI assistance and checked against the sources above. SciQuest could not open the main PNAS text; claims come from the abstract, the open supplement and the published dataset. Calculations by SciQuest: the mercury median and peak ratio come from the 48 values in Dataset S1 (the supplement reports 51 samples analysed); 252.1 minus 251.941 Ma gives 0.16 million years; 13.6 cm per thousand years equals 0.136 mm a year. The graphic is a SciQuest original built from the published ages. Spotted an error? See our corrections policy.