The Day the Dinosaurs Died: Inside the Extraordinary — and Contested — Fossil Site That May Preserve the Chicxulub Impact Hour by Hour

In the badlands of southwestern North Dakota, within a stretch of private ranchland near the town of Bowman, a 1.3-meter-thick slab of ancient sediment holds what may be the most detailed snapshot of a single catastrophic day in Earth's 4.5-billion-year history. The site, known as Tanis, sits within the Hell Creek Formation — the same geological unit that has produced Tyrannosaurus rex skeletons for more than a century — and its discoverers argue it preserves the immediate aftermath of the Chicxulub asteroid impact, the event that ended the age of dinosaurs roughly 66 million years ago [1][5].

The claim is staggering: hundreds of articulated fish, their gills packed with tiny glass beads — impact spherules — that rained down from the sky while the animals were still breathing. Marine ammonite shells carried inland by a wall of water. A turtle impaled by a wooden stake during the chaos. And, layered above it all, the thin iridium-rich clay that marks the Cretaceous-Paleogene (K-Pg) boundary worldwide [1][3].

If the interpretation holds, Tanis is not just another fossil site. It is a crime scene preserved in geological amber — a record not of millennia or centuries, but of hours and minutes. Yet the site has also become a flashpoint for debates about scientific publishing norms, private fossil ownership, research misconduct, and the limits of what any single deposit can tell us about the deep past.

The Physical Evidence: Clocking the Catastrophe

The chain of evidence tying Tanis to the Chicxulub impact rests on multiple independent lines of analysis, each with its own precision and margin of error.

Radiometric dating. The most direct clock comes from argon-argon (⁴⁰Ar/³⁹Ar) dating of unaltered impact glass — the spherules themselves — recovered from the deposit. This analysis yields an age of 66.043 ± 0.043 million years, placing the deposit squarely at the K-Pg boundary [1]. A separate analysis yielded 65.76 ± 0.15 million years, a broader but overlapping range [1]. The margin of error on the tighter measurement — roughly 43,000 years — is too wide to prove a single day, but it confirms the deposit is contemporaneous with the impact event rather than a later reworking.

Spherule chemistry. The impact spherules range from 0.3 to 1.4 millimeters in diameter and are composed of altered glass with chemical signatures consistent with Chicxulub ejecta [5]. More than 50% of the acipenseriform fish — sturgeons and paddlefish — preserved at Tanis have these spherules embedded in their gill rakers, the comb-like structures fish use to filter food from water [3]. Paddlefish, which feed by swimming open-mouthed, would have inadvertently inhaled the spherules as ejecta settled through the water column [3]. This is the strongest direct evidence for temporal correlation: the animals were alive and breathing when the impact debris arrived.

Stratigraphy. The event deposit itself — approximately 1.3 meters thick — interrupts the orderly architecture of a Late Cretaceous point bar (a river channel deposit). The basal unit, about 50 centimeters of coarse sand with rip-up clasts, grades upward into sand-silt laminations and then fine silt and mud [1][5]. Directly above this chaotic layer sits a 1-to-2-centimeter-thick tonstein (an altered volcanic ash layer) containing 3.8 parts per billion of iridium — the telltale signature of the K-Pg boundary [1]. Shocked quartz grains with planar deformation features, a hallmark of hypervelocity impact, are also present [5].

The seiche mechanism. The proposed depositional model holds that seismic waves from the impact — traveling at roughly 5 kilometers per second through the crust — reached North Dakota within 10 to 13 minutes, despite the impact site being more than 2,800 kilometers away in what is now the Gulf of Mexico [5]. The resulting magnitude-10-to-11 earthquake generated seiches (standing waves in enclosed or semi-enclosed bodies of water) with a minimum runup height of 10 meters, which surged upriver and deposited the entire Tanis assemblage within one to two hours [1][5]. The analogy is not hypothetical: the 2011 Tohoku earthquake in Japan triggered 1.5-meter seiches in Norwegian fjords 8,000 kilometers away [4].

What Has Been Found — and What Remains in the Ground

Tanis has yielded hundreds of exceptionally preserved, often articulated specimens across a wide taxonomic range [1]. The fish assemblage includes the sturgeon Acipenser praeparatorum and the paddlefish Parapsephurus willybemisi, both described as new species from the site, along with bowfin (Amia sp.) and other acipenseriform and amiiform taxa [1]. Terrestrial vertebrates include a partial leg of the ornithopod dinosaur Thescelosaurus neglectus, preserving skin, tendons, and soft tissues; fragmentary hadrosaurid material; pterosaur remains including a near-hatching embryo inside its egg; small mammal fragments including a possible marsupial jawbone; and turtle and crocodilian remains [1][7]. The invertebrate record includes the ammonite Sphenodiscus lobatus, unionoid bivalve mussels, and hymenopteran insects [1].

The deposit is described as covering roughly 1 to 2 square kilometers, but no public disclosure has been made of what fraction has been excavated since work began in 2012 [1]. This means the published fossil record from Tanis may represent a small window into a much larger assemblage — or, conversely, that the richest portions may already have been sampled. Without transparent reporting of excavation extent, it is impossible for outside researchers to estimate how much of the record remains in the ground.

Comparing Tanis to Other K-Pg Boundary Sites

Tanis is not the only locality that preserves the K-Pg transition, but it differs from others in preservation quality and taxonomic breadth.

Source: Compiled from published literature

Data as of Jan 1, 2025CSV

The Brazos River section in Falls County, Texas, located roughly 1,500 kilometers from the Chicxulub crater, preserves a marine K-Pg sequence with an abundant fauna of ammonites, foraminifera, and other marine invertebrates [9][10]. But Brazos River records primarily marine microfossil turnover across the boundary — a record measured in relative abundance shifts over thousands of years, not hours. Its temporal resolution is orders of magnitude coarser than what Tanis claims.

El Kef in Tunisia, the Global Boundary Stratotype Section and Point (GSSP) for the K-Pg boundary, provides the international reference standard but is likewise a marine section dominated by microfossils [10]. Caravaca in Spain and Stevns Klint in Denmark similarly record the boundary in marine chalk or marl sequences [10].

What Tanis offers that none of these sites do is three-dimensional preservation of articulated vertebrate carcasses — a Konservat-Lagerstätte, or deposit of exceptional preservation — directly linked to the impact event by spherules physically lodged in the organisms' bodies [1][5]. No other K-Pg site has produced articulated fish with impact ejecta in their gills. No other site has yielded dinosaur skin, pterosaur embryos, or mammal remains in direct stratigraphic association with the boundary clay.

The Skeptics' Case: A Secondary Event?

Not all paleontologists accept the "Day One" interpretation. A vocal minority argues that the Tanis deposit could represent a secondary event — a later tsunami surge, river flood, or reworked sediment package — rather than the immediate post-impact seiche.

Kirk Johnson, Sant Director of the Smithsonian National Museum of Natural History, criticized the gap between claims made in media accounts and those supported in the peer-reviewed paper, calling it "a sloppy way to conduct science" [6]. Pat Holroyd of UC Berkeley questioned the methodology for determining exact timeframes, arguing that there is no conclusive way to establish the precise duration represented in the deposit [6]. Geochemist Shaena Montanari described the geochemical data as "scant and in some cases being stretched a bit to make interpretations," calling for more rigorous hypothesis testing [6].

James Witts raised a specific concern about the ammonite shells found at Tanis: they "might not have come from a nearby remnant of sea but could have been fossils when the asteroid struck" — in other words, reworked from older marine bedrock rather than carried in alive by the seiche [6]. If the ammonites were already dead and fossilized before the event deposit formed, one key piece of evidence for a marine incursion would collapse.

More broadly, critics point to the potential for reworking of the event deposit and note that the tektites and ejecta "may not be in situ but rather sourced externally" [1]. The private management of the site, with restricted access, has prevented independent verification of many claims, amplifying skepticism.

DePalma and colleagues have responded by pointing to the spherules-in-gills evidence as proof of contemporaneity: the glass beads could not have been aspirated by dead fish [3][5]. They also cite the undisturbed stratigraphy — the graded bedding, the intact tonstein cap — as evidence against later reworking.

The New Yorker, the PNAS Paper, and Scientific Gatekeeping

The timeline of Tanis's public reveal is itself a case study in the tension between scientific norms and public communication.

The site was first identified in 2008 by Steve Nicklas and Rob Sula, with preliminary surveys continuing through 2011 [1]. Robert DePalma assumed management of excavations in 2012 [1]. But the site remained largely unknown outside a small circle until March 2019, when journalist Douglas Preston published a lengthy feature in The New Yorker describing Tanis and its implications in vivid detail [4][6].

The problem: the peer-reviewed paper — DePalma et al. in Proceedings of the National Academy of Sciences — appeared only days after the magazine article, and it did not include evidence for several of the most dramatic claims reported in The New Yorker, including dinosaur skin and feathers [4][6]. The paper had cleared peer review in roughly four months, but the public announcement preceded the scientific record.

This sequencing drew sharp criticism. Johnson noted that "many interesting aspects of this study appear only in the New Yorker article and not in the scientific paper" [6]. The episode crystallized a persistent anxiety in paleontology: that media-friendly discoveries, especially those controlled by a single researcher on private land, can bypass the vetting process that science relies on to separate extraordinary claims from wishful thinking.

Subsequent publications followed: a 2021 paper in Scientific Reports on fish isotope analysis concluded that the asteroid struck during boreal spring [8]. In 2022, the BBC documentary "Dinosaurs: The Final Day" brought further public attention. A January 2025 study by Kaskes et al. in Nature Communications used Tanis geochemical data to model sulfur's role in the mass extinction [1].

Source: OpenAlex

Data as of Jan 1, 2026CSV

Academic interest in the Chicxulub impact has grown steadily, with publications peaking at 160 papers in 2024, up from 33 in 2011 — reflecting both Tanis-driven attention and broader advances in impact science.

Misconduct, Accusations, and an Incomplete Resolution

In late 2022, Melanie During, a paleontologist at Uppsala University, accused DePalma of fabricating isotopic data in the 2021 Scientific Reports paper — specifically, of manipulating graphs and using nonexistent measurements to establish that the impact occurred in spring [2][11]. During alleged that DePalma rushed to publish ahead of her own independent work on the same question.

The accusations triggered a formal investigation by the University of Manchester, where DePalma was a doctoral student. The investigation, completed in 2023, reached a split verdict: DePalma did not fabricate data, but there were "several instances of poor research practice in the way the isotope data was managed and presented, which together constituted research misconduct" [11][12]. The investigation found "compelling evidence that DePalma was already working on seasonality prior to Melanie During's introduction to the Tanis site" [12].

Critically, the investigation did not invalidate the paper's central conclusion — that the impact occurred in spring — but the finding of research misconduct, combined with DePalma's inability to specify where key isotope measurements were produced, damaged confidence in the data provenance [11]. The journal added an editor's note to the paper flagging the data as under review [2].

Private Land, Private Fossils: The Ownership Question

Tanis sits on private ranchland, and DePalma holds exclusive lease rights from the landowner [1][13]. Under United States law, fossils found on private land belong to the landowner — and, by extension, to whoever the landowner authorizes to collect them. No federal or North Dakota state statute requires that scientifically significant fossils from private land be deposited in public repositories [13].

This legal framework means DePalma controls access to the site, determines who can visit, and reportedly stores at least some fossils in his personal collections rather than in accredited museum repositories [1][13]. Other paleontologists have criticized this arrangement as effectively restricting independent verification. The Society of Vertebrate Paleontology's code of ethics calls for fossils described in the scientific literature to be deposited in permanent, accessible collections — but compliance is voluntary [13].

The commercial value of specimens like the Thescelosaurus leg or the pterosaur embryo is difficult to estimate precisely, but comparable specimens have sold at auction for millions of dollars. The 2020 sale of a T. rex skeleton ("Stan") for $31.8 million underscores the financial stakes. No public information is available about specific funding sources, institutional support agreements, or the long-term disposition of Tanis fossils [1].

The parallel to prior fossil ownership disputes is direct. The Tyrannosaurus rex specimen "Sue," discovered in 1990 on Cheyenne River Sioux land in South Dakota, triggered a decade-long legal battle involving the landowner, the collecting institution, and the federal government. The Supreme Court ultimately ruled that Sue was held in trust by the U.S. government on behalf of the tribal landowner [13]. Tanis, on non-tribal private land, faces a different legal landscape — but the underlying tension between scientific access and property rights is the same.

What Is Missing — and What That Tells Us

The Tanis assemblage, for all its richness, shows notable gaps. The deposit is dominated by aquatic and semi-aquatic fauna — fish, turtles, crocodilians — with terrestrial organisms appearing as fragmentary intrusions rather than a representative cross-section of the Hell Creek ecosystem [1][7].

Large-bodied dinosaurs are conspicuously underrepresented. The Thescelosaurus leg and fragmentary hadrosaurid material are the only confirmed dinosaur body fossils; no Tyrannosaurus, Triceratops, or Ankylosaurus skeletal material has been reported from the event deposit itself, despite these being common in the surrounding Hell Creek Formation [1][7]. Feathers and skin impressions have been described in media accounts but have not appeared in the peer-reviewed literature with formal identifications [6].

This pattern is consistent with the depositional model: a seiche surging up a river channel would preferentially capture organisms already in or near the water — fish, turtles, aquatic plants — while leaving large terrestrial animals on higher ground. The absence of megafauna skeletons does not contradict the "Day One" hypothesis, but it does constrain what Tanis can tell us about the terrestrial extinction. The site records who was in the river when the wave hit, not necessarily who was alive in the broader landscape.

Small mammals, birds, and flying insects — groups that would go on to dominate the Paleogene — are likewise rare in the deposit. Whether this reflects genuine absence from the local ecosystem, taphonomic bias (the physics of what gets buried and preserved), or incomplete excavation remains an open question [1].

Where This Stands

Tanis remains, seven years after its public debut, a site of genuine scientific importance and unresolved controversy. The spherules-in-gills evidence is difficult to explain away: it places living, breathing fish in a rain of impact ejecta, linking the deposit to the Chicxulub event with a directness no other K-Pg site can match [3][5]. The radiometric ages are consistent. The stratigraphy is coherent.

But the site is controlled by a single researcher who has been found to have committed research misconduct in the handling of related data [11]. Access is restricted. Key specimens may not be in public repositories. And the most dramatic claims — dinosaur skin, pterosaur embryos, the precise hour of death — have yet to receive the independent scrutiny that extraordinary claims require.

The tension at Tanis is not between good science and bad science. It is between the desire to tell a complete story — the day the dinosaurs died — and the slower, more frustrating work of proving it, piece by piece, under conditions that the broader scientific community can verify. Until the site is opened to independent investigation, its full significance will remain, like much of the deposit itself, still in the ground.