Peng Zhou's team documented both five million year old fossils and a living chemosynthetic community on a recent Antarctic minke carcass in the same hadal seafloor, making the site a rare deep time baseline for whale fall ecology.
Seven kilometres below the surface of the southern Indian Ocean, where sunlight never reaches and pressure exceeds 700 atmospheres, the Chinese crewed submersible Fendouzhe has returned with evidence of a seafloor where whale bones have been accumulating, fossilizing, and rotting in place for at least five million years. The site, identified during the Global Hadal Exploration Programme's 2023 field season in the Diamantina Zone, is the deepest known whale-fall assemblage on Earth, and the same expedition found both a five-million-year-old bone bed and a living chemosynthetic community in the same place.
The find was reported by Peng Zhou, a paleontologist at the Chinese Academy of Sciences' Institute of Deep-Sea Science and Engineering (IDSSE), and his colleagues. Across 32 crewed dives covering roughly 1,200 kilometres of seafloor southwest of Perth, Australia, Zhou's team located a fossil whale graveyard in the Dordrecht Deep at approximately 7,002 metres depth, the deepest such site yet documented. According to a New Scientist write-up of the work, the recovered material includes fossils older than five million years, possible baleen-whale ribs recovered at around 5,656 metres, and a putative new species of extinct beaked whale.
What makes the discovery scientifically unusual is not just the depth. It is that the same expedition captured two different kinds of whale-fall evidence in one place. The ancient bones are paleontological archive: a record of cetacean populations and ocean conditions from the early Pliocene, preserved at pressures and temperatures that have, until recently, been inaccessible to sampling. Sitting alongside them, the team documented a far more recent carcass, an Antarctic minke whale roughly five metres long, on which a dense community of chemosynthetic invertebrates is currently feeding. The contrast is the story. One part of the seafloor is a five-million-year-old bone bed. Another is an active whale-fall ecosystem, the kind of cold-seep community that ecologists usually have to wait decades to observe developing.
The whale-fall mechanism is the part most readers will need explained, because the term obscures more than it reveals. When a whale dies at the surface, its body often sinks. On the way down, scavengers strip the soft tissue. What remains is a skeleton, sometimes tens of tonnes of bone, landing on the seafloor. In waters shallow enough for oxygen and currents to do their work, that skeleton fuels a localized ecosystem for decades: first scavengers, then a "sulfophilic stage" in which chemosynthetic bacteria oxidize sulfides released by the decomposing bones, then a slow, decades-long phase of microbial and animal colonization. In hadal trenches, where no light penetrates and pressure reshapes the chemistry of decay, the same bones can persist for millions of years rather than decades, and the fossil record they leave is unusually complete.
Zhou's team estimates the assemblage at densities of up to roughly 760 whale individuals per square kilometre, and the invertebrate communities colonizing recent carcasses at up to about 2,800 animals per square metre. Those figures are extrapolations from dive surveys, not direct counts of every bone on the seafloor, and the headline-level "millions of bones" framing is the publication's construction, not a census. The new beaked-whale species, similarly, is reported in the write-up as "thought to be" new rather than formally described, and the dating of the oldest fossils carries the precision limits of any radiometric work done at hadal depth, where sample size is small and contamination is hard to rule out. These are not takedowns of the result. They are the honest method language of deep-sea paleontology, and they belong in the lede, not the footnotes.
The Diamantina Zone has been on oceanographers' maps for decades as a complex of troughs and fracture ridges in the southern Indian Ocean, but hadal biological sampling there is recent. Fendouzhe, China's deep-diving crewed submersible, and the Global Hadal Exploration Programme that runs it, are the reason this kind of dense, repeated, photographed survey is now possible. Before 2023, the deepest known whale-fall communities had been documented in the hundreds to low thousands of metres. The Dordrecht Deep find pushes that floor down by kilometres and reframes the hadal zone, the part of the ocean below 6,000 metres, as a potential paleontological archive rather than a biological desert.
What to watch next is the primary literature. The New Scientist piece is a popular-science summary of work that is expected to appear in a peer-reviewed journal; until the underlying paper is in hand, specific species identifications, the dating method used, and the exact density estimates should be treated as provisional. If Zhou and colleagues publish the assemblage-level survey, including dive tracks, sample coordinates, and the full invertebrate inventory from the active minke-fall site, the Diamantina find will move from "deepest known whale graveyard" to a baseline dataset for an entire depth band of oceanography. That is the part of the story most worth following.