The doomsday fish that predicts nothing

It is the longest bony fish in the world: a silver ribbon up to eight metres long, with a crimson fin running the length of its back, living half a kilometre down in permanent darkness. When one washes ashore, it makes the news on three continents. In Japan it has a name that is hard to improve on — リュウグウノツカイ, ryūgū no tsukai, the messenger from the palace of the Dragon King, sent to warn people that the earth is about to shake.

It is a beautiful story. It is also, as far as anyone can measure, false — and the way it was taken apart is a small masterclass in how to think about warning signs. Which is why an early-warning project is the right place to tell it.

The fish

Two species carry the reputation: the giant oarfish Regalecus glesne, verified to about 8 m, and the slender oarfish Regalecus russelii, the one usually found off Japan and California. They live in the twilight zone, roughly 200–1000 m down, drifting vertically and feeding on krill. They are almost certainly the original of some old sea-serpent sightings. They are also poor swimmers: the body is a ribbon with very little muscle, built for hanging in still, dark water — not for fighting a current.

The story that was too good to check

In the year before the magnitude 9.1 Tōhoku earthquake of March 2011, roughly a dozen oarfish washed up along the Japanese coast. Afterwards, everyone remembered it. The folklore had a fresh, terrible confirmation, and it spread far beyond Japan — every subsequent stranding, anywhere in the world, arrived pre-labelled as an omen. In late 2024 three oarfish came ashore in southern California within a few months, and the phrase "doomsday fish" went around the world again.

Yoshiaki Orihara, a seismologist at Tokai University, had the reaction a scientist is supposed to have. If the folklore were true, he pointed out, it would be genuinely useful: a free, distributed, biological early-warning system. So it deserved to be tested rather than repeated.

Someone finally counted

The reason nobody had tested it was mundane: there was no database. Rare fish make good local news, so Orihara and his colleagues built one — from Japanese regional newspapers, academic papers and a marine museum's records — covering oarfish, several ribbonfishes, dealfish and the unicorn crestfish. Then they laid it against the earthquake catalogue.1

The result, published in the Bulletin of the Seismological Society of America in 2019:

The authors' own verdict is blunter than most papers allow themselves: the folklore is "a superstition attributed to the illusory correlation between the two events."1 The same group had already looked at a related claim — anomalous fish catches before Tōhoku — and found no support there either.3

So why do they come ashore?

Nobody has a single answer, and the honest sources say so. The leading explanations are all oceanographic. A deep-water fish with almost no muscle that gets caught in an upwelling current has no good way of getting back down; once at the surface it is, in the words of one biologist, out of options. Layered on top of that are temperature anomalies — strandings in California clustered with the North Pacific warm "blob" of 2013–2014 and with the strong El Niño of 2015–16 — plus red tides, storms and simple illness.4

The Californian record puts the rarity in perspective: about 22 strandings in 123 years, and then three in a few months of 2024. Staff at Scripps Institution of Oceanography, who recovered them, were careful to say that changing ocean conditions are the likely cause, that the El Niño link is real but inconsistent, and that they do not know for certain.

And yet: fish really do hear infrasound

Here the story turns, because the sensory half of the folklore is not nonsense at all. It is measured, repeatedly, in three species, by one research group in Oslo — and the numbers are remarkable.

In 1986 Olav Sand and Hans Erik Karlsen put Atlantic cod in a standing-wave acoustic tube and used cardiac conditioning to find their threshold. The cod responded to infrasound down to 0.1 Hz — one cycle every ten seconds — with a threshold, expressed as particle acceleration, of about 10⁻⁵ m/s². Sensitivity kept improving as the frequency fell below 10 Hz. Since ocean background noise rises steeply toward the lowest frequencies, the authors proposed something lovely: that fish may read the sea's infrasound field as an aid to navigation on migration.2

Karlsen then answered the obvious objection — maybe it is the lateral line, the row of flow detectors along a fish's flank, rather than hearing. He tested perch with their lateral-line organs chemically blocked, and the infrasound thresholds did not budge: down to 0.3 Hz at about 2×10⁻⁴ m/s², blocked or not. The organ responsible is the inner ear.5 In plaice, in a seawater chamber swung on steel wires, he measured about 4×10⁻⁵ m/s² at 0.1 Hz and concluded the otolith organs behave as slightly underdamped harmonic oscillators.6

Behaviourally it is just as clear, and frequency-specific in a way that is hard to explain away. Juvenile salmon exposed to 10 Hz flee immediately; at 150 Hz they show no visible reaction at all, even a few centimetres from the source.7