Nobody expected the barrels to last.

That was no secret. Steel drums of laboratory waste, weighted with concrete and tipped off the stern of a ship twenty-five to sixty miles west of San Francisco, were only ever asked to survive the trip down. A 1981 review by the General Accounting Office set out the reasoning without embarrassment: the containers were not intended to hold anything permanently, only to carry the waste to the bottom, where currents were expected to dilute it until it stopped mattering.

Between 1946 and 1970, roughly 47,800 of them went over the side.

What was in the drums

Cold War paperwork was never designed to be audited, so every tally comes with a shrug built in. A U.S. Geological Survey open-file report puts the total at about 47,800 containers, mostly 55-gallon steel drums, alongside concrete blocks and other packaging. Records place them across three designated sites: roughly 150 drums at 90 metres depth, 3,600 at 900 metres, and 44,000 at a nominal 1,800 metres. Excluding tritium, the material carried an estimated 540 terabecquerels of thorium, uranium, transuranic elements and mixed fission products. Nobody knows the precise breakdown by isotope.

“Low-level” is a leftover category rather than a reassurance. It means anything that is neither high-level nor transuranic waste, which in practice covered broken glassware, ashes, animal carcasses and assorted laboratory equipment, plus whatever else had been near a reactor or a research bench.

Missing the target

Navigation in the 1940s and 1950s ran on dead reckoning and optimism. Crews aimed for the designated sites and often missed, and since detailed record-keeping was not required, no one was checking their work. The Navy, which handled most of the disposal on behalf of the Atomic Energy Commission, kept no detailed information at all on its Pacific dumping runs.

What exists now is spread thin across shelf and slope. Drums are thought to litter about 1,400 square kilometres of sea floor at depths from 100 metres to more than 2,000 metres, an area catalogued as the Farallon Island Radioactive Waste Dump. Much of it falls inside the Greater Farallones National Marine Sanctuary, designated in 1981, eleven years after the last container was tipped overboard.

Sonar, then eyes

In 1990 the Geological Survey and the sanctuary ran sidescan sonar across part of the dump, and image enhancement let analysts pick out barrel-shaped targets over roughly 125 square kilometres. Confirming them meant going down. A camera sled made the first pass, and in 1994 the Navy deep submergence vehicle Sea Cliff and an unmanned Advanced Tethered Vehicle visited targets directly.

They found a spectrum, from drums sitting perfectly intact on the sediment to others that had disintegrated entirely, leaving a smear of rust and concrete where a container used to be.

Whether any of it matters

The official answer has been consistent and faintly impatient. That 1981 review canvassed more than thirty scientists and concluded that public concern had been overstated. By its reckoning, an estimated 90 per cent of the dumped radioactivity had already decayed to negligible levels, and radionuclides in fish sampled near the site could not be told apart from weapons-test fallout.

One paper complicates the picture. Writing in Health Physics in 1996, Suchanek and colleagues reported isotope signatures in fishes and mussels from the dump site that were badly out of step with fallout. Plutonium-238 came in at about four times plutonium-239/240, where weapons fallout gives a ratio nearer 0.03. Americium-241 ran to roughly 30 times plutonium, against a fallout figure of 0.43. Projected doses to people eating Dover sole, sablefish and thornyheads still averaged only about 0.03 millisieverts a year. The study drew a published rebuttal in the same journal, and one study is one study.

Radiation is not the only worry down there. Sanctuary records note that cyanides, mercury, beryllium and other heavy metals were tipped into the same waters, along with explosives, dredge spoils and ordinary rubbish. A scientific panel convened by the Oceanic Society in 1980 set out to demonstrate that the drums were a hazard, failed to find the evidence, and finished by recommending that regulators pay more attention to the toxic chemicals than to the radioactivity.

The carrier everyone worried about

For decades the most alarming object off San Francisco was a light aircraft carrier. USS Independence survived the Bikini Atoll atomic tests, served afterwards as a floating laboratory for decontamination experiments, and was scuttled in 1951 with radioactive material still aboard.

In 2015 NOAA, Boeing and Coda Octopus surveyed the wreck using an autonomous underwater vehicle called Echo Ranger, and found the hull remarkably intact in about 790 metres of water inside the neighbouring Monterey Bay sanctuary. A write-up in Frontiers in Marine Science reported no detectable radioactivity at the site, while declassified documents confirmed the ship had genuinely been used as a store for radioactive waste when it was sunk. Six decades of rumour, settled by one robot and an archive request.

What the ban did and did not do

Ocean disposal of radioactive waste is now prohibited outright. An amendment to the London Convention, adopted in 1993, entered into force in February 1994 and shut the practice down worldwide.

It did not retrieve a single drum.

No treaty obliges anyone to go and collect them, and no agency has proposed it, partly because pinpointing every container would cost far more than the measured hazard appears to justify. So the sanctuary now protects a seabed whose contents were settled long before anyone thought to protect it, and the working policy is to leave the drums where they lie and keep watching. A decision taken by ships’ crews working from approximate positions has become permanent, and nobody has had to make it again.