On the sonar map they look manufactured.
Circles, thousands of them, each about the length of two football fields across and sunk maybe sixteen feet into the mud.
They are not scattered. They sit at regular spacing, spread across five hundred square miles of seabed.
Gas eruptions were the obvious reading, and for decades that is what they were.
Then the vehicles went down and sampled everything in reach.
There was no methane. Not in one sample.
Why the spacing was the thing that gave it away
Gas comes up where the plumbing lets it.
Faults, permeable layers, weak spots in the cap. A field driven by seepage is patchy, because the geology underneath is patchy.
What is on that seabed is the opposite. The depressions are evenly spaced across the whole field, which is a pattern you get from something sweeping across a surface rather than bursting up through it.
That something is a sediment gravity flow, which is an underwater avalanche of mud, sand and water running along the bottom.
The difference matters more than it sounds.
A gas field is chemical, continuous and awake underneath anything anchored above it. An avalanche is mechanical, episodic and passes through.
Same holes in the ground, two different problems for anyone planning to put something on top of them.
The record stacked underneath every crater
Under the seabed the sediment reads as a layer cake.
Fine mud settling out of the water column over long quiet stretches, cut over and over by bands of coarse sand that arrived fast.
Each sand band is one flow event, laid down in an afternoon rather than a millennium.
The deepest of them go back at least 280,000 years.
The detail that closes the argument is that the same bands show up in craters far apart from one another.
These were not thousands of separate local events. They were regional flows crossing the whole field at once, scouring each depression a little deeper as they went and keeping it open.
What actually went down there
The mapping came from autonomous vehicles flying low over the seabed with sonar and sub bottom profilers.
The sampling came from a remotely operated vehicle working the bottom directly.
Between them the team pulled 107 vibracores and 433 pushcores, plus piston and gravity cores running up to about twenty five feet.
Every one of those was checked for gas. Nothing.
The field sits in 1,600 to 5,200 feet of water off Big Sur, and the results came out in an earth surface journal in May of 2024, with Eve Lundsten of the research institute that ran it as lead author.
Funding came from the ocean energy regulator, the geological survey and a private foundation, which is worth knowing because of what the answer touches.
What the robots still could not work out
Nobody knows how the craters started.
Lundsten says it plainly in the write up, that the team was unable to determine how the pockmarks were initially formed.
The flows explain why they are still there. They do not explain the first hole.
What sets the flows off is also unresolved. Earthquakes, storm loading, slope failure upslope, all plausible and none demonstrated.
And the timing cuts both ways. The most recent event in that record ran roughly 14,000 years ago, which is either reassuring or simply outside the window anyone has instruments for.
A thirty year asset standing on ground that resets every few tens of thousands of years is not obviously exposed. It is also not obviously safe, because a record read at that resolution cannot rule out smaller flows arriving more often.
Reading a seabed this closely is still new work, the same way surveys of the ocean floor keep turning up structures nobody had counted.
Why any of this was funded in the first place
In late 2018 the federal ocean energy regulator named lease areas off central California for floating wind.
The field is in that water, and a seabed with thousands of unexplained holes is not something you anchor to without asking what made them.
Each of those seafloor craters runs about 512 feet across and sixteen feet deep, and there are more than five thousand of them.
That is why the survey happened at all, and the answer changes the engineering question rather than removing it.
Active gas would mean pressure and chemistry working against a foundation continuously.
Recurrent sediment flows mean a rare, violent, mechanical load that a mooring has to survive rather than resist daily, which is a design case a floating platform can actually be written around.
The craters were never erupting. They were being swept clean, every few tens of thousands of years, by something passing overhead.
Read the whole thing?
Get the week's signal, not the noise
Our sharpest reporting on energy, climate and nature — free, once a week.
