A boat basin on the far north coast of California, a little before three in the morning.
The water in it is not rising or falling so much as sliding, the whole basin drawing sideways.
Along one pier a line of concrete floats tips, goes under, and comes back up in pieces.
Somewhere in the dark a cable shorts and throws sparks over the water.
The ocean here is doing something unusual.
It has done it many times before.
Why this one harbor keeps getting hit hardest
A tsunami in open ocean is almost nothing. A few feet of height spread across a wave hundreds of miles long, moving at the speed of an airliner.
What turns that into damage is the shape of the ground it runs into at the far end.
Off this stretch of coast the continental shelf is broad and shallow, and it acts as a ramp that slows the wave and stacks its energy into less water.
The coastline then does the rest. Offshore ridges and the bend of the shore steer wave energy toward the same point rather than spreading it along the beach.
Inside the basin the water then sloshes back and forth at the basin’s own rhythm, which keeps the motion going long after the first surge.
So the harbor is not unlucky. It is shaped wrong.
What was in the water that night
The quake struck offshore of the Kamchatka Peninsula at a depth of about 22 miles, on a subduction zone that produces the largest earthquakes on the planet.
Near the source the waves were enormous, with Russian institutions reporting splash heights up to about 62 feet in the northern Kuriles.
Across the Pacific the picture was different. Deep water buoys caught the wave within the hour, and the largest of those readings was later called the second biggest ever recorded on that network.
The basin on the California end had been rebuilt after the Pacific crossing of 2011, and that rebuild is the reason no boats were lost and nobody was hurt.
What did fail was a single long structure serving the larger vessels, which came apart in segments and sank along its own pilings.
The surge reached the basin near two forty in the morning and kept coming in sets.
Which number means what
Two different figures circulated for the same wave, and both are correct.
The height above the normal water level was about 3.7 feet. The full swing from trough to crest was 7.42 feet, which is the number the instrument reports.
A tide gauge measures the whole oscillation rather than a single crest, so a peak to trough figure is always the larger one.
Timing then did the rest of the damage. The surge arrived on a high tide of roughly the same size, so the crest rode on water that was already up.
For comparison the largest American reading of the event was in Hawaii at 11.44 feet, and a gauge in American Samoa read 6.25 feet.
Federal readings of this kind are instrument output, not eyewitness estimates.
What actually pulled the dock under
The day after the event the harbor described the lost structure as a wave and current attenuator that had played a sacrificial role as designed.
A week later the district published an engineering analysis that told a more specific story, and a more interesting one.
The failure mechanism was negative lift. Fast current accelerating underneath a floating concrete dock drops the pressure beneath it, and the float is pulled down rather than pushed aside.
Once the deck is below the surface the water simply runs over the top of it, and the structure loses along its length instead of at one joint.
The district called it a previously undocumented vulnerability, and proposed remedies as ordinary as sealed buoyancy tanks or a changed cross section.
Careful measurement keeps replacing folklore with mechanism, the way a catalog of sightings found no earthquake link for giant oarfish.
The engineering findings and the damage estimate are set out in the district release.
A structure can be sacrificial by intent and still fail by surprise.
What the record does and does not say
The reading was the largest on the continental coast for this event, and that qualifier carries the whole claim.
The same harbor has seen far worse. The Alaska earthquake of the 1960s sent water through the town that killed eleven people.
Damage this time ran to roughly a million dollars, concentrated in water lines, fire suppression and electrical service under the decking.
The instruments are the part worth watching, because a network that reads a wave mid ocean is what turns a distant rupture into a usable warning time.
It is the same discipline that put sonar over 27,000 barrels on the seafloor off Los Angeles.
The gauge readings and the continental comparison come from the federal summary.
The warning worked, and what failed was a detail nobody had modeled yet.
Read the whole thing?
Get the week's signal, not the noise
Our sharpest reporting on energy, climate and nature — free, once a week.
