Mysterious earthquake reveals Cascadia’s hidden dangers

The origin of the 1954 magnitude 6.5 earthquake that rattled residents around Humboldt Bay has been unclear, but a new study now suggests a surprising source: the Cascadia subduction interface.

In the Bulletin of the Seismological Society of America, a team of researchers shares the scientific sleuthing that led them to that conclusion, in a tale that combines fading paper records, modern methods and modeling, and eyewitness accounts of the event.

The December 21, 1954, event occurred in a region where earthquakes are common. This part of coastal northern California includes the Mendocino Triple Junction, where the Pacific, Gorda and North American tectonic plates meet and is the most seismically active area of the lower 48 states.

But the 1954 earthquake was unusual in its location, magnitude and shaking intensity. Historically, most large earthquakes in the area are located within the Gorda Plate, either offshore or the portion of it that is subducted beneath the North American plate. There have been no large earthquakes on surface faults in the North American plate in the region in instrumental times, although the faults have been mapped as active.

Peggy Hellweg, a retired seismologist at the University of California, Berkeley's Seismological Laboratory and colleagues conclude that the earthquake was a thrust event, located about 11 kilometers deep below Fickle Hill to the east of the city of Arcata. These characteristics taken together suggest that the most likely source of the earthquake is the Cascadia subduction interface, they write.

The Cascadia Subduction Zone along the Pacific Northwest coast looms large in the scientific and public minds, as it has the potential to generate great earthquakes. The magnitude 9.0 Cascadia earthquake in 1700 drowned forests, sunk coastlines by six feet and led to a massive tsunami that caused damage as far away as Japan.

The Fickle Hill earthquake could help answer questions that seismologists have been working diligently to solve: does the Cascadia subduction interface only rupture in large, 1700-style earthquakes? Does the entire interface always rupture, or can smaller parts of it rupture on their own?

There's only one other large recorded earthquake in the area -- the 1992 magnitude 7.2 Cape Mendocino event -- that may have its origins on the subduction interface, said Hellweg.

"And then we have the big one from 1700 when it was the entire fault," she said. "But we really don't know of any earthquakes that we've measured with instruments that were on the interface. And people have postulated that it is locked and nothing's going to happen until the next big one comes."

"Cascadia is really unusual in that in the instrumental era, it has been eerily quiet," said Lori Dengler, a retired seismologist from Cal Poly Humboldt and one of the study's co-authors. "We don't have smaller earthquakes, and that's not something you usually see in subduction zones."

In Humboldt County, Dengler added, there's the question of whether mapped faults in the overlying North American plate that are related to the subduction interface "rupture on their own or do they only rupture as part of a megathrust event? It looks like this is a little patch on the megathrust that did rupture. So this is really new in terms of our understanding of how Cascadia works."

Hellweg and colleagues spent three years revisiting the enigma of the 1954 event, which has gone by many names over the years. They analyzed published earthquake catalogs, unpublished data from the Berkeley archives and newly identified data from accelerometers that were operated at the time of the earthquake by the United States Coast and Geodetic Survey (USCGS).

Along the way, Hellweg recruited colleagues to contribute their expertise in locating and digitizing records, creating a probability cloud for the earthquake's hypocenter using modern software, and determining a mechanism for the earthquake.

It was especially helpful to find records of how these data were collected, including how the relevant stations and instruments operated, and what calculations were made with these data throughout the years, Hellweg said, noting the importance of preserving those types of records.

"Even when we think about our modern data collection, and what we preserve, we need to think about it in terms of somebody who in 50 years might want to go back and look at it," she said. "The metadata are really important."

The researchers also revisited estimates of the earthquake's intensity, with the help of reports that detailed damaging and felt shaking that had been gathered by the USCGS, newspaper archives, photos, maps of damage to the water supply for the nearby town of Eureka, and newly collected eyewitness accounts.

As part of the study, the researchers placed a call for earthquake stories in local newspapers and Facebook groups. Stories came in from people who were children when the earthquake happened 71 years ago, but they had remarkably consistent memories of sloshing bathtubs, toppling chimneys and rolling ground that allowed Hellweg and colleagues to estimate the earthquake's intensity.

One 11-year-old girl was riding her bicycle with a friend when they felt the shaking, and the two of them immediately dropped to the ground and covered their heads, doing what they had been taught in their school's atomic bomb drills.

She remembers rolling ground, toppling chimneys and sparking power lines, but one of the images that stuck with her was the unheard-of sight (in 1954) of a woman coming out of her home with her hair still in curlers.