Simulating 800,000 Years Of California Earthquake History To Pinpoint Risks

From The Texas Advanced Computing Center

Published on January 25, 2021 by Aaron Dubrow

A randomly selected 3,000-year segment of the physics-based simulated catalog of earthquakes in California, created on Frontera. [Credit: Kevin Milner, University of Southern California]

Massive earthquakes are, fortunately, rare events. But that scarcity of information blinds us in some ways to their risks, especially when it comes to determining the risk for a specific location or structure.

“We haven’t observed most of the possible events that could cause large damage,” explained Kevin Milner, a computer scientist and seismology researcher at the Southern California Earthquake Center (SCEC) at the University of Southern California.

“Using Southern California as an example, we haven’t had a truly big earthquake since 1857 — that was the last time the southern San Andreas broke into a massive magnitude 7.9 earthquake. A San Andreas earthquake could impact a much larger area than the 1994 Northridge earthquake, and other large earthquakes can occur too. That’s what we’re worried about.”

The traditional way of getting around this lack of data involves digging trenches to learn more about past ruptures, collating information from lots of earthquakes all around the world and creating a statistical model of hazard, or using supercomputers to simulate a specific earthquake in a specific place with a high degree of fidelity.

3D view of one especially complex multi-fault rupture from the synthetic earthquake catalog. [Credit: Kevin Milner, University of Southern California]

However, a new framework for predicting the likelihood and impact of earthquakes over an entire region, developed by a team of researchers associated with SCEC over the past decade, has found a middle ground and perhaps a better way to ascertain risk.

A new study led by Milner and Bruce Shaw of Columbia University, published in the Bulletin of the Seismological Society of America in January 2021, presents results from a prototype Rate-State earthquake simulator, or RSQSim, that simulates hundreds of thousands of years of seismic history in California. Coupled with another code, CyberShake, the framework can calculate the amount of shaking that would occur for each quake. Their results compare well with historical earthquakes and the results of other methods, and display a realistic distribution of earthquake probabilities.

According to the developers, the new approach improves the ability to pinpoint how big an earthquake might occur in a given location, allowing building code developers, architects, and structural engineers to design more resilient buildings that can survive earthquakes at a specific site.

“For the first time, we have a whole pipeline from start to finish where earthquake occurrence and ground-motion simulation are physics-based,” Milner said. “It can simulate up to 100,000s of years on a really complicated fault system.”


RSQSim transforms mathematical representations of the geophysical forces at play in earthquakes — the standard model of how ruptures nucleate and propagate — into algorithms, and then solves them on some of the most powerful supercomputers on the planet. The computationally-intensive research was enabled over several years by government-sponsored supercomputers at the Texas Advanced Computing Center, including Frontera — the most powerful system at any university in the world — Blue Waters at the National Center for Supercomputing Applications, and Summit at the Oak Ridge Leadership Computing Facility.

“One way we might be able to do better in predicting risk is through physics-based modeling, by harnessing the power of systems like Frontera to run simulations,” said Milner. “Instead of an empirical statistical distribution, we simulate the occurrence of earthquakes and the propagation of its waves.”

“We’ve made a lot of progress on Frontera in determining what kind of earthquakes we can expect, on which fault, and how often,” said Christine Goulet, Executive Director for Applied Science at SCEC, also involved in the work. “We don’t prescribe or tell the code when the earthquakes are going to happen. We launch a simulation of hundreds of thousands of years, and just let the code transfer the stress from one fault to another.”

The simulations began with the geological topography of California and simulated over 800,000 virtual years how stresses form and dissipate as tectonic forces act on the Earth. From these simulations, the framework generated a catalogue — a record that an earthquake occurred at a certain place with a certain magnitude and attributes at a given time. The catalog that the SCEC team produced on Frontera and Blue Waters was among the largest ever made, Goulet said. The outputs of RSQSim were then fed into CyberShake that again used computer models of geophysics to predict how much shaking (in terms of ground acceleration, or velocity, and duration) would occur as a result of each quake.

“The framework outputs a full slip-time history: where a rupture occurs and how it grew,” Milner explained. “We found it produces realistic ground motions, which tells us that the physics implemented in the model is working as intended.” They have more work planned for validation of the results, which is critical before acceptance for design applications.

Read the full article here.

5 6 votes
Article Rating
Newest Most Voted
Inline Feedbacks
View all comments
Joel O'Bryan
January 25, 2021 10:38 pm

How many reasons do I need not to live in SoCal???? 7.9 reasons.

John Galt III
Reply to  Joel O'Bryan
January 26, 2021 6:01 pm

California Dreamin’

So an Earthquake will destroy your house
Or the idiotic forest practices will burn your house
Or the Democrats will destroy your job, make you foreclose and they give your house to an illegal alien.
Oh Sweet Saint of San Andreas, hear my prayer.

January 26, 2021 1:19 am

There were 6 Italian scientists convicted of manslaughter in relation to a 2009 earthquake because they had prior declared no danger as energy was serially discharging. They were to pay victims’ families ~US$6 million & serve 6 year prison sentences; in 2014 the convictions were overturned freeing the scientists.

The U.S.A. is litigation crazy. The OriginalPost describes wanting “design applications”, which sounds like a lawyer’s dream scenario.

I’ve been in numerous earthquakes. I built my house in rural Gringolandia with local stone, tree rafters, cement, rebar/steel & tin roof. During 1 big provincial earthquake I could hear the force shift sounding like sequentially marching wall column to wall column going once around the rectangular building. The tin roof has a central ridge with 4 sides of roof pitches coming off of that allowing it kind of flop up & down in place, unlike the easier/cheaper to build dual slope tin roof which fail easier (an earthquake can un-align a wall & if only 1 roof angles catching support it can collapse).

Peta of Newark
Reply to  gringojay
January 26, 2021 2:17 am

Nailed it gringo.
a) They are gonna fall foul of the Jevon’s Paradox – they’re going to increase the number of buildings/homes/whatevers in places where quakes happen.
i.e. An increase in building efficiency will create greater overall consumption of buildings.

Folks will be given a fake sense of security by the imposition of Government diktat which was delivered by the authority of the ubiquitous SUPER computer.

And get The Hubris. The lack of self-awareness,
That they really do believe it is THEY who are Super and are so dumb they feel they must constantly tell everyone as much.

If they really were/are ‘super’ wouldn’t it be obvious?
(The girls know quite different, the complete lack of ‘super-ness’ leading hence the lack of babies. And notice how effectively they keep their mouths shut as well as keeping their legs crossed. Multi-tasking is real. haha)

b) Buildings are thus going to be built DOWN to A Standard.
And will Fall Down like perfect synchronised dominoes and thus: a minor catastrophe will become major
As every human endeavour has done previously in all of history, it will end in calamity & tears.
Just ask NASA, Boeing or Boris Johnson as recent examples

In places like they’re talking about, build as strong as you can (afford) to.
Always expect The Worst in those situations.
Then the lawyers/parasites won’t be able to get a handle on you when it does go pear-shaped, AND, you’ll have some money/resource available to rebuild/relocate as is your whim. Not theirs.

It is exactly, in fact, The First Law Of Insurance.
Buy insurance > you won’t need it
Don’t buy insurance > you’ll need it tomorrow
(For those who’ve not had any coffee, cocaine or nicotine yet today, insurance is not necessarily ‘monetary’)

Steven Johnson
January 26, 2021 1:57 am

“The framework outputs a full slip-time history: where a rupture occurs and how it grew,” wow

Ron Long
January 26, 2021 2:27 am

Interesting supercomputer presentation, but not the predictive tool they hope. The capture of the North American plate by the pacific plate began in the Cretaceous along the Wasatch Front in Utah. At 40 million years ago it was in north-central Nevada, where south-migrating magmatism/volcanism, with tendencies to form gold deposits) also arrived at the same time. So, us gold exploration geologists from Nevada study these strike-slip faults, and all of their associated effects, like R1 shears, extension zones, step-overs, rebound, etc, because the gold deposits are controlled by these fracture systems. I held my handy (plasticized, so you don’t drip sweat or beer on it) strike-slip fault diagram to the video and the figure and can readily see the right-lateral San Andreas, the antithetic left-lateral strike slip-faults, the extensional zones, and the reverse faults. What these studies and data, as presented, is lacking is crustal inhomogenaity (causing step-overs) positive or negative flower structures at transpressive or transgressive bends, rebound, and activation of older structures conveniently positioned to relieve stress. The destruction from fault movement is due to two events, the primary is compressive and radiated out perpendicular tot he fault plane, and the secondary effect is sine wave movement that radiates in all directions (the primary feels explosive if it hits you and the secondary produces the rocking back-and-forth commonly seen on TV), so any analysis of potential destruction needs to separate the two waves into probability scenarios, good luck with that. This theme is far more complex than the ability of a supercomputer to understand it.

John Galt III
Reply to  Ron Long
January 26, 2021 6:26 am

The underlying type of rock is also important. If you go to Parkfield, CA in Monterey County CA smack dab on the San Andreas fault and there are signs on the bridge there denoting the North American Plate and Pacific Plate, earthquakes don’t cause much damage because the underlying rock is Serpentinite – think jello. Then go to Southern CA and there is granite and similar rocks at depth – think Velcro. The amount of displacement built up in Velcro type rocks is huge but eventually the build up is too great and the Velcro gives way – now you have an 7.5 to 8.0 Moment Magnitude 90 second or longer disaster, Last time I read there is 30 feet of displacement built up in So Cal. By comparison San Francisco’s 1906 big one had a max of 22 feet displacement after the fact.

Ron Long
Reply to  John Galt III
January 26, 2021 10:01 am

John Galt III, you are correct about the underlying type of rock being important. My reference to that is “crustal inhomogeneity”. The build up of stress certainly shows there is a need for some strain to occur, however it might manage to creep and not rupture dramatically, although 30 feet of unresolved plate movement across a major fault is a lot of destructive potential.

Richard Page
January 26, 2021 3:17 am

Hmph. Well someone certainly didn’t want to leave their nice cosy air conditioned office to go tramping around muddy fields getting their shoes dirty and callouses on their hands did they? Anything to get out of doing an honest days work!

Mad Mac
January 26, 2021 4:50 am

We lived 6 miles from the epicenter of the Northridge quake fortunately on bedrock in Topanga Canyon. Scary way to wake up in the morning! Several large after shocks also. Lived near LAX during the Whittier Quake. I was coming back from a bike ride in the morning and I reached for the door knob and the door started shaking and I turned and looked down the street and the pavement was making rolling waves. We experienced a number of quakes in So Cal including the Landers quake in the Coachella valley.

Abolition Man
Reply to  Mad Mac
January 26, 2021 9:44 am

I was still in the Bay Area for the Loma Prieta quake of 1989. I was working on a major seismic remodel of an old theater when the quake hit; fortunately about 90% of our structural upgrades were complete so the theater rode it out with no problems! Our boss told us that David Packard, Jr. came out after with a big smile on his face as the only damage he saw was some cracks in the plaster!
About a year later I was working on a remodel of a home in Los Altos Hills that was severely damaged; apparently it was situated right over a small branch of the San Andreas. The house was supposedly being rebuilt to withstand an 8.0 quake! My favorite part was the new footing under the hot tub. The tub was place on a new concrete column about three feet in diameter, resting on bedrock, with a square pad about 12-15” thick on top. When we placed the hot tub I checked with the engineer what kind of fastening system he wanted because I couldn’t find anything noted on the plans. He told me to just leave it unattached; I guess it was the worlds first hot tub launching pad!

Tombstone Gabby
Reply to  Mad Mac
January 26, 2021 7:58 pm

A small correction if I may – Landers is in the high desert, north of Yucca Valley, not “down-below” in the Coachella Valley.

There’s a spot in Whitewater Canyon (off Interstate 10) where it is possible to stand and place your hands on a rock face – astride the San Andres fault. A different type of rock under each hand.

January 26, 2021 5:04 am

However, the ultimate output of the model said that global warming was responsible for all earthquakes and that we are safe now that Biden has rejoined Paris.

January 26, 2021 6:16 am

Maybe they should try a sorter period of time. They might gave better results.

January 26, 2021 2:46 pm

I read somewhere recently (sorry can’t remember the source) that there is now evidence that strong quakes in either the northern end of the San Andreas fault, or the southern end of the Cascadia Subduction fault can trigger the other fault to break as well.

Paul of Alexandria
January 26, 2021 3:51 pm

Personally, I’m more worried about the New Madrid fault. When that thing slips again the whole East Coast is toast.

John Tillman
January 26, 2021 4:28 pm

In Chile, I just missed two terremotos over Mag 8.0, ie the southern Bicentennial 8.8 in 2010 and northern 8.3 in 2015. I flew out the day before the Bicentennial, and would have been stuck for two weeks had I not. The latter lasted so long that my wife called me during it.

I was walking in Temuco during a 7.2 and didn’t notice it. The Acongaqua region of the Andes between Santiago and Mendoza had daily 5s and 6s last week.

Big Ring of Fire quakes come in storms. The 1960s were active.

The highest magnitude earthquake ever recorded, Mw 9.4-9.6, occurred in May 1960 near Valdivia, Chile. It was preceded the day before by three strong quakes starting at Concepcion, Mw 8.1, and heading south.

Locale Mapuches sacrificed a five year-old boy.

The tsunami hit Hawaii.

The probably second most powerful, Mw 9.2, struck Prince William Sound, Alaska in 1964. In the next year, the Rat Islands, Alaska experienced an 8.7.

The worst year for recorded East Pacific earthquakes might have been 1906. Both San Francisco, CA and the SF of South America, Valparaiso, Chile, from where I write, suffered Mw 7.9 and 8.2 temblors, respectively. For insurance reasons, in SF, the disaster is known as the Fire rather than the Earthquake.

Earlier on the same day as the Valparaiso terremoto, August 17, a Mw 8.3 quake hit Alaska’s Aleutian Rat Islands. It’s thought to be a coincidence. The HI tsunami came from Chile.

Ecuador and Colombia were struck by an 8.8 that year as well.

Like Chilean and Indonesian megathrust quakes, the Cascadian zone off the Pacific NW coast results from a triple plate junction. The last big one there was in 1700. Another is overdue.

If just the southern portion of Cascadia goes, the magnitude will be 8-9. If the north as well, then 9-10.

While densely built-up and populated SoCal is indeed in danger of massive death and destruction from the next Big One on the San Andreas and its outlying faults, Cascadia could be the Biggest One in magnitude.

January 28, 2021 12:02 pm

Interesting but how can the model be validated?
If there’s anything we have seen at WuWT – it is that computer models aren’t reliable.

Verified by MonsterInsights