San Andreas—a Hollywood action-adventure film set in California amid not one, but two magnitude 9+ earthquakes in quick succession and the destruction that follows—was released worldwide today. As the movie trailers made clear, this spectacle is meant to be a blockbuster: death-defying heroics, eye-popping explosions, and a sentimental father-daughter relationship. What the movie doesn’t have is a basis in scientific reality.
Are magnitude 9+ earthquakes possible on the San Andreas Fault?
Thanks to the recent publication of the third Uniform California Earthquake Rupture Forecast (UCERF3), which represents the latest model from the Working Group on California Earthquake Probabilities, an answer is readily available: no. The consensus among earth scientists is that the largest magnitude events expected on the San Andreas Fault system are around M8.3, forecast in UCERF3 to occur less frequently than about once every 1 million years. To put this in context, an asteroid with a diameter of 1,000 meters is expected to strike the Earth about once every 440,000 years. Magnitude 9+ earthquakes on the San Andreas are essentially impossible because the crustal fault zone isn’t long or deep enough to accumulate and release such enormous levels of energy.
My colleague Delphine Fitzenz, an earthquake scientist, in her work exploring UCERF3, has found that, ironically, the largest loss-causing event in California isn’t even on the San Andreas Fault, which passes about 50 km east of Los Angeles. Instead, the largest loss-causing event in California is one that spans the Elsinore Fault and runs up one of the blind thrusts, like the Compton or Puente Hills faults, that cuts directly below Los Angeles. But the title Elsinore+ Puente Hills doesn’t evoke fear to the same degree as San Andreas.
Will skyscrapers disintegrate and topple over from very strong shaking?
Short answer: No.
In a major California earthquake, some older buildings, such as those made of non-ductile reinforced concrete, that weren’t designed to modern building codes and that haven’t been retrofitted might collapse and many buildings (even newer ones) would be significantly damaged. But buildings would not disintegrate and topple over in the dramatic and sensational fashion seen in the movie trailers. California has one of the world’s strictest seismic building codes, with the first version published in the early part of the 20th century following the 1925 Santa Barbara Earthquake. The trailers’ collapse scenes are good examples of what happens when Hollywood drinks too much coffee.
A character played by Paul Giamatti says that people will feel shaking on the East Coast of the U.S. Is this possible?
First off, why is the movie’s scientist played by a goofy Paul Giamatti while the search-and-rescue character is played by the muscle-ridden actor Dwayne “The Rock” Johnson? I know earth scientists. A whole pack of them sit not far from my desk, and I promise you that besides big brains, these people have panache.
As to the question: even if we pretend that a M9+ earthquake were to occur in California, the shaking would not be felt on the East Coast, more than 4000 km away. California’s geologic features are such that they attenuate earthquake shaking over short distances. For example, the 1906 M7.8 San Francisco Earthquake, which ruptured 477 km of the San Andreas Fault, was only felt as far east as central Nevada.
Do earthquakes cause enormous cracks in the earth’s surface?
I think my colleague Emel Seyhan, a geotechnical engineer who specializes in engineering seismology, summed it up well when she described this crater from a trailer as “too long, too wide, and too deep” to be caused by an earthquake on the San Andreas Fault and like nothing she had ever seen in nature. San Andreas is a strike-slip fault; so shearing forces cause slip during an earthquake. One side of the fault grinds horizontally past the other side. But in this photo, the two sides have pulled apart, as if the Earth’s crust were in a tug-of-war and one side had just lost. This type of ground failure, where the cracks open at the surface, has been observed in earthquakes but is shallow and often due to the complexity of the fault system underneath. The magnitude of the ground failure in real instances, while impressive, is much less dramatic and typically less than a few meters wide. Tamer images would not have been so good for ticket sales.
Will a San Andreas earthquake cause a tsunami to strike San Francisco?
San Andreas is a strike-slip fault, and the horizontal motion of these fault systems does not produce large tsunami. Instead, most destructive tsunami are generated by offshore subduction zones that displace huge amounts of water as a result of deformation of the sea floor when they rupture. That said, tsunami have been observed along California’s coast, triggered mostly by distant earthquakes and limited to a few meters or less. For example, the 2011 M9 Tohoku, Japan, earthquake was strong enough to generate tsunami waves that caused one death and more than $100 million in damages to 27 harbors statewide.
One of the largest tsunami threats to California’s northern coastline is from the Cascadia Subduction Zone, stretching from Cape Mendocino in northern California to Vancouver Island in British Colombia. In 1700, a massive Cascadia quake likely caused a 50-foot tsunami in parts of northern California, and scientists believe that the fault has produced 19 earthquakes in the 8.7-9.2 magnitude range over the past 10,000 years. Because Cascadia is just offshore California, many residents would have little warning time to evacuate.
I hope San Andreas prompts some viewers in earthquake-prone regions to take steps to prepare themselves, their families, and their communities for disasters. It wouldn’t be the first time that cinema has spurred social action. But any positive impact will likely be tempered because the movie’s producers played so fast and loose with reality. Viewers will figure this out. I wonder how much more powerful the movie would have been had it been based on a more realistic earthquake scenario, like the M7.8 rupture along the southernmost section of the San Andreas Fault developed for the Great Southern California ShakeOut. Were such an earthquake to occur, RMS estimates that it would cause close to 2,000 fatalities and some $150 billion in direct damage, as well as significant disruption due to fault offsets and secondary perils, including fire following, liquefaction, and landslide impacts. Now that’s truly frightening and should motivate Californians to prepare.
You May Also Like
November 13, 2017
Puebla Earthquake: New Insights from RMS Reconnaissance
This blog is a reprint of an article published in Canadian Underwriter
New insights often challenge the established view. The view of earthquakes in Canada is changing, including shifts in the seismic risk within the greater Metro Vancouver area and in the balance of seismic risk between the east and west.
Starting with Metro Vancouver, insured seismic risk was previously viewed as being more heavily concentrated in the city proper, given the exposure concentration, including a prevalence of high-value buildings. But based on insights, the product of a new RMS model focused on earthquake risk in Canada, it appears insured seismic risk is driven more by exposure in the expansive region to the south of Vancouver, which straddles the main arm of the Fraser River.
Why is this region in focus? It is worth examining how it was created. For thousands of years, the Fraser River carried sand, silt, and other sediments westward towards the Georgia Straight, helping to create a network of islands. Today, these islands and their surroundings are home to the fast-growing City of Richmond, the Vancouver International Airport, Deltaport container terminals and more, all of which sit on top of what geologists call a sedimentary basin. The portion of the basin to the south of Vancouver, call it the Vancouver Basin, has deposits of soft soils as deep as 300 meters.
The Vancouver Basin accounts for about 10 percent of the total earthquake-exposed value within British Columbia, but contributes about 20 percent of the province’s modeled insured (or gross) average annual loss and 500-year return period loss (for this particular article, earthquake-exposed value — that is, after taking into account penetration rates, but before applying policy terms — and modeled loss considers all lines of business and coverages, although not high-value specialty occupancies, such as power plants, which were removed to generate results that reflect more typical insurance portfolios).
In contrast, the City of Vancouver accounts for about 13 percent of British Columbia’s total earthquake-exposed value while contributing about 15 percent of the modeled insured average annual loss and 500-year return period loss.
It is tempting to attribute this shift to differences in growth patterns between the City of Vancouver and the Vancouver Basin. But analysis indicates the new view of risk is driven by improvements in seismic risk modeling and not changes in insured exposure.
Ground motion amplification that occurs during earthquakes is a key consideration. Softer site conditions, such as those made up of silts and clays, tend to amplify shaking compared with stiffer site conditions, such as those made of rock. This amplification has been observed in past earthquakes and contributed, for example, to the heavy damage in the Marina District of San Francisco, following the magnitude 6.9 Loma Prieta Earthquake in 1989.
Although ground motion amplification can occur anywhere there are soil deposits, the magnitude of the amplification is even more pronounced for buildings on sedimentary basins, such as is the case in the Vancouver Basin.
Figure 1: Map of Vancouver Basin. Darker colors indicate deeper deposits of soft soils.The geologic structure and deep deposits of soft soils within a basin amplify shaking several times greater than would otherwise be observed for taller buildings. For example, the sedimentary basin beneath Mexico City, coupled with inadequate design and construction standards, resulted in the collapse of more than 400 buildings following the magnitude 8.0 Michoacán Earthquake in 1985.
The magnitude of ground motion amplification depends on a number of factors, including the magnitude of an earthquake, the distance between the earthquake source and the site, the depth of the basin and the characteristics of a particular building. In general, shaking can be as much as two to four times greater for taller buildings within basins compared with those located on rock.
Modeling techniques that only consider the surface soil conditions generally underestimate the ground motion amplification within basins. But recent research by Sheri Molnar at the University of Victoria and scientists at Natural Resources Canada and elsewhere have provided important data to improve modeling the potential impact of amplification.
Predicting where liquefaction will occur is another important feature of modeling to identify risk. Liquefaction — a process in which loose and saturated sandy soil transforms into a semi-liquid state during strong earthquake shaking — was a major cause of damage during the 2010-2011 Canterbury earthquake sequence in New Zealand. One of the tremors, the magnitude 6.3 Lyttelton Earthquake in 2011, led to widespread liquefaction in the City of Christchurch, especially along the Avon River, which cuts through the city. The liquefaction caused the ground to vertically settle and, especially closer to the river, move laterally, undermining building foundations.
The destruction and loss of life from these earthquakes was tragic, but the events provided a huge opportunity to learn, acting as a laboratory for the earthquake engineering community. The billions of dollars in claims data, in addition to numerous studies by academics and practitioners, has resulted in a vastly improved understanding of the conditions necessary to cause liquefaction and its impact on the built environment.
One observation from Christchurch is that shallow ground water acts as a sort of fuse for liquefaction, and without it, regardless of other conditions, the process will not occur.
Another observation is that the severity of liquefaction varies significantly over short distances. Maps of liquefaction severity following the Lyttelton Earthquake show dramatic differences in liquefaction over areas just tens of meters apart. This is especially true close to water channels because of the prevalence of loose soil and their “open faces,” which allow the ground to displace laterally.
Figure 2: Liquefaction contribution to modeled insured average annual loss in the City of Vancouver, the Vancouver Basin and surrounding area.The Fraser River Delta in British Columbia has all of these liquefaction triggers present: shallow ground water; loose, sandy soil; and a web of river channels. It also has the most important trigger of them all, nearby faults and the Cascadia Subduction Zone, with the capacity to produce large earthquakes.
Lower Insured Seismic Risk
Looking north of the Fraser River, towards the City of Vancouver, modeled insured seismic risk has reduced. Ironically, this is partly driven by reduced modeled ground motion amplification and liquefaction in the city, but there are other important factors at play. For example, some types of reinforced concrete buildings that are common in the city are now viewed as less vulnerable than they were previously because engineers now better understand their seismic performance.
Looking towards the east, to the provinces of Ontario and Quebec, modeled insured loss (average annual loss and the 500-year return period loss) for these provinces has reduced by more than 40 percent. This change is driven primarily by an improved understanding of seismic hazard within the St. Lawrence River Valley, home to the cities of Montreal and Quebec City, among others.
This leads to another important shift in Canada’s overall modeled insured seismic risk, considering a nationwide perspective. With seismic risk reducing in Ontario and Quebec while remaining relatively stable in British Columbia as a whole, seismic risk is even more heavily concentrated in the west. British Columbia now accounts for almost 65 percent of the country’s insured modeled average annual loss and about 75 percent of the 500-year return period loss.
While changes in risk modeling drive this new view, it is important to acknowledge that earthquake insurance take-up is much lower in Ontario and Quebec than it is in British Columbia. Risk modeling techniques have improved in recent years, thanks to the advances in earthquake science and engineering, as well as lessons learned from past earthquakes. These advances provide insights into a changing insured seismic risk landscape across Canada.
The insurance industry should be aware of these changes and strive to use these emerging insights in their business decisions.…
Senior Product Manager, Model Product Management, RMS
Justin is a senior product manager in the Model Product Management team, focusing on RMS Latin America, Caribbean and Canada earthquake models. He supports the product definition and change management activities for updates to these models. Prior to RMS, Justin managed engineering and research projects focused on community resilience to earthquakes for GeoHazards International, and he has experience conducting site-specific seismic evaluations of existing buildings. Justin is a registered civil engineer in the state of California and holds a bachelor's degree in structural engineering from the University of California, San Diego, a master's degree in structural engineering from the University of California, Berkeley, and a post-graduate diploma from the London School of Journalism. Justin is a member of the Earthquake Engineering Research Institute.