Seismo Blog
New Algorithm GFAST Enhances the ShakeAlert Earthquake Early Warning System
June 5, 2024
by Kelly Missett
The Pacific Northwest is earthquake country. Oregonians and Washingtonians may experience multiple types of earthquakes in their lifetimes, from moderate-sized earthquakes like the 2001 Nisqually event to a devastating megathrust earthquake on the Cascadia Subduction Zone. But the ShakeAlert® Earthquake Early Warning System can alert people before dangerous shaking arrives, potentially saving lives.
The ShakeAlert System, managed by the United States Geological Survey (USGS), is constantly being upgraded to improve alert speed and accuracy. It is extremely important to be able to quickly estimate the location and magnitude of major earthquakes so that alerts can rapidly be delivered to everyone in the affected areas. Scientists at the Pacific Northwest Seismic Network (PNSN) have developed a new geodetic algorithm called GFAST (Geodetic First Approximation of Size and Time) that will improve ShakeAlert’s ability to quickly characterize the magnitude of the next Big One. ShakeAlert is the world’s first earthquake early warning system that incorporates geodetic data and algorithms.
ShakeAlert uses real-time data from over 1500 seismic sensors and now, with the incorporation of the GFAST algorithm, utilizes data from over 760 Global Navigation Satellite System (GNSS) sensors for rapid earthquake detection. GNSS includes the well-known US-based Global Positioning System (GPS) as well as other satellite-based positioning systems used across the globe. Seismometers measure how quickly the earth is shaking in terms of velocity or acceleration. GNSS geodetic sensors measure how far the ground moves up, down, or sideways as a result of an earthquake. Small earthquakes cause small localized shifts. Larger magnitude earthquakes have greater velocity and acceleration of shaking as well as cause more extensive permanent ground displacement. During a Cascadia Subduction Zone earthquake, the ground could shift down and westward by several meters!
The Pacific Northwest Seismic Network operates most of the seismic stations in the Pacific Northwest. The geodetic stations used for GFAST are maintained by a range of other partners, including the EarthScope Consortium and Central Washington University.
a Regional seismic stations alone cannot accurately estimate the magnitudes (M) of the largest earthquakes. They saturate around M7, meaning they have difficulty discerning a M7 earthquake from a M9 earthquake using initial measurements. That could lead the ShakeAlert System to initially underestimate the magnitude of a Cascadia Subduction Zone earthquake, which would cause the ShakeAlert System to alert fewer people than needed. Because geodetic stations measure ground displacement, they can provide more accurate magnitude estimates for very large earthquakes. In a Cascadia Subduction Zone earthquake, this means a broader area and more people would be alerted, enabling them to take immediate life-saving protective actions.
GFAST was developed by PNSN researchers Brendan Crowell and Carl Ulberg, who work at the University of Washington. Research was also contributed by Diego Melgar, who directs the Cascadia Region Earthquake Science Center at the University of Oregon, and Jessica Murray of the USGS. Crowell and Melgar began developing the theory that peak ground displacement is analogous to earthquake magnitude in the early 2010s as part of their PhD research at the Scripps Institution of Oceanography. Crowell then used funds donated to the PNSN by the Gordon and Betty Moore Foundation and theAmazon Catalyst Program to develop the methodology and code base for the early GFAST algorithm and convert it to a format that was compatible with the ShakeAlert System. Additional funds from the USGS Earthquake Hazards and NASA Disasters programs helped to further refine the GFAST architecture.
In the early 2020s, Ulberg and others in ShakeAlert’s Software Management Working Group began integrating GFAST messages into the ShakeAlert solution aggregator. This aggregator was designed to evaluate the results of the two other earthquake detection algorithms already used by ShakeAlert and provide a definitive estimate of the detected earthquake’s size and location. Those existing algorithms, EPIC (Earthquake Point‐source Integrated Code) and FinDer (Finite‐Fault Rupture Detector), solely use data from seismic stations.
Ulberg created new logic within the ShakeAlert solution aggregator and GFAST itself in order to decide when and how to use GFAST in addition to EPIC and FinDer. For example, GFAST had to understand how many geodetic stations needed to provide data and how much displacement was required before an estimated magnitude would be considered reliable. It also had to weed out background “noise” from geodetic stations that could lead to false alerts. GFAST cannot reliably detect small and moderate earthquakes and works best in conjunction with seismic algorithms which can determine when and where earthquakes originate. For this reason, GFAST will start augmenting EPIC and FinDer for M7+ earthquakes and will become increasingly important the larger an earthquake grows.
Ulberg was also involved with testing the algorithm alongside colleagues from the USGS. The System Testing and Performance Group generated dozens of earthquake simulations to ensure that GFAST would perform well. It took around 10 years to create and test GFAST before it was finally ready to be fully integrated into the ShakeAlert System.
The earthquake early warning alerts received by end-users will look exactly the same and the speed of earthquake detection will not change. The public will not notice a difference but will benefit from the improved event size accuracy and more comprehensive alerting regions.
There are several ways that Oregonians and Washingtonians can receive earthquake early warning alerts on their cell phones. Alerts can be delivered via texts from the Wireless Emergency Alert System, push notifications from the Android Operating System, and push notifications from free ShakeAlert-powered apps like MyShake. To ensure alert delivery, make sure that your phone’s operating system is updated and that Emergency Alerts are turned on in its settings. iPhone users should also turn on Local Awareness. Learn more from theUSGS website.
During an earthquake seconds matter! Immediate action can save lives. As soon as you feel shaking or receive an alert, protect yourself!
- DROP where you are onto your hands and knees. This position protects you from being knocked down and reduces your chances of being hit by falling or flying objects. >
- COVER your head and neck with both arms and hands and bend over to protect your vital organs. If a sturdy table or desk is nearby, crawl underneath it for shelter. If no shelter is nearby, crawl next to an interior wall away from windows, hanging objects, and tall furniture.
- HOLD ON to your shelter until shaking stops. Be prepared to move with it if it shifts. If you have no shelter, hold on to your head and neck with both arms and hands.
If you are near or on the coast, or in a tsunami prone area, follow evacuation routes to higher ground or inland as soon as you are able after the shaking stops. Remember, the earthquake itself is the warning that a tsunami may be coming.
Medford Schools Now Use Earthquake Early Warning Technology - And Yours Could Too!
December 13, 2022
by Kelly Missett
The MyShake app is now delivering ShakeAlert-powered alerts in Washington
January 26, 2022
by Gabriel Lotto
What can the recent Kentucky tornado disaster teach us about earthquake early warning?
December 17, 2021
by Paul Bodin
Blanco Fracture Zone swarm: Active, unusual, interesting... but not concerning
December 10, 2021
by Alex Hutko
The CAscadia Seismic Imaging Experiment 2021 (CASIE21): All Aboard the R/V Marcus G. Langseth!
August 31, 2021
by Madeleine Lucas
December 2017 Oregon Tremor Event
Over the past 9-10 days, it appears that tremor in central Oregon has picked up (Figure 1). The last slow slip and tremor event was in February 2016, 22 months ago.
Figure 1
Figure 1. Age progression of tremor in central Oregon for the past 9 days. Earliest tremor locations start from 12/5/2017 and propagate roughly outward, clustering near Salem and Roseburg. Last update was December 14, 2017.
Tremor is the release of seismic noise from slow slip along the interface of the Juan de Fuca and North American plates and lasts for several weeks to months. This process is known as Episodic Tremor and Slip (ETS). Slow slip happens down-dip of the locked zone (Figure 2). The locked zone is where tectonic stress builds up until it releases in a great earthquake or megaquake. The recurrence interval of slow slip and tremor varies at different regions along the Cascadia Subduction Zone.
Figure 2
Figure 2. Cross section of the subducting Juan de Fuca Plate. Figure from Vidale, J. and Houston H. (2012) Slow slip: A new kind of earthquake (Physics Today, 2012 pages 38-43).
The last ETS event in Cascadia started in February 2017 around the western edge of the Olympic Mountains. The duration was approximately 35 days with a two-week quiescent period. Prior ETS events in northern Washington/Vancouver Island area was approximately December 2015.
The last ETS event in central Oregon was 2016 and lasted just over a week before it stopped on March 1, 2016.
ETS events are still being studied to understand the processes about slow slip and megathrust earthquakes.
More information about slow slip and tremor can be found here on the PNSN website.
Tremor has continued in Oregon since the last post on December 15th. Current tremor activity has been ongoing since about 12/5/2017 (figure 1).
Figure 1
Figure 1. Age progression of tremor in central Oregon for the past two weeks. Earliest tremor locations start from 12/5/2017 and propagate northerly and southerly. Last update was December 26, 2017.
Since December 19th, tremor has now migrated northerly toward Portland and southerly toward Medford (figure 2).
Figure 2
Figure 2. Tremor activity from 12/19 to 12/26 showing progression in a northern and southerly direction.
More FAQs on Slow Slip and Tremor
On our previous blog post, we briefly discussed what ETS (episodic tremor and slip) is. Let’s go through a couple of more frequently asked questions.
1.What is tremor?
Tremor in the Cascadia Subduction Zone is the seismic noise of slow moving earthquake along the interface of the subducting Juan de Fuca Plate and the North American plates. Compared to normal earthquakes, tremor has lower frequency energy and can last for minutes, hours or weeks.
2. What about volcanic tremor?
Tremor can also be volcanic. But ETS is deep, non volcanic signatures that are a result of plate motion, not magmatic movement.
3. How deep are the tremors?
As it states on our website - “This is a topic of ongoing research.” But research suggests that it occurs near the plate interface at approximately 30 - 40 km deep.
4. What is the magnitude of tremor?
Tremor is probably made up of many tiny individual earthquake-like sources each with a "magnitude" of less than 1. Since tremor is an on-going continuous signal assigning a magnitude to it is never done.
Check out the map on our web page:
In the previous blog post about The M9 Project, we talked about how the Cascadia Subduction Zone can generate an M9.0 earthquake. However, our understanding of what an earthquake of this scale would actually look like is less advanced. While we have evidence of past earthquakes (e.g., native oral histories, tsunami records), we have no quantitative observations of how strong the shaking would be during a megathrust earthquake in the Pacific Northwest.
To address this problem, researchers with The M9 Project used 3D computer simulations to help understand what 50 different realizations of an M9.0 earthquake could look like in Cascadia. To create these simulations, The M9 Project researchers used multiple supercomputers: Stampede (University of Texas - Austin), Constance (Pacific Northwest National Lab), and Hyak (University of Washington). A single earthquake simulation took up to 46 hours to complete. If it was possible to run these earthquake models on a personal computer (many of which have a mere 2 processors, compared to the 576 processors used to run these simulations on a supercomputer), it would take about 522 days to complete one simulation.
Why are earthquake simulations important?
The unique properties of the Cascadia Subduction Zone prevents a side-by-side comparison between a future Cascadia earthquake, and other earthquakes that have occured around the world. For instance, an M9.0 earthquake in Japan, Chile, or Indonesia may look very different from an M9.0 in the Pacific Northwest.
Scientists have developed equations that can estimate the strength of ground shaking based on an earthquake’s magnitude and a specific location’s distance from the fault. However, these equations still rely on averages, and do not fully account for location specific 3-D effects (i.e., “How will seismic waves bounce around in the Seattle basin?”). Conversely, the supercomputer earthquake simulations, while still having some unknowns, can estimate shaking at every point in the Pacific Northwest, and are specific to the geologic conditions of the Cascadia Subduction Zone.
How are these earthquake simulations created?
Out of an infinite number of possibilities, 50 simulations of an M9.0 earthquake were run by The M9 Project team. The individual earthquake scenarios had a few important variations between them, that made each earthquake source unique: (1) the hypocenter location (i.e., where the earthquake starts), (2) how far inland the rupture extends (i.e., how close the earthquake gets to major inland cities, such as Seattle), (3) the location of “sticky patches” on the fault, that generate the strongest ground shaking, and (4) the slip distribution on the fault (i.e., how far certain areas on the fault move during the earthquake).
Are certain earthquake scenarios “better” or “worse”?
The area affected by a megathrust earthquake is large enough that the outcome is going vary by location. A “best-case” scenario for one area in the Pacific Northwest could be a “worst-case” scenario somewhere else in the region.
One of the results of the computer simulations showed that when an M9.0 earthquake occurs on the Cascadia Subduction Zone, less violent shaking may be felt closer to the epicenter. This is because an earthquake on the Cascadia Subduction Zone will not occur at a single point -- instead, it will rupture a very large area. As the rupture moves along the fault, the seismic waves will start to “pile up,” similar to the Doppler Effect.
As the waves at the front of the rupture combine, their amplitudes get larger and create more violent ground motion. Therefore, locations closer to the hypocenter may receive less complex and destructive seismic waves than locations that are farther along the rupture and experience this “piling-up” of seismic energy.
In these two videos, notice how Seattle's mock seismogram has larger spikes (which denotes stronger ground motion) when the earthquake source is farther south, and the fault ruptures north.
This variation by location makes it virtually impossible to award a scenario the title “best-case” for the entire Pacific Northwest.
Can we do even better?
These computer simulations are the most accurate representations of what an M9.0 earthquake would look like in the Pacific Northwest. Unfortunately, there is a lot of variability in these calculations because there are still too many unknowns. An increase in seismic and GPS instrumentation throughout the Pacific Northwest, especially offshore, will help us identify more specifics about the Cascadia Subduction Zone and improve future computer simulations. For instance, we may be able to determine where “sticky patches” are located on the fault and obtain a more detailed image of the 3D structure of the subduction zone. Further constraining these variables in the computer simulations will ultimately help us refine our estimates of seismic hazards in the Pacific Northwest.
Special Thanks To
Dr. Erin Wirth, UW Affiliate Assistant Professor
M9 Simulation coverage from UW News
We are going to have The Big One. That’s a fact.
The final blog about The M9 Project is going to focus on you. What are you going to experience during a megathrust earthquake? How do we connect science and community? What should you do to be prepared?
The Next Stages of The M9 Project
Seismologists are not the only contributors to The M9 Project. Civil engineers, urban design and planners, statisticians, social scientists and public policy researchers also play a role in determining earthquake risk, safety measures, and public response to hazards.
Understanding the hazards and mechanisms of an earthquake is one thing, communicating effectively to the public is a completely different ball game. The next steps of The M9 Project focus on how we define and discuss hazards with communities.
For example, how does the way we design hazard maps affect how communities approach hazard planning? (See photo below) Or, how can hazards planning be steered towards rebuilding to community-specific values?
From assessing the utility of hazard maps (see image below), to hosting community planning workshops , The M9 Project’s research into “long term” preparedness- mitigation, response, and recovery focuses on how to best help you. We will discuss what to expect when the big one hits, as well as some resources so you can take steps to prepare .
Learning about Cascadia from other Large Earthquakes
The last megathrust earthquake on the Cascadia Subduction Zone occurred in 1700 AD, before written records were kept in the region. In addition to The M9 Project research at UW, we can also look to observations of other major earthquakes worldwide, to help us predict what The Big One may look like in the Pacific Northwest.
Ground Shaking
A magnitude 9 earthquake will generate very strong shaking for several minutes. The intensity, measured by the Modified Mercalli Intensity Scale (MMI), is determined by observations during an earthquake. Shaking tends to decrease farther away from the fault and will vary with local soil conditions, so intensity will vary by location. A more detailed description on intensity can be found here.
Below is a comparison of the shaking intensity from the 2001 Nisqually earthquake, compared to a hypothetical M9.0 earthquake scenario. A megathrust earthquake will be felt over a much larger area, and generate stronger shaking.
You can find more about how magnitude and intensity are related here.
As part of The M9 Project, UW civil engineers are researching building response to strong ground shaking from a magnitude 9.0 earthquake in Seattle. This video from Kinetica Dynamics shows skyscrapers in Tokyo shaking from the 2011 M9.0 Japan earthquake.
Tsunami
Large earthquakes on a subduction zone are capable of generating large tsunamis. For example, the 2004 M9.1 Sumatra earthquake resulted in a 30+ meter high tsunami on the west coast of Sumatra (source). For more about tsunamis, visit our tsunami overview page.
This NOAA video models the tsunami from the 1700 Cascadia earthquake, which caused damage and loss of life as close as the west coast of North America, and as far away as Japan.
We expect the next great Cascadia earthquake to be similar. As mentioned in our first M9 Project blog post, the record of past megathrust earthquakes can be found in muddy estuaries on the coast of the Pacific Northwest. In the layers of coast that have subsided and been filled again, there are bands of sand brought inland by tsunami waves, time and time again. Here is a article written by the American Museum of Natural History with more information on the Ghost Forests on the PNW.
Liquefaction
For liquefaction to occur, three things must happen. (1) Young, loose and grainy soil (2) needs to be saturated with water, and (3) experience strong ground shaking. The USGS, in coordination with California Geological Survey, give a summary of liquefaction and its effects here.
This video from the 2011 M9.0 Japan earthquake shows dramatic cracks in the ground, as well as liquefaction.
Our webpage on liquefaction includes video of the 2011 Christchurch earthquake, as well as links to liquefaction hazard maps for Washington and Oregon.
Landslides
Strong shaking can increase susceptibility to landslides.
This blog from the American Geophysical Union details some of the significant landslides in Paupa New Guniea that were triggered by a M7.5 earthquake on February 25th, 2011.
The Pacific Northwest is susceptible to landslides due to seasonal conditions, and strong shaking will increase landslide risk. Here are some resources from the states of Washington and Oregon.
Building Damage and Fires
The 1989 Loma Prieta earthquake caused widespread damage to infrastructure, such as the collapse of the Cypress Viaduct in Oakland, and fires in the San Francisco area.
Major cities in the Pacific Northwest would be just as susceptible to fire damage, and the construction of the Cypress Viaduct bears a striking resemblance to Seattle’s Alaskan Way Viaduct.
So, What Can I Do?
In order to prepare effectively, it is important to be aware of all earthquake-related hazards, such as the ones listed above. The following websites are great resources for earthquake hazards and preparedness in the Pacific Northwest. We encourage you to know your risks, be prepared, mitigate against hazards, respond safely, and enagae in holistic recovery planning.
- Cascadia Regional Earthquake Workgroup (CREW)
- Washington State Emergency Management Division
- State of Oregon Emergency Management
- Emergency Management British Columbia
- Become a Community Emergency Response Teams member
- Become a Red Cross Volunteer
- Check your local emergency management office for information specific to your community!
In the event of an earthquake, don't forget to drop, cover and hold!
Special Thanks To
Dr. Erin Wirth, Affiliate Professor, University of Washington
Lan T. Nguyen, Doctoral Student, Interdisciplinary Urban Design and Planning, University of Washington
On May 18th, 1980, at 8:32 AM, the landscape in Southwestern Washington was forever changed by an explosive eruption of Mount St. Helens. This was the most deadly volcanic event in US history.
Mount St. Helens is part of the Cascade Range, a chain of volcanoes from British Columbia to Northern California. The PNSN and the Cascades Volcano Observatory cooperatively operate 21 seismometers on or near Mount St. Helens, the most historically active volcano in the Cascade Range.
Seismogram of May 18th from one of our seismic stations.
Main PNSN Page on Mount St. Helens
Main CVO Page on Mount St. Helens
On the anniversary of the eruption of Mount St. Helens, earth science gets its day in the spotlight. Here's a collection of news stories about the eruption and what we have learned about volcanoes since 1980.
38 years later: What's changed since the Mount St. Helens Eruption?
- Q13 News
Features an interview with PNSN DIrector Emeritus Steve Malone
Mount St. Helens: Remembering the deadliest U.S. eruption 38 years later
- USA Today, King 5 News
Summary of events on May 18th, 1980 with photo galleries
Lessons Learned: Mount St. Helens to Kilauea
- King 5 News
Interview filmed in the PNSN Seismology Lab with Director Emeritus Steve Malone
Remembering Mt. St. Helens as Cascade event looms
- KOIN 6
Event overview, brief discussion of Cascade hazards, and dispels concern of a Kilauea-triggered event in the Cascades.
Photos: The Mt. St. Helens eruption of 1980
- KOIN 6
Includes an interview with CVO Scientist-In-Charge Seth Moran
Scientists Reflect on the Catastrophic 1980 Mount St. Helens Eruption
- Ashley Williams, AccuWeather
Details about the activity that led to the eruption, how life in the MSH area fared, and another Steve Malone feature.
How Dangerous are the Northwest's Volcanoes?
-KUOW, Oregon Public Broadcasting
Interview with CVO Scientist-In-Charge Seth Moran, discusses Oregon volcano hazards
Background: The Challenges of Predicting Seismic Hazard
We all want to know when the next big earthquake will happen. But because we can’t predict earthquakes, the next best thing we can do is figure out how strong shaking could be during future quakes. To do this, earth scientists can reconstruct an area’s historic earthquake record using geologic markers and any available written accounts, which allows them to constrain recurrence and magnitude ranges for quakes on local faults. Scientists then combine these estimates with what we know about an area’s surface geology to predict how often and how strong the ground will shake during an earthquake. This sort of analysis is the basis for the US’ National Seismic Hazard Mapping Project, and it is critical for making sure we build structures that can withstand earthquake shaking.
Predicting how an earthquake will shake an area is not a trivial task, however. Besides the challenges of figuring out the historical earthquake record from sparse geologic markers, understanding how conditions at a given location will affect earthquake shaking is very difficult. In Seattle, for example, there are many interesting geologic features that change and amplify ground motion during earthquakes; large areas of water-logged, artificial fill under Pioneer Square and SODO have amplified ground motion and even liquefied during historic earthquakes; the deep Seattle Basin, which contains thick layers of relatively soft sediments, traps seismic waves and “sloshes” like a big bowl of jelly in some earthquakes. Taking these effects into consideration during seismic hazard analysis is important, but requires a detailed understanding of the local geology and how it affects seismic waves.
The hills are alive with the sound of … earthquakes?
A geographic component that is not often considered during seismic hazard analysis is topography. As seismic waves from an earthquake propagate through the earth, they interact with subsurface geology, like faults and basins; however, the waves also interact with surface features, like hills, cliffs, and valleys. Depending on the wavelength of the seismic waves and the direction they’re traveling, the shaking felt on these features can be much stronger (or weaker!) than in surrounding flat areas. This amplification happens because these features scatter and trap seismic energy; they can also experience a sort of structural “resonance”, similar to the way buildings and bridges can. Studies of ground motion from the 2009 L’Aquila and 2010 Haiti earthquakes [1, 2] have shown that topography significantly amplified ground motion at certain locations, sometimes shaking twice as strong as surrounding areas!
With these findings in mind, we are interested in seeing how topography might affect earthquake shaking here in Seattle. While the city doesn’t have any particularly tall or prominent ridges, it does have many steep bluffs and cliffs overlooking Elliott Bay and Puget Sound. We know from past earthquakes that these features are prone to land-sliding, and we would like to know how topographic amplification might contribute to that hazard, as well as to the general safety of structures built on the bluffs.
How will the 70-90m tall bluffs around the city behave in an earthquake?
How will things shake out in West Seattle?
So, as part of a graduate-student-led research project, we are looking for volunteers in West Seattle to host seismometers for a small experiment. In this project, portable seismometers will be placed along and down the bluffs facing Puget Sound. Over the course of a few hours, these seismometers will record the ambient vibrations caused by ocean waves, weather systems, and even car traffic. Together, these sources create what’s known as ambient seismic noise. By comparing recordings of this noise from points along the bluff, we can understand how the topography amplifies ground shaking during earthquakes.
For our experiment, we are looking for hosts in West Seattle living near the bluffs along Alki. Ideally, hosts would live on or between Sunset Ave. SW and Alki Ave. SW. We also need a few sites in the neighborhood away from the sea bluff (see the map below). The experiment itself will last just 1 day, and only requires access to your yard, where we will plop-down a coffee-can sized seismometer for a few hours. The experiment will take place some time between early October and late November.
So, if you are a citizen scientist who would like to help us better understand how earthquakes rattle the hills around our city, and you live on or near the Sound-facing bluffs in West Seattle, please consider filling out the form linked below!
If you have any questions about the experiment, you can contact the project lead Ian Stone at ipstone@uw.edu
Link to Sign-up Form: https://forms.gle/vfSBgpZUapbKs5eo7
References
[1] Massa, M., Lovati, S., D’Alema, E., Ferretti, G., and M. Bakavoli, 2010. An experimental approach for estimating seismic amplification effects at the top of a ridge, and the implication for ground-motion predictions: The case of Narni, Central Italy. Bul. Seis. Soc. Amer., 100 (6) 3020-3034.
[2] Hough, S. E., Altidor, J. R., Anglade, D., Given, D., Janvier, M. G., Maharrey, J. Z., Meremonte, M., Mildor, B. S. L., Prepetit, C., and A. Yong, 2010. Localized damage caused by topographic amplification during the 2010 M 7.0 Haiti earthquake. Nature Geoscience, 3. 778-782.