Cascadia Subduction Zone: The Pacific Northwest's Hidden Threat

For most of the 20th century, the Pacific Northwest was thought to be seismically quiet — a rare gap in the Ring of Fire where large earthquakes simply did not happen. That understanding was catastrophically wrong. Beginning in the 1980s, a series of geological discoveries revealed that the region sits atop one of the most dangerous fault zones on Earth: the Cascadia Subduction Zone.

The Cascadia Subduction Zone (CSZ) is a 1,000-kilometer megathrust fault running from Cape Mendocino in northern California to Vancouver Island in British Columbia. Along this boundary, the small Juan de Fuca Plate and its associated fragments (the Gorda and Explorer plates) are being forced beneath the much larger North American Plate at a rate of approximately 30–40 mm per year. The fault is currently locked, accumulating strain energy that will eventually be released in a major earthquake — possibly a magnitude 9.0 event comparable to the 2011 Tōhoku earthquake in Japan.

The scientific community now considers a great Cascadia earthquake to be not a question of "if" but "when." The paleoseismic record is unambiguous: full-margin ruptures have occurred repeatedly throughout the Holocene, and the current interval since the last great earthquake in 1700 already exceeds the average recurrence time. Understanding the Cascadia Subduction Zone — its history, mechanics, and potential impacts — is essential for the millions of people living in its shadow.

Geography and Geology

The Fault Zone

The Cascadia Subduction Zone extends roughly 1,000 km (620 miles) along the coast of the Pacific Northwest, from the Mendocino Triple Junction near Cape Mendocino, California, northward through Oregon and Washington to the Nootka Fault zone near northern Vancouver Island. The actual plate boundary — the megathrust fault surface — lies approximately 80–100 km (50–60 miles) offshore beneath the continental shelf and slope.

The CSZ is where the Juan de Fuca Plate, a relatively small remnant of the once-vast Farallon Plate, subducts beneath the North American Plate. The Juan de Fuca Plate is generated at the Juan de Fuca Ridge, a spreading center located roughly 400–500 km offshore. As the plate moves eastward, it cools, becomes denser, and eventually plunges beneath the continent at the subduction zone. The rate of convergence varies along the fault: approximately 30 mm/year in the north (Vancouver Island) to roughly 40 mm/year in the south (northern California), according to GPS measurements compiled by the Pacific Northwest Geodetic Array.

Two smaller plate fragments are associated with the Juan de Fuca system: the Gorda Plate to the south (off northern California) and the Explorer Plate to the north (off northern Vancouver Island). Both are geologically young and relatively thin, which affects the earthquake behavior of their respective fault segments.

The Locked Zone

The most dangerous portion of the CSZ is the "locked zone" — the shallow section of the megathrust where the two plates are stuck together by friction. The locked zone extends from the seafloor at the deformation front (where the Juan de Fuca Plate first contacts the North American Plate) to a depth of roughly 15–25 km inland. Below the locked zone, the transition zone (or partially locked zone) extends to depths of about 30–40 km, where the plates begin to slide past each other more freely.

While the plates are locked, the overriding North American Plate is being compressed and dragged downward at the coast. GPS measurements show that the Oregon and Washington coastlines are currently moving eastward (landward) and subsiding at rates of several millimeters per year — accumulating the strain that will be released in the next earthquake. When the fault ruptures, the coast will suddenly rebound westward and uplift, while areas further inland may subside.

[MAP: Cascadia Subduction Zone showing the plate boundary, locked and transition zones, major cities at risk (Vancouver, Seattle, Portland, Eugene, Salem, Olympia), and modeled tsunami inundation zones] Data source: USGS, Pacific Northwest Seismic Network, Oregon Department of Geology and Mineral Industries Features: Juan de Fuca Ridge, Juan de Fuca/Gorda/Explorer plates, deformation front, locked zone (dark red), transition zone (orange), Cascade volcanic arc, major population centers with population estimates, coastal tsunami evacuation zones

Discovery: Uncovering a Hidden Catastrophe

Brian Atwater and the Ghost Forests (1987)

The modern understanding of the Cascadia Subduction Zone began with the work of USGS geologist Brian Atwater. In 1987, Atwater was investigating tidal marshes along the coast of Washington State when he discovered something startling: buried beneath the surface were layers of marsh soil that had abruptly dropped below sea level, killing forests of western red cedar. The dead trees — "ghost forests" — were preserved in tidal flats, their roots entombed in mud.

Atwater recognized that these buried soils recorded sudden coastal subsidence — the kind of ground dropping that occurs when a subduction zone megathrust fault ruptures and the overriding plate snaps back. He found multiple buried soil layers stacked on top of each other, indicating that this had happened repeatedly. Each subsidence event was overlain by a layer of sand deposited by a tsunami that followed the earthquake. The implications were enormous: the Pacific Northwest had experienced great subduction zone earthquakes, and would again.

The Orphan Tsunami of 1700

The breakthrough in dating the most recent Cascadia earthquake came from Japan. Japanese historical records documented a mysterious tsunami that struck the Pacific coast of Honshu on January 27–28, 1700 (local time). The tsunami was notable because no locally felt earthquake preceded it — making it an "orphan tsunami" that must have originated from a distant source.

In a landmark study published in 2003, researchers Kenji Satake, Brian Atwater, and colleagues worked backward from the Japanese tsunami records, accounting for Pacific Ocean travel times, to determine that the source was a great earthquake on the Cascadia Subduction Zone on the evening of January 26, 1700 (Pacific time). Radiocarbon dating of the ghost forests and buried soils in Washington confirmed this date. The earthquake was estimated at M~9.0 based on the size of the tsunami in Japan and the extent of coastal subsidence in the Pacific Northwest.

Native oral histories from the Pacific Northwest also preserve accounts consistent with a great earthquake and tsunami around this time. The Yurok, Huu-ay-aht, and other coastal peoples maintained stories of the ground shaking violently and the sea withdrawing and returning with devastating force.

The Turbidite Record

Further evidence for past Cascadia earthquakes came from deep-sea sediment cores. Oregon State University geologist Chris Goldfinger and colleagues collected cores from submarine channels along the Cascadia margin, identifying turbidites — layers of sediment deposited by underwater landslides triggered by earthquake shaking. By dating these turbidite layers using radiocarbon methods, they constructed a 10,000-year record of earthquake activity on the CSZ.

The turbidite record, published in a series of papers beginning in 2003, revealed at least 41 major earthquakes in the past 10,000 years. Nineteen of these were full-margin ruptures (affecting the entire 1,000 km fault), while the remainder were partial ruptures of the southern portion. This record provides the most detailed chronology of earthquake recurrence on any subduction zone worldwide.

Earthquake Recurrence: How Often Does Cascadia Rupture?

The paleoseismic record shows that Cascadia earthquakes do not occur at regular intervals. Instead, recurrence times vary, and the fault produces two distinct types of earthquakes:

Full-margin ruptures (M~9.0): These involve the entire 1,000 km fault and produce earthquakes estimated at M8.7–9.2. Based on the turbidite record, full-margin ruptures have occurred roughly every 200–600 years, with an average recurrence interval of approximately 240 years for the full margin. The most recent was in 1700 — more than 325 years ago.

Partial-margin ruptures (M~8.0–8.5): These involve only the southern segment of the fault (roughly southern Oregon and northern California) and occur more frequently. The southern CSZ has an average recurrence interval of roughly 220–240 years for events of any type.

Table based on Goldfinger et al. (2012) turbidite chronology. Dates are approximate with uncertainties of ±50–100 years.

[CHART: Timeline — Past Cascadia Subduction Zone Earthquakes (10,000 years to present)] Data: Vertical lines representing each of the 41 documented turbidite events over 10,000 years, with full-margin events (~19) shown in red and partial-margin events (~22) shown in orange. The 1700 event and the current 325+ year gap since the last event highlighted. Average recurrence interval of ~240 years marked. Source: Goldfinger et al., 2012, Oregon State University turbidite chronology

What a Future M9.0 Cascadia Earthquake Would Look Like

Seismologists, engineers, and emergency planners have developed detailed scenarios for a future full-margin Cascadia earthquake. The picture is sobering.

The Shaking

A M9.0 Cascadia earthquake would produce shaking lasting approximately 3–5 minutes — dramatically longer than the roughly 45 seconds of strong shaking expected from a major San Andreas Fault earthquake. This prolonged duration reflects the enormous length of the fault: the rupture would propagate along 1,000 km of the megathrust over the course of several minutes.

The shaking would be felt across a vast area, from northern California to British Columbia and from the coast well into the interior. Coastal areas closest to the fault would experience the most intense shaking (Modified Mercalli Intensity VIII–IX), but even cities 200+ km from the fault — Seattle, Portland, Vancouver — would experience strong, prolonged ground motion sufficient to damage buildings and infrastructure.

Soil liquefaction would be widespread in river valleys and filled land, particularly in Portland (Willamette River valley), Seattle (filled tidelands along the waterfront), and the Fraser River delta near Vancouver. Landslides would block roads and railways throughout the region's mountainous terrain.

The Tsunami

A full-margin Cascadia rupture would generate a major tsunami. Modeling by NOAA, the Oregon Department of Geology and Mineral Industries (DOGAMI), and university researchers indicates wave heights of 6–12 meters (20–40 feet) along much of the Oregon and Washington coast, with locally amplified heights exceeding 15 meters in some bays and inlets.

The first tsunami wave would reach the nearest coastline in approximately 15–30 minutes — possibly less for communities closest to the fault. This leaves very little time for evacuation. Many small coastal communities in Oregon and Washington are located at sea level with limited high ground nearby and roads that could be blocked by earthquake damage.

For more on how earthquake-generated tsunamis work, see Earthquakes and Tsunamis.

The Aftermath

FEMA's Cascadia Rising exercise in June 2016 — the most comprehensive federal emergency exercise ever conducted for an earthquake scenario — modeled the impacts of a M9.0 Cascadia event. The exercise scenario projected:

• More than 13,000 fatalities
• Approximately 27,000 injuries requiring hospitalization
• Nearly 1 million displaced people
• More than 1 million buildings damaged or destroyed
• Collapse of critical transportation links, including bridges over the Willamette and Columbia rivers
• Disruption of water, power, and sewer systems for weeks to months
• Economic losses potentially exceeding $30 billion

Recovery was estimated to take years to decades, with some infrastructure requiring complete replacement rather than repair.

Cascadia Compared: How Does It Stack Up?

The Cascadia Subduction Zone shares fundamental characteristics with other great subduction zones that have produced devastating earthquakes in recent decades. The comparison provides a lens for understanding what the Pacific Northwest could face.

The comparison with the 2011 Tōhoku earthquake is particularly instructive. Japan, one of the world's best-prepared nations for earthquakes and tsunamis, suffered nearly 20,000 deaths despite decades of investment in seawalls, early warning systems, and building codes. The Cascadia region, while increasingly aware of its risk, has significantly less tsunami-specific infrastructure than Japan had in 2011.

Slow-Slip Events and Episodic Tremor and Slip (ETS)

One of the most important discoveries about the Cascadia Subduction Zone in recent decades has been the detection of slow-slip events — also called episodic tremor and slip (ETS).

Beginning with observations published by Herb Dragert of the Geological Survey of Canada in 2001 and refined by subsequent studies, scientists discovered that the deep portion of the CSZ megathrust (beneath and downdip of the locked zone, at depths of roughly 25–45 km) periodically "slips" over the course of several weeks, releasing accumulated strain without producing a conventional earthquake. These events are accompanied by a distinctive type of seismic signal called "tectonic tremor" — a low-frequency, continuous vibration that differs from the sharp signals of ordinary earthquakes.

Cascadia ETS events recur with remarkable regularity, approximately every 14 months along the northern segment (beneath Puget Sound and southern Vancouver Island) and at somewhat different intervals along other segments. Each ETS event involves slip equivalent to a M~6.5–6.8 earthquake, but released so slowly that it produces no damaging ground motion.

The significance of ETS for earthquake hazard is actively debated. Some researchers hypothesize that slow-slip events transfer stress to the locked zone above, potentially bringing it incrementally closer to failure. Others suggest that ETS acts as a "pressure relief valve," partially relieving accumulated strain. The question of whether ETS events could trigger — or be triggered by — a megathrust earthquake remains one of the most important open questions in Cascadia seismology.

The Cascadia Subduction Zone was one of the first places where ETS was discovered, and it remains one of the best-studied natural laboratories for this phenomenon.

Current Monitoring

The CSZ is monitored by a network of agencies and instruments that has expanded significantly since the 1990s.

Pacific Northwest Seismic Network (PNSN): Operated by the University of Washington and the University of Oregon, the PNSN maintains seismograph stations throughout Washington, Oregon, and portions of neighboring states. It provides real-time earthquake detection, location, and magnitude estimation for the region. Pacific Northwest Seismic Network

Geodetic Networks: Continuous GPS stations throughout the Pacific Northwest measure the slow eastward compression of the coast — the direct evidence that the fault is locked and accumulating strain. The EarthScope Plate Boundary Observatory (now part of SAGE/GAGE facilities) deployed hundreds of GPS stations and strainmeters across the Cascadia region.

Seafloor Instruments: Because the locked zone is offshore, land-based instruments provide only an indirect view. Efforts to place pressure sensors, seismometers, and GPS-acoustic instruments on the seafloor above the locked zone are ongoing, through projects like the Ocean Observatories Initiative (OOI) cabled array off the Oregon coast.

ShakeAlert: The USGS ShakeAlert Earthquake Early Warning system, which became publicly operational in stages between 2018 and 2021 for the West Coast, can provide seconds to tens of seconds of warning before strong shaking arrives from a Cascadia earthquake. However, because the fault is offshore, the warning time for coastal communities closest to the epicenter may be very limited.

Preparedness and Mitigation

Recognition of the Cascadia threat has spurred significant — but still insufficient — preparedness efforts across the Pacific Northwest.

Building Codes and Retrofitting

Oregon and Washington have updated building codes to account for subduction zone earthquakes, but many older structures — including unreinforced masonry buildings, pre-1970s concrete buildings, and older bridges — remain highly vulnerable. Oregon's inventory of unreinforced masonry buildings alone includes thousands of structures, many in downtown Portland. The Oregon Resilience Plan (2013) estimated that bringing the state's critical infrastructure to a resilient standard would cost billions of dollars and take decades.

Tsunami Preparedness

Oregon and Washington have invested in tsunami hazard mapping, evacuation route signage, and public education. Blue "Tsunami Evacuation Route" signs are now visible throughout coastal communities. Oregon has led the development of tsunami "beat the wave" evacuation models and has considered vertical evacuation structures — reinforced towers or berms that allow people to reach safety above tsunami inundation levels when no natural high ground is within reach.

The town of Westport, Washington, built one of the first purpose-designed tsunami vertical evacuation structures in the United States, completed in 2016. The Ocosta Elementary School in Westport features a rooftop tsunami refuge capable of holding students, staff, and community members above projected wave heights. See Tsunami Safety Guide for detailed preparedness guidance.

Individual and Community Preparedness

Emergency management agencies across the region consistently emphasize:

• Keeping emergency supplies (food, water, medications, first aid) for at least two weeks, as infrastructure damage may prevent resupply for extended periods
• Knowing tsunami evacuation routes in coastal areas
• Securing heavy furniture and water heaters to reduce injury risk during shaking
• Understanding that the earthquake itself is the tsunami warning — if you feel strong, prolonged shaking on the coast, move immediately to high ground without waiting for an official alert
• Participating in annual earthquake drills such as the Great ShakeOut

For region-specific earthquake information, see Pacific Northwest Earthquakes and Seattle Earthquake Risk.

The Science Continues

Research on the Cascadia Subduction Zone is advancing rapidly. Key areas of active investigation include:

Fault geometry and segmentation: High-resolution seismic imaging is refining understanding of the megathrust's three-dimensional shape, the extent of the locked zone, and potential barriers to rupture propagation that could determine whether the next earthquake ruptures the full margin or only a segment.

Tsunami modeling: Increasingly detailed computer simulations, incorporating high-resolution coastal topography and realistic earthquake source models, are improving tsunami inundation maps that guide evacuation planning. Oregon Department of Geology and Mineral Industries (DOGAMI) produces some of the most detailed tsunami inundation maps in the world.

Paleoseismology: Ongoing studies of coastal subsidence records, offshore turbidites, and lake sediments continue to refine the earthquake chronology and identify possible patterns in recurrence behavior.

Slow-slip dynamics: Understanding the relationship between ETS events and megathrust earthquake potential remains a frontier research area, with implications for whether any precursory signals might be detectable before a great earthquake.

Offshore monitoring: Expanding instrumentation on the seafloor is expected to dramatically improve detection capabilities for both slow-slip events and any precursory activity on the locked megathrust.