A massive fault sits just off the Pacific Northwest—and in some parts of Oregon, a tsunami could arrive only 10 to 20 minutes after the shaking begins.
In this DiscoverTube investigation, we examine the Cascadia Subduction Zone, Oregon and Washington tsunami risk, ghost forests from the 1700 earthquake, tsunami inundation maps, vertical evacuation towers, ShakeAlert and the new seafloor sensors being installed offshore.
Scientists now know Cascadia can produce magnitude 8 to 9 earthquakes, but the biggest challenge may be time: in some coastal communities, the earthquake itself could consume most of the available evacuation window.
Oregon says it needs far more vertical evacuation structures, while researchers are expanding offshore monitoring to detect future ruptures faster.
Subscribe to DiscoverTube for more earthquake science, tsunami research, hidden geology and discoveries across America.
The Cascadia Subduction Zone is a 700-mile fault capable of producing magnitude 8 or 9 earthquakes and devastating tsunamis. With waves arriving within 20 minutes of shaking, coastal communities face extreme risks. The region relies on a mix of vertical evacuation structures and early warning systems, though significant funding and infrastructure gaps remain between Washington and Oregon.
Stand on the beach at Cannon Beach, or Long Beach, or Ocean Shores, and look west.
Somewhere, out past the surf line, roughly 50 to 80 miles offshore, the seafloor is being shoved under North America.
That boundary runs about 700 miles, from Cape Mendocino in Northern California, all the way up past Vancouver Island.
It is the largest fault in the lower 48 states, and for most of modern history, nobody believed it could do anything at all.
Stories like this are what we dig into on Discover Tube, where our documentaries go past the headline and keep you updated as new developments change the picture.
And the reason this fault matters isn't just its size.
It sits close enough to shore that state emergency managers now describe tsunami waves reaching parts of the Oregon coast in as little as 10 to 20 minutes after the shaking starts.
That is not a forecasting problem. That is a distance problem. The fault nobody wrote down.
That 10-minute figure comes directly from the geography, so the geography has to come first.
What sits offshore is called the Cascadia Subduction Zone.
NOAA describes it as extending roughly 700 miles, or about 1,100 kilometers, from Cape Mendocino to Northern Vancouver Island.
You'll sometimes hear 600 miles instead, and that's not a contradiction.
That shorter number simply stops at the Canadian border and leaves out the northern segment.
The mechanics are straightforward, even if the scale isn't.
Oceanic crust is being driven underneath the North American plate.
And it isn't one plate doing the pushing. It's three.
The Juan de Fuca Plate in the middle, the Explorer Plate to the north, and the Gorda Plate down near California.
Most coverage flattens that into a single slab.
But the segmentation matters because different pieces of the boundary can behave differently.
Where those plates meet the continent, the two surfaces lock together.
They don't slide smoothly. Friction holds them, strain builds in the rock, and eventually the lock fails.
When it fails across a boundary this long, you get what geologists call a Great Earthquake, magnitude 8 or 9.
NOAA now places Cascadia in the same category as the faults that ruptured off Indonesia and off Japan.
Here's the part that makes this story genuinely strange.
Before the mid-1980s, that wasn't the scientific consensus at all.
The prevailing view was that the Pacific Northwest faced tsunamis only from distant sources,
the kind that cross an ocean and give you hours of warning.
Locally generated tsunamis were considered a minor concern.
There were two reasons for that.
First, Cascadia was thought incapable of producing Great Earthquakes in the first place.
Second, and more simply, there was no written record of one ever happening.
That second reason is the trap.
European record-keeping on this coastline only goes back a couple of centuries.
A fault that ruptures every few hundred years can easily sit quiet through that entire window and look harmless.
Silence in the archive was read as safety in the rock.
The fault was eventually proven to have ruptured, and we'll get to exactly how.
But the more urgent problem isn't the size of what it can do.
It's how little time it leaves.
Start the clock at the moment the ground begins moving.
Oregon's emergency management officials estimate the shaking itself along the coastline
would last five to seven minutes, weakening as you move inland.
That's an unusual number for people who've only felt ordinary earthquakes, which are over in seconds.
Five minutes of continuous ground motion is long enough to bring down unreinforced buildings,
break water mains, and drop bridges.
Now layer the water on top of that.
Oregon's own state guidance states that a Cascadia rupture could send waves onto the coast
in as little as 10 to 20 minutes after shaking begins.
Bridget Good, a mitigation analyst with the Oregon Department of Emergency Management, has put it more bluntly.
Some communities could see the first waves within 10 minutes.
Do the subtraction.
If the shaking runs five to seven minutes and the water arrives at 10,
the shaking has consumed most of the escape window before anyone has taken a step.
Arrival times aren't uniform, and that's a geography story, too.
Washington's Department of Natural Resources models some locations at 30 to 35 minutes
and others at 10 to 15.
Low-lying, exposed communities like Long Beach and Westport sit at the short end,
waves potentially reaching them within about 15 minutes of the earthquake starting.
This is why the usual mental model of tsunami warning breaks down here.
When a tsunami crosses the Pacific from Alaska or Japan, buoys detect it,
centers issue alerts, and coastal towns get hours.
Cascadia doesn't work that way.
The rupture is offshore, but it's close.
The wave doesn't have to travel far,
and nothing in the warning chain can outrun a distance that short.
You can't be told about something you are already standing in.
The strain driving all of this accumulates at a rate that sounds almost gentle.
The plates converge at roughly 1 1⁄4 to 1 1⁄2 inches per year,
about the pace your fingernails grow.
That slow, relentless loading is exactly what makes the release so violent.
Centuries of quiet accumulation, spent in minutes.
So the practical question stops being how to warn people
and becomes what they can physically reach in the time available.
But before we get to that, there's an obvious hole in the story.
If nobody wrote any of this down, if there was no record at all,
how do we know the fault has ever ruptured, and how do we know the date?
The ghost forests and the orphan wave.
The answer is standing upright in the tidal marshes
along the Washington and Oregon coast, dead but still rooted.
Walk the banks of the Copalis River in southwest Washington,
and you'll find the stumps of western red cedars sitting in salt marsh,
trees that once grew on dry, forested ground and are now flooded by every tide.
These are called ghost forests.
A forest doesn't drown by accident.
The ground it was growing on dropped.
That drop is the fingerprint of the fault.
When a locked subduction zone finally slips,
the coastline above it can subside abruptly, in some places by a meter or more.
Salt water then pushes into soil that had never held it,
and the trees die where they stand.
The dead wood preserved something better than a memory.
Because those cedars kept growing right up until the moment the land sank,
the outermost growth rings record the last full season they were alive.
Tree ring dating narrowed the death of those forests
to between the growing seasons of a single winter, the winter of 1700.
The matching evidence came from across the ocean.
Japanese officials kept careful written records of coastal flooding,
and their archives describe a tsunami that arrived without any felt earthquake beforehand.
Historians called it an orphan tsunami because it had no local parent.
Cascadia was the parent.
The wave crossed the Pacific and reported the rupture
to the only people on Earth who were writing things down.
So the fault does rupture, and it can drop the coast when it does.
That same subsidence matters enormously for what happens next time,
because lowering the land before the water arrives
changes how far inland the water gets to travel.
What the water actually does.
Take Siletz Bay on the central Oregon coast just south of Lincoln City,
a shallow, low-lying estuary with homes and roads ringing the water.
Modeling of a Cascadia rupture there
projects open coast run-up of roughly 30 to 35 feet in a likely scenario.
In a worst-case version of the same event,
that figure rises to somewhere around 50 to 55 feet.
That contrast is worth pausing on,
because it's the honest version of a number you've probably heard.
Oregon state guidance describes waves ranging from several tens of feet
up to more than a hundred feet.
Both statements can be true at once.
The hundred-foot figure sits at the extreme end of the modeled range,
usually in narrow inlets and channels where water gets funneled and stacked.
The 30 to 50 foot band is the realistic working expectation for open coastline.
These are scenarios, not predictions.
There's also a measurement people routinely mix up.
Run-up describes how high the water climbs.
Inundation depth describes how deep the water stands once it's there.
And in flat coastal towns, that number can be brutal.
Corina Allen, Chief Hazards Geologist with Washington's Department of Natural Resources,
has stated that in low-lying communities like Long Beach and ocean shores,
inundation can reach up to 60 feet in some locations.
That's standing water deeper than a five- or six-story building is tall.
Now, add distance.
Stephen Richardson, Oregon's State Hazard Mitigation Officer,
has pointed out that the inundation modeling shows depth extending a mile or more inland.
Those are not steep places.
Long Beach and ocean shores sit on narrow sand peninsulas with almost no natural elevation.
Water crossing a mile of flat sand doesn't slow down in any way that helps a person running.
And it doesn't happen once.
Washington's geologists stress that tsunamis are multi-wave events
affecting the coast for many hours and potentially days after the earthquake.
The first wave is frequently not the largest.
Anyone who returns to check a house between waves is walking back into the hazard.
The mapping also broke an assumption that it held for decades.
Washington's inundation modeling showed that a large Cascadia earthquake
sends water into Puget Sound and Hood Canal,
with some low-lying areas ending up under roughly 14 feet.
A decade before those maps existed,
most people believed the sound was sheltered from Cascadia entirely.
It isn't.
Protected inland water still moves when the ocean floor does.
If maps like these are the kind of thing you want explained properly,
subscribe to Discover Tube.
So the water reaches a mile inland within 10 to 20 minutes
in communities built on flat ground.
That turns the question from prediction into logistics.
Where exactly is a person supposed to go?
One tower, 20 needed.
Oregon's answer currently stands on the bayfront in Newport.
At Oregon State University's Hatfield Marine Science Center,
there's a building engineered so that people can climb it
and survive a tsunami rather than trying to outrun one.
It is the only verified structure of its kind in the entire state,
with two more in development.
In August, the Oregon Department of Emergency Management,
working with the state's Department of Geology and Mineral Industries,
published Oregon's first statewide guidance
on tsunami vertical evacuation structures.
It was developed with support from NOAA's
National Tsunami Hazard Mitigation Program and a FEMA grant program.
The department's director, Aaron McMahon,
framed the stakes plainly, saying Oregon's coastal communities
face some of the most urgent tsunami risks in the country.
The logic behind vertical evacuation is simple.
Horizontal evacuation only works if high ground is inside a 10-minute walk.
Where it isn't, the only remaining option is to go up,
into a tower, a berm, a reinforced parking garage,
or a building designed to take the load of moving water.
Then came the number.
In September, reporting from Oregon Public Broadcasting
and Jefferson Public Radio put the shortfall in concrete terms.
The state says it needs at least 20 of these structures along the coast.
It has one finished.
The same coverage carried the state's worst-case analysis,
which estimates as many as 40,000 deaths along the Oregon coast
in the most severe scenario.
That is the upper bound, not the expected outcome.
And other state assessments, using different scopes, land far lower.
But it explains the urgency behind a document that is, on paper, a citing manual.
Oregon is publishing guidance.
The state to the north already poured the concrete
and can say exactly what it cost.
Auntie Lee's Tower and Project Safe Haven.
The Shoalwater Bay Indian Tribe lives at Toakland, Washington,
on a low spit of land reaching into Willapa Bay.
High ground is not a realistic walk from there.
So the tribe built the first freestanding tsunami evacuation tower in the United States
and named it after the elder who made it happen.
That elder was Lee Shipman, who led the tribe's effort to secure funding.
The structure is known locally as Auntie Lee Tower,
and building it took a federal award of roughly $2.2 million.
The first time FEMA construction money in Washington
went toward a vertical evacuation structure.
It wasn't the state's first refuge, though.
That distinction belongs to Acosta Elementary School near Westport,
which in 2016 opened with a reinforced rooftop platform
engineered to hold a community above the water.
It was the first tsunami vertical evacuation structure in North America,
and it settled an argument that had been theoretical until then.
A school gymnasium can be a life safety refuge
if you design the columns and foundations for the load of moving water and debris.
Washington got there because it did the siting work first,
under a program called Project Safe Haven,
and in a later state assessment covering Pacific, Grays Harbor, and Clallam counties.
Planners modeled walk times under a magnitude 9 Cascadia scenario.
They asked a narrow, answerable question.
From any given doorway, can a person reach survivable elevation before the water arrives?
Wherever the answer was no, that became a candidate site.
The scale of that answer is sobering.
Washington's Emergency Management Division
counts more than 175,000 people inside its tsunami inundation zone,
and estimates the state needs at least 58 structures,
possibly more than 80, for everyone to have somewhere to go.
Karina Allen, the state's chief hazards geologist, has been direct about it,
saying vertical evacuation structures are really the only hope
for saving the number of people living on these peninsulas.
Do you live inside a coastal evacuation zone,
or within walking distance of a designated refuge?
Tell us where, because the geography of that answer
is exactly what we try to map out on Discover Tube.
Oregon's new guidance is honest about that comparison.
The report states that Oregon has not completed a statewide planning effort
comparable to Project Safe Haven,
which means individual coastal communities may have to run their own evaluations
to work out whether a refuge makes sense and where it should stand.
What the guidance does provide is a menu.
It identifies multi-story parking garages, community buildings, schools,
retrofitted existing structures, stand-alone towers,
and engineered earth berms as workable options,
with the right choice depending on local geography, population density,
evacuation routes, available land, and money.
That list is worth noticing because none of it is exotic.
A berm is a hill you build.
A parking garage is a parking garage, with stronger columns and a deeper foundation.
The engineering problem was solved at Acosta and Toakland.
The remaining problem is paying for it.
Steven Richardson, Oregon's State Hazard Mitigation Officer,
has said the obstacles are compounded by uncertainty
around federal emergency management funding,
along with the question of whether local governments commit to seeing projects through.
Cost is the biggest barrier for small coastal towns,
where one structure can exceed anything in the municipal budget.
So the gap between one tower and 20 is not a scientific gap.
It's a procurement schedule, and a refuge only works if people start climbing immediately,
which means recognizing within seconds that the shaking is a Cascadia rupture
rather than an ordinary earthquake.
That recognition depends on instruments,
and the fault is offshore, where instruments barely exist.
Listening to the seafloor.
The system meant to provide that recognition is called ShakeAlert.
It's the only public earthquake early warning system in the country,
operated by the USGS, covering more than 50 million residents
and visitors across California, Oregon, and Washington.
Seismologists describe it as stable and mature on the West Coast,
which is why the same technology is now being extended inland into Nevada.
But maturity on land does not solve Cascadia, and the USGS says so plainly.
The agency acknowledges that there will almost always be a zone near the epicenter
where the alert arrives after damaging shaking has already started.
For a fault that ruptures offshore and shakes the coastline first,
that zone is the coastline itself.
The deeper limitation is sensor placement.
A Cascadia megathrust rupture unzips mostly beneath the seafloor,
yet ShakeAlert draws on no seafloor sensors at all.
It infers an offshore event from onshore instruments,
which costs seconds and blurs the estimate of how big the rupture is and how far it has spread.
That's the gap the Cascadia Offshore Subduction Zone Observatory is trying to narrow.
The project, run by the University of Washington and Scripps,
with National Science Foundation funding,
is adding geophysical instruments to the Ocean Observatory's initiative's regional cabled array off Oregon,
fiber-optic cable already lying on the seafloor, feeding continuous data back to shore.
Consider the current density.
There are just two cabled ocean-bottom seismometers offshore Oregon,
along a fault roughly 700 miles long.
Four more are being installed through this year,
which is meaningful progress and still a very thin line of instruments.
Modeling suggests what that buys.
Incorporating the existing ocean-bottom seismometers
could improve detection time for nearby ruptures by about 5 to 9 seconds,
and using the fiber-optic cable itself as a sensing array could add up to roughly 10.
The project team is also candid that the footprint of both the Oregon array
and Ocean Networks Canada's Neptune cable off Vancouver Island
remains small enough to limit their present value for early warning.
Tsunami detection is harder still.
In a modeled magnitude 9 scenario,
the offshore dark buoys register the wave in about 15 minutes,
roughly when it's already reaching the beach.
So the hazard nobody could see has been located, dated, mapped and costed.
What remains unresolved is construction.
Watch three things.
Whether Oregon funds a statewide siting effort like Washington's.
Whether the additional seafloor instruments come online
and feed ShakeAlert.
And whether federal funding holds steady enough for small towns to build.
Subscribe for more and tell us in the comments
of everything Discover Tube covered here, which should be funded first.
Towers, seafloor sensors or siting studies.