The Big One's Secret: Two Faults May Strike Back to Back
🕐 9 min read | 🌍 Natural Wonders
🔒 Key Takeaways
- High-precision tree-ring dating published in Science Advances in 2024 shows the Seattle Fault and the Saddle Mountain fault zone both ruptured in 923–924 CE, in the same growing season.
- Combined, the two ruptures would have released moment equivalent to a single magnitude 7.8 earthquake — larger than either fault is modelled to produce on its own.
- The Seattle Fault runs beneath downtown Seattle and Bellevue, inside a metropolitan area that held about 4.0 million people at the 2020 census.
- Offshore turbidite cores record roughly 19–20 full-margin Cascadia megathrust ruptures in 10,000 years, the most recent on 26 January 1700.
- Modern analogues are documented: Turkey's 6 February 2023 sequence paired a magnitude 7.8 rupture with a magnitude 7.5 on a neighbouring fault just nine hours later.
Most West Coast earthquake drills assume a single catastrophic jolt, then recovery. The rocks hint at something more unsettling: back-to-back West Coast earthquakes, in which one fault's rupture reloads the stress on its neighbour. Ghost forests drowned beneath Puget Sound lakes now indicate that around 1,100 years ago two major faults let go within the same year — and the mechanics that allowed it have not changed.
The 923 CE Double Rupture Beneath Puget Sound
The Seattle Fault is a shallow, east–west crustal fault system that passes beneath Seattle's downtown, Elliott Bay and Bellevue, with slip surfaces reaching to within a few kilometres of the ground. Roughly 60 kilometres west, across Puget Sound in the Olympic foothills near Lake Cushman, lies the Saddle Mountain fault zone. Both are known to have ruptured about 1,100 years ago: the Seattle Fault lifted Restoration Point on Bainbridge Island some 5 to 7 metres out of the sea, while landsliding and faulting dropped whole stands of forest into lakes on both sides of the Sound. For decades the two were treated as separate earthquakes that happened to fall inside the same radiocarbon window of roughly 900–930 CE, because carbon dating at that age carries uncertainties of several decades. A 2024 tree-ring study in Science Advances narrowed that window to a single year — 923 or 924 CE — and found that trees killed on both sides of Puget Sound stopped growing in the same season. If the two faults slipped simultaneously, the combined moment release is equivalent to a magnitude 7.8 event; if they slipped weeks or months apart, the first rupture is the most plausible trigger for the second.
How Dead Trees Dated an Earthquake to the Season
When coseismic subsidence drops the land or a landslide dams a valley, living conifers are suddenly plunged into standing water and die with their trunks upright and intact. Cold, oxygen-poor lake water — often below 10 °C at depth — preserves that wood for more than a millennium, and because Douglas-fir (Pseudotsuga menziesii) and western redcedar (Thuja plicata) add one ring per year, each trunk carries a barcode of wet and dry summers. Dendrochronologists cross-date that pattern against master chronologies built from old living trees and archaeological timbers that extend well beyond 1,000 years in the Pacific Northwest, pinning the outermost ring to an exact calendar year. Researchers cored drowned trees in Lake Sammamish, Lake Washington and Price Lake plus submerged stumps on the Olympic Peninsula, and found the final rings terminated in the same year, with partial latewood indicating death late in the growing season. That is a precision radiocarbon dating cannot reach, since its uncertainty at this age spans roughly 30 to 60 years — long enough to hide or invent a connection between two faults. In effect, the trees converted a vague geological estimate into something close to an eyewitness timestamp.
🤔 Did You Know?
Drowned Douglas-fir (Pseudotsuga menziesii) snags still standing in Lake Sammamish laid down their final ring in the same season as trees killed about 60 km away on the Olympic Peninsula — silent evidence that two separate faults tore open within a single year.
The Physics of Fault Triggering: Static and Dynamic Stress
Faults do not exist in isolation; they sit in a shared crustal stress field, so when one slips it instantly rearranges the forces acting on its neighbours. Static stress transfer, mapped with Coulomb failure models, typically changes stress on nearby faults by only 0.1 to 1 bar (10–100 kPa), yet that is demonstrably enough to advance rupture on a fault already close to failure. Dynamic triggering works differently: passing seismic waves, especially long-period surface waves, can briefly unclamp faults or pressurise their pore fluids, and documented cases span thousands of kilometres from the source. In the Puget Lowland, the Seattle and Saddle Mountain faults are both products of the same north–south crustal shortening, driven by the Juan de Fuca plate converging obliquely with North America at roughly 35 to 45 millimetres per year, with a few millimetres per year of that shortening absorbed across the lowland itself. That shared loading is why geologists increasingly treat Pacific Northwest hazard as a connected network rather than a list of independent faults. The uncomfortable implication is that the probability of a second damaging quake is at its highest in the hours and days immediately after the first.
Cascadia and the San Andreas: A Longer, Deeper Pattern
The Puget Sound pair is not the only suspicious coincidence on the West Coast. Deep-sea cores from the Cascadia margin preserve turbidites — underwater sediment flows triggered by strong shaking — that record roughly 19 to 20 full-margin megathrust ruptures over the past 10,000 years, an average recurrence near 500 years, the most recent on 26 January 1700. Paleoseismic work led by Chris Goldfinger and colleagues at Oregon State University reported that several northern San Andreas turbidites sit stratigraphically just above Cascadia layers, implying San Andreas ruptures followed Cascadia events within decades or less. Mechanically this is plausible: the two systems physically meet at the Mendocino Triple Junction near 40.4°N, and a magnitude 9 Cascadia rupture would subject the northern San Andreas to minutes of intense dynamic stress. The correlation remains contested, with critics arguing that sediment dating resolution of ±50 years or worse cannot demonstrate causation. Even sceptics, however, accept that a Cascadia megaquake is an efficient mechanism for perturbing stress on faults across western North America.
Modern Doublets: Turkey, Kaikoura, and New Madrid
Earthquake doublets and multi-fault cascades are observed behaviour, not hypothetical curiosities. On 6 February 2023, a magnitude 7.8 rupture on the East Anatolian Fault was followed nine hours later by a magnitude 7.5 on the adjacent Sürgü–Çardak fault, roughly doubling the area of severe shaking. New Zealand's 14 November 2016 Kaikōura earthquake broke at least 12 mapped faults in a single magnitude 7.8 event, producing about 180 kilometres of surface rupture with local slip up to 12 metres and overturning the assumption that ruptures stop at fault boundaries. North America's 1811–1812 New Madrid sequence delivered three major shocks estimated between magnitude 7.0 and 7.5 on 16 December 1811, 23 January 1812 and 7 February 1812, with shaking reported as far east as the Atlantic seaboard. Christchurch learned the cruellest version of the lesson on 22 February 2011, when a magnitude 6.2 event on a previously unmapped fault killed 185 people in a city already damaged by the 2010 mainshock. The consistent pattern is that a first rupture rarely relieves all stored strain, and can redistribute it onto a neighbour closer to a population centre.
What a Back-to-Back Rupture Would Mean for Seattle Today
A magnitude 7-class rupture on the Seattle Fault could produce stronger shaking in central Seattle than a distant Cascadia megathrust event, simply because the source would lie directly beneath the city at shallow depth. The fault crosses beneath Elliott Bay, the industrial tideflats and thousands of older unreinforced masonry buildings, and much of that ground is soft glacial sediment or artificial fill that amplifies ground motion and liquefies. A published scenario study for a magnitude 6.7 Seattle Fault earthquake estimated on the order of 1,600 deaths and about $33 billion in losses, with widespread liquefaction in SoDo and on Harbor Island and tsunami or landslide-generated waves reaching nearby shorelines in roughly three minutes — far too fast for warning systems designed for an offshore Cascadia source. Add a second magnitude 7 on the Saddle Mountain zone or another Olympic Peninsula fault days later, and the losses fall precisely on the infrastructure recovery depends on: bridges, ferry terminals, port cranes and already overloaded hospitals. Emergency planners in Washington increasingly design for a sequence rather than a single event, because contingency plans built around one shock can fail at the second. Seismic retrofitting of masonry buildings, flexible water-main connections and detailed liquefaction mapping remain the cheapest available mitigation.
Can Scientists Forecast the Second Shock?
No one can predict the date of an earthquake, but conditional probabilities after a mainshock are genuinely forecastable. The U.S. Geological Survey now issues aftershock forecasts within about 20 minutes of significant U.S. earthquakes, using statistical models such as ETAS and Reasenberg–Jones that quantify how strongly one rupture clusters further events in space and time. Those models typically put the chance of another magnitude 7 or larger within a week of a magnitude 7 at a few percent — small in absolute terms, but orders of magnitude above the background rate. Research priorities include denser offshore seismometer and seafloor geodesy arrays along Cascadia, updated Coulomb stress maps of Puget Lowland faults, and more tree-ring and turbidite chronologies to test whether paired ruptures are typical or exceptional. The 2023 USGS National Seismic Hazard Model already permits ruptures to jump between connected faults, a direct acknowledgement of cascade behaviour observed at Kaikōura and in Turkey. For residents the practical guidance is blunt: when the shaking stops, treat the following 72 hours as part of the same earthquake.
Final Thoughts
The ghost forests of Puget Sound are a record written in wood: a major West Coast earthquake can arrive with a sequel on a neighbouring fault. Take three concrete steps this week — look up your address on the Washington Geological Survey's seismic hazard and liquefaction maps, enable ShakeAlert notifications on your phone, and build a two-week water and medication supply that assumes a second strong shock before help arrives.
🌍 Explore More Earth Wonders
Frequently Asked Questions
Can the Cascadia and San Andreas faults rupture at the same time?
They are separate fault systems, but they meet at the Mendocino Triple Junction, so a Cascadia megaquake would load the northern San Andreas with large dynamic and static stresses. Offshore turbidite records suggest several northern San Andreas ruptures closely followed Cascadia events, though the dating is too coarse to say whether the gap was hours or decades.
How dangerous is the Seattle Fault earthquake risk?
The Seattle Fault is a shallow crustal fault running beneath downtown Seattle and Bellevue, and hazard models treat it as capable of roughly magnitude 7 to 7.5 earthquakes. Its rupture around 923 CE raised Restoration Point on Bainbridge Island about 5 to 7 metres and generated waves inside Puget Sound that today would reach nearby shorelines within minutes.
Has a back-to-back earthquake ever happened on the West Coast?
Paleoseismic evidence points that way: tree rings from drowned forests date ruptures on both the Seattle Fault and the Saddle Mountain fault zone to the same growing season in 923 or 924 CE. Elsewhere, doublets are firmly documented, including Turkey's magnitude 7.8 and 7.5 pair just nine hours apart on 6 February 2023.
When is the next Cascadia subduction zone earthquake expected?
The last full-margin Cascadia rupture was on 26 January 1700, and turbidite records show roughly 19 to 20 such events in 10,000 years, averaging about 500 years apart. Scientists commonly cite a 10 to 15 percent chance of a magnitude 8 or larger Cascadia earthquake in the next 50 years, with higher odds quoted for the southern margin.
How do scientists date ancient earthquakes using tree rings?
Trees killed by subsidence or landslide-dammed water preserve their final growth ring, which dendrochronologists cross-date against regional master chronologies to get an exact calendar year. In the Puget Sound case this narrowed a radiocarbon window of about 900–930 CE down to 923–924 CE and even identified the season of death.
📚 Further Reading & Research Sources
The following journals and institutions publish peer-reviewed research on the topics covered in this article:
🎉 Did this blow your mind?
Share it with someone who loves Earth’s wonders! What natural phenomenon do you want us to cover next? Leave a comment below.
Images via USGS, NASA Earth Observatory and Wikimedia Commons (public domain / CC BY)
Comments
Post a Comment