A major earthquake on the Seattle Fault Zone would likely cause more destruction in Seattle than the Cascadia subduction zone event that has dominated Pacific Northwest disaster planning for decades. That assessment does not come from a fringe source or an outlier model. It comes from the lead researcher of a peer-reviewed study published in GSA Bulletin in January 2026, a geologist at the U.S. Geological Survey’s Earthquake Science Center who has spent years excavating the ground beneath Seattle to read what the fault system left behind.
Findings published in GSA Bulletin in January 2026 quantify the rupture history of two newly mapped secondary faults within the Seattle Fault Zone, establishing that these structures have fired repeatedly since the end of the last ice age and most recently ruptured around 1833, producing physical changes to the landscape that persisted for centuries.
The Seattle Fault Zone is an east-west reverse fault system running directly beneath Bainbridge Island and the urban core of Seattle, extending approximately 70 kilometres under the densely populated Puget Sound region. It is not a single clean break in the earth. It is a layered system of primary blind faults, meaning faults that never reach the surface, sitting beneath a complex network of shallower secondary structures that fold, bend, and fracture the rock above them as strain accumulates. The primary faults sit kilometres underground and are estimated to have accumulated more than 7 kilometres of displacement since the Eocene epoch. The secondary faults sit close enough to the surface that when they rupture, the ground moves, shorelines shift, and forests die.
The fault system accommodates roughly 15 percent of the total north-south compressional strain measured across the Cascadia subduction zone forearc, the wedge of crust caught between the offshore subducting plate and the stable interior of North America. That strain does not dissipate. It builds continuously and releases intermittently. The question has always been where, how large, and how often.
For the primary faults, the answer to how often is sobering but not immediately alarming. The most recent major rupture on the primary Frontal fault occurred around 923 CE, producing an estimated magnitude 7.5 earthquake that lifted the shoreline of Bainbridge Island by 6 to 7 metres, generated a tsunami in Puget Sound, triggered widespread landslides, and killed trees across the region. No comparable rupture on the primary fault has occurred in the 1,100 years since. The recurrence interval for that scale of event on the primary fault is estimated at 5,000 years or more.
The secondary faults tell a different story entirely. These structures, shallower and shorter than the primary faults, were built by the same folding process that drives the whole system. As the primary blind faults push rock upward and northward, the overlying strata bend. Where those strata bend sharply enough, they crack. The cracks become faults. Those faults accumulate strain independently of the primary structure beneath them, and they release it on their own schedule. Over the past 2,500 years, the earthquake history of the Seattle Fault Zone has been dominated not by the rare catastrophic primary fault rupture but by these smaller, shallower, more frequent secondary events firing every few hundred years directly beneath the city.
Two of these secondary faults, the Lytle Beach fault on southern Bainbridge Island and the Vasa Park fault near Lake Sammamish, were mapped in detail for the first time using a combination of lidar imaging, ground-based magnetic surveys, trench excavations, radiocarbon dating, and tree-ring analysis. Lidar, a laser-based technology that strips away vegetation cover to reveal the bare ground surface beneath, identified subtle north-facing scarps, steps in the terrain formed by fault movement, that had been invisible under the dense Pacific Northwest forest canopy. Trenches cut across those scarps exposed the sediment layers beneath, preserving a readable record of past ruptures in the geometry of deformed, displaced, and buried soils.
The Lytle Beach fault produced at least two surface-rupturing earthquakes. The older event, designated RH1, occurred between approximately 11,240 and 10,430 years ago, in the immediate aftermath of the last glaciation. The more recent event, designated RH2, is constrained by radiocarbon dating to after 1663 CE, and by tree-ring evidence from a stand of western red cedars that drowned in a pond on Bainbridge Island to approximately 1833 CE. Those trees did not die from logging. They died from flooding triggered by fault-driven ground deformation. Their jagged, broken tops are still visible above the water surface today. The RH2 rupture uplifted the adjacent shoreline by at least 0.7 metres and left physical evidence of displacement that persisted long enough to be measured by modern survey equipment nearly two centuries later.
The Vasa Park fault produced one documented surface rupture, designated VP1, constrained to between 11,380 and 7,400 years ago. The scarp it left behind stands approximately 2 metres high. The fault-parallel slip estimated from that scarp height is approximately 2.5 metres, with around 1.4 metres of horizontal shortening across the fault plane in a single event.
Neither of these faults appeared in the national seismic hazard models used to set building codes, design bridges, plan emergency response infrastructure, and calculate insurance risk across the Seattle metropolitan area. They were excluded not through negligence but through a structural limitation of the modelling process: faults below a minimum length threshold are not included because they are considered incapable of generating large enough earthquakes to warrant the complexity of inclusion. In the Seattle Fault Zone, that threshold filters out the structures that have been most active, most recently, closest to the surface, and directly beneath the city’s core.
The gap between what the models contain and what the ground records is not academic. Seattle’s bridges, hospitals, highway overpasses, water mains, and high-rise towers were designed to withstand shaking calculated from hazard models that do not fully account for the fault network underneath them. The 923 CE earthquake on the primary Frontal fault generated more than 15 metres of slip at depth and a tsunami that deposited sediment across Puget Sound. A comparable event on the primary fault remains possible within the current geological timeframe. But it is the secondary faults, firing more frequently, sitting shallower, and positioned directly under the urban core, that present the more immediate and less-modelled threat.
The distinction between Cascadia and the Seattle Fault matters because proximity and depth determine the character of shaking. The Cascadia subduction zone sits offshore, and its ruptures generate long, rolling, low-frequency waves that travel hundreds of kilometres before reaching Seattle. The Seattle Fault sits at most a few kilometres beneath the streets. Its ruptures generate short, sharp, high-frequency shaking concentrated directly at the source. High-frequency shaking at close range is what collapses buildings, drops bridges, and ruptures pipelines. It is what liquefies saturated soils. It is what the urban infrastructure of Seattle is least equipped to absorb.
Four million people live within the impact zone of the Seattle Fault Zone. The secondary faults within that system have ruptured repeatedly, have done so as recently as approximately 190 years ago, and are still not fully represented in the models that govern how the city is built and how its emergencies are planned. The physical evidence is in the ground. It is in the shifted shorelines, the drowned forests, the deformed sediment layers visible in trench walls on Bainbridge Island. The fault left a record. The record has now been read.
Source:
Angster, S.J., Sherrod, B.L., Pearl, J.K., Staisch, L.M., Johns, W., and Blakely, R.J., 2026, Latest Pleistocene to nineteenth-century earthquakes on bending-moment reverse faults of the Seattle fault zone, Washington: GSA Bulletin, https://doi.org/10.1130/B38333.1
