Pacific Northwest Seismic Network

Cascadia Subduction Zone Megathrust

The Cascadia Subduction Zone (CSZ) "megathrust" fault is a 1,000 km long plate boundary that stretches from Northern Vancouver Island, British Columbia to Cape Mendocino, California. It separates the Juan de Fuca and North America plates. New oceanic crust is created offshore along the Juan de Fuca ridge, contributing to both the Juan de Fuca and Pacific Plates. The production of new, warm, buoyant crust moves cooler and denser older crust away from spreading ridges. The Juan de Fuca plate moves towards the North America Plate at about 4 cm per year, causing it to slowly subduct beneath North America.

Earthquake sources in the Pacific Northwest, highlighting the Cascadia subduction zone. Image from Washington State Department of Natural Resources.

At depths shallower than around 30 km, the two plates of the CSZ are locked together by friction. Strain (deformation) slowly builds as the subduction forces continue to act upon the locked plates. Once the fault's frictional strength is exceeded, the rocks slip past each other along the fault in a "megathrust" earthquake.

The fault's frictional properties change with depth. Immediately below the locked part is an area that slides in slow slip events ("slow slip events" ) that slip a few centimeters every dozen months or so. This relieves the plate boundary stresses there, but adds to the stress on the locked part of the fault. Below this transition zone, geodetic (GPS) evidence suggests that the fault slides continuously and silently at long term plate slip rate. From its surface trace offshore to a depth of possibly 5 km, all remote from land, observations are minimal. It remains unknown whether the fault is stuck or slipping silently.

Graphic from IRIS showing how different frictional "zones" vary with depth in the CSZ. Great megathrust earthquakes occur in the “Locked" Zone", slow slip events occur in the “Episodic Tremor and Slip” zone, and the plates move continuously past each other in the “Continuous Slip” zone.

Subduction zone megathrust faults are the only faults on Earth that can produce earthquakes greater than M8.5. The Cascadia Subduction Zone has produced magnitude 9.0 or greater earthquakes in the past, and undoubtedly will in the future.

The last known megathrust earthquake in the Pacific Northwest was on January 26th, 1700, just over 300 years ago, with an estimated magnitude between 8.7 and 9.2 (). Geological evidence shows at least 19 great earthquakes (M8+) occurring over the past ~10,000 years in the Pacific Northwest, with an average recurrence interval of ~500 years. There is evidence for both full-margin ruptures (~M9), where the entire coastline from Canada to California experiences an earthquake, and partial-margin ruptures (~M8), where only part of the coastline experiences an earthquake. The USGS estimates a 10-15% chance of a full-margin ~M9 earthquake occurring on the Cascadia Subduction Zone in the next 50 years.

To learn more about the history of the Cascadia Subduction Zone and the science that led to the discovery of it, see the sections below, which describe land level changes and turbidites created by Cascadia Subduction Zone earthquakes.

In the time between subduction zone earthquakes, when the two converging plates are locked, internal stress stored by the plates slowly deforms the land, pushing it upward and in the direction of motion of the subducting plate. When the plates slip past each other in a major earthquake, the upper plate experiences subsidence: the toe of the upper plate moves up and back, and the previously uplifted land drops down to a lower position.

Between large subduction zone earthquakes, the upper plate is "locked" to the lower plate, which causes it to shorten as strain accumulates. When a large subduction zone earthquake occurs, the plates slide past each other, which relieves the shortening and leads to subsidence in the upper plate. Image from Leonard et al., 2004 (https://doi.org/10.1130/B25369.1).

This subsidence can be recorded in sediment layers. When the land surface drops abruptly during great earthquakes, large numbers of freshwater coastal plants die when they suddenly drop at or below sea level and become inundated with salt water. The dead plants leave evidence in the form of dark soils, called "peat". At low tide, when the soil layers are exposed in riverbanks, you can see the "peaty horizon" exposed. Over time, accumulated sediment and vegetation rebuild the land surface. In the photo below from Oregon's Salmon River, the peaty layer is overlain by tsunami sands which are then covered by gray-green layers of marine plants and sediments. You can also see two Native American campfire pits buried.

Source: U.S. Geological Survey.

Flooding from this subsidence can lead to "ghost forests", which also help record the history of great earthquakes. In this video developed by IRIS (Incorporated Research Institutions of Seismology) — now EarthScope — and others, the process that creates buried soils and ghost forests along our coast is illustrated.

Every day, rivers carry sediment — soil and other debris — into the ocean where it collects on the continental shelf (submerged gently-sloping continental crust of the North American Plate) and continental slope (steep transition to deep sea floor which is the Juan de Fuca plate). Over time, sediment accumulates on the shelf until it becomes unstable and slides down the continental slope like an avalanche. This is called a turbidity current. A number of events can trigger turbidity currents, including tsunamis, storm-induced waves, and slope failures, but earthquakes are inferred to typically be the cause when there is evidence of past turbidity currents. The resulting layer of sediment that the turbidity current deposits is called a turbidite.

Source: NOAA.

Identifying turbidites became a key process for scientists to uncover the history of Cascadia earthquakes. Geologist Gary B. Griggs studied sediment core samples taken from various drainage channels offshore Washington and Oregon, and all samples showed that 13 turbidites had been deposited since the eruption on Mount Mazama (a well-dated geologic event that produced a highly visible layer in all of the samples). To determine if these turbidites were formed during a Cascadia megathrust earthquake, other events had to be ruled out. 

An example of a core sample showing 13 turbidites since the eruption of Mount Mazama. From Hyndman and Rogers (2010), after Adams (1990).

John Adams of the Geological Survey Canada suggested in 1990 that turbidity currents originating from different locations occurred simultaneously during great subduction zone earthquakes. When simultaneous turbidity currents from different side channels merge, the main channel can be expected to show a single large turbidite.  If the turbidites originated at different times in the side channels, the main channel would record each separate turbidite event.  The consistent number of turbidites in core samples from the side and main channels indicate that the turbidity currents were likely caused at the same time, by the same event. 

Large storms are not a likely source of a coast-wide event because these storms produce waves that are not much larger than common storms. If large storms were the trigger, the turbidite record should reflect more than 13 events in the last 5,000 years.

The 1964 Alaska earthquake generated the most recent damaging tsunami that struck the Oregon-Washington coast. Although this earthquake is one of the largest seismic events of the 20th century, it did not produce any known turbidites. If this large tsunami did not trigger a turbidity current, it is highly unlikely the turbidite record reflects the occurrence of distance-source tsunamis.

In a slope failure, so much sediment develops on the inclined continental slope that it slips, much like an avalanche triggered by excessive snowfall. When enough sediment accumulates at a given point on a coastal slope, slope failure occurs. This underwater avalanche can cause turbidity currents to spread sediment throughout the underwater sea channels. Although these kinds of currents are likely to occur spontaneously given enough time, the different rates of sedimentation and inclination of coastal regions make it unlikely that spontaneous turbidity currents across the entire coastline would be synchronized, as is seen in the core samples.

Cascadia Subduction Zone earthquakes, on the other hand, likely provide enough force and affect a large enough region of coast to cause the observed turbidite deposits. Great Cascadia Subduction Zone earthquakes have recurrence intervals of ~500 years, as do turbidity currents. Radiocarbon dating of each turbidite in Adams's core samples show a recurrence interval of about 590 years, closely matching the interval of coastal subsidence observed in Washington. 

In 1996, Chris Goldfinger of Oregon State University began investigating the paleoseismic history of the Cascadia margin based on turbidite records with a team of scientists based at OSU. They conducted their first major cruise to test this hypothesis in 1999 aboard the R/V Melville. One notable conclusion asserted by the OSU research group is that the southern portion of the Cascadia Subduction Zone periodically produces magnitude ~8 earthquakes more frequently and in addition to participating in the full-rupture M9 earthquakes. Most recent work from this group suggests that turbidites indicate a recurrence interval of 480-530 years for full or near-full M9 earthquakes (https://doi.org/10.1016/j.margeo.2016.06.008).

More recently, coastal lakes in the Pacific Northwest, including Ozette Lake, have also been shown to contain records of turbidites. These lake deposits, which are typically less disturbed than marine deposits in the ocean, also suggest ~500 year recurrence intervals for great Cascadia earthquakes. For more information on lake turbidites, see this page from the USGS.