More than 1,000 earthquakes occur in the state annually. Washington has a record of at least 20 damaging earthquakes during the past 125 years. Large earthquakes in 1946, 1949, and 1965 killed 15 people and caused more than $200 million (1984 dollars) in property damage. Most of these earthquakes were in western Washington, but several, including the largest historic earthquake in Washington (1872), occurred east of the Cascade crest. Earthquake histories spanning thousands of years from Japan, China, Turkey, and Iran show that large earthquakes recur there on the order of hundreds or thousands of years. Washington's short historical record (starting about 1833) is inadequate to sample its earthquake record. Using a branch of geology called paleoseismology to extend the historical record, geologists have found evidence of large, prehistoric earthquakes in areas where there have been no large historic events, suggesting that most of the state is at risk.
Washington is situated at a convergent continental margin, the collisional boundary between two tectonic plates. The Cascadia subduction zone, which is the convergent boundary between the North America plate and the Juan de Fuca plate, lies offshore from northernmost California to southernmost British Columbia. The two plates are converging at a rate of about 3-4 centimeters per year (about 2 inches per year); in addition, the northward-moving Pacific plate is pushing the Juan de Fuca plate north, causing complex seismic strain to accumulate. Earthquakes are caused by the abrupt release of this slowly accumulated strain.
EARTHQUAKE TYPES IN WASHINGTON
Intraplate or Benioff Zone Earthquakes
Intraplate or Benioff zone earthquakes occur in the subducting Juan de Fuca plate at depths of 25-100 km. The largest of these recorded were the magnitude (M) 7.1 Olympia earthquake in 1949, the M6.5 Seattle-Tacoma earthquake in 1965, the M5.1 Satsop earthquake [p.28-29] in 1999, and now the M6.8 Nisqually earthquake of 2001. Strong shaking during the 1949 Olympia earthquake lasted about 20 seconds; during the 2001 Nisqually earthquake, about 40 seconds. Since 1870, there have been six earthquakes in the Puget Sound basin with measured or estimated magnitudes of 6.0 or larger, making the quiescence from 1965 to 2001 one of the longest in the region's history.
As the Juan de Fuca plate subducts under the North America plate, earthquakes are caused by the abrupt release of slowly accumulated strain. Benioff zone ruptures usually have dip-slip or normal faulting and produce no large aftershocks. These earthquakes are caused by mineral changes as the plate moves deeper into the mantle. Temperature and pressure increase, and the minerals making up the plate alter to denser forms that are more stable at the increased temperature and pressure. The plate shrinks and stresses build up that pull the plate apart.
For the February 28, 2001, Nisqually earthquake, the hypocenter, or point beneath the surface at which the rupture starts, was at 52 kilometers (32 miles). The area of rupture was approximately 30 kilometers by 10 kilometers (18 miles by 6 miles) and slipped approximately one yard. The epicenter was just off the Nisqually delta in Puget Sound. The quake was felt as far north as Vancouver, British Columbia, as far south as Salem, Oregon, as far east as Spokane, Wash., and as far southeast as Salt Lake City, Utah. Most of the damage was sustained in the Olympia and Seattle areas.
Shallow Crustal Earthquakes
Shallow crustal earthquakes occur within about 30 km of the surface. Recent examples occurred near Bremerton in 1997, near Duvall in 1996, off Maury Island in 1995, near Deming in 1990, near North Bend in 1945, just north of Portland in 1962, and on the St. Helens seismic zone (a fault zone running north-northwest through Mount St. Helens) in 1981. All these earthquakes were about M5–5.5. In Oregon, historically a low-seismicity state, crustal earthquakes have recently occurred just south of Portland (M5.7) and in Klamath Falls (M6.0). The largest historic earthquake in Washington (estimated at M7.4), the North Cascades earthquake of 1872, is also thought to have been shallow. It may rank as Washington’s most widely felt earthquake. Because of its remote location and the relatively small population in the region, though, damage was light.
Recent paleoseismology studies are demonstrating previously unrecognized fault hazards. New evidence for a fault system that runs east–west through south Seattle (the Seattle fault) suggests that a major earthquake, M7 or greater, affected the area about 1,000 years ago. Similar large faults occur elsewhere in the Puget Sound but have not been studied in detail.
Subduction Zone (Interplate) Earthquakes
Subduction zone (interplate) earthquakes occur along the interface between tectonic plates. Compelling evidence for great-magnitude earthquakes along the Cascadia subduction zone has recently been discovered. These earthquakes were evidently enormous (M8–9+) and recurred on average every 550 years. The recurrence interval, however, has apparently been irregular, as short as about 100 years and as long as about 1,100 years. The last of these great earthquakes struck Washington about 300 years ago.
HOW EARTHQUAKES CAUSE DAMAGE
The principal ways in which earthquakes cause damage are by strong ground shaking, by the secondary effects of ground failures (surface rupture, ground cracking, landslides, liquefaction, subsidence), or by tsunamis and seiches. Most building damage is caused by ground shaking.
The strength of ground shaking (strong motion) generally decreases with distance from the earthquake source (attenuation), but locally can be much higher than adjacent areas, due to amplification (an increase in strength of shaking for some range of frequencies). At the same time, there is a decrease, or deamplification, in strength of shaking for other frequencies. Amplification occurs where earthquake waves pass from bedrock into softer geologic materials such as sediments.
Strong shaking of long duration is one of the most damaging characteristics of great subduction zone earthquakes. Strong shaking during the 1964 Alaska earthquake lasted about 90 seconds with an additional 90 seconds of strong ground motions “still of alarming magnitude”, followed by “swaying…and shaking a little”. The total time of shaking was about 3 minutes 40 seconds. Strong shaking is a hazard both near the epicenter of an earthquake and in areas where amplification occurs. West Seattle and certain areas of downtown Olympia are examples of places where ground motion has been documented as being significantly stronger than in adjacent areas during the same earthquake. The extensive damage to the Cypress Structure viaduct in Oakland, California, was a classic example of strong ground motion damage during the M7.1 Loma Prieta earthquake of 1989. Most of the damage and deaths in earthquakes are caused by strong ground motion.
Ground failures accompanying earthquakes include fault rupture (surface faulting), ground cracking, subsidence, liquefaction, and landslides.
Fault rupture occurs as offsets of the ground surface and is limited to the immediate area of the fault. Other ground failures can occur over a wide area and can have several causes.
Landslides, including debris avalanches from volcanoes, have been caused by earthquakes. Earthquake-induced acceleration can produce additional downslope force, causing otherwise stable or marginally stable slopes to fail. In the 1964 Alaska earthquake, for instance, most rockfalls and debris avalanches were associated with bedding plane failures in bedrock, probably triggered by this mechanism. In addition, liquefaction of sand lenses or changes in pore pressure in sediments trigger many coastal bluff slides. Rockfalls, such as those that caused two deaths in the 1993 Klamath Falls earthquake in Oregon, can be triggered at great distances from earthquake epicenters.
Liquefaction occurs when water-saturated sands, silts, or (less commonly) gravels are shaken so violently that the grains rearrange and the sediment loses strength, begins to flow out as sand boils (also called sand blows or volcanoes), or causes lateral spreading of overlying layers. Ground failures, such as ground cracking or lateral spreads (landslides on very shallow slopes) commonly occur above liquefied layers. Noteworthy liquefaction took place in Puyallup during the 1949 earthquake. The sands that failed in many cases were sand deposits from Mount Rainier debris flows; similar hazards could be expected in other valley floors downstream from other stratovolcanoes, such as Mount Baker, Mount St. Helens, and Mount Adams.
Subsidence (including differential ground settlement) can result in the flooding and (or) sedimentation of subsided areas, as occurred over broad areas in Chile (1960) and Alaska (1964).
Tsunamis and Seiches
Tsunamis (seismic sea waves) are long-wavelength (large distance between wave crests), long-period (several minutes to several hours between wave crests) sea waves that can be triggered by earthquakes or by landslides into a body of water. These are erroneously called tidal waves even though they are not caused by tides because they are sometimes preceded by a recession of water resembling an extreme low tide. Tsunamis are more damaging when they strike a coastline that has suffered earthquake-induced subsidence.
Seiches resemble tsunamis but occur as standing waves (or sloshes) in enclosed or partially enclosed bodies of water.
CAN EARTHQUAKES BE PREDICTED?
Many precursors to earthquakes have been studied in the hope that they will allow us to predict the size, location, and time of an earthquake, all of which must be accurately predicted simultaneously to be useful in preparing for and responding to earthquakes. Some of the precursors studied are small magnitude earthquakes, water levels in wells, concentrations of radon and helium in ground water, changes in natural electromagnetic radiation, and animal behavior. Psychics and amateur scientists frequently claim (without verification) to be able to predict earthquakes. However, as yet, none of the precursors or other prediction methods have been consistent. Consequently, in the United States, more effort is directed toward understanding earthquake sources and effects than toward prediction.
Prepared by Timothy J. Walsh, Wendy J. Gerstel, Patrick T. Pringle, and Stephen P. Palmer
Washington Division of Geology and Earth Resources