The short answer is that we don't know, but several lines of evidence now seem to limit the number of possibilities. Ideally a prediction would specify the time, place, and magnitude of the next earthquake. With current information, this is about as easy as predicting the next traffic accident in a large city like Seattle. A specific traffic accident cannot be predicted, but experience shows that accidents are more likely during evening rush hours. They also occur more frequently on certain streets and more often involve two rather than ten cars. Similarly, the pattern of past earthquake activity can be used to estimate how often and where earthquakes of a given magnitude and location are likely to occur in the future. With the addition of geologic information, we can refine the estimate of where these earthquakes might occur.

Stepp (1973), for example, calculated earthquake return times versus earthquake size in the Puget Sound region by tabulating the maximum intensities reported from past earthquakes. (He used earthquake intensities rather than magnitudes because magnitudes were not generally calculated before 1960.) From this recurrence pattern, Stepp calculated that an earthquake of maximum intensity VII would occur in the Puget Sound region every 7 years on the average. Similarly, he estimated that the average return time of intensity VIII and IX earthquakes is 23 years and 73 years respectively; such earthquakes would be roughly equivalent to the ones in 1965 and 1949 (Table 2).

In the Puget Sound area, the activity of the deep earthquake zone in the subducted Juan de Fuca plate has differed significantly from the activity of the shallow earthquake zone in the overlying North America plate (depths less than 30 kilometers). Because earthquakes in Washington can now be clearly grouped into either the shallow or deep zone, seismologists have tried to make separate forecasts for each group. Puget Sound earthquakes recorded by the University of Washington network having magnitudes greater than 4 have been more numerous in the subducting Juan de Fuca plate. This observation agrees with studies suggesting that older large Puget Sound earthquakes are deep. This agreement supports the idea that an earthquake of magnitude 7 or greater in the Puget Sound region is much more likely to be deep than shallow. Similarly, a study by the U.S. Geological Survey (1975) proposed that the largest earthquake likely to occur in the Puget Sound region would have a depth of 50 kilometers and a magnitude as large as 7.5.

There is still no consensus on the probability of a large shallow earthquake, especially in the Puget Sound region. Based on information from the last 150 years, Rasmussen and others (1974) proposed that shallow earthquakes in the Puget Sound area would probably not have magnitudes exceeding 6.5. Only one of the largest Puget Sound earthquakes was reported to have had aftershocks; this event, in 1880, is very poorly documented. So far, however, estimates of the largest shallow Puget Sound earthquake suggest that the size would probably be smaller than the largest deep Puget Sound earthquake.

Large shallow earthquakes have occurred in the Cascades. The 1872 North Cascades earthquake was probably shallow. Weaver and Smith (1983) argue that the St. Helens Seismic Zone in the Cascades of southern Washington may also be capable of producing a large shallow earthquake having a magnitude of 7.0. Therefore, future earthquakes in the Cascades will likely be shallow and could have a magnitude of 7 or greater. In eastern Washington and Oregon, all earthquakes have been shallow, but no earthquakes having a magnitude exceeding 5.7 have occurred there in the past 150 years,

The largest earthquake proposed for Washington and Oregon is a subduction earthquake exceeding magnitude 8 located between the Juan de Fuca plate and the overlying North America plate (Heaton and Kanamori, 1984). Since the Juan de Fuca and North America plates are converging at a rate of 3 to 4 centimeters a year, Heaton and Hartzell (1987) have estimated an average return time of 400 to 500 years for a great subduction earthquake in the Pacific Northwest. Because of such a long return time, it is not surprising that we have not experienced such an earthquake within the last 150 years. Buried tidal marshes, newly discovered in bays and rivers along Washington's coast, were possibly submerged suddenly during large subduction earthquakes that struck western Washington as recently as 300 years ago. Atwater and others (1987) and Hull (1987) have used geologic evidence to show that such large earthquakes may have occurred at least eight times in the past 5,000 years.

To summarize: the maximum probable earthquake in Washington would be a subduction earthquake having a magnitude exceeding 8 and an epicenter near the coast; it would be caused by sudden slip between the Juan de Fuca and North America plates. From the dating of organic material preserved in rocks along the coast, some scientists believe that such earthquakes have occurred every 300 to 1,000 years. Other large earthquakes in the Puget Sound region can be expected to have magnitudes of at least 6.5 to 7.5 and depths greater than 40 kilometers. Rasmussen and others (1974) estimate 10year return times for magnitude 6 earthquakes in the Puget Sound area. They also propose return times of 35 and 110 years for magnitude 6.5 and 7.0 earthquakes, respectively.