limit the ages of the prehistoric ruptures, and can be used to determine a local recurrence interval and slip rate for the fault. Although the 1886 Charleston earthquakes apparently did not rupture the ground surface, preventing a direct analysis of recurrence rate, studies of liquefaction-related sand blows in the Charleston area by Obermeier et al. (1985) suggest at least two prehistoric events occurred that may be used to establish the recurrence interval of large earthquakes for the area. Thatcher (1984) noted examples in which geodetic measurements may also lead to recurrence estimates.
Mogi (1979) pointed out that the term “seismic gap” has been used to describe two different phenomena. Mogi termed a seismic gap of the first kind as a gap in the spatial distribution of rupture zones of the largest earthquakes in a seismic belt. A second kind of seismic gap is a gap in the seismicity of smaller-magnitude earthquakes before larger earthquakes.
Wallace (1981) described seismic gaps as active-tectonic zones between recently active fault segments with high potential for reactivation in the near future. Wallace and Whitney (1984) examined the paleoseismic history of three segments of the Central Nevada Seismic Belt—the Dixie Valley Fault segment, the Stillwater Fault segment, and the Pleasant Valley Fault zone segment. They found that scarps 104 to 105 yr of age and approximately Holocene age (less than 104 yr) are present in places along all three segments. However, within historical time, only the Pleasant Valley faults in 1915 and the Dixie Valley faults in 1954 ruptured. Within the intervening Stillwater seismic gap there are no free faces preserved on Holocene scarps, indicating that they are probably older than 300 yr. Wallace and Whitney (1984) commented that “the Stillwater gap is a likely site for future major faulting, but the low level of seismicity in the gap area suggests that the next major earthquake is not imminent.”
Other fault systems where seismic gaps of the first kind have been identified are plate boundary systems. McCann et al. (1979) conducted a comprehensive study of large earthquakes and seismic gaps along major plate boundaries. Figure 3.7 shows large earthquakes and seismic gaps along the major plate boundaries near Alaska. The second type of seismic gap applies mainly to the earthquake prediction and management.
Several lines of evidence suggest the Juan de Fuca and Gorda plates are being subducted underneath the North American plate in the northwestern United States. These include (1) the Cascade Range, an active andesitic chain of volcanoes; (2) seismicity related to the Benioff-Wadati zone (Smith and Knapp, 1980; Cockerham, 1984; Tabor and Smith, 1985); (3) geodetic deformation consistent with subduction (Ando and Balazs, 1979; Savage et al., 1981); and (4) deformation of marine terraces consistent with subduction (Adams, 1984). Although active subduction seems clear, this system has been relatively aseismic with respect to great subduction-zone earthquakes. Subduction is either proceeding with relatively little coupling with the overriding plate and thus large elastic strains are not being stored, or the subducting Juan de Fuca/Gorda plate is strongly coupled with the overriding plate, producing conditions in which a large earthquake could occur in the future and the historical period is a period between