is offset 5.5 m (dip-slip component), which yields an average long-term slip rate of 0.05 m/ka. More importantly, however, scarp-derived talus that accumulated on the marine platform below the periodically rejuvinated fault scarp records at least nine slip events over the past 104 ka (Figure 6.22B) (Weber and Cotton, 1981). If coseismic, these slip events, which averaged 0.6 m, yield an average earthquake recurrence interval of about 12 ka. It should be stressed, however, that if total fault movement was restricted to relatively short periods of time (Figure 6.22B), average recurrence intervals are of little use in predicting the time of the next event. Each event must be dated independently to make meaningful predictions of future events.
As indicated previously, lateral fault movement is rarely recorded by marine strandlines, even where strandlines clearly cross active strike-slip faults. Usually, the amount of offset is ambiguous because of irregularities in or poor preservation of the paleoshoreline. For example, at Point Año Nuevo on the central California coastline, two Pleistocene strandlines that cross a major right-lateral fault system appear to be offset across several fault strands (Weber and Lajoie, 1977; Weber and Cotton, 1981). However, because these strandlines must be projected considerable distances (up to 1.0 km) at low angles across some of the fault strands, the individual and cumulative offsets of each strandline are too uncertain to yield meaningful rates of lateral fault movement.
In some areas, the sense but not the amount of lateral fault movement is expressed by the orientation of drag folds recorded by marine strandlines. For example, the 82-ka strandline at Half Moon Bay about 60 km northwest of Point Año Nuevo on the central California coastline records broad crustal warps on both sides of a major fault strand within the coastal fault system (Figure 6.21) (Lajoie et al., 1979b; Kennedy et al., 1981). Structural contours on the warped wave-cut platform define a broad syncline that plunges obliquely into the vertical fault plane from the east and a broad anticline that plunges away from the fault plane to the west. These structural relationships suggest that the syncline and anticline are drag folds related to right-lateral fault movement, which is consistent with regional fault movements. Unfortunately, the strandline of this deformed marine platform does not intersect the fault plane and, therefore, neither the local slip vector nor the slip rate is known. However, where latest Pleistocene stream courses cross the warped marine terrace northeast of the fault they are deflected toward the axis of the syncline (Figure 6.21), which suggests that fault movement and related crustal warping were continual over the past 82 ka.
In several seismically active coastal areas historical coseismic uplift has produced conspicuous emergent marine strandlines 1–15 m above sea level. The best documented examples of historical coseismic strandlines are found in Japan (Sugimura and Naruse, 1954; Nakamura et al., 1965; Ota et al., 1976; Matsuda et al., 1978; Shimazaki and Nakata, 1980), New Zealand (Wellman, 1969; Stevens, 1973), Alaska (Plafker, 1965, 1972), Chile (Plafker and Rubin, 1967; Plafker and Savage, 1970; Plafker, 1972), and Iran (Page et al., 1979). The historical coseismic strandlines in these areas suggest that similar higher (older) Holocene strandlines in these and other active-tectonic areas are also of earthquake origin. If this interpretation is correct, a sequence of emergent Holocene strandlines is a physical record of past earthquakes from which it should be possible to predict the size and date of the next seismic event.
If the average uplift rate is constant, coseismic uplift events may follow a time-predictable or displacement-predictable pattern (Figure 6.23) (Shimazaki and Nakata, 1980). In a time-predictable pattern the time between events is proportional to the size of the preceding event, and, therefore, the date, but not the size, of the next event can be predicted. In a displacement-predictable pattern the time between events is proportional to the size of the succeeding event, and the size of the next event can be predicted for any future date, but that date cannot be predicted. Of course, if the period and size of uplift events are regular, both can be predicted for the next event. In practice, however, strandline elevations and dates are usually too variable or imprecise to fit any pattern exactly, and, therefore, only average displacements and recurrence intervals can be derived from sequences of emergent Holocene strandlines. In some cases, the uplift rate is so variable that no reasonable predictions can be made.
A few examples of coseismic uplift from Japan, Alaska, and New Zealand illustrate some of the possibilities and limitations in deriving detailed seismic histories from sequences of emergent Holocene strandlines.
The Japanese archipelago, which lies along the leading, overthrust margin of the Eurasian tectonic plate in the western Pacific Ocean, is one of the most seismically