High-resolution seismic-reflection data show that the Seattle fault comprises a zone of at least 3 or 4 splays that extends across the Puget Lowland (Fig. 6). The zone is bounded on the north by fault "A" and the syncline on the southern margin of the Seattle basin, which have a similar and distinctive geometry for about 40 km from Lake Washington to Dyes Inlet. Quaternary deformation appears concentrated on fault "A", consistent with northward migration of the thrust front (Johnson et al., 1994). Fault "A" coincides with an anticlinal axis (steeply dipping synclinal beds are in the footwall), and thus represents an "anticlinal breakthrough" (Suppe and Medwedeff, 1990). South of fault "A", the network of structures labeled "B" and "C" are shown to be relatively continuous across the Puget Lowland (Fig. 6) , but it is also possible that these structures are a set of anastomosing, discontinuous faults. Seismic-reflection images of the "B" faults typically show minor truncation and warping of Quaternary strata. Fault "C" at the south end of the zone is locally imaged as both a fault and an unbroken fold axis, and is inferred to be a pre-Quaternary structure with some local and minor Quaternary reactivation.
Largely from gravity surveys (e.g., Finn et al., 1991), previous workers (e.g., Gower et al., 1985; Yount and Gower, 1991; Johnson et al., 1994) have interpreted the Seattle fault as a continuous band across Puget Sound and the Puget Lowland. Our mapping, based on the high-resolution seismic-reflection survey, reveals a more complex, segmented zone (Fig. 6). Starting on the east, the faults (and the syncline at the northern edge of the fault zone) imaged on Lake Washington profiles are inferred to extend along the strike of the local gravity anomaly to eastern Puget Sound. Fault "A" and the adjacent asymmetric syncline on the margin of the Seattle basin are clearly the same basin-margin structures in both areas (Figs. 5A, C). The Seattle fault therefore does not extend through Elliot Bay (Fig. 5B, 6) as previously interpreted, but rather lies on its southern margin. The location of the fault at Alki Point coincides with the northern margin of the terrace uplifted in the ~900 AD earthquake (R. C. Bucknam, written commun., 1997). Fault "C" on each line forms the northern margin of a Quaternary sedimentary basin. East of Lake Washington, Yount and Gower (1991) have suggested that fault "A" extends through southern Lake Samammish, consistent with seismic-reflection profiling of Prunier and others (1996).
On central Puget Sound seismic profiles (e.g., Fig. 5D), the location of faults "A" and "B" and the adjacent fold in the Seattle basin lie about 1,200 m north of their position in eastern Puget Sound (Fig. 6). This displacement also corresponds to truncation of a high-resolution aeromagnetic anomaly in the Seattle fault zone (R.J. Blakely, unpublished data). A second abrupt northward displacement of the leading edge of the fault zone, also about 1,200 m, occurs between seismic profiles in central and western Puget Sound (Figs. 5D-E, 6). This offset is perplexing in that it also appears to dextrally displace faults "B" and "C" and structure contours in the Seattle basin, but does not significantly displace (> 200 m?) either a prominent aeromagnetic anomaly in the Seattle fault zone (R.J. Blakely, unpublished data) or the syncline axis at the southern margin of the Seattle basin.
Fault "A" in western Puget Sound coincides with the northern limit of late Holocene uplift recognized by Bucknam and others (1992) to the west on Bainbridge Island. The fault zone can be traced westward from western Puget Sound to Port Orchard with no apparent displacement of fault splays (Figs. 5E-F, 6). The width of the Seattle fault zone thus varies from about 4 km in eastern Puget Sound and Lake Washington to 5.8 km in Port Orchard and Sinclair Inlet (Fig. 6). Farther west, we match and connect fault "A" with the northern fault on the Dyes Inlet profile (Fig. 5G) which similarly forms the southern margin of a sedimentary basin. Fault "B" is tentatively matched with the southern fault on this profile, suggesting westward narrowing of the zone. As described above, the fault zone does not extend west to our Hood Canal profiles.
The abrupt northward displacements in the Seattle fault zone in Puget Sound could result from the occurrence of lateral ramps or tear faults in the thrust system, or from crosscutting dextral strike-slip faults. This distinction is important because ramps and tears would tend to slip at the same time as other elements of the Seattle fault zone, whereas crosscutting strike-slip faults could move independently and represent an additional potential earthquake source. These possibilities are tested below by examination of east-west seismic profiles in Puget Sound.
A mix of east-west oriented industry and high-resolution seismic-reflection profiles extending from Kingston to offshore of northern Vashon Island in Puget Sound were examined to evaluate the abrupt displacements in the location of the Seattle fault zone (Figs. 2, 6, 7). Most crossings display evidence of high-angle faults. Eight discontinuous fault strands were mapped and are inferred to represent components of a complex dextral shear zone (Fig. 6). Unless noted, each of these mapped structures is recognized on all crossing seismic profiles. The expression of these faults (Fig. 7) on some profiles is more subtle than for the faults in the Seattle fault zone (Fig. 5) because the predominant motion on these structures may be largely lateral and displacements (potentially as much as 1,200 m) are relatively small. The best images of these faults are on profiles where (1) the lateral displacement has juxtaposed markedly different seismic facies or beds; (2) the lateral displacement has juxtaposed parts of folds with different dips; (3) the lateral displacement involves a component of transpressional deformation so that beds adjacent to the fault or between fault splays are warped or arched; (4) vertical displacement is large enough to truncate continuous reflections across the fault. Where lateral displacement may have juxtaposed similar seismic facies and vertical uplift is minimal, faults will be difficult to recognize on seismic data.
Fault "1" coincides with the eastern displacement in the location of fault "A" of the Seattle fault zone. Line P330 (Figs. 2A, 7F) extends west across this fault in the southern part of the Seattle basin. This structure truncates and warps prominent reflections in the Pleistocene section and propagates up to the upper part of the latest Pleistocene to Holocene basin fill. A prominent basin-fill reflection within 40 m of the seafloor appears to be vertically displaced 10 to 15 m. On an industry profile a few km to the north (Fig. 7D), fault "1" is characterized by a zone of truncated and warped reflections and local dip change that extends well upward into the Quaternary section. Farther north, Fault "1" could not be identified on an east-trending high-resolution seismic profile east of northernmost Bainbridge Island (fig. 2A).
Figure 7H shows high-resolution seismic profile P326 that crosses fault "1" a few km south of the Seattle fault zone (Fig. 2A). The fault truncates, warps, and arches reflections on the west limb of an anticline that underlies the central Puget Sound trough, and extends upward to at least the middle of the Pleistocene section before being obscured by multiples. On a parallel industry profile (Fig. 7G), the fault is characterized by a dip change in the inferred uppermost T ertiary beds and by truncated reflections in the lower part of the Pleistocene section. Farther south on the east end of line P324 east of northern Vashon Island (Figs. 2A, 7I), fault "1" cuts east-dipping reflections in Tertiary strata but multiples obscure the Pleistocene section. On both P324 and P326, the latest Pleistocene to Holocene basin fill on the west flank of the central Puget Sound trough dips east as much as 3° - 5°, indicating east-west Holocene contraction that could either be associated with regional deformation or with more local transpression. South of P324, fault "1" strikes onto Vashon Island in an area of extensive landsliding (Booth, 1991). Fault "1" therefore appears to have a minimum length of about 24 km.
Fault "2" was only recognized on line P324 (Fig. 2A, 7I) east of Vashon Island and thus can be no longer than about 2 km. This structure appears to displace the floor of the latest Pleistocene to Holocene basin between 15 and 30 m, depending on how truncated reflections are matched. Although it is conceivable that this offset results from relict glaciofluvial topography, similar relief is rare elsewhere at the base of the central Puget Sound trough and this alternative hypothesis is considered less likely. Slightly lower in the seismic profile (~ 0.45 s TWT), the inferred Tertiary-Quaternary contact appears to be displaced about 20-25 m in an opposite sense, up to the east. Lower in the profile, the fault truncates and warps reflections from inferred Tertiary strata, including a gentle anticline east of the fault (best seen at 0.5-1.0 s TWT).
Fault "3" comprises 2 splays (Fig. 2A, 7J). The western splay truncates a zone of east-dipping reflections in inferred Tertiary strata and juxtaposes it with flat-lying beds east of the splay. At 0.5 s, reflections in Tertiary strata at the base of the latest Pleistocene to Holocene basin are markedly downwarped along the fault. The eastern splay similarly cuts and warps reflections along its trace and appears to displace the basin floor about 40 m. Because the basin floor at this locality probably consists of Tertiary strata, the timing of basin-floor displacement could be pre-Quaternary (i.e., relict relief filled in by latest Pleistocene to Holocene deposits). However, reflections in the lower part of the basin fill on the west flank of the basin-floor uplift converge toward and over the uplift, indicating that at least some faulting and uplift coincided with latest Pleistocene to Holocene deposition. Fault "3" was not identified on line P324 (Fig. 2A, 7I) and its southern extent has not been mapped. Thus, it has a minimum length of 2-3 km.
Figure 7A shows an industry seismic profile that crosses fault "4" east of Kingston (Fig. 2B), about 20 km north of the Seattle fault zone. The fault is imaged as a high-angle, upward-diverging structure that truncates and warps reflections. Strata east of fault "4" form a gentle, north-trending (based on analysis of additional industry profiles) anticline, further evidence for east-west contraction in the Puget Lowland. Line P338 (Fig. 2A, 7B) shows fault "4" as 2 splays that truncate and warp reflections and similarly cut the west limb of a gentle anticline. There is a significant contrast in reflection amplitude and density for the strata between the two splays and strata imaged to the east and west of the two splays. Fault "4" occurs along the same trend as fault "1", however we could not identify a fault that would link these structures on a high-resolution profile between their mapped traces (Figs. 2A, 6). Fault "4" extends for at least 6 km north of Kingston based on analysis of additional seismic profiles and thus has a minimum length of about 13 km.
Fault "5" coincides with the western abrupt displacement in the location of fault "A" of the Seattle fault zone (Fig. 6) and with a ~200-m-high bathymetric lineament that forms the western boundary of the central Puget Sound trough (fig. 4). Figure 7F shows an east-trending profile that extends from the northern part of the Seattle uplift (hanging wall of the Seattle fault) on the west across fault "5" and into the southernmost Seattle basin (Figs. 2A, 6). The fault is not well imaged on this profile because of the multiples and diffractions associated with steeply dipping bedrock of the Seattle uplift. However, across this fault there is more than 500 m of vertical relief on the base of the Quaternary.
South of Line P330 (Fig. 7F) but within the Seattle fault zone, there is no significant dextral displacement (> 200 m?) of a continuous east-trending aeromagnetic anomaly (R.J. Blakely, unpublished data) along the projection of fault "5". North of the Line P330, dextral displacement of the axis of the syncline at the southern margin of the Seattle basin also does not appear to exceed 200 m, although this feature is difficult to map precisely. Thus, fault "5" appears to be largely confined to the northernmost part of the hanging wall of the Seattle fault, and should be regarded as a tear fault within the Seattle fault zone. This structure has a length of about 1 to 1.5 km.
Evidence from seismic-reflection data
Fault "6" extends south from the southern part of the Seattle fault zone in western Puget Sound (Fig. 6). On the industry profile shown in Figure 7G, the fault is marked by an abrupt change in dip in Tertiary beds (steeper west dips on the west, shallower east dips on the east) as well as truncated and arched truncations along the inferred fault plane. Reflections in the Pleistocene section are cut and warped within about 100-150 m of the seafloor. On high-resolution seismic-reflection profiles (e.g., line P326, Fig. 7H), fault "6" is characterized by two splays that truncate and warp reflections in the lower part of the Quaternary section. Arched reflections between the two splays are juxtaposed with reflections of contrasting geometry and density. To the north, fault "6" appears to extend into the southern part of the Seattle fault zone. To the south, the fault projects onto northern Vashon Island.
Investigation of land exposures
Fault "6" projects south along the ~3 km, north-trending western shoreline of northern Vashon Island, intersecting the coastline at Fern Cove (Figs. 6, 8). The projected trace then continues south across the Island and passes offshore through the mouth of Quartermaster Harbor. With the exception of a few gravel pits, virtually all of the outcrop on Vashon Island occurs in discontinuous bluffs along shorelines. West of the projected trace of fault "6", Booth (1991) correlated virtually all shoreline exposures with pre ~28 ka glacial and interglacial units (Fig. 8). East of the projected fault trace, Booth correlated virtually all shoreline exposures with post ~28 ka glacial and interglacial units.
On the north end of Vashon Island, there is no exposure along the projected fault trace but the nearest shoreline outcrops on either side of the projection contain evidence of late Quaternary deformation. Quaternary strata in coastal bluffs at Vashon Heights (Fig. 8) on the north coast of the Island about 600 m east of the projected fault consist of ~10-15 m of glaciolacustrine deposits of the ~20 ka Lawton Clay overlain by >50 m of sandy outwash of the ~15-18 ka Esperance Sand (Fig. 3; Booth, 1991; Easterbrook, 1994a). Exposures are restricted to the lower ~3-4 m of the bluffs and a few landslide scarps. Exposed beds are warped (dips of ~4°) into a gentle north-trending anticline and the westernmost exposures are cut by high-angle faults and fractures with vertical displacements of as much as 40 cm.
About 1 m of Quaternary strata are exposed during the lowest tides along a 45-m-wide beach west of Fern Cove, about 500 m west of the projection of inferred fault "6" (Fig. 8). These strata consist of glaciolacustrine silt and clay with common dropstones, and Booth (1991) considered them part of a pre-28 ka sequence of glacial drift. Beds have a general westerly strike (~260°), dip about 20° to the south, and are cut by several discontinuous, north-trending, high-angle shears bounded by meter-scale drag folds.
The geologic relationships outlined above are consistent with a projection of fault "6" across Vashon Island (Figs. 6, 8). Stratigraphic units and (or) facies exposed along shorelines are apparently juxtaposed across the projected trace on both a local and island scale. The east-west distribution of stratigraphic units suggests west-side-up displacement. The amount of possible vertical displacement can not be easily determined from these data because of probable stratigraphic complexity (e.g., large-scale glacial scour and fill), but Booth (1991) noted about 100 m of relief from east to west on the base of the ~15-18 Ka Esperance Sand along contours subparallel to the projected fault trend. The gentle folding and small scale structural disruption (faults and fractures) noted in strata adjacent to the projected fault trace is uncommon within Quaternary exposures along Puget Sound and suggests that shearing and deformation associated with the fault may be distributed across several hundred meters.
Alternatively, the east-west contrast in relative ages of shoreline exposures could reflect large-scale scour and fill associated with repeated Puget Sound glaciations (Fig. 3), and some of the anomalous structural deformation could have a glaciotectonic origin (e.g., Aber et al., 1989). Thus, geological data on Vashon Island are consistent with but do not absolutely demand continuation of fault "6" to the south.
On Figure 7D, a northwest-trending industry profile (Fig. 2B), fault "7" is characterized by truncations and abrupt dip change (steeper east dips east of fault, flat dips west of fault) in Tertiary strata but does not obviously penetrate the Quaternary section. Two high-resolution seismic profiles from the same area (Fig. 2A) as this industry profile, however, clearly show displacement and warping of Quaternary strata on two structures considered splays of fault "7" (e.g., Fig.7E). The eastern splay juxtaposes and cuts reflections of contrasting amplitude and density in the Pleistocene section, appears to displace the floor of the latest Pleistocene to Holocene basin, gently warps Holocene basin reflections, and projects upward to a notch in the seafloor. The western splay is harder to interpret because of multiples, but also cuts Pleistocene reflections and juxtaposes disparate seismic facies. No Quaternary deformation along the fault "7" trend can be identified on the next two high-resolution seismic profiles to the north (Fig. 2A), and evidence for Quaternary deformation on the next high-resolution profiles to the south is ambiguous. Thus, fault "7" has an inferred minimum length of about 3 km.
Evidence from seismic-reflection data
Fault "8" extends northward in western Puget Sound from offshore of northern Bainbridge Island to the Kitsap Peninsula (Fig. 6). On an industry profile (Fig. 7C), fault "8" is characterized by truncated and warped reflections, local dip changes, and an upward divergence of splays, and propagates up into the lower part of the Quaternary section. On high-resolution seismic-reflection profiles (e.g., line P338, Fig. 7B), fault "8" is characterized by two splays that cut and warp Quaternary reflections and cause juxtaposition of reflections with contrasting reflection geometry and amplitude. Fault 8 must end north of the location of two high-resolution seismic profiles offshore of northeastern Bainbridge Island, on which the fault was not recognized.
Investigation of land exposures
North of the seismic profiles shown in Figures 7B and 7C,, fault "8" projects onshore into the mouth of a 500-m-wide, alluvium-filled, north-trending valley on the Kitsap Peninsula (Fig. 6). The fault trace is covered, and strata exposed in shoreline bluffs in the pre-40,000 ka Quaternary section on the eastern and western flanks of the valley are different (Yount et al., 1993). The ~15-m-thick section on the east flank of the valley consists of a ~3 m-thick-lower unit of flat-bedded sandy and silty glaciolacustrine deposits and an overlying unit (~ 12 m thick) of massive to locally crudely stratified pebbly till. The lower unit has a variable west dip that is typically less than 5° and the contact between the two units is locally an angular unconformity. On the west flank of the valley, a gently west-dipping (< 5°), ~20-m-thick Quaternary strata exposed in coastal bluffs consist of a lower unit (5-8 m thick) of flat- and cross-stratified pebbly outwash and an upper unit (12-15 m thick) of sandy outwash. The gentle folding and stratigraphic contrasts across this covered area are consistent with but do not demand the presence of a fault at this locality. Fault "8" thus has a minimum length of about 4-5 km.
Seismic-reflection data indicate that a zone of en-echelon, north-trending, high-angle faults are present in central Puget Sound. These faults truncate and warp reflections, commonly juxtapose strata with different dips and contrasting seismic facies, and display some evidence of dip reversal (Fig. 7I). These characteristics typify oblique strike-slip faults (e.g., Harding et al., 1983), consistent with the dextral displacement of piercing points in the Seattle fault zone.
Faults "1" and "6" are inferred to be the longest and most significant structures. An extensional stepover (Aydin and Nur, 1985) between these faults (area of overlap between northern Vashon Island and the southern Seattle fault zone) and associated normal faulting may be partly responsible for the significant relief (~500 m) on the base of the Quaternary within the Seattle uplift from the center of Puget Sound to its east and west flanks (Fig. 6). At its southern end, our data suggest offset on fault "1" may transfer to parallel structures to the east (faults "2" and "3"). To the north, we think it likely that faults "1" and "4" are connected, despite the lack of evidence on one crossing seismic profile.
On the west flank of Puget Sound, the alignment of discontinuous faults "6", "5", "7", and "8" provides an indication that these structures may have once formed a continuous fault. In this scenario, this continuous fault may have become fragmented with increasing transfer of offset to structures in eastern Puget Sound. The coincidence of these structures with the prominent bathymetric lineament on the west flank of the central Puget Sound trough (Fig. 4) suggests this discontinuous fault trend has provided an ongoing control on the location of glaciofluvial channels and erosion.
The north-trending zone of faults in central Puget Sound appear to be discrete structures at the surface but are probably connected at depth. This zone lies a few km west of and parallel to the Coast Range Boundary fault, an inferred regional crustal block boundary (Fig. 1; Johnson et al., 1996), and may be rooted within that contact. In that scenario, this north-trending zone of strike-slip and normal faulting represents a portion of a regional distributed shear zone along which the Washington Coast Range is moving northward relative to the eastern Puget Lowland and Cascade Range.
These faults must be considered active or potentially active because (1) almost all profiles show the faults extending upward into the Quaternary section and (2) these faults displace the northernmost active strand of the Seattle fault. Focal mechanisms from the central Puget Lowland indicate a mix of thrust, strike-slip, and normal faulting (Ma et al., 1996), consistent with the complex pattern of faulting shown in Figure 6. The north-trending zone extends up the axis of the Puget Lowland trough (Fig. 6) and has probably provided an element of structural control (zones of lithologic weakness more erodible than adjacent blocks) on the development of Puget Sound.