Click on image for larger version
[Next Section] [Previous Section] [Contents]
In order to address how the deformational structures underlying Hood Canal evolved, the convergent margin tectonics that control this forearc region must be considered. The schematic flow chart shown in Figure 34. The question is whether bedrock deformation patterns in the Hood Canal area reflects strike-slip faulting along the margins of a décollement (Pratt et al., 1997), thrust faulting as suggested by Wells and Snavely (1996), or normal faulting (e.g. trackline 18; Figure 25). Faulting patterns presently observed in the Crescent Formation volcanics reflect long-lived and varying periods of deformation.
Overall, the fault structures in Hood Canal are extensional and defined by horst and graben and half-graben topography. In the northern study area (tracklines 26-50), bedrock structures are dominantly extensional, however there are shortening structures
Click on image for larger version
Figure 34: A generalized regional deformational history. This chart combines and clarifies some possible relationships of deformation, regional tectonism, and their combined effect on the observed structure of Crescent Formation bedrock shown in the interpreted seismic records. The present-day tectonic framework is given at the bottom of the page.
present (Figure 35). In particular, 0.5-2.0 km east and northeast of the Dosewallips River delta (Figure 22) and west of the Hood Canal Discovery Bay fault line (Figure 35). Here inverted V-shaped pressure ridge (‘pop-up’) features are observed on tracklines 40 (Appendix), 41 (Figure 27), and 48 (Figure 31). These shortening features are acoustically continuous and can be traced laterally across several records. Although it is possible these structures are related to a large landslide along the delta front, their seismic continuity and relatively undisturbed internal nature argues against it. A large landslide deposit however cannot be ruled out without further study of the area adjacent to the delta. A general relationship exists between significant bedrock vertical displacement and the Hood Canal Discovery Bay fault trace (e.g. Figure 29). This relationship can be used to generally define the orientation of major structural offset along Hood Canal.. For example, north of Quatsap Point (Figure 1) the northeast-trending fault trace makes a wide-arcing northward bend up into Dabob Bay (Figure 3). If the Hood Canal-Discovery Bay fault (HDF) exhibits dextral strike-slip displacement, structures might be observed that resemble those seen along westward-directed restraining bends of the dextral San Andreas Fault system. The shortening structures on the west side of the HDF shown in Figure 35 may have resulted from compression along a restraining bend. This pattern is typically seen where dextral strike-slip faults step or bend left. Extensional faulting seen on the northern (lines 26-50) seismic profiles is
Click on image for larger version
Figure 35: This map is a compilation of bedrock structural patterns observed on seismic tracklines 26-50. East of the Dosewallips River delta and west of the Hood Canal-Discovery Bay fault, are ellipses representing zones of shortening. These zones are inferred from folds and pressure ridges on the seismic profiles. Horst and graben structures are noted elsewhere and indicate overall extension.
widespread. In fact, the area between Quatsap Point and the southern shoreline of Toandos Peninsula might be best characterized as a pull-apart basin. A right step-over or bend along a dextral high-angle strike-slip fault trending up northern Hood Canal (east side of Toandos Peninsula) could cause a releasing bend and pull-apart basin to develop (Figure 35). Although sufficient focal mechanisms are not available to characterize fault motion in Hood Canal (personal communication; Ruth Ludwin, 1997), the right lateral sense of motion is favored based on the N-S and NE-SW compression shown in the Puget Lowland (Pratt et al., 1997). Combined the extensional and minor compressional structures seen in Hood Canal dataset indicate possible strike-slip displacement and argue against thrust faulting indicated by Snavely and Wells (1996). High angle strike-slip faults are common in forearc settings especially where the angle of convergence is strongly oblique like the Pacific-North American system (McCalpin, 1996).
Sediment deformational patterns are pervasive throughout the Hood Canal dataset. Near-vertical blind faults, many terminating within 10 to 50 m of the sediment-water interface, originate deep in the seismic section and migrate vertically up to 400 meters. A review of both current and paleo-glacial seismic-stratigraphic unit interpretations (e.g. Syvitski et al., 1997) did not reveal an analogue for these deep-seated faults. They do not appear to resemble glaciogenic sediment deformation related to ice-terminus thrust, near-surface liquefaction, or ice-contact structures. The faulted, high relief bedrock erosional surface beneath these sediments undoubtedly reflects pre-existing and possibly long-lived tectonic faulting patterns. What then is the driving force for these deep-seated deformational structures?
There are several possibilities: 1) glacial and post-glacial uplift-related deformation; 2) sediment depositional or erosional structures; 3) tectonically related coseismic or post-seismic fault structures, or 4) shear structures related to bedrock displacement. As previously discussed, post-glacial crustal uplift was completed by 9 to 10 ka (Thorson, 1989). Therefore, any uplift-related sediment shear structures would have to be preserved in the geologic record for at least that time period. This is not a likely considering the low preservation potential of unconsolidated sediment structures in an inherently erosive glacial environment.
Are these blind faults more ‘naturally-occurring’ sediment load structures or simply depositional and erosional contacts? The majority of Quaternary deposits mapped along the western shoreline of Hood Canal are heterogeneous mixtures of coarse-grained sand and gravel with minor silt and clay components (Carson, 1976a, b). This type of deposit is generally unaffected by post-depositional compaction (Maltman, 1994). Overlying these coarser deposits are finer-grained marine, ice-marginal delta and lake deposits, and organic-rich sediments. These finer deposits may experience considerable compaction and dewatering (Maltman, 1994). The surface expression of this gravitational compaction may be the closed contour bathymetric depressions noted at the distal edges of the modern river and remnant Fulton Creek and Eldon Creek ice-marginal deltas (Figure 20 and Figure 21). Relatively high sedimentation rates due to riverine input, delta slumping, and the remobilization of remnant glacial deposits in these areas likely contributes toward dynamic loading. These bathymetric depressions may also be tied to a much broader interplay between Hood Canal’s existing bathymetry and bedrock topography (Figure 33).
In areas dominated by crustal extension McCalpin (1996) notes that erosional or depositional contacts are relatively easy to distinguish from normal faults. Relative to erosional contacts, normal faults invariably have steeper slopes (55-90º), are generally much straighter, and steepen upward.
In areas dominated by thrust faulting, fault dip angles can also vary widely. Studies in subaerial environments show that high-angle thrust faults (> 45º) are often interspersed with folds, tend to be segmented in regional fault zones, and typically form scarps steeper than the faulted material’s angle of repose (McCalpin, 1997). Thus, erosional or depositional contacts are typically of a shallower angle than those observed in normal, thrust, and near-vertical strike-slip fault zones.
Are the observed faulting patterns seen in sheared Quaternary sediments infilling Hood Canal coseismic (instantaneous) or post-seismic (delayed response) in origin? Simply being able to recognize such faulting patterns at a seismic record resolution threshold of 2.3 m/pixel suggests that they may be associated with a large seismic event. McCalpin (1996) points out that earthquake-induced deformation preserved in the geologic record is usually a result of large (> M 6.5) or great (> M 7.8) earthquakes. If significant late Holocene fault displacement did occur along crustal structures underlying Hood Canal or the Seattle Fault, this movement should be recorded in the sediments above.
If the Seattle fault displacement about 1100 years ago was accompanied by a M 7.0 earthquake (Pratt et al., 1997), it is reasonable to infer that considerable seismic energy propagated to the 10-400 m-thick unconsolidated and semi-consolidated sediment in the Hood Canal. At 0.7 km below mean sea level, these deposits are saturated with the marine waters of Hood Canal. In this subaqueous environment, cyclic loading caused by propagating earthquake seismic waves add to the shearing stress acting on these sediments. This can cause over-pressurization, a reduction in sediment strength, and ultimately failure. This can occur even in relatively dense sediments by a process termed ‘cyclic mobility’ (Maltman, 1994).
Cohesionless gravel deposits along active faults typically form discrete fractures patterns while cohesive materials (i.e. finer grained materials) generally form monoclines when ruptured (McCalpin, 1996). In unconsolidated alluvium, an approximate analogue to sediments in Hood Canal, it has been shown that earthquake-induced fault ruptures tend to propagate preferentially along pre-existing planes of weakness (McCalpin, 1996). Discrete faulting patterns and folds recognized on the seismic profiles (e.g. trackline 41; Figure 27) might be the result of multiple earthquakes.
The geomorphic relationship between the bathymetric and bedrock topography lows (Figure 33) may indicate bedrock displacement and settlement is taking place at a greater rate than sediment can infill and effectively hide its bathymetric expression. Dynamic interactions with near-bottom tidal currents and the existing bottom topography also may play an active role in scouring out bathymetric lows and redistributing sediments in Hood Canal. Their relative influences need to be considered as well.
Blind faults are observed throughout the dataset terminating 10-50 m below the sediment-water interface. This relatively thin fault-capping sediment layer is assumed to be late Holocene in age, an assumption supported by Carpenter et al.’s (1984) Puget Sound regional sedimentation study. Their study included two core sites in the Dabob Bay area. Of these two sites, the southern core was taken 2 km north of trackline 45 and at a water depth of 190 m. Pb210 derived sedimentation rates from this site yielded a maximum of 2.4 cm/yr. This rate was the highest recorded in Dabob Bay, and compares well with the highest sedimentation rates of > 4.0 cm/yr from the deeper main basin of Puget Sound basin. The 2.4 cm/yr sedimentation rate can at best represent sedimentation rates over a period spanning the last 100 years (M. L. Holmes, personal communication, 1997). During the Holocene it is likely that this rate was much higher due to increased submarine slumping, shoreline landsliding of unconsolidated glacial sediments, and riverine deltaic sediments. A more representative sedimentation rate for Dabob Bay during the late Holocene might be in the range of 2 to 5 cm/yr.
An important question is whether this dataset can provide paleoseismic evidence for either catastrophic and/or smaller-scale fault displacement in Hood Canal. A late Holocene coseismic faulting model can be proposed using blind faults observed on the seismic records, the estimated sediment thickness covering their upward termination, and an estimated sedimentation rate.
The trackline 44 seismic profile, oriented perpendicular to the projected Seattle Fault, was used to create a generalized fault sequence (A-C) diagram shown in Figure 36. The upper drawing (Figure 36A) shows what the high relief bedrock erosional surface may have looked like following the completion of post-glacial isostatic
rebound about 9-10 ka (Thorson, 1989). Figure 36B illustrates the vertical propagation of bedrock faulting through the overlying glacial deposits during the M 7.0 earthquake that accompanied displacement on Seattle Fault 1100 ca. It is assumed in Figure 36B that during the Seattle Fault rupture or shortly thereafter, shallow-terminating blind faults would have broken or terminated very close to the sediment-water interface. Figure 36C represents the interpreted trackline 44 seismic record (Figure 28) which shows near-surface blind faults covered by a 10-50 m layer of sediment. To temporally constrain the deposition of this sediment layer the estimated late Holocene sedimentation rate of 2-5 cm/yr can be used. Over a period of 1100 years, this rate would yield a thickness of 22 to 55 m (2-5 cm/yr ´ 1100 yrs). This closely reflects the actual thickness of sediment overlying blind faults on the trackline 44 seismic record (Figure 28).
Click on image for larger version
Figure 36: Diagram showing possible faulting sequence. A ¾ Faulted bedrock and bathymetric relief following 9 to 10 ka post glacial isostatic rebound; B ¾ Coseismic faulting following 1100 yr Seattle Fault earthquake; C ¾ Present-day Hood Canal following deposition of 10 to 50 m layer of glacio-marine sediments. This figure modeled after trackline 44 (07:06 to 07:22; Figure 28).
Using the simplified parameters of this model, coseismic or post-seismic deformation is inferred to have occurred along Hood Canal during the 1100 year Seattle fault earthquake. In fact, these near surface faulting patterns may represent more recent late Holocene Seattle fault movement (Figure 37) or seismicity associated with a large Cascadia subduction zone earthquake. This inference would be further supported if the sedimentation rates in Hood Canal were much higher than 5 cm/yr.
Figure 37: Seattle Fault splays? This figure is the same bedrock surface plot shown in Figure 33 rotated CW to allow a view NW toward the Dosewallips River delta and inline with a variety of possible projections for the Seattle Fault (SFZ). The arrow indicates a zone of likely crossing points. Trackline 44 (Figure 28) is detailed in the upper trackline plot and runs along the near edge of the lower surface plot.
