[Next Section] [Previous Section] [Contents]

Regional Geology

The Puget Lowland lies approximately 230 km east of the active Cascadia subduction zone and within a broad forearc basin that extends south from the Fraser Lowland of British Columbia to Oregon’s Willamette Basin (Figure 5). Bordering the Puget Lowland to the west are Paleocene to late Eocene rocks of the Coast Range Volcanic Province (CRVP); to the east are the Mesozoic and Paleozoic terranes of the active Cascade Range (Figure 8). The Coast Range Volcanic Province extends 650 km north-south along the Oregon-Washington-British Columbia coastline and includes thick sections of submarine tholeiitic and sub-aerial basalt flows with minor sedimentary interbeds (Wells et al., 1984; Babcock et al., 1992). The total magmatic volume of this province is estimated at approximately 250,00⁰ km3 and rivals that of the world’s great continental flood basalt provinces and rift zones (Babcock et al., 1992). With a total thickness estimated to exceed 25 km (Pratt et al., 1997), volcanic flow sequences in western Washington’s Coast Range make the most voluminous contribution to this magmatic pile.

Click on image for larger version

Figure 8: Geology map of western Washington. The expansive 100 km-wide low-gradient plain of unconsolidated glacial outwash and alluvial deposits (Qg) define the Fraser glacial maximum (Vashon Stade; 15 ka). Tertiary basalt of the Crescent Formation (lTv) is shown lining the western shoreline of Hood Canal. East of the Puget Lowland are the diverse suites of volcanic rocks comprising the Cascade Range (modified from Schuster, 1992).

Late Paleocene to middle Eocene (62-48 Ma) submarine and subaerial basalt flows of the Crescent Formation (Figure 9) are among the oldest basalt sequences in the Coast Range. Along with the Metchosin Igneous Complex of British Columbia (Figure 5) and Oregon’s Roseburg and Siletz River Volcanics, this formation makes up the Coast Range ‘basement’ and represents the earliest of three main eruptive periods (Snavely and MacLoed, 1974). The Crescent Formation underlies most of western Washington, and based on gravity and aeromagnetic studies, is thought to comprise a continuous crustal block that extends eastward from the Olympic Peninsula to the longitude of Seattle, Washington (Finn, 1990; Finn and Stanley, 1997). The easternmost edge of this crustal block lies concealed beneath thick glacial deposits and forms a major north-trending structural boundary separating rocks from the Coast Range and Cascade terranes. This boundary is approximately located by juxtaposing fault lines of the north-northwest trending ‘Coast Range Boundary’ and ‘Southern Whidbey Island’ faults (CRBF and SWIF respectively; Figure 6).

Click on image for larger version

Figure 9: Composite stratigraphic section from the Dosewallips River Valley along Hood Canal’s western shoreline (see Figure 4). It is located approximately perpendicular to the strike of the thickest Crescent sequences and flows are oriented near vertical with tops facing east. The thickness of sills and sedimentary interbeds are exaggerated. Babcock et al. (1992) indicate that thickening of the 16 km-thick flow sequence has likely occurred, however no field evidence for repetition was found. (modified from Babcock et al., 1992).

The second period of Coast Range volcanism is comprised of lower Eocene-to-early Oligocene alkalic basalts with minor andesites and dacites. This volcanic group includes the 40 Ma Goble Volcanics of southwestern Washington along with Oregon’s Tillamook Volcanics (44 Ma) and Yachats Basalt (37-34 Ma; Barnes and Barnes, 1992). In the northern Coast Range, dacitic intrusives and volcanic rocks erupted onto older basement rocks signifying early development of the Cascadia arc at 42 Ma (Wells et al., 1984). The third period of early to middle Oligocene (35-30 Ma) igneous activity was generally confined to the Oregon Coast Range and consisted of ‘end-member’ basalt flows of alkalic camptonite and intrusions of syenite and ferrogabbro (Wells et al., 1984).

Prior to Babcock et al.’s (1992) study of the Crescent Formation, the Coast Range basement was interpreted as a line of seamounts which erupted onto oceanic crust and were subsequently accreted to the continental margin during the mid-to-late Eocene (Snavely and MacLoed, 1974; Wells et al., 1984). However, this allochthonous accretion model has some difficulties; particularly explaining the occurrence of inter-bedded continentally derived sediments and the production of anomalously thick Coast Range volcanic sequences over a relatively short period of 5-9 million-years.

Babcock et al. (1992) conducted paleomagnetic and geochemical studies on a 16 km-thick composite section of Crescent basalt exposed along western Hood Canal’s Dosewallips River Valley (Figure 4). They used convincing arguments to show that the Coast Range basement volcanic sequence did not originate as a seamount chain, but rather was erupted into a subsiding basin formed by rifting along the margin of North America. Rifting was initiated about 60 Ma following increases in the velocity and obliquity of Kula-Farallon plate convergence and provided conduits for the mantle plume-driven extrusion of the anomalously thick Coast Range basement sequence. Zones of pillow basalt and interbedded marine sedimentary rocks observed in the uppermost Crescent Formation resulted from lower volume eruptive stages coincident with subsidence of the rifted basin (Babcock et al., 1992). Wells et al. (1984) were not convinced that a continental margin rifting model could explain observed compressional deformation structures in the southern Oregon Coast Range, which they attributed to island arc accretion. However, Babcock et al. (1992) found no evidence in the Crescent Formation’s lower member (Figure 9) for intense compressional-transtensional deformation or ophiolite lithologies, which would be expected in oceanic crust subjected to extensive folding and thrusting during tectonic accretion. In fact, paleomagnetic data from rocks in the northern Coast Range revealed no components indicating appreciable northward translation or rotation (Babcock et al., 1992). Snavely and Wells (1996) have since refined their model for the origin of the Coast Range, however they make no mention of the Babcock et al. (1992) study.

Marine sedimentary and volcanic rocks of the Olympic Peninsula cover nearly 15,000 km² of northwestern Washington. They reach elevations of about 2500 m, and form a crude triangle bordered by the Pacific Ocean to the west, Hood Canal to the east, and the Strait of Juan de Fuca to the north (Figure 10). The peninsula has two major geologic terranes, the ‘peripheral’ and ‘core’ rocks which are distinguished primarily by lithology (Figure 10) (Tabor and Cady, 1978). Making up the horseshoe-shaped belt of peripheral rocks are thick sequences of lower-to-mid Eocene Crescent basalt overlain on the north, east, and south by upper Eocene to Miocene and minor Pliocene marine sedimentary rocks. The western and eastern core rocks (Figure 10) include Eocene to middle-Miocene marine sedimentary rocks with minor interbeds of pillow basalt, some of which are metamorphosed to prehnite-pumpellyite and greenschist facies (Babcock et al., 1992). The core rocks are in general stratigraphically continuous and decrease in age westward (Tabor and Cady, 1978).

For at least the last 40 m.y. the mountainous Olympic core rocks have been subjected to accretionary prism-style deformation; a direct result of underthrusting by the Juan de Fuca plate beneath North America (Wang et. al., 1995). Horizontally shortened structures are pervasive throughout the core rocks, particularly in the eastern core region where zones of anastomosing, steeply-dipping folded thrust faults are over-turned, and sheared bedding planes are present (Figure 10). The core rocks are separated from the peripheral rocks by the near-vertical Hurricane Ridge thrust fault (Figure 10). This fault defines the eastern boundary of the Olympic core terrain. Resistivity modeling by Aprea (1996) shows that this fault may extend through the entire crust. Seismic velocity studies by Dewberry (1996) and Symons and Crosson (1997; unpublished data) show that a ‘wedge’ of Olympic Peninsula core rocks may be thrust beneath peripheral Crescent basalt (Figure 11). The emplacement of this lower density and more buoyant rock wedge under the eastward dipping Crescent basalt may contribute to tectonic stresses observed in the Cascadia forearc.

Click on image for larger version

Figure 10: Geologic terranes and structure of the Olympic Mountains. Deformational structures are shown increasing eastward. Along the western shoreline of Dabob Bay sedimentary cover rocks pinch-out exposing outcrops of Crescent basalt. Holocene faults NE of Lake Cushman mapped by Wilson et al. (1979) are not shown (modified from Tabor and Cady, 1978).

Relative to the marine sedimentary Olympic core rocks, the peripheral Crescent Formation is a mechanically strong lithologic unit. This can be inferred from its great stratigraphic thickness, low historical seismicity (Dewberry, 1996), and general lack of deformational structures (Figure 10). However, this volcanic sequence endured Paleogene development of the Olympic accretionary complex and associated isostatic uplift of the Olympics from about 40 Ma to present, the combination of which undoubtedly caused compensating fault structures to develop. Hood Canal may define the western limit of long-lived fault structures and the start of more geologically recent deformation to the east where Crescent Formation underlying the Puget Lowland was subject to periodic glacial loading throughout the Quaternary.

Click on image for larger version

Figure 11: 3-D seismic velocity structural model of Puget Sound using about 3000 well located > M_c 2.0 1980-1996 earthquakes. Both map and cross section use 4 km horizontal by 2 km vertical grid elements. The shallow depth map (upper left) and A-A¢ cross section show a westward transition from the Seattle-Tacoma low velocity sediments to the higher velocity Eocene Crescent Formation volcanics. The Olympic core rock wedge (lower cross section) may be thrust beneath older Crescent volcanics or reflect modified rock properties in the subducted slab (modified from Symons and Crosson, 1997; University of Washington Geophysics Department, unpublished data).

Cordilleran Ice Sheet advances into western Washington occurred at least five times during the late Quaternary (Thorson, 1996). Fed by alpine glaciers in the Canadian Coast Range, the tidewater Juan de Fuca lobe advanced westward into the Strait of Georgia and the Strait of Juan de Fuca (Figure 5). Eventually the coalescing ice blocked this seaway and advanced south as the Puget lobe completely filling the Puget Lowland between the Olympic Mountains and Cascade Range (Figure 5). The modeled geometry and volume of the Puget ice lobe during the late Pleistocene Vashon Stade glacial maximum 15 ka is well studied (e.g. Thorson, 1980; Booth,

1987). Puget ice lobe reconstructions show glacial ice reached a maximum thickness of about 1.3 km (assuming a 200 m isostatic depression), an estimated volume of 10⁴ to 10⁵ km³, and had a terminus that extended 80 km south of Seattle, Washington (Figure 12) (Thorson, 1980; Booth, 1987). A modern temperate glacier analogue might be the broad piedmont Malaspina Glacier of Alaska. While the Puget lobe occupied the Puget Lowland, it generated large volumes of glacial meltwater, fast channelized ice flow, and during its retreat vast ice-marginal lakes (Thorson, 1980). Combined, these hydrologic components promoted the rapid removal of sediment fill and extensive erosion of sub-glacial topography (Mullins and Hinchey, 1989; Eyles et al., 1990; Booth and Hallet, 1993). The deep and relatively narrow glacial trough of

Click on image for larger version

Figure 12: Ice marginal limits of the Puget Ice lobe at about 15 ka. Contours denoting post-glacial uplift (upper figure) are derived from geodetic data along remnant proglacial lake shorelines formed during deglaciation of the Juan de Fuca/Puget lobe system. Note the asymmetrical uplift along B-B¢ (lower figure) and relative position of Hood Canal about 5 km east of the western ice margin. Net uplift assumes a 200 m isostatic crustal depression (modified from Thorson, 1989).

Hood Canal lay both along the former western margin of the Puget lobe and was generally aligned with regional ice flow. These two spatial parameters would have subjected pre-existing glacial outwash plain deposits and Tertiary bedrock along Hood Canal to concentrated sub-glacial erosion. Erosional troughs resembling Hood Canal in other glaciated regions of North America, such as southeastern Canada’s intermontane fiord lakes (Mullins et al., 1990) and New York State’s finger lakes (Mullins et al., 1997), are thought to have been formed primarily by subglacial erosion.

Deglaciation of the Juan de Fuca-Puget ice lobe system was rapid (an estimated 300 m/yr), completed by 13.5 ka, and left behind an extensive low-gradient outwash plain referred to as the "great Lowland fill" by Booth (1994). During glacial unloadingcrustal blocks located along pre-existing structural discontinuities, particularly those oriented perpendicular to ice flow directions and parallel to ice margins, probably experienced failure (Thorson, 1996). Present-day Hood Canal, oriented parallel and just up-glacier from the former Puget ice lobe’s western margin (Figure 12), would have been ideally suited for the maximum expression of postglacial fault displacement (Thorson, 1996). Thorson (1989) used geodetic data from remnant proglacial lake shorelines to calculate net post-glacial uplift, which was complete throughout most of the Puget Lowland by 9 to 10 ka. An east-west profile of net vertical uplift (tectonic and isostatic) versus distance from the 15 ka western Vashon limit shows net uplift decreasing asymmetrically toward the ice margins (Figure 12). This difference in east-west isostatic response (up doming) is only partially explained by more rapidly thinning glacial ice to the east and probably indicates significant crustal heterogeneity or post-isostatic tectonic warping (R. Thorson, 1997; personal communication).

Current tectonic uplift in western Washington is apparently unrelated to post-glacial rebound and driven primarily by crustal shortening/thickening in the direction of Juan de Fuca plate convergence. Uplift varies across the Olympic Peninsula from about 4 mm/yr on the Pacific Coast to 0 mm/yr near Seattle (Savage et al., 1991).

Today, the Puget Lowland displays two general zones of seismicity (Figure 7). The first, a relatively shallow (£ 20 km) and diffuse zone within the North American plate; the second, a deeper (40-70 km) zone within the Juan de Fuca plate which is obliquely subducting at a rate of 40 mm/yr in a direction of N68° E (Savage et al., 1991). Beneath Puget Sound crustal earthquakes are generally confined to the glaciated area with a strong correlation between earthquake depth and glacial loading (Thorson, 1996). Thorson (1996) hypothesizes that the present interglacial may have been seismically inactive. Furthermore, stresses in the Puget Lowland may be returning to pre-glacial ‘background’ levels where the transfer of subduction zone stress was not modulated or stabilized by periodic glacial ice loading. Late Holocene fault displacement along the Seattle fault zone may be evidence for this. Aprea et al. (1998) point out that crustal seismicity occurs almost solely within the Crescent Formation volcanics underlying the Seattle basin (Figure 7). They propose that this ³ 20 km-thick volcanic sequence may effectively prevent the subduction of Olympic core rocks, contribute to the Puget Lowland’s historical pattern of crustal seismicity, and possibly cause a coupling of the Juan de Fuca and North American plates (Figure 7).

Although typically represented on small-scale geologic maps as a single fault trace (e.g. Figure 6) the Seattle fault zone consists of four parallel, south-dipping, high-angle, generally east-west trending fault strands with dominantly reverse displacement (Johnson et al., 1994). This fault shows both direct and paleoseismic evidence for up to 7 m of uplift near Restoration Point during the late Holocene (Figure 13) (Bucknam et al., 1992). The estimated M > 7.0 earthquake (Pratt et al., 1997) that accompanied this displacement about 1,100 years ago generated numerous coseismic geologic features in the Puget Sound area (Figure 13). These included tsunami deposits in Puget Sound (Atwater and Moore, 1992), rock avalanches in the Olympic Mountains (Schuster et al., 1992), anomalous lake turbidites in Lake Washington (Karlin and Abella, 1996), and forests submerged during large coseismic landsliding along Mercer Island (Figure 13) (Jacoby et al., 1992).

Late Holocene faulting has also been documented along the Saddle Mountain East fault located approximately 5 km west of Hood Canal’s western shoreline (Figure 3 and Figure 13). This fault trends N22° E, dips 75° SE, and is one of several crustal faults in this area thought to have developed during complex faulting and doming of the Olympic Peninsula (Wilson et al., 1979). A 1.8 km-long scarp along the Saddle Mountain East fault exhibits 3.5 m of reverse dip-slip offset dated at 1,160 ¹⁴C yrs B. P. and possibly concurrent with displacement along the Seattle fault zone (Wilson et al., 1979).

[Next Section] [Previous Section] [Contents]