Abstract

The Tualatin basin, west of Portland (Oregon, USA), coincides with a 110 mGal gravity low along the Puget-Willamette lowland. New gravity measurements (n = 3000) reveal a three-dimensional (3-D) subsurface geometry suggesting early development as a fault-bounded pull-apart basin. A strong northwest-trending gravity gradient coincides with the Gales Creek fault, which forms the southwestern boundary of the Tualatin basin. Faults along the northeastern margin in the Portland Hills and the northeast-trending Sherwood fault along the southeastern basin margin are also associated with gravity gradients, but of smaller magnitude. The gravity low reflects the large density contrast between basin fill and the mafic crust of the Siletz terrane composing basement. Inversions of gravity data indicate that the Tualatin basin is ∼6 km deep, therefore 6 times deeper than the 1 km maximum depth of the Miocene Columba River Basalt Group (CRBG) in the basin, implying that the basin contains several kilometers of low-density pre-CRBG sediments and so formed primarily before the 15 Ma emplacement of the CRBG. The shape of the basin and the location of parallel, linear basin-bounding faults along the southwest and northeast margins suggest that the Tualatin basin originated as a pull-apart rhombochasm. Pre-CRBG extension in the Tualatin basin is consistent with an episode of late Eocene extension documented elsewhere in the Coast Ranges. The present fold and thrust geometry of the Tualatin basin, the result of Neogene compression, is superimposed on the ancestral pull-apart basin. The present 3-D basin geometry may imply stronger ground shaking along basin edges, particularly along the concealed northeast edge of the Tualatin basin beneath the greater Portland area.

INTRODUCTION

The Tualatin basin, west of Portland (Oregon, USA), is one of several structural basins within the Puget-Willamette lowland, a 600-km-long structural and topographic trough between the Coast Range and Cascade volcanic arc (Fig. 1). The Puget-Willamette lowland is seismically active and home to most of the Oregon and Washington State populations and major cities. In Puget Sound, active faulting accommodates margin-parallel shortening (Nelson et al., 2003; Brocher et al., 2004) driven by the northward motion of Oregon and Washington against Canada (Wells et al., 1998; McCaffrey et al., 2007). Although the level of seismicity is much lower in northwest Oregon, the Tualatin basin and northern Willamette basin appear to be responding to the same north-south compression. Beeson et al. (1985) argued that northwest-striking, dextral strike-slip faults played an important role in the tectonic evolution of the northern Willamette basin. These structures are difficult to map, however, because much of the region is covered with Missoula flood deposits that largely conceal geomorphic and geologic evidence of tectonism older than 15,000 yr. Yeats et al. (1996), Popowski (1996), and Wilson (1998) used seismic reflection and well data to interpret shallow (upper few kilometers) structure within the Tualatin basin, particularly faults and folds affecting the Columbia River Basalt Group (CRBG) and younger fluvial deposits. Popowski (1996) used potential field and seismic reflection data to interpret a two-stage history of the Tualatin basin, with early, left-lateral shear accommodating block rotation and basin formation. In Blakely et al. (2000), high-resolution aeromagnetic data were used to infer that the Tualatin basin is bounded by northwest-trending dextral strike-slip faults.

Here we describe a three-dimensional (3-D) model of the Tualatin basin based on more than 3000 newly acquired gravity measurements in the Tualatin, Portland, and northern Willamette basins (Morin et al., 2007). We show that basin geometry is consistent with an early history of transtensional extension between two overlapping, northwest-striking strike-slip fault zones.

GEOLOGIC SETTING

The Tualatin basin is between the Cascade volcanic arc and the Cascadia subduction zone (Fig. 1A). The basement beneath the basin is not well known, but likely consists mostly of the Siletz terrane, an oceanic plateau or seamount chain accreted to the continental margin ca. 50 Ma (Duncan, 1982). The Siletz terrane is exposed in the Coast Range west of the Tualatin basin (Fig. 1B), where it is composed of late Paleocene to middle Eocene oceanic basalt flows, intrusions, and breccias of the Siletz River Volcanics (Snavely et al., 1968; Wells et al., 1995, 2000). Seismic data show that the Siletz terrane consists of high-velocity (and likely dense) mafic crust 20–35 km thick beneath the Coast Range and Willamette basin ∼75 km south of the Tualatin basin (Tréhu et al., 1994).

East of the Tualatin basin, basement is more difficult to define due to the lack of exposures and overlying volcanics of the Cascades volcanic arc. Our best candidate for pre-Oligocene basement east of the Tualatin basin is the middle to late Eocene basalt of Waverly Heights exposed in a small area at the southeast end of the Portland Hills, between Portland and Oregon City (Fig. 1B). Waverly Heights basalts are also thought to have been part of an oceanic island accreted to western Oregon during the Eocene (Wilson, 1998; Beeson et al., 1989a, 1989b) and may be part of the Siletz terrane.

We also include the voluminous, middle to late Eocene Tillamook Volcanics (Fig. 1B) in our basement assemblage because of their compositional similarity to underlying Siletz terrane (Wells et al., 1995). The Tillamook Volcanics form a >3-km-thick sequence of dominantly basalt flows and breccias with minor siltstone and conglomerate (Wells et al., 1995) and are locally separated from the underlying Siletz terrane by thin sedimentary interbeds.

As a group, the Eocene Siletz terrane, Waverly Heights basalt, and Tillamook Volcanics are considerably denser than overlying Eocene and younger sediments that compose the Tualatin basin fill. Several hydrocarbon exploration wells and industry seismic data show that at least 2 km of Eocene and Oligocene marine sedimentary rocks fill the Tualatin basin (Oregon Department of Geology and Mineral Industries, 2013; Popowski, 1996). Two of the deeper wells in the Tualatin basin, the Texaco Cooper Mountain #1 and the Nahama and Weagant Klohs #1 (Table 1), show ∼700 to 1100 m of marine sedimentary rocks overlying tens to hundreds of meters of Eocene mudstone and siltstone, and then an additional 500 to 1300 m of tuffaceous mudstone, siltstone, and minor sandstone of the Yamhill Formation (Yeats et al., 1996; Oregon Department of Geology and Mineral Industries, 2013). The Texaco Cooper Mountain #1 well encountered Siletz River Volcanics at ∼2600 m (Table 1). CRBG flows entered northwestern Oregon between 16 and 14.5 Ma (Yeats et al., 1996; Beeson et al., 1989a) and range in thickness from a few meters to 300 m (Newton, 1969; Oregon Water Science Center, 2013) within the Tualatin basin and surrounding uplands (Fig. 1). In the Tualatin basin, CRBG flows occupy only a small fraction (maximum of 300 m) of the total vertical section of post-Eocene basin fill (minimum of 2 km).

The CRBG in the Tualatin basin is overlain unconformably by late Miocene and younger fluvial siltstone, sandstone, and mudstone, and the Hillsboro Formation of Wilson (1998), deposited under low-energy conditions (Yeats et al., 1996; Wilson, 1998), creating an extensive, relatively flat plain. Locally, Pliocene and Pleistocene Boring Lava were erupted from vents in the Portland basin, Tualatin basin, and northernmost Willamette basin between ca. 3 and 1 Ma (Evarts et al., 2009; Ma et al., 2012; Madin et al., 2008). The youngest deposits in the basin consist of several meters of rhythmically bedded fine sand and silt deposited by numerous Missoula floods, which entered the basin through the Lake Oswego flood channel (Trimble, 1963; Schlicker and Deacon, 1967) at 22–15 ka (O’Connor and Benito, 2009; Madin et al., 2008).

Late Neogene folding uplifted the CRBG around the Tualatin basin margin, producing the Portland Hills (formally the Tualatin Mountains) to the northeast, Parrett Mountain to the south, the Chehalem Mountains to the southwest, and the Coast Range to the northwest (Fig. 1B). In the center of the basin, the top of the CRBG is ∼300 m below sea level (Wilson, 1998). Post-CRBG folds within the basin (e.g., Cooper Mountain, Fig. 1B) have east-west–trending axes, consistent with present-day north-south maximum horizontal compressive stress (Yeats et al., 1996; Werner et al., 1991). A major east-west structure within the Tualatin basin, the Beaverton fault, extends ∼30 km east-west through the middle of the basin and may consist of more than one strand (Yeats et al., 1996), although it was originally mapped as a single, down-to-the-north fault (Madin, 1990). The westward extent of the fault is poorly constrained. The northeast-striking Sherwood fault bounds the southeastern margin of the Tualatin basin along the northern front of the Parrett Mountain homocline. In the Portland Hills to the northeast, the northwest-striking Sylvan-Oatfield, Portland Hills, and East Bank faults (Ma et al., 2012) accommodate post-CRBG deformation (Blakely et al., 1995, 2004; Walsh et al., 2011). Boring Lava exposures (Fig. 1) are located along this northeastern margin.

The northwest-trending Gales Creek fault bounds the Coast Range uplift (Schlicker and Deacon, 1967; Beeson et al., 1985; Wells et al., 1995, 2009) and marks the southwestern boundary of the Tualatin basin (Fig. 1). This fault may connect with the Mount Angel fault zone to the southeast, forming a structural zone more than 100 km long (Beeson et al., 1985; Blakely et al., 2000). Geologic mapping suggests that the Gales Creek fault is a northwest-trending, steeply west dipping fault that has accommodated dextral slip (Wells et al., 1995, 2009). Offset magnetic anomalies across northwest linear gradients show evidence for ∼10 km of dextral slip on the Gales Creek–Mount Angel fault zone (Blakely et al., 2000). Beeson et al. (1985) argued that the Gales Creek structural zone and the parallel Portland Hills–Clackamas River structural zone to the northeast are loci of significant dextral shear. Motion on these structures was a factor in the formation of the Portland and Tualatin basins in the middle Miocene, albeit direct evidence for the existence of major strike-slip faults in the Portland Hills was minimal. Yeats et al. (1996) and Popowski (1996) suggested that reactivation of the Gales Creek fault system and new faulting resulted in the deepening of the Tualatin basin after the emplacement of the CRBG. On the basis of gravity data, we argue for the existence of an ancestral Tualatin basin that involved significant pre-Miocene faulting on the Gales Creek fault.

We define the Tualatin basin in a geophysical sense, as a tectonic depression floored by Eocene, largely basaltic basement, covered with younger sedimentary deposits and volcanic rocks. We recognize that our simple definition encompasses a complex geologic framework. Basement, for example, includes the Siletz terrane, Waverly Heights basalt, and Tillamook Volcanics, generally separated from each other by thin sedimentary layers, presumably lower in density. Basin-fill lithology is similarly complex, including sedimentary rocks and basalt flows, notably flows of the CRBG. Although the CRBG may be as dense as Eocene basement, these flows are thin (<300 m) relative to the total vertical section of basin fill (>2 km). Thus, we argue that the largely volcanic basement lithology is higher in density than the largely sedimentary overlying basin fill, and that this density contrast is responsible for the Tualatin basin gravity anomaly described in the next section.

GRAVITY DATA AND INTERPRETATION

Since 2004, we have systematically acquired more than 3200 gravity stations in the Tualatin and northern Willamette basins, more than tripling the existing coverage and providing an average spacing of 1 station per 2.6 km2 over the region (Fig. 2). Gravity data used in this study as well as detailed information about the collection and processing of the data were provided by Morin et al. (2007). Observed gravity values were reduced to isostatic anomalies using standard formulas (Blakely, 1995).

Isostatic gravity values reach +72 mGal over Eocene basement rocks to the west of the Tualatin basin (Fig. 2). East of the Tualatin basin, substantially lower values of +8 to +12 mGal occur over sparse Eocene basement exposures composed of Waverly Heights basalt. We note, however, that the Tualatin basin is within a significant west-east gravity gradient that extends more than 600 km along the length of the Oregon and Washington Coast Range. As discussed herein, this regional west-east gradient has a number of explanations unrelated to basement exposures. The Tualatin basin, with low-density basin-filling sedimentary rocks, coincides with a large gravity low (values of -36 to −44 mGal; Fig. 2) superimposed on the broad 70 mGal, west-east gravity gradient. The lowest gravity values (−44 mGal) occur in the central part of the Tualatin basin, north of the Beaverton fault (Fig. 2).

Gravity data clearly delineate the southwest, southeast, and northeast margins of the Tualatin basin and suggest a rhombic basin shape (Fig. 2). The steepest gravity gradient in northwest Oregon (∼100 mGal over ∼25 km) extends for more than 40 km coincident with the Gales Creek fault. An apparent kink in this sharp gradient occurs at the northwest corner of the Tualatin basin, where the gradient broadens and weakens to the north (Fig. 2). The steep gravity gradient also weakens to the south of the Tualatin basin, where it spans a wide western margin of the northern Willamette basin. A broad, northeast-trending gravity high associated with the Sherwood fault separates the Tualatin basin from the northern Willamette basin to the southeast. This relative gravity high (values of -16 to -12 mGal), suggestive of a concealed, fault-bounded basement ridge, coincides with the Parrett Mountain uplift. The northern Willamette basin is characterized by gravity values of ∼-24 mGal, suggesting that it is shallower than the Tualatin basin, which has gravity values of −44 mGal. The Portland Hills near the northeast margin of the Tualatin basin also correspond with low gravity values (-28 mGal), suggesting that dense basement rocks are not a significant part of the Portland Hills uplift. The gravity low extends northeastward into the Portland basin, with modest gravity values (∼-24 mGal) similar to those in the northern Willamette basin.

Gravity data do not define a distinct northwestern boundary of the Tualatin basin, but rather show subtle gravity highs and lows along post-CRBG anticlines and synclines (Dinterman and Duvall, 2009; Madin and Niewendorp, 2009; R.E. Wells, 2010–2011, field observations).

3-D BASIN GEOMETRY

We inverted the gravity data to estimate the shape of the basement surface for the Tualatin basin and surrounding areas (Fig. 3A). The inversion method (modified from Jachens and Moring, 1990) uses an initial basement gravity field, calculated from gravity measurements on exposed Eocene basement rocks, and an assumed variation of the density of younger basin deposits with depth. This inversion method is useful for predicting the relative shapes of basins, although determination of absolute basement depths is less certain because of uncertainties in the local density-depth relation for overlying sediments, and because basin-fill density is assumed to vary only in the vertical direction (Phelps et al., 1999). A larger issue in our case is the sparse basement exposures (critical constraints for the separation of the basement and basin-fill gravity fields) on the east side of the basin. Our inversion depends on the few exposures of Waverly Heights basalt between Portland and Oregon City (Fig. 1B).

We used three wells that intersected basement and a fourth well that bottomed above basement to help constrain the basement inversion (wells 1–4, Fig. 3; Table 1). In addition, we used a density-depth profile (Table 2) based on velocity-density relationships calculated from sonic velocity logs from the Nehama and Weagant Klohs #1 oil and gas well (Oregon Department of Geology and Mineral Industries, 2013) located in the Chehalem Mountains (well 5, Fig. 3).

Complicating our inversion is the large west-east gravity gradient across the Tualatin basin, as exemplified by large differences in isostatic gravity values at our three basement wells. There are a number of explanations for the west-east gravity gradient. (1) Some of the observed gravity gradient is caused by the Juan de Fuca plate, which is at 44 km depth beneath the study area and dips eastward ∼20° (Romanyuk et al., 1998). (2) The mantle wedge beneath the Puget-Willamette lowland may be significantly serpentinized (Bostock et al., 2002; Blakely et al., 2005). (3) At Roseburg and in Puget Sound, Siletzia appears to be obducted onto continental basement, producing an eastward-declining gradient (Wells et al., 2000; Anderson et al., 2008); similar conditions may exist beneath the Tualatin basin. (4) Basement composition varies from west to east in the vicinity of the Tualatin basin, as it does in the Siletz terrane from north to south (Wells et al., 1995). (5) Basement exposures in the southern Portland Hills may not root into underlying basement. However, where we can observe stratigraphic relations, high-titanium lavas like those of the Waverly Heights basalt are closely associated with oceanic basement. The contributions to gravity from thin sedimentary interbeds are thought to be small, given the modest north-south gradients in the Coast Range.

Based on the assumption that the large west-east gravity gradient across the study area (Fig. 2) is due to deep crustal and mantle effects and possibly a change in basement composition, we removed a linear east-west regional trend (Fig. 4A) based on isostatic gravity values measured on basement outcrops. The residual gravity field (Fig. 4B) was then inverted to estimate depth to basement (Fig. 3A). Removal of a regional field results in more similar basement gravity values on both sides of the basin (Fig. 3B).

The presence of concealed CRBG in the basin fill has little effect on our inversion, despite having greater density than surrounding sediments (Fig. 5). First, the total thickness of the CRBG section in the Tualatin basin is generally <300 m, a small fraction of the overall basin fill. Second, the CRBG layer is gently dipping and nearly flat-lying with respect to the overall basement geometry. Horizontal slabs produce constant gravity values, so adding a 300-m-thick, horizontal layer of CRBG to our density-depth profile would not greatly affect the determined map-view shape of the basin, although it would increase absolute depth-to-basement values slightly; the maximum depth of the central part of the basin would increase by ∼1 km due to the addition of a 300 m layer of higher density material.

The depth-to-basement inversion (Fig. 3A) shows an ∼6 km-deep asymmetric basin with a steep (∼40°) southwestern margin along the Gales Creek fault and complex basement topography. Our basin depth is several kilometers deeper than determined by Popowski (1996) on the basis of seismic reflection and well data along a few select profiles. The basin is bounded by the Gales Creek fault on the southwest. On its northeast flank, a northwest-trending basement high follows the trend of the Sylvan-Oatfield, Portland Hills, and East Bank faults. The Sherwood fault to the southeast completes a largely fault-bounded rhombic shape. The northwestern margin is broad and not associated with any major mapped faults.

The northwest-striking Sylvan-Oatfield, Portland Hills, and East Bank faults are within the northwest-trending basement high associated with the Portland Hills uplift (Fig. 3A). These faults extend northwestward into the Portland Hills gravity low, where Eocene basalt is not exposed. The Portland Hills consist of a fault-bounded anticline formed in the CRBG and underlying marine sedimentary rocks (Blakely et al., 2004; Madin and Niewendorp, 2009; Walsh et al., 2011). The low gravity values suggest that Eocene basement may not be involved in the anticline in the northern Portland Hills. The Richfield Barber #1 well (Table 1; Fig. 3A), located in the Portland Hills, extends to a depth of ∼2.4 km through mostly volcanic rocks, including significant thicknesses of volcanic agglomerate and sands. The interpretation of the top of Eocene basement is uncertain for this well; in Plate 2A of Yeats et al. (1996), basement is shown at 626 m, whereas in the text of Yeats et al. (1996), the top of the Tillamook Volcanics may be at a depth of 2124 m. According to our depth to basement (Fig. 3A), which does not incorporate the Barber well as a constraint, Eocene basement is predicted to be at 1977 m. Incorporating the Barber well with a basement top of 2124 m as a constraint does not change the inversion results significantly, with a maximum change in basement depth and basement gravity of 150 m and 0.7 mGal, respectively. However, using a basement depth of 626 m causes a significant basement gravity low compared to values measured on basement exposures to the southeast. A basement depth of ∼2 km is more consistent with gravity measurements on basement on the east side of the basin, and the dominantly volcanic section encountered in the well has densities that are more similar to the basin fill than to the Siletz River volcanics.

The deepest part of the basin is bounded on the northeast by a significant basement step that is subparallel to the Portland Hills, but has no surface expression (Fig. 3A). It may represent a buried fault or fold in the basement. Basement under the southwest Portland basin appears to be relatively shallow (∼2 km) and featureless, although our determination of basement depth is poorly constrained in this region. Paleogene Cascade Arc volcanic rocks, which are less dense than the basement assumed in the inversion, crop out to the east of the study area and may underlie much of the Portland basin.

To the southeast, the northeast-striking Sherwood fault bounds the northwest side of a broad basement ridge that separates the Tualatin basin and shallower (2.4–3 km) northern Willamette basin. The basement to the northwest of the Sherwood fault is several hundred meters lower than to the south, consistent with a previous interpretation of ∼300 m of vertical offset on the Sherwood fault that decreases to the northeast (Miller et al., 1994). Farther north, the concealed east-striking Beaverton fault cuts across the Tualatin basin, and basement drops abruptly, more than 1 km from south to north in the central part of the basin, consistent with a north-verging reverse fault.

The Tualatin basin is physiographically defined by downwarped CRBG flows (Fig. 1B; Yeats et al., 1996; Gannett and Caldwell, 1998), but gravity data show that the Eocene to middle Miocene basin fill is significantly deeper than the CRBG. We simultaneously modeled residual gravity and aeromagnetic data along a profile across the Tualatin basin, using available well and surficial geologic constraints (Fig. 5). Although the CRBG has little effect on the gravity field, it produces prominent magnetic anomalies, which we have modeled with thin, shallow magnetic layers, consistent with well data and expected CRBG magnetic properties (Fig. 5B). For simplicity, our model assumes that basement has uniform magnetization and density.

We used a 2.5-D simultaneous forward gravity and magnetic modeling program. The user varies selected parameters (depth, shape, magnetization, and density of suspected sources) in an attempt to reduce the weighted root mean square (rms) error (Fig. 5A) between the observed and calculated potential fields. The interpretation of potential field data yields nonunique solutions because an infinite number of geometrical models will have an associated field that closely matches the measured field. Our preferred model is consistent with known physical properties, well information, and surficial geology.

Our forward model indicates ∼200 m of vertical offset, down to the north, along the shallow parts of the Beaverton fault (Fig. 5A), comparable to 285 m of vertical offset observed farther to the east and north of Cooper Mountain (Yeats et al. 1996). The Gales Creek, Sylvan-Oatfield, and Portland Hills faults also bound CRBG flows of varying thickness (to ∼300 m thick) and vertically offset these flows by as much as 300 m (Fig. 5A). The 2.5-D model illustrates that the buried CRBG, despite its strong magnetic signature, constitutes a relatively minor component of the post-Eocene fill of the Tualatin basin (Fig. 5A).

DISCUSSION

The gravity inversion reveals a deep, asymmetric, fault-bounded Tualatin basin (Fig. 6). The Gales Creek and Sherwood faults are both associated with gravity gradients along the southwest and southeast margins of the basin, respectively. The northeast margin is bounded by the Portland Hills and its faults. The shape of the gravity-defined basin suggests that the Tualatin basin is a pull-apart rhombochasm bounded on the southwest by the dextral Gales Creek fault (Blakely et al., 2000; Wells et al., 2009; Beeson et al., 1989a). This origin implies dextral slip along the northeast margin of the basin. Although little evidence exists for sense of motion on the Portland Hills fault, Walsh et al. (2011) observed horizontal slickensides exposed in the TriMet light rail tunnel along the Sylvan-Oatfield fault zone, indicating a combination of thrusting and dextral motion on northwest-striking Quaternary faults. In addition, dextral offset of basement magnetic anomalies along strike to the southeast, along the Canby-Molalla fault, was interpreted (Blakely et al., 2000). Although focal mechanisms in northwest Oregon are, in general, based on small earthquakes and show a combination of strike-slip and thrust faulting (McCaffrey et al., 2007), most of the structures that accommodate north-south shortening in the region are small and discontinuous. The larger, more continuous northwest-striking structures have accommodated strike-slip motion and are thought to have significant slip, including the faults along the northeastern margin of the Tualatin basin.

The Tualatin basin inferred from our gravity inversion is six times deeper than the maximum depth of the CRBG (∼1 km) (Oregon Water Science Center, 2013), implying that Eocene basement was downfaulted substantially prior to emplacement of the CRBG ca. 15 Ma. We suggest that the Gales Creek fault played an important role in the first stage of basin formation and, consequently, has been active for more than 15 m.y., as also inferred by Yeats et al. (1996). Since its emplacement, north-south shortening (Werner et al., 1991) has folded and faulted the CRBG throughout the Tualatin basin. The east-striking Beaverton fault is likely to be a north-verging thrust or reverse fault responding to north-south maximum horizontal compressive stress of the region. The gravity defined basement shows that the Beaverton fault offsets Eocene basement >1 km down to the north, with significantly less (∼200 m) offset on the overlying CRBG, suggesting an earlier history of deformation on the Beaverton fault.

The rhombic shape of the early basin may simply reflect a releasing-bend step-over between the Gales Creek–Mount Angel and Portland Hills fault systems during basin filling between middle Eocene and middle Miocene time. Opening of the basin may have been augmented by north-northwest extension in the Coast Range during the late Eocene, indicated by abundant Eocene dike swarms (Wells et al., 1984) and normal faulting in the Mist Gas Field (Niem et al., 1994; Eriksson, 2002) located northwest of the Tualatin basin. The Beaverton and Sherwood faults thus may have originated as normal faults in the Eocene.

A new understanding of the Tualatin basin geometry may help us better understand shaking hazards from basin edge effects. Strong ground shaking during earthquakes is amplified along the margins of sedimentary basins, where basement underlies low-velocity fill at shallow depth. Previous studies (McPhee et al., 2007; Frankel and Vidale, 1992; Graves et al., 1998) have shown a correlation between 3-D basin geometry and observed shaking in earthquakes. Thus, basin edge effects in populated regions could pose a significant seismic hazard. Our basin model indicates that the greater Portland area is situated along the northeast edge of the Tualatin basin in an area where the basement drops ∼2 km over a distance of ∼5 km. In addition, preliminary site response studies between the Portland and Tualatin basins show no sign of long-period (1–2 s) seismic wave amplification for sites in the Portland basin, whereas there is an indication of basin-edge–generated surface waves at sites near the northeast edge of the Tualatin basin (Frankel et al., 2010; Art Frankel, 2013, written commun.), suggesting that the Tualatin basin has an effect on seismic wave amplification. Although the faults bounding the Tualatin basin cannot be directly associated with the diffuse seismicity in the region (McCaffrey et al., 2007), the potential for enhanced ground shaking should be further investigated with additional study, such as seismic tomography, to constrain P- and S-wave velocities for earthquake ground motion simulations, not only from upper crustal earthquakes, but also from deeper earthquakes from the subduction megathrust.

CONCLUSIONS

Inversion of gravity data shows that the Tualatin basin is a deep (∼6 km), asymmetric rhombochasm with low-density sedimentary fill overlying Eocene and older basement rocks. The existence of a large dextral fault that bounds the southwest margin and possible dextral faulting at the northeast margin suggests a right-lateral pull-apart origin. Geophysical modeling constrained by geology and well data indicates that the Tualatin basin contains at least 2–5 km of sediment older than emplacement of the CRBG at 15 Ma, and thus basin evolution began before middle Miocene time. Opening of the basin may have been augmented by north-northwest extension in the Coast Range during the late Eocene. Basin edge effects could give rise to significant seismic wave amplification, and future studies characterizing these basin effects and seismic hazards in the region are advised.

We thank Bob Morin, Karen Wheeler, and Janet Watt for their efforts in data collection, and Karen Wheeler for her help with preparation of the geologic map. Russ Evarts and Bob Jachens provided helpful comments. Reviews by Mike Cheadle, Ian Madin, Bob Yeats, and two anonymous reviewers helped significantly clarify and improve the manuscript, and we thank them for their valuable suggestions and comments.