ABSTRACT The rifted Precambrian margin of western Laurentia is hypothesized to have consisted of a series of ~330°-oriented rift segments and ~060°-oriented transform segments. One difficulty with this idea is that the 87 Sr/ 86 Sr i = 0.706 isopleth, which is inferred to coincide with the trace of this rifted margin, is oriented approximately N-S along the western edge of the Idaho batholith and E-W in northern Idaho; the transition between the N-S– and E-W–oriented segments occurs near Orofino, Idaho. We present new paleomagnetic and geochronologic evidence that indicates that the area around Orofino, Idaho, has rotated ~30° clockwise since ca. 85 Ma. Consequently, we interpret the current N-S–oriented margin as originally oriented ~330°, consistent with a Precambrian rift segment, and the E-W margin as originally oriented ~060°, consistent with a transform segment. Independent geochemical and seismic evidence corroborates this interpretation of rotation of Blue Mountains terranes and adjacent Laurentian block. Left-lateral motion along the Lewis and Clark zone during Late Cretaceous–Paleogene time likely accommodated this rotation. The clockwise rotation partially explains the presence of the Columbia embayment, as Laurentian lithosphere was located further west. Restoration of the rotation results in a reconstructed Neoproterozoic margin with a distinct promontory and embayment, and it constrains the rifting direction as SW oriented. The rigid Precambrian rift-transform corner created a transpressional syntaxis during middle Cretaceous deformation associated with the western Idaho and Ahsahka shear zones. During the late Miocene to present, the Precambrian rift-transform corner has acted as a fulcrum, with the Blue Mountains terranes as the lever arm. This motion also explains the paired fan-shaped contractional deformation of the Yakima fold-and-thrust belt and fan-shaped extensional deformation in the Hells Canyon extensional province.

The Columbia River flood basalt province covers an area greater than 210,000 km 2 in the Pacific Northwest. The province is subdivided into the Oregon Plateau and the Columbia Basin based on significant differences in the style of deformation. The Oregon Plateau contains four structural-tectonic regions: (1) the northern Basin and Range, (2) the High Lava Plains, (3) the Owyhee Plateau, and (4) the Oregon-Idaho graben. The Columbia Basin covers a broader region and consists mainly of the Yakima Fold Belt and the Palouse Slope. Volcanism began in the Oregon Plateau and quickly spread north to the Columbia Basin. In the Oregon Plateau, flood basalt eruptions were contemporaneous with rhyolitic volcanism at the western end of the Snake River Plain hotspot track and with a major period of crustal extension in northern Nevada that began at ca. 16–17 Ma. In the Columbia Basin, a new phase of rapid subsidence folding and faulting of the basalt commenced with the initiation of volcanism but declined as volcanism waned. The coeval development of broad uplifts, subsiding basins, and flood basalt volcanism in the province is consistent with geodynamic models of plume emplacement. However, more specific structures in the province can be linked to older structures in the prebasalt basement. We attribute mid-Miocene deformation and the northward migration of volcanism to a rapidly spreading plume head that reactivated these preexisting structures. Exploitation of such structures may have also played a role in the orientation of many fissure dikes, including rapid eruption of the Steens Mountain shield volcano.

The pattern of deformation in the western part of the Columbia River flood basalt province contains two key components: (1) anticlinal uplifts of the Yakima Fold Belt with east-northeast to west-southwest trends, and (2) strike-slip fault zones with dominantly northwest trends. It is the abundance and regional extent of the latter that distinguish this area from other parts of the province. There are many northwest-striking, right-lateral, strike-slip faults in the interval from the Willamette Valley eastward to Umatilla (123°W to 119°W longitude). Some of these faults are only a few kilometers long, whereas others are of regional extent (>100 km). Conjugate northeast-striking, left-lateral, strike-slip faults have also been identified but are far less numerous. Local variations in the stress field within basins have produced sets of subsidiary structures by transtension and transpression. These occur where fault zones change trend with respect to the NNW-SSE–oriented maximum principal compressive stress. Strike-slip faulting was active early in the history of the Yakima Fold Belt uplifts, at least by emplacement of the Columbia River Basalt Group lavas, but after the Yakima Fold Belt uplift, spacing had already been firmly established. It is probable that many of these faults are episodically reactivated basement structures that have repeatedly undergone cycles of emergence, burial by flood basalts, and reemergence. Strike-slip deformation appears to have happened simultaneously within the Yakima Fold Belt uplifts and adjacent synclinal basins. However, the pattern and magnitude of deformation differ significantly in the basins compared to the uplifts. The Yakima Fold Belt uplifts have been segmented and shifted many kilometers by strike-slip faults, while displacements within adjacent basins are orders of magnitude less. Within Yakima Fold Belt uplifts, reversals of vergence sometimes occur wherein the frontal (forelimb) thrusts and fold asymmetry switch from one side of the uplift to the other. These changes are accommodated by cross-trending, right-lateral, strike-slip faults of regional extent. The pattern of strike-slip deformation as mapped within basins in many cases appears to be immature and lacking in interconnection. Eruptive vents in the Simcoe backarc volcanic field and Boring lavas are often aligned along strike-slip faults. Pliocene-age Simcoe lava flows have been deformed by both folding and strike-slip faulting within the Klickitat Valley basin. Pleistocene-age deposits are known to be cut by both the Luna Butte and Portland Hills faults. Strike-slip earthquake focal mechanisms have also been determined for some faults.

The Lewiston Structure is located in southeastern Washington and west-central Idaho and is a generally east-west–trending (~075°), asymmetric, noncylindrical anticline in the Columbia River Basalt Group that transfers displacement into the Limekiln fault system to the southeast and the Silcott fault system to the southwest. A serial cross-section analysis and three-dimensional (3-D) construction of this structure show how the fold varies along its trend and shed light on the deformational history of the Lewiston Basin. Construction of the fold’s 3-D form shows that the fold’s wavelength increases and amplitude decreases near its eastern and western boundaries. Balanced cross sections show ~5% shortening across the structure, which is consistent with the Yakima Fold Belt. An angular unconformity below the Grande Ronde Basalt N1 magnetostratigraphic unit, in addition to a variation of N1 unit thickness across the structure, suggests that the fold was forming before N1 time. Analysis of structural data using the Gauss method for heterogeneous fault-slip data indicates north-south (~350°) shortening prior to and after N1 emplacement. The presence of a reverse fault on the southern limb of the Lewiston Structure is controversial. This fault crops out to the east of the field area where Grande Ronde Basalt magnetostratigraphic unit R2 is thrust over Pliocene(?) gravels. However, better control on unit thicknesses and map contacts rules out an exposed reverse fault on the southern limb of the fold west of the Washington-Idaho border, suggesting the fault either dies out or becomes blind.

Recent mapping of pre-basalt rocks along the northwestern Columbia River basalt margin and well logs from Shell Oil Company gas wells provide new information about the rocks and structure underlying the Yakima fold belt. Pre-basalt rocks along the margin range in age from Jurassic to lower Miocene, with early to middle Tertiary arkosic and volcaniclastic strata concentrated in three fault-bounded basins. With one exception, pre-basalt rocks cut by the Shell Oil Company wells (Yakima Minerals, Bissa, and Saddle Mountains) can be correlated with rocks found in the basins along the margin. These rocks extend under the Columbia River Basalt Group almost to the center of the Columbia Basin. Two major features, the Leavenworth–Hog Ranch cross-structure and the White River–Naches River fault zone, affect the distribution of sedimentary rock types. Based on well and geophysical data, the Columbia River Basalt Group thins across the Hog Ranch–Naneum Ridge structure, suggesting that this feature was active during Miocene time. The northwestern Columbia River basalt margin is the focus of major structural elements that converge on the Yakima fold belt, including the Olympic-Wallowa lineament (OWL), the Cle Elum–Wallula lineament (CLEW), the Hog Ranch–Naneum Ridge cross-structure, the Chiwaukum graben, and the White River–Naches River fault zone. In the area of CLEW, splays of the Straight Creek fault turn southeast and pass under the Columbia River Basalt Group, aligning with folds of the Yakima fold belt. Elsewhere along the margin, there is little expression of sub-basalt structure in the overlying Columbia River basalt. The Columbia River Basalt Group, at the margin, exhibits an absence of faulting and displays only broad, gentle folds. Closely spaced, tight folds and associated faults in the interior of the Yakima fold belt either die out before reaching the margin or become broad, gentle flexures.

Topography and ground conditions were important factors in controlling the distribution of individual Columbia River Basalt Group (CRBG) flows in western Oregon. The Columbia trans-arc lowland, the Yakima fold belt, the Portland Hills–Clackamas River structural zone, and Cascadian volcanism largely controlled the distribution of CRBG flows across the Miocene Cascade Range. The first flows to cross the Miocene Cascades into the Willamette Valley encroached onto a low-relief topography generally consisting of eroded Tertiary-age marine sedimentary rocks deformed along northwest-trending structural zones, volcanic highs, and estuaries. No north-south trough affected the distribution and thickness of the CRBG in the Willamette Valley, but an incipient Coast Range acted as a leaky barrier to the Oregon coast. Water-saturated sediments rapidly extracted heat from advancing CRBG lava flows, producing narrow, abnormally thick lobes extending along existing topographic lows. Deformation along the northwest-trending Portland Hills–Clackamas River structural zone produced a major topographic barrier early and late in the incursion of CRBG flows. The CRBG thins across this zone from 600 to 150 m. This zone diverted the earliest Grande Ronde flows into and through the Portland Basin. Some of the succeeding R 2 and N 2 Grande Ronde flows were able to cross this zone and followed another structural low, the Sherwood trough, to the Oregon coast. The total thickness of CRBG along the Sherwood trough is approximately 300 m, about twice that on either side. Paleodrainage developed during time intervals between emplacement of CRBG flows. The positions of these drainage courses were influenced by the position of the CRBG flow margins and/or structural lows. A longer hiatus between flows (> 100,000 yr) enabled rivers to develop major canyons by headward erosion, which served to channelize subsequent CRBG flows.

Tectonic development and evolution of the central Columbia Plateau since middle Miocene time is a product of dynamic interplay among (1) the eruption and emplacement of the Columbia River Basalt Group (CRBG), (2) the subsidence of the area encompassing the Yakima fold belt subprovince, (3) the growth of the Yakima folds, and (4) the influence of regional structures transecting the fold belt, specifically the Hog Ranch-Naneum Ridge anticline and the Cle Elum–Wallula disturbed zone. Subsidence of the Yakima fold belt subprovince began prior to the eruption of the CRBG and has continued from Miocene time to the present. The rate of subsidence kept pace with the rate of CRBG flow emplacement, decreasing as CRBG volcanism waned. Simultaneously, anticlinal fold growth within the Yakima fold belt occurred under north–south compression and also decreased as the rates of subsidence and eruptions of lava declined. Paleomagnetic data indicate fold growth was accompanied by local clockwise rotation of basalt within the anticlines. The tectonic and volcanic histories of the central Columbia Plateau are interrelated and imply a common cause. The structural rotation and north-south compression, and thus fold growth, are interpreted to result from oblique subduction along a converging plate margin. The coincidence of the timing and rates of fold growth, subsidence of the central Columbia Plateau, and basalt production rates suggest that CRBG volcanism is primarily a product of oblique subduction off western North America.

Deformation of the continental flood-basalt in the westernmost portion of the Columbia Plateau has resulted in regularly spaced anticlinal ridges. The periodic nature of the anticlines is characterized by dividing the Yakima fold belt into three domains on the basis of spacings and orientations: (1) the northern domain, made up of the eastern segments of Umtanum Ridge, the Saddle Mountains, and the Frenchman Hills; (2) the central domain, made up of segments of Rattlesnake Ridge, the eastern segments of Horse Heaven Hills, Yakima Ridge, the western segments of Umtanum Ridge, Cleman Mountain, Bethel Ridge, and Manastash Ridge; and (3) the southern domain, made up of Gordon Ridge, the Columbia Hills, the western segment of Horse Heaven Hills, Toppenish Ridge, and Ahtanum Ridge. The northern, central, and southern domains have mean spacings of 19.6, 11.6, and 27.6 km, respectively, with a total range of 4 to 36 km and a mean of 20.4 km ( n = 203). The basalts are modeled as a multilayer of thin linear elastic plates with frictionless contacts, resting on a mechanically weak elastic substrate of finite thickness, that has buckled at a critical wavelength of folding. Free slip between layers is assumed, based on the presence of thin sedimentary interbeds in the Grande Ronde Basalt separating groups of flows with an average thickness of roughly 280 m. Many of the observed spacings can be explained by this model, given that: (1) the ratio in Young’s modulus between the basalt and underlying sediments E/E o ⩾ 1,000, (2) the thickness of the Grande Ronde Basalt was between 1,200 and 2,300 m when the present wavelengths were established, and (3) the average thickness of a layer in the multilayer is between 200 and 400 m. The lack of well-developed anticline-syncline pairs in the shape of a sinusoid may be the result of plastic yielding in the cores of the anticlines after initial deformation of the basalts into low amplitude folds. Elastic buckling coupled with plastic yielding confined to the hinge area could account for the asymmetric fold geometry of many of the anticlines.