Introduction

Magmatism in intracontinental settings represents a major unresolved problem in continental dynamics. Unlike mid-ocean ridge and subduction-zone settings where we have a first-order understanding of the dynamic processes responsible for magma generation and crust production, we have no such governing paradigms for the tectono-magmatic processes at intraplate settings. This proposal outlines a multi-disciplinary study of an excellent (archetypal) example of an intraplate magmatic environment that is one of the great continental volcanic provinces on Earth (Fig. 1). Our area of interest spans about 400 km from the Cascades arc eastward across an amalgam of accreted terranes in eastern Oregon to the Precambrian craton in Idaho and from nonextended and amagmatic accreted crust in central Oregon about 250 km to the south into the northern Basin and Range Province of southern Oregon and adjacent Nevada. We call this vast area the High Lava Plains [sensu lato; cf., Christiansen and McKee , 1978], which, along with the Snake River Plain, has been the most volcanically active part of North America in the late Cenozoic (Fig. 1).

Figure 1: Volcanic [after Smith and Luedke, 1984] and tectonic [after Streck et al. , 1999] elements of the Pacific Northwest. Left panel shows post 17 Ma volcanic deposits to illustrate the tremendous volcanic activity east of the Cascades in the northern reaches of the Basin and Range (outlined by thick grey lines on right panel). The right panel shows only sub-5 Ma volcanic fields to illustrate the continuing activity in the Cascades and along both the High Lava Plains (HLP – brown field) and the Eastern Snake River Plain (ESRP). Short curves along the ESRP and HLP are isochrons (ages in Ma) for the migrating silicic volcanism along each volcanic trace. Flood basalt activity was fed from dike systems in the Northern Nevada Rift (NNR), Steens Mtn. (SM), the Western Snake River Plain (WSRP) and the Chief Joseph (CJ) and Cornucopia (C) dike swarms of the Columbia River basalts. With the exception of the Cornucopia swarm, these dikes occur near the western border of Precambrian North America as defined by the 0.706 line (large dot-dash line). Dotted lines show NW trending fault systems: Olympic-Wallowa Lineament (OWL), Vale (V), Brothers (B), Eugene-Denio (ED) and McLoughlin (Mc). Additional features shown include Newberry Volcano (NB), the Owyhee Plateau (OP), Juan de Fuca Plate, and San Andreas Fault (SAF).

The complex tectonomagmatic evolution east of the Cascades has been variously ascribed to a variety of plate-boundary causes including back-arc spreading [ Christiansen and McKee , 1978], Basin and Range extension [ Cross and Pilger , 1978] and asthenospheric inflow behind a steepening subducting Juan de Fuca plate [ Carlson and Hart , 1987]. Given the extraordinary volume of volcanism in this area and the plate motion directed migration of volcanism up the Snake River Plain, the notion of a deep mantle plume has gained popularity as an alternative "external" cause of the volcanism [e.g. Geist and Richards , 1993; Camp and Ross , 2004]. All these various models have been developed over the last 30 years based largely on the geologic, geochemical and geochronologic study of the surface volcanism in this area. What is notably absent, but is critical for resolving the true cause of magmatism in this area, is information on the crustal basement and upper mantle structure that underlies the veneer of late Cenozoic volcanic rocks that completely covers most of southeastern Oregon.The four broad questions of concern in explaining the magmatism in southeastern Oregon are the same as those related to the general causes of continental intraplate tectonomagmatism:

1. Is a plume the only way to get large-volume aerially extensive intraplate volcanism?
2. Does flow of mantle around the edges of a retreating or "dying" subducting plate instigate focused volcanism in the overlying crust of the back arc?
3. Can the bottom topography of the lithospheric mantle influence flow in the underlying mantle to the point of localizing tectonomagmatism in the overlying crust?
4. Is crustal extension the cause or expression of continental magmatism?

Overview of project structure, goals, and integration

Existing geologic, geochemical, and geochronologic information on the surface volcanism provides the basis to suggest that the High Lava Plains is the best, currently active, example worldwide in which to study the causes of widespread intraplate continental volcanism. This four-year, multi-institutional, interdisciplinary project will have four major components; seismology, experimental petrology, geochemistry and geodynamic modeling. The collaborative research will produce:

• The first high resolution images of crustal and upper mantle structure beneath the High Lava Plains
• The first experimental petrological study of primitive basalts spanning the wide compositional range observed from the Cascades to the Snake River Plain
• New geochemical and geochronologic studies targeted specifically to improve estimates of spatio-temporal variations in volcanic volume and the role of subduction-derived fluids/melts in magma genesis
• Geodynamic models of mantle flow patterns expected for the complicated mantle structure present in this area that may include any or all of: a plume, a retreating slab, a migrating southern edge to the slab, and significant topography on the base of the lithospheric mantle

The way we view the various disciplinary components of this project interacting is shown schematically in Table 1.

Table 1: What process(es) lead to voluminous continental intraplate volcanism?

The Investigative Tools

The greatest hindrance to understanding and testing ideas about the processes causing widespread magmatism in the High Lava Plains is the lack of detailed geophysical information on crust and upper mantle structure and how these structures relate to the Cenozoic volcanic record. Other regions of the western US, particularly the Snake River Plain - Yellowstone area, have been the subjects of several recent geophysical imaging projects [ Saltzer and Humphreys , 1997; Schutt et al. , 1997, 1998; Peng and Humphreys , 1998; Humphreys et al. , 2000; Christiansen et al. , 2002], but eastern Oregon remains largely uninvestigated geophysically. We propose to carry out seismic imaging at three levels of resolution (Fig. 2). Broad regional coverage will be provided by analysis of data from the transportable component of USArray [ Meltzer et al. , 1999]. This wide area image with relatively low resolution will illuminate features of the High Lava Plains lithosphere as they relate to broader scale features of the western US including the subducting Juan de Fuca plate. Higher resolution images of upper mantle structure and anisotropy will be provided by dense arrays, both linear and 2-D, of broadband seismometers (Fig. 2).

Figure 2: Map showing station locations overlain on color-coded topography in southeastern Oregon. A total of 92 broadband (REFTEK/STS-2 or equivalent) seismograph units will be used in two separate deployments over a two-year period. The broadband instrumentation consists of 32 telemetered stations (white stars) and 60 stand-alone stations (white triangles). (Note: for ease of discussion we use the term “telemetered array” for these deployments. Logistical restrictions make it likely that these dense profiles will consist of both telemetered and stand-alone instruments). Approximately 170 broadband sites will be occupied during the course of the experiment, with ~30 sites duplicated to provide anchor points at the crossing between the two transects. The first year deployment will be along the NW-SE transect, the second year deployment will be along the crossing N-S transect. The axes of the two broadband transects (instrument spacing ~15 km) coincide with and complement the refraction/reflection lines. This geometry will allow for continuous, unaliased, high-resolution imaging of discontinuity and anisotropy structure within both crust and upper mantle. The 2-D array is of sufficient aerial extent (~500x350 km) to assure that tomographic structures can be well resolved from the uppermost mantle and lowermost crust to depths in excess of 300 km. The refraction/low-fold reflection profiles are shown as heavy black dashed lines. Black squares denote USArray Transportable Array (Bigfoot) stations, scheduled for installation in 2006. The closely spaced open circles that form the east-west profile from the Pacific coast across the Cascades mark stations of the Cascadia broadband experiment [ Nabelek et al. , 1993]. Small black circles are nominal sites of a reconnaissance broadband experiment led by Richard Allen (U. Wisconsin). Small white diamonds denote regional network sites comprised mostly of short period instruments. Labeled dotted yellow lines indicate previous active source experiments across Newberry and northern Nevada.

These dense arrays include: (1) two closely spaced (~15 km) linear deployments to enable continuous and unaliased imaging of deep discontinuity structure; and (2) sixty stand-alone broadband deployments in densified swaths (station spacing ~ 20–30 km) along the two major transects to enable high-resolution tomography to depths in excess of 300 km (Fig. 3).

Figure 3: Horizontal cross-sections through hypothetical P-wave velocity models showing percent slowness velocity perturbations at 50 km (top 3 panels) and 100 km depth (bottom 3 panels). The resolution tests are based on an assumed two years of teleseismic events with body wave magnitude conservatively restricted to events ≥5.8. A relatively large P-wave rms noise level of 0.05s has been added to the synthetic data. Leftmost panels are map views of a starting model that includes: a hypothetical 2% low velocity tabular anomaly beneath the High Lava Plains region to a depth of 125 km; a 50 km diameter sphere of 5% negative anomaly beneath the Newberry region centered at a depth of 75 km; a Proterozoic North American lithosphere 3% faster than that of the surrounding upper mantle beneath central and eastern Oregon to a depth of 175 km; a hypothetical Blue Mountain lithosphere represented by a weakly positive anomaly of 1.5% to 125 km. In the western part of the region, a dipping 100 km slab with a tapered positive 4% velocity anomaly represents a descending plate at 45º, but is not well imaged in these particular examples. We note that these features are for the purposes of illustration only and do not reflect any preconceived notion of velocity structure in the region.

Middle panels are the output models for linear inversions based on data from USArray Transportable stations alone (black dots; see also Fig. 2). Rightmost panels are the output models for linear inversions based on data from a combination of USArray stations and the additional densified arrays proposed for this study (black dots; see also Fig. 2). These resolution tests clearly demonstrate that the proposed dense broadband array is critical for imaging structures in the upper 100 km of the mantle, where the smaller-scale distribution of thermal and chemical anomalies likely present in this region is directly applicable to the surface geology, petrology and geochemistry. Conversely, USArray stations alone provide little resolution of small features in the upper 100 km. The hypothetical boundary with Proterozoic North America is best defined where the NW-SE densified swath of stations crosses near the Oregon-Idaho border. These examples demonstrate a characteristic problem with tomographic models where velocity anomalies tend to “smear out” with depth, demonstrating the need to combine tomography with receiver function imaging and array-based processing techniques to map discontinuities precisely. The tomography does an extremely good job, however, of imaging the geographic boundaries and horizontal extent of the anomalous bodies, which will be further enhanced by the nonlinear inversions that will be performed on the real data.

An embedded active-source refraction array will provide still higher resolution images of crustal velocity structure. Both the broadband arrays and embedded active-source refraction lines are sited to provide the best imaging of the mantle and crust directly beneath the trace of volcanism along the High Lava Plains and across the four distinct terrane boundaries surrounding the High Lava Plains (Fig. 4).

Figure 4: Schematic showing the strategy for multidisciplinary imaging of the crust and upper mantle.

To assist in interpreting the seismic structures imaged in the crust, gravity measurements along the seismic refraction lines will further constrain the density and hence compositional structure of the crust, particularly near major volcanic centers. This information is essential in order to obtain accurate estimates of spatio-temporal variations in the volume of magmatism; a key input parameter to geodynamic modeling that will examine the mantle melting consequences of flow in the mantle caused by a plume, a retreating slab, and around the southern edge of the "dying" Juan de Fuca plate, in an attempt to correlate the surface volcanism with features of the underlying mantle that ultimately may be responsible for the volcanism.

Geodynamic studies will focus on the potential role of the subducting Juan de Fuca plate on influencing flow in the surrounding mantle and how this may translate into enhanced volcanism in the overlying crust of the back arc, including the possibility of slab-plume interaction. An example of the general approach to be undertaken in the laboratory component of this effort is shown schematically in figure 5.

Figure 5: Combination of previous laboratory results (top row) with conceptual models for melt production (bottom row) within the wedge for a time evolution in subduction styles from a) shallow downdip motion of the slab to b) sudden onset of slab steepening to c) the initiation of slab retreat.In (a), flowlines in the wedge are shallow (nearly horizontal, unfavorable for decompression melting). In (b), the change in subduction mode produces rapid steepening of flowlines in the wedge and a long, linear pulse of decompression melting.In (c), the zone of residual melt (perhaps with increased viscosity) and the zones of active melting are deformed by the strong lateral return flow (or shear flow) to the wedge produced by slab retreat. This model will be tested in the proposed work by running similar subduction scenarios with added model aspects (variable topography for the base of the overriding plate; introduction of a increased viscosity region to represent the residuum). A 3-D network of gridded streak images will be analyzed with DPIV software to produce map-view slices of vertical fluid velocity from different depth horizons, for incorporation into melting models. After Kincaid and Hall (2003) and Hall and Kincaid (2003, 2004).

These geophysical tools, guided by the physical modeling component of this project, are directed at key targets using the background of geological, geochemical, and geochronological information obtained by a subset of the PI's, and others, through past work in this area. We do not intend to duplicate previous field and geochemical study of the volcanic rocks in this area, but we do intend to add some components to this dataset that are of critical importance as input into the geophysical and geodynamic aspects of this project. One such component is improved geochronology of the young volcanism in order to better define the pattern of volcanic volume versus time and position that will be an important parameter to compare with both seismic images and particularly geodynamic models. Another component is experimental petrological investigations of primitive basalts coupled with measure of their volatile (H 2 O, CO 2 ) abundances and other tracers (Ba, Nb, Ta, Cl abundances and B isotopic composition) sensitive to the presence of subduction fluids/melts in the source of these magmas (Fig. 6). Information on the pressure, temperature, and water content in the magma's mantle sources will provide a critical calibration of the temperature structure in the upper mantle as inferred from teleseismic tomography. Finally, we will obtain data on subsurface crustal rocks through analysis of the rare xenoliths of upper crust, xenocrysts of zircon and feldspar contained in some of the silicic volcanic rocks, and petrologic modeling of crustal magma sources and their compositional variation with time. These data will constrain seismically determined crustal velocity/density structure and how the crust of the High Lava Plains has been modified by the extensive Cenozoic magmatism.

Figure 6: Schematic cross-section across the Cascade arc to the south of High Lava Plains study area showing petrologic estimates of depth of mantle melting and crustal fractionation. The position of the slab and the thickness of the crust are inferred from geophysical evidence. Squares show maximum depth of fractional crystallization of olivine and plagioclase and circles show mantle melting depth. Temperature of mantle melting also varies systematically across the arc from 1300 o C in the east to 1450 o C in the west. Two vents, Tennant and Yellowjacket, fall off the trend, because these lavas were modified by the addition of a slab fluid. After Elkins-Tanton et al. (2001)

References

Camp, V.E. and M.E. Ross, Mantle dynamics and genesis of mafic magmatism in the intermontane Pacific Northwest, J. Geophys. Res. 109, doi:10.129/2003JB002838, 2004.

Carlson, R.W., and W.K. Hart, Crustal genesis on the Oregon Plateau, J. Geophys. Res. 92 , 6191-6206, 1987.

Christiansen, R.L., G.R. Foulger, and J.R. Evans, Upper-mantle origin of the Yellowstone hotspot, Geol. Soc. Amer. Bull. 114 , 1245-1256, 2002.

Christiansen, R.L., and E.H. McKee, Late Cenozoic volcanic and tectonic evolution of the Great Basin and Columbia Intermontane regions, in Geol. Soc. Amer. Mem. , edited by R.B. Smith, and G.P. Eaton, pp. 283-311, 1978.

Cross, T.A., and R.H. Pilger, Constraints on absolute motion and plate interaction inferred from Cenozoic igneous activity in the western United States, Am.J.Sci. 278 , 865-902, 1978.

Elkins-Tanton, L.T., T.L. Grove, and J. Donnelly-Nolan, Hot, shallow mantle melting under the Cascades volcanic arc, Geology 29 , 631-634, 2001.

Geist, D., and M. Richards, Origin of the Columbia Plateau and Snake River Plain: deflection of the Yellowstone plume, Geology 21 , 789-792, 1993.

Hall, P. and C. Kincaid, Melting, dehydration and the dynamics of off-axis plume-ridge interaction, Geochem. Geophys. Geosyst ., 4(9), 8510, 2003.

Hall, P., and C. Kincaid, Melting, dehydration and the geochemistry of off-axis plume-ridge interaction, Geochem. Geophys. Geosyst ., in press, 2004.

Kincaid, C. and P. Hall, The role of back-arc spreading in circulation and melting at subduction zones, J. Geophy. Res , 208, 2240-2254, 2003.

Meltzer, A., R. Rudnick, P. Zeitler, A. Levander, G. Humphreys, K. Karlstrom, G. Ekstrom, R. Carlson, T. Dixon, M. Gurnis, P. Shearer, and R. van der Hilst, USArray Initiative, GSAToday 9 , 8-10, 1999.

Nabelek, J.L., X.-Q. Li, S. Azevado, J. Braunmiller, A. Fabritius, B. Leitner, A. Trehu, and G. Zandt, A high-resolution image of the Cascadia subduction zone from teleseismic converted phases recorded by a broadband seismic array, EOS, Trans. Amer. Geophys. Union 74 , 431, 1993.

Peng, X., and E.D. Humphreys, Crustal velocity structure of the eastern Snake River Plain and the Yellowstone swell, J. Geophys. Res. 103 , 7171-7186, 1998.

Saltzer, R., and E.D. Humphreys, Upper mantle P-wave structure of the eastern Snake River Plain and its relationship to geodynamic models of the region, J. Geophys. Res. 102 , 11,829-11,841, 1997.

Schutt, D.L., E.D. Humphreys, and K.G. Dueker, Vp/Vs structure beneath the Snake River Plain; imaging upper mantle partial melt, EOS, Trans. Amer. Geophys. Union 78 , 492-493, 1997.

Schutt, D.L., E.D. Humphreys, and K. Dueker, Anisotropy of the Yellowstone hot spot wake, eastern Snake River Plain, Idaho, Pure and Applied Geophysics 151 , 443-462, 1998.

Smith, R.L., and R.G. Luedke, Potentially active volcanic lineaments and loci in western conterminous United States, in Explosive Volcanism , pp. 47-66, National Academy Press, Washington, 1984.

Streck, M.J., and A.L. Grunder, Enrichment of basalt and mixing of dacite in the rootzone of a large rhyolite chamber: inclusions and pumices from the Rattlesnake Tuff, Oregon, Contrib. Mineral. Petrol. 136 , 193-212, 1999.

Streck, M.J., J.A. Johnson, and A.L. Grunder, Field Guide to the geology of the Rattlesnake Tuff and High Lava Plains, eastern Oregon, Oregon Geology 61 , 64-76, 1999.