Subsections

Western USA mantle structure and its implications for mantle convection processes

Mei Xue and Richard M. Allen

Introduction

The western USA is on the margin of the North American plate and has complicated and active tectonics. The Juan de Fuca plate, the Pacific plate, and the North American plate meet in this region and form the Mendocino Triple Junction just offshore Northern California. In addition to these primary tectonic objects, many other complicated geologic features are also observed (Figure 2.25). Their corresponding velocity structures have not been well resolved so far and many are still under debate, e.g., the depths the Juan De Fuca plate reaches and the Yellowstone plume originates (e.g., Humphreys, et al., 2000; Jordan, et al.,2004; Yuan and Dueker, 2005; Waite, et al., 2006; Geist and Richards, 1993). Here we incorporate the Transportable Array data with all other available networks, resulting in an unprecedented dense distribution of stations in the western USA. This allows us not only to fill the gaps in the resolution of previous studies, but also to see deeper into the mantle, revealing new features. We refer to our seismic velocity models as DNA07-P for P-wave and DNA07-S for S-wave, where DNA07 represents the Dynamic North America model of 2007. Due to the limited space, we only show DNA07-S here, which reveals an extremely heterogeneous mantle structure and provides important clues to mantle convection processes in this tectonically active region.

Figure: Tectonic map for the study region. Labeled features (Humphreys and Dueker, 1994b) are OH, Okanogan Highlands; OM, Olympic Mountains; OCR, Oregon Coast Ranges; CCR, California Coast Ranges; KM, Klamath Mountains; MP, Modoc Plateau; MTJ, Mendocino Triple Junction; CV, Central Valley; SN, Sierra Nevada; SAF, San Andreas Fault; TR, Transverse Ranges; ST, Salton Trough; CRB, Columbia River Basalts; BM, Blue Mountains; WM, Wallowa Mountains; NC, Newberry Caldera; MC, McDermitt Caldera; YC, Yellowstone Caldera; YHT, Yellowstone hotspot track along the eastern Snake River Plain; B&R, Basin and Range; SB&R, southern Basin and Range; TMC, Timber Mountain Caldera; SGVT, Saint George Volcanic Trend; CP, Colorado Plateau; WF, Wasatch Front; RMF, Rocky Mountain Caldera. Black lines trending northwest across Oregon indicate right-lateral strike-slip faults. Dike swarms associated with the 17 Ma basaltic outpourings are shown in gold (Christiansen, et al., 2002). Plate motions from HS3-NUVEL 1A are shown as black arrows (Gripp and Gordon, 2002). Horizontal black lines indicate the locations of the vertical slices shown in Figure 2.28
\begin{figure}
\begin{center}
\epsfig{file=meixue08_1_1.eps, width=8cm}
\end{center}
\end{figure}

Data and Method

The total number of stations we used is 809, and most are from the Transportable Array, with an average station spacing of 70 km (Figure 2.26). The 26 permanent networks are composed of the following: (1) 2 Global Seismograph Networks: (IRIS/IDA and IRIS/USGS); (2) 5 Federal Digital Seismic Networks: the Canadian National Seismograph Network (CNSN), GEOSCOPE (GEO), International Miscellaneous Stations (IMS), Leo Brady Network (LB), and the United States National Seismic Network (USNSN); (3) 14 regional networks: the ANZA Regional Network (ANZA), Berkeley Digital Seismograph Network (BDSN), Cascade Chain Volcano Monitoring (CC), Caltech Regional Seismic Network (CRSN), Montana Regional Seismic Network (MRSN), Northern California Seismic Network (NCSN), Western Great Basin/Eastern Sierra Nevada (WGB/ESN), Princeton Earth Physics Project-Indiana (PEPP), US Bureau of Reclamation Seismic Networks (USBR), Southern California Seismic Network TERRAscope (TERRA), University of Oregon Regional Network (UO), University of Utah Regional Network (UURN), Pacific Northwest Regional Seismic Network (PNSN), and the Yellowstone Wyoming Seismic Network (YWSN); (4) 3 temporary networks: the North Bay Seismic Experiment (NBSE), DELTA LEVY Northern California (DLNC), and the Wallowa TA 2006-2008 (WTA); and (5) 2 other networks: the Laser Interferometer Gravitational-Wave Experiment (LIGO) and the Network of Autonomously Recording Seismographs (NARS). We use a technique of teleseismic body wave traveltime tomography and follow the procedure of (Allen, et al., 2002). To correct for source effects and crustal structures, event and station corrections are included in the inversion as a set of free parameters. Rather than leave the station corrections unconstrained, we use corrections calculated from the crustal model of CRUST2.0 as a reference (Bassin, et al., 2000). Station elevations are also corrected. The initial RMS residual is 1.83 sec and is reduced to 0.49 sec after inversion, corresponding to a variance reduction of 73$\%$.

Figure 2.26: The seismic stations used in this study, with a total number of 809. The inset shows the distribution of 88 events and 23233 rays used in the DNA07-S model inversion. The red and yellow dots indicate events providing good direct S phases and SKS phases, respectively.
\begin{figure}
\begin{center}
\epsfig{file=meixue08_1_2.eps, width=8cm, bb=0 72 560 670 , clip }
\end{center}
\end{figure}

Tomographic results and interpretation

Figure 2.27: Depth slices through the DNA07-S model from 100 km to 800 km depth with an interval of 100 km. Areas with ray hits smaller than 10 are shaded gray. On the right side of each velocity slice is the corresponding ray density plot, where white indicates zero hits and black indicates 100$+$ hits
\begin{figure}
\begin{center}
\epsfig{file=meixue08_1_3.eps, width=8cm}
\end{center}
\end{figure}

Figure 2.28: Vertical slices through the DNA07-S models. The locations of the cross-sections are shown in Figure 2.25. The color scale is the same as in Figure 2.27
\begin{figure}
\begin{center}
\epsfig{file=meixue08_1_4.eps, width=8cm}
\end{center}
\end{figure}

Perhaps one of the most striking observations is just how heterogeneous mantle structure is beneath the western USA. Despite this heterogeneity, there is a very strong correspondence with the complicated tectonics of the region. The main features of the velocity models and their implications, shown in Figure 2.27 and Figure 2.28, are listed as follows: North of the Mendocino Triple Junction: (1) The Juan de Fuca subduction system stops at $\sim{400}$ km, and is disrupted in Oregon, which we interpret as being due to interaction with the Yellowstone plume head. (2) West of the Cascades the forearc is imaged as a low velocity zone beneath the Coastal Ranges with the strongest velocity anomaly beneath the Olympic Mountains and Northern California. (3) East of the Cascades and above the Juan de Fuca slab, a north-south trending low velocity zone is imaged from southern Washington to northern Nevada. (4) A high velocity region is imaged from central Washington, through northern Oregon, and into Idaho. Beneath Washington the anomalies reach 250 to 300 km depth and deeper, extending to $\sim{400}$ km, beneath the Wallowa Mountains of northeast Oregon. These are likely due to a combination of a cold and thick lithosphere and melt extraction during the eruption of the Columbia River Basalts. (5) The low velocity anomaly beneath Yellowstone dips towards the northwest and stops at 500 km depth. (6) A shallow low velocity zone to $\sim{200}$ km depth lies beneath the Eastern Snake River Plain and does not appear connected to a deeper low velocity zone at the top of the lower mantle. (7) We do not detect a low velocity conduit reaching greater than 500 km depth beneath Yellowstone implying that either (a) any plume was short-lived; or (b) the conduit is $<$ 50 km in diameter and/or the velocity perturbation is less than 1.5% for S and 0.75% for P and therefore unresolved; or (c) there was no deep mantle plume. We prefer the short-lived plume model as it best explains many of the imaged features the Pacific Northwest. (8) There are only shallow low velocity anomalies ($<$150 km depth) along the Newberry hotspot track indicate no deep source. South of the Mendocino Triple Junction: (1) In California, the high velocities of the Pacific plate are imaged abutting against the low velocity North American plate. (2) We image the "slab gap" as low velocity anomalies extending to 400 km depth from the southern end of the Juan de Fuca subduction system to the southern end of the Sierra Nevada. These anomalies are particularly strong just south of the Mendocino Triple Junction. (3) High velocity bodies are imaged beneath the southern tip of the Central Valley/Sierra Nevada and the Transverse Ranges with dips to the east. These may be part of a fossil Farallon subduction system. (4) The Basin and Range is a region of low velocities to a depth of $\sim{300}$ km. In the middle of the Basin and Range, in central Nevada, a high velocity feature is imaged extending to 300 km depth. (5) A zone of low velocity is observed to 200 km depth under the Salton Trough consistent with ongoing rifting and small scale convection in the region. While the upper $\sim{400}$ km of the DNA07 models correlate well with surface tectonics and geologic provinces, the deeper structure (400-750km) is equally complex and not easily explained in terms of either existing geologic or geodynamic models. Further investigation is therefore warranted.

Acknowledgements

We thank Ana Luz Acevedo-Cabrera for preprocessing some of the data and thank Greg Waite for providing crust correction codes. We thank Doug Dreger, Barbara Romanowicz, and Ved Lekic for beneficial discussions/suggestions. The IRIS DMC provided seismic data. This work was supported by the NSF (EAR-0539987). The figures were produced with SAC and GMT (Wessel and Smith, 1995).

References

Bassin, C., Laske, G. and Masters, G., The Current Limits of Resolution for Surface Wave Tomography in North America, EOS Trans AGU, 81, F897, 2000. Christiansen, R. L., et al., Upper-mantle origin of the Yellowstone hotspot, Geological Society of America Bulletin, 114, 1245-1256, 2002. Geist, D. and M. Richards, Origin of the Columbia Plateau and Snake River Plain - Deflection of the Yellowstone Plume, Geology, 21, 789-792, 1993. Gripp, A. E. and R. G. Gordon, Young tracks of hotspots and current plate velocities, Geophysical Journal International, 150, 321-361, 2002. Hales, T. C., et al., A lithospheric instability origin for Columbia River flood basalts and Wallowa Mountains uplift in northeast Oregon, Nature, 438, 842-845, 2005. Humphreys, E. D. and K. G. Dueker, Western United-States Upper-Mantle Structure, Journal of Geophysical Research-Solid Earth, 99, 9615-9634, 1994b. Humphreys, E. D., et al., Bemeath Yellowstone: Evaluating Plume and Noneplume Models Using Teleseismic Images of the Upper Mantle, GSA Today, 10, 1-7, 2000. Jordan, B. T., et al., Geochronology of age-progressive volcanism of the Oregon High Lava Plains: Implications for the plume interpretation of Yellowstone, Journal of Geophysical Research, 109, doi:1029/2003JB002776, 2004. Waite, G. P., et al., V-pp and Vs structure of the Yellowstone hot spot from teleseismic tomography: Evidence for an upper mantle plume, Journal of Geophysical Research-Solid Earth, 111,, 2006. Yuan, H. Y. and K. Dueker, Teleseismic P-wave tomogram of the Yellowstone plume, Geophysical Research Letters, 32, 2005.

Berkeley Seismological Laboratory
215 McCone Hall, UC Berkeley, Berkeley, CA 94720-4760
Questions or comments? Send e-mail: www@seismo.berkeley.edu
© 2007, The Regents of the University of California