S32B-01 INVITED
Seismic Azimuthal Anisotropy Beneath the Western United States from Ambient Noise Tomography and Shear Wave Splitting
Constraining the nature of deformation in the crust and upper mantle is a key element behind the motivation for the EarthScope Project, and is an essential component toward a better understanding of plate tectonic driving forces. Data from Earthscope's USArray Transportable Array is providing an unprecedented opportunity to image Earth's crust and mantle fabric in great detail via imaging of seismic anisotropy. In this study, we examine seismic azimuthal anisotropy through a multi-pronged analysis of shear wave splitting, which provides good lateral resolution, and ambient noise tomography, which provides improved resolution with depth. To provide constraints on the lateral distribution of anisotropy, we determined shear wave splitting parameters for over 400 broadband stations in the western United States and found clear variations with geologic terrane. In the Pacific Northwest, splitting times are large (2.25+ sec) and fast directions are oriented ~E-W with limited variability. Beneath the southern Basin and Range/Colorado Plateau region, splitting times are also large (~1.75+ sec) and fast directions are oriented ~NE-SW (similar to absolute plate motion). Away from the Pacific-North American plate boundary, and sandwiched between broad regions of simple (i.e., regionally similar fast directions) and strong (i.e., large splitting times) azimuthal anisotropy, stations within the Great Basin exhibit significant complexity. Fast directions show a clear rotation from E-W in the northern Great Basin, to N-S in the eastern Great Basin, to NE-SW in the southeastern Great Basin. Splitting times reduce dramatically, approaching zero within the central Great Basin. To provide constraints on the depth distribution of shear wave anisotropy, we inverted Rayleigh wave phase velocity profiles ranging from 10 to 34 sec. The velocity profiles were computed from cross-correlation of ambient noise to determine velocity and azimuthal anisotropy structure of the crust and uppermost mantle across the western U.S. using a similar distribution of seismic stations applied for the shear wave splitting analysis. Overall, fast directions and relative strengths of anisotropy of the uppermost mantle layer in the surface wave inversion exhibit a similar pattern compared to the shear wave splitting results, while fast directions and relative strengths of anisotropy in the crust exhibit a weaker correlation with shear wave splitting variations. In order to compare these datasets more directly, we will use the results of the surface wave inversion to predict shear wave splitting at each station used in the splitting analysis, and will assess the lateral correlation of both fast directions and strength of azimuthal anisotropy. Further, we will quantify the scale of lateral heterogeneity in bulk upper mantle anisotropy before and after corrections for crustal fabric. We also will use the results of each dataset as the starting model in an inversion for the companion dataset. Results from these analyses will permit a direct assessment of crustal anisotropy on shear wave splitting measurements and the depth distribution of azimuthal anisotropy. In turn, the results from this effort will enable us to assess more directly the relationship of crustal anisotropy and geologic terrane, and will provide an improved understanding of the sublithospheric mantle flow field.
S32B-02
3-D Isotropic and Anisotropic S-velocity Structure in the North American Upper Mantle
The tectonic diversity of the North American continent has led to a number of geological, tectonic and geodynamical models, many of which can be better tested with high resolution 3-d tomographic models of the isotropic and anisotropic mantle structure of the continent. In the framework of non-linear asymptotic coupling theory (NACT), we recently developed tools to invert long period seismic waveforms combined with SKS splitting data, for both isotropic and radial and azimuthal anisotropic S-wave velocity structure in the upper mantle at the continental scale (Marone et al., 2007; Marone and Romanowicz, 2007). Striking differences in both isotropic and anisotropic velocity structure were observed: beneath the high velocity stable cratonic region a distinct two-layer anisotropic domain is present, with the bottom layer fast axis direction aligned with the absolute plate motion, and a shallower lithospheric layer with north pointing fast axis most likely showing records of past tectonic history; under the active western US the direction of tomographically inferred anisotropy is stable with depth and compatible with the absolute plate motion direction. Here we present an updated model which includes nearly five more years of data, including data from newly operative USArray stations, and a somewhat more extended frequency band. Our new model confirms our previous results, and reveals greater yet complex details of the anisotropic velocity structure beneath the western U.S.. We also show initial results of incorporating constraints on the depth to the lithosphere-asthenosphere boundary (LAB) using teleseismic receiver functions. We discuss the different anisotropic domains resolved both laterally and in depth, in the context of tectonic history of the north American continent.
S32B-03
Crustal Anisotropy in the Western US from a Joint Study of Receiver Functions and Surface Waves
I present a comparison of crustal anisotropy in the western US measured with receiver functions versus surface waves. Recent USArray data and technique developments (Lin et al., session S21) have resulted in new surface-wave based maps of radial and azimuthal anisotropy (Moschetti et al., this session). Longer period surface wave azimuthal anisotropy, with sensitivity kernels peaking in the upper mantle, shows general correlation with SKS splitting parameters (see Fouch et al., this session), while crustal kernel short-period surface waves show different patterns. The analysis of azimuthal variations of receiver functions offers a way to measure crustal anisotropy with high lateral and depth resolution. Crustal anisotropy with horizontal or plunging symmetry axes leads to characteristic azimuthal patterns in radial and tangential component receiver functions. One drawback of the technique is that the determination of anisotropic parameters is often nonunique. Geological and petrological a priori information can be used to constrain the models. Alternatively, the short period surface wave azimuthal anisotropy map can be used as a smoothed version (lateral resolution of about 200 km, depth sensitivity kernel spanning the entire crust with increased sensitivity in lower crust) of the crustal anisotropy measurable with receiver functions. Depending on the azimuthal coverage and noise conditions at the station, different approaches can be used to detect and measure anisotropy with receiver functions. Optimal data can be fit with synthetic seismograms to determine the magnitude, depth, and orientation of anisotropy and place constraints on details of symmetry. Where azimuthal coverage is insufficient to quantify anisotropic parameters, I calculate the power as a function of depth of arrivals with a periodicity of 180 deg (due to horizontal symmetry axis anisotropy) or 360 deg (due to plunging symmetry axis or dipping isotropic interfaces, distinguishable by phase) to indicate anisotropic crustal layers. Surface wave results can be used as a basis for hypothesis testing (do the receiver functions allow the orientation and magnitude of anisotropy suggested by the surface waves?). I interpret the results in the context of the tectonic evolution of the western US and the nature of coupling between crustal and mantle deformation.
S32B-04
Crustal Anisotropy in Southern California: Evidence for a Fossilized Detachment?
Receiver functions were calculated for sixteen stations in southwestern California and are currently being calculated for eight more stations in the Mojave Desert and surrounding area to explore a proposed fossilized mid/lower crustal regional detachment related to past subduction in the region (Ozacar and Zandt, 2008). All of the stations exhibit large azimuthal variations in receiver function polarity and amplitude on both the tangential and radial components. A trend seen regionally on radial components is an arrival immediately before the Moho which appears as a high amplitude positive arrival. This pre-Moho arrival peaks in amplitude as a negative arrival at ~200 ° and is a low amplitude positive at ~40°. On the tangential component this arrival appears as a negative arrival from ~0° to ~200° where it becomes a positive arrival to ~360°. The move out pattern and consistency of this signal over a large geographic area suggests that it stems from an anisotropic lower-crustal low velocity zone. While dipping interfaces may contribute to the complexity of the signal seen in individual stations, it is unlikely that similar dipping interfaces could be present throughout the region. A neighborhood algorithm search was run on stations RPV, MWC, CHF, PHL, CIA, VTV and PKD to invert for interface depth, strike and dip, percent anisotropy and orientation of anisotropy. The inversion for all of the stations (with the exception of VTV and CIA) resulted in an anisotropic low velocity layer that range in depth from ~14 to 20 km with anisotropy ranging from -14% to -20% with a unique axis oriented at ~240 °. We are currently calculating receiver functions and running inversions for the Mojave stations. These initial results support the idea of a regionally pervasive subduction related detachment underlying the Salinian, San Gabriel and related terranes. Further work will help to confirm the existence of this detachment and determine its extent.
S32B-05
Seismic character of the crust and upper mantle beneath the Sierra Nevada
Recent geophysical studies of the Southern Sierra Nevada suggest that the removal of a gravitationally unstable, eclogitic residue links to recent volcanism and uplift in the Eastern Sierra. The Sierra Nevada EarthScope Project (SNEP) investigates the extent of this process beneath Central and Northern Sierra Nevada. We present receiver functions, which provide estimates of crustal thickness and Vp/Vs and image the response of the crust and upper mantle to lithospheric removal. For completeness this study combines data from the 2005-2007 SNEP broadband experiment, EarthScope's BigFoot Array, regional backbone stations, and earlier PASSCAL deployments. We analyze transects of teleseismic receiver functions generated using a common-conversion-point stacking algorithm. These identify a narrow, "bright" conversion from the Moho at depths of ~25-35 km along the crest of the Eastern Sierra and adjacent Basin and Range northward to the Cascade Arc. Trade-off analysis using the primary conversion and reverberations shows a high Vp/Vs (~1.9) throughout the Eastern Sierra, which may relate to partial melt present in the lower crust. To the west the crust-mantle boundary vanishes beneath the western foothills. However, low frequency receiver functions do image the crust-mantle boundary exceeding 50 km depth along the foothills to the west and south of Yosemite National Park. Unusually deep, intraplate earthquakes (Ryan et al., this session) occur in the center of this region. The frequency dependence of the Moho conversion implies a gradational increase from crust to mantle wavespeeds over a significant depth interval. The transition from a sharp to gradational Moho probably relates to the change from a delaminated granitic crust to crust with an intact, dense, eclogitic residue. The spatial correlation and focal mechanisms of the deep earthquakes suggest that a segment of this still intact residue is currently delaminating.
S32B-06
Tomographic observations connecting convective downwellings with lithospheric source regions, Sierra Nevada, California
Considerable speculation has focused on the possible existence of convective downwellings associated with the Sierra Nevada, California. The 2005-2007 Sierra Nevada Earthscope Project (SNEP) occupied ~100 sites within the broader EarthScope Transportable Array using EarthScope FlexArray equipment. We observed 2000 events at 95 SNEP stations and 164 TA, permanent, and pre-SNEP Sierran experiment stations, yielding over 81,000 teleseismic P-wave arrival times picked with G. Pavlis's dbxcor waveform picking algorithm. We selected 27,000 arrivals for inversion both to equalize representation of different backazimuths and accommodate computational limitations. Using a teleseismic inversion code developed by S. Roecker that uses wavespeed gradients between nodes and calculates 3-D raypaths using a finite- difference algorithm, we find that we can recover lateral variations in wavespeed with very high resolution but the extent of sharp anomalies can become smeared vertically as far as one node spacing (~50 km). As expected, we image the large high-velocity anomalies previously seen in California, including the Isabella Anomaly (San Joaquin Valley) between about 70 and 250 km depth, the Redding anomaly under the eastern Sacramento Valley above about 200 km depth, and a Foothills Anomaly near the Moho under much of the western Sierra. The Foothills anomaly extends between the Redding and Isabella anomalies. At each end of the Foothills anomaly, the high-velocity body bends down to connect with the deeper, more vertical anomaly at its end. This is most striking at the north end, where a peculiar convex-upward portion of the anomalies appears to represent interaction of a convective downwelling like that at the south end of the Sierra with the clearly visible Gorda plate. This suggests that some active foundering of lithospheric material occurs in these locations. The eastern, high Sierra are underlain by lower velocity mantle; this mantle increases in velocity from south to north, suggesting more vigorous upwelling to the south. Whether or not the Foothills Anomaly represents material that formed in situ or has been thickened with material originally under the eastern Sierra remains unclear. These results strongly indicate that convective processes under continents are asymmetric and prone to complex interactions with other geologic entities.
S32B-07
Three-Dimensional Passive Seismic Imaging around the SAFOD Site, California, Using the Generalized Radon Transform
The generalized Radon transform (GRT) has been successful in exploration seismology in search for hydrocarbon reservoirs such as gas and oil. Recently, combined with the modern statistical methods, the GRT has been applied to image the structures in the interior of the Earth all the way from the upper mantle to the core-mantle boundary using global passive seismic data excited by large natural earthquakes [Wang et al., JGR, 2006; Ma et al., JGR, 2007; Van der Hilst et al., SCIENCE, 2007]. In this study, we applied a similar approach, i.e., GRT plus statistical models, to characterize the structure near the San Andreas Fault Observatory at Depth (SAFOD) site using ~560 local earthquakes recorded by PASO and HRSN network stations. We calculated travel time tables for each event and station using the Podvin and Lecomte (1991) finite-difference method and an updated version of the P-wave velocity model around the SAFOD site. An across-fault section at the SAFOD site obtained using GRT is similar to that found by steep-dip pre-stack seismic migration using active seismic reflection and refraction data [Bleibinhaus et al., 2007]. This result demonstrates that passive source GRT imaging (with local seismicity) can provide 3D images of similar quality as 2D or 3D active source surveys, but at a fraction of the cost for the 3D case. The steep reflectors to the southwest of the San Andreas fault, some of which continue along the strike of the fault, may indicate multiple intrusive cycles within the granite.
S32B-08
Studies of Atmospheric Sources and Propagation Using the USArray
Although the USArray was designed to illuminate the Earth's internal structure, it also can be useful for studies of the atmosphere and atmospheric phenomena. Specifically, studies of infrasound events, or subaudible acoustic events, must contend with the variable structure of the atmosphere. Atmospheric propagation is determined mostly by the atmospheric temperature and the wind, both of which vary spatially and continuously with time. The global suite of infrasound stations remains sparse, complicating efforts to understand propagation and thus use infrasound stations to characterize sources and image the atmospheric velocity structure. A key to solving this problem is recording infrasound signals at more points on the ground. The relatively densely spaced and extensive USArray may be ideal for this task. To demonstrate the value of the seismometers in the USArray for atmospheric infrasound studies, we report the results of a study of the shuttle Atlantis as it passed over the USArray in southern California in 2007 at supersonic speed. Seismic signals from the shuttle were recorded by over 100 stations in the USArray and are presumed to be due to acoustic-to-seismic coupling. We also report preliminary findings from a study of a large bolide that exploded over Oregon and was recorded by over 100 seismic stations in the USArray and four infrasound stations. Most of the analyzed data in these studies come from seismometers. Although we gained knowledge from the two studies using travel times from the seismic data, because the efficiency of the conversion from acoustic-to-seismic is dependent on many factors that are usually unknown a priori, the amplitude and detailed character of the associated acoustic signal is generally unknown. Collocation of the USArray seismometers with infrasound sensors could lead to an unprecedented jump in our understanding of the atmosphere, atmospheric phenomena, and the mechanical coupling between the atmosphere and solid Earth.