S41B-0555
Crustal structure of the northern Basin and Range province, Nevada: A new look
A major objective of the EarthScope Transportable Array (TA) is to investigate the crustal structure of the United States with uniformly spaced seismic coverage across the entire continent. Receiver functions (RFs) provide one means of estimating bulk crustal values of Vp/Vs and crustal thickness from the TA teleseismic data. Although the ~70 km average station spacing is too sparse to stack the data, the uniform coverage provides an opportunity to correlate crustal Ps phases and their multiples across a tectonic province such as the Basin and Range. We have calculated receiver functions from ~1 year of data for the TA stations in Nevada between latitude 37--42° N. The best results obtained were from the O and N lines located in northern Nevada. This particular corridor has been investigated in detail by the 1986 COCORP seismic reflection transect and the 1988-- 1989 PASSCAL experiments. This provides us with an opportunity to compare the correlated RF results to prior work in the area as well as crustal thickness estimates calculated by the EarthScope Automated Receiver Survey (EARS) system. Preliminary results show that most of the stations in the O and N lines have a high amplitude positive arrival, interpreted as the Moho Ps phase, that correlates across the entire state. The 2P1S Moho multiple is strong enough to correlate across the lines and the two phases can be used to estimate Vp/Vs and Moho depth across the transect. The Vp/Vs estimates range between 1.66 and 1.88, with most lying between 1.73 and 1.84. Crustal thickness estimates range from 28 km to 41 km, with most stations exhibiting a depth of 31--34 km. Other crustal features observed in the RFs include an upper crustal low-velocity layer at ~10 km depth and a mid-crustal high velocity layer located between 15-25 km depth. A new result is the observation of a lower crustal low-velocity layer at many stations. The top of this layer may correlate with the zone of lower crustal reflectivity observed in the COCORP lines. Our results roughly agree with previous active source experiments in that Moho depth increases from west to east and from south to north. With a few exceptions the EARS results for crustal thickness are generally within ~2 km of our estimates.
S41B-0556
Shear-Wave Splitting in the Great Basin
The Great Basin region of western North America has undergone as much as 200% extension over the past 20 M.y., in a predominantly east-west direction. The goal of this study is to better understand the causes and processes for wide spread extension in the region. To this end, we present results of anisotropic structure imaging via shear wave splitting analyses using both long-lived stations and the USArray Transportable Array (TA). Over 75 TA stations exist in the region with data spanning back to February 2006, while 8 permanent broadband stations in the Great Basin have recorded data for over 14 years. We processed seismic waveforms for SKS splitting analysis using a combination of the rotation-correlation, minimum energy, and eigenvalue methods. Initial results for this study exhibit predominantly east-west fast polarization directions across the northern and central Great Basin, consistent with both the primary direction of extension and with the direction of plate motion in a global hotspot reference frame. However, fast directions at most stations exhibit significant variations over short interstation distances (16 km to 24 km) and over small changes in back azimuth. Splitting times range from 0.1 to 2.2 seconds, and average approximately 1.0 sec. Some variability in splitting times between stations was observed; southerly stations (TPH and TPNV) exhibited significantly smaller splitting times than were observed elsewhere. The combination of strongly varying fast polarization directions and large splitting times is indicative of a complex anisotropic structure that varies both laterally and with depth, likely caused by a combination of lithospheric and asthenospheric structures. Candidate complexity styles include multiple layers of anisotropy, dipping layers, and/or small scale heterogeneities in subsurface structure. To elucidate the nature of the anisotropic complexity, we are currently engaged in further study by expanding our shear-wave splitting analysis to include recent data from all USArray stations in the Great Basin, as well as using receiver function analysis to provide better constraints on the nature of vertically varying structures.
S41B-0557
Complex Upper Mantle Seismic Structure Across the Southern Colorado Plateau / Basin and Range I: Results from Shear Wave Splitting Analysis
The origin of the stark transition from the highly extended and deformed geologic terrane of the southern Basin and Range to the stable, laterally undeformed region beneath the Colorado Plateau has been a long-standing question in understanding the tectonic evolution of the southwestern United States. To provide new constraints for this enigmatic region, we are utilizing the excellent regional coverage afforded by the USArray Transportable Array (TA) and other permanent and temporary broadband seismic stations to image anisotropic crust and mantle structure. This study focuses on our results from shear wave splitting analyses, but is part of a broader study of regional seismic anisotropy that includes receiver function analyses. As of summer 2007, we have performed shear wave splitting analyses on over 7,500 broadband waveforms. Our results provide the first clear evidence for anisotropic complexity laterally and with depth for this region. These variations appear to be correlated with regional geologic terranes. Splitting times range from 0.6 sec to 2.7 sec and are generally larger for Basin and Range stations. For stations located within the southern Basin and Range, fast directions are oriented from ENE-WSW to N-S. While regional similarities exist for some areas, significant variations exist over spatial scales of 70 km or less for events from similar backazimuths. We also observe well-resolved differences in fast direction with event backazimuth. Conversely, fast directions for stations located on the Colorado exhibit a more limited range of fast directions from NNE-SSW to NE-SW. Splitting for stations near the Arizona Transition Zone, between the Colorado Plateau and southern Basin and Range, tend to reflect the range in fast directions exhibited by the southern Basin and Range. The regional variations in shear wave splitting likely result from a combination of crust, mantle lithosphere, and asthenospheric fabric. The large splitting times suggest that sublithospheric anisotropy plays a strong role. Additionally, we note that receiver function imaging of the 660 km mantle discontinuity in this area exhibits backazimuthal dependence, suggesting contributions from deep upper mantle structure to the observed splitting complexity. Future work will include the analysis of additional TA waveforms, the distribution of null splitting results, and understanding the relationship between shear wave splitting results and anisotropy apparent from receiver function analyses at these stations.
S41B-0558
Complex Upper Mantle Seismic Structure Across the Southern Colorado Plateau / Basin and Range II: Results from Receiver Function Analysis
Connections between surface tectonics and mantle dynamics have long been investigated. The current configuration of the Transportable Array (TA) component of USArray provides nearly uniform sampling of the upper mantle that underlies much of the tectonically active western United States. The ~70 km spacing of TA stations densely sample the upper mantle discontinuities at the nominal depths of 410 and 660 km with seismicity broadly distributed over a range of azimuths and distances providing excellent coverage. The upper mantle discontinuities are thought to mark phase transition of minerals within the mantle to denser configurations. Changes in the depths of the discontinuities can thus be used to infer thermal variations at specific depths in the mantle based on the pressure-temperature relation of the responsible phase transitions. Here we use receiver functions calculated by teleseismic events recorded by the TA to investigate the depths of the discontinuities in the southwestern United States focusing on the transition between the Basin and Range and the Colorado Plateau. Previous studies of the depths of these discontinuities in the western United States have found each interface to exhibit ~30 km of topography. Observations presented here indicate a similar amount of discontinuity topography with depths of the 660-km discontinuity varying from less than 640 km in the central Colorado Plateau to greater than 670 km in the southern Basin and Range. The 410-km discontinuity shallows to less than 405 km depth in the southern Basin and Range and reaches near 420 km depth in the Colorado Plateau. Exploring azimuth fluctuations in receiver functions sampling discontinuity structures from different directions indicates a greater amount of variability in the depth of the 660-km discontinuity when sampled from eastern versus western azimuths in the southern Basin and Range than in the Colorado Plateau. Attributing variations in discontinuity depths to anisotropy suggests a more complex pattern of anisotropy beneath the southern Basin and Range than the Colorado Plateau and that a significant component of that anisotropy may be focused towards the base of the upper mantle.
S41B-0559
Receiver Function Imaging of Upper Mantle Discontinuities Beneath the Oregon High Lava Plains and Surrounding Regions
The Pacific Northwestern United States has experienced wide-spread tectonomagmatism from the late Cenozoic to present-day, the reasons for which are likely linked to complex interactions between subduction zone processes near the edge of a Precambrian craton, possible hotspot effects, and mantle flow within the asthenosphere. The goal of this study is to examine upper mantle seismic velocity discontinuities beneath this region to investigate the connection between these geodynamic processes. We present results from analysis of Ps receiver functions using stations from the High Lava Plains temporary broadband seismic experiment, the USArray Transportable Array (TA), and other regional broadband stations for a total of over 108 stations in the region. We compute receiver functions at individual stations using an iterative deconvolution approach, and generate common conversion point (CCP) stacks using moveout corrections defined by the Tectonic North America (TNA) shear wave velocity model. Results from our imaging show that conversions from the 410 km discontinuity have narrow, high amplitude peaks across most of the region, with local complexity in peaks for some areas. Conversely, most conversions near the 660 km discontinuity exhibit peaks that are broad and less well-defined. Estimates in transition zone thickness show local variability, with notable areas of thickening in northeastern Oregon into Idaho and thinning under the coastal regions. The complexity in upper mantle discontinuity structure in this region is likely due to a combination of strong variations in isotropic and anisotropic seismic wavespeeds. The local complexities of the 410 km discontinuity correlate well with seismic wavespeed anomalies imaged by regional P-wave tomography. The source of the variations in the 660 km discontinuity are less apparent, but we note that seismic anisotropy is strong and regionally homogenous across much of the region, perhaps producing backazimuthal variations in migrated conversion depths. Transition zone thickness variations likely indicate that thermal properties of the upper mantle are related to the subducting Juan de Fuca slab and to the presence of warmer mantle under areas of the High Lava Plains. Our images are consistent with the suggestion of a zone of low seismic wavespeed anomalies associated with the slab window in central Oregon that is likely related to extensive magmatism in the back-arc region.
S41B-0560
Shear-wave splitting and seismic anisotropy in Oregon's High Lava Plains
The High Lava Plains (HLP) of Oregon constitutes a region of young (< 15 Ma), age-progressive volcanism that approximately mirrors the track of the Yellowstone hot spot. This age progression is oblique to North American absolute plate motion. Different models have been proposed to explain the age progression, generally invoking either lithospheric processes or flow in the asthenosphere linked to a Yellowstone plume head. In order to assess models of mantle flow that might explain HLP volcanism, we analyzed SKS splitting for 16 stations from the first phase of a large deployment of broadband instruments in the HLP. We also examined data from a number of other sites in the northwestern US, including USArray and permanent stations. We interpret shear wave splitting in terms of upper mantle deformation, with the fast polarization direction inferred to be parallel to mantle flow and delay time proportional to the strength of anisotropy. We have found significant splitting in the HLP, with large split times (up to ~ 2-2.5 sec) and generally east-west fast directions. The splitting pattern througout the region is simple, with no evidence for multiple layers of anisotropy and little geographical variation in fast directions. Since the lithosphere under the HLP is thin and the split times are large, the splitting signal is likely dominated by anisotropy in the asthenosphere. The east-west splitting direction is oblique to both the trace of the HLP and to absolute plate motion. Thus, our splitting observations are not consistent with models of HLP- parallel asthenospheric flow driving age-progressive volcanism. Any model for HLP formation and local upper mantle processes must therefore be consistent with both the NW-trending progression of recent volcanism and the large E-W deformation in the asthenosphere that is indicated by our splitting results.
S41B-0561
Mapping Lateral Heterogeneity with Wavefront Modeling of Rayleigh Waves in the High Lava Plains, Oregon
We have developed an analytical procedure that takes advantage of high-density broadband seismic networks to test directly for lateral homogeneity, the results from which can be used to map boundaries of laterally homogeneous subregions. This procedure, termed wavefront modeling, is a variant of the first phase of Forsyth's two-plane-wave modeling. In both methods, for a well-recorded teleseism, corrections for wavefront perturbations caused by heterogeneities along the path between the epicenter and the study area are calculated at a fixed period. The two methods differ as to how the perturbed wavefront is modeled. In our wavefront modeling, relative phase delays are calculated among all instrumentally-corrected vertical-component waveforms recorded in the study region. We do a nonlinear inversion of the phase delays to determine the azimuth of the best-fit single-plane-wave wavefront and the average phase velocity across the station array. Contour plots of calculated-minus-predicted wavefront phase delays provide a simple visual display of the degree of lateral homogeneity within the recording network. A high density of stations allows us to subdivide the study area into smaller regions and solve for the wavefront perturbations in each of these, with the objective of identifying laterally homogeneous subregions. 2-D phase-velocity maps derived by this procedure provide a robust input dataset for inversion to obtain the 3-D S-wave velocity structure. We are currently applying this analysis procedure to data recorded on and around the High Lava Plains (HLP) of central and eastern Oregon, southwestern Idaho, and northern Nevada. The HLP "hotspot" track is a prominent volcanic lineament that trends oblique to plate motion from the southeast corner of Oregon in the northern Great Basin to Newberry volcano in the eastern Cascades. The HLP region is instrumented by the 2004--2007 deployment of the USArray Transportable Array, with its ~70 km station spacing on a regular grid, and the 2006--2009 high-density HLP seismic deployment with ~25 km spacing. Preliminary results find that phase velocities in the HLP region are distinct from and slower than the Proterozoic Blue Mountain block to the north at a period of 68 s. Based on the sensitivity kernels, this corresponds to a similar relationship between the S-wave velocities at a depth of about 90 km.
S41B-0562
Upper mantle S-velocity variance beneath North America
With the implementation of continental-scale geologic initiatives such as EarthScope, increasing amounts of seismic data are continuously becoming available. We have analyzed and fit regional S and Rayleigh wave trains generated by events from the years 2000 through 2006, including waveforms from Transportable Array stations of USArray. We have combined the constraints provided by the new fits with those used for the 3D S-velocity model NA04 to create a new North American model, NA07. Additionally, we investigate the effects of data heterogeneity and model parameterization (ie. node spacing, flattening, damping) on the tomographic model. This allows for an appreciation and quantification of reasonable variances for NA07. Alternate models within NA07's estimated variance do not cover the entire null space, but they do provide a useful guide for model uncertainty, allowing for improved interpretations of the compositional, mineralogical and thermal state of the crust and mantle, which will ultimately provide more accurate insights concerning the causes and effects of upper mantle heterogeneity. Here we present model NA07, as well as the alternate models, which are created by removing varying sets of waveforms (such as USArray data) from our data set before inversion. Robust velocity features are identified as those features consistently modeled with similar properties in all models in the suite. The most prominently robust feature modeled is the North American Craton. Our current seismic-model suites define the craton with an average velocity of 4.71 km/s at a depth of 90 km. At 90 km, the suite has an average standard deviation of 17 m/s in the craton. At 200 km the craton averages 4.59 km/s with the model suite varying by 10 m/s. This significant decrease with depth within the lithosphere in the cratonic average for the S-velocity is consistent with an average steady-state geotherm inferred for regions with low heat flow around 40 mW/m2. The North American craton appears to be no thicker than about 250 km. The western United States exhibits a very thin lithosphere, underlain by a very low S-velocity asthenosphere that appears to extend down to depths around 150 km. Its average velocities range from 4.12 km/s at a depth of 80 km to 4.25 km/s at 120 km, but the lateral variations have standard deviations of 150 m/s at 80 km and 90 m/s at 120 km. In this region, the slowest velocities are modeled beneath the Northern Basin and Range and Gulf of California.
S41B-0563
Rayleigh and Love wave tomography of the western United States from ambient seismic noise using USArray
We present the most recent results of Rayleigh wave and Love wave tomography in the western United States using ambient seismic noise observed at 424 broadband stations from the EarthScope/USArray Transportable Array and regional networks. All available three-component time series between 1 October 2005 and 31 Aug 2007 have been cross-correlated to yield estimated empirical Rayleigh and Love wave Green's functions. Phase and group velocity dispersion curves for both Rayleigh and Love waves between 5 and 40 sec period are measured for each inter-station path by applying frequency-time analysis and are then used to invert for Rayleigh and Love wave speed maps. The significant velocity variations observed on the short period maps suggest that it may not be appropriate to use straight ray theory in these inversions. To investigate both the off-great-circle and finite- frequency effects, we apply a 2D finite difference wave propagation simulation combined with the adjoint method. We investigate the effects on ray geometry, travel time measurement, and tomography.
S41B-0564
Ambient seismic noise and teleseismic tomography in the western USA: High-resolution 3-D model of the crust and upper mantle from Earthscope/USArray
This study applies two new complementary methods of surface wave tomography, ambient noise tomography (ANT) and two-plane wave teleseismic tomography (TPWT), to the rapidly accruing data resources in the western US, predominantly from the Transportable Array (TA) component of EarthScope/USArray. Ambient noise tomography (ANT) is based on the extraction of empirical surface-wave Green functions by cross-correlating long sequences of ambient seismic noise. Two plane-wave tomography (TPWT) interprets the variation in amplitude and phase of teleseismic surface waves observed across a regional seismic array in terms of phase velocity variations within the foot-print of the array. Both methods measure surface wave dispersion, but in complementary period bands: ANT (6 - 40 sec) and TPWT (25 - 150 sec). Used in combination, the methods produce surface wave dispersion maps (Rayleigh and Love wave, group and phase velocity) across the western US from about 6 sec to 150 sec period on a 25-50 km geographic grid. Using the combined phase velocity dispersions at periods from 6 sec to 150 sec, we perform 3-D shear velocity inversion to obtain a high-resolution shear velocity model from surface to ~150 km in the W. US. The model possesses a wealth of features. We list only three examples here: the high velocity slab subducting beneath the Cascades in northern California and Oregon overlain by a low velocity upper mantle wedge; a low velocity anomaly beneath the eastern edge of the southern Sierra Nevada and the adjoining Walker Lane region underlain by a high velocity anomaly; and the well-known high velocity anomaly associated with downwelling lithosphere beneath the Transverse Range.
S41B-0565
Lg Amplitude Tomography in the Western United States
Lg Q can be used to isolate source effects, and determine magnitude (mbLg), and patterns have been shown to correlate with deformation age, geothermal circulation, and possibly partial melt within the crust. Tomographic studies of Lg Q are typically performed using a few stations and many earthquakes, the density of earthquakes providing spatial resolution. The unprecedented breadth and density of the USArray should allow more balanced data sets, and better resolution in aseismic regions. To test this, we processed amplitude data from 515 events recorded by the broad-band vertical channels of 395 stations within the initial footprint of the USArray. Events were obtained from PDE monthly and weekly catalogs, and were chosen to maximize station coverage, including consideration of the USArray deployment schedule, and a magnitude-distance criteria, selecting the potentially best recorded event per quarter-degree block. The IRIS DMC provided waveform data and instrument response information. We measured RMS amplitudes in overlapping octave width bands between 0.5 and 8 Hz, for windows defined by group velocities 3.6 to 3.0 km/s. Over 30,000 amplitudes passed a signal-to- noise threshold of 2. We then solved for laterally varying crustal Q, as well as site and source terms. Results for the 1-Hz band (0.75 – 1.5 Hz) show dramatic variations in Q ranging from below 100 to over 600. Low Q is observed throughout the Pacific region, including Coast and Cascade ranges. Lowest Q is found at the Geysers and in the Salton Trough. A large inland region of low Q extends from the Yellowstone area, through the Snake River Plain, to the Great Salt Lake and Sevier basins. High Q crust is observed for the Colorado Plateau, and the Columbia Basin. Smaller islands of high Q include the Sierra Nevada, higher altitude regions of Nevada, such as the Toiyabe, Toquima, Monitor, Hot Creek, Snake, Wilson Creek and Jarbidge ranges, the Bitterroot Range, the Snoqualmie Batholith, the central Willamette Valley, Antelope Valley in the western Mojave, and the Sierra Juarez. Feature outlines as small as 50 km appear to be resolved via visual correlation with topography; however, resolution calculations remain to be performed. Residuals average 0.16 log10 amplitude units. This is higher than for comparable studies in Asia, perhaps due to source radiation effects for the shorter USArray paths.
S41B-0566
Wave Gadiometry for USArray
Wave gradiometry is a new array analysis technique that makes use of the wave amplitude variation across a small aperture array to find the phase velocity, wave directionality and geometrical spreading. In this study we apply wave gradiometry to USArray data for the Western North America. A weighted inversion method is introduced to compute spatial gradients that are linearly related to the station spacing and wave amplitude differences based on a first order of Taylor series approximation. The amplitude differences of station pairs with large truncation errors due to higher order terms are down-weighted in the inversion equations. The reduced time method is also effectively applied in the data processing under the assumption that the media inside of the sub- array is homogeneous. Synthetic waveforms using a wave number integration method and Gaussian plane waves are used to successfully test the process. Waveforms from the Kurile Islands earthquake of 13 January 2007 are analyzed. Preliminary results show that the average phase velocity of 100-125 seconds Rayleigh waves is about 4.0 km/s with the prominent feature that the Basin and Range and the west coast are dominated by velocity lows and highs, respectively. The wave directionality map reveals refraction into the continent at the plate boundary. Future work will include using waveforms of earthquakes from different directions and distances to understand the variations seen in the parameter maps derived by wave gradiometry and then use results in tomographic structure studies.
S41B-0567
Scientific Visualizations of Data Collected From EarthScope's Seismic Observatory (USArray) and San Andreas Fault Observatory at Depth (SAFOD)
Looking at data from perspectives other than map view, or standard cross sections, can help researchers with their science. Interactively exploring visualizations of multi-dimensional data allows scientists to assess the quality of their data, identify links between different data types, assist with project planning, refine their hypotheses and more easily convey research findings to a wide range of audiences. Working with EarthScope scientists we explore ways to use visualization techniques to help researchers explore their data and explain key concepts and theories. Examples of our visualizations include: (1) Movies of the temporal evolution of earthquakes, detected and recorded by USArray stations, juxtaposed with the progress of USArray station deployment. (2) Using the USArray station spacing as an irregular grid we create a 3D mesh depicting displacements generated by teleseismic waves. (3) An interactive 3D visualization of data pertaining to the SAFOD observatory (i.e., drill hole plans, side tracks, surface and borehole experiment locations, geologic cross-sections, seismicity and fault planes). (4) Exploration of the temporal evolution of the Rayleigh wave group velocity dispersion throughout the California region. (5) Interactive 3D visualizations of notable earthquakes that include, but are not limited to, the location of the mainshock epicenter and hypocenter, historical seismicity, USArray seismic station locations and station codes, geographic boundaries and topography of the region. We make these visualizations available for free download on the web within a day or two of the mainshock event so they can be used in classrooms, outreach venues and for media response. These visualizations can be accessed from the visual objects library at the Scripps Institution of Oceanography's Visualization Center (http://siovizcenter.ucsd.edu/library.php). They include 3D interactive visualizations, Quicktime movies and online tools and can be explored using freeware that runs on multiple platforms (e.g., Windows, Mac OS X, Linux, SGI Irix). If you would like help visualizing your EarthScope data please contact us at vizinfo@ucsd.edu. http://siovizcenter.ucsd.edu
S41B-0568
Microseism Noise Sources observed by Earthscope USArray Transportable Array
The NSF Earthscope USArray Transportable Array deployment of 400+ high quality broadband telemetered seismic stations has created a unique opportunity to study the temporal and spatial microseism distribution as a function of time across the deployment from coastal regions to the continental interior. We examine the effects of geologic provinces and regional differences on power observed in the microseismic bands. We also integrate our results with satellite derived surface wave heights to identify microseismic source regions for different regions of the array.
S41B-0569
Mining Data, What a Blast!
Seismic network data processing involves a number of critical decisions which are a balance of available funding and manpower vs the amount and extent of data being processed. In an ideal world, any event detected by a given network would have associated arrivals and usually an associated origin. In this world of decreasing telecommunications costs, that has resulted in an ever increasing number of sensors and stations along with accessibility to ever expanding real-time data flow, this complete human data review is no longer a feasible reality with the existing personnel support. Decisions on catalogue inclusiveness are being made based on expediency and budget constraints rather than on a scientific or technical basis. One of the critical time sinks for an analyst is the location and discrimination of the large number of daily man-made blasts, whether they be from road construction, quarries, or mines. Given that mines exist in a given location it is possible to first, automatically assign event locations to blast sites in real-time, and second, to provide quick mine site associations on the post-real-time processing level. This reduces the analyst's job from a complete event location to simply verifying and correcting automatic detections. A study has been carried out using a grid of mine locations and running an event associator with automatic detections over this grid. Mine blasts are automatically located at the grid mine sites. This has been particularly successful with large blasts outside the network which were previously creating poor locations and necessitated analyst involvement to ensure that these events were not a seismic event within or near the boundaries of the network.