S44B-01
USArray Receiver Function Images of the Lithosphere in the Western U.S.
We have begun making P- and S-wave receiver functions from the data available from USArray Transportable Array stations to examine lithospheric and upper mantle structure beneath the western US. At present the Transportable Array extends as far west as ~108o to 113oW, crossing the Snake River Plain and Columbia flood basalts, most of the Basin and Range, part of the Rocky Mountains, and about half of the Colorado Plateau. Our preliminary results are for PdS receiver functions: We constructed a common conversion point (CCP) stacked volume for the western U.S. using a standard reference velocity model for ray tracing and a 4th root stacking method to enhance lateral correlation. The stack contains ~12,000 receiver functions from 175 earthquakes recorded at 331 stations. The preliminary PdS stack shows a number of interesting lithospheric features, particularly the lithosphere- asthenosphere boundary (LAB). We identify the LAB as a negative polarity event at various depths beneath the Moho, above the zone of contamination from crustal multiples. In northwestern Nevada near (120o to 117o W, 40o to 42o N), the LAB appears to shallow almost to Moho depths (~30 km) in the region of modern Basin and Range extension. Beneath the rest of the Basin and Range the LAB is at about 60km depth. These results are in good agreement with SdP receiver functions by Li et al., 2007. To the north beneath the Snake River plain and Columbia plateau, the LAB is deeper, broader, and centered at about ~75 km depth. The southernmost Sierra Nevada, Mojave block, and southern Basin and Range have an eastward thickening lithosphere, with the LAB increasing in depth from ~50km to 70km under the southwestern edge of the Colorado Plateau. The Sierra Nevada has complications in the lower crust and upper mantle previously identified by Zandt et al., 2004. Further work will include incorporating a laterally variable crustal model derived from active source studies to recalculate the CCP stacks, and comparison of linear and 4th root CCP stacks. Working with other datasets we have observed that an accurate crustal velocity model can reposition Moho depth by several times the formal depth error obtained from a general reference model. This is also presumably true for the depth to the LAB.
S44B-02 INVITED
USArray Enables Improved Resolution Studies of Earth's Upper and Lower Mantle
The first full deployment of EarthScope's USArray has now been completed, with over 400 high-quality broadband stations distributed over a 70-km grid covering the western U.S. The openly-available USArray data can be obtained in near real-time using convenient IRIS-hosted data retrieval tools (e.g., SOD, JWeed). The dramatic increase in number of seismograms for each event, along with the unusual combination of large aperture and dense wave field sampling, are enabling improved resolution of deep Earth structure in diverse applications. Automation is now important for most initial data processing steps, in order to handle the large numbers of records that quickly accumulate. We highlight several applications of USArray data to Earth structure investigations in locations remote from North American lithosphere where the bulk of applications are focused. Earthquake source distributions relative to the USArray define several target study regions. For example, deep focus South American and Fiji-Tonga earthquakes allow high-resolution analyses of lower mantle structure beneath the Cocos Plate and the central Pacific, respectively. The increasingly detailed models that are inferred have many thermal, compositional and dynamical implications. Fiji-Tonga events recorded by USArray are enabling detailed mapping of upper mantle discontinuity structures in the vicinity of the Tonga slab. Images of upper mantle discontinuities beneath South America and Hawaii are brought into sharper focus using teleseismic data recorded by USArray. Various data processing approaches, such as stacking, migration, and source wavelet deconvolution are now stabilized by the dense USArray data sets. The increasingly detailed characterization of deep structure is motivating both advances in 3D wave propagation methods and collaborative interdisciplinary efforts to provide viable interpretations.
S44B-03 INVITED
Exploring the Seismic Wavefield with EarthScope's Transportable Array
The unprecedented aperture and dense station spacing of EarthScope's Transportable Array (TA) provide opportunities to explore subtle features of the seismic wave field using array methods, including stacking to enhance signals from small, distant events, and using spatial animations and slowness spectra to investigate the origin and nature of long-period (tens to hundreds of seconds) seismic coda. We illustrate the TA's capabilities using teleseismic short-period P waves and complete broadband ground motion recordings. The high quality and uniform installation procedures of the TA produce many stations with superb short-period signal- to-noise ratios enabling direct detection of tiny amplitude (~2nm) P-waves from the mb 4.2, 09 October, 2006 North Korean nuclear test. Stacking of many TA signals resolves subtle complexity in the source signals. Guided by seismic wavefield animations from large earthquakes, which display many wavefronts sweeping across the array, visual inspection and seismic slowness spectra of the extended coda resolve off-great circle Rayleigh-wave arrivals. The scattered waves approach the array from preferred directions depending on the source location relative to the TA. For example, scattered signals from the 01 April, 2007 Solomon Islands earthquakes are preferentially located to the west and northwest, with timing that suggests scattering from subduction zones along the margin of the western and north Pacific ocean. For the 16 August, 2007 Central Peru earthquake, Rayleigh wave coda contains significant energy arriving from the north and northeast, scattered near the North American margin. Analysis of scattered arrivals may lead to improved constraints on lateral gradients in lithospheric structures.
S44B-04
New Data Products Available at the IRIS Data Management Center
With USArray data processing and services in full production, the IRIS Data Management Center (DMC) has developed an online tool to function as a library for products derived from raw USArray observations as well as other catalogs with broad interest produced and supplied by the geoscience community. The service, called SPADE, is a flexible, annotated cataloging system for storing heterogeneous data products supplied by registered providers. SPADE includes a web interface that allows users to search across the spectrum of varied data products using geographic, temporal, and/or keyword parameters to locate products of specific interest. In addition to cooperative product submissions from outside institutions, the IRIS DMC is collaborating and developing frameworks to produce and to provide interesting products that take advantage of data from USArray sensors. These products are automatically produced in near- real time following interesting seismic events, and broaden the range of information that IRIS delivers to the geoscience community. In this presentation we will demonstrate the SPADE query interface and present examples of the available data products such as Global Centroid Moment Tensor solutions from Harvard and Columbia Universities, historic earthquake data scanned from select WWSSN film chips, as well as IRIS-produced seismic record sections, and ground-motion animations derived from Transportable Array observations. http://www.iris.edu/spade/
S44B-05
Calibration of USArray-TA Stations using Earth Tides
Full deployment of the 400 stations comprising the transportable array (TA) portion of the USArray component of the NSF Earthscope program was accomplished during the summer of 2007 and equipment rollover to new sites has become a reality. These stations currently form a grid covering the continental US from the Pacific coast to the eastern state borders of Arizona, Utah and Idaho. Seismic data from TA stations is transmitted from the field to the Array Network Facility (ANF) located at UCSD and then onwards to the IRIS Data Management Center which distributes it to the scientific community at large. The current and future data collected by this mostly homogeneous network, consisting of 3-component broadband seismometers and high dynamic range digitizers installed within weather-proof shallow vaults with sensors deployed about 2 m below the surface, will constitute a great data set for years to come. Currently the instrument response of TA stations is derived from the specifications of the equipment manufacturers, which is believed to be reliable. However, in attempt to assess the accuracy of the published response information of these stations we will measure the amplitude and phase of the Earth's tides recorded by TA stations by comparing these measurements with values expected from the models of the elastic, solid earth derived from tidal and free oscillation measurements and from modern ocean tidal models based upon analysis of satellite data. We will present results from stations with continuous recordings of at least 60 days from each station in which the earth's tides are observed.
S44B-06
Seismic Attenuation Observations: Signals and Artifacts
In this study we estimate to what degree seismic velocity heterogeneity contaminates seismic attenuation measurements for body waves. Comparisons between anelastic attenuation (per cycle energy loss) and seismic velocity can be used to infer temperature or chemical heterogeneity in the Earth. Unfortunately, many observations of seismic attenuation may not represent true anelastic attenuation if proper precautions are not taken. By measuring the seismic attenuation for attenuation-free 2.5D axisymetric synthetic S waves we can estimate how much elastic velocity heterogeneity contaminates attenuation observations. We perform multiple tests representing varying idealized and seismically observed structures in the Earth. While all S-wave attenuation measurements for seismic waves that travel through a heterogeneous body are contaminated to some degree, the severity of the contamination can be greatly reduced by taking three steps: (1) Determine differential attenuation for as many phase pair combinations as possible to increase reliability and redundancy. (2) Properly align/window each wave prior to measuring differential attenuation. (3) Only accept attenuation measurements for highly correlated (R2 > 0.9) wave pairs. These three steps minimize the observed contamination in synthetic S waves, even for cases of severe velocity heterogeneity. We have developed a fast and easy interactive tool for measuring differential attenuation with large seismic networks (e.g., EarthScope) following the three criteria listed above. Application of this program to USArray broadband data for both P and S waves provides sufficient data to invert for spatial variations in the quality factor Q. We present these attenuation results for the Western United States.
S44B-07
Attenuation Tomography of Western North America Using USArray Body Waves
We present P- and S-wave attenuation models for western North America, using the latest up-to-date EarthScope USArray seismic data. Tomographic models of P and S waves are done separately, using a least-squares inversion with a volume parameterization beneath the USArray stations. We find that there is strong correlation between the P and S models, as well as with corresponding velocity models, which match up well with the tectonic structure of the western US. For instance, both show high-attenuation beneath Nevada down to 150 km, in agreement with previous attenuation models done with techniques like Lg-wave attenuation. The seismic attenuation measurements are found in a relative sense, through a multi-waveform cross- correlation method that compares the t* value of a set of like phases (such as S, ScS, SS, P, PP, sS, pP, etc.). The process is automated, so that large numbers of phases can be analyzed quickly. This automation involves removing waveforms with low signal-to-noise ratios by making sure that they reasonably correlate with some minimum number of other phases. This is done in both the time and spectral domains. We also determine velocity anomalies using the same data set, and use the relative t* values to correct the travel-time delays, as we did in Fisher et al. [2003]. In addition, we also use differential t* values by comparing different phases from the same seismograms (like ScS/S, SS/S, PP/P, etc.), as we did in Lawrence and Wysession [2006]. As USArray moves across the country, we will continue to extend our tomographic imaging eastward.