S22A-01
Strong Southern California Crustal Heterogeneity Revealed by Adjoint Tomography
Adjoint tomography utilizes 3D simulations of seismic wave propagation in conjunction with a tomographic technique based on adjoint methods. We begin with an initial 3D model of shear and compressional wavespeeds for southern California provided by the Southern California Earthquake Center (SCEC; model CVM-H), extending to a depth of 60~km. We use the spectral-element method to simulate 140 good-quality local earthquakes, each recorded by as many as 160 stations. We compute misfits between observed and synthetic seismograms by using a new automated time-window selection algorithm that picks any time window within which the data and 3D synthetics are reasonably similar (e.g., P, S, Love, and Rayleigh waves). Within each time window we measure a frequency-dependent traveltime anomaly. For each record with a measurement, we compute an adjoint source that is used to create an adjoint wavefield. The interaction between the adjoint wavefield and the regular wavefield forms the gradient of the misfit function for one event. These gradients are combined using a source subspace projection method to compute a model update. We are presently on the tenth iteration of a southern California crust and upper mantle model. Thus far we have applied changes in excess of ± 20% from the initial 3D model. With each iteration, the changes in wavespeeds have improved the data fit, and we are able to include additional seismograms whose fits to the data for previous model iterations were too poor for selection. In general, the tomographic results compare well with surface geology, the most striking features being the low wavespeeds of the southern San Joaquin basin, the high wavespeeds and depth extent of the Santa Monica Mountains, the low wavespeeds in the Coast Ranges, the low wavespeeds in the eastern Mojave shear zone, and the sharp contrast at the eastern front of the Sierra Nevada due to volcanism in the Coso Junction area. Having applied relatively large-scale, large-amplitude changes, we can now exploit more of the shorter-period measurements to resolve and reveal km-scale features in the southern California crust.
S22A-02
A California Statewide Three-dimensional Seismic Velocity Model From Both Absolute and Differential Times
We present a tomographic model of the P- and S-wave velocity structure of the crust and uppermost mantle of California. The dataset combines first-arrival times from natural earthquakes (including composite events derived from many nearby earthquakes), explosions, and identified quarry blasts, with differential times from both catalog picks and waveform cross-correlation data. We apply a regional-scale double-difference tomography algorithm, which incorporates a finite-difference travel-time calculator and spatial smoothing constraints. We start with a coarse 3D velocity model with uniform 30 km horizontal node spacing and depth nodes positioned at -1, 1, 4, 8, 14, 20, 27, 35 and 45 km (relative to mean sea level). This coarse model is computed using only absolute picks starting from a 1D velocity model and an a priori Moho boundary based on previous studies. In order to use differential times to obtain a finer-scale model, we split the whole state into 5 sub-regions, each of which has a horizontal gridding of 10 km and the same depth layers as the coarse model. Starting with the coarse 3D model, we invert the differential times (both catalog differential times and waveform cross-correlation data) in each sub-region and all the absolute times from the entire state. Our final statewide model is a "stitched" version of the 5 sub-region models. Our model extends 540 km in the SW-NE direction and 1320 km in the NW-SE direction. We are able to image the principal features present in previous separate regional models for northern and southern California, such as the high-velocity subducting Gorda Plate, the deep low-velocity basins of the Great Valley, the Ventura basin, and the Los Angeles basin, and a high-velocity body in the lower crust underlying the Great Valley. The new statewide model has improved areal coverage compared to previous models, and extends to greater depth due to the inclusion of substantial data at large epicentral distances. This model can be applied to a variety of regional-scale studies in California, such as developing a unified statewide earthquake location catalog and performing regional waveform modeling.
S22A-03
New Results from WOMBAT: an Ongoing Program of Passive Seismic Array Deployment in Australia
The WOMBAT experiment is an ambitious program of rolling passive seismic array deployments designed to cover a significant portion of the Australian continent. Over the last decade, ~450 sites have been occupied in southeast Australia during the course of 11 separate deployments lasting up to one year. In each case, short period seismometers with a natural frequency of 1 Hz are deployed with spacing varying between 15 and 50 km. The large volume of recorded data is ideal for various classes of study, including teleseismic tomography, crustal receiver functions, ambient noise tomography and array seismology (e.g. probing fine scale core structure). In this presentation, a variety of new results from WOMBAT will be presented. These include (1) joint passive and active source tomography of the Tasmanian lithosphere; (2) combined array teleseismic tomography from the mainland; and (3) ambient noise tomography from multiple arrays. A limitation of conventional teleseismic tomography is that crustal structure is usually poorly resolved, which requires mitigation via station correction terms or the inclusion of a priori structure. In Tasmania, both teleseismic data from WOMBAT and 3-D wide-angle data from a previous experiment are available, a combination which promises good resolution throughout the full lithospheric thickness. We simultaneously invert both passive and active source data for the 3-D structure of the Tasmanian lithosphere, and show that there is little evidence for Tasmania comprising two continental fragments that were juxtaposed during the Phanerozoic. To date, teleseismic data from five mainland arrays that span much of Victoria, southern NSW and eastern South Australia have been combined in a single inversion for 3-D P-wavespeed. The resulting model contains many well resolved features, but the most pronounced is a strong velocity contrast between the Proterozoic lithosphere that underpins the Delamarian Orogen in the west and the Proterozoic lithosphere beneath the Lachlan Orogen in the east. Rayleigh wave group velocity maps of the mid-upper crust derived from ambient noise tomography reveal significant variations in wavespeed near the edges of the Murray Basin, but show little evidence for a transition between the two orogens at shallow depths.
S22A-04
A Full-Wave Seismic Tomography for the Crustal Structure in the Metropolitan Beijing Region
The greater Beijing metropolitan region is located in an old cratonic block in northeast China with complex geology and several large historic earthquakes, such as the Sanhe-Pinggu earthquake (~M8.0) in 1679, the Xingtai earthquake (M7.2) in 1966, and the Tangshan earthquake (M7.8) in 1976. To enhance our understanding of the crustal structure and the seismotectonics under this region, we conduct a full-wave three-dimensional (3D) tomographic study of this region using the waveforms recorded by the newly established Beijing metropolitan digital seismic network. Since the Beijing network was put into operation in October 2001, there have been 89 local earthquakes of magnitude 3.0 and above. From these, we selected 23 events of magnitude 3.2 and above and obtained their waveform records at 50 stations within our area of interest. The types of instruments at these stations include broadband, short-period and very broadband. First-motion focal mechanisms were determined for these events. We used a regional 3D model obtained by seismic reflection surveys as the reference model and calculated the synthetic seismograms by the finite-difference method. In this first attempt at finite- frequency tomography for the Beijing region, we focus on the variation of the P-wave speed using the first- arriving P waves. We measure the frequency-dependent traveltime anomalies of the P waves by the cross- correlation between observed and synthetic P waveforms within several discrete frequency bands between 20-sec and 5-sec periods. The sensitivity or Fr¨Śchet kernels of these measurements for the perturbations in P-wave speed were computed by the same finite-difference method. We will present the preliminary result in our full-wave seismic tomography for the Beijing region.
S22A-05
Joint Inversion of P- and S-Wave Receiver Functions and Surface-Wave Dispersion Velocities
Joint inversion of P-wave receiver functions and surface-wave dispersion velocities is commonly utilized to obtain the S-velocity variation with depth beneath a recording seismic station. P-wave receiver functions contain secondary phases generated by the impingement of a teleseismic P-wavefront with subsurface seismic discontinuities, and these secondary phases constrain S-P travel-times and S-wave velocity contrasts across the discontinuities. Dispersion velocities are obtained from surface-waves propagating at frequency- dependent velocities and constrain absolute S-wave velocity averages within depth-ranges that depend on frequency. The combination bridges resolution gaps between the data sets while integrating their independent information into robust estimates for the S-velocity structure of the crust and uppermost mantle. Constraints on deeper structure, however, are generally less robust. Lithospheric thickness, for instance, is weakly constrained by the combined data set as P-to-S refracted waves from the base of the lithosphere generally interfere with reverberations in the crustal layer, and dispersion velocities do not constrain sharp discontinuities. We have modified the joint inversion procedure to include S-wave receiver function waveforms. Similar to P-wave receiver functions, S-wave receiver functions also contain secondary seismic phases, which result from the interaction of a teleseismic S-wavefront with seismic discontinuities under the recording station. However, in contrast to P-wave receiver functions, S-to-P refracted conversions on S-wave receiver functions are recorded before the incident S-wave and they never interfere with crustal reverberations. We have tested our modified procedure with synthetics and at a number of stations in South Africa. Our results show that the addition of S-wave receiver functions into the joint inversion scheme effectively decouples the signature from the base of the lithosphere from the crust and provides robust constraints on the depth, size, and sharpness of the lithosphere-asthenosphere boundary.
S22A-06
Tomographic Errors From Wavefront Healing
Despite recent advances in full-waveform modeling ray theory is still, for good reasons, the preferred method in global tomography. It is well known that ray theory is most accurate for anomalies that are large compared to the wavelength. Exactly what errors result from the failure of this assumption is less well understood, in spite of the fact that anomalies found in the Earth from ray-based tomography methods are often outside the regime in which ray theory is known to be valid. Using the spectral element method, we have computed exact delay times and compared them to ray-theoretical traveltimes for two classic anomalies, one large and disk-shaped near the core mantle boundary, and the other a plume-like structure extending throughout the mantle. Wavefront healing is apparent in the traveltime anomalies generated by these structures; its effects are strongly asymmetric between P and S arrivals due to wavelength differences and source directionality. Simple computations in two dimensions allow us to develop the intuition necessary to understand how diffractions around the anomalies explain these results. When inverting the exact travel time anomalies with ray theory we expect wavefront healing to have a strong influence on the resulting structures. We anticipate that the asymmetry will be of particular importance in anomalies in the bulk velocity structure.
S22A-07
Finite-frequency effects in surface wave tomography in southern Africa
Ryaleigh wave phase velocities in southern Africa are computed from a two-plane-wave method using 2-D sensitivity kernels. Lower phase velocities are continuously imaged in the Namaqua-Natal Belt and the Kheiss Belt from 30 s to 167 s. Phase velocity in the Kaapvaal craton is generally higher than the average. In the Bushveld Complex, phase velocity anomaly is negative at periods below 59 s and positive above 100 s. Compared with results from ray-theoretical tomography, phase velocity images from finite frequency tomography are more coherent among adjacent periods, especially for those that are greater than 100 s. The positive correlation between the images from the finite-frequency tomography and the ray-theoretical tomography holds well to 125 s, although the correlation becomes worse at longer periods. Unlike in global surface wave tomography where magnitude of anomaly is stronger in the finite-frequency model, the variation of phase velocity anomaly in southern Africa is largely compatible in the finite-frequency and the ray- theoretical tomography models. Given that the periods used in the two-plane-wave surface wave tomography are generally below 150 s, finite frequency effects in those tomography studies are not significant.
S22A-08
Parameterizing surface-wave tomographic models with harmonic spherical splines
We present a mathematical framework and a new methodology for the
parameterization of surface-wave phase-speed models based on
travel-time data. Our method is neither purely local like block-based
approaches, nor is it purely global like those based on spherical
harmonic basis functions. Rather, it combines the well-known theory
and practical utility of the spherical harmonics with the spatial
localization properties of spline basis functions. We derive the
theoretical foundations for the application of harmonic spherical
splines to surface-wave tomography and summarize the results of
numerous numerical tests illustrating the performance of a practical
inversion scheme based upon them. Our presentation is based on the
notion of reproducing-kernel Hilbert spaces, which lends itself to the
parameterization of fully-three dimensional tomographic earth models
including body waves as well.
http://geoweb.princeton.edu/people/simons/Amirbekyan+2008-GJI.html