S24A-01 INVITED
Ambient Noise Interferometry and Surface Wave Array Tomography: Promises and Problems
In the late 1990ies most seismologists would have frowned at the possibility of doing high-resolution surface wave tomography with noise instead of with signal associated with ballistic source-receiver propagation. Some may still do, but surface wave tomography with Green's functions estimated through ambient noise interferometry ('sourceless tomography') has transformed from a curiosity into one of the (almost) standard tools for analysis of data from dense seismograph arrays. Indeed, spectacular applications of ambient noise surface wave tomography have recently been published. For example, application to data from arrays in SE Tibet revealed structures in the crust beneath the Tibetan plateau that could not be resolved by traditional tomography (Yao et al., GJI, 2006, 2008). While the approach is conceptually simple, in application the proverbial devil is in the detail. Full reconstruction of the Green's function requires that the wavefields used are diffusive and that ambient noise energy is evenly distributed in the spatial dimensions of interest. In the field, these conditions are not usually met, and (frequency dependent) non-uniformity of the noise sources may lead to incomplete reconstruction of the Green's function. Furthermore, ambient noise distributions can be time-dependent, and seasonal variations have been documented. Naive use of empirical Green's functions may produce (unknown) bias in the tomographic models. The degrading effect on EGFs of the directionality of noise distribution forms particular challenges for applications beyond isotropic surface wave inversions, such as inversions for (azimuthal) anisotropy and attempts to use higher modes (or body waves). Incomplete Green's function reconstruction can (probably) not be prevented, but it may be possible to reduce the problem and - at least - understand the degree of incomplete reconstruction and prevent it from degrading the tomographic model. We will present examples of Rayleigh wave inversions and discuss strategies to mitigate effects of incomplete Green's function reconstruction on tomographic images.
S24A-02
Crustal and Uppermost Mantle Shear Wave Azimuthal Anisotropy in the Western United States Based on Ambient Noise Cross Correlation and Eikonal Tomography
We investigate crustal and uppermost mantle shear wave azimuthal anisotropy in the western US based on 10 to 34 sec period Rayleigh wave phase velocity measurements derived from ambient noise cross- correlations. To date, more than 500 stations, mainly EarthScope/USArray Transportable array (TA) stations, have been deployed in the region and all available vertical component ambient noise records are cross- correlated to obtain the empirical Rayleigh wave Green¡¦s function as well as the phase travel time between each station pair. Taking advantage of the TA, we have developed a new tomography method, Eikonal tomography, which estimates the azimuthally dependent phase velocity at each location based on calculating the gradient of the empirical phase travel time surface constructed from the available phase travel time measurements. The most significant advantages of Eikonal tomography compared with traditional straight-ray tomography is its ability to produce meaningful uncertainty information about the inferred phase speed maps and its production of more reliable information about azimuthal anisotropy. The 2-psi azimuthal anisotropy signals are clearly observed for Rayleigh wave phase velocity throughout the region and the spatial variation of both the fast directions and the strengths of anisotropy require a structural cause. To investigate the depth distribution of shear velocity anisotropy, we inverted for a 3D shear wave velocity model with azimuthal anisotropy included in the middle/lower crust and the uppermost mantle. While the observed anisotropy appears to be decoupled between the two layers, the fast directions and the strengths of anisotropy in the uppermost mantle are broadly consistent with shear wave splitting measurements. This suggests that uppermost mantle structure plays a significant role in shear wave splitting measurements.
S24A-03
Anelastic Earth Structure from the Coherency of the Ambient Seismic Field
Cross correlation of the ambient seismic field is now routinely used to measure seismic wave travel times; however, relatively little attention has been paid to other information that could be extracted from these signals. In this paper we demonstrate the relationship between the spatial coherency of the ambient field and the elastodynamic Green's function in both time and frequency domains. Through measurement of the frequency domain coherency as a function of distance, we sequentially recover phase velocities and attenuation coefficients. From these measurements we generate 1D shear wave velocity and attenuation models for southern California. The ambient field measurements of attenuation and the exceptional path coverage that results from the many possible inter-station measurements, allow us to extend Q estimates over a range of frequencies that has previously been difficult to analyze using earthquake data. Measurements from paths that cross major sedimentary basins show both lower wave speeds and lower quality factors than other paths, as expected. Our results indicate that there is a wealth of information available in the spatial coherency of the ambient seismic field.
S24A-04
Attenuation tomography using ambient noise correlation
We present a method for resolving the seismic attenuation (Q) in the crust and upper mantle using the estimated Green's functions (EGF) obtained by the cross correlation of ambient noise data. Using the data from a PASSCAL experiment, we demonstrate that high resolution of the crust and upper mantle is possible using as little as a few months worth of data and that the resulting Q tomogram rivals the velocity tomogram obtained from the same data set. Finally, we validate the methodology using synthetic seismograms and investigate the contributions of scattering and intrinsic attenuation to the observed result. Measurement of Q at regional distances typically involves calculating spectral amplitude ratios of seismic waves radiated by earthquakes. This requires either a good knowledge of the source (including location, depth, mechanism, radiation pattern, geometrical spreading function and initial amplitude) or a choice of geometries in which source effects cancel out. Using the noise correlation methodology all of these terms are either known exactly or well defined. We calculated the EGF's for a year's worth of data between each pair of stations in the Ristra experiment (a linear array which stretched from Texas to Utah). We inverted the resulting waveforms for shear velocity and also measured the time-domain RMS amplitudes to invert for Q. Combining all our amplitude measurements, we were able to create a detailed profile of attenuation beneath the array. Over 1200 unique amplitude measurements went into the inversion (the others failed based on signal to noise criteria). High attenuation directly beneath the Rio Grande Rift and at depth beneath the Laramide and Paleozoic uplifts correlate precisely with low velocity features in the velocity tomogram. In addition, the upwelling mantle beneath the rift is particularly evident and highly structured in the attenuation result. Especially interesting is a sharp, thin, highly attenuating feature directly beneath the Jemez Lineament (the most active volcanic feature in the Southwest US) which is connected to the upwelling mantle. This feature is not well resolved on the velocity tomogram, suggesting that we have better resolution from the amplitude measurements than from the surface wave inversion.