Multichannel inversion of scattered teleseismic body waves: Practical considerations and applicability

Abstract

We investigate the resolving power and applicability of a recently developed technique for multichannel inversion of scattered teleseismic body waves recorded at dense seismic arrays. The problem is posed for forward-and back-scattered wavefields generated at discontinuities in a 2D isotropic medium, with the backprojection operator cast as a generalized Radon transform (GRT). The approach allows for the treatment of incident plane waves from arbitrary backazimuths, and recovers estimates of material property perturbations about a smoothly varying reference model. An investigation of the main factors affecting resolution indicates that: (1) comprehensive source/station coverage is necessary to optimize geometrical resolution and recover accurate material property perturbations; (2) the range in dip resolution diminishes with increasing depth and is inversely proportional to array width (e.g., reaches [−45°,45°] at depths equivalent to~1/2 array width); (3) distortion of the image due to spatial aliasing is only significant at depths ≤2 × [station spacing]; and (4) unaccounted for departures from model assumptions (i.e., isotropy and 2D geometry) result in defocusing and mismapping of structure. Two applications to field data are presented. The first considers data from the Abitibi 1996 broadband array, in which stations were deployed at~20 km intervals. Imaging results show that this level of spatial sampling, which is characteristic of modern broadband arrays, is sufficient to adequately resolve structure below mid-crustal depths. For these data, we introduce a new preprocessing algorithm that uses eigenimage decomposition of seismic sections to suppress wavefield contamination by PcP and PP phases. The second application involves short period data from the Los Angeles Region Seismic Experiment and shows that images obtained from high frequency records are subject to significant contamination by scattered surface waves.

Distribution of hydrous minerals in the subduction system beneath Mexico

“…4a, the candidate phase assemblages which explain our deeper-depth ( 4 $ 100 km) data are gabbro ( $ 024% porosity) and eclogite ( $ 6% porosity). Observed dominant gabbroic phase at deeper depth ( 4 $ 100 km) is supported by the seismic image showing the sharp boundary between the top interface of the steeply subducting slab and the peridotitic mantle, which extends down to a depth of 200 km although the image does not show the oceanic Moho (bottom of the oceanic crust) due to the low frequency range in the analysis (0.03-0.3 Hz) and resolution limit (Rondenay et al, 2005;MacKenzie et al, 2010). It is generally understood that large amount of fluid is expelled from the downgoing hydrated crust as it undergoes subduction-zone metamorphism with increasing P and T, and the degree of dehydration strongly controls the reflectivity at the top and bottom of the oceanic crust (Kawakatsu and Watada, 2007).…”

“…The subducted Cocos plate beneath central Mexico, as imaged with both RFs and teleseismic migration (Bostock et al, 2001;Rondenay et al, 2005) Waveform from teleseismic events, as shown in Fig. 1 upper-left inset, are first filtered at 0.01-1.0 Hz and processed by the time domain deconvolution (Kikuchi and Kanamori, 1982;Ligorria and Ammon, 1999) with a Gaussian filter parameter of 4.…”

Upper Mantle Imaging with Array Recordings of Converted and Scattered Teleseismic Waves

“…Multichannel preprocessing approaches rely on the assumption that the incident and scattered wavefields have distinct moveouts (arrival times vs offset), and usually comprise the following steps (e.g., Rondenay et al 2005): (1) The traces are aligned relative to the incident wavefield identified on the L/P components. (2) The incident wavefield is extracted either by stacking the aligned L/P components, or by conducting Principal Component Analysis (PCA) on the aligned L/P components and retaining the first or first few components.…”

Multiple seismic reflectors in Earth’s lowermost mantle

“…The data used (∼20% of all available traces) passed selection criteria on the basis of signal-to-noise and multichannel cross-correlation values. Source signatures are estimated through principal component analysis and removed from the data through Wiener deconvolution (27,28). With GRT, we estimate (from scattered energy) elasticity contrasts at nodes of a 3D mesh (10 km vertical, 1°lateral spacing); spatial alignment of such contrasts indicates the presence of an interface.…”