← Back to Sessions

Finite-domain Full-waveform Ambient Noise Inversion for Structure and Source Parameters

Description: 

Ambient seismic vibrations are a rich source of information about subsurface structures that can be probed by inversion approaches. Where noise fields are homogeneous, empirical Green’s function retrieval methods can be used to extract this subsurface information. In heterogeneous noise fields, however, significant amplitude and phase errors can arise, presenting a serious obstacle for full-waveform applications.

Full-waveform ambient noise inversion bypasses emprical Green’s function retrieval by treating noise correlations as observables in inversion. Unfortunately, existing formulations of full-waveform ambient noise inversion necessarily involve numerical simulation on the scale of noise source distribution rather than domain scale, an orders-of-magnitude discrepancy that makes small-scale applications impractical. This is a significant obstacle for reservoir monitoring applications, where ambient noise monitoring would otherwise be ideal.

We present an approach for achieving full-waveform ambient noise inversion on spatial scales that are small relative to the distribution of noise sources. This is achieved by reframing the problem in terms of boundary wavefield operations in place of out-of-domain wave simulations. When the out-of-domain medium is relatively simple, these boundary operations take the form of a relatively small number of simple correlations. We present numerical examples in which both material properties within a domain and direction of noise contributions outside a domain are simultaneously inverted for. We demonstrate that we are able to accurately recover both sets of properties without the need for out-of-domain wavefield simulations, even when all noise sources considered are external to the domain of interest.

Session: Numerical Modeling in Seismology: Theory, Algorithms and Applications - II

Type: Oral

Date: 4/17/2025

Presentation Time: 10:45 AM (local time)

Presenting Author: Scott

Student Presenter: No

Invited Presentation: 

Poster Number:

Authors