Adjoint Tomography Based on the Differential Measurements of Ambient Noise Correlations

Session: Full-Waveform Inversion: Recent Advances and Applications [Poster]
Type: Poster
Date: 4/29/2020
Time: 08:00 AM
Room: Ballroom
Description: 

Ambient noise adjoint tomography benefits from better data availability and denser spatial coverage compared with earthquake adjoint tomography, but the tomographic results are subject to strong bias due to the non-uniform noise source distribution and the temporal variation of noise source strength. Without the exact knowledge on the spatio-temporal evolution of noise sources, it is difficult to constrain and account for the bias in the resulting image. To overcome these limitations, we recently developed a new approach towards ambient noise adjoint tomography using differential measurements with a dense seismic array. The differential measurements can be either differential time or amplitude ratio, corresponding to velocity or attenuation tomography, respectively. Tomographic inversions on synthetic data have shown that this new method is insensitive to the spatial distribution and temporal variation of noise sources and converges much faster than previous approaches. In this work, we apply it to a dataset in Netherlands and expand it to both velocity and attenuation inversion. We show the improvements on tomographic images compared with the traditional approach using travel time measurements from ambient noise data. We note that this method should be broadly applicable to applications in adjoint tomography across different spatial scales.

Presenting Author: Xin Liu

Authors