← Back to Sessions

Full Waveform Tomography for the Upper South Island Region, New Zealand Using Ambient Noise Data

Session: Advances in Seismic Interferometry: Theory, Computation and Applications [Poster]

Type: Poster

Date: 4/22/2021

Presentation Time: 11:30 AM Pacific

Description: 

In this research, we apply Full Waveform Tomography (FWT) based on the Adjoint-Wavefield (AW) method using ambient noise data to iteratively invert an existing 3-D velocity model for the upper South Island region, New Zealand. Through the AW method, gradients of model parameters (compressional and shear wave velocity) were computed by implementing the cross-adjoint of forward and backward wavefields simulated for each station acting like a source. We used L-BFGS/ line search algorithms to obtain a model update. Continuous time-series data recorded for one year by 10 seismic broadband stations in South Island was downloaded, and cross-correlations were computed and stacked to extract the empirical Green’s functions (EGFs). Multi-taper misfits between the EGFs and the simulated Green’s functions (SGFs) were measured at periods from 10 - 40s and minimized along the inversion to iteratively update the velocity model. To improve the existing velocity structure, we can start the inversion from a New Zealand wide velocity model (NZVM 2020) or from an inverted velocity model using FWT of the earthquake events. The simulated seismograms computed using our inverted velocity models and earthquake events show the relative waveform misfit to the earthquake recorded data has been reduced consistently along the iterations. This suggests that using ambient noise data in FWT can supplement for the inversion with only earthquake data.

Presenting Author: Trung Dung Nguyen

Student Presenter: No

Authors