← Back to Sessions

Shear Wave Velocity Structure Beneath a Dense Seismic Array in the Presence of Local Noise Sources Using Matched Field Processing

Description: 

The presence of local sources in the ambient seismic wavefield causes bias in the estimation of surface wave dispersion via seismic interferometry (SI). Here we use Matched Field Processing (MFP) to capture local noise sources and then extract more accurate surface wave phase velocity dispersion curves from data recorded by a dense array of 2639 Fairfield Nodal Zland 1C nodes deployed near Sweetwater, TX. These short period instruments recorded continuously for 11 days, including signals from Vibroseis trucks at various locations.

MFP is an array processing method that conducts a grid search of correlations between a recorded wavefield and a synthetic wavefield generated by forward modeling. For each potential source location, an estimate of the phase match is computed between both wavefields over a certain frequency band, taking coherency of the wavefields across stations into account. The relative spatial phase differences between all sensors of the array patch can be expressed as a Cross Spectral Density Matrix (CSDM). A 3D energy spectrum matrix is obtained at each frequency and the peak of the energy spectrum matrix corresponds to the location of the dominant noise source. The velocity vector at the location of the noise source is extracted along the velocity axis as the velocity distribution for the current frequency. Velocity vectors that correspond to each frequency are then normalized and stitched together along the frequency axis to obtain the frequency-velocity energy spectrum, which is the dispersion energy map of the current data segment. Finally, we stack all the dispersion energy images from a single data segment to obtain a more stable dispersion energy map. This dispersion curve is then inverted using a Markov Chain Monte Carlo (MCMC) approach to retrieve the 1D shear wave velocity model beneath the patch of the array under consideration. Preliminary results show that MFP produces a more accurate dispersion curve in the 4-20 Hz frequency range than SI-based surface wave methods in the presence of local noise sources.

Session: How Well Can We Predict Broadband Site-Specific Ground Motion and Its Spatial Variability So Far? [Poster Session]

Type: Poster

Date: 5/1/2024

Presentation Time: 08:00 AM (local time)

Presenting Author: Yashwant

Student Presenter: Yes

Invited Presentation: 

Authors