Poster #049, Seismology

Monitoring seismic velocity changes in 18 years at Parkfield using ambient seismic noise

Poster Image: 

Poster Presentation

2020 SCEC Annual Meeting, Poster #049, SCEC Contribution #10356
Both transient tectonic events such as fast and slow earthquakes, non-volcanic tremors, and periodic deformation caused by tidal and thermoelastic strains play a role in the evolution of seismic velocity change (dv/v) around faults. Our study aims to decipher these signals the dv/v time history in order to better resolve the long-term tectonic interseismic process by monitoring the coda of ambient seismic noise correlation functions. We conduct auto- and cross- correlation of the continuous noise from the High Resolution Seismic Network (HRSN) at Parkfield, CA from 2002-2020. We remove transient signals from continuous seismic waveforms using STA/LTA and kurtosis to minimize the bias in dv/v... caused by the earthquakes and tremors. We then use a wavelet transform to decompose the correlation functions into narrow frequency bands, and explore various stacking methods (linear, selective, robust, phase-weighted) to enhance the convergence of stacked correlation functions and continuous dv/v measurement. Our noise processing is performed using Julia-based tools, SeisIO and SeisNoise, to handle the large dataset. To remove the seasonal variation caused by atmospheric temperature change from dv/v history, we use the thermoelastic strain model in an elastic half space [Ben-Zion and Leary, 1986]. The long-term post earthquake loading continues until today, where the current dv/v is equivalent to, or larger than, the level before the Parkfield earthquake. We also investigate the coseismic velocity change associated with recent earthquakes such as the 2014 South Napa earthquake and the 2019 Ridgecrest earthquake sequence.