Using Noise Correlation to Improve the Seismic Velocity Model of the Seattle Basin and 3D Simulations of Large Earthquakes

Date: 4/24/2019

Time: 09:00 AM

Room: Pine

Cross correlation waveforms of seismic noise in the Seattle basin were analyzed to determine the group velocities of surface waves and improve the velocity model used in 3D simulations of earthquakes. Twenty broadband seismometers were deployed for about three weeks by personnel from the U.S. Geological Survey and the Pacific Northwest Seismic Network of the University of Washington. The instruments were arranged in three dense arrays separated by about 5 km, with minimum intra-array station spacing of about 0.5 km. We found good correlations of seismic noise between the arrays at periods of 2 to 10 s using only 9 days of noise recordings. Usable noise correlations up to 2 Hz were determined for sites separated by 0.5-1.0 km. The cross correlation waveforms were typically larger for the transverse component of motion, indicating better correlation than the radial and vertical components. We used the cross correlation waveforms to determine Love wave group velocities at 0.5-10 s for paths within the Seattle basin and paths crossing the southern edge of the basin. We compared these observed group velocities in the basin to those calculated using a flat-layered velocity model approximated from the 3D velocity model of the Seattle basin derived by Stephenson et al., 2017. The noise correlation results indicate that the shear-wave velocity (V_s) increases rapidly with depth in the top 0.3 km of the glacial sediments and then is relatively constant down to about 1 km depth, the base of the sediments. V_s values in the basin at depths of 1-5 km were found to be about 10-15% lower than in the model. We evaluate the effects of these differences in V_s on synthetic seismograms derived from 3D simulations of the M6.8 Nisqually earthquake and M9 Cascadia earthquakes.

Presenting Author: Arthur Frankel

Authors