Machine Learning-Based Tomography of the Ridgecrest Region

Zheng Zhou, Peter Gerstoft, Michael Bianco, & Kim B. Olsen

Submitted August 15, 2020, SCEC Contribution #10701, 2020 SCEC Annual Meeting Poster #064

We perform ambient noise tomography using data recorded from Nodal arrays within a ~50 km by 50 km area including the July 2019 M7.1 and M6.4 Ridgecrest, CA, earthquakes. Our preliminary imaging using a conventional least squares approach reveals a 1-6 km wide low-velocity zone (LVZ) around the fault traces for the M7.1 and M6.4 events, surprisingly well correlated with the width of recorded aftershocks. We find an average reduction in Rayleigh wave velocity within the LVZ (~1.9 km/s) relative to the surrounding area (~2.6 km/s) of about 25-30% within the depth range resolved in our inversion (~0.1-1 km). The relatively large width of the LVZ is consistent with the complex and distributed active faulting observed along the M7.1 and M6.4 rupture zones. Moreover, the presence of a LVZ may be consistent with studies reporting relatively slow rupture velocities for the Ridgecrest events. The SCEC Community Velocity Models (CVMs) include seismic velocities close to our imaged values for the host rock around the fault, but do not resolve the LVZ. This LVZ should be included in the CVMs in future work for more accurate ground motion predictions.

Citation
Zhou, Z., Gerstoft, P., Bianco, M., & Olsen, K. B. (2020, 08). Machine Learning-Based Tomography of the Ridgecrest Region. Poster Presentation at 2020 SCEC Annual Meeting.


Related Projects & Working Groups
Seismology