Exciting news! We're transitioning to the Statewide California Earthquake Center. Our new website is under construction, but we'll continue using this website for SCEC business in the meantime. We're also archiving the Southern Center site to preserve its rich history. A new and improved platform is coming soon!

Poster #236, Seismology

2019 Ridgecrest earthquake sequence: RAPID seismic deployment and a new aftershock catalog based on machine learning

Abhijit Ghosh, Yijian Zhou, Shankho Niyogi, Manuel M. Mendoza, Kuntal Chaudhuri, Baoning Wu, Han Yue, Lihua Fang, Shiyong Zhou, & Andrew Birkey
The poster PDF is private. For more information, please contact the author(s).

Poster Presentation

2021 SCEC Annual Meeting, Poster #236, SCEC Contribution #11585
We installed a temporary seismic network immediately after the Mw 7.1 Ridgecrest earthquake in 2019. The first station went in within a day of the mainshock. There are 20 seismic stations that were operated in continuous mode for up to 1 year. They surround the aftershock area in all directions, and complement other temporary deployments. In this study, we have used permanent seismic network and USGS temporary stations near the aftershock area to make an earthquake catalog using a new algorithm based on machine learning (Zhou et al., 2019). We have analyzed three weeks of seismic data starting from Mw 6.4 foreshock on July 4th. We first use a kurtosis-based phase picker to pick P- and S-phas...es. Then we associate them and determine location of events, first absolute and then relative relocation. Our catalog contains 66% more earthquakes compared to the catalog produced by the Southern California Seismic Network. We use our catalog to train the machine learning algorithm. We use a convolutional neural network to detect events and recurrent neural network to pick phases using this algorithm. Association and location are similar to the previous stage. This new earthquake catalog based on machine-learning algorithm contains 97,855 events detected and located during the three weeks of data analyzed so far. The level of detection is similar to Ross et al., 2019, and 73% higher than Shelly, 2020, when compared to the same time period. Our catalog generally shows similar features as compared to other catalogs, but appears to be more clustered. There are, however, some remarkable differences in details. For example, we observe interesting differences in pattern of earthquake distribution near the peak slip patch, and southeastern end of the main fault. We plan to update our catalog including data from RAPID deployment, and expect to observe more details of the structure and aftershock dynamics.
SHOW MORE

Full Presentation

Citation: Ghosh, A., Zhou, Y., Niyogi, S., Mendoza, M. M., Chaudhuri, K., Wu, B., Yue, H., Fang, L., Zhou, S., & Birkey, A. (2021, 08). 2019 Ridgecrest earthquake sequence: RAPID seismic deployment and a new aftershock catalog based on machine learning. Poster Presentation at 2021 SCEC Annual Meeting.
Keywords: Ridgecrest, earthquake, aftershock, machine learning
Relevant SCEC Themes:
Stress and Deformation Over Time
Post-Earthquake Rapid Response