Stress Drop in Ridgecrest Sequence Events from the Generalized Inversion Technique

Emma G. Devin, Grace A. Parker, Annemarie S. Baltay, Tara Nye, & Valerie J. Sahakian

Submitted August 12, 2021, SCEC Contribution #11311, 2021 SCEC Annual Meeting Poster #198

We estimate stress drops for earthquakes in the 2019 Ridgecrest sequence using the generalized inversion technique of Andrews (1986). Stress drop is an earthquake parameter that characterizes high frequency seismic radiation and the amount of energy released during an earthquake. Stress drop is notoriously difficult to measure, and estimates made using different methods and datasets tend to be variable, making it difficult to understand physical trends. The Community Stress Drop Validation Study aims to understand differences in stress drops by examining estimates made by many researchers using a consistent dataset from the Ridgecrest sequence. That dataset contains ground motion time series from over 12,000 events M≥1 that occurred from 4 – 18 July 2019, recorded at distances up to 1°. We use the Obspy Python library to linearly detrend the sequence data, correct for instrument response, and apply a tapered bandpass filter with corners at 0Hz, 0.001Hz, 35Hz, and the Nyquist frequency. Following Klimasewski et al., (2019) we compute record spectra, invert for event and station spectra, and fit theoretical Brune (1970) model to the event spectra. Fourier amplitude spectra with 75 frequency bins are obtained using multitaper spectra estimation (Prieto et al., 2009). The spectra are corrected for geometrical spreading using a simple 1/R relation for a homogenous elastic half-space, as the spectral amplitudes with hypocentral distance do not support more complexity. To determine how sensitive our results are to the chosen path correction, we also use the geometrical spreading term from Bayless and Abrahamson (2019), which depends on magnitude and frequency. No correction is applied for anelastic attenuation because all the records are from stations less than 110 km away from the source where anelastic attenuation effects are negligible. Record power spectra are then inverted for station and event spectra using a singular value decomposition and constrained using an individual earthquake with the most Brune-like spectra (Klimasewski et al., 2019). We also test the site response at a reference station as a constraint, as in Andrews (1986). Lastly, a Brune spectrum is fit to each source spectrum to find the corner frequency and thus stress drop for each event, and we examine spatial, temporal, and parametric trends.

Citation
Devin, E. G., Parker, G. A., Baltay, A. S., Nye, T., & Sahakian, V. J. (2021, 08). Stress Drop in Ridgecrest Sequence Events from the Generalized Inversion Technique. Poster Presentation at 2021 SCEC Annual Meeting.


Related Projects & Working Groups
Seismology