11:30 AM - 11:45 AM
[SSS06-04] Stress drops and uncertainty resolution of repeating earthquakes at Parkfield
Keywords:Stress drop, repeating earthquakes, uncertainty resolution, Sampling rates
Earthquake stress drop is an important parameter that is closely associated with ground motion predictions and earthquake source properties. Estimating stress drop from corner frequency is superficially easy but subject to significant uncertainty. In the Parkfield segment, Abercrombie [2014] analyzed three repeating clusters, and found that the sampling rate and source complexity significantly affect the reliability of source parameters. Specifically speaking, low sampling rate from the surface stations is insufficient to resolve corner frequencies for small earthquakes when the corner frequency is beyond the Nyquist frequency. In this study, we aim at resolving the uncertainty in source parameters systematically by comparing results from stacking methods that averaging over many event-station pairs, and results from individual pair analysis methods that look into details for each earthquake.
We begin with the double-difference catalog in Northern California, and an existed catalog with several clusters of repeating earthquakes. Then we search for suitable EGF events for each cluster. We are now studying two event groups, one from the proposed clusters with <20 repeating earthquakes, and the other selected ourselves manually which consists of 406 events with numerous distinct repeating clusters. For our own clusters, we set 0.99 as the cross-correlation threshold following a bandpass filter of 2~40 Hz for repeating pairs that will potentially be utilized for spectral analysis. Both event groups have wide magnitude range from 0 to 3, and we only include channels in different networks with high sampling rates (for borehole network sampling rate is 250 Hz and for surface network only 100 Hz channels). Low sampling rate channels are excluded so that it becomes feasible to apply a wider extension of bandpass filter. Then we will choose a cluster with a rational number of repeating events, apply both improved stacking analysis methods and individual pair analysis to the selected clusters, and systematically assess their consistency and difference. Updated results will be reported at the meeting.
We begin with the double-difference catalog in Northern California, and an existed catalog with several clusters of repeating earthquakes. Then we search for suitable EGF events for each cluster. We are now studying two event groups, one from the proposed clusters with <20 repeating earthquakes, and the other selected ourselves manually which consists of 406 events with numerous distinct repeating clusters. For our own clusters, we set 0.99 as the cross-correlation threshold following a bandpass filter of 2~40 Hz for repeating pairs that will potentially be utilized for spectral analysis. Both event groups have wide magnitude range from 0 to 3, and we only include channels in different networks with high sampling rates (for borehole network sampling rate is 250 Hz and for surface network only 100 Hz channels). Low sampling rate channels are excluded so that it becomes feasible to apply a wider extension of bandpass filter. Then we will choose a cluster with a rational number of repeating events, apply both improved stacking analysis methods and individual pair analysis to the selected clusters, and systematically assess their consistency and difference. Updated results will be reported at the meeting.