Stress drop of earthquakes from the Multi-Window Spectral Ratio method in a regional network
Abstract
Insight into the mechanics of faulting comes from seismologically determined parameters such as earthquake stress drop. To produce higher precision estimates of stress drop on regional networks like the Parkfield HRSN borehole network, we use the multi-window spectral ratio (MWSR) method to determine the corner frequency and relative moment of earthquake pairs. This method assumes that the frequency distribution of spectral power is the same for direct arrivals and later arriving scattered energy. By taking the power spectra of individual time windows, we collect multiple, quasi-independent estimates of the power spectra for a given earthquake. For pairs of earthquakes located within a few hundred meters of one another, the ratio of their power spectra at a given station provides better estimates of their relative moments and corner frequencies by eliminating some of the path-dependent effects. We calculate the spectral ratio for each time window at each station in a regional borehole network and stack the ratios to cancel out noise and determine a robust estimate of the true power spectral ratio. Systematic differences between the spectral ratio at stations with different azimuths can be interpreted as evidence for directivity in the earthquake pulses or frequency- dependent attenuation along the source-receiver path. Because the method is based on a collection of quasi- independent spectral ratio observations, we can quantify the uncertainty in the estimates of corner frequency and moment for each earthquake pair. We use a bootstrap analysis with noise selected from the distribution of scatter from individual time windows about the mean spectral ratio to quantify the range of moment and corner frequencies that fit each station. By assuming that the earthquakes are simple circular cracks, we convert these parameters into stress drop with estimates of uncertainty. We evaluate the reliability of the technique in a regional network for different earthquake spacings (from tens to hundreds of meters). We hope to constrain the variability of source parameters for repeating microearthquakes and explore the distribution of stress drop as a function of depth. Preliminary results show that the method works well with the HRSN borehole network at Parkfield.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2006
- Bibcode:
- 2006AGUFM.S23C0186D
- Keywords:
-
- 7215 Earthquake source observations (1240);
- 8020 Mechanics;
- theory;
- and modeling