We used HypoDD (17) to better constrain the locations of the cataloged and newly detected seismicity. We included both shallow and deep earthquakes, but those with depths shallower than 6 km were filtered to require ≥M1 so as not to dominate the inversion but still provide a common reference frame between shallow and deep seismicity. Differential times from analyst picks and subsample waveform cross-correlation were used for both P and S arrivals and assumed common starting location and pick times between new detections and their best matching catalog event. For cross-correlation differential times, we assigned weight as (cccmax)2 * min(1, 10*(cccmax − cccnextmax)), where cccmax is the absolute maximum normalized cross-correlation coefficient (minimum, 0.4), and cccnextmax is the next highest absolute peak in the cross-correlation function. This permits well-determined differential times to carry greater weight in the inversion while allowing yet punishing differential times that have a high chance of cycle skipping. In addition, allowing for negative cross-correlation coefficients enables more high-quality pairs between earthquakes with different focal mechanisms. We began the first iterations of the double-difference inversion with maximum weight on catalog differential times to define the macroscale structure and then increased the weight on waveform cross-correlation differential times in later iterations to resolve finer-scale features. We then required final earthquake locations to be constrained by at least 20 differential time pairs, although most of the best constrained locations have several hundred observations. Of the 9531 input events, 6491 were successfully relocated with HypoDD, and 4419 met the constraints for plotting. Finally, we further increased the explicit depth separation between shallow and deep events to 8 km for cleaner plotting in map view once it became clear where the shallowest LP events were finally located. The velocity model used for relocation is included in table S1 (30), with Vp/Vs = 1.6, as suggested by previous tomographic studies (14). Relocations presented here were insensitive to the choice of the velocity model. Figure S1 shows the small difference in relocations when using a higher Vp/Vs ratio of 1.73. The absolute locations of most of the seismicity at depth were shifted approximately 400 to 500 m, but all of the features we interpreted in the text remained intact.

Note: The content above has been extracted from a research article, so it may not display correctly.

Please log in to submit your questions online.
Your question will be posted on the Bio-101 website. We will send your questions to the authors of this protocol and Bio-protocol community members who are experienced with this method. you will be informed using the email address associated with your Bio-protocol account.

We use cookies on this site to enhance your user experience. By using our website, you are agreeing to allow the storage of cookies on your computer.