Earthquake detectability and depth resolution with dense arrays

Asaf Inbal, Jean-Paul Ampuero, & Robert W. Clayton

Published August 15, 2021, SCEC Contribution #11475, 2021 SCEC Annual Meeting Poster #222

Dense array seismology is an emerging field which is well-suited for signal detection in noisy environments, and hence may be used effectively for characterizing earthquake activity in urban areas. This application frequently employs detection and location schemes that are based on statistical analysis of back-projection (BP) images. However, the imaging procedure, and hence the detection and location reliability, are strongly dependent on the source-array geometry and SNR levels. Here, we use synthetic tests and raw dense array data to quantify the performance of BP-based dense array processing methodologies, and assess their depth resolution.

We test the capabilities of the 5200-sensors Long Beach (LB) and 2600-sensors Extended Long Beach (ELB) arrays, deployed for several-month periods along the Newport-Inglewood Fault (NIF) within the LA basin. This fault segment hosts deep (>25 km) earthquakes, which are recorded in extremely noisy environments. As a result, the NIF microearthquake detection rates are the lowest in southern California, despite the fact that the NIF is one of the best instrumented faults in that region. Inbal et al. (2015, 2016) used the LB dataset to compile a catalog for the portion of the NIF in LB, and found abundant seismicity occurring in the lower crust and upper mantle. The results were recently contested by Yang et al (2021), who introduced a new method for discriminating coherent seismic sources from noise sources, and found little evidence for deep NIF seismicity. We benchmark this new scheme by using synthetic data, and find that its sensitivity to the array-aperture to source-depth ratio results in strong bias towards shallow signals, and that it cannot reliably discriminate between a deep (>8 km) coherent source and random noise. We quantify the source depth resolution using point spread functions, and show that the larger aperture of the LB array results in a significant increase in the depth resolution power over the ELB array. For signals with SNR>1, the LB array maintains its resolution even if only 1% of its sensors are used for BP. Analysis of the regional and array-derived NIF catalogs, suggests mantle seismicity beneath LB is a long-lived feature of this fault. We discuss this result in light of newly acquired Moho depths beneath the NIF.

Citation
Inbal, A., Ampuero, J., & Clayton, R. W. (2021, 08). Earthquake detectability and depth resolution with dense arrays. Poster Presentation at 2021 SCEC Annual Meeting.

Related Projects & Working Groups
Seismology