Evaluating the Reliability of Reported Deep Seismicity Beneath Long Beach by Back-Projection of Randomized Traces

Date: 4/24/2019

Time: 06:00 PM

Room: Grand Ballroom

Precise earthquake locations are important for monitoring the location and geometry of active faults, yet it is difficult to monitor earthquakes in urban settings due to high levels of cultural noise. The dense array deployments in 2011 and 2012 at Long Beach provide a unique opportunity in terms of spatial coverage and density for monitoring local seismicity in an urban setting. Applying a combination of downward continuation and back-projection to this dataset, Inbal et al. (2015; 2016) found widespread seismicity with depths greater than 20 km, which is much deeper than the conventionally determined seismogenic depth beneath the western Los Angeles. A later study by Li et al. (2018), however, found that most of the earthquakes have shallow origins compared with Inbal et al.'s results. We test the reliability of the deep imaged sources determined by Inbal et al. by randomizing the traces among different receivers and back-projecting the randomized data to the source locations to compare with Inbal et al.'s results. In addition, we experimented with enhancing the earthquake signals by exploring the newly developed DeepDenoiser algorithm based on deep learning of earthquake waveform and noise characteristics (Zhu et al., 2018). Our results should inform the understanding of the reliability of the deep earthquake locations under the Long Beach array.

Presenting Author: Lei Yang

Authors