Earthquake Detection with Sparse Networks

Abstract:

The detection of earthquakes and low magnitude seismic tremors are of significant importance in seismic hazard response and mitigation. In regions with sparse seismic networks, the detection and location problem is particularly challenging because of a paucity of data. In this study, we implement an approach to improving earthquake detection and location using a combination of back-projection and polarization techniques. Our approach is important for identifying and studying seismicity in the regions where the events catalog is less complete and accurate. Wavefield back-projection incorporates timing refinements using both a global body-wave (LLNL-G3D) and a surface wave model (GDM52). Rapid detection during grid search on the globe will incorporate wave polarization and arrival angle measurements. We will test our detector on a decade-long record investigating accuracy using a handful of seismic stations. Results demonstrate that incorporating surface wave timing adjustments is crucial for detecting high magnitude events. Future study will explore machine learning algorithms.