Accurate Depth Determination for Moderate-Magnitude Earthquakes Using Global Teleseismic Data

The determination of accurate source depths for globally observed earthquakes has long been one of the most problematic issues in earthquake source seismology. The optimal method for constraining the depths of teleseismically observed earthquakes is through the identification of near-source surface reflectionsdepth phases. However, observing such phases is complicated by the relatively small amplitude of these arrivals compared to the background noise, particularly for earthquakes with M-w<5.5. In this study, I present a methodology for leveraging the recent expansion in global seismic network coverage to enhance the identification of coherent depth phases through an automated stacking routine using a globally distributed array. While depth solutions for each individual depth phase are often nonunique, the identification of potential depths across multiple different phases, where each shows the correct distance-dependent delay, provides a robust way to semiautomatically determine the source depth of an earthquake. I present a range of examples for the processing routine developed, along with an example for its regional application, to the North Chilean subduction zone. The technique presented offers an opportunity to improve depth estimates for earthquakes down to M(w)4.9 and, requiring significantly less analyst input than other techniques offering a similar resolution, has the potential to be applied to large earthquake data sets. Plain Language Summary The determination of accurate depths for globally observed earthquakes is critical for our understanding of the structure and deformation of tectonic plates but has proven to be one of the most intractable problems in earthquake seismology. Here, I present a method aimed at using the recent vast expansion in global seismic data coverage to enhance the detection of a suite of seismic phases that reflect from the Earths surface directly above the earthquake sourcethese are often referred to as depth phases. These phases provide the best known constraint on earthquake depths for remotely observed earthquakes, especially in cases where several of these phases can be observed for a given earthquake. However, they are often difficult to detect, due to their low amplitude, and the complexity of the associated waveform. While limitations remain on the magnitude and source depth ranges over which the technique presented here can be applied, this approach enhances the detection of these low-amplitude phases, allowing them to be identified for smaller earthquakes than has routinely been possible in the past. As this approach uses data observed from around the world, it can potentially be applied to large earthquake data sets in remote areas, where near-field data are unavailable.