Earthquake relocation at intermediate depths using automatically detected teleseismic depth phases

Intermediate-depth earthquakes, accommodating intra-slab deformation, typically occur within subduction zone settings at depths between 40-350 km. High magnitude events can pose a significant hazard to populations, and increase the risk of damaged infrastructure, injury and fatality. Despite improvements in recorded seismic data density and quality, the distribution and controls of these events remain poorly understood. Here, we demonstrate an automatic method for the detection of depth phases from these intermediate-depth earthquakes using seismic array data, with the aim of determining their hypocentral depths and locations to a greater level of accuracy. These will allow new comparisons and insights into the governing controls on the distribution of earthquakes in subducted slabs.

Depth phases (near-source surface reflections, e.g. pP and sP) are crucial for the accurate determination of earthquake source depth using global seismic data, however, they suffer from poor signal-to-noise ratios in the P-wave coda. This reduces the ability to systematically measure differential travel times to the corresponding direct arrival, particularly for the frequent lower-magnitude seismicity which highlights considerable seismogenic regions of the subducted slabs. To address this limitation, we have developed an automated approach to group globally distributed stations at teleseismic distances into sub-arrays, before optimising and applying phase-weighted beamforming techniques to each sub-array. Resultant vespagrams allow automated picking algorithms to determine differential arrival times between the depth phases (pP, sP) and their corresponding direct P arrival. These are subsequently used to invert for a new hypocentre location. We demonstrate this method by relocating intermediate-depth events associated with the Peruvian flat slab region of the subducting Nazca plate.