Teleseismic Waveform Complexities Caused by Near Trench Structures and Their Impacts on Earthquake Source Study: Application to the 2015 Illapel Aftershocks (Central Chile)

Abstract

The complex near trench velocity structures, characterized by strongly varying bathymetry along with seawater, can produce substantial waveform complexities for near trench earthquakes, which makes it difficult to study earthquake source parameters through waveform modeling/inversion. Here we explore these wavefield complexities via modeling teleseismic records of a Mw6.6 near coast event and a Mw6.8 near trench event in the 2015 Illapel earthquake sequence. For the near coast event, the waveforms of direct P waves at teleseismic/diffracted distances are simple, and we obtain consistent source parameters between 1D regional and teleseismic waveform inversions. In contrast, the near trench event produces stronger and longer P coda waves (>100 s), resulting in many dramatic discrepancies between the regional and teleseismic inversions, in particular for the centroid depth. We adopt a spectral element method-direct solution method hybrid approach to simulate synthetics with the complex source-side 3-D structures (bathymetry and water layer in this case) and investigate their roles in the genesis of strong P coda waves. Compared with the 1-and 2-D synthetics, the 3-D synthetics significantly improve the waveform fits up to 0.1 Hz when the source is placed at the preferred horizontal (~30 km from the trench axis) and vertical (~5 km beneath the ocean bottom) location. The refined location of the earthquake indicates that the plate interface is probably locked at very shallow depths and capable of nucleating strong earthquakes. We highlight the need for considering near trench 3-D structures in seismic waveform analysis of near trench earthquakes, many of which are tsunamigenic.Key Points:• The persisting (>100 s) and strong P coda waves of a Mw6.8 near trench event are well reproduced by the 3D modelling that incorporates bathymetry and water layer • 3D waveform modelling is used to relocate the centroid of the near trench earthquake to be 30 km away from the trench and 5 km beneath the ocean floor • Strong and long lasted coda waves of the 2015 Mw8.3 Illapel mainshock could be largely explained by the near trench structure effectsSupporting Information:• Supporting Information S1

Evaluating the Updip Extent of Large Megathrust Ruptures Using Pcoda Levels

“…Nonetheless, ambiguity exists regarding which source rupture attributes and source region structural features are important for exciting strong P coda amplitudes (Fan & Shearer, ; Ihmlé & Madariaga, ; Ward, ; Wu et al, ; Yue et al, ). Recent three‐dimensional studies by Qian et al () and Wu et al () conclude that the bathymetry and sedimentary layers have strong effects on enhancing P wave coda oscillations in addition to the location of earthquake slip.…”

“…For example, a proposed second-stage tsunami earthquake at shallow depth during the source rupture process for the 2015 Illapel (M W 8.4) earthquake was inferred based on observation of prolonged teleseismic P wave ground motions (Lee et al, 2016). However, Lay et al (2016), An et al (2017), and Qian et al (2019) demonstrate that the P coda could be explained without any prolonged rupture by improved modeling of the water reverberations generated by shallow slip. Nonetheless, ambiguity exists regarding which source rupture attributes and source region structural features are important for exciting strong P coda amplitudes (Fan & Shearer, 2018;Ihmlé & Madariaga, 1996;Ward, 1979;Wu et al, 2018;Yue et al, 2017).…”

“…Nonetheless, ambiguity exists regarding which source rupture attributes and source region structural features are important for exciting strong P coda amplitudes (Fan & Shearer, 2018;Ihmlé & Madariaga, 1996;Ward, 1979;Wu et al, 2018;Yue et al, 2017). Recent three-dimensional studies by Qian et al (2019) and Wu et al (2018) conclude that the bathymetry and sedimentary layers have strong effects on enhancing P wave coda oscillations in addition to the location of earthquake slip. Figure 1 provides an intuitive framework for anticipating that the updip extent of slip on the megathrust during a large earthquake will affect the strength and azimuthal pattern of P coda generation.…”

Oceanic high-frequency global seismic wave propagation with realistic bathymetry

“…These include how P waves which are refracted to near-normal angles at the seafloor are converted to horizontally travelling t-waves in the SOFAR channel (de Groot-Hedlin & Orcutt 2001), how primary microseismic noise with a transverse component is generated (Nishida & Takagi 2016), and how best to isolate or account for water column reverberations, especially over areas rough bathymetry (Blackman et al 1995). The latter of these can significantly complicate understanding of earthquake dynamics (Yue et al (2017), Qian et al (2019)), especially of the largest earthquakes which occur at subduction zones with rough bathymetry (Lay & Rhode (2019), Wu et al (2020)).…”

Insights Into the Kinematic Rupture of the 2015 Mw 8.3 Illapel, Chile, Earthquake From Joint Analysis of Geodetic, Seismological, Tsunami, and Superconductive Gravimeter Observations

“…The model of Heidarzadeh et al (2016) revealed a broadened region of coseismic slip, and this could be caused by their preset of rupture speed, because the rupture distribution expands proportionally to the rupture velocity. However, it is a well-recognized fact that teleseismic data themselves are insufficient to resolve rupture processes in detail (Delouis et al, 2010), and also, the teleseismic waveforms for sea-land subduction earthquake are vulnerable to water reverberation, bathymetry of the seafloor, and 3-D heterogeneity near trench, resulting in unexplainable phenomena if these effects are ignored during the rupture process inversion (Yue, Castellanos, et al, 2017;Qian et al, 2017). Barnhart et al (2016), Grandin et al (2016), and Klein et al (2017) jointly used GPS static displacements and InSAR data to explore the coseismic slip distribution.…”