Localized water reverberation phases and its impact on backprojection images

Abstract

Coherent radiators imaged by backprojections (BP) are commonly interpreted as part of the rupture process. Nevertheless, artifacts introduced by structure related phases are rarely discriminated from the rupture process. In this study, we use a calibration event to discriminate between rupture and structure effects. We reexamine the waveforms and BP images of the 2012 Mw 7.2 Indian Ocean earthquake and a calibration event (Mw 6.2). The P wave codas of both events present similar shape with characteristic period of approximately 10 s, which are backprojected as coherent radiators near the trench. S wave BP does not image energy radiation near the trench. We interpret those coda waves as localized water reverberation phases excited near the trench. We perform a 2‐D waveform modeling using realistic bathymetry model and find that the steep near‐trench bathymetry traps the acoustic water waves forming localized reverberation phases. These waves can be imaged as coherent near‐trench radiators with similar features as that in the observations. We present a set of methodologies to discriminate between the rupture and propagation effects in BP images, which can serve as a criterion of subevent identification.

Coherent Seismic Arrivals in the P Wave Coda of the 2012 Mw 7.2 Sumatra Earthquake: Water Reverberations or an Early Aftershock?

“…However, back‐projection can still suffer from imaging artifacts, depending upon the data coverage and quality, as well as details of the data processing (e.g., Kiser & Ishii, ; Meng et al, ; Xu et al, ). In some cases, depth phases or water reverberations may cause coherent radiation, contaminating the back‐projection images (e.g., Xu et al, ; Yue et al, ). Thus, evaluating the resolution and robustness of back‐projection results is important and can be done using synthetic experiments (e.g., Koper et al, ) or back‐projecting smaller earthquakes near the mainshock rupture to estimate the imaging kernel for a point source (e.g., Wang & Mori, ).…”

“…Recently, Yue et al () examined the anomalous arrivals in the P wave trains of the 2012 M w 7.2 Sumatra earthquake and concluded that they probably originated from mainshock water reverberations rather than from an early aftershock, based on both observational details and a 2‐D synthetic simulation. A key part of their argument involves a comparison between the mainshock and a nearby M 6 earthquake, which they claim contains anomalous late P wave arrivals similar in character to those seen for the mainshock, and yields back‐projection images with energy near the location of the early aftershock(s) imaged by Fan and Shearer () for the mainshock.…”

“…Here we study the M w 7.2 Sumatra earthquake wave trains in more detail, by examining three calibration (EGF, empirical Green's function) events near the mainshock, including the event analyzed by Yue et al (), as well as three additional EGF events near the location of our previously inferred early aftershock (Figure ). Our results indicate that the oscillatory character of the anomalous mainshock arrivals suggests water reverberations, but that the data are more consistent with water reverberations excited by an early aftershock located landward of the trench, than with delayed and displaced reverberations from the mainshock.…”

“…Azimuthal variations in these arrivals, together with back‐projection analysis, led Fan and Shearer (, https://doi.org/10.1002/2016GL067785) to conclude that they originated from early aftershock(s), located ∼150 km northeast of the mainshock and landward of the trench. However, recently, Yue et al (, https://doi.org/10.1002/2017GL073254) argued that the anomalous arrivals are more likely water reverberations from the mainshock, based mostly on empirical Green's function analysis of a M 6 earthquake near the mainshock and a water phase synthetic test. Here we present detailed back‐projection and waveform analyses of three M 6 earthquakes within 100 km of the M w 7.2 earthquake, including the empirical Green's function event analyzed in Yue et al (, https://doi.org/10.1002/2017GL073254).…”

“…However, recently, Yue et al (, https://doi.org/10.1002/2017GL073254) argued that the anomalous arrivals are more likely water reverberations from the mainshock, based mostly on empirical Green's function analysis of a M 6 earthquake near the mainshock and a water phase synthetic test. Here we present detailed back‐projection and waveform analyses of three M 6 earthquakes within 100 km of the M w 7.2 earthquake, including the empirical Green's function event analyzed in Yue et al (, https://doi.org/10.1002/2017GL073254). In addition, we examine the waveforms of three M 5.5 reverse‐faulting earthquakes close to the inferred early aftershock location in Fan and Shearer (, https://doi.org/10.1002/2016GL067785).…”