Characterizing the Seafloor Sediment Layer Using Teleseismic Body Waves Recorded by Ocean Bottom Seismometers

Abstract

This study presents an approach to better characterize the P‐wave and S‐wave velocity structure of the seafloor sediment layer using ocean bottom seismometers. The presence of low‐velocity seafloor sediment layers influences the observed seismic record at the seafloor over a broad frequency range, such that detailed knowledge of this sediment structure is essential to predict its effect on teleseismic records. We use the radial component of teleseismic P waves and autocorrelation functions of the radial, vertical, and pressure components of teleseismic P and S waves to obtain sediment layer models using the Markov chain Monte Carlo approach with parallel tempering. Synthetic tests show that the body waves constrain the P‐ and S‐wave impedances and travel times and the P‐ to S‐wave velocity ratio of the sediment layers. The proposed method resolves thin layers at a high resolution, including the uppermost thin (∼50 m to a few hundred meters) low S‐wave velocity layer. Real data applications at sites across the Pacific Ocean that are coincident with previous in situ studies demonstrate the effectiveness of this method in characterizing the seafloor sediment unit. The sediment models characterized by this new approach will allow us to more accurately predict and correct the effects of sediment layers in generating P‐ and S‐wave reverberations.

A Mid‐Lithospheric Discontinuity Detected Beneath 155 Ma Western Pacific Seafloor Using Sp Receiver Functions

“…We first tested simple models that include a ∼6 km water column, a 50-m-thick sedimentary layer with Vs of 0.1 km/s and Vp of 1.8 km/s (Kim et al, 2022(Kim et al, , 2023, a crust and uppermost mantle model from ambient noise cross-correlation (Takeo et al, 2014) and two step-function velocity boundaries in the mantle. We determined the depths and velocity contrasts of the two mantle boundaries that minimize misfit with the observed stack through forward modeling (Figure 3a), finding values of 44 km ( 3%) and 146 km (3%).…”