Hydrous state of the subducting Philippine Sea plate inferred from receiver function image using onshore and offshore data

Abstract

Exploring the hydrous state of subducting oceanic crust is intriguing because it is considered to affect the strength of megathrust faults that cause various types of earthquakes; however, its state beneath offshore regions remains unclear. In this study, we investigated fluid contents along the subducting Philippine Sea plate around the Kii Peninsula by receiver function (RF) analysis using data from both on‐land stations and ocean bottom seismometers (OBSs). The vertical component of OBS records contains dominant water reverberations, and thus, conventional methods fail to estimate RFs correctly. We therefore developed a method to calculate RFs that removes such reverberations. The RFs calculated by our method showed considerable improvement for later phase identification, compared with those obtained using a conventional method. Resultant RF amplitudes suggest the existence of low‐velocity zones directly beneath the plate interface of both onshore and offshore regions. We interpreted this as evidence of hydrous oceanic crust, which extends from 5 km to 35 km depth to the plate interface. Reduction of RF amplitudes beneath the Kii Peninsula suggests that dehydration of the oceanic crust increases the seismic velocity, and the accompanying densification makes the plate interface permeable. This permeable plate interface may characterize the location of non‐volcanic tremors. This contrasts with long‐term slow slip events because it is believed that they occur along the sealed plate interface. Comparison between the plate geometry and local earthquakes reveals the paucity of earthquakes in the oceanic crust below a certain depth, which provides further insight into the dehydration process in the oceanic crust.

A fluid‐rich layer along the Nankai trough megathrust fault off the Kii Peninsula inferred from receiver function inversion

“…Ensemble of preferable S wave velocity structures obtained by inversion analysis (brown color). White dashed lines denote depths of the plate interface and the oceanic Moho estimated by Akuhara and Mochizuki []. The area not sampled by the preferable models is masked by gray.…”

“…Nowadays, large-scale scientific projects involving passive seismic observations at offshore sites have been launched or are undergoing at subduction zones worldwide [e.g., Scherwath et al, 2011;Kaneda, 2014;Toomey et al, 2014;Harris et al, 2016]. Unfortunately, the application of RF analysis to such offshore data are limited to date [e.g., Akuhara and Mochizuki, 2015;Janiszewski and Abers, 2015;Akuhara et al, 2016;Audet, 2016]. Our analysis method has overcome severe problem specific to OBS data by taking multiple phases within seawater and sediment layers into consideration.…”

“…For overriding and oceanic crust layers, we treated the layer bottom depths ( z o and z c , respectively) and V p / V s ratio ( κ o and κ c , respectively) as unknown parameters (Figure ). The search range of z o and z c are set to ±2 km wide around the depth of the plate interface and the oceanic Moho, respectively, which were estimated previously [ Akuhara and Mochizuki , ]. Search ranges of κ o and κ c are 1.6 to 1.8 and 1.6 to 2.0, respectively.…”

“…In addition, high‐frequency analysis would be required so that the resolution of RF analysis can be comparable to that of active source surveys. Fortunately, such challenges can be addressed, because previous studies have found solutions to calculate RFs stably at high frequencies [ Park and Levin , , ] and with data recorded by ocean bottom seismometers (OBSs) where strong water reverberations dominate vertical component records [ Akuhara and Mochizuki , ; Akuhara et al ., ]. In this study, we perform RF analysis with OBSs deployed off the Kii Peninsula, located in southwest Japan to estimate seismic property of megathrust fault at seismogenic zone depth (Figure ).…”

“…In addition, high-frequency analysis would be required so that the resolution of RF analysis can be comparable to that of active source surveys. Fortunately, such challenges can be addressed, because previous studies have found solutions to calculate RFs stably at high frequencies Levin, 2000, 2016] and with data recorded by ocean bottom seismometers (OBSs) where strong water reverberations dominate vertical component records [Akuhara and Mochizuki, 2015;Akuhara et al, 2016].…”