[S-SS02] Seismological advances in the ocean

[SSS02-12] Precise local earthquake locations relative to a reflection interface accompanying a seamount subduction at Off-Ibaraki region

*Shinji Yoneshima¹, Kimihiro Mochizuki¹ (1.Earthquake Prediction Research Center, Earthquake Research Institute, University of Tokyo)

Keywords:seamount subduction, off-Ibaraki, OBS, event location

A subducting seamount is supposed to cause strong stress heterogeneity to its circumstances, likely generating small earthquakes. To understand the dynamic behavior of the subducting seamount, the determination of precise event locations might be a key. Especially the location of hypocenters relative to a plate boundary is crucial. In this study, at off-Ibaraki region where a seamount subducts, we focus on a precise hypocenter determination and validation enough to classify whether these events are above or below the plate boundary, or along the plate boundary. For this goal, we have had mainly three obstacles so far; (i) seismicity around the seamount at off-Ibaraki region was quite low (ii) conventional event location might not be enough but would need a well-qualified time pick and reliable 3D velocity model (iii) a reasonable validation may not be available to assess a relative event location from a reflection interface. Authors have resolved the aforementioned obstacles (i) and (ii) by an OBS experiment conducted from 2010 to 2011 at Off-Ibaraki region since the seismicity became quite active after the 2011 Tohoku-oki earthquake. Using this dataset, we performed a double difference event location analysis for an unprecedented number of over 20,000 events with 3D velocity model embedding a seamount structure with a qualified P and S time pick. The resultant dense event distribution implies that a large number of events are located in the oceanic plate below the plate boundary. Nevertheless, while the relative locations of events are rather accurate using the double difference approach, the event location accuracy relative to a plate boundary has not been qualified yet. Because of this uncertainty, the obstacle (iii) is still an open question, and non-trivial; due to a complex event distribution around the seamount, a validation of the interplate/intraplate event classification is not straight-forward since any conventional approaches of a focal mechanism analysis would not be of help to gain an insight about the event location against the plate boundary. In this study, we take a novel approach to validate the event location whether events lie along the plate interface or not. As an attempt, we exploited a vertically aligned event array which was extracted from the relocated events ranging from about 8 km to 25 km presumably intersecting a plate boundary (figure 1). We see that this vertically aligned event geometry is analogous with that of a source array in a reverse VSP at which the source is located vertically in the underground and the station is on the seafloor. With this configuration, the relative time difference between the direct arrival and the reflected arrival constrains the distance between the event location and the reflection interface. Thus this reflected arrival is thought to be a depth phase, providing us evidence against the event location relative to a reflection interface. We are developing a characteristic function to normalize waveform and to enhance the later phases from a trace gather of natural earthquakes. Figure 2 exhibits the preliminary result of the identified reflection phases. Several reflections are visually identified in figure 2 left, suggesting an existence of several reflection interfaces around the plate boundary. In this presentation, we will show a detail of these reflected phases and will discuss the event classification result whether events are dominant along the plate interface or not in this particular off-Ibaraki region.