Development of a SeisComP3 module for detecting micro earthquakes and its application to ocean bottom network data

Abstract

Seismic data from local size networks and a semi-global size network are being received at JAMSTEC. In order to systematically handle the data of the different network sizes, we attempted to apply SeisComP3, which is an earthquake monitoring system originally designed for detecting earthquakes or tsunamis in global or semi-global scales. On the other hand, it has been reported that many micro earthquakes are recorded in one of the networks, DONET. DONET is a local size network of ocean bottom seismometers deployed offshore of Kii peninsula. It is important to efficiently detect and list these earthquakes for the full utilization of the network data toward understanding the dynamics of the shallow part of the Nankai Trough. However, since the earthquake detection algorithm used in SeisComP3 focuses more on locating relatively large size earthquakes and minimizing the chance of detecting false events, some of the micro earthquakes do not get detected by the existing function of SeisComP3. Here, we developed a new SeisComP3 module, which binds signals detected at multiple stations within a certain distance and time span. We applied this module, which we call "binder", to one daylong DONET data. We examined four types of bandpass filter, which have their bandwidth somewhere between 2 and 30 Hz, and concluded that the filter with corner frequencies at 4 and 20 Hz most efficiently detects the signal of the micro earthquakes. We confirmed that the binder is useful for detecting micro earthquakes. Moreover, one of the binder's advantages is that it works without assumptions on the velocity profile, thus it detected enigmatic events whose epicenters are located on the Kumano Basin and whose signal propagates at the velocity of around 1.5 km/s throughout the stations placed on the basin. No clear P arrivals were observed for these events. The finding demonstrates the possibility of the binder to become a useful tool to further explore the data and to find out more of different types of events, which have not been identified yet.