T003-0013
Low-frequency tremor activity along northern Japan Trench before the 2011 Tohoku-Oki earthquake

Monday, 7 December 2020
Poster
Hidenobu Takahashi1, Ryota Hino1, Naoki Uchida1, Kazuaki Ohta2 and Masanao Shinohara3, (1)Tohoku University, Graduate School of Science, Sendai, Japan, (2)NIED National Research Institute for Earth Science and Disaster Prevention, Tsukuba, Japan, (3)Univ Tokyo, Tokyo, Japan
Abstract:
Recent studies revealed that various kinds of slow earthquakes occur broadly along the Japan Trench (e.g., Nishikawa et al., 2019), but their activity is significantly low in the coseismic rupture zone of the 2011 Tohoku-Oki earthquake. However, direct observations of slow-earthquake phenomena, e.g. low-frequency tremors (LFTs) and very low frequency earthquakes (VLFEs), are limited prior to the 2011 earthquake and the slow earthquake distribution has been mostly discussed based on small repeating earthquakes or on earthquake swarms.

Here, we report on the LFT activity along the northern Japan Trench before the 2011 Tohoku-Oki earthquake. We used 48 ocean bottom seismometers (OBSs) deployed from Oct. 2007 to May 2008. An envelope correlation method (e.g., Obara, 2002) was applied to the OBS seismograms. About 500 LFTs were detected with reliable hypocenter locations. Generally, the hypocenter distribution resembles well to that after 2016 estimated by the seafloor cabled network (S-net) observation. The LFT epicenters show along-strike migration during multiple episodes, each of which has ~1 week duration. Regular earthquakes next to the LFTs were detected as events of short duration. It is notable that the estimated focal depths of detected events are almost the same regardless of durations, meaning that LFTs and regular earthquakes occur at the identical depth range, probably on the plate interface. We relocated the regular earthquakes by using P & S-arrivals to confirm that their focal depths are consistent with the known plate boundary depth. We identified an intensive cluster of LFTs at the end of 2007 within the rupture area of the 2011 earthquake, whereas few events were detected by S-net. This suggests that LFTs were much more active before the 2011 earthquake around the cluster location. The LFT activity may have been calmed down reflecting the change in slip-rate on the plate boundary before and after the massive interplate earthquake.