日本地球惑星科学連合2025年大会

講演情報

[E] 口頭発表

セッション記号 S (固体地球科学) » S-CG 固体地球科学複合領域・一般

[S-CG45] Science of slow-to-fast earthquakes

2025年5月28日(水) 10:45 〜 12:15 国際会議室 (IC) (幕張メッセ国際会議場)

コンビーナ:加藤 愛太郎(東京大学地震研究所)、山口 飛鳥(東京大学大気海洋研究所)、中田 令子(東京大学大学院理学系研究科)、大久保 蔵馬(防災科学技術研究所)、座長:竹尾 明子(東京大学地震研究所)、利根川 貴志(海洋研究開発機構 地震津波海域観測研究開発センター)

11:30 〜 11:45

[SCG45-46] Very low-frequency earthquakes in the eastern Shikoku observed by campaign deployment of broadband seismometers

*竹尾 明子1加藤 愛太郎1 (1.東京大学地震研究所)

Tremors or very low frequency (VLF) earthquakes are seismic signals of slow earthquakes observed at frequency ranges of 1–10 or 0.01–0.1 Hz, respectively. These signals are often observed during the geodetic slow slip events in subduction zones such as Nankai and Cascadia. Mathematically, those seismic signals can be attributed to the perturbation and/or acceleration in the moment release rate (and thus the slip rate) of geodetic slow slip at the frequency range of observations. However, the quantitative evaluation of the slip rate is still limited due to the difficulty of estimating the precise moment release rate and the slip area. To evaluate the moment rate during the geodetic slow slip, we deployed two broadband seismometers in November 2024 in the eastern Shikoku area, where slow slip occurs every ~3 months and where the migration of tremor will be resolved by the simultaneous observation of distributed acoustic sensing (DAS).

We installed Trillium Horizon broadband seismometers of Nanometrics, whose natural period is 120 s. The seismometers were buried in the soil for ~80 cm near heliports for emergencies. The data has been sent to an online server by a mobile network almost instantaneously. We analyzed the three-component data in early January, when a tremor episode occurred near our observation network. The details of the analyses are similar to those of Takeo et al. (Workshop on Slow Earthquakes, 2021). We down-sampled the data to 1 Hz, filtered in 20–50 s, and grid-searched the epicenter by assuming the focal depth of 35 km and determining each grid's CMT solutions. The waveform length of analysis is 120 s, and we got solutions and a variance reduction value every 1 s. We compared the variance reduction value with that obtained by the incidence of plane waves to distinguish local events from teleseismic events. As a result, we first detected three VLF events with moment magnitudes (Mw) of 3.05–3.25. We lowered the detection threshold to detect more events and found 20 events with Mw 2.87–3.25 by visual checking. All events coincided with the tremor episode.

Several events show signals in a longer period range of 50–100 s, which indicates the possibility of estimating a more accurate moment release rate. We will continue the analysis of those signals and will compare the result with the area of sources determined from tremor records by DAS to discuss the slip rate of seismic slow signals in the future