[M-IS04] Interdisciplinary studies on pre-earthquake processes

[MIS04-15] A Method for Earthquake Prediction Using Quasi-four Components Strain-meter

Yukio Fujinawa¹, *Yoichi NODA², Chie Yoshino³, Katsumi Hattori^3,4,5 (1.Organization for Development of Resilient Communities, 2.Graduate School of Science and Engineering, Chiba University, 3.Graduate School of Science, Chiba University, 4.Center for Environmental Remote Sensing Chiba University, 5.Research Institute of Disaster Medicine Chiba University )

Keywords:Four-component Strain Meter, GNSS, Earthquake Precursors, Short-term earthquake prediction

The substantial damage assumed by major earthquakes requires the preparation of prevention methods especially short- and long-term prediction. Many kinds of research except ordinal observation by seismometer have been tried to develop practical methods without results except for two rare cases some 50 years. Among many phenomena, electromagnetic phenomena, crustal deformation and others have been widely investigated. For instance, Kamiyama et al. (2017) detected precursory phenomena before Kumamoto and Tohoku earthquake by analysing GNSS network, and Kong et al. (2018) find anomalous phenomena before large earthquakes in Yunnan province using four component strain meters in borehole. We need practical observation network covering whole nation. In Japan we can use the NEAD seismometer network using some 800 sites and MLIT GNSS network of some 1,300 sites.
Su (1997) find that that sum of strain of odd sensors of the four-component strain meter is same as that of even sensors in case of calm state. Using this property clear precursor phenomena were detected before several large earthquakes in Yunnan province (Kong et al., 2018).
Owing to this property, the correlation coefficient between odd and even sensors is for any time length during calm state. However, layout of four observation sites is rarely chosen square. We use approximate square for the four observation sites and find strain of baseline using the tensor transformation with the result to detect anomalous precursors with higher level of resolving power.
Here we try to check if or not to find short term precursors of large earthquake using several large earthquakes, starting the Kumamoto earthquake. We get the results that this new method shows similar result for six weeks before major earthquake period compared with the result of high frequency tremor (Fujinawa et al., 2022) to be used for social implementation.