5:15 PM - 6:45 PM
[SSS05-P05] Validating the Method of Estimating Rake Angles Using Regional Stress Field with the Wallace-Bott Hypothesis for Focal Mechanism Solutions
Keywords:Tectonic stress field, The Wallace-Bott hypothesis, Fault slip angles, Focal mechanism solutions for microearthquakes
For realistic evaluation of strong ground motions and tsunamis from future earthquakes, it is required that the input data, such as fault geometry, fault slip angle (direction of displacement), and amount of fault slip, do not deviate from those of actual earthquakes. Recently, based on the hypothesis that the slip direction is parallel to the direction of shear traction on the fault plane (Wallace-Bott hypothesis; WBH), the fault slip angle was estimated from regional stress fields estimated from seismological data and the estimated fault slip angles were used for the evaluations of strong ground motion and tsunami (e.g., Satake et al., 2022, EPS). However, the validity and accuracy of the above methods have not been sufficiently examined. In addition, it is necessary to examine whether the regional stress field estimated from approximately ten years of seismological data could be used as a proxy for the long-term tectonic stress field.
As a first attempt for such validation, our previous study (Ishibe et al., 2024; I2024, EPS) validated the method by comparing the calculated rake angles with the observed (determined) rake angles for focal mechanism solutions of earthquakes from the two catalogs: i.e. F-net mechanism solutions (NIED 2024) and JUNEC FM2 (Ishibe et al. 2014, BSSA), with calculated rake angles according to the method. I2024 also indicated that the calculated rake angles were generally consistent with geomorphologically and geologically evaluated fault types for Quaternary active faults (zones) in southwestern Japan by HERP (e.g., HERP 2013).
In recent years, a deep learning method using a conventional neural network (CNN) has been developed to pick first motion polarities from abundant seismic waveform records and applied to the Hi-net data and then, the focal mechanism solutions for numerous microearthquakes under the Japanese islands were estimated (Uchide et al., 2022, U2022) In the present study, we validated the effectiveness and availability of the method using focal mechanism solutions for abundant microearthquakes by U2022. We calculated rake angles from the tectonic stress field of Terakawa and Matsu'ura (2010) according to WBH by fixing the fault strike and dip angles to those of the focal mechanism solutions. The calculated rake angles were then compared with the determined rake angles from U2022 by calculating misfit angles (hereafter, denoted by λ). The smaller absolute misfit angle (hereafter, denoted by |λ|) between the first and second nodal planes was adopted as the representative value.
As a result, the calculated rake angles were mostly consistent with the observed ones with |λ| <30°. Among 18,310 earthquakes for the earthquakes which occurred form January 2005 to just before the occurrence of the 2011 Tohoku-oki earthquake, |λ| was =<30° for 12,250 (~66.9 %) earthquakes, while 2,418 (~13.2 %), 1,015 (~5.5 %) and 2,627 (~14.3 %) earthquakes have 30°<|λ|=<60°, 60°<|λ|=<90° and |λ|>90°, respectively. In addition, basic characteristics of the fault types were well reproduced by the method: i.e., reverse-type faulting in Hokkaido and Tohoku districts, a mixture of reverse and strike-slip types in Kanto, Chubu and Kinki districts, predominantly strike-slip faulting in Chugoku and Shikoku districts, and a mixture of strike-slip and normal faulting in Kyushu district. However, the relatively large misfit angles (>30°) were concentrated in several regions such as the source and surrounding regions of major earthquakes and the region along the outer zone of southwestern Japan. According to the results in the present study in addition to I2024, we conclude that the rake angles of future large earthquakes can be estimated from regional tectonic stress fields according to WBH, while the careful attention is required for applying the method to the regions where the stress fields have been perturbed and low seismicity area with the number of available focal mechanism data to invert stress fields are limited.
Acknowledgements
This study has been supported by the Headquarters for Earthquake Research Promotion (HERP) of the Ministry of Education, Culture, Sports, Science, and Technology (MEXT) of Japan, and MEXT of Japan, under its The Second Earthquake and Volcano Hazards Observation and Research Program (Earthquake and Volcano Hazard Reduction Research).
As a first attempt for such validation, our previous study (Ishibe et al., 2024; I2024, EPS) validated the method by comparing the calculated rake angles with the observed (determined) rake angles for focal mechanism solutions of earthquakes from the two catalogs: i.e. F-net mechanism solutions (NIED 2024) and JUNEC FM2 (Ishibe et al. 2014, BSSA), with calculated rake angles according to the method. I2024 also indicated that the calculated rake angles were generally consistent with geomorphologically and geologically evaluated fault types for Quaternary active faults (zones) in southwestern Japan by HERP (e.g., HERP 2013).
In recent years, a deep learning method using a conventional neural network (CNN) has been developed to pick first motion polarities from abundant seismic waveform records and applied to the Hi-net data and then, the focal mechanism solutions for numerous microearthquakes under the Japanese islands were estimated (Uchide et al., 2022, U2022) In the present study, we validated the effectiveness and availability of the method using focal mechanism solutions for abundant microearthquakes by U2022. We calculated rake angles from the tectonic stress field of Terakawa and Matsu'ura (2010) according to WBH by fixing the fault strike and dip angles to those of the focal mechanism solutions. The calculated rake angles were then compared with the determined rake angles from U2022 by calculating misfit angles (hereafter, denoted by λ). The smaller absolute misfit angle (hereafter, denoted by |λ|) between the first and second nodal planes was adopted as the representative value.
As a result, the calculated rake angles were mostly consistent with the observed ones with |λ| <30°. Among 18,310 earthquakes for the earthquakes which occurred form January 2005 to just before the occurrence of the 2011 Tohoku-oki earthquake, |λ| was =<30° for 12,250 (~66.9 %) earthquakes, while 2,418 (~13.2 %), 1,015 (~5.5 %) and 2,627 (~14.3 %) earthquakes have 30°<|λ|=<60°, 60°<|λ|=<90° and |λ|>90°, respectively. In addition, basic characteristics of the fault types were well reproduced by the method: i.e., reverse-type faulting in Hokkaido and Tohoku districts, a mixture of reverse and strike-slip types in Kanto, Chubu and Kinki districts, predominantly strike-slip faulting in Chugoku and Shikoku districts, and a mixture of strike-slip and normal faulting in Kyushu district. However, the relatively large misfit angles (>30°) were concentrated in several regions such as the source and surrounding regions of major earthquakes and the region along the outer zone of southwestern Japan. According to the results in the present study in addition to I2024, we conclude that the rake angles of future large earthquakes can be estimated from regional tectonic stress fields according to WBH, while the careful attention is required for applying the method to the regions where the stress fields have been perturbed and low seismicity area with the number of available focal mechanism data to invert stress fields are limited.
Acknowledgements
This study has been supported by the Headquarters for Earthquake Research Promotion (HERP) of the Ministry of Education, Culture, Sports, Science, and Technology (MEXT) of Japan, and MEXT of Japan, under its The Second Earthquake and Volcano Hazards Observation and Research Program (Earthquake and Volcano Hazard Reduction Research).