5:15 PM - 6:45 PM
[SSS08-P01] Probabilistic seismic hazard analysis of the eastern margin of the Japan Sea: Considerations regarding the settings of the area sources.
Keywords:Probabilistic seismic hazard analysis , Eastern Margin of the Japan Sea, Seismicity declustering, Area sources
Introduction
In probabilistic seismic hazard analysis, seismic sources are designated as “characteristic sources” or “area sources”. In sea areas such as the eastern margin of the Japan Sea, active faults that become characteristic sources are extracted mainly using reflection sections obtained through acoustic wave exploration, and fault models are established (e.g., Ministry of Land, Infrastructure, 2014). Area sources are used to cover the uncertainties. Area sources are often established using the G-R law based on seismic observation data, but removing aftershocks from the observation data accurately and representing multiple sources efficiently is a challenge. In this study, we conducted a comparison of multiple declustering methods and a basic investigation of the mechanism of seismic sources to evaluate the impact of these challenges on seismic hazard.
Data
The earthquake catalog published by the Japan Meteorological Agency was used to set the occurrence frequency, with data from 1919 to 2019. In addition, in Investigation 2 (B) below, information on the mechanisms from F-net was also used.
Study Contents and Results
In this study, we investigated the following two items and their effects:
(Investigation 1) Effects of aftershock removal methods on seismic hazard
The national seismic hazard map published by HERP (e.g., HERP, 2009) removes aftershocks from the earthquake catalog using the window function proposed by the Ministry of Construction (1983). However, other aftershock removal methods, such as Gardner and Knopoff (1974) and Musson (1999), have been proposed. Therefore, we investigated the differences in the trends of aftershock removal for each method and their impact on seismic hazard. As a result, we found that the method of Gardner and Knopoff (1974) removed more aftershocks than that of HERP (2009), while method of Musson (1999) removed fewer. Differences were also observed in the PGA hazard curve. Therefore, it is desirable to include aftershock removal methods as epistemic uncertainty in the logic tree.
(Investigation 2) Differences in hazard curves between point and area sources
The HERP uses a set of point sources with occurrence frequencies set at 0.1-degree intervals, based on the study by the National Research Institute for Earth Science and Disaster Resilience (2003), as area sources. However, in cases of large maximum magnitudes, area sources may not be regarded as point sources. Therefore, in this study, sensitivity analyses were conducted from three perspectives: (A) area sources and point sources, (B) mechanism of the area source, and (C) scaling laws for setting area sources.In (B), we examined the difference in seismic hazard between a uniform distribution of the strike, dip, and depth directions of the area source and a probability-weighted setting based on observed mechanisms. In (C), we examined the difference in seismic hazard using the scaling laws proposed by Utsu and Seki (1955) and Wells and Coppersmith (1994).As a result of the investigation, (A) showed that the hazard increased with increasing seismic motion amplitude for area sources compared to point sources. (B) and (C) did not show significant differences at the targeted locations in this study.
Acknowledgments
This study was conducted as part of the commissioned research work of the Nuclear Regulation Authority (NRA), "Preliminary Survey of Probabilistic Seismic Hazard Evaluation for Nuclear Facilities and Disaster Prevention Measures in Sea Areas in FY2023." We would like to express our sincere appreciation to Dr. Morikawa of the National Research Institute for Earth Science and Disaster Resilience (NIED) for his guidance on the research contents. We would also like to thank the Japan Meteorological Agency for providing the earthquake catalog and the NIED for providing the source information from F-net. We are grateful to all those involved in their efforts.
In probabilistic seismic hazard analysis, seismic sources are designated as “characteristic sources” or “area sources”. In sea areas such as the eastern margin of the Japan Sea, active faults that become characteristic sources are extracted mainly using reflection sections obtained through acoustic wave exploration, and fault models are established (e.g., Ministry of Land, Infrastructure, 2014). Area sources are used to cover the uncertainties. Area sources are often established using the G-R law based on seismic observation data, but removing aftershocks from the observation data accurately and representing multiple sources efficiently is a challenge. In this study, we conducted a comparison of multiple declustering methods and a basic investigation of the mechanism of seismic sources to evaluate the impact of these challenges on seismic hazard.
Data
The earthquake catalog published by the Japan Meteorological Agency was used to set the occurrence frequency, with data from 1919 to 2019. In addition, in Investigation 2 (B) below, information on the mechanisms from F-net was also used.
Study Contents and Results
In this study, we investigated the following two items and their effects:
(Investigation 1) Effects of aftershock removal methods on seismic hazard
The national seismic hazard map published by HERP (e.g., HERP, 2009) removes aftershocks from the earthquake catalog using the window function proposed by the Ministry of Construction (1983). However, other aftershock removal methods, such as Gardner and Knopoff (1974) and Musson (1999), have been proposed. Therefore, we investigated the differences in the trends of aftershock removal for each method and their impact on seismic hazard. As a result, we found that the method of Gardner and Knopoff (1974) removed more aftershocks than that of HERP (2009), while method of Musson (1999) removed fewer. Differences were also observed in the PGA hazard curve. Therefore, it is desirable to include aftershock removal methods as epistemic uncertainty in the logic tree.
(Investigation 2) Differences in hazard curves between point and area sources
The HERP uses a set of point sources with occurrence frequencies set at 0.1-degree intervals, based on the study by the National Research Institute for Earth Science and Disaster Resilience (2003), as area sources. However, in cases of large maximum magnitudes, area sources may not be regarded as point sources. Therefore, in this study, sensitivity analyses were conducted from three perspectives: (A) area sources and point sources, (B) mechanism of the area source, and (C) scaling laws for setting area sources.In (B), we examined the difference in seismic hazard between a uniform distribution of the strike, dip, and depth directions of the area source and a probability-weighted setting based on observed mechanisms. In (C), we examined the difference in seismic hazard using the scaling laws proposed by Utsu and Seki (1955) and Wells and Coppersmith (1994).As a result of the investigation, (A) showed that the hazard increased with increasing seismic motion amplitude for area sources compared to point sources. (B) and (C) did not show significant differences at the targeted locations in this study.
Acknowledgments
This study was conducted as part of the commissioned research work of the Nuclear Regulation Authority (NRA), "Preliminary Survey of Probabilistic Seismic Hazard Evaluation for Nuclear Facilities and Disaster Prevention Measures in Sea Areas in FY2023." We would like to express our sincere appreciation to Dr. Morikawa of the National Research Institute for Earth Science and Disaster Resilience (NIED) for his guidance on the research contents. We would also like to thank the Japan Meteorological Agency for providing the earthquake catalog and the NIED for providing the source information from F-net. We are grateful to all those involved in their efforts.