日本地球惑星科学連合2025年大会

講演情報

[J] 口頭発表

セッション記号 S (固体地球科学) » S-TT 計測技術・研究手法

[S-TT43] 最先端ベイズ統計学が拓く地震ビッグデータ解析

2025年5月26日(月) 10:45 〜 12:15 201A (幕張メッセ国際会議場)

コンビーナ:長尾 大道(東京大学地震研究所)、加藤 愛太郎(東京大学地震研究所)、矢野 恵佑(統計数理研究所)、椎名 高裕(産業技術総合研究所)、座長:長尾 大道(東京大学地震研究所)、加藤 愛太郎(東京大学地震研究所)、矢野 恵佑(統計数理研究所)、椎名 高裕(産業技術総合研究所)

10:45 〜 11:00

[STT43-01] Slip-size-dependent Brownian passage time model with slip-size uncertainties

★Invited Papers

*寺田 吉壱1 (1.大阪大学)

This study focuses on the evaluation of long-term earthquake probabilities. For recurrence interval data, we estimate the probability of the next earthquake using non-negative distributions, such as the Brownian passage time (BPT) distribution. When slip size data are available, the time-predictable model is often used to estimate the timing of the next earthquake. However, this model has faced significant criticism. For example, Rubinstein et al. (2012) demonstrated its limitations for short-term predictions. Statistically, the time-predictable model has a fundamental drawback: it fails to account for random noise in recurrence interval data.
To address this issue, Ogata (2002) introduced a stochastic extension of the time-predictable model, called the slip-size-dependent Brownian passage time (SSD-BPT) model. Bayesian inference for the SSD-BPT model has been developed to incorporate uncertainties in hazard estimates.
However, obtaining accurate slip size data from historical earthquakes remains a significant challenge. To overcome this limitation, we propose a Bayesian framework for the SSD-BPT model incorporating slip-size uncertainties. Additionally, we extend the SSD-BPT model to incorporate older recurrence interval data that lack slip size records, enhancing its applicability for seismic hazard assessment.