昨年の発表(H. Tanaka-K. Umeno, AGU24)に引き続き、今年も、世界で初めて検出したとする大地震発生前の”プレスリップ”について、世界最大の地球物理学の国際会議AGUで口頭発表します。

Whether crustal deformation known as “preslip” that precedes earthquakes exists or not is a crucial issue linked to the realization of earthquake prediction. Precursory signals from GPS data were reported approximately two hours before Mw ≥ 7.0 earthquakes worldwide [1], but a subsequent study raised concerns about network-wide noise—Common Mode Error (CME) [2]. The method used in [1] is based on focal mechanism solutions and discusses the statistical properties of major earthquakes, so it is unsuitable for real-time forecasting of individual events.
To address these issues, a correlation analysis method (CRA) was developed to analyze the spatiotemporal correlation of GNSS station displacements [3]. This method detected precursory deformation prior to the 2011 Tohoku-Oki mainshock using only pre-event data. However, the distinction between CME and crustal deformation signals remained qualitative, merely suggesting the crustal deformation signal’s dependence on epicentral distance without a quantitative assessment [4].
In this study, we conduct a crustal deformation analysis that builds upon the correlation analysis method by incorporating a CME filtering process and a quantitative assessment of the anomaly’s distance dependence. We quantitatively evaluate whether preslip signals exhibit a power-law decay with distance, drawing on theoretical and observational evidence that shows this characteristic for coseismic displacement [5-6]. We statistically evaluate the significance of preslip signals preceding the 2011 Tohoku-Oki earthquake using long-term data.
In this presentation, we will describe the CME filtering method, the evaluation of distance-dependent features, and the analysis results, highlighting the potential to detect crustal deformation anomalies several hours before an earthquake. We consider this study an important step toward the development of a real-time earthquake forecasting system.
[1] Q. Bletery and J.-M. Nocquet, Science 381 (6655), 297-301 (2023)
[2] Q. Bletery and J.-M. Nocquet, EarthArXiv (2024) DOI:10.31223/X5RT3N
[3] H. Tanaka and K. Umeno, AGU24, December (2024)
[4] H. Tanaka and K. Umeno, Slow- and pre-slip signals detected in crustal deformation before Mw9.0 Tohoku-Oki earthquake (2025-11-12)
doi: https://doi.org/10.22541/essoar.176296628.85631921/v1
[5] K. Aki and P. Richards, Quantitative Seismology: Theory and Methods, W.H. Freeman and Company (1980)
[6] P. Xu, Tectonophysics, 907, 230742 (2025)