[M-IS09] Interdisciplinary studies on pre-earthquake processes

[MIS09-03] Changes in permutation entropy based on seismic noise prior to the 2024 Mw 7.4 Hualien Earthquake

*Cholisina Anik Perwita¹、Konstantinos I Konstantinou²、Wen Tzong Liang³ (1.Taiwan International Graduate Program (TIGP)–Earth System Science Program, (ESS) Academia Sinica and National Central University, Academia Sinica, Taipei 11529, Taiwan、2.Department of Earth Sciences, National Central University, Jhongli 320, Taiwan、3.Institute of Earth Sciences, Academia Sinica, Taipei 11529, Taiwan)

キーワード：permutation entropy, seismic noise, earthquake, Hualien

Permutation entropy (PE) integrates the principles of entropy and dynamics and offers a way to measure randomness in time series by identifying permutation patterns. PE will take on values that fluctuate between 0 and 1, where 0 denotes periodic signals and 1 denotes stochastic ones. On April 3, 2024, a Mw ~7.4 earthquake struck off the coast of Hualien County, Taiwan. It was felt island-wide and became the most powerful seismic event since the 1999 ChiChi earthquake. Previously, PE has been successfully used to detect precursors to volcanic eruptions. This study investigates its potential use as a tool for identifying precursory changes in seismic noise prior to the 2024 Hualien earthquake. PE was calculated from the vertical component of continuous waveforms recorded by 10 broadband seismometers provided by the CWASN and BATS networks, primarily located along Taiwan's east coast. The analysis covered four months from 1 December 2023 to 5 April 2024. PE calculation was performed for 20-minute windows, using an embedding dimension of 5 and a delay time of 3. To identify changepoints and nonlinear trends in the PE time series, we used the BEAST (Bayesian Estimator of Abrupt change, Seasonal change, and Trend) algorithm. Applying BEAST to the PE time series revealed a significant upward trend in the majority of stations located within 62.62 km from the earthquake epicenter, starting 60 to 40 days before the onset. The observed rise in PE can likely be attributed to the accumulation of elastic energy before the earthquake. Strain buildup might have initiated slow micro-fracturing, which would amplify scattering strength and increase high-frequency amplitudes, ultimately resulting in elevated PE values.