Skip to content

Detection and classification of pre-eruptive tilt changes for understanding immediate precursory processes of volcanoes

Yuta Maeda , Misaki Asai

Abstract

Tilt changes sometimes precede volcanic eruptions and are essential for understanding the immediate precursory processes. Maeda (2023) identified 3,976 waveforms that included tilt changes before eruptions at ten volcanoes in Japan. We present subsequent surveys of detecting pre-eruptive tilt changes worldwide and classifying them based on waveform similarity to understand the mechanisms of tilt changes. We surveyed the official websites of monitoring agencies and reviewed papers to identify the times of 281 eruptions at 20 monitored volcanoes in seven countries with a 1-min temporal resolution. We downloaded broadband seismic records from the EarthScope, IPGP, and Geofon data centers and applied the same procedure as Maeda (2023) to examine tilt changes before each eruption. The analysis is in progress; so far, we identified pre-eruptive tilt changes in six waveforms at Pagan, Semisopochnoi, Kanaga, and Cleveland volcanoes. Combined with the tilt changes in Japan identified by Maeda (2023), we classified the waveforms based on a matrix of misfits between all waveform pairs. The results consisted of 75 groups with  waveforms. We stacked the waveforms in each group and categorized them based on acceleration-deceleration patterns. Some groups showed acceleration that typically evolved to t1--2, where t is time. Only two groups showed deceleration (t<0.5). Mixed types, composed of acceleration followed by deceleration or vice versa, were the most frequent. This work was supported by JSPS KAKENHI Grant Number JP21H05203.