Applying unsupervised machine-learning algorithms and MUSIC back-projection to characterize 2018–2022 Hualien earthquake sequence

The city of Hualien in eastern Taiwan is located at the junction of the Luzon arc–Eurasian continent collision and the oblique subduction of the Philippine Sea plate along the Ryukyu Trench. This complex tectonic setting creates one of the most intense seismic hazards in the world. Since the mid-twe...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:TAO : Terrestrial, atmospheric, and oceanic sciences Ročník 33; číslo 1; s. 1 - 18
Hlavní autoři: Jian, Pei-Ru, Wang, Yu
Médium: Journal Article
Jazyk:angličtina
Vydáno: Singapore Springer Nature Singapore 01.12.2022
Springer
Témata:
Atmospheric Sciences
Central Range fault
Coastal rotation structure
DBSCAN
Earth and Environmental Science
Earth Sciences
Hualien earthquakes
Oceanography
Original Article
PCA
Unsupervised machine learning
Unsupervised machine learning
Hualien earthquakes
DBSCAN
Central Range fault
Coastal rotation structure
PCA
ISSN:1017-0839, 2311-7680
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:The city of Hualien in eastern Taiwan is located at the junction of the Luzon arc–Eurasian continent collision and the oblique subduction of the Philippine Sea plate along the Ryukyu Trench. This complex tectonic setting creates one of the most intense seismic hazards in the world. Since the mid-twentieth century, three earthquake sequences (1951, 1986–1992, and 2018–2022) have caused severe damages and revealed a diversity of focal mechanisms, demonstrating a complex seismogenic tectonic interactions of collisions and subductions. In this study, we used the two-dimensional back-projection method to examine rupture characteristics of three strong earthquakes during the latest sequence (2018–2022). We also studied the same earthquake sequence using unsupervised machine-learning via density-based spatial clustering of applications with noise (DBSCAN) and principal component analysis (PCA). Our results revealed two westward dipping seismogenic structures. The deeper one lies beneath the eastern flank of the Central Range and the shallower one rotates clockwise along the coastline from south to north. Overall, focal mechanisms within these two structures are consistent with planar geometries obtained using PCA. The deeper structure beneath the Central Range is likely an imbricated fault system. The coastline structure is characterized by a transition of strike and dip angles from south to north. We suggest that the subhorizontal nodal planes in earthquakes offshore Hualien are related to the shallow northward dipping patch in the northern part of the coastal structure. Our Coulomb stress change results revealed strong linkages between these two structures. Key points We used unsupervised machine-learning algorithms DBSCAN and PCA to study the 2018–2022 Hualien earthquake sequence. A deep westward-dipping and a shallow rotation structure system are revealed from earthquake clusters close to the northernmost Longitudinal Valley. Coulomb stress change is used to ascertain cascaded triggering on these two structures.
ISSN:1017-0839
2311-7680
DOI:10.1007/s44195-022-00026-y