Fault Geometries of the 2024 Mw 7.5 Noto Peninsula Earthquake From Hypocenter-Based Hierarchical Clustering of Point-Cloud Normal Vectors

Abstract

The elucidation of intricate fault geometries provides fundamental and essential information regarding seismology and other fields of solid Earth sciences. Hypocenter alignments typically reflect complex crustal fault structures, so spatial clustering of hypocenter distributions has been used to construct planar fault geometries. However, conventional spatial clustering inherently struggles with the complexity of hypocenter distributions. In this study, we integrated point-cloud normal vectors, commonly used in object recognition to reflect the local surface geometry of an object, into a hypocenter-based hierarchical clustering to construct intricate planar fault models. We applied this method to the aftershock sequences of the Mw 7.5 Noto Peninsula earthquake in central Japan on 1 January 2024, which caused notable crustal deformation. We identified fault planes aligning with the coastline from the western to northern coast. A southeast-dipping plane was located between the two south-southeast-dipping planes along the northern coast, correlating with gravity anomalies and surface geology or reflecting the complexity of fault ruptures and dynamic stress perturbations. The east-dipping fault in the southwestern area showed a different distribution from the aftershocks of the 2007 Mw 6.7 earthquake, suggesting that the 2024 earthquake did not reactivate the 2007's fault plane. The NS-trending aftershock focal mechanisms in the southwestern area suggest that a reverse-fault slip probably occurred on the plane. Further investigations based on the intricate fault planes will contribute to a deeper understanding of the spatial characteristics of the coseismic slip of the 2024 earthquake and seismotectonics of the Noto Peninsula.