Defining Surface Deformation Boundaries Using ICP and Clustering: 2024 Noto Peninsula Earthquake

Abstract

The 2024 Noto Peninsula Earthquake caused widespread slope deformations in northeastern Noto, where complex geological formations and weak sedimentary layers triggered large-scale mass movements. This study proposes a quantitative and reproducible method for analyzing internal deformation structures by combining Iterative Closest Point (ICP) analysis and Hierarchical clustering using multi-temporal S-DEM data. The ICP analysis was used to align pre- and post-event point clouds and derive high-resolution displacement vectors on a 10M grid, enabling the extraction of fine-scale deformation features. Hierarchical clustering was then applied to classify the displacement field based on spatial and kinematic parameters, objectively delineating coherent deformation blocks. The resulting clusters correspond well with geomorphologically interpreted scarps and sliding masses, demonstrating that the proposed approach effectively identifies block boundaries and internal variations. Patterns of uplift and subsidence further suggest secondary sliding behavior within the moving body. This ICP-clustering framework provides a robust basis for quantitatively segmenting slope deformations and statistically interpreting their internal kinematics following large earthquakes.