top of page

Science Corner Mar 2025

by Yihe Huang (University of Michigan)

Mar 22, 2025

Advancing Earthquake Physics through Direct Observations of Fault Zone Structures in Subduction Zones

While earthquakes in subduction zones seem to occur randomly in time, they tend to cluster in localized areas, which are likely controlled by fine-scale fault zone structures. Such structures can be seen from exhumed subduction zone rocks in certain regions, but direct observations of subduction zones at the depths of earthquakes and at the scale relevant to earthquake sources are still rare. As the size and spatial distribution of these structures could control the maximum size of subduction zone earthquakes, high-resolution imaging of the kilometer-size fault zone structures will significantly contribute to the assessment of earthquake hazards in subduction zones.


In this paper, we directly measure the in-situ material properties of fault zones that host frequent earthquakes at depths of 60 to 70 km along the plate interface between the Pacific plate and the Philippine plate in Japan. We find localized, anisotropic structures with a near-zero Poisson’s ratio (a median Vp/Vs ratio of 1.44). The heterogenous fault material properties, combined with stress drop estimates of deep interplate earthquakes, suggest that the fault medium is damaged, foliated, and enriched with fluid. Such small-scale material heterogeneity may cause faults stress perturbations that in turn favor the frequent occurrence of deep interplate earthquakes.


Huang, Y., Ide, S., Kato, A., Yoshida, K., Jiang, C., & Zhai, P. (2025). Fault material heterogeneity controls deep interplate earthquakes. Science Advances, 11(9), eadr9353. https://doi.org/10.1126/sciadv.adr9353


Vp/Vs ratios of earthquake patches and stress drop of M3.5 to M4.3 earthquakes in the patches. (A) shows the locations of earthquakes colored by the Vp/Vs ratios. The circle size is scaled by the rupture radius calculated from a stress drop of 3 MPa. The gray dots show the locations of M2 to M4 earthquakes. Depth contours of the PAC are obtained through linear interpolation of relocated earthquake depths and denoted by gray solid lines with numbers. (B) demonstrates the variations of stress drop. The circle size corresponds to eight times the rupture radius. The inset figures show the separation of high-stress-drop and low-stress-drop earthquakes on a real scale. (C) and (D) show the Vp/Vs ratios of earthquake patches at different depths and the stress drops of M3.5 to M4.3 earthquakes, respectively, with the dashed lines showing the median values and the same marker color corresponding to earthquakes in the same patch. The error bars in Vp/Vs ratios represent the median absolute deviation of the bootstrapping distribution, and the error bars in stress drop values represent the stress drops estimated from different eGfs.
Vp/Vs ratios of earthquake patches and stress drop of M3.5 to M4.3 earthquakes in the patches. (A) shows the locations of earthquakes colored by the Vp/Vs ratios. The circle size is scaled by the rupture radius calculated from a stress drop of 3 MPa. The gray dots show the locations of M2 to M4 earthquakes. Depth contours of the PAC are obtained through linear interpolation of relocated earthquake depths and denoted by gray solid lines with numbers. (B) demonstrates the variations of stress drop. The circle size corresponds to eight times the rupture radius. The inset figures show the separation of high-stress-drop and low-stress-drop earthquakes on a real scale. (C) and (D) show the Vp/Vs ratios of earthquake patches at different depths and the stress drops of M3.5 to M4.3 earthquakes, respectively, with the dashed lines showing the median values and the same marker color corresponding to earthquakes in the same patch. The error bars in Vp/Vs ratios represent the median absolute deviation of the bootstrapping distribution, and the error bars in stress drop values represent the stress drops estimated from different eGfs.

bottom of page