The 2017 Mw 6.5 Jiuzhaigou earthquake (JE) struck a rugged area of the Jiuzhaigou Valley in eastern Tibet that has experienced frequent seismic activity over the last few decades. We use finite-element models (FEMs) and Sentinel-1 Interferometric Synthetic Aperture Radar observations to characterize the earthquake source. The FEM domain accommodates a heterogeneous (HET) distribution of realistic crustal materials inferred by regional seismic tomography data. The HET-derived source configurations yield a significantly smaller misfit, at the 95% confidence level, than that estimated for a homogeneous (HOM) half-space. The former generally requires a lower degree of smoothing constraint, highlighting that the HET solutions are systematically more compatible with the surface observations than the HOM solutions. The magnitudes of induced Coulomb failure stress change (ΔCFS) estimated by the HET solution drastically differ (by >0:1 MPa) from those calculated by the HOM solution. The postearthquake stability of near-field faults is generally overestimated by the HOM estimations, whereas some localities of negative ΔCFSHOM are predicted with positive ΔCFSHET. These results highlight the sensitivities of both slip and stress estimations to the complexity of the adopted elastic modeling domain, leading to more accurate aftershock hazard assessments. The HET-resolved seismic rupture reveals two major slip asperities of magnitude up to 0.83 m distributed along the fault strike, which is coherent with the aftershock distribution. Two aftershock clusters are consistently found near or below these two peak-slip zones, which are imaged by the HET model but absent in the HOM solution. The JE hypocenter and aftershocks are bounded below by a negative velocity anomaly (ΔVP, ΔVS down to -4%) at ∼18 km depth. Such low-velocity layers of reduced strength may be relevant to the vertical distribution of seismicity and earthquake slip, which provide insights into assessing the seismic hazards and aftershock-prone areas of the eastern Tibetan margin.
All Science Journal Classification (ASJC) codes