Bayesian inversion and quantitative comparison for bilaterally quasi-symmetric rupture processes on a multisegment fault in the 2021 Mw7.4 Maduo earthquake

Abstract

Summary On May 21st, 2021, the Mw7.4 Maduo earthquake occurred in the southern Qinghai Province, China. This earthquake ruptured approximately 160 km along the Kunlunshankou-Jiangcuo fault, an east-west trending fault located in the middle of the Bayan Har Block (BHB). The seismogenic fault exhibits an apparent simple geometry, characterized by fault branches to the east and a splay fault to the west. Despite the apparent simplicity of the fault’s structure, a noteworthy level of variability and inconsistencies persist in the representations of fault geometry in published rupture models of the earthquake. Our study employs a Bayesian approach to elucidate both the fault geometry and kinematic rupture parameters of the earthquake. We use three-dimensional displacements obtained from synthetic aperture radar (SAR) images and teleseismic data to quantify the rupture kinematics. We then conduct three separate finite-fault inversions (FFI) using individual datasets, and perform a subsequent joint inversion for a comprehensive analysis. Additionally, we employ teleseismic backprojection (BP) to complement the kinematic analysis of the earthquake rupture. Both kinematic finite-fault inversion and backprojection unveil a nearly symmetric bilateral rupture in the east-west direction, characterized by an average rupture speed of approximately 2.7 km/s. The rupture to the east displays a heightened level of complexity, manifested in at least five discernible stages, whereas the rupture to the west is comparatively simpler. The eastward rupture directly triggered the southern branch of the bifurcating fault, with a notable delay of approximately 3 s on the northern branch. Several studies have presented coseismic slip models for the earthquake. An analysis of variability among 10 slip models, including our preferred model, highlights that fault geometry and inversion strategy (e.g., fault discretization, smoothing factor) contribute to considerable variability in both slip magnitude and slip extent on the fault, despite similar data types being used in the inversions. Furthermore, the finite-fault model acquired through slip inversion plays a crucial role in calculating Coulomb failure stress change (ΔCFS) transmitted from the source fault to neighboring receiver faults. Understanding how the variability in slip models influences ΔCFS calculations is essential for conducting comprehensive analyses in seismic hazard studies. Our findings highlight that discrepancies in fault geometry contribute to the variance of ΔCFS in the regions delineating positive and negative stress change. Meanwhile, variability in slip magnitude substantially impacts the variability of ΔCFS in the vicinity of the source fault. Furthermore, our analysis of ΔCFS calculations using our preferred slip models indicates that a major event on the Maqin-Maqu segment, a well-recognized seismic gap on the East Kunlun Fault (EKF), could potentially be advanced in time.