Kinematics of the Southeastern Tibetan Plateau from Sentinel-1 InSAR and GNSS: Implications for Seismic Hazard Analysis

Abstract

Earthquakes release strain energy that has accumulated between seismic events. Measuring strain accumulation rates is critical for understanding earthquake cycle and assessing earthquake potential, with fault slip rates serving as essential inputs for seismic hazard models. However, the Tibetan Plateau has been lacking comprehensive estimates of geologic slip rates on numerous faults. To address this gap, geodetic data have been invoked to derive fault slip (or slip deficit) rates using various methodologies. These include the commonly adopted classic and deformable block modelling approaches (Meade & Loveless, 2009) and the newly developed direct inversion of geodetic strain rates (Johnson et al., 2022), which has the advantage of not requiring blocks to be defined.  A comprehensive comparison of slip rates obtained from these different geodetic methods has been notably absent. In this study, we focus on the southeastern Tibetan Plateau, utilising Sentinel-1 satellite data from 35 ascending and 32 descending frames spanning the period between 2014 and 2023, along with published GNSS velocities. We constructed high-resolution (1 km) maps of velocity and strain rate fields covering 1.3 million km2. Using these maps, we derived slip rates on newly mapped faults (Styron, 2022) using classic block modelling, “deformable block” modelling, and by the direct inversion of strain rates. Our strain rate fields reveal a partition through focused shear on the Kunlun fault, the Xianshuihe-Xiaojiang fault system, the Longriba fault, the Longmenshan fault possibly influenced by the ongoing postseismic deformation of the 2008 Mw 7.9 Wenchuan earthquake, and the Lijiang-Xiaojinhe fault. On the deforming plateau there is diffuse deformation away from the major faults, with average shear strain and dilatation rates of 14.3 and 13.1 nanostrain/year, compared to 9.4 and 11.1 nanostrain/year in the Sichuan basin (which likely reflects the noise floor in the data). The geodetically-determined slip rates from the three methods generally align with available geologic rates, particularly along-strike variations on the Kunlun fault and the Xianshuihe-Xiaojiang fault system. Our block model consists of 103 blocks bounded by 326 fault sections in the southeastern Tibetan Plateau. The model is constrained by the combined geodetic horizontal velocities from 6617 observation points. Classic block modelling without considering internal strain tends to overestimate slip rates on faults that slip faster than 5 mm/yr, compared to deformable block model that accounts for homogeneous intrablock strain, constituting 5% of the total. The two block models explain approximately 45-50% of the geodetic strain, predicting focused strain on block boundaries even in the absence of observed strain concentrations. By directly inverting strain rates, we suggest that 40-50% of the geodetic strain is attributable to elastic coupling (back slip) on faults, while the remaining can be explained by off-fault distributed moment sources (body forces) in a thin elastic plate. We discuss limitations of different geodetic approaches in modelling deformation (velocities or strain rates) and implications for seismic hazard by comparing the seismic moment release rate from earthquakes and the geodetic moment accumulation rate from our geodetic models.