Objectives Source parameters in earthquake catalogs are commonly derived from inversions of seismic-wave observations, but their reliability is often limited by the number and spatial distribution of local seismic stations and uncertainties in crustal velocity models.Interferometry synthetic aperture radar (InSAR) provides high-spatial-resolution, centimeter-level coseismic deformation measurements, offering an effective way to better constrain the source parameters of shallow earthquakes. However, the most existing InSAR-based studies have focused on larger events (Mw>6.0), whereas moderate earthquakes, which occur more frequently, remain comparatively underexplored, and a comprehensive, systematic InSAR earthquake catalog has not yet been established. The Tibetan Plateau is an ideal region for developing an InSAR earthquake catalog, given its limited seismic-station density, commonly low seismic signal-to-noise ratios, and generally good InSAR coherence for coseismic deformation monitoring. Here, we develop and analyze an InSAR-based catalog of moderate earthquakes across the Tibetan Plateau and evaluate its differences from conventional earthquake catalogs.

Methods We constructed a catalog of source parameters for 42 moderate-to-strong shallow earthquakes (Mw 4.8-6.6) on the Tibetan Plateau from October 2014 to October 2022. We first assessed the quality of coseismic Sentinel-1 interferograms available from the Alaska Satellite Facility (ASF) platform. Depending on interferogram quality, we applied two processing strategies.For high-quality events, we utilize generic atmospheric correction online service for InSAR (GACOS) assisted differential InSAR (D-InSAR) approach. For low-quality events, we adopted a phase-stacking strategy. These procedures yielded coseismic deformation fields for all 42 earthquakes, which were then used to invert earthquake source parameters within a Bayesian framework.

Results The inverted events comprise 13 thrust, 10 normal, and 19 strike earthquakes. Overall, the model fits are good, with root mean square (RMS) misfits below 2 cm for the majority of events. Parameter reliability and uncertainty were further evaluated through comparisons with published studies and by examining the posterior credible intervals from the Bayesian inversions. Compare to conventional earthquake catalogs, the InSAR-derived catalog provides more accurate centroid depths and horizontal locations. Depth discrepancies reported in conventional catalogs can reach 4-11 km. Among the existing catalogs, the China Earthquake Networks Center (CENC) solutions show higher horizontal location accuracy than United States Geological Survey (USGS) and Global Centroid Moment Tensor (GCMT), with a mean offset of 7.4 km, whereas the GCMT locations exhibit a 11.9 km systematic bias and a distinct spatial pattern that correlates with topography, trending from high-elevation regions toward lower elevations. In contrast, fault-geometry parameters reported by conventional catalogs are broadly consistent with the InSAR solutions. The absolute differences are 10°-11° in dip and 9°-10° in strike. This study further integrates the InSAR-based catalog with previously published source datasets to derive empirical scale relationships for earthquakes over a broader magnitude range (Mw 4.8-8.03) on the Tibetan Plateau.

Conclusions The study shows that the InSAR-derived catalog significantly supplements the existing global catalog and can improve empirical scale relationships, facilitating scientific reference for regional seismic risk assessment.