The ambiguous fault geometry derived from InSAR measurements of buried thrust earthquakes: a synthetic data based study

Abstract

SUMMARY The challenge of ruling out potential rupture nodal planes with opposite dip orientations during interferometric synthetic aperture radar (InSAR)-based kinematic inversions has been widely reported. Typically, slip on two or more different fault planes can match the surface deformation measurements equally well. The ambiguous choice of the nodal plane for the InSAR-based models was thought to be caused by InSAR's 1-D measurement and polar orbiting direction, leading to its poor sensitivity to north–south crustal motion. Through synthetic experiments and simulations, this paper quantitatively demonstrates the main reason of the ambiguous InSAR-based models, which confuse researchers in the small-to-moderate thrust earthquake cases investigation. We propose the inherent 1-D measurement is not the principle cause of the fault plane ambiguity, since models derived from the same InSAR data predict similar, but not identical, 3-D deformation patterns. They key to differentiating between these different models is to be able to resolve the small asymmetry in the surface deformation pattern, which may be smaller in amplitude than the typical noise levels in InSAR measurements. We investigate the fault geometry resolvability when using InSAR data with different noise levels through ‘R’ value. We find that the resolvability does not only rely on the InSAR noise, but also on the fault geometry itself (i.e. depth, dips angle and strike). Our result shows that it is impossible to uniquely determine the dip orientation of thrust earthquakes with Mw < 6.0 and depth > 5.0 km with InSAR data at a noise level that is typical for mountain belts. This inference is independent from the specific data set (i.e. interferogram or time-series) and allows one to assess if one can expect to be able to resolve the correct fault plane at all.