Frequency-Difference Back-projection of Earthquakes

Abstract:

Back-projection has proven useful to image large earthquake rupture processes. It utilizes array techniques to estimate the spatial and temporal evolution of earthquake rupture over time, and can help us identify interesting earthquake phenomena like supershear rupture. However, the method does not directly solve an inverse problem and has difficulty in quantifying epistemic uncertainties, which can be caused by seismic array configurations, structural heterogeneities in the Earth’s crust, unknown seismic phases, and variations in the focal mechanism. These uncertainties may cause erroneous interpretations of earthquake physics, which is particularly challenging to distinguish for complex earthquake rupture processes.

In this study, we develop and apply a new frequency-difference back-projection method to image earthquake rupture processes. The method is developed from a frequency-difference beamforming approach, which was originally designed to locate acoustic sources. Frequency-difference beamforming utilizes frequencies below the bandwidth of the signal, which are less affected by multipathing and structural inhomogeneities. This method can potentially allow us to locate sources more accurately even in the presence of strong scattering, albeit with lower resolution. We will first verify robustness of the method through both synthetic simulations and analysis of M6 earthquakes. We will then apply the method to the 2015 M7.8 Nepal earthquake to unveil its early rupture evolution. Our research aims to address how Nepal earthquake rupture evolved during the first 20s and overcame a geometric fault barrier, which is challenging to resolve because of the difficulties in imaging early rupture propagation process using other kinds of back-projection approaches.