Joint regional waveform and surface displacement inversion for the seismic moment tensor: Application to September 3, 2017 declared North Korean nuclear test

Abstract:

On 3 September 2017, the Democratic People’s Republic of Korea conducted its largest nuclear test (DPRK2017). This event is the largest man-made explosion ever recorded by orbital radar systems and provides a unique dataset with which to develop new methods of source inversion. Even though regional distance seismic MT inversion methods have identified this event as an explosion (Chiang et al., 2018) , we have found that the best surface wave and first motion solution fails to fit the static deformation field from the SAR data (e.g. Dreger et al., 2017). However we have found that a subspace of best fitting models can satisfy the surface deformation field and reduce the overall uncertainty in the moment tensor solution. To develop a joint inversion method we use the SW4 finite-difference code that utilizes a curvilinear mesh to accurately model surface topography effects on the seismic wave field to develop static displacement Green’s functions for the geodetic data inversion. We have generated Green’s functions for a source volume comprised of XYZ sources. We perform a grid search of moment tensor solutions over the source volume to find the best fitting location, depth and moment tensor for the event. Our initial results show that we can explain 65% horizontal static ground deformation using a composition of 3 difference sources that accounts for the finite extent of the explosive source. We will present static field only, joint static, regional waveform and first-motion point-source solutions, and multiple point-source solutions to account for source finiteness.