Modeling of earthquake source parameters, such as the orientation of the fault plane and the direction of the fault slip, is important for understanding the physics of earthquake source processes, determining the stress-strain state of the geological medium and seismic hazard estimation. For modeling source parameters of the earthquake on December 12, 2018 at 08:49:56,16 (UTC) in Japan (36,4478° N, 140,5788° E, Northern Ibaraki Pref region) at a depth of 62 km with a magnitude of Mw = 4.3, the waveforms inversion was used to determine seismic moment tensor and representation it through a focal mechanism. The earthquake source is considered as a point source of seismic waves which propagate in a medium represented by a set of horizontally homogeneous elastic layers. An algorithm for determining seismic tensor components based on the forward problem solved by the matrix method, and using the generalized inverse solution, selecting only direct waves is applied. The input data for determining seismic moment components are data of only direct P waves selected from the observed records at six seismic stations of the Japanese local network NIED F-net: TSK, YMZ, ASI, ONS, SBT, KSK. The seismic moment tensor components were determined through waveform inversion using the matrix method. The obtained results, presented through a focal mechanism, are compared to the results obtained by the National Research Institute of Earth Sciences and Resistance to Natural Disasters (NIED CMT solutions). As a result of focal mechanisms comparison, it is concluded that the proposed algorithm for determining seismic moment tensor components can be used if it is impossible to use another method, or requires some refinement for another method. This approach is especially relevant for regions with low seismicity and insufficient number of stations. In addition, this method reduces the effects of an inaccurate medium model, because direct waves are much less distorted than reflected and converted, and that increases the accuracy and reliability of the method.
Read full abstract