8:30 AM - 8:45 AM
[S13-1-01] Challenges in moment tensor resolution: collapses, explosions and shallow earthquakes
Regional moment tensor inversion is extensively used to model the earthquake rupture geometry and estimate earthquake source parameters. Typically, inversion algorithms derive the moment tensor configuration by fitting low frequency full waveforms or amplitude spectra. Resulting moment tensor catalogues provide a valuable information on seismic source processes. They can be used for seismotectonic interpretation, to discuss non shear ruptures or as a base to investigate the stress orientation. The resolution of the moment tensor can be reduced by several causes, such as waveform mismodeling due to a poor knowledge of the crustal structure, low signal-to-noise ration due to seismic noise and poor azimuthal coverage due to an unfavourable network geometry. Poor conditions may bias moment tensor solutions, for example introducing non double couple terms. Beside these known problems, the procedures of moment tensor inversion, decomposition and interpretation are also affected by intrinsic ambiguities, which can become critical for shallow sources or specific source-network geometries. First, different uncertainties affect different moment tensor entries, resulting in an uneven resolution of moment tensor components. In addition, moment tensor components trade-offs affect the uniqueness of the solution, so that different moment tensor configurations may fit well the data. Finally, the interpretation of the moment tensor solution may be hindered by the chosen decomposition. We discuss and investigate some of these often neglected problems and provide example for each of them, combining synthetic tests with the discussion of significant real data applications. Among the real data cases, the analysis of seismic signals produced by recent volcanic collapse and shallow underground explosions in North Korea helps to illustrate the problem of moment tensor uncertainties, trade-off and decomposition.