Resolution of isotropic percentage in moment tensor inversion of tensile sources

Abstract

Moment tensor solutions are commonly used in order to understand earthquake source mechanism. Moment tensor can be decomposed into three components, namely isotropic (ISO), double-couple (DC) and compensated linear vector dipole (CLVD). It is well-known fact that tensile sources generate non-DC components and those earth quakes can be defined as the combination of both tensile and shear motion on a fault. In this thesis, resolution of the isotropic part in the moment tensor are considered for tensile sources. For that reason, synthetic waveforms are created by using full moment tensors with different isotropic percentages and those waveforms are inverted with gCAP method. Afterwards, a range of different isotropic values, with a step of t = 0.1, are forced in the moment tensor inversion process in order to investigate the change in variance reduction as the isotropic percentage deviate from its true value. Inversions of the full waveform are performed in different distances and depths for three moment tensors with different isotropic percentages, namely 2%, 5% and 14%. Inver sions results of those original moment tensors and moment tensors with manipulated isotropic percentages are expressed. Those results are compared to each other in terms of changing isotropic percentages, depth and variance reduction in different stations. The results can be summarized as firstly, inversion is not really sensitive to the isotropic component of the moment tensor because isotropic component has small en ergy compared to the whole waveform. Secondly, earthquakes with relatively high isotropic percentages are less sensitive when inversions are performed for high values of manipulated isotropy. Finally, it is observed that the error in the depth of the earthquake is very sensitive to isotropic percentage.