Characterizing Lithospheric Transverse Isotropy Using Moment Tensors—On the Origin of the Non–Double-Couple Components of Global CMT Solutions

ABSTRACT The seismic moment tensor, which represents the equivalent body-force system of the seismic source, may exhibit non–double-couple components (NDCs) when the earthquake occurs on a planer fault if the source medium is anisotropic. Kawakatsu (1991a) reported that the NDCs of the moment tensors (MTs) for shallow earthquakes from the Harvard Centroid Moment Tensor (HCMT) catalog exhibited a systematic characteristic dependent on faulting types: the sign of the NDC parameter ϵ on average systematically changes for thrust fault (TF) (negative), normal fault (NF) (positive), and strike-slip (SS) (positive) fault. The behavior of TF and NF that shows opposite sign can be explained if the source region is transversely isotropic with a vertical symmetry axis (VTI, radially anisotropic). In fact, the VTI model of the Preliminary Reference Earth Model (PREM) at a sub-Moho depth predicts the observed systematic NDC pattern, although the magnitude is underestimated, indicating the potential to enhance our understanding of the lithospheric transverse isotropy using the NDC of the MTs. To investigate the lithospheric transverse isotropy structure using the NDCs of the MTs, we propose a novel inversion scheme, building on the approaches used previously for deep and intermediate-depth earthquakes but with necessary modifications to address shallow sources. Synthetic tests conducted under conditions of random faulting indicate the potential to constrain the S-wave anisotropy ξ and the fifth parameter ηκ, but with moderate to severe correlation depending on data types. The application of this method to real data sourced from the Global Centroid Moment Tensor catalog suggests that the lithospheric transverse isotropy of PREM at the sub-Moho depth serves as a suitable initial model. However, some adjustments may be necessary, particularly regarding the fifth parameter, to enhance the model’s fidelity in representing observed NDCs of the MTs. Finally, the behavior of the SS faults that cannot be explained by the VTI may deserve further attention.