A family of 37 most reliable moment tensors of microearthquakes that occurred at a depth level of 5.4 km, having been induced during the 2000 injection experiment at the KTB deep drilling borehole, contain significant non-double-couple (non-DC) components. The DC is on average 60% and the non-DC is 40%. Fault plane solutions computed from the DC part show preferred strike-slip mechanisms with small normal or reverse components. A predominant azimuth of P and T axes is in the range of N140°-160°E and of N50°-70°E, respectively. The optimum principal stress directions, inferred from the focal mechanisms, are (azimuth/plunge): σ1=335°/15°, σ2=110°/70°, σ3=240°/15°, and the shape ratio is 0.55. The errors in the plunge and azimuth are about 10°. The non-DC components contain both the isotropic (ISO) and compensated linear dipole (CLVD) components. The mean value of ISO is 1.5%, the mean value of CLVD is -5.7%. The predominantly negative CLVD components are inconsistent with the concept of the non-DC mechanisms as a result of tensile faulting due to fluid injection into the rock. The non-DC components have probably three other major origins: random numerical errors produced by the moment tensor inversion due to noise and limitations of input data, systematic errors produced by mismodeling of the medium when calculating the Green functions, and anisotropy in the focal area. Anisotropy is a very likely origin of non-DC components, because a rather 247 strong seismic anisotropy has been observed at the KTB site with strength up to 20% for P waves, and attributed to foliated crystalline rocks composed of gneiss and amphibolite. Adopting four alternative models of anisotropy obtained by other seismic measurements at the KTB, we have employed the non-DC components for estimating an optimum orientation of anisotropy in the focal area. The optimum orientation of the symmetry plane of anisotropy is nearly vertical with a strike of N335-340°E. This strike coincides well with the strike of 330° typical for many major lithological units and faults and with the orientation of the transversely isotropic model inferred by other authors. A successful inversion for anisotropy orientation confirms that a significant part of the non-DC components of the moment tensors under study is produced by anisotropy, and that the moment tensors can be used to estimate the anisotropy.
The image of the paper in GIF 150dpi (1741 kB !) is designed for an instant screen preview.
The paper is available in PostScript (8843 kB !) and GZIPped PostScript (1153 kB !).