A new method for calculating seismic velocities in rocks containing strongly dimensionally anisotropic mineral grains and its application to antigorite-bearing serpentinite mylonites

Seismic velocity is one of the most important sources of information about the Earth's interior. For its proper interpretation, we must have a thorough understanding of the dependence of seismic velocity on microstructural elements, including the modal composition, the crystal preferred orientation (CPO), the grain shape, the spatial distribution of mineral phases, etc. The conventional Voigt, Reuss and Hill averaging schemes take into account only the modal composition and the CPO. The information about the Earth's interior is thus poorly constrained. For a better interpretation, it is critical to have a calculation method which accounts for the grain shape and the spatial distribution of mineral phases, etc. We propose a calculation method which accounts for the grain shape of strongly dimensionally anisotropic minerals like micas and serpentines. Our method can be applied to a distributed geometrical orientation of mineral grains. Comparison was made between calculated and measured velocities in three antigorite–serpentinite mylonites. Judging from the root mean square relative error, our method provides velocities closer to measured values than the Voigt, Reuss and Hill averaging schemes. The input of the grain shape considerably improves the prediction of seismic properties. However, large discrepancies (>0.1 km/s) between measured and calculated velocities can be seen in some directions. The discrepancies might come from microstructural elements which were not considered in the calculation (layer structures and cracks). •We propose a new method for calculating seismic properties of rocks.•Our method takes into account the grain shape of dimensionally anisotropic minerals.•Our method can be applied to a distributed geometrical orientation of mineral grains.•Comparison was made between measured and calculated velocities in serpentinites.•The input of grain shape considerably improves the prediction of seismic properties.