Modeling Rayleigh Wave Azimuthal Anisotropy in the Upper Mantle Beneath Hawaii

We present results from the forward modeling of frequency-dependent Rayleigh wave azimuthal anisotropy observed at the Hawaiian Plume-Lithosphere Undersea Mantle Experiment (PLUME) that was deployed from January 2005 through May 2007. We create models of the mantle consisting of one or two layers of constant anisotropy derived from pyrolite, which has one symmetry axis, in order to understand how mantle flow affects observations of Rayleigh wave azimuthal anisotropy. We perform forward modeling experiments in which we vary the depth and thickness of the layer(s), as well as the inclination and azimuth of the symmetry axis in each of the layers. We use mode summation code MINOS to calculate predictions for these models. Then we compare the predictions with frequency-dependent average phase velocities within a series of station triangles. We use a chi-squared test for frequencies between 10 and 50 mHz to model the two principal parameters of azimuthal anisotropy: the frequency-dependent strength of azimuthal anisotropy and the frequency-dependent direction of fast phase velocity. For the former, we fix the azimuth in the synthetic model and compare the strength of anisotropy. For one-layer models, we find that most station triangles are best fit by fixing the inclination between 20 and 35° from the vertical, suggesting that a significant amount of vertical mineral alignment is present. Almost all station triangles require an anisotropic layer beginning no deeper than 50 km below the surface, and a thinner layer is present beneath the Island of Hawaii and to the east of the islands. A thicker layer is present beneath and to the west of the older islands. First results from two-layer modeling suggest that a layer in the asthenosphere must consist of near horizontal flow while a layer in the lithosphere must contain a significant amount of vertical anisotropy.