The lithosphere-asthenosphere boundary: Clues from joint interpretation of surface-wave velocity and attenuation models

Combined interpretation of seismic velocity and attenuation models can provide a powerful tool to investigate the physical state of the upper mantle. In this study, we consider the global shear velocity model of Kustowski et al. (2006) and the global shear attenuation model of Dalton et al. (2007). The velocity model contains 3-D variations in radially anisotropic shear-wave speed and is constructed from large data sets of surface-wave phase anomalies, long-period waveforms, and body-wave travel times. The attenuation model is derived from > 30,000 fundamental-mode Rayleigh wave amplitude measurements at each period in the period range from 50-250 s. The amplitudes are inverted simultaneously for the coefficients of the 3-D attenuation model as well as frequency-dependent correction factors for each source and receiver. Focusing effects due to elastic heterogeneities are accounted for from jointly determined phase velocity maps.
In plots of attenuation (1/Q) vs velocity at fixed depths the model data form distinct trends. We compare these seismologically determined trends to the fit of experimental shear modulus and attenuation data of Faul and Jackson (2005). At a depth of 100 km the seismological data cuts obliquely across the experimental fit. When the seismological model data are separated into continental and oceanic regions, at depths of 150 and 200 km the oceanic data form a trend that is close to the experimentally predicted one, while the continental trend is still oblique to the experimental trend. At 250 km depth both trends are similar to the experimentally derived trend. At those depths where the seismological trends are oblique to the experimental relationship, the oceanic velocities generally are slower for a given attenuation value and the continental velocities generally are faster for a given attenuation value relative to the experimental relationship. Since the experimental relationship describes the anelastic behavior of olivine at temperatures above ~1000°C, the oblique trends of the seismic data relative to the experimental fit imply non-thermal contributions to velocity and/or attenuation, for example variations in composition (major and trace elements) or melt. At the depths where the seismic trends coincide with the experimental fit, the range in velocity and attenuation can be ascribed mostly to lateral temperature variations. It seems plausible, therefore, that the depth at which the change in slope occurs indicates the average depth of the lithosphere-asthenosphere boundary.