Seismic attenuation structure of the Pacific upper mantle in the NoMelt experiment region as constrained by Rayleigh waves from earthquakes

The NoMelt experiment deployed an ocean bottom seismograph and magneto-telluric arrays in the central Pacific Ocean on 70 Ma seafloor. It is designed to study the evolution of lithosphere and asthenosphere in the "normal" portion of the Pacific Ocean.

We use the long period (10mHz-30mHz) Rayleigh waves recorded by this array to derive the mantle attenuation structure. After the removal of tilt and compliance noise, we were able to measure Rayleigh wave phase and amplitude from over 100 earthquakes at the frequency range of interest. The compliance correction for ocean wave pressure on the seafloor is particularly important at frequencies lower than 17 mHz.

The pattern of the measured Rayleigh wave amplitudes from many earthquakes show a strong azimuthal variation. This pattern can be well modeled by the interference effect from two plane waves entering the array at different arrival angles. An inversion for 2D phase velocity perturbations, average azimuthal anisotropy, and the average attenuation within the array was performed by accounting for the interference effect. We have inverted for the average Rayleigh wave attenuation at 10, 15, 20, 25 and 30mHz. The average attenuation values at all five frequencies are then used to invert for the intrinsic shear attenuation structure with depth. Our preliminary result shows that only two layers of Q are required by the data, a shallow one with low attenuation (Q>300) and a deeper one with high attenuation (Q 100). The transition between the layers happens in the depth range 50-100km. A more quantitative inversion result and error analysis will be shown in the meeting. In conjunction with previous estimates of shear velocity and electrical conductivity, we interpret the observed attenuation structure in terms of the temperature structure and volatile content of the upper mantle in the NoMelt region.