Upper mantle Q structure beneath old seafloor in the western Pacific

Abstract

The Pacific Lithosphere Anisotropy and Thickness Experiment ocean bottom seismometer array south of the Shatsky Rise allows imaging the attenuation structure of the upper mantle in the oldest parts of the Pacific Ocean. The array consisted of eight seismometers with a lateral extent of 200 km by 600 km. Intermediate-and deep-focus earthquake sources in the Mariana and Izu-Bonin subduction zones provide paths that probe different depths beneath the seafloor. Acceleration spectra from the vertical component of P waves are inverted for the attenuation operator and corner frequency. P n propagation indicates very high Q p values in the lithosphere, of order 1000, while path-average Q p values in the upper mantle are on the order of 400. The variety of source epicentral distances and depths enables us to deduce the vertical Q p À1 structure of the upper mantle. Using the amplitude spectra directly in a weighted least squares problem that accounts for the effects of triplications in the traveltime curves, we invert for Q p À1 as a function of depth. We demonstrate that Q p is frequency dependent and adopt a power law dependence with coefficient α = 0.27. We resolve three distinct layers using a reference frequency of 1 Hz with Q p on the order of 145 at depths 100-250 km, Q p on the order of 350 at depths 250-410 km, and very high Q p with attenuation indistinguishable from zero for depths below 410 km. The P waves have a component of scattering, and thus, Q p À1 represents the maximum intrinsic attenuation beneath normal undisturbed oceanic lithosphere.

Reconciling mantle attenuation-temperature relationships from seismology, petrology, and laboratory measurements

“…Booth et al . [] examine regional P waves recorded on the western Pacific seafloor from Mariana slab earthquakes, with spectral fitting and source correction methods analogous to those applied here. However, the Booth et al .…”

“…However, the Booth et al . [] paths are longer than in the arc/back‐arc regions discussed above, and their Q estimates involve nonlinear corrections to triplicated arrivals in the transition zone. These corrections and parameterization result in large adjustments of Q P from path‐average values near 400 to inferred asthenospheric minimum values of 128–165 at 100–250 km depth, beneath a very low attenuation lithospheric layer assumed to be 100 km thick.…”

“…Taken at face value, the Booth et al . [] results imply a Q S of 57–73 at 1 Hz, assuming Q P / Q S of 2.25, or 75–97 for a Q P / Q S of 1.7. Owing to the pressure dependence of attenuation, this Q S range would decrease an additional ∼30% if it was extrapolated to the 50–80 km depth ranges where our subduction zone Q S values are constrained (Table ).…”

“…This Q P / Q S suggests the presence of bulk attenuation (possibly due to melt), a source of attenuation that could be absent in the suboceanic asthenosphere sampled in Booth et al . []. Taken at face value, the Booth et al .…”

“…Given these assumptions, the Booth et al . [] Q S estimates for old oceanic mantle fall within the range of minimum subduction zone attenuation we infer for Central America, the Marianas, and the Lau Basin.…”