Abstract

We present a new methodology for constraining mantle composition directly from observed seismic travel times. By measuring travel times of PcP, PcS, ScS, and PKiKP phases generated by the Chinese nuclear explosion of 21 May 1992 and recorded at epicentral distances of ∼10°, we directly determine bulk sound travel times (TΦ' for the mantle and outer core. This eliminates the need to combine independent VP and VS profiles, characterizing different paths and frequencies, to obtain a reference VΦ profile. Our observed PKiKP‐PcP differential travel time indicates that the outer core model of AK135 is slow by 2.1 s (that of IASP91 by 2.3 s' while that of PREM is slow by only 0.7 s and suggests, through comparison with other data, relative lateral homogeneity of the outer core. Furthermore, we find that T0Φ for the crust and mantle in this region exceeds the AK135 prediction by 3.5 s (IASP91 by 3.9 s' ±1.0 s. Since the lower mantle contribution to this delay depends upon how much arises in the upper mantle, we present two end‐member models for such a slower mantle VΦ, in which the delay resides entirely above or below 660 km depth. The latter yields best‐fitting lower mantle compositions which are less enriched in Si than those fit to T0Φ for AK135 (or IASP91' for the former, and neither supports Feenrichment of the lower mantle. Applicability of such localized observations to the larger mantle is suggested by tomographic results indicating that lateral heterogeneity in VΦ is very small relative to that in VS. Our results illustrate the potential utility of T0Φ in constraining mantle composition.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call