Small-scale mineralogical heterogeneity from variations in phase assemblages in the transition zone and D" layer predicted by convection modelling

Arie P. van den Berg, David A. Yuen, Michael H.G. Jacobs, Maarten V. de Hoop

Research output: Contribution to journalArticlepeer-review

1 Scopus citations

Abstract

Small-scale heterogeneity in the deep mantle is concentrated in the upper-mantle transition zone (TZ), in the depth range 410-660 km and also at the bottom 250 km D" region. This encourages a more detailed investigation of the potential for seismic reflectivity imaging by modelling heterogeneous structures in mantle convection models including phase transitions of the TZ and D" regions. We applied finite elements with variable spacing near the boundary layers in 2-D cylindrical geometry that allow for sufficient spatial resolution. We investigated several models including an extended Boussinesq (EBA) model, focused on the D" region, and a compressible (ALA) model for the TZ region. The results for the D" region show typical lens-shaped structures of post-perovskite (PPV) embedded in the perovskite (PV) background mantle, where the thickness of the lenses, at 200-400 km, strongly depends on the Clapeyron slope of the PV-PPV transition. A second phase transition (double crossing) occurs in case the core temperature is higher than the intercept temperature Ti. Our phase-dependent rheology results in contrasting effective viscosity between PV and PPV. Our model results reveal a distinctly clear mechanical weakening of the PPV lenses with about an order of magnitude lower viscosity. The shear wave-speed distributions computed from our convection results are strongly correlated with the heterogeneous distribution of the mineral phase. Gradients in the seismic wave-speed that are the target of seismological reflectivity imaging are clearly revealed. The wave-speed results show a clear resolution of the top and bottom interfaces of the PPV lenses. Our ALA model for the TZ is based on a thermodynamical model for the magnesium end-member of an olivine-pyroxene mantle. The model predicts a much more complex distribution of mineral phases, compared to our D" results, in agreement with the greater number of mineral phases involved in the olivine-pyroxene phase diagram for the P, T conditions of the transition zone. Near cold downwelling flows representing subducting lithospheric slabs, where the local geotherm can differ by up to 1 000 K from the horizontal average, and small-scale lateral variations in the mineral phases can occur.

Original languageEnglish (US)
Pages (from-to)160-168
Number of pages9
JournalJournal of Earth Science
Volume22
Issue number2
DOIs
StatePublished - Apr 2011

Bibliographical note

Funding Information:
This study was supported by the CMG Program of NSF, Senior Visiting Professorship by the Chinese Academy of Sciences, The Netherlands Research Center for Integrated Solid Earth Science (ISES 3.2.5), and the 216 through ISES Project ME-2.7. *Corresponding author: berg@geo.uu.nl © China University of Geosciences and Springer-Verlag Berlin Heidelberg 2011

Funding Information:
We acknowledge constructive reviews from two anonymous reviewers and stimulating discussion with Rob van der Hilst, Reinhard Boehler and Ladislav Hanyk. Rob van der Hilst is also thanked for making available unpublished results. David A Yuen thanks the CMG Program of NSF for support and for a Senior Visiting Professorship by the Chinese Academy of Sciences. Computational resources for this work were made available by The Netherlands Research Center for Integrated Solid Earth Science (ISES 3.2.5). Collaboration between Arie P van den Berg and Michael H G Jacobs has been supported by ISES Project ME-2.7.

Keywords

  • mantle convection
  • mineralogical heterogeneity
  • phase transition
  • post-perovskite

Fingerprint

Dive into the research topics of 'Small-scale mineralogical heterogeneity from variations in phase assemblages in the transition zone and D" layer predicted by convection modelling'. Together they form a unique fingerprint.

Cite this