Inference of viscosity jump at 670 km depth and lower mantle viscosity structure from GIA observations

Masao Nakada, Jun'ichi Okuno, Yoshiya Irie

Research output: Contribution to journalArticlepeer-review

6 Citations (Scopus)


A viscosity model with an exponential profile described by temperature (T) and pressure (P) distributions and constant activation energy (Eum* for the upper mantle and Elm* for the lower mantle) and volume (Vum* and Vlm*) is employed in inferring the viscosity structure of the Earth's mantle from observations of glacial isostatic adjustment (GIA). We first construct standard viscosity models with an average upper-mantle viscosity (ηum) of 2 × 1020 Pa s, a typical value for the oceanic upper-mantle viscosity, satisfying the observationally derived three GIA-related observables, GIA-induced rate of change of the degree-two zonal harmonic of the geopotential, J2, and differential relative sea level (RSL) changes for the Last Glacial Maximum sea levels at Barbados and Bonaparte Gulf in Australia and for RSL changes at 6 kyr BP for Karumba and Halifax Bay in Australia. Standard viscosity models inferred from three GIA-related observables are characterized by a viscosity of ~1023 Pa s in the deep mantle for an assumed viscosity at 670 km depth, ηlm(670), of (1-50) × 1021 Pa s. Postglacial RSL changes at Southport, Bermuda and Everglades in the intermediate region of the North American ice sheet, largely dependent on its gross melting history, have a crucial potential for inference of a viscosity jump at 670 km depth. The analyses of these RSL changes based on the viscosity models with ηum ≥ 2 × 1020 Pa s and lower-mantle viscosity structures for the standard models yield permissible ηum and ηlm (670) values, although there is a trade-offbetween the viscosity and ice history models. Our preferred ηum and ηlm (670) values are ~(7-9) × 1020 and ~1022 Pa s, respectively, and the ηum is higher than that for the typical value of oceanic upper mantle, which may reflect a moderate laterally heterogeneous uppermantle viscosity. The mantle viscosity structure adopted in this study depends on temperature distribution and activation energy and volume, and it is difficult to discuss the impact of each quantity on the inferred lower-mantle viscosity model. We conclude that models of smooth depth variation in the lower-mantle viscosity following η(z) α exp[(Elm* + P(z)Vlm*)/RT (z)] with constant Elm* and Vlm* are consistent with the GIA observations.

Original languageEnglish
Pages (from-to)2206-2225
Number of pages20
JournalGeophysical Journal International
Issue number3
Publication statusPublished - Mar 1 2018

All Science Journal Classification (ASJC) codes

  • Geophysics
  • Geochemistry and Petrology


Dive into the research topics of 'Inference of viscosity jump at 670 km depth and lower mantle viscosity structure from GIA observations'. Together they form a unique fingerprint.

Cite this