 |
| Titel |
The thermo-chemical structure of the deep mantle: observations and models |
| VerfasserIn |
Frédéric Deschamps, Jeannot Trampert, Paul J. Tackley, Joe Resovsky |
| Konferenz |
EGU General Assembly 2010
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 12 (2010) |
| Datensatznummer |
250036059
|
|
|
|
|
|
| Zusammenfassung |
| Our description of the Earth’s mantle structure has considerably changed during
the past decade. The hypothesis that large chemical heterogeneities (whose nature
and origin is still to be determined) gained increasing support, in particular from
normal mode seismology observations (including probabilistic tomography). The key
point in these studies is that the mapped density anomalies and shear-wave velocity
anomalies are not correlated, which is inconsistent with a purely thermal origin of
the observed seismic velocity anomalies. Interestingly, other studies, using other
observations and arguments (e.g., the seismic ratio between the relative anomalies of
compressional and shear wave velocities) arrive to similar conclusions. The classical
interpretation that the large low shear-wave velocity provinces seen in the deep
mantle (LLVSZ) are hot and buoyant must be revisited. Instead, LLVSZ are likely
the result of superposed thermal and chemical heterogeneities. Using a careful
equation of state modeling, it is possible to estimate the distributions of temperature
and composition (e.g., the volume fractions of perovskite and iron oxide) from the
distributions of seismic velocities and density. This operation also requires an appropriate
treatment of the various sources of uncertainties in the thermo-elastic parameters of the
Earth’s mantle minerals through Monte-Carlo searches. Sensitivities of seismic
velocity and density to temperature and composition indicate that in the deep mantle
shear-wave velocity is sensitive to both the temperature and the volume fraction of
iron. Furthermore, density is mainly sensitive to variations in the volume fraction
of iron oxide, and bulk velocity anomalies are a good proxy for variations in the
volume fraction of perovskite. Additional data and observations independent from
seismological are needed to further break the trade-off between temperature and
composition. A promising constraint is electrical conductivity. Unlike thermo-elastic data
and density, electrical conductivity increases with increasing temperature. Using
available mineral physics data, the thermo-chemical structure predicted by probabilistic
tomography results in an equatorial belt of high electrical conductivity in the deep
mantle.
The thermo-chemical structure and dynamics of the mantle are intimately linked, and a
successful model of mantle convection must explain seismological observations. A
key question is therefore to build models of mantle dynamics that can maintain
large pools of dense material at the bottom of a convective layer for a long period
of time. So far, two independent sources of chemical heterogeneities have been
investigated by models of thermo-chemical convection: the interaction of primitive
reservoir(s) of dense material with mantle convection, and the production and recycling of
MORBs. Models with an initial basal layer of dense material indicate that strong
thermal viscosity contrasts are able to create and maintain large pools of dense
material at the bottom of the system, and that an endothermic phase transition at
660-km depth prevents the dense material to massively flow into the upper part
of the system. Models that include MORBs recycling indicate that the formation
and survival of pools of dense material at the bottom of the mantle is sensitive to
the buoyancy ratio of MORBs, and to the properties of the post-perovskite phase
transition. Individually, primitive reservoirs and MORBs do not fully explain the
available seismological observations. It is likely that the chemical contribution to
the observed seismic velocity and density anomalies originate from two or more
sources. |
|
|
|
|
|
|