‘Observations’, Geotherms and Adiabats
• What do observations tell us
about whole vs layered mantle
• What is the temperature profile
in the Mantle? Whole vs Layered Convection
• We have seen from numerical models that depending on the viscosity profile (which
depends on pressure, temperature and chemistry), the mantle may either be convecting
as a single layer or as two separate layers
• Are there any observations we can make that determine which it is?
Main argument for layered convection:
• The chemistry of the upper mantle is not “chondritic”. Lava flows from mid ocean
ridges (which sample the upper mantle) seem to be missing some fraction of
elements and isotopes we would expect if the Earth formed from the solar nebula.
People have therefore suggested that the Earth has a hidden reservoir of these
elements. Because of the phase boundary at 670km depth, the layer below
(the lower mantle) would be an ideal location for this hidden reservoir.
• Hot Spot lava also has a different chemistry than mid ocean ridge lava. Since the
source of hot spots is believed to be plumes that come from deeper in the mantle,
this also suggests that the mantle may have hidden chemical reservoirs.
Main argument for whole mantle convection:
• Its simpler and the chemistry issues may be solvable other ways
• Other elements are proper representations of “chondritic” abundances
• D” layer at base of mantle believed to be a “slab graveyard”
• Seismic tomography: Seismic Tomography
• Because seismology can be used to determine wave speeds and hence density and
temperatures in the mantle, we can, in theory, track subducting slabs or hot spots and see how
far down they go
• Geophysical analog of a CAT scan
• Remember slabs are a part of mantle convection: the cool upper boundary layer that sinks
into the mantle. If the slabs stop at 670km, then that suggests mantle convection is
layered, if they penetrate 670km and go to the core-mantle boundary, then that suggests
whole mantle convection
• Hot spots are also part of mantle convection: the hot lower boundary layer that rises to
the surface. If the lower boundary layer is the 670km discontinuity, then that suggests
mantle convection is layered, if it’s the core mantle boundary, then that suggests whole
Whole Layered Subducting Slabs:
Japan Central America
Indonesia Izu Bonin r
… neither strong layering nor simple whole mantle convection Hawaii
How about Hot Spots?
More difficult to image but with
better data, better coverage, and
better inversion theory the next
decade may show major progress …
(Montelli et al., GJI, Science, 2004) Mantle Geotherm
• According to our mantle convection discussions, if the mantle convects vigorously, then
most of the interior should be “isothermal” (at constant temperature).
• Is this the case in the Earth? NO! The mantle convection models we discussed did not
include one important effect: COMPRESSION.
• All the numerical models (and experiments) involve “incompressible” fluids. That is,
the density of the fluid does not change with pressure. The mantle however, is
compressible, so we do need to take this into account.
• In the interior of a vigorously convecting fluid which is COMPRESSIBLE, the mean
temperature increases with depth approximately along the ADIABAT.
• The adiabatic temperature gradient in the mantle is the rate of increase of temperature
with depth as the result of compression of the rock by the weight of the overlying
4 equal mass
elements at T T4>T 3T >2 1