Constraining Convection:

1
Constraining Convection

• Electrical Conductivity and the Geomagnetic
  Response Function.

2
Earth is a heat-engine
HEAT SOURCE Gravitational Chemical Radiogenic
ENGINE Core Convection Mantle Convection
EXPRESSION Magnetic Field Volcanism Seismicity Co
ntinental Drift
www.keveney.com
3
Seismic Wave Propagation
http//epsc.wustl.edu/seismology/michael/michael.h
tml
4
Seismic tomography
http//eri-ndc.eri.u-tokyo.ac.jp/jp/ohrc/ken1/top1
-img.jpg
5
(No Transcript)
6
Geochemistry
http//pmlgw.misasa.okayama-u.ac.jp
http//www.geo.vuw.ac.nz
7
Layered convection?
http//rmocfis.uprm.edu
8
Convection models
9
Geomagnetism Secular variations
10
Response function Al
Observed

11
Previous work Earth profileXu et al., 2000
12
Mineral physics and petrology
www.es.ucl.ac.uk
13
EC Cell (18/8)
14
Electrical conductivity

• ss0e-(DH/RT)
• DH EPDV

15
Sample data (Mg,Fe)O
16
Sample Data PerovskiteXu et al., 2000
17
How Does Conduction Occur?

• snem
• nnumber of charge carriers
• echarge
• mmobility

18
Previous work Earth profileXu et al., 2000
19
Global Induction Forward Model Constable 1993

• Data
• Globally averaged long period (3 days 3
  months)
• Stacked data from 38 stations
• Model
• Spherical shell model
• Finite shell thickness
• Finite conductivity
• (depends on -mineralogy
• -chemistry
• -temperature)

20
Geotherms
21
Earth Conductivity Profile
22
Response function Al
Observed

23
Response function No Al

24
Response function Al

25
Best fit response functions

26
Summary 1

• The global response function is consistent with
  either
• A cool lower mantle with aluminous perovskite or
• A hot lower mantle with no Al in perovskite.
• Layered convection models result in a lower
  mantle with more Fe and Al.
• Interconnected magnesiowustite in the lower
  mantle would favour cool geotherms
• Convection is most likely whole-mantle to at
  least
• 1800 km

27
Bercovici and Karato model
28

• Transition zone minerals contain more water than
  olivine saturation.
• Upwelling mantle dehydrates at 410 km producing
  small volumes of hydrous melt
• The wet melt density keeps it at 410 km
• Normal upper mantle has been through this
  filter and shows depleted trace
  element/isotopic ratios
• Plumes come up fast enough to entrain the melt
  and so show undepleted ratios.

29
Bercovici and Karato model
30

• Subduction pumps material to CMB
• Mass balance makes mantle dilate
• Subduction rate plate velocity 10cm/year
• Dilation rate (from mass balance) 10cm/yeararea
  of slab/area of 410 km 10106/(4.5108)
  210-2cm/year

31

• Fast subduction slow dilation
• Subduction drags down melt (percolation too slow)
• Dilation leaves melt behind (percolation faster
  than dilation rate)
• Water is recycled into deep mantle via subduction
  and so conserved

32

• Plume rates comparable to subduction rates
• Melt is generated in plumes but percolation is
  slow
• melt within the plume is dragged up with the
  plume, maintaining the undepleted geochemistry.

33
Assumptions???