SUMMARY OF MANTLE TEMPERATURES

1
SUMMARY OF MANTLE TEMPERATURES

• DON L. ANDERSON
• 2006

2
Bottom lines

• Geophysical global estimates of mantle
  temperature are slightly higher and have a larger
  range than petrological estimates from mature
  spreading ridges (away from plume influence)
• The conduction geotherm may extend deeper than
  the average depth of MORB extraction (280 vs
  100 km)
• The deeper geotherm is subadiabatic
• Middles of long-lived plates can be 30-50C
  hotter than at mature ridges new ridges can
  give hotter MORB

3
POTENTIAL TEMPERATURES (Tp)

• Global Geophysical Inversion heatflow, spreading
  rates
• 1410 180C Kaula, 1983 (entire range)
• Ridges
• melt petrology 137070C Asimow, 2006 (2
  sigma)
• peridotites 100C Bonatti et al., 1993
• subsidence rates 100C Perrot et al., 1998
• Kolbeinsey Ridge
• 1270-1360C Korenaga, Kodaira
• Lower mantle
• 1500-1730C Zhao, Anderson, Stacey, Stixrude
• Compare with McKenzie and Bickle 1988
• Tp 1280C 20C for normal mantle (thereby
  implying plumes for Tpgt1300 C)

4
TEMPERATURE BUMP IN UPPER MANTLE

• Internal heating and secular cooling are expected
  to decrease the radial geothermal gradient away
  from an adiabat. Modeling shows that the average
  thermal gradient is expected to be significantly
  subadiabatic through much of the interior of the
  mantle
• There may be a maximum in T near 100-200 km below
  the plate and below the depth of MORB extraction
• The geotherm is unlikely to be a TBL joining with
  an adiabat at the lithosphere-asthenosphere
  boundary (ala McKenzie, who uses the term
  lithosphere for the TBL )

5
(No Transcript)
6

more realistically, 30-50 C
7
McKenzie and colleagues assume the upper mantle
to be homogeneous and isothermal. They adopt a
cold subsolidus potential temperature of 1280C
20C for normal mantle. Sleep adopted ?Ts of
gt200 C to represent plumes
Most other hotspots and LIPs have no thermal or
heat flow anomaly (see www.mantleplumes.org)
8
Temperature Iceland
Foulger et al. (2005)
9

• The total range sampled by normal ridges
  inferred from petrology is 1250-1450C (Asimow,
  2006) or 1475C if Iceland is a ridge and is not
  built on a continental fragment (this includes
  crustal thicknesses of 3-11 km and includes
  'ridge-like' ridges, away from complexities that
  are likely to confound simple relationships
  between potential temperature, crustal thickness,
  and melt composition such as active flow, fertile
  sources, non-steady flow, focusing, EDGE
  effects).

10

• These petrological estimates are now consistent
  with long-standing geophysical estimates. Kaula
  (1983) estimated the minimal upper mantle
  temperature variations that are consistent with
  observed heat flow and plate velocities. At the
  fully convective level, about 280 km depth,
  temperature variations are at least 180C,
  averaged over 500 km spatial dimensions. Tp under
  ridges was estimated at 1410C. There is some
  indication that MORB at the onset of spreading
  are 50 hotter

11
Melting Temperatures(solidi extrapolated to P0)

• Eclogite 1100C (extrapolated from 1 MPa)
• Peridotite 1300C (from 1 MPa)
• Melting anomalies may be due to fertility streaks

12

• The potential temperature of the present upper
  mantle is 1400200C based on bathymetry,
  subsidence, heat flow, tomography, plate motions,
  discontinuity depths (Anderson, 2000). Temporal
  variations of 200C over 200 Myr are expected.
  Secular cooling of 100C in 1 Ga is plausible.
• Temperatures at onset of spreading may be 50C
  warmer
• McKenzie and Bickle (1988) assumed the upper
  mantle to be homogeneous and isothermal. They
  adopted a cold subsolidus potential temperature
  of 1280 20C for normal mantle and thus require
  hot plumes elsewhere.
• If normal mantle temperatures are 1400 200C,
  or even 1350 150C, there is no thermal
  requirement for hot mantle plumes.
• Convection simulations without plumes give the
  above ranges in temperature

13
RIDGES
Peridotite liquidus
Eclogite liquidus
Eclogite solidus
HISTOGRAM FROM ASIMOW (2006)
14
Cold eclogite can be negatively buoyant but it
can have low shear wave velocities low melting
point. Fertile eclogite blobs can be brought into
shallow mantle by entrainment or displacement or
by melting
15
Dry peridotite can only melt in shallow mantle
16
COLD ECLOGITE CAN MIMIC HOT UPWELLING
I
ridges
Presnell, Gudfinnsson, Herzberg
Dense cold eclogite can have low seismic
velocities
17
ARCS the hot mantle wedge paradox
Kelemen et al, 2002
Extreme case of subadiabatic gradient
18

• Figure 1. Predicted geotherms beneath arcs from
  thermal modeling (small symbols and fine lines),
  compared to petrological estimates of PT
  conditions in the uppermost mantle and lowermost
  crust in arcs (large symbols and thick lines).
• Most petrological estimates are several 100
  hotter than the highest temperature thermal
  models at a given depth.
• Wide grey lines illustrate a plausible thermal
  structure consistent with the petrological
  estimates.
• Such a thermal structure requires adiabatic
  mantle convection beneath the arc to a depth of
  50 km, instead of minimum depths of 80 km or
  more in most thermal models.
• Deeper mantle may be hotter than usually modeled.

19
Mantle Temperature Variations Beneath Back-arc
SpreadingCenters Inferred from Seismology,
Petrology, and BathymetryDouglas A. Wiens,
Katherine Kelley. Terry PlankEarth and Planetary
Science Letters
Compare max T with Hawaii
20
The currently high flux at Hawaii is unusualVan
Ark Lin, 2004
The quasi-periodic variations in the flux along
the Hawaiian ridge may be due to fertile streaks
or stress variations rather than pulsation of a
plume. The highest flux is on the young
lithosphere between the Murray and Molokai FZ
21
A fertility streak can be due to subduction of an
aseismic ridge or seamount chain (about 20 are
currently entering subduction zones)

• Hawaiian swell can be due to a buried buoyant
  load at 120 km
• depth (Van Ark Lin, 2004).

22

• In an internally heated mantle or in a mantle
  that is cooled by cold slabs, the geotherm
  becomes subadiabatic.
• This means that shallow mantle temperatures can
  be hotter than at 600 km.
• Actual mantle temperatures and their variations
  are greater, and the melting temperatures can be
  less, than assumed in plume modelling.