Transcript Hawaiian Plume geochemistry: HSDP data and models

Outline
1. Properties of silicate liquids
2. Adiabatic decompression melting
•
Melting temperature(s) of lherzolite
•
Model for mid-ocean ridges
3. Melting in mantle plumes
4. Effects of water and pyroxenite/eclogite veins
5. Phase equilibria of melting (only the basics)
6. Melt percolation models..... U series isotopes
Some questions about plumes in general
and the Hawaiian plume in particular
1. How big is the plume in horizontal dimensions?
•
Width of temperature anomaly
•
Width of upwelling velocity anomaly
•
Width of melting region
•
Width(s) of isotopic anomalies
•
Widths at depth >200km versus width in the melting
region (90 - 150km)
2. How hot is the plume (and why do we think so)
•
At its core
•
At the fringes of the melting region
Mid-ocean Ridge
1800
Temperature (°C)
Hawaii (90 Ma lithosphere)
2000
1600
mantle solidus
10% melting
1580°C
1480°C
1380°C
1400
1280°C
1200
1000
0
10
20
30
40
Pressure (kb)
50
60
70
80
There is currently one available
model for the Hawaiian plume
Ribe and
Christensen (1994)
R&C (1999) q at 110km depth
0
800km
QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture.
-200
0
200
400km
R&C (1999) generate the plume with a radial thermal anomaly at the
bottom of the box (400km depth). At 170 km depth T is approx.:
 r 2 
q (r)  q o exp 2 
 aq 
QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture.
Where aq ≈ 65 km, and ∆qo = 300K. These values depend in detail
on the viscosity structure of the plume. The width of the thermal
anomaly is about 130km (1s at170km), which is the full width at
about17% of ∆qmax. Constraints are (a) buoyancy flux, (b) H/W of
swell, (c) rate of magma production
QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture.
Integrated
amount of
melting
Melting rate
In the R&C (1999) model, the maximum degree of melting reached is
about 20%. The core of the plume does not melt above 90km depth. At
the edges of the melting region, melting stops at about 120km depth. At
120 km depth the plume width has increased by about 1.6 times as result
of spreading beneath the lithosphere
QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture.
In the R&C (1999) model, the width of the melting region at 120km
depth is about 120 km, and the total melt production varies horizontally
as:
 r 2 
G(r)  Go exp 2 
 aG 
where Go is about 0.05 m3/m2/yr, and aG ≈ 30 km. The width of the
melting region is about 1/3 the radius of the thermal anomaly (which has
spread to ≈100km width at 120 km depth).

QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture.
R&C (1999) do not specifically mention the upwelling velocities in their
model plume. From the melting rate at the core of the plume it can be
inferred that the maximum upwelling rate is about ≈ 35 cm/yr. This value
is similar to that obtained by Watson and McKenzie (1991) with a simpler
axisymmetric model.
Upwelling rate, plume width and temperature are all interrelated. The
relationships, and the effects of the assumptions about mantle rheology,
were investigated by Hauri et al. (1994)
Arrangement a la
DePaolo et al. 2001
Magma capture area
G. Hart and DePaolo (in prep), solidus from Hirschmann (2000). Equations from Asimow et al
Dry melting
120 km
90 km
Dry melting
MORB
Hawaii
Extent of Melting (F)
0.4
F1280
F1380
F1480
F1580
F1680
0.3
0.2
Core of plume
0.1
0
0
10
20
30
40
Pressure (kb)
50
60
70
Edge of melting
region
Effect of H2O on melting temperature, from Hirschmann et al. (2001)
QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture.
Effect of H2O on melting temperature, from Hirschmann et al. (2001)
Model
700ppm
H2O
QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture.
Model - dry
@Higher - P
Adiabatic
decompression
melting
2000
mantle solidus
700ppm H O
2
1800
dry
1% melting
Plume
Core
1580°C
1600
1480°C
1380°C
Melting
"edge"
M ORB
1400
1280°C
1200
90km
Temperature (°C)
20% melting
1000
0
10
20
30
40
Pressure (kb)
50
60
70
80
Magma generation and composition summary
• Two main types of basalt
– Tholeiitic
– Alkaline
• Tholeiitic basalt forms at lower pressure and with higher melt
fractions
• Alkali basalt forms only at higher pressure and with small melt
fractions
• Andesite forms in island arcs at high pressure because of the
abundacne of H2O
(P.S. The following is qualitative, but at least it may be understandable)
The difference between Tholeiitic and Alkali basalt can be
understood
withoxe
this
ne diagram.
Clinopyr
Diopside
Low Pressure, dry
Pressure < 1 kbar
CPX
Olivine
Oli
Or thopyr oxe ne
QUART
OLIVINE
OPX
Z
P
Quartz
Tholeiitic basalt will crystallize OPX and with extensive
fractional crystallization Clinopyr
can eventually
crystallize quartz
oxe ne
Diopside
Low Pressure, dry
Pressure < 1 kbar
CPX
Olivine
Or thopyr oxe ne
QUART
OLIVINE
OPX
Z
P
Tholeiitic
Crystallization
(OL, OPX,
OPX+CPX,
CPX+Q)
Quartz
Alkali basalt will not crystallize OPX or Quartz
Clinopyr oxe ne
Diopside
Low Pressure, dry
Pressure < 1 kbar
CPX
Alkaline
Crystallization
(OL, OL+CPX)
Olivine
Or thopyr oxe ne
QUART
OLIVINE
OPX
Z
P
Quartz
The high pressure version of the phase diagram shows
why Alkali basalt is formed at high pressure
Clinopyr
oxe ne
Diopside
High Pressure, dry
Pressure 20 kbar
CPX
Olivine
Or thopyroxe ne
Q U ARTZ
OL
OPX
P
Quartz
Mantle peridotite is OL-rich. This version of the diagram
ne
shows approximateClinopyr
SiO2 oxe
contents
of rocks
Diopside
High Pressure, dry
Pressure 20 kbar
CPX
P
Mantle
Peridotite
OPX
OL
65%
45%
Olivine
QUARTZ
55%
Or thopyroxe ne
Quartz
Clinopyr oxe ne
Diopside
High Pressure, dry
Pressure 20 kbar
CPX
First liquid to
form as
peridotite
begins to melt
P
Mantle
Peridotite
OPX
OL
65%
45%
Olivine
QUARTZ
55%
Or thopyroxe ne
Quartz
Diopside
Clinopyr
oxe ne
High Pressure, dry
Pressure 20 kbar
CPX
Mantle
Peridotite
P
OL
40
OPX
30
Q U ARTZ
0-20
Up to 20%
melting, the
liquid
composition
stays the same
and the solid
still has all 3
minerals
50
Olivine
Or thopyroxe ne
Quartz
Diopside
oxe ne
Clinopyr
High Pressure, dry
Pressure 20 kbar
CPX
Mantle
Peridotite
P
OL
40
OPX
30
50
Olivine
Or thopyroxe ne
Q U ARTZ
0-20
For >20%
melting, CPX is
eliminated
from the solid,
and the liquid
composition
becomes more
and more
OPX-rich
(more
tholeiitic)
Quartz
Diopside
oxe ne
Clinopyr
High Pressure, dry
Pressure 20 kbar
CPX
Mantle
Peridotite
P
OL
40
OPX
30
Q U ARTZ
0-20
The solid
gradually loses
CPX, then
OPX until just
OL remains
50
0
40
Olivine
20
Or thopyroxe ne
Quartz
When liquids formed at P>20kb are brought to the surface and crystallize,
those corresponding to lower % melts are seen to be alkali basalt, those at
higher % melts are tholeiitic Clinopyr oxe ne
Diopside
Low Pressure, dry
Pressure < 1 kbar
CPX
P
OPX
40
OLIVINE
Z
30
50
Olivine
Or thopyr oxe ne
QUART
0-20
Quartz
Summary diagram
50
Alkali basalt
Old lithosphere
thickness
Tholeiitic basalt
10
0
% melting
25
Examples - Hawaii has high melt fractions at high pressure and hence
the most common lavas are tholeiitic. Tahiti is mostly alkaline (it is a
weaker plume). MORB are always tholeiitic
50
Alkali basalt
Tahiti
Hawaii
Old lithosphere
thickness
Tholeiitic basalt
MORB
10
0
% melting
25
QuickTime™ and a
TIFF (LZW) decompressor
are needed to see this picture.
Backscattered electron image of an experiment “charge”
showing vitreous carbon spheres. During the experiment the
peridotite melts and the liquid is squeezed into to the space
between the spheres (P = 1.5 Gpa)
QuickTime™ and a
TIFF (LZW) decompressor
are needed to see this picture.
Lherzolite w/
more CPX
QuickTime™ and a
TIFF (LZW) decompressor
are needed to see this picture.
0
crust
20%
15%
20
5%
Depth (km)
10%
40
60
80
Linear depletion
in CPX for simple
100
MOR melting
0
model
Std. Lherz.
Depleted Lherz.
2
4
6
8
10 12 14 16
CPX (wt.% )
Lherzolite: Peridotite with Olivine > Opx + Cpx
Olivine
Pyrolite/
Primitive UM
Dunite
90
Peridotites
Lherzolite
40
Pyroxenites
Olivine Websterite
Orthopyroxenite
10
10
Orthopyroxene
Websterite
Clinopyroxenite
Clinopyroxene
In plumes, depletion isopleths correspond to CPX* mode
0%
QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture.
CPX*
8%
15%
(*Need to include garnet in the
model here.)
For trace element partitioning we use the expression:
Cliq
1

C0 Di  F(1 Di )
But, what is F for real situations?

For MOR model mean F
is about 0.05 to 0.10.
Varies from 0.20 to
0.00x over the melt zone.
QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture.
For Hawaii model mean
F is about 0.04 to 0.05;
averaged over entire
melting region. Varies
from 0.2 to 0.00x.
Cliq
1

C0 Di  F(1 Di )
C liq
C0



1
F
For Di=0
For weaker plumes,
mean F may be
≤0.01; but also lavas
may represent small
F as volcano drifts
off of the plume.
Other experiments showing difference
between tholeiitic and alkalic basalt
20
0.5 GPa
1.0 GPa
1.5 GPa
2.0 GPa
2.5 GPa
3.0 GPa
Normative Ne-Hy-Q
10
0
Alkali
basalt
Tholeiitic
basalt
-10
-20
-30
Kushiro (1996)
-40
0
5
10
15
20
25
% melting
30
35
40
QuickTime™ and a
TIFF (LZW) decompressor
are needed to see this picture.
Veins of eclogite/pyroxenite in the mantle: melting effects
Model for melt percolation; 1D steady state upwelling
Fmax
d
z
F (fraction melted)
v melt 
(solid  liq ) 2
liq
Fmax

d
F
z
For F/=10, w/W0 ≈ 12
Porosity(=melt fraction)
Melts produced at higher-P are not in equilibrium with lherzolite or
harzburgite at lower P (they are too SiO2-poor). They react with the solid
as they pass thru, precipitating OL
and dissolving
pyroxene.
oxe ne
Clinopyr
Diopside
Low Pressure, dry
Pressure < 1 kbar
CPX
0-20
P
OPX
40
OLIVINE
Z
30
50
Olivine
Or thopyr oxe ne
QUART
Liquid
produced
at 20kb
Quartz
Melt-filled channels
Computer models of melt migration
with formation of melt channels Marc Spiegelman, LDEO
Ridge
Channels containing upward flowing liquid
Mantle "depletion map"
Mid Ocean Ridge computer models Marc Spiegelman, LDEO