Download Eclogite formation and the rheology, buoyancy, seismicity, and H2O

EclogiteFormationand the Rheology,Buoyancy,
Seismicity,andH20 Contentof OceanicCrust
Bradley
R. Hacker
1
Departmentof Geologicaland EnvironmentalSciences,StanfordUniversity,Stanford,California
A broad spectrumof variably altered igneous rocks with a wide range of grain sizes are
compressedand heated over a wide range of pressure-temperature
paths in subductionzones.
Although experimentalkinetic data cannotbe extrapolatedto predictthe ratesof blueschistand
eclogite formation in nature, textural data from rocks indicate that transformationbelow
temperaturesof 150øCis minimal. Completetransformationof volcanicrocksoccursby •-250øC,
but incompletetransformationof gabbroicrocks heatedto 800øC has been observed.There are
important consequencesto the rapid transformationof volcanic rocks and the metastable
persistence
of gabbroicrocksinto theblueschist
andeclogitestabilityfields.Fastseismicvelocities
shouldbe evident first in the upper oceaniccrust and may be substantiallyretardedin the lower
oceaniccrust.The upperoceaniccrustwill be denserthan asthenosphere
beforethe lower oceanic
crust.Early in the processof eclogiteformation,volcanicrockswill be placedin deviatorictension
andthe underlyingcoarsergrainedrocksin compression;
with furtherreaction,the stateof stressin
gabbroicrockswill changefrom compressiveto tensile.Earthquakesat shallowdepthsshouldbe
extensional in basalt and contractionalin gabbro, changing at deeper levels to extensional
throughoutthe crust.
INTRODUCTION
Formationof eclogiteis a key factor in plate tectonics,
influencingthe size and shapeof platesas well astheir age
and rate of disappearance from the Earth's surface.
Subductionzoneshave long beenof interestbecausethey
are the buildersof continents,collectingmaterialfrom the
oceanbasinsandconstructing
magmaticarcs.In spiteof the
dynamic nature of subduction zones, much of our
understandingof their behavior stemsfrom equilibrium
thermodynamics.The equilibriumviewpointhasprovena
valuablefirst meansof addressing
processes
in subduction
zones, but a more in-depth comprehension requires
considerationof the ratesand mechanismsby which phase
changesoccurduringsubduction.
AlsoatU.S.Geological
Survey,
MenloPark,California
This paper summarizes the rates and textures of
densificationreactionsin subductionzonesby examining
experimental and field studies, and then considersthe
effects of these findings on the rheology, buoyancy,
seismicity,andH20 contentof subducting
oceaniccrust.
WHAT
ARE BLUESCHIST
AND
ECLOGITE?
Blueschistfacies mafic rocks are characterizedby the
coexistenceof sodic amphibole and lawsonite at low
temperatureor epidote at high temperature[Ernst, 1963].
They may alsocontaingarnetor omphacite.Eclogitesensu
stricto is mafic rock consistingof garnetand omphacitic
clinopyroxenewith or without additional minor phases
[Coleman et al., 1965]. The geophysicalimportanceof
eclogiteis that it represents
the highestdensitycommonly
attained by crustal rocks. Omphacite and garnet have
densities
of3.2-3.4gm/cm
3 and3.6-4.0gm/cm
3,amarked
Subduction:Top to Bottom
GeophysicalMonograph96
This paperis not subjectto U.S. copyright.
Publishedin 1996 by theAmericanGeophysical
Union
elevation beyond the density of a typical basalt at 2.9
gm/cm
3 [Carmichael,
1989].Thisshiftfrompositive
to
negativebuoyancyrelative to asthenosphere,
which has a
337
338
ECLOGITE
FORMATION
IN OCEANIC
CRUST
densityof roughly3.23 gm/cm
3 [Cloos,1993],is as
important a factor in controlling plate subductionas are
conductive cooling of the lithosphereand the olivine -->
spinel transformation in subductingmantle [Ahrens and
2.6
':iiiiiiiiiiii::.:.0::::•i•i•i!ii!!111iiiiiiiiiii!!iiiiii?iiiii½•::::•
..................
2.4
2.2
'?iiiilili•
3•7
':iiiiiii!iiiiii:. eclogite
.. ..........
::::.:::::::•.
....
,,•,•,:•,•,,..
3.24
..:.:...: .......
Schubert, 1975].
••'
3.10
••
•
PRINCIPAL
DENSIFICATION
REACTIONS
of basaltic
rocks to blueschist
1986].
Eclogite facies conditionsexist at temperatures>500øC
andpressures> 1.2 GPa (Figure 1). Eclogitefaciesrocksdo
not necessarilyexist wherever such conditionsprevail in
the Earth becauseof sluggishkinetics.All reactionsrequire
oversteppingof equilibriumto overcomelocal free energy
to attachment
or detachment
of atoms at
interfaces, and all reactions involve diffusion or atom-atom
bond rearrangement.Perceptible reaction only occurs at
sufficientlyhigh diffusivitiesand reactionfree energies.
HOW
FAST
ARE
REACTIONS
IN THE
..........................
........
•....
..... ;:•:•:•
...........
•-•:•:::•::,:.•'-----1!
..............................
•:•'-'::ep•dota•i•i•i•
......
•:•:•i•i•i•11!'•'•:g:'•i...'.
•:•
............
......:...:.:
....:.:.....:.:.:.:.:.:....•[•?.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:
.....
.........
•::•
?•
..........
•?::•
......................
......
•?
....................
•...........
•:•
....................
0.6
...................
ß
..............
2 93 ............
•"v ....... ,
0.4 subgreenschist
.....
•?' 3.5 .:•?
0.2
0
2.80
....
ii•i•}i•?"
5.2
1•
::•!•i•i:'"
•
3•
..•}•?"
-':3•iiii
..........
Tem•ramre (øC)
;.9ff .•{•?
•-'
0
:•{{•}•'
•:•
7•
8•
9•
Fig. 1. Metamorphic facies and reactions pertinent to eclogite
formation, after [Liou, 1971], [Maresch, 1977], [Brown, 1977;
Brown and Ghent, 1983; Evans, 1990], [Green and Ringwood,
1967; Ito and Kennedy,1971;Liu et al., 1993; Oh and Liou, 1990;
Poli, 1993]. Three- and two-digit numberfor eachfaciesindicate
density(g/cm3) and H20 content(wt%) calculated
using
NCMASH model of Peacock [1993] assumingequilibrium (i.e.,
no kinetic hindrance).
and
then eclogite at the low temperaturesthat prevail in most
subduction zones is more complicated because of the
prevalenceof disequilibriumin natureandexperiment.One
reactioninferredto have transformedblueschistto eclogite
at severallocalities worldwide is epidote + glaucophane
--> garnet + omphacite+ paragaonite+ quartz + H20 [ElShazly et al., 1990; Ridley and Dixon, 1984; Schliestedt,
increases related
....... .............
==========================
:::::::::::::::::::::::::::::::::::::
ß
A broad spectrumof reactionscauserocks to become
more densein subductionzones.At the high temperatures
found in very hot subductionzonesthe transformationof
basalticrocksto garnetgranuliteandthen eclogiteinvolves
the breakdown of plagioclase, addition of Na20 to
cliopyroxene and the growth of garnet (Figure 1). PT
conditions for the appearance of garnet have been
determinedby experimentalstudy [Green and Ringwood,
1967; Ito and Kennedy, 1971; Liu et al., 1993; Poli, 1993],
and occursatlower pressures/higher
temperaturesin SiO2undersaturated[Ito and Kennedy, 1971] and MgO-rich
rocks [Green and Ringwood,1967]. The disappearance
of
plagioclase is less well defined by the same studiesand
dependsmore stronglyon rock composition.
The transformation
......
.....: ......
.::::::::::::.::::::::::::
EARTH?
Experimental and TheoreticalInferences
How long can rocks remain out of equilibrium at
elevated pressuresand temperatures?Specifically, how
long can rocks remain within the eclogite stability field
before transformingpartially or completely to the stable
phase assemblage?Information relevant to this question
derivesfrom two sources:experimentand studyof natural
rocks[Rubie, 1990]. Experimentalkinetic datacollectedon
powderedmaterialsmay not be directlyapplicableto the
majority of geologic situationswhere rocks are tightly
packed crystallineaggregates--principallybecausethe
interfacialfree energyof a rock is muchlessthanthatof a
powder.For extrapolationto geologicconditions,dataon
phase transformationsin unpowderedmaterials are
required--information
thatis exceedinglysparse.
Recent experimentson reactionsknown to occur in
subduction zones demonstrate that most kinetic data cannot
be extrapolated
to theEarthandthatH20 hasmajoreffects
on reaction rate and texture. For instance,Holland [1980]
observedjadeite growthin 24 hr from powderedalbite +
quartz at 900øC and a pressureoverstepof 50 MPa. In
contrast,albite -->jadeite+ quartzexperiments
by Hacker
et al. [1993] on albiterock showedno transformation
in 24
hoursat 1100øCat a pressureoverstepof 500 MPa!
Quantitative transformation-rate measurements are
available for the calcite <--->aragonitereaction. As with
albite, the calcite --->aragonitetransformationin Carrara
marble[e.g., Hacker et al., 1992] is muchslowerthanin
calcite powder. For example, at 600øC and 2.0 GPa
confiningpressure,powderedcalcitetransforms50% to
aragonitein <1 hour[Brar and Schloessin,
1979,p. 1409],
whereas192 hourswere requiredfor similarconversionin
unpowdered
marble.CarlsonandRosenfeld[1981]andLtu
and Yund [1993] measuredthe rate of the aragonite --->
calcite reaction;extrapolatedto geologictimescalesand
grainsizes,theirdataimply that 1 mm aragonitegrainswill
HACKER
revert completely to calcite in less than 1 m.y. at
temperatures as low as 200øC. Hacker et al. [1992]
exploredthe reversetransformation
in marbleand founda
similar result, that the calcite --> aragonitetransformation
occurswithin less than 1 m.y. at 200øC for overstepsof
<100 MPa. The reactiondependsstronglyon temperature,
such that even several GPa overstepat temperaturesof
<100øCareunlikely to promotetransformation.
The catalyticactionof H20 is well known from kinetic
studiesof powders and rocks. Hacker and Kirby [1993]
deformed polycrystalline, vacuum-driedalbite rocks at
strain
ratesof 10-4-10
-6s-1 attemperatures
of 600-800øC
andconfiningpressures
of 1.0-2.0 GPa. In spiteof extreme
pressureoverstepping
andextremesamplestrain,deformed
anhydrous samples contained no jadeite. Differential
stressesas high as 2000 MPa producedmaximumnormal
stressesup to 4.0 GPa--all at temperatureswhere the
equilibriumpressureis <2.0 GPa. Someof thesesamples
were strained70%, and individual grains attainedaspect
3.0
339
•-,
c.....•o_.esite
eclogite
-:
:.•feeble,
c.
25%'::'""'"•:'""'"•i•:...,
•:i'/!111['"'":'""'""
'"'
"'"'""'":•:'
........
2.8
2.6
;"[]moderate,
c.50%
2.4
I extensive,
c.75%
2.2-
I complete,
100%
2.0-
'?'"'"'•jiii:.
eclogite
'"':":::½'{'"'"'"':'
g-."i:-•i
!•,•;•;*:'
:'
'--•cooling
:.'.-':•:•:•:•:•.'.-':•:i:.'.
""•i: ......... '
lawsonite
blueschist
2:iiii::
::•::5"
i•5::i•5•!'1•.
.:i::::::::ii
•$;";.'•:
t'•;J
.•i......
..si::::::::::::::ii
[g'-:$•
t'•-":•
•
"....
::--'•".:"-..--'i--'•::•
-:•:..........
••,.•.::lim•'-'
ß
:>.-r.
,'-•m:
:•,-•.,-.amet
granuh.t...e.....
•,,.,...:.©
•½.•-•...-•::•
..... :,•..'p•.•'./•.,.•-..•j
• ::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
.................
•:•:•:•,
..........
•.-.:--•:•:•:.-',•.-.::i'..:,•.-.:.,•.--:.•
.................
:.....
ß ........:::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
1.0
• •'..•.,x
-. .....!,x.•qx
B .:•5{'
0.8
IIi-'•
•i•'•:
'......
• ; • ....'•
/i:'.i:'.....'.-':•:•:5
:'••I'•
•-•i .'-:::'-.:i:
*!•½:.':
0.6
.-•
.... : :' .-::::::
•
"'"":::
5:5':"
:':':':'
•{•'"'&'•ø'•:e'
:'":•'
hi 1'
subgreenschi•'t
.."'"'•½':'gree
'"'"'
""'"'
'
nschist.':•i'-"?
':'
0.4
•.:-'•i•i.-'..:
::]-'.-'-'•:•:•-•
:::::.-::-'.--'..
.
5:•5•'.':•:•:
granuhte
'"'•?'
':'"'
'"'
'
0.2
0.0
0
1•0 2•)03•)0
Temperature
(oC)6•)0
7•)08•)09&) 1000
Fig. 2. Reported extents of transformation, pressures, and
temperatures,for blueschistand eclogite facies metamorphism
worldwide. References and data used to compile figure are
availablefrom the authoron request.
ratios
ofgreater
than10:1.In marked
contrast
tothis,the
addition of 1 wt% H20 produced partial reaction at
temperatures as low as 600øC, even in undeformed
samples.
More complicatedexperimentsmeasuredthe olivine --->
spinel transformationin hot-pressedMg2GeO4 [Burnley,
1990] and Ni2SiO4 [Rubie et al., 1990] at pressuresand
temperatures of 800-1900øC and 1-15 GPa. When
extrapolatedto geologic conditionsthese investigations
suggestthat transformationmay be suppressedto depths
150-200 km below the equilibrium boundaryin rapidly
subductingslabs.
In summary, existing experimentaldata indicate that
H20 has a major catalyticaffect on reactionratesand that
some reactionsare geologicallyfast (e.g., calcite-->aragonile) whereasothersare probablyslow (olivine --> spinel).
We have essentiallyno experimentalinformationon reactionsformingblueschistandeclogite.
Metamorphic Textures
Observations
on rocks
show
that transformation
to
blueschistand eclogitemay be suppressed
long afterrocks
have passedinto the respectivestability fields of these
assemblages.Figure 2 was constructedfrom studies
worldwide that reported not only peak metamorphic
pressuresand temperaturesbut also textural information
about the degree of reaction. Rocks reported as feebly
recrystallizedor whosehigh-pressure
phasesweresominor
as to requiredetectionby electronmicroscopyare shownin
Figure2 as"feeble,25%." Rockscontainingabundantrelict
phases (usually clinopyroxene) or partially developed
reaction coronas are denoted as "moderate, 50%", and those
with rare relict phases are labeled "extensive, 75%".
Metamafic rocks specificallydescribedas lacking relict
phasesare shownas"complete,100%".
There are some important limitations to Figure 2. 1)
Reportedpeak pressuresand temperatures
may not have
occurred at the same time and, at best, representonly a
singleway-station
on thePT pathexperienced
by therocks.
2) How long the rockswere at the reportedpressures
and
temperatures
is unknown.3) Pressureestimatesfor many
rocksare minimabecauseof the highvarianceof the phase
assemblages;
theseare denotedby U-shapedlines rather
thanrectangles.4) Many high-pressure
rocksexperienced
substantial post-high pressure alteration that likely
consumedexistingrelict phases;studiesreportingsevere
alterationof thistypewereexcludedfromconsideration.
5)
Fine-grainedrocksreact fasterthan coarse-grained
rocks
becauseof increasedinterfacialfree energy;gabbroicrocks
areonly partiallytransformed
at conditions
wherevolcanics
arecompletelyrecrystallized.
Numerouspapersthatdo not
specifically
statewhetherrelictphases
arepresent
or absent
had to be excludedfrom Figure 2, but this may simply
mean that the absenceof relics was not significantfor the
purposesof the study.ThusFigure2 shouldbe takenas an
illustration
of the transformation
behavior
of coarse-
grained,i.e., gabbroicrocks.
With these limitations in mind, Figure 2 indicatesthat
feebly metamorphosed rocks generally persist to
temperaturesof 200-250øC. Moderate transformation
beginsat 200-250øC andis foundin rocksthatwereheated
to just above 600øC. There are a few casesof extremely
340 ECLOGITE FORMATION
IN OCEANIC CRUST
limitedreactionat highte•nperature
shownby dashedboxes
in Figure2. Incompleteconversion
of gabbroto eclogiteis
known to have happenedat temperatures>700øC and
pressures
of 1.5-2.0 GPa in the presence
of aqueous
fluid
[MOrk,1985]. Suppression
of transformation
of gabbroto
eclogitewasreportedfrom rocksthatreached800-900øC
andpressures
of >2.8 GPain the absence
of fluid [Zhanget
al., 1995]. Oddly enough,thereis little differencein thePT
range over which extensiveto completetransformation
have been found. Both have been observed from rocks
heatedto lessthan300øCandyet partialtransformation
of
gabbroto eclogitehasoccurredat 760-850øCand 1.6-2.0
GPa [Indares and Rivers, 1995]. In general,500-600øC
marks the transitionto extensiveor completereactionfor
most rocks. A few cases, however, indicate that the
transformation
of gabbroto eclogitemay be suppressed
to
muchhighertemperatures.
Fe-rich rocks transform to eclogite along high P/T
trajectoriesmore readily than Mg-rich rocks because
reactionsinvolvingFe-richphasesgenerallyoccurat lower
temperatures
[Bohlenet al., 1983].Fieldevidenceof this-MgO-rich rockswith abundantigneousrelicsadjacentto
Fe-rich rocks with rare relict phases--hasbeen noted at
BardoneyValley in theAlps [ReynardandBalleyre,1988],
at Flems0y, Norway [MOrk, 1985], and in Piedmont
ophiolites[Pognante,1991;Sandroneet al., 1986].
Eclogiteformationis fasterin finer grainedrocks.For
example,in the BardoneyValley of the Alps, abundant
relict augites are present in gabbros, but absent in
metavolcanics[Reynardand Balleyre, 1988]. In the Sesia
Lanzo zone farther east, amphibolite transformed to
eclogitecontainsrelics in coarserocksbut nonein finegrainedrocks[Lardeauxand Spalla, 1991].
Not only is H20 provento acceleratereactionsin the
laboratory,numerouslocalitiesdemonstrate
thatthisis true
in natureas well. Mafic granulitein the MusgraveRanges
of Australia cooled into the eclogite stability field and,
thoughthe bulk of the rock is unaltered,local shearzones
contain omphacite + garnet + minor zoisite [Ellis and
Maboko, 1992]; the presenceof zoisite implies that the
reaction was fluxed by H20. At Holsn0y in the Bergen
Arcs, zoisite-bearing
eclogitehaloesin undeformedrock
surrounding
deformedveinsalsoindicatethat fluid drove
reaction[Boundyet al., 1992; Klaper, 1990]. Deformation
is involved
in
these
fluid-mediated
transformations
inasmuch
asthefluid mustmakeits way to thereactionsite
via cracks.
structure of oceanic crust. Oceanic crust has been studied
by dredging and drilling in ocean basins,and by the
examination
of ophiolites.The Samailophiolitein Omanis
the best-exposed,
largest,least-deformed,and perhaps
most-studied
ophiolitein the world.Plutonicrocksof the
Samailophiolitearedivisibleintoa lowerlayeredsequence
andan upperisotropicsection.The lower2.3 km (avg.)is
well-foliated and lineated, layered ultramafic (chiefly
wehrlite, dunite, and clinopyroxenite) through mafic
cumulates [Pallister and Hopson, 1981]. Olivine
predominates
at thebaseof thesection,whereas
plagioclase
is dominantat the top [Juteauet al., 1988b].Three-quarters
of the sectionis gabbro--typicallyolivine-clinopyroxene
or clinopyroxenegabbrowith minorhornblendeandFeTi
oxides--but at rare localitiesit is gabbronorite[Juteauet
al., 1988a].The gabbroconsists
of mm-sizegrainsof 55%
plagioclase,35% clinopyroxene,10% olivine, andrarely
Ti-pargasite [Browning, 1984]. Overlying the layered
sectionare 200-900 m of isotropichigh-levelmaficrocks,
chiefly hornblende-clinopyroxene
gabbro [Pallister and
Hopson,1981]. Grain sizesrangefrom mm to pegmatitic
[Juteauet al., 1988a]. Sheeteddikes containabundantto
sparseplagioclaseand augite [Hopsonet al., 1981]. The
volcanic rocks are non- to moderately vesicular,almost
aphyric pillows and rare brecciatedflows and massive
flows [Ernewein et al., 1988].
Mid-ocean-ridge
basalts(MORBs) typicallyareglassy
or containplagioclasewith lessolivine [Hekinian,1982],
while rockscrystallizedin intraoceanicmagmaticarcsgenerally containplagioclaseand augite [Ewart, 1982]. Midocean-ridgebasalts (MORBs) are subalkalinetholeiites
that typically contain plagioclase with less olivine
[Hekinian, 1982]. Oceanic plateaus, which make up a
sizeableproportionof oceaniccrust, are similarin their
majorelementchemistryto MORB, but are 10-40 km thick
[Cloos, 1993]. Hotspot lavas and propagatingrifts are
enriched in Fe and Ti relative to MORB [Flower, 1991].
Lavasproducedat fastspreading
ridgessuchasthePacific
have lower Mg/Fe ratios than MORB eruptedat slow
spreadingridges[Batiza,1991].
ALTERATION
OF IGNEOUS
OCEANIC
CRUST
Because
phasetransformations
are stronglyinfluenced
by theavailabilityof H20 andtheextentandtypeof preexistingalteration,understanding
the alteration
of oceanic
crust is critical to evaluatingwhen and where phase
transformations occur. Variable alteration of oceanic crust
COMPOSITION
OF IGNEOUS
OCEANIC
To understand
the transformation
CRUST
of oceanic crust to
blueschistandeclogitewe mustknowthe composition
and
to lower grade mineral assemblages
occursby active
hydrothermal
circulationat spreading
centersandby later
weathering
on theseafloor[Altet al., 1986;Humphrisand
Thompson,
1978]. Somerocksescapealterationentirely,
HACKER
while others are wholly recrystallized to new minerals.
Most typically, lavas are variably metamorphosedto
subgreenschist
facies,dikeschangepartially to greenschist
facies, and gabbro recrystallizesweakly to amphibolitefaciesminerals.The pressuresat which thesemetamorphic
assemblages form are relatively low (<300 MPa;
determinedby ocean depth and the thicknessof oceanic
crust) and the formation temperaturesmay range up to
A
Equilibriummodel
2.8
ß, base of crust
2.6
--
2.4
best information
about the alteration
of oceanic
lavas and dikes comes from DSDP hole 504B [Alt et al.,
1989; Alt et al., 1986]. There the lavas contain -10%
alterationhaloesaroundveins and cracks,leadingto about
1 wt% H20 in the bulk rock. Veins and vugs containFehydroxides, clay, celadonite, smectite, zeolites, and
carbonate; groundmass minerals are altered to Fehydroxides and clay; plagioclase is partly altered to
smectite;elivine is partly to totally replacedby smectite
and minor calcite;glassis partly to completelyreplacedby
clay; and pyroxenesare unaltered.Alteration in the dikes
increases down section to a maximum
of-50
•
::..::•..i•}•.
3.25
........
!•.:...iii
'"•
'•"
/3.02
,]Z
.:i:.:i?
:'"i{{i":":'"'"•J:-5•!!:
.,..-:•---'-'--}i'ii
'""'"'"'"'"'"'
':'•':':•:"•••••
'
2.0-
.•?;•?.":.-:-:iii:":-'?.::.
O.9
OF SLOW
/
1.0
.:g?siiiii!}.::..,•:i•i::iii{ii?.--•ii•¾-3':
-:' 1.2:g-':.-.'-:....-'•
.....
........
xD;•
,:•557:..:'-"....'5•i
.................
'"••5!5"'"•:':"
/
0.8
........
/ •/'"":•:%•i;':-::•?
.... ' 2.91........... '•::'"":5•ii
2.89
0.6
2.99
.........
0.4
0.2
0.0
0
1'00 200 •00Temperature
(øC)
B
2.8
•
Kineticmodelwith unaltered
basalticandgabbroicprotolith
•
.
[......::.-..-:3.06
3.23 3.34 ..........
2.6.
/'k
base
Of
crust
/ ':':•--"'"J•ii:•:0•
-O•*.i;-"•'"'"'•
•--::""•"'•::i5
-"•g"i•i'•'•*
.............
0.3
':O.....-•
3.37
ß
'::"?•:•.
:.ii-i':"i•,•i
'•:'::
'••'••--..........
" --- - •::
:'•
............
-!!i.
5.:65.77.--.--:•--.•
2.4.
V-',-,7
top
elcrust
/
2.2.
i /
':':•..:.•"3.16:
..•
•'
•U
. :!ii::'".-"-5:..-;3•i•i0.7
3.25
•.--:-:•:.....
:
.;...../
0
.. 294
'g•':"i"'93
........
•:a:..;•:.::"k.
. ' i['"'"
"•'•'"'"2
':j:":=8
i
..........
:.-!•.:-'•i
'-"'
"•'
':"'"•'"'/:•
'""'""
' ' .....
::0.6
/2.3:. ß'?•;•i-.:'
'-'.'•'
'"'"'--'-'":-'-'•--'•••••iim"•-•":-----'"•....:j•i
1.0•
,'
/
":•'"••'•"'•i'":•:•iii•?-'-ili•
"•'""
............
"' ' .,.:.::s::.-::.-::.-e,%-.'<-:.:e
0. 9'
............
'•.*:':.: .......
•::-:::-x
u 1.2 .:..:'..-':..-':..:•:•
:::s::::.'-•
•.:.:...:,:s:...::.-:j;-'.:4-'.i'"'"'""
..-•:---•:.:-%-.':i:Z:•::'-'•::':i
0.6 2.89 4'00
., /-:...-:o.....-.•......-•-::.'•;?:
2.0
' 2f90
:?:iiil..-"ii?
i. 98' z.:,v
.,,,,, ":'•--'"d•il}i
2ß89
:':";"'"":•'
oh'::ii{
:::•--:".'?'
3 '::2:--•"2
:-:..-'•.:'.:'•i.-'::•
0.4
o
•.oo ,,.,.o,,i.'.i-..? ß --•;•i:.-ii•:'
ß7 _'
• n '::":J•..:'."{.-::0
•
.0.5 / 2.{I ""
_ ._/
;:..:.-'.•?
: 1.3
ß-'
-":-..-":..':-'-_•?
0.2 gabbro:!2S
"'"0/"":'•:•.
-S'O/o
rea'"•h i7S0/o:
100%
:::'"'"'"/'::•ction
0.0 bas91t:
:50%rxn. •00%reaction
i
0
C
100' 200' 300Temperature(oC)g00
000 100 •00 1000
Kineticmodelwithzeolite-facies
basaltandamphibolite-facies
gabbro
protolith
2.6•kbase
ofcrust // ':•-;"•iii•i•O.5
...0.....:::::•,:..•
•":•:""-'"•!:'i:i•5':.•"•i•11
t}..............i'"':':":'"'"::•:•'::'"•
.............
'*:?'":'
:•••••i!i!i::ii[.-'•i•::•--"..-':-•'•
'•'':"'- ' ''..
2.4T top
ofcrust
' /
":!{3'i•:•'
?
..........
'
2.8t
' :' .' !.:.::::.:..•:.3
3.23
3.34
............
;::-----...
,
/
.......!
3.31
:!"":•
'
•/o:•' "::]•ii:i3.16
3.25
2.2
9
3 02
/
,
/
/
0
35/'"'""•73•t.5
....
-'"d..-'3";!•,
.......
,.-..':•C'"'"'"'•'"'•"'•'•••••
.........
.
/
.,
•"
.' i:::::::i:
ß
B::::::::::::
:::.-:
::::::•:-:-:-:...
:-,•-:.::•:.::!:-:[.?.i:i:.::---:-:-:....
;...•:.5:...-?. ""'"'::'•.-'-':":?:
3 O1
.:i•i:i.:i:';-i
•
?'"'::'•:?
.....
*'•':::;{':'"":::"'""•!','*'
..... 1'2 iii::}:{
\1/
u .......;.:-:½;•:--..!
.......
...:•
"""'":•':•?"':":•:•,,,.•g
.i:.iii?..,.:•?
:•!:i:.:..-::.......•'.
-.-•.
• ?.:.•:...--.•i:
..........
.;.'.
................
?•.
:•:.:'•..-.•::•-.'.-..',
ß
:-..:i?':•.
..'•:ii:..:•:..211ii?'
'•...•
':'-':.
::•
?'•"
ßß
'-ß-----?-':'y-':•:'":'
0.6299/254
29
0.4 . ' ß3. / 6.7. _-- "'"'•-:'•i!-':':"
i ß' - '"':'--"-'-"-•57
0.2
gabbro:25'0/":'"'•0""::'•.
50%readi--':.;.•:n
:75% 100%::::::::•ction
0.0 basalt:
i$0%
rxn.100%
re?lion
i
[Robinsonet al., 1991].
EFFECTS
•:.:,•;ii::::.:'.:'giii?..':•.
:_gig
':;':::::':'
3 4 ..:::::5..`..}iiii•.........•.!{iii•..`.:iii.......•i•i•.•*•::•`.`...:.....
':.:'i!•{•.--:-.:...:......--:::-:ggiiii.::'m"'---:-:-:-:.-:•..-•..,..:.:.'-.-:.,i,e
/.-2..-.-'•?
3.4
.....
3.!0
vol%. elivine
is replaced by chicrite, clay, and talc; plagioclase is
partially to totally replacedby albite + zeolites,chicrite,or
clay; and augitcis unalteredhigh in the sectionandrimmed
by actinolitedeeperin the section.The H20 contentof the
dikesaverages2 wt% [Alt et al., 1989;Alt et al., 1986].
Good information on the alteration of plutonic rocks
comes from the Samail ophiolite and ODP core holes.
Alterationof the Samaillayeredplutonicrocksis limitedto
rare replacementof clinopyroxeneby actinolite[Lippardet
al., 1986] and alterationof the isotropicplutonic rocks is
limited to the growth of up to 10% actinolite and FeTi
oxides,andthe infilling of cavitieswith epidote+ sphene+
prehnite + quartz [Ernewein et al., 1988]. In contrast,
gabbroicrocks drilled from ODP hole 735B in the Indian
Oceanshowwidespreadamphibolite-facies
alterationfrom
0-100%, with an averageof perhaps20-30% [Dick et al.,
1991]. elivine and pyroxeneare replacedby hornblende,
thereis partial to completealterationof elivine to colorless
amphiboleand talc, and 2.4 vol% of the rock consistsof
hydrothermal and late magmatic veins composed of
hornblende or oligoclase + diopside + epidote. H20
contentsof the 735B gabbros are mostly 0.5-1.5 wt%
-----....
--..:•:!•i•:::.-::-:::-'
...........
3.31
';iii?iii'.:•.
0
top of crust
2.2-
-650øC.
Our
341
0
TRANSFORMATION
ON
DENSITY AND H20 CONTENT
When oceaniccrust is subductedit is compressedand
heatedalong one of a wide variety of possiblePT paths
(Fig 3). The highest-temperature
path into the eclogite
•60' 2•' 360Temp•a•a,e{oC)Oh0
7b0 860 900•000
Fig.3. Densities
(g/cm
3) andH20 contents
(wt%)forsubducted
oceaniccrust calculatedusing NCMASH model of Peacockand
kinetic data discussedin text. A) Equilibriummodel.B) Kinetic
model with unalteredbasalt/gabbroprotolith. C) Kinetic model
with subgreenschist-facies
basalticprotolith and amphibolitefaciesgabbroicprotolith.Where two numbersare given,the first
refersto gabbroandthe secondto basalt.
342
ECLOGITE
FORMATION
IN OCEANIC
CRUST
stability field is from the granulitefacies. In the hottestof
subductionzonesinvolvinglithospherelessthana few m.y.
old, it is possiblefor oceanicigneousrocksto passinto the
eclogitestability field by this path [Hacker, 1991]. Rocks
that recrystallized in the subgreenschistfacies during
hydrothermalalteration at the mid-oceanridge will pass
through the greenschistand amphibolite facies prior to
enteringgranuliteconditions,and may transformvariably
to greenschist,amphibolite,or granulitefaciesassemblages
prior to entering the stability field of eclogite. At the
opposite end of the temperature spectrum,rocks that
previouslyrecrystallizedin the prehnite-pumpellyitefacies
will pass through the blueschistfacies prior to entering
eclogite conditions [Peacock, 1990], and may thus
transform variably to blueschist prior to entering the
eclogitestabilityfield.
Peacock[ 1993] usedthe restrictedcompositionalsystem
NCMASH (Na-Ca-Mg-AI-Si-H-O) to calculatethe mineral
modes and maximum H20 contentsof subductingmafic
rocks. He demonstratedthat moderately altered oceanic
crustcontaining 1-2 wt% H20 beginsto dehydrateat the
onsetof eclogiteor amphibolitefaciesmetamorphism,and
suggestedthat the transition from blueschistto eclogite
facies, associated with the breakdown of lawsonite or
clinozoisite,releasesthe mostH20 duringsubduction.
Peacockcalculateda rangeof pathsthattraversemostof
the regionof PT spacerelevantto subductionzones.Figure
3 illustrates two paths for the uppermostand lowermost
partsof subducted7-km thick oceaniccrustmcalculatedfor
a subductionvelocity of 50 mm/a and shearstresses
of 100
and 33 MPa. Most of the PT paths followed by the
uppermostlayer of the crust (basalt) are limited by the
paths identified by upright triangles, and most of paths
followed by the lowermost layer of the crust (gabbro) are
limited by the paths with invertedtriangles.Intermediate
levelsin the crustfollow intermediatePT paths.For path 1,
the uppercrustpassesprogressively
throughthe greenschist
to amphiboliteto eclogitefaciesconditions,while the lower
crustevolvesfrom blueschistto eclogitefacies.Along path
3, all the crust remains at blueschist facies conditions to
pressuresof 2.5 GPa. For pathsintermediatebetween1 and
3, most of the crust passes through the blueschist and
eclogitefacies.
In the spirit of Peacock's calculationsand using the
information presentedin Figure 2, assumethat coarsegrainedrocks(i.e., gabbro)transform25%, 50%, 75%, and
100% at 150øC,250øC, 500øC and 550øC,respectively,in
accordance
with
field
observations
described
earlier.
Further assume that fine-grained rocks (i.e., basalt)
transform twice as fast--such that they reach 100%
transformation
at 250øC. Assume also that oceanic crust is
made of 3 km of basalt and 4 km of gabbro.Given these
assumptions,
whatconclusions
canbe drawnabouttherates
of eclogiteformationin subductionzonesandits effecton
rock physical properties? Possible variables to be
consideredincludethe amountof H20 andthe composition
of the crust.
Figure 3A and 4A illustratethe predicteddensitiesand
H20 contentsof NCMASH assemblages
determinedby
Peacock[ 1993] for the variousmetamorphicfacies.Figures
3B and 4B repeatthis theme,usingthe kineticinferences
discussedabove.Figures3C and 4C use the samekinetic
hindranceconsiderations
but begin with alteredprotoliths.
Because these figures are based on modeling real
assemblageswith a subsetdefined by the NCMASH
system,the predicteddensitiesandH20 contentscannotbe
consideredexact; however, several interestingfeaturesare
apparent.
Figures 3B and 4B predict the behaviorof subducted
basaltand gabbrousing the mineral assemblageolivine +
orthopyroxene
+ clinopyroxene+ plagioclaseasa protolith.
In the upper, volcanic crust, 50% reactionis assumedto
occur at 150øC and complete reaction at 250øC; in the
lower plutonic crust, 25, 50, 75, and 100% reaction is
modeled as happening at 150, 250, 500, and 550øC,
respectively.
Forexample,
thedensity
of 2.97g/cm
3 for
lawsonite-blueschist facies gabbro in Figure 3B was
calculated
using
unaltered
gabbro
(2.89g/cm
3)transformed
50% to lawsonite
blueschist
(3.10 g/cm3).The H20
requiredfor hydrationin Figures3B and 4B is assumedto
derivefrom sedimentsor hydratedmantle.In the absenceof
sufficientH20, all faciesotherthanCE, EC, GG, and GN
haveadensity
of2.89g/cm
3 and0 wt%H20.
The resultsfor volcanicrocksarecloseto thepredictions
of Peacock'sequilibrium model (Figure 1) becausethe
transformationrate is rapid. Basaltstravelingvia path 1
undergoonly slightchangesin densityprior to the appearance of garnet and the disappearanceof plagioclase.In
contrast, Path 3 indicates that 5-10% volume loss occursin
basaltsupon enteringboth the blueschistand eclogitefacies.As in the equilibriumcase(Figures3A and 4A), the
biggeststepin dehydrationoccursat the blueschist/eclogite
facies boundary;dehydrationalong path 1 is split subequally betweenthe greenschist/amphibolite
and amphibolite/granulitetransitions.Basaltsmoving alongpath 3 or
slightlywarmertrajectoriesloseroughlyequalamountsof
H20 when entering and leaving the epidote blueschist
stability field. Comparedto basalt, where volume loss
occursfairly evenlyoverthe 150-450øCtemperature
range,
half the volume loss of gabbrooccursin the 500-550 ø
interval. In contrast to basalt, which may undergo
substantialhydration by 250øC, gabbro is predictedto
HACKER
A
Equilibrtum
model
with
unaltered
basalt
and
gabbro
protolith
C
Kinetic
model
with
prehnite-pumpellyite
and
amphibolite
facies
protoliths
distance
(kin)
0
50
100
343
distance
(kin)
150
200
Path
1 •
= 2ocm/a
0
0
100
MPa.•
50
100
150
200
Path
1 • •n•
v = 2ocm/a •_1%u __•..._
0
'c= 100
MPa .•l 5.7_
,os ••
AM
•
• •o
GG
•
•
50
•
311 •
Pa•3 •
•sGS'"•••••
//
•4 ///
=33
MPa
[ 2•9
GG
•I0
•• •299
X•"•• 311 50•
__
Pa•3
•
-••••••
324
/
//
/ 1•
$0
I0•
$0
I0
Fig.4. Densities
(g/cm
3) andH20 contents
(wt%)of subducted
B
crust, basedon Figure 3. A) Equilibrium model with essentially
Kineticmodelwith unalteredbasaltandgabbroprotolith
unaltered crust. B) Kinetic model with unaltered crust. C) Kinetic
distance(kin)
0
Path
•
•
v = 10 crn/a
•c
-- MPa 2i89
50
2.80
100
150
200
0
modelwith crustalteredto subgreenschist-facies
andamphibolitefaciesminerals.Comparewith [Peacock,1993] Figure4. Thicker
lines are facies boundariesand thinner lines representchanges
from 25% to 50% to 75% to complete transformation. The
thickness of the crust is stretchedby a factor of 5 for clarity.
Elongate rectangle at bottom of each panel shows density
difference
relative
toasthenosphere
(3.23gm/cm3).
DISCUSSION:
GEODYNAMIC
IMPLICATIONS
The retardationof reactionduring subductionillustrated
in Figures 3 and 4 will affect 1) how the structure of
subductedcrust is interpreted from seismic data, 2) the
buoyancyforces that producesinking and bendingof the
slab, 3) the distribution of stressesinduced by volume
changes, 4) seismicity in the slab induced by volumechange-relatedstresses,5) the mechanicalbehaviorof the
slab, 6) the distributionof H20 in the slab, and 7) the
undergohydrationto temperatures
up to 550øCmalthough generationof arc magmatism.All thesefactorsdiffer from
at any givenconditionthe amountof H20 in gabbrois less the equilibriumstatebecausethe densificationreactionsin
thanin basalt.Again, as in the equilibriummodel,the bulk
descendingslabsdependon grain size, H20 content,bulk
of the dehydration occurs at the blueschist/eclogite composition,and temperature.Relative to the equilibrium
transition.
model, transformationto blueschistand eclogiteis always
Figures 3C and 4C predict the behavior of subducted delayedto greaterdepthsin the kineticmodel.
altered basalt and gabbro using the prehnite-pumpellyite
The first rocks to transform will be glassy basalts,
faciesmineralassemblage
chlorite+ albite+ pumpellyite+
followed by holocrystalline basalts, then diabases,and
quartz+ calcitefor alteredbasaltand the amphibolite-facies finally gabbros.N9te that this derivesfrom thermalas well
assemblagehornblende+ clinozoisite+ chlorite + quartz as kinetic causes,as the upper crust is warmer than the
for alteredgabbro.The predictionsare qualitativelysimilar lower crust as well as being finer grained. The seismic
to Figures3B and 4B, exceptthat the H20 contentat any velocity signatureof denserock shouldbe evidentfirst in
chosenP andT is higher.
the upper crust and may be substantiallyretardedin the
•0•
344 ECLOGITE FORMATION IN OCEANIC CRUST
lower crust.The absenceof fast velocitiesat eclogitefacies
conditions cannot be taken to indicate the absence of
oceaniccrust.High densitieswill appearfirst in the upper
crustand may be considerablysuppressed
in the lower
crust.In Figure4, thelowercrustis notonlylessdensethan
asthenosphere,
but alsolessdensethanthe basalticupper
crust.In thesesituationsthe uppercrustmay sink in the
asthenosphere
butthelowercrustcannot.Moreover,Kirby
et al. [1996] calculated that densification reactions in
Kirby and Hacker[1991] proposeda link betweenarc
volcanism,intermediatedepthearthquakes,
and eclogite
formation.They notedthathypocenters
of subduction
zone
earthquakes
in the depthrange50-250 km frequently
coincide with arc volcanoes,and proposedthat these
intermediate-depth
earthquakes
mightbecaused
by delayed
eclogiteformation
withinsubducting
crust.Thishypothesis
requiresthat eclogitedoesnot form at shallower
depths
becauselow temperatures
and low H20 activityhinder
subducting
crustcan placethe crustin deviatorictension
andthe mantlelithospherein deviatoriccompression.
The
transformation,and that the devolatilizationof amphibole
differencein reactionrate betweenbasaltandgabbromeans
that their calculationsare oversimplified.The initiationof
densificationin the volcanicrockswill initially placethose
rocksin tensionandthe underlyingcoarsergrainedrocksin
and releasesvolatilesthat lead to the overlyingvolcanoes.
compression.
As reactionprogresses
intothecoarserlower
crustmustbe essentiallyanhydrous.
Calculationsand conclusionspresentedabove apply
crust,the stateof deviatoricstressin the gabbroicrockswill
changefromcompressive
to tensile.If therewereno other
forcesat work in subductionzones,earthquakes
at shallow
depthswouldbe extensional
in basaltandcontractional
in
in the 50-250 km depthrangetriggerseclogiteformation
Figure2 suggests
thattransformation
to eclogiteoccursat
much shallowerdepthsin most subductionzones.For
KirbyandHacker'shypothesis
to becorrect,
thesubducting
principallyto normalmid-ocean
ridgecrust.Oceancrust
that is anomalously
thickor of unusualcomposition
will
transformdifferently.Oceanicplateaus,
whichareup to 6
gabbro,changingasthe rocksdescenddeeperin subduction times thicker than mid-oceancrust, but compositionally
similar, will have somewhat retarded transformation
zonesto extensionalthroughoutthe oceaniccrust.
Little is known about the mechanical
behavior of
becauseof the additionaltime requiredfor the subducting
materialto heat;this may be partly offsethoweverby the
blueschistfaciesminerals,but clinopyroxeneandgarnetare
of volcanicrocks,whichwill transform
two of the strongestphasesin the crust[Ji and Martignole, greaterthickness
morerapidly.Field observations
discussed
aboveindicate
1994; Kolle and Blacic, 1983] and eclogite should be
that Fe-rich rockstransformnotablyfasterthanMgO-rich
strongcomparedto alteredbasalt.The replacement
of finecentersoften
grainedvolcanic phasesby blueschistor eclogitefacies rocks.Hydrothermalalterationat spreading
markedMgO enrichment
in lavas[Humphris
and
mineralsmay resultin strengthening,
whereasthe alteration produces
Thompson,1978], which would be expectedto slow
of coarse-grainedminerals in gabbro to blueschistor
transformation.Oceanicislandsbuilt by hotspotsandcrust
eclogitephasesmay enhancedeformation[Rubie, 1983].
rifts are bothenrichedin Fe relativeto
The volcanic layer, thoughhotter, may actuallybecome from propagating
strongerthan the gabbroiclayer as a result of eclogite MORB, andwill transformmorerapidlythanstandardmidocean-ridgecrust.Crust eruptedat fast spreadingridges
formation. It is conceivablethat suchrheologicallayering
could result in a downward jump of the subduction (>60 mm/a) is alsoricher in Fe than crusteruptedat slow
spreading
(<50 mm/a)ridges[Niu andBatiza,1993;Sinton
decollement from the volcanic into the gabbroic layer,
resultingin underplatingof eclogitizeduppercrustto the and Detrick, 1992], and is expectedto transformmore
rapidly.
mantle hanging wall and continued subductionof the
CONCLUSIONS
gabbroiclayer.
Crustthat undergoesslowtransformation
is lesscapable
The transformation of mafic rocks to blueschist and
of containing H20. Figure 4 shows that a slowly
zonesis complex,encompassing
a
transforminggabbrolayer has a maximum H20 content eclogitein subduction
and grain
-1/2 thatof the equilibriummodel.The flux of H20 carried wide varietyof igneousmineralassemblages
intothemantle
beyond
adepth
of70kmis 1.3x 108grams sizes,variablydevelopedandequilibratedmetamorphic
permeterper yearin theequilibriummodel,butonly 0.48 x
108grams
permeterperyearin thekineticmodelwith
unalteredprotolith.Note thatif the only sourceof H20 in a
subductionzone is the igneouscrust, situationssuch as
Figure4C, wherethe H20 contentincreases
with depth,are
impossible.Only if there is anothersourceof H20 in the
mineralassemblages,
all pressurized
andheatedvia a broad
rangeof P T paths.Existingexperimental
kineticdata
cannot
be extrapolated
to actualpet•rotectonic
settings
in
orderto predictwhererockstransformto blueschist
and
eclogite.However,texturaldatafromexhumed
subduction
zones worldwide indicate that little transformation occurs at
temperatures
below150øC.Volcanicrocksarecompletely
by perhaps250øC.Coarsergabbroicrocks
the H20 content of igneouscrust increasein subduction transformed
zones.
rarelyavoidcompleteeclogitization
at temperatures
above
subductionzone, from subductedsedimentsor mantle, can
HACKER
550øC,although
examples
of incomplete
transformation
are
knownfromrocksheatedat temperatures
ashighas800øC.
345
doteandassociated
mineralsin low grademetamorphic
rocks,
Contributions
toMineralogy
andPetrology,
64, 123-136,1977.
Brown,E.H., andE.D. Ghent,Mineralogy
andphaserelations
in
The rapidtransformation
of volcanicrocksandthe
the blueschistfacies of the Black Butte and Ball Rock areas,
metastable
persistence
of gabbroic
rocksintotheblueschist NorthernCaliforniaCoastRanges,
American
Mineralogist,
68,
365-372, 1983.
andeclogitestabilityfieldshasseveral
implications.
The
P.,Cryptic
variation
withinthecumulate
sequence
of
seismic
velocitysignature
of dense
rocksshould
beevident Browning,
the
Oman
ophiolite:
magma
chamber
depth
and
petrological
firstin theuppercrustandmaybesubstantially
retarded
in
implications,
in Ophiolites
andoceanic
lithosphere,
Geological
the lower crust.The buoyancyforcescausingthe slabto
Society
Special
Publications,
edited
byI.G.Gass,
S.J.Lippard,
sink will appearfirst in the uppercrustand may be
and A.W. Shelton,
pp.71-82,Geological
Societyof London,
London, 1984.
considerablysuppressed
in the lower crust.Earlier
Opticaldetermination
of
formation
of garnetin theuppercrustmaycausetheupper Carlson,W.D., andJ.L. Rosenfeld,
topotactic
aragonite-calcite
growth
kinetics:
metamorphic
imcrustto becomestronger
thanthegabbroic
layer,andmay
plications,
Journal
ofGeology,
89,615--638,
1981.
lead to underplating
of eclogitizeduppercrustand Carmichael,
R.S.,CRCpracticalhandbook
ofphysical
properties
continuedsubductionof the lower crust.The initiation of
of rocksandminerals,
741pp.,CRCPress,
BocaRaton,FL,
densificationin the volcanicrockswill initially placethose
rocks in deviatoric tension and the underlying coarser
1989.
Cloos,
M., Lithospheric
buoyancy
andcollisional
orogenesis:
subductionof oceanicplateaus,
continental
margins,
islandarcs,
grained
rocksin compression.
As reactions
progress
into
spreading
ridges,and seamounts,
Geological
Societyof
the coarserlower crust,the stateof stressin the gabbroic
America Bulletin, 105, 715-737, 1993.
rockswill change
fromcompressive
totensile.
If therewere Coleman,R.G., D.E. Lee, L.B. Beatty,andW.W. Brannock,
no otherforcesat work in subductionzones,earthquakes
at
Eclogites
andeclogites--Their
differences
andsimilarities,
shallowdepthswouldbe extensional
in the basaltand
contractional
in the gabbro,changingat deeperlevelsto
extensional
throughout
the crust.All theseeffectswill be
minimizedin Fe-richcrustproduced
at oceanicislandsand
Geol. Soc.AmericaBull., 76, 483-508, 1965.
Dick,H.J.B.,P.S.Meyer,S.H.Bloomer,
S.H.Kirby,D.S.Stakes,
andC. Mawer,Lithostratigraphic
evolution
of anin-situsection
of oceanic
layer3, Proceedings
oftheOceanDrillingProgram,
Scientific
Results,
118,439-538,1991.
A.E.-D.,R.G.Coleman,
andJ.G.Liou,Eclogites
and
fastspreading
ridgesandmostpronounced
in Mg-richcrust E1-Shazly,
formedat slow spreadingcenters.
blueschists
fromnortheastern
Oman:petrologyandP-T evolution,Journalof Petrology,
31,629--666,1990.
tectonics
andthe
Acknowledgments.
Reviewed
by S.H.Bloomer,S.R.Bohlen,S. - Ellis,D.J.,andM.A.H.Maboko,Precambrian
DeBari,W.G. Ernst,J.G. Liou. A particularlythoroughreview
wasgivenby J.H. Natland.
physicochemical
evolution
ofthecontinental
crust.
I. Thegabbro-eclogite
transition
revisited,
Precambrian
Research,
55,
491-506, 1992.
REFERENCES
Ernewein,
M., C. Pflumio,andH. Whitechurch,
Thedeathof an
accretionzoneas evidenced
by the magmatichistoryof the
Sumail
ophiolite
(Oman),
Tectonophysics,
151,247-274,
1988.
of glaucophane
schists,
J. Petrology,
4,
itsgeophysical
significance,
Reviews
of Geophysics
andSpace Ernst,W.G.,Petrogenesis
Ahrens,T.J., and G. Schubert,Gabbro-eclogite
reactionrateand
Physics,13,383-400, 1975.
1-30, 1963.
B.W.,Phase
relations
of epidote-blueschists,
Lithos,
25,3Alt, J.C., T.F. Anderson,L. Bonnell,and K. Muehlenbachs, Evans,
23, 1990.
Mineralogy,
chemistry
andstableisotopic
compositions
of hyandpetrology
of Tertiary-Recent
orodrothermally
alteredsheeted
dikes;ODPHole504B,Leg 111, Ewart,A., Themineralogy
genicvolcanic
rocks:
withspecial
reference
to theandesiteProceedings
of theOceanDrillingProgram,Scientific
Results,
basaltic
andesire
compositional
range,in Andesites:
orogenic
111, 27-40, 1989.
Alt, J.C., J. Honnorez, C. Laverne, and R. Emmermann,
andesites
and relatedrocks,editedby R.S.Thorpe,pp. 26-87,
Wiley, Chichester,1982.
Hydrothermal
alteration
of a 1-kmsection
through
theupper
processes
in oceanic
ridgeandintraplate
oceaniccrust,DeepSeaDrillingProjectHole504B;mineral- Flower,M., Magmatic
in Oceanic
basalts,
edited
byP.A.Floyd,pp.116-147,
ogy,chemistry,
andevolution
of seawater-basalt
interactions, settings,
BlackieandSon,Glasgow,1991.
Journalof Geophysical
Research,
91, 10,309-10,335,
1986.
An experimental
investigation
Batiza,R., PacificOceancrust,in Oceanicbasalts,
editedby P.A. Green,D.H., andA.E. Ringwood,
ofthegabbro-eclogite
transformation
anditspetrological
impliFloyd,pp.264-288,BlackieandSon,Glasgow,
1991.
cations,
Geochimica
Cosmochimica
etActa,31,767-833,1967.
Bohlen,S.R., V.J. Wall, andA.L. Boettcher,
Geobarometry
in
in theformation
of metagranulites,
in Kineticsandequilibrium
in mineralreactions, Hacker,B.R.,The roleof deformation
morphic
field
gradients:
Ridge
subduction
beneath
the
Oman
editedby S.K. Saxena,pp. 141-171,Springer-Verlag,
New
ophiolite,
Tectonics,
10,455-473,1991.
York, 1983.
Boundy,
T.M., D.M. Fountain,
andH. Austrheim,
Structural
de- Hacker,B.R., S.R. Bohlen,andS.H. Kirby,Albite-->jadeite+
quartztransformation
in albitite,
Eos,Transactions
American
velopment
andpetrofabrics
of eclogitefaciesshearzones,
Geophysical
Union,74,611, 1993.
Bergen
arcs,western
Norway;
implications
fordeepcrustal
deformational
processes,
Journalof Metamorphic
Geology,
10, Hacker,B.R., S.H. Kirby,andS.R.Bohlen,Time andmetamorphicpetrology:
thecalcite
-->aragonite
transformation,
Science,
127-146, 1992.
258, 110-113, 1992.
Brown,E.H.,Phase
equilibria
among
pumpellyite,
lawsonite,
epi-
346
ECLOGITE
FORMATION
IN OCEANIC
CRUST
Hekinian, R., Petrology of the oceanfloor, 393 pp., Elsevier,
Amsterdam, 1982.
Holland, T.J.B., The reactionalbite = jadeitc + quartzdetermined
experimentally in the range 600-1200øC, American
Mineralogist, 65, 129-134, 1980.
Hopson, C.A., R.G. Coleman, R.T. Gregory, J.S. Pallister, and
E.H. Bailey, Geologicsectionthroughthe SamailOphioliteand
associatedrocks along a Miscat-Ibra transect, southeastern
Oman Mountains,Journal of GeophysicalResearch,86, 25272544, 1981.
Humphris, S.E., and G. Thompson, Hydrothermal alteration of
oceanic basalts by seawater, Geochimica Cosmochimicaet
Acta, 42, 107-125, 1978.
Indares,A., and T. Rivers, Textures, metamorphicreactionsand
thermobarometry
of eclo•itizedmetagabbros;
a Proterozoic
example,EuropeanJournalof Mineralogy,7, 43-56, 1995.
Ito, K., and G.C. Kennedy,An experimentalstudyof the basaltgarnet granulite-eclogite transition, in The Structure and
Physical Properties of the Earth's Crust, edited by J.G.
Heacock, pp. 303-314, American Geophysical Union,
WashingtonD.C., 1971.
Ji, S., and J. Martignole,Ductility of garnetas an indicatorof extremely high temperaturedeformation,Journal of Structural
Geology,16, 985-996, 1994.
Juteau,T., M. Beurrier,R. Dahl, andP. Nehlig, Segmentation
at a
fossil spreadingaxis: the plutonic sequenceof the Wadi
Haymiliyah area (Haylayn block, Sumail nappe, Oman),
Tectonophysics,
1881, 167-197, 1988a.
Juteau, T., M. Ernewein, I. Reuber, H. Whitechurch, and R. Dahl,
Duality of magmatismin the plutonicsequenceof the Sumall
nappe,Oman, Tectonophysics,
151, 107-136, 1988b.
Kirby, S.H., E.R. Engdahl, and R. Denlinger, Intraslab earthquakesand arc volcanism:dual physicalexpressionsof crustal
and uppermostmantle metamorphismin subductingslabs,in
Dynamics of Subduction,GeophysicalMonograph, edited by
G.E. Bebout,D. Scholl,and S. Kirby, AGU, Washington,D.C.,
1996.
Kirby, S.H., and B.R. Hacker, Intermediate-depthearthquakes,
crustal phase changesand the roots of arc volcanoes,Eos,
Transactions
AmericanGeophysicalUnion,72, 481, 1991.
Klaper, E.M., Reaction-enhanced
formation of eclogite-facies
shear zones in granulite-faciesanorthosites,in Deformation
Mechanisms,Rheologyand Tectonics,GeologicalSocietyof
London Special Publication, edited by R.J. Knipe, and E.H.
Rutter,pp. 167-173, London,1990.
Kolle, J.J., and J.D. Blacic, Deformationof single-crystalclinopyroxenes;2, Dislocation-controlledflow processesin hedenbergite, Journalof GeophysicalResearch,88, 2381-2393, 1983.
Lardeaux,J.-M., and M.I. Spalla, From granulitesto eclogitesin
the SesiaZone (Italian WesternAlps); a recordof the opening
and closure of the Piedmont ocean, Journal of Metamorphic
Geology,9, 35-59, 1991.
Liou, J.G., P-T stabilities of laumontite, wairakite, lawsonitc, and
relatedmineralsin the systemCaA12Si208-SiO2-H20, Journal
of Petrology, 12,379-411, 1971.
Lippard, S.J., A.W. Shelton, and I.G. Gass, The Ophiolites of
Northern Oman, 178 pp., 1986.
Maresch,W.V., Experimentalstudieson glaucophane;
ananalysis
of presentknowledge,Tectonophysics,
43, 109-125,1977.
M0rk, M.B.E., A gabbro to eclogite transition on Flems0y,
Sunnm0re,westernNorway, ChemicalGeology,50, 283-310,
1985.
Niu, Y., and R. Batiza, Chemical variation trendsat fast and slow
spreadingmid-oceanridges,Journalof Geophysical
Research,
98, 7887-7902, 1993.
Oh, C.W., andJ.G. Liou, Metamorphicevolutionof two different
eclogitesin the FranciscanComplex,California,USA, Lithos,
25, 41-53, 1990.
Pallister,J.S., and C.A. Hopson,Samailophioliteplutonicsuite:
field relations,phasevariation,crypticvariationandlayering,
and a modelof a spreadingridgemagmachamber,Journalof
GeophysicalResearch,86, 2673-2645, 1981.
Peacock,S.M., Numerical simulationof metamorphicpressuretemperature-timepaths and fluid productionin subducting
slabs, Tectonics, 9, 1197, 1990.
Peacock,S.M., The importanceof blueschist-• eclogitedehydration reactionsin subducting
oceaniccrust,GeologicalSocietyof
America Bulletin, 105, 684-694, 1993.
Pognante,U., Petrological constraintson the eclogite- and
blueschist-facies
metamorphismandP-T-t pathsin the Western
Alps,Journalof MetamorphicGeology,9, 5-17, 1991.
Poli, S., The amphibolite-eclogite
transformation;
anexperimental
studyon basalt,AmericanJournalof Science,293, 1061-1107,
1993.
Reynard,B., andM. Balleyre,Coexistingamphiboles
in aneclogitc from the WesternAlps; new constraintson the miscibility
gap between sodic and calcic amphiboles, Journal of
MetamorphicGeology,6, 333-350, 1988.
Ridley,J., andJ.E. Dixon, Reactionpathwaysduringtheprogressive deformation of a blueschist metabasite; the role of chemi-
cal disequilibrium
andrestrictedrangeequilibrium,Journalof
MetamorphicGeology,2, 115-128, 1984.
Robinson, P.T., H.J.B. Dick, V. Herzen, and R. Pierre,
Metamorphismand alterationin oceaniclayer 3; Hole 735B,
Proceedingsof the OceanDrilling Program,ScientificResults,
118, 541-562, 1991.
Rubie, D.C., Reaction-enhanced
ductility:the role of solid-solid
univariant reactions in deformation
of the crust and mantle,
Tectonophysics,
96, 331-352, 1983.
Rubie,D.C., Role of kineticsin the formationandpreservation
of
eclogites,in EclogiteFaciesRocks,editedby D.A. Carswell,
pp. 111-140, Blackie,Glasgow,1990.
Sandrone,R., L. Leardi, P. Rossetti, and R. Compagnoni,P-T
conditionsfor the eclogiticre-equilibrationof the metaophiolites from Val d'Ala di Lanzo (internal Piemontese zone,
WesternAlps), Journalof MetamorphicGeology,4, 161-178,
1986.
Schliestedt,
M., Eclogite-blueschist
relationships
asevidencedby
mineral equilibria in the high-pressuremetabasicrocks of
Sifnos(CycladicIslands),Greece,Journal of ?errology,27,
1437-1459, 1986.
Sinton,J.M., andR.S. Detrick, Mid-oceanridgemagmachambers,
Journalof Geophysical
Research,97, 197-216,1992.
Zhang,R.Y., J.G.Liou, T.F. Yui, andD. Rumble,Transformation
Liu, J., W.G. Ernst,J.G. Liou, andS.R. Bohlen,Experimental of gabbroandgranitoidto coesite-bearing
eclogitefaciesrock
constraintson the amphibolite-eclogitetransition,Geological
from Yangkou,the SuluTerrane,easternChina,International
Societyof AmericaAbstractswith Programs,25, 213, 1993.
GeologicalCongress,in press,1995.
Liu, M., and R.A. Yund, Transformation kinetics of polycrystalline aragoniteto calcite: new experimentaldata, modelling,
and implications,Contributionsto Mineralogyand Petrology,
114,465-478,
1993.
B. R. Hacker, Departmentof Geologicaland Environmental
Sciences,StanfordUniversity,Stanford,California94305-2115.
(e-mail: [email protected])