Transcript Fluid Processes in Subduction Zones Hydrous Minerals and

Fluid Processes in Subduction Zones
Hydrous Minerals and
Dehydration Reactions
Simon M. Peacock
Dept. of Earth and Ocean Sciences
University of British Columbia
Effects of H2O on Subduction Zones
• H2O and hydrous minerals weaken plate
interface permitting subduction and plate
tectonics to occur
• H2O lowers melting temperature of mantle,
generates explosive arc magmas
• Fluids released by slab dehydration promote
brittle behavior and may trigger earthquakes
• Hydration structure, rheological structure, and
thermal structure of subduction zones are
strongly coupled
Selected Characteristics of Volcanic Arcs
• Depth to underlying Benioff zone is ~100 – 200 km
• Depth to underlying Benioff zone beneath volcanic
front is 124± 38 km (Gill, 1981)
– suggestive of P-dependent melting or slab
dehydration rxn
– or may be necessary depth for slab in order to be
overlain by hot asthenosphere such that addition
of H2O triggers melting
• Several places where active arc is missing (or feeble)
– Peru, central Chile, SW Japan
• T eruptions – 1050-1100°C
• T mantle equil – commonly 1300°C+ for basalts
Geochemistry of Arc Magmas
Compared to MORBs, arc magmas are:
• Fractionated
•
basalts andesites dacites rhyolites
• Wet
• Enriched in many trace elements, such as
Rb, that appear to be derived from the
subducting slab
H2O Content of Arc Magmas
• Significantly more H2O than MORBs, OIBs
• Explosive eruptions (more H2O degassing,
higher SiO2 content)
• Hornblende (amphibole) is common
phenocryst
• H2O dramatically lowers melting T of rocks,
mantle by 100s of degrees.
 key to arc magma genesis
H2O measurements of volcanic glasses
• For submarine basaltic glasses, we
can measure H2O contents directly
because ocean pressure prevents
H2O from exsolving.
• For subaerially erupted glasses,
measure glass inclusions.
MORB glasses < 0.5 wt % H2O
Back-arc basin glasses
0 – 2.5 wt % (ave = 1.1 wt %) H2O
Arc glass inclusions
0 – 6 wt % (ave = 3.4 wt%) H2O
Stern (2002)
Arc Magmas
Distinctive Trace Element Characteristics
• Arc basalts, as compared to MORBs, are
– Enriched in large-ion lithophile elements (LILE)
K, Rb, Cs, Sr, Ba, and Pb, U, B, Be
– Depleted in high field strength elements (HFSE)
Y, Zr, Hf, HREE, Nb, Ta (controversial niobium anomaly: Ti-phase
in source?, prev. melting event?)
• H2O-rich fluids will preferentially contain more mobile
LILE – derived from slab
• Be and Th are relatively enriched in arc basalts, but
believed to be relatively insoluble in H2O-rich fluids
 suggests sediment melting
Variably hydrated
oceanic crust and
mantle
Zandt (2002)
Hydration of Oceanic Crust & Mantle
Hydrothermal circulation at mid-ocean ridges
• Black smokers
• Adds H2O, CO2 to crust
• Basalt, gabbro  hi T, lo P minerals (amph, chl, epi)
• Believed to be limited to oceanic crust, but could go
deeper at slow-spreading ridges
• Sea-floor weathering
• Low T alteration
Likely additional hydration at trench - outer-rise, including
serpentinization of uppermost mantle
Global Subduction H2O Fluxes
Input fluxes:
Pore H2O
Sediments
Oceanic crust
[1012 kg/yr]
1
0.1
Chemically bound H2O
Sediments
0.1
Oceanic crust
1 to 2
Oceanic mantle
0.1 to 1
Expelled by
porosity collapse
Expelled by
metamorphic
dehydration rxns
Output flux to arc magmas: 0.1 to 0.4 (5-20% of input flux)
Mass of hydrosphere = 1.4 x 1021 kg
• Oceanic crust:
basalts = 1-3 wt % H2O
gabbros = 0.5-1 wt % H2O
• High-pressure metabasalts
contain more bound H2O
than altered basalts
recovered in DSDP/ODP drill
core.
Peacock (2004)
Porosity collapse:
Sediment compaction
Basalts ~ 200-400 °C
Slab dehydration reactions
Temperature-dependent
Hyndman and Peacock (1999)
Thermal structure of a subduction zone determines
where H2O is released from slab and where hydrous
minerals may be stable in the overlying plate.
Thermal structure of subduction zones
• Subducting slab drags down isotherms
• Inverted isotherms beneath mantle wedge
• Mantle-wedge isotherms parallel to flow lines
• Tinterface ~ 0.5 Tmantle
Cool vs. Warm Subduction Zones
after Peacock and Wang (1999)
In warm subduction
zones, H2O is liberated
from the slab by
metamorphic
dehydration reactions
and possibly by the
collapse of porosity in
the upper crust.
The amount of H2O
released is predicted to
be small:
0.1 x 10-3 m3 / (m2 yr) =
100 milliliters of H2O per
m2 column per year
Hyndman and Peacock (1999)
Metamorphic facies
– diagnostic mineral assemblage
indicative of region in P-T space
(particularly used in mafic systems)
Facies boundaries – complex reactions
Dehydration of subducting oceanic crust
blueschist
eclogite
• Metabasalt dehydration reactions are generally continuous
reactions, which are smeared out in P-T space
• Metabasalt  eclogite rxns release large amounts of H2O,
increase density, and increase seismic velocity
Crust subducted in warm
subduction zones
passes through
greenschist  epidote
bluseschist/greenschist
eclogite facies.
And transformation of
metabasalt to eclogite
(garnet + cpx) occurs at
~50 km.
Thermal models:
Peacock and Wang (1999)
Currie et al. (2000)
Peacock et al. (2002)
Cascadia (central Oregon)
• Serpentinized forearc mantle
• Metabasalt --> eclogite
Bostock et al. (2002)
For a given P, T, and composition (X),
what is the mineralogy of the rock?
• P=rgz
• T = thermal models
• X = bulk composition, including H2O
content - fully hydrated or anhydrous?
Phase diagram for SiO2
coesite
~2.5 GPa
P
quartz
T
qtz = coes
DGrxn = 0
dP/dT = DS/DV
Clapeyron slope
Higher pressure
 higher r
Higher temperature
 higher S
Dehydration Reactions
dP/dT = DS/DV = + / - = negative
because H2O is compressible
P
serpentine
olivine +
orthopyroxene +
H2O
dP/dT = DS/DV = + / + = positive
T
Dehydration rxns:
In nature, many reactions are “smeared
out” in P-T space because of:
• Solid solutions – particularly in mafic rocks
(e.g., Fe-Mg)
• Variable fluid compositions
• Kinetics
Dehydration rxns (cont.):
Phase diagrams can be based on :
• Experiments (sluggish at T < 600°C)
• Thermodynamic calculations (must know
all phases, solid solutions)
• Field-based petrologic studies
(observations + P-T-ometry)
• Combinations of all 3
Experimentally
determined phase
relations in H2O
saturated MORB
Schmidt +
Poli (1998)
Maximum H2O
contents bound in
hydrous minerals in
MORB
Schmidt + Poli (1998)
Schmidt + Poli (1998)
Phase diagram for H2O-saturated
peridotite and maximum H2O content
Stern (2002) based on Schmidt and Poli (1998)
Important points of Schmidt & Poli (1998)
• Numerous hydrous minerals stable in
subduction zones. Amphibole is not the key
• Dehydration of slab is ~continuous due to
smeared out P-T rxns + isotherm / isobaric
structure
• Volcanic front controlled by wedge isotherm
(not a specific dehydration rxn)
• Serpentine  phase A may transport H2O to
great depth in cool s.z.
• Distribution + amount of H2O incoming
lithosphere is critical unknown
Phase Diagram for Metabasalt
Based on field
observations and thermo
calculations
Facies boundaries are
broader than depicted
P
Different parts of
subducted crust intersect
facies boundaries at
different places
T
Hacker et al. (2003)
Hacker et al.
(2003)
Hacker et al.
(2003, JGR)
Hacker et al.
(2003, JGR)
Hacker et al.
(2003, JGR)
Where does H2O expelled from the subducting slab go?
mid-ocean ridge
trench
outer rise
1
2
oceanic
crust
?
crust
3
oceanic
mantle
mantle
wedge
4
50
km
approximate
scale
0
5
0
50 km
(1) Updip fluid flow along faults to seafloor/surface
(2) Incorporated into forearc crust (hydration)
(3) Incorporated into forearc mantle (hydration)
(4) Incorporated into arc magmas
(5) Subducted past volcanic arc
Zandt (2002)
Example of retrograde metamorphic reaction:
Serpentinization of the forearc mantle wedge
Mg2SiO4 + MgSiO3 + 2H2O = Mg3Si2O5(OH)4
olivine
pyroxene
fluid
serpentine
(13 wt% H2O)
Serpentine mud volcanoes in
Mariana forearc (Fryer et al.)
Evidence for Serpentinized
Forearc Mantle
• Active serpentine mud volcanoes (Mariana
forearc)
• Hydrated ultramafic hanging walls of
paleosubduction zones
• Low seismic velocities, high Poisson’s ratios
observed in forearc mantle (Alaska,
Aleutians, Andes, Cascadia, Costa Rica, IzuBonin, Japan, Mariana)
Consequences of serpentinized forearc mantle
Zandt (2002)
• “Weak” serpentinite may control downdip limit of seismogenic
zone and reduce mechanical coupling b/w slab and wedge
• Buoyancy will tend to isolate forearc wedge from corner flow
• Heating of hydrated forearc mantle (e.g., ridge subduction,
post-subduction) will release significant amounts of H2O
Cascadia (central Oregon)
Bostock et al. (2002, Nature)
Isoviscous vs. Olivine Rheology
van Keken, Kiefer, and Peacock (2002)
Mantle Wedge and Interface Rheology
Isoviscous wedge
Non-Newtonian wedge
Abers et al. (2006)
Non-Newtonian
Cold nose (0% coupling)
Non-Newtonian
Cold nose (10% coupling)
Ranero et al. (2005)
Zandt (2002)
The End