Transcript Slide 1
Petrologic processes that generate the intermediate to felsic plutonic core of island arcs
Susan DeBari & Michael Johnsen, Department of Geology, Western Washington University, Bellingham, WA
In exhumed arc sections worldwide, the upper mid-crust is composed dominantly of
hornblende-bearing tonalite, quartz diorite, diorite, and gabbro (49-76 wt.% SiO2) whose
compositions would correspond to an in situ Vp in the range of 6.0-6.3 km/s. This is in
contrast to a more mafic, cumulate lower crust composed dominantly of two-pyroxene
gabbro (±hornblende, ±garnet) and pyroxenite, (43-52 wt.% SiO2) whose compositions
would correspond to an in situ Vp ~7.0 km/s. This grossly simplified crustal structure is
surprisingly similar to many modern arcs whose seismic velocity structures have been
determined (IBM, Tonga, Kurile, Aleutians, North Honshu, Cascades). In all of these
modern arcs, an upper mid-crust with Vp 6.0-6.5 km/s is present, corresponding to velocities
calculated for exhumed arc upper mid-crust lithologies. If we presume that modern arcs and
exhumed arcs all contain upper mid-crust with intermediate to felsic plutonic rocks (an
unsubductable nucleus), we must be able to model how these rocks are generated.
In general, we have discerned two chemically distinct groups of tonalite/diorite in the
exhumed arc sections. The first compositional group (Type I) typically has flat to slightly
LREE enriched rare earth element patterns where REE abundances increase with increasing
SiO2. The second compositional group (Type II) shows trends of LREE enrichment and
HREE depletion, where both LREE and HREE abundances decrease with increasing SiO2.
They are also depleted in Y and enriched in Sr. The more felsic members of this group
generally exhibit concave-up patterns of HREE depletion. Most exhumed arcs show one or
the other of these trends, but some, including Talkeetna, show both, but at different times in
the arc’s history.
In the Talkeetna arc, least squares calculations and REE Rayleigh fractionation modeling
indicate that Type I tonalite/diorite (55-76 wt.% SiO2) form via fractional crystallization
from basalt to dacite. Type II tonalite/diorite (56-74 wt.% SiO2) must be produced by more
complicated means that involve some component of cannibalization of lower crust, either by
partial melting, or by assimilation. Type II tonalites in the Talkeetna arc can be effectively
modeled as a result of magma mixing between an andesitic parental liquid (presumably
formed by fractional crystallization) and felsic partial melts of hornblende-bearing mafic
rock (amphibolite, hornblende gabbro cumulates). In the Talkeentna arc, these Type II rocks
post-date the Type I rocks, and were formed after the arc had matured and (presumably)
thickened.
These mechanisms provide a testable hypothesis for modern arcs. If the arcs are
relatively young and thin, then tonalite/diorite should have geochemical characteristics of
Type I (fractionation only). If the arcs are more mature and thicker, then tonalite/diorite may
have geochemical characteristics of Type II (some component of lower crustal melting).
S. Coast Plutonic Cplx
(Washington)
Kohistan arc
(Pakistan)
Jurassic x-section
Jurassic x-section
Cretaceous x-section
Cretaceous x-section
granitoid
batholith
+ mingled
mafic
+ mingled
mafic
Interlayered
intermediate
plutons
5 kb amphibolite
(~15 km) crustal melting
• Intermediate plutonic
layer
• No obvious exposure of
crustal melting
• Crustal melting
• Crustal melting
• Crustal melting
• Older oceanic basement
• Older oceanic basement
• Older oceanic basement
• Unknown bulk
composition
• Int. bulk composition to
~30 km depth
• Unknown bulk
composition
• No oceanic basement
• Mafic bulk composition
These sections have been color-coded to their expected seismic velocities based on
lithology (velocities calculated using formulation of Behn and Kelemen, 2003)
6
REE/Chondrite
10
Type 1
(65% SiO2)
92-274
66% SiO2
(69% SiO2)
Type 2
15
120
[La/Yb]N
concentration / C1 chondrite
La Ce
Sr/Y
1.1
1.0
0.9
0.8
20% partial melt
of 5710J10B
compared to
5711J01
92-274
66% SiO2
Type 2
40
20
0
0
50
55
60
65
70
75
45
50
SiO2 (wt %)
55
60
65
70
75
plutonic samples (observed)
volcanic samples (observed)
modeled REE (mixing)
interm. end-member (5721J03)
felsic end-member (5711J01)
(69% SiO2)
Decreasing
SiO2
Lu
Yb
Ho
Dy
Tb
Gd
Eu
Sm
Nd
Ce
37c (melanosome)
10
10
Type 2 (dacite)
37c - 'residue'
1 - leucocratic diorite
model results
1
La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Yb Lu
10
90
% of felsic
end-member
Felsic
end-member
(5711J01)
71.2 wt
% SiO2
Nd Sm Eu Gd Tb Dy Ho Er Yb Lu
20% 5721J03
80% 5711J01
compared
65.8 wt
SiO2
to%5712J07
1.1
0.9
0.7
0.5
La Ce
Pr
Nd Sm Eu Gd Tb Dy Ho Er Yb Lu
Concluding hypothesis:
35% amphibole
45% plagioclase
10% magnetite
10% clinopyroxene
~25% melting
1
Pr
Intermediate
end-member
(5721J03)
RESIDUE
60
5
10
60.4 wt
% SiO2
In the Talkeetna Arc, the older plutonic rocks are Type 1. The
youngest rocks are Type 2 (thicker, more mature crust?)
Batch modal melting
100
80
Type 2 also have
high Sr/Y with
La
Ce
Pr
Nd
Sm
Eu
Gd
Tb
Dy
1
Pr Nd Sm Eu Gd Tb Dy Ho Er Yb Lu
Example 2: Bonanza arc (field exposure of leucosome &
melanosome)
100
10
Pr Nd Sm Eu Gd Tb Dy Ho Er Yb Lu
Mafic lavas that
backmix with Type 2)
(65% SiO2)
1
REE normalized to C1 chondrite
(Sun & McDonough, 1989)
~20% fractional melting
Felsic
end-member
(5711J01)
71.2 wt% SiO2
1
100
10
Lu
Yb
Ho
Dy
Tb
Gd
Eu
Sm
1
>54 wt.% SiO2
(diorite/tonalite)
Increasing
SiO2
REE/Chondrite
Increasing
SiO2
Nd
REE/Chondrite
REE/Chondrite
A Cascade volcano with a
similar pattern: Glacier
Peak volcano
(56% SiO2)
1
Residual
Cpx 10%
Opx 5%
Plag 45%
Amphibole 35%
Magnetite 5%
100
La Ce
Diorites, Tonalites
& Leucotonalites
(65% SiO2)
(60% SiO2)
(58% SiO2)
10
10
Example: Talkeetna arc (see Johnsen et al poster)
44.8 wt% SiO2
Gabbroic
composition
(5710J10B)
Cc/Cm
100
(56% SiO2)
Bonanza arc
migmatite
plutonic samples (observed)
volcanic samples (observed)
modeled REE (melting)
gabbroic rock (5710J10B)
felsic end-member (5711J01)
Bonanza Arc
La
Kohistan arc
migmatite
7
Mixing of the crustal melts with
mantle-derived magmas can
produce the Type 2 trend of
decreasing REE with increasing
SiO2
100
La Ce
Diorites, Tonalites
& Leucotonalites
Type 1
(60% SiO2)
(58% SiO2)
Produce a Si-rich, Type 2
magma by crustal melting
Example 1: Talkeetna arc (see Johnsen et al poster)
Cc/Cm
• Type 2 - REE abundances decrease with increasing SiO2 content (and decreasing Mg#).
This cannot be modeled as fractional crystallization (even taking into account observed
quantities of apatite). This is best modeled as melting of a distinct low LREE source
(cumulates?) coupled with mixing.
(65% SiO2)
Moho
pyroxenite,
dunite
+/- garnet
• Intermediate to >30 km
(10 kbar)!
• Type 1 - REE abundances increase with increasing SiO2 content (and decreasing Mg#).
This can be easily modeled as fractional crystallization, typically involving cpx + plag +
amphibole + Fe-Ti oxide (see Johnsen et al. poster for Talkeetna detailed example).
Western Talkeetna Arc
10 kb
(~30 km)
• Intermediate plutonic
layer
The intermediate to felsic plutonic rocks of the mid crust display distinct REE patterns that fall
into two categories
100
10 kb
(~30 km)
• Intermediate plutonic
layer
How is the intermediate-felsic mid crust generated?
Using geochemical signatures to discern magmatic processes:
100
garnet
gabbro
Questionable
stratigraphy……
(bottom and top
may actually be
two sections
based on ages)
Crust at least
60 km thick
• What lithologies make up this upper crust? Can we make
generalizations based on arc crustal sections?
2. Crustal melting - snapshots of this process observed in the
Bonanza arc and the Kohistan arc in the mid-crust to upper parts of
the lower crust
?
tonalite, diorite
hornblende
gabbro
layered
2-px
gabbros
pyroxenite
duniteharzburgite
• Even Izu Bonin and Aleutians are not that different in the upper crust.
Close to homogenization?
?
2-px gabbro
amphibolite
crustal melting
Mafic plutonic
rocks
10 kb garnet gabbro
(~30 km)
Moho
• Modern arcs in a gross sense have similar seismic velocity structures
person for scale
5 kb
(~15 km)
Intermediate-felsic
plutonic
rocks
Uppermid
crust
Fine-scale mingling
and
Volcanic and
volcaniclastic
rocks
Upper
crust
5-~7 km
Shillington et al (2004)
5 kb
(~15 km)
metamorphic
rocks
(schist/
amphibolite)
layered
2-px
gabbro
Iwasaki
Nakanishi
et al.
et al.
(2001)
(2007)
Parsons et al. (1998)
volcanic
rocks
tonalite
diorite
5 kb
(~15 km)
Suyehiro et
al. (1996)
0
?
volcanic
rocks
?
5
45
0
0
tonalite
diorite
+mingled
mafic
Crawford et
al. (2003)
Depth
Depth
Depth
volcanic
rocks
Physical mixing of crystals
Mingling on the large scale
Bonanza arc
(Vancouver Island)
0
1. Magma mingling (and mixing) is pervasive in the middle crust in
all arc sections. Extreme heterogeneity at all scales.
Break-up and mingling of
basaltic sill
Talkeetna arc
(Alaska)
Depth
Ce
4
Common processes observed in the mid
crust of exposed arc sections
3
The crustal sections
concentration / C1 chondrite
2
ABSTRACT
La
1
Seismic stratigraphy of modern arcs - more
similarities than differences?
La Ce
Nd
Sm Eu Gd
Dy Ho
Yb Lu
Type 1: Increasing REE with increasing
SiO2. Process is fractionation (young
thin arcs?)
Type 2: Decreasing REE with increasing
SiO2. Process is crustal melting and is
often coupled with mixing with mantlederived magmas and fractionation (older
thicker arcs?)
Both processes produce the nonsubductable nucleus of continental
crust
La
Ce
Pr
Nd
Sm
Eu
Gd
Tb
Dy