major elements in continental crust like “calc-alkaline” arc andesite & dacite

major elements in continental crust like “calc-alkaline”arc andesite & dacite

Kelemen CMP 1995

.

BaTh

UK

TaNb

LaCe

PbPr

SrNd

ZrHf

SmEu

GdTb

TiDy

HoEr

TmYb

LuY

CoNi

0.1

1

10

100 Taylor & McL andesite modelTaylor & McL bulk crustTaylor & McL Archean crustWeaver & Tarney bulk crustRudnick & Fountain bulk crust

continental crust

Kelemen et al., 2001, figure 6A

Kelemen et al., AGU Mon 2003

compared to igneous protolith,

Sr removed by chemical weathering,

Zr, Ti, … enriched in detrital zircon & rutile

Kel

em

en e

t al

., A

GU

Mon

200

3

some island arcs have lavas ~ continental crust

^

igneous

3 recipes for continental crust

1. crystal fractionation of primitive basalt at high fO2 and/or H2O

plus “delamination” of dense stuff*

2. mixing of primitive and evolved igneous rocks

plus “delamination” of dense stuff*

3. crystal fractionation of primitive andesite

plus “delamination” of dense stuff*

*: or cumulates lie below Moho, or “relamination” of buoyant stuff

in this talk:

1. delamination?

2. use trace elements to choose 1-3

Region map

redrawn from DeBari & Coleman, 1989

Aleut I-B

Mar

Tonga

S Sand

Chugach Mtns

SE Talkeetna Mtns

NW Talkeetna Mtns

Talkeetna arc did not intrude pre-existing oceanic or continental crust

instead, like an ophiolite or oceanic crust, the arc was emplaced “into itself” Rioux et al, GSA Bull 2007

magma mixing, AFC, …

crystal fractionation & melting

numbers of samples,not necessarily representativeof proportions of rock types

crystal fractionation model of Greene et al., J. Petrol., 2006

. .

1500

1000

A A'

harzburgite

gabbronorite

garnet gabbro

pyroxeniteserp

melange

antigorite sch

A

A'145°10' 145°05'

61°35'

BERNARD MTN

1 km

gabbronorite

pyroxenite

harzburgite

serpentinitemélange

gabbronorite

BORDERRANGESFAULT

gabbronorite

30% xl’nCpx Mg# 92 to 85

Bernard map & crustal sectionfrom DeBari & Coleman, 1989

first 30% crystallization only represented by a few hundred meters of cumulate pyroxenite and gabbro near Moho

Delta rho cpx’te & arc lcJull & Kelemen, JGR 2001

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

Behn & Kelemen JGR 2006

D C BA

if you “like” A or B, then lots more magma mixing or strongly calc-alkaline crystal fractionation than we initially expected or modeled, or …

D C BA

A or B … substantially more SiO2-rich primitive magmas than we initially expected or modeled … maybe early in arc history? volcanics mostly eroded, plutonics largely preserved?

Region map

approxlength of SNB

approx length of Alaska Peninsula batholith

Takahashi et al Geology 2007

Behn & Kelemen JGR 2006

A B C D

1974

two stage distillation:

melt at ridge

melt again in subduction

enrich low melting pt components

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

MORB

cont crust

2% fluid

2% melt

A

M

arc/MORB enrichments via melt, not aqueous fluid

boninitesremoved

here

boninitesremoved

here

Kelemen et al Treatise on Geochemistry 2003

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

Peacock et al. PEPI 2005

pre-2000style

post-2000models

VERY simple geodynamic modelsWhy so sure?Geochemists say it’s right

VERY simple geochemical modelsWhy so sure?Geophysicists say it’s right

Kel

em

en e

t al

., T

reat

ise

on G

eoch

em 2

003

Hac

ker

et a

l., 2

004

Rho granulite & arc lcJull & Kelemen, JGR 2001

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

most models don’t satisfy constraintswhich basically require T’ at about 45 km

drips

Jull & Kelemen, JGR 2001

sorting products of calc-alkaline differentiation: 1. throw dense rocks off the bottom

QuickTime™ and aTIFF (LZW) decompressor

are needed to see this picture.

QuickTime™ and aTIFF (LZW) decompressor

are needed to see this picture.

QuickTime™ and aTIFF (LZW) decompressor

are needed to see this picture.

sorting products of calc-alkaline differentiation: 2. hide mafic rocks below the Moho

Santos et al J Pet 2002

Brad Hacker and his fan club, in prep.Hannes Broeckner & colleagues

sorting products of calc-alkaline differentiation: 3. “relamination” of buoyant products

veryrare

Kelemen et al., Treatise on Geochem, 2003

mixing

60-80% “cumulates”

wet and oxidized

50-60% “cumulates”

primitive andesite

~ 20-30% “cumulates”

making lavas ~ continental crust

numbers of samples,not necessarily representativeof proportions of rock types

crystalfractionation

& melting

crystal fractionation & melting

Kelemen et al., Treatise on Geochem 2003

all samples, 0.5 < Mg# < 0.7oceanic arc lavas

continental arc lavas

arc plutons

wt% SiO2

wt% SiO2

wt% SiO2

mid-ocean ridge lavas don’t “go there”

Behn & Kelemen JGR 2006