DISS. ETH NO:.

15823

Zoned Ultramafic Complexes ofthe Kohistan paleo-island

Are
Petrological, structural and geochemicalevidence for kilometer-scalemelt-
conduits

A dissertation submitted to the

SWISS FEDERALINSTITUTE OF TECHNOLOGY ZURTCH
for the degree of

Doctor ofNatural Sciences

Presented by

Oliver Jagoutz
Diplom-Geologe,
GeologischesInstitut der
Johannes Gutenberg-Universität, Mainz

Born 02.04.1971
Citizen of Austria

Accepted on the recommendation of

Prof. Jean-Pierre Burg ETH Zürich examiner

Prof. Peter Ulmer ETH Zürich co-examiner

Prof. Othmar Müntcner Universität Bern co-examiner
Prof. Jean-Louis Bodinier Universite Montpellier co-examiner

Zürich, 2004
 Table of Contents

Abstract ____________________________________________________________ 7

Zusammenfassung________________________________________________________ 9

1 Introduction ________________________________________________________ 12

1.1 Aims and Project Background__________________________________________ 12

1.2 Geographic outline of the study area ____________________________________ 12

1.3 The Himalayas: a geological overview ___________________________________ 13

1.3.1 Tectonic setting ___________________________________________________________ 13
1.3.2 Karakoram _______________________________________________________________ 16
1.3.3 Kohistan_________________________________________________________________ 17
1.3.3.1 Yasin Group Sediments____________________________________________________ 18
1.3.3.2 Chalt and Shamran (Teru) Volcanites _________________________________________ 18
1.3.3.3 Kohistan Batholith _______________________________________________________ 19
1.3.3.4 Dir and Utror Volcanites ___________________________________________________ 19
1.3.3.5 Chilas Complex__________________________________________________________ 19
1.3.3.6 Southern Amphibolites ____________________________________________________ 20
1.3.3.6.1 Southern metavolcanic Complex_____________________________________________ 21
1.3.3.6.2 Southern metaplutonic Complex _____________________________________________ 21
1.3.3.7 Jijal Complex ___________________________________________________________ 21
1.3.4 Indian plate ______________________________________________________________ 22

1.4 A brief history of time: A summary of the tectono-metamorphic evolution of the

Kohistan arc. ________________________________________________________________ 22

2 Geology of the Chilas area ____________________________________________ 33

2.1 Technique __________________________________________________________ 33

2.2 Regional structures in the mapped area __________________________________ 35

2.3 Lithology ___________________________________________________________ 35

2.3.1 Kohistan Batholith_________________________________________________________ 35
2.3.2 Gabbro __________________________________________________________________ 35
2.3.3 Sheared amphibolites_______________________________________________________ 36
2.3.4 Metasediments ____________________________________________________________ 37
 3

2.3.5 Chilas Complex ___________________________________________________________ 38

2.3.5.1 Main Gabbronorite sequence _______________________________________________ 39
2.3.5.1.1 Diorite _________________________________________________________________ 39
2.3.5.1.2 Gabbronorite ____________________________________________________________ 39
2.3.5.1.3 Tonalite ________________________________________________________________ 41
2.3.5.2 Ultramafites_____________________________________________________________ 41
2.3.5.2.1 Porphyroclastic lherzolite __________________________________________________ 41
2.3.5.2.2 Dunite _________________________________________________________________ 41
2.3.5.2.3 Lherzolite ______________________________________________________________ 41
2.3.5.2.4 Pyroxenite ______________________________________________________________ 42
2.3.5.3 Hornblendite ____________________________________________________________ 42
2.3.6 Fine grained amphibolites dykes ______________________________________________ 43
2.3.7 Thak amphibolite__________________________________________________________ 43
2.3.8 Thak Diorite______________________________________________________________ 44
2.3.9 Southern metavolcanic Complex ______________________________________________ 44
2.3.10 Southern Metaplutonic Complex_____________________________________________ 44
2.3.11 Greenshists _____________________________________________________________ 44

2.4 Ultramafic bodies ____________________________________________________ 45

2.4.1 Darel Body_______________________________________________________________ 45
2.4.2 Basha Body ______________________________________________________________ 45
2.4.3 Khanbari body ____________________________________________________________ 45
2.4.4 Thurli body ______________________________________________________________ 46
2.4.5 Giche body_______________________________________________________________ 47
2.4.6 Butho body ______________________________________________________________ 47
2.4.7 Upper Thak body __________________________________________________________ 48
2.4.8 Lower Thak body__________________________________________________________ 50
2.4.9 Gine Body _______________________________________________________________ 51

2.5 Contact relationship between ultramafite and gabbronorite sequence _________ 51

2.5.1 Contact relationships of the hornblendite _______________________________________ 53
2.5.2 Kinematics_______________________________________________________________ 59

2.6 Discussion of the field relationship ______________________________________ 62

2.6.1 Origin of the zonation ______________________________________________________ 62
2.6.2 Origin of the dunite core ____________________________________________________ 63
2.6.3 Rheological consideration ___________________________________________________ 64
2.6.4 Interpretation of the intrusive relationship_______________________________________ 65

2.7 Conclusion __________________________________________________________ 65

 4

3 Pre-collision tilt of crustal blocks in extended island arcs: structural evidence in the
Kohistan Arc ___________________________________________________________ 69

3.1 Introduction_________________________________________________________ 70

3.2 Kohistan Arc: Outline ________________________________________________ 71

3.3 Magmatic structures in the Chilas Complex ______________________________ 72

3.3.1 Description ______________________________________________________________ 72
3.3.2 Interpretation _____________________________________________________________ 76

3.4 Hornblende-plagioclase pegmatite veins__________________________________ 79

3.4.1 Description ______________________________________________________________ 79
3.4.2 Interpretation _____________________________________________________________ 80

3.5 Discussion __________________________________________________________ 81

3.6 Conclusion __________________________________________________________ 82

4 Petrology __________________________________________________________ 89

4.1 Texture_____________________________________________________________ 89

4.2 Mineral Chemistry of rocks within the Chilas Complex_____________________ 93

4.2.1 Method__________________________________________________________________ 93
4.2.2 Results __________________________________________________________________ 95
4.2.3 Plagioclase_______________________________________________________________ 95
4.2.4 Olivine __________________________________________________________________ 96
4.2.5 Clinopyroxene ____________________________________________________________ 98
4.2.6 Orthopyroxene___________________________________________________________ 101
4.2.7 Spinel__________________________________________________________________ 103
4.2.8 Amphibole ______________________________________________________________ 106

4.3 Discussion _________________________________________________________ 108

4.3.1 P-T conditions during intrusion ______________________________________________ 108
4.3.2 A composite Chilas Complex? ______________________________________________ 108
4.3.2.1 Mineral trace element modelling____________________________________________ 109
4.3.2.2 Results of mineral trace element modelling ___________________________________ 111
4.3.3 Water content____________________________________________________________ 111
4.3.4 Parental liquid composition _________________________________________________ 111

4.4 Conclusion _________________________________________________________ 113

 5

5 Lower continental crust formation through focused flow in km-scale melt conduits:
The zoned ultramafic bodies of the Chilas Complex in the Kohistan Island arc (NW
Pakistan) __________________________________________________________ 120

5.1 Introduction________________________________________________________ 121

5.1.1 Present day origin of continental crust ________________________________________ 121

5.2 Geological setting ___________________________________________________ 122

5.2.1 Chilas Complex __________________________________________________________ 123

5.3 Field and petrological observations_____________________________________ 124

5.3.1 Field relations ___________________________________________________________ 124
5.3.2 Petrological description ____________________________________________________ 126

5.4 Geochemical data ___________________________________________________ 127

5.4.1 Whole rock geochemistry __________________________________________________ 127
5.4.2 Clinopyroxene chemistry___________________________________________________ 131
5.4.3 Nd Isotope geochemistry ___________________________________________________ 135

5.5 Discussion _________________________________________________________ 136

5.5.1 Crystallisation mechanism of the main gabbronorite _____________________________ 136
5.5.1.1 Results________________________________________________________________ 138
5.5.1.2 Implication of the modelling _______________________________________________ 140
5.5.2 Origin of the dunitic core in the ultramafite bodies _______________________________ 140
5.5.3 Origin of the zonation within the ultramafite bodies ______________________________ 142
5.5.4 The origin of the ultramafite bodies __________________________________________ 143
5.5.5 Basaltic or andesitic flux out of the mantle? ____________________________________ 145
5.5.6 A model for the origin of the lower continental crust _____________________________ 148

5.6 Conclusion _________________________________________________________ 148

6 The plutonic crust of an island arcs: History from LA-ICPMS U-Pb zircon ages of
intrusive units from the Kohistan paleo-island arc (NW Pakistan). _______________ 158

Abstract _______________________________________________________________ 159

6.1 Introduction________________________________________________________ 159

6.2 Geological Setting ___________________________________________________ 160

6.2.1 The Kohistan arc _________________________________________________________ 160
6.2.2 Previous geochronological constraints ________________________________________ 161

6.3 Analytical Method___________________________________________________ 162

6.3.1 Sample description _______________________________________________________ 163
 6

6.4 Results ____________________________________________________________ 164

6.4.1 Cathodoluminescence _____________________________________________________ 164
6.4.2 Zircon chronology ________________________________________________________ 165

6.5 Discussion _________________________________________________________ 169

6.6 Implication for the development of the Kohistan arc ______________________ 172

6.7 Conclusion _________________________________________________________ 173

7 Summary and Conclusions ___________________________________________ 181

7.1 Introduction________________________________________________________ 181

7.2 Relationships between the ultramafic units and the surrounding gabbronorite 181

7.3 Effects of the intrusion of the Chilas Complex on the tectono-magmatic evolution of
the Kohistan arc _____________________________________________________________ 185

7.4 Time relationship between accretionary / obduction events and deformed

plutonites of the Kohistan Batholith_____________________________________________ 186

7.5 Tectono-magmatic evolution of the Kohistan arc _________________________ 187

7.6 Remaining open questions and outlook for future work____________________ 188

8 Appendix IV _______________________________________________________ 194

List of Figures _________________________________________________________ 209

Curriculum Vitae _______________________________________________________ 223

 7

Abstract
The Kohistan Island Arc Complex was initiated some time in the Jurassic above a north
dipping intra-oceanic subduction zone, somewhere in the equatorial area of the Tethyan
Ocean. The arc, remnants of it are now located in NW Pakistan, got embedded between the
colliding Indian and Eurasian plate during the Himalayan orogeny. It is one of the few
places in the world where the lower arc crust is fully exposed and lower crustal arc
building mechanism can be studied.
This research work aimed at elucidating three main questions: 1) What are the mutual
relationships between the ultramafic units of the so-called Chilas Complex and the
surrounding large mass of gabbronorite; 2) Which effects had this large intrusion on the
tectono-magmatic evolution of the Kohistan arc; 3) What is the time relationship between
accretionary / obduction events and deformed plutonites of the Kohistan Batholith? To
answer the first two question detailed field mapping and structural traverses were carried
out in the area around the town Chilas in NW Pakistan. Petrological and geochemical
analyses were conducted at key lithologies within the Chilas Complex. Geochronological
measurement (U-Pb zircon ages) on various intrusive units exposed in the vicinity of the
town Dir were carried out to reveal the intrusion and deformation history of the so-called
Kohistan Batholith and the associated Dir-Utror volcanite.
The so-called Chilas Complex is a large (<250 km long and 30 km wide) mafic-ultramafic
intrusion. The vast mafic part, a monotonous, homogeneous, dominantly gabbronoritic to
gabbroic sequence is associated with numerous meter- to kilometer scale ultramafic bodies.
Field relations indicate intrusion of the ultramafic rocks under sub-solidus conditions into a
partially consolidated gabbronorite mush, which in turn has intruded and reacted with the
ultrabasic rocks. The mineral major and trace element chemistry and the Nd isotopic
signature of the ultramafite and the surrounding gabbronorite sequence indicate
fractionation of the both from a common magma. The trace elements of the melt parental
to the entire Chilas Complex indicates melting within the spinel stability field in the upper
mantel (<1.5 GPa). An external component is necessary to explain the LREE signature of
the parental melt. The calculated component is most similar to a low percentage partial
melt of an eclogitised MORB probably derived from a subducted slab. Based on field
observation and geochemical arguments the dunitic core of the ultramafite bodies, are
interpreted as replacive in origin due to extensive interaction between the melt parental to
 8

the Chilas Complex and ultramafite rocks. Therefore the ultramafite bodies are thought to
be surface exposure of vertically continuous melt channels through which the mafic part of
the Chilas Complex was fed.
Since the base of the Kohistan arc prior to intrusion of Chilas Complex reached down to
pressures of ~1.8 GPa, intrusion of the Chilas Complex was associated with significant
decompression and associated uplift of the Kohistan arc base. The uplift is interpreted to
result from arc wide intra-arc extension.
Syn-magmatic structures within the mafic part of the Chilas Complex additional indicate
an extensional setting during the intrusion. The dominantly south dipping orientation of
planar magmatic fabric implies intrusion along a south-dipping, listric, normal fault.
Vertical extensional veins of late hornblende pegmatite suggest that original orientations
have been preserved. The generally north dipping attitude of the southern part of the
Kohistan arc predates the intrusion and can be explained by antithetic rotation of the
hanging wall during the intra arc extension.
The intrusion age of various plutons near the town Dir span a continuous spectrum of ages
ranging from 75 Ma to 44 Ma. 72 Ma old granites intrusive into the so-called Utror-Dir
volcanite imply a significantly older minimum age for the volcanite than previously
assumed. The chemical characteristic of some Dir-Utror volcanite is similar to the mafic
sequence of the Chilas Complex. The Dir-Utror volcanite are interpreted, at least in part, as
the effusive equivalent of the Chilas Complex intrusion. They are deposited in an intra arc
basin associated with the intra-arc rifting period.
 9

Zusammenfassung
Der Kohistan Inselbogen Komplex entstand, vermutlich im Jura, durch nordwärts
gerichtete intraozeanische Subduktion im äquatorialen Bereich der Tethys. Der Inselbogen,
von dem Teile heute im NE von Pakistan aufgeschlossen sind, wurde während der
Gebirgsbildung des Himalaya in die Kollisonszone zwischen der indischen und
eurasischen Kontinentalplatte miteinbezogen. Aufgrund der hervorragenden
Aufschlussverhältnisse bietet der Kohistan Inselbogen Komplex die weltweit seltene
Möglichkeit tiefkrustale Prozesse, die seinerzeit zur Bildung des Inselbogen führten, zu
studieren.
Diese Arbeit zielt darauf ab drei essentielle Fragen zu erörtern: 1) Welche gegenseitigen
Beziehungen bestehen zwischen den ultramafischen Gesteinen des Chilas Komplexes und
den sie umschließenden Gabbronoriten ? 2) Welche Effekte hatte die Intrusion dieses
riesigen Komplexes auf die tectonisch-magmatische Entwickelung des Inselbogens ? 3) Im
welchen zeitlichen Verhältnis stehen Akkretion/Obduktion des Inselbogens und
Deformation der Plutonite des Kohistan Batholithes ?
Um die beiden ersten Fragen zu beantworten wurde in der Umgebung der Stadt Chilas
detailliert geologisch kartiert. Anschließend wurden petrologische und geochemische
Untersuchungen an wichtigen Lithologien des so genannten Chilas Komplexes
durchgeführt. Um die dritte frage zu beantworten wurden geochronologische
Bestimmungen (U-Pb an Zirkonen) von verschiedenen intrusiven Gesteinen,
aufgeschlossen in der Nähe der Stadt Dir, wurden durchgeführt, um die
Deformationsgeschichte der so-genannten Kohistan Batholithe und der assoziierten Dir-
Utror Vulkanite zu erleuchten.
Der Chilas Komplex ist eine riesige (<250 km lang und 30 km breit) mafische-
ultramafische Intrusion. Der beherrschende mafische Teil, bestehend aus eintönigen,
homogenen, hauptsächlich gabbronoritischen bis gabbroischen Gesteinen, ist assoziiert mit
einer Vielzahl von ultramafischen Komplexen, deren Ausmaße von wenigen Metern bis zu
einigen Kilometern reichen. Die Geländeverhältnisse deuten eine Intrusion der
ultramafischen Komplexe unter sub-solidus Bedingungen in partiell verfestigten
gabbronoritischen Einheiten an, die im Gegenzug die Ultramafite intrudieren und mit ihnen
reagieren. Die Mineralchemie der ultramafischen und mafischen Gesteine und ihre Nd
isotopische Signaturen deuten darauf hin, dass beide Serien durch Fraktionierung eines
 10

gemeinsamen Magmas entstanden sind. Die Konzentrationen von Spurenelementen der

Schmelze, welche den gesamten Chilas Komplex geschaffen hat, indizieren
Aufschmelzung des oberen Mantels innerhalb des Spinel Stabilitäts-Feldes (<1.5 GPa).
Eine externe Komponente ist jedoch notwendig um die leichten Spurenelemente der
Schmelze zu erklären. Die Chemie dieser berechneten Komponente hat große Ähnlichkeit
mit einer Schmelze die durch geringe Aufschmelzung eines eklogitisierten MOR Basaltes
entstanden ist. Diese Komponente könnte von der subduzierten ozeanischen Platte
stammen. Feldbeobachtungen und geochemischen Argumenten lassen darauf schliessen,
dass die dunitischen Kerne der ultramafischen Körper durch Interaktion zwischen der
Schmelze die den Chilas Komplex gebildet hat und ultramafischen Gesteinen entstanden
sind. Deshalb werden die Ultramafite als das Oberflächen Abbild eines in die Tiefe
fortlaufenden Förderkanals, durch den die Schmelze, die später den mafischen Teil des
Chilas Komplexes geformt hat, aufgestiegen ist, interpretiert.
Da der untere Teil des Kohistan Inselbogens, vor der Intrusion des Chilas Komplexes,
schon Tiefen, die einem Druck von ca. 1.8 GPa entsprechen, erreichte, muss im
Zusammenhang mit der Intrusion der untere Teil des Inselbogens signifikant angehoben
worden sein. Diese Anhebung wird interpretiert als das Ergebnis einer Extension die den
gesamten Inselbogen erfasst hat.
Syn-magmatische Strukturen innerhalb des mafischen Teiles des Chlias Komplexes deuten
zusätzliche extensionale Spannungen während der Intrusionsphase an. Die dominant nach
Süden einfallenden, magmatischen, planaren Gefüge indizieren, dass der Chilas Komplex
entlang einer nach Süden einfallenden, listrisch verlaufenden Abschiebung intrudiert ist.
Die heutzutage vertikale Orientierung von spät-magmatischen Hornblende-Pegmatiten
lässt vermuten, dass der Komplex noch immer seine originale Orientierung im Raum
beibehalten hat. Die Rotierung der im Süden des Chilas Komplex aufgeschlossenen Serien,
die vornehmlich nach Norden einfallen, muss dementsprechend vor der Intrusion
stattgefunden haben. Eine anithetische Rotierung des Hangendens, entlang einer listrischen
Abschiebung, könnte eine Erklärung für diese Rotation liefern.
Die Intrusionsalter verschiedener Plutonen, die in der Nähe der Stadt Dir aufgeschlossen
sind, ergeben ein kontinuierliches Spektrum, das von ca. 75 Ma bis 44 Ma reicht. Intrusive
Granite innerhalb der sogenannten Dir-Utror Vulkanite sind 72 Ma alt und deuten damit
ein signifikant älteres Alter der Vulkanite an als ursprünglich angenommen. Die
chemischen Charakteristiken einiger Vulkanite kann man mit denjenigen der mafischen
Sequenz des Chilas Komplexes vergleichen. Dementsprechend werden die Dir-Utror
 11

Vulkanite, zumindest zum Teil, als vulkanische Äquivalente des Chilas Komplexes
interpretiert. Sie wurden in einem Graben innerhalb eines Bogens abgelagert, der während
der Extensionsphase entstanden ist.
 12

1 Introduction

1.1 Aims and Project Background

This thesis is part of the research project 2-77771-00 supported by the Swiss National
Science Foundation, which was focused on an interdisciplinary investigation to understand
the nature and structure of the Kohistan arc + remnant arc Complex, the tectonic
significance of magmatic events and the role of melt rock interaction in island arc and
continental crust formation.
In the frame of this Thesis, integrated field, petrological and geochemical analyses were
focused on two main units of the Kohistan Complex: 1) the so-called Chilas Complex and
2) the so-called Kohistan Batholith. Three main questions were of particular concern: 1)
What are the mutual relationships between the ultramafic units of the Chilas Complex and
the surrounding large mass of gabbronorite; 2) Which effects had this large intrusion on the
tectono-magmatic evolution of the Kohistan arc; 3) What is the time relationship between
accretionary / obduction events and deformed plutonites of the Kohistan Batholith? To
answer the first two questions, the structural and petrographical relationships between
ultramafic rocks, gabbronorites and associated plutonic rocks have been documented
through detailed mapping in the vicinity of the town Chilas; petrology, geochemistry and
Neodymium (Nd) isotope techniques have been applied to samples collected in-situ during
fieldwork. The third question has been approached differently: since detailed maps of the
Kohistan Batholith are missing and are extremely difficult to produce, due to the high
mountainous area and unwelcoming population, remote sensing of Landsat 7 ETM +
pictures has been applied to comprehend the distribution of the various plutonic bodies and
the regional trends of the Kohistan arc Complex. In addition the age of some plutons and
associated volcanites have been investigated in the Dir area, in the south western part of
the arc, with the U-Pb technique.

1.2 Geographic outline of the study area

The Chilas township is located in NW Pakistan on the Indus River in the eastern part of
Kohistan (Fig 1.1). The mapped area covers ca. 2000 km2 and provides excellent exposure
of the so-called Chilas Complex, a central unit of the former Kohistan island arc (Bard,
1983; Coward et al., 1986; Tahirkheli et al., 1979). The morphology of the Indus valley is
characterized by gentle slopes, whereas the tributary rivers are characterized by steeper
 13

slopes. Vegetation is in general sparse and localized within the water-rich river area.
Denser forests are preserved at higher altitude.

Figure 1.1: False color Landsat ETM+ Picture of NE Pakistan. The box indicates the extents of the
geological map in .

1.3 The Himalayas: a geological overview

1.3.1 Tectonic setting

The Himalayan orogen results from subduction of the Tethys Ocean and subsequent
collision of the Indian continent with Eurasia. The Indian subcontinent separated ca. 130
Ma ago from E-Gondwana (Johnson et al., 1976) and drifted northward (Patriat and
Achache, 1984). Paleomagnetic data indicate that India converged towards Asia at a
velocity of 15-20 cm/yr between 83-52 Ma and less than 10cm/yr between 52-36 Ma,
while the drift direction changed several times. Since 36 Ma the convergence rate is about
5 cm/yr. The drop in drift velocity around 52 Ma is attributed to initial collision between
the Indian continental plate and Asia, which may have began in the west and propagated
eastwards.
 14

Figure 1.2: Mid-Cretaceous (100-80 Ma) reconstruction of palaeopositions of the Kohistan (Koh) and
Ladakh (Lad) arc complexes, the Indian, African and E-Gondwana continental plates (Zaman and
Torii, 1999)
 15

The resulting gross structure of the Himalayan mountain chain is characterized by four
major tectonic units separated by major fault systems (Gansser, 1964):
Major tectonic units Major fault systems
Lhasa Block
Indus-Yarlung Tsangpo Suture Zone
Tethyan Himalaya
Southern Tibet Detachment
High Himalaya
Main Central Thrust
Lesser Himalaya
Main Boundary Thrust
Sub-Himalaya
Main Frontal Thrust

The amount of total shortening between the Tethys Himalaya and the indenting Indian
plate is estimated to be, since the Paleocene, 2700-2800 km in the east, 2475km in the
centre and 1800-2100 km in the west, based on magnetic anomalies and palaeomagnetic
data (Molnar and Tapponnier, 1975) and volumetric balancing studies (Le Pichon et al.,
1992).
The laterally very consistent structure of the Himalayan chain is different in NE Pakistan
due to the presence of the Kohistan arc Complex. From N to S:

Major tectonic units Major fault systems

Karakoram
Kohistan-Karakoram Suture
Kohistan arc Complexes
Indus Suture
High Himalaya (? disputed)
Main Central Thrust (? disputed)
Lesser Himalaya
Main Boundary Thrust
Sub-Himalaya
Main Frontal Thrust

In the following section the geology of the main tectonic units is described:
 16

1.3.2 Karakoram

Figure 1.3: Index-map of the Western Himalaya, with the major tectono-stratigraphic blocks and
sutures. Hatched area = Karakoram block (after Gaetani, 1997)

The N Karakoram (Figure 1.3) is a large and remote region in central Asia and its geology
is still partly unknown. The Karakoram microplate is separated from the Pamir range to the
north by the Tirich Mir and the Kikil fault zones (Gaetani, 1997). The southern
Karakoram, dominated by Cretaceous calc-alkaline plutonic rocks and metamorphosed
sediments (Searle and Tirrul, 1991) is separated from the northern Ordovician to Tertiary
sedimentary belt (Gaetani and Garzanti, 1991) by the Karakoram Batholith (Searle et al.,
1989). The northern sedimentary belt records a 400 Ma year depositional history and is
dominated by marine type sediments (Gaetani, 1997). Zircon U-Pb ages reveal intrusion
ages between 105 ± 0.5 Ma (Fraser et al., 1999) and 95 ± 5 Ma (Le Fort et al., 1983) for
the plutonic Hunza unit of the Karakoram Batholith. Magmatism is related to northward
subduction of the north Tethys lithosphere below an Andean type margin (Debon et al.,
1987). Metasediments of southern Karakoram record three stages of metamorphism: a
Jurassic (?) to lower Cretaceous andalusite-garnet grade, a Barrovian type metamorphism
with a minimum age of 37 Ma and contact metamorphism around the large early Miocene
Baltoro plutonic unit (Searle and Tirrul, 1991). In the south, the Karakoram Block is
bounded against the Kohistan arc by the Kohistan-Karakoram Suture zone (Tahirkheli et
al., 1979). The timing of the closure of the Kohistan-Karakoram Suture is bracketed
 17

between 105 and 85 Ma based on Rb-Sr whole rock ages of deformed and undeformed
plutonic rocks (Petterson and Windley, 1985).

1.3.3 Kohistan
The Kohistan arc (Figure 1.4) is generally regarded as a fossil Jurassic/Cretaceous island
arc that was sandwiched between the Indian and Asian plates during the Himalayan
collision (Tahirkheli et al., 1979); (Bard, 1983). In the east the Kohistan arc is separated
from the Ladakh arc by the Nanga Parbat Syntaxis, a halfwindow of Indian Plate gneisses
undergoing rapid denudation (Zeitler et al., 1993; Zeitler et al., 1989).

Figure 1.4: Geology of Kohistan based on interpretation of Landsat ETM + pictures.

It is generally believed that the intra-oceanic arc originated through a northward dipping
subduction somewhere in the equatorial area of the Tethys Ocean, near the Eurasian
continent (Figure 1.2) (Bignold and Treloar, 2003; Coward et al., 1987; Khan et al., 1993).
The Kohistan arc is composed of Jurassic/Cretaceous to Tertiary igneous, volcanic and
sedimentary rocks. The large quantity of basic to intermediate rocks led Tahirkheli et al.
(1979) and Bard (1983) to the conclusion that Kohistan represents a paleo intra-oceanic
island arc. It is one of the few places in the world where the full structure of an arc – from
oceanic upper mantle to the uppermost volcanic and sedimentary sequences is preserved
(Coward et al., 1986; Searle et al., 1999; Treloar et al., 1996). As such, the Kohistan
represents an unrivalled opportunity to investigate subduction zone processes.
 18

It is generally assumed that the Kohistan arc evolved through three main stages 1) as an
intra-oceanic island arc separated from Eurasia by a backarc basin of unknown size, 2) as
an Andean type active continental margin after accretion with Eurasia and 3) as the leading
edge of a continent-continent collision between India and Eurasia (Petterson and Windley,
1985; Searle et al., 1999; Treloar et al., 1996).
In the southernmost part of the Kohistan arc Complex, the Jjial Complex represents the
deepest part of the Kohistan arc mantle and crust (Jan and Howie, 1981; Jan and Windley,
1990; Ringuette et al., 1999) overlain by a thick pile of metaplutonites the so-called
Southern metaplutonic Complex and the fine grained Southern (Kamila) Amphibolites.
These units are separated by the Chilas Complex, intruded during incipient back arc rifting
(Khan et al., 1989; Khan et al., 1997), from the Gilgit domain in the north, which is
composed dominantly of upper crustal plutonic, volcanic and sedimentary rocks and their
metamorphosed equivalent (Petterson and Windley, 1985; Pudsey et al., 1985). In general,
the proportional amount of volcano-sedimentary rocks in the Gilgit domain diminishes
from north to south while plutonic units become more abundant. Accordingly, the domain
has been subdivided into the so-called Kohistan Batholith in the south and volcano-
sedimentary series (Chalt, Shamran and Utor volcanites and Yasin sediments) in the north.
In the next subparagraphs the main Kohistan units are described from north to south:

1.3.3.1 Yasin Group Sediments

The Yasin group is a Cretaceous volcano-sedimentary unit cropping out directly south of
the Kohistan-Karakoram Suture. It is a succession of volcanoclastic sediments, volcanite
and sediments composed of pelite to sandstone with Aptian-Albian rudist limestone.
Sedimentation most likely occurred in an intra-arc or back arc basin (Pudsey et al., 1985;
Robertson and Collins, 2002).

1.3.3.2 Chalt and Shamran (Teru) Volcanites

The Chalt and Shamran volcanites are tectono-stratigraphically below the Yasin group.
The mid-Cretaceous, island-arc related Chalt Group is divided into the backarc-related sub-
aqueous Hunza formation and the dominantly sub-aerial Ghizar formation. The Hunza
formation is composed of boninitic, basaltic and andesitic volcanites whereas the Gizar
formation comprises basaltic to andesitic volcanites. (Bignold and Treloar, 2003; Petterson
and Treloar, 2004; Petterson and Windley, 1991; Petterson et al., 1991). The Eocene-
Oligocene Shamran volcanites (also termed Teru volcanites) is comprised of a 3 km thick
 19

sequence of sub-aerially erupted mafic to felsic volcanic rocks along with their
volcanoclastic equivalents (Danishwar et al., 2001; Petterson and Treloar, 2004)

1.3.3.3 Kohistan Batholith

The Kohistan Batholith is the largest unit of the Kohistan arc. It is a composite suite of
various plutonic, volcanic and sedimentary units. Plutonic units are grouped into three
intrusive stages: Stage 1 plutons (110-85 Ma) make a deformed suite ranging from gabbros
to quartz-rich diorites. Stage 2 (60-40 Ma) are undeformed, more acid, gabbroic to granitic
plutons. Stage 3 (40-26 Ma) are leucocratic sills (Petterson and Windley, 1985, 1986;
Petterson and Windley, 1991). The Nd and Sr isotopic signatures indicate an increasing
crustal component for successively younger plutons, with the latest magma being crustal
derived. This progression results from crustal thickening and subsequent remelting of the
lower crust (Petterson et al., 1993). The deformation fabric within stage 1 plutons is
attributed to the suturing of the Kohistan arc Complex to Asia between 104 – 85 Ma.

1.3.3.4 Dir and Utror Volcanites

The Dir and Utror volcanites form a NE-SW trending belt in the central part of the arc and
are dominated by basaltic to basaltic-andesitic volcanites with subordinate siliciclastic
sediments and Eocene limestones (Sullivan et al., 1993). The volcanites show lower grade
metamorphism and yielded an 39Ar/40Ar age of 40 Ma (Treloar et al., 1989b) .

1.3.3.5 Chilas Complex

The focus of the research conducted in this thesis deals mainly with the 85 Ma old calc-
alkaline Chilas Complex (Schaltegger et al., 2002; Zeitler, 1985). This complex is
considered to be one large gabbronoritic intrusion with intercalated ultramafic units. With
the dimensions of > 250 km in length and more than 30 km in width, it is one of the largest
intrusions worldwide. It is composed of layered to homogeneous gabbronorite to quartz-
diorite, the so-called “main gabbronorite”, and associated kilometer sized ultramafic bodies
composed of dunite-peridotite-pyroxenite-gabbronorite-anorthosite, the so-called
ultramafic-mafic-anorthositic (UMA) association (Khan et al., 1993; Khan et al., 1989).
The main mineralogical difference is the abundance of olivine and Cr-spinel in the latter
and the absence of olivine and the presence of magnetite-illmenite in the former. Pressure
and temperature estimates indicate equilibration around 600-800 °C and 0.6-0.8 GPa
(Bard, 1983; Jan and Howie, 1980). Originally, the Chilas Complex has been interpreted
 20

either as the remnant of the gabbroic to ultramafic cumulate sequence of an oceanic crust
(Asrarullah et al., 1979; Butt et al., 1980; Chaudry et al., 1983) or as a layered intrusion at
the base of the island arc (Bard, 1983; Coward et al., 1986; Jan et al., 1984; Khan et al.,
1985). Geochemical and petrological data indicate an island arc affinity (Khan et al.,
1989), ruling out the oceanic origin of the Chilas Complex. Nevertheless, the large volume
of mafic magma needed to produce the Chilas Complex led Khan et al.(1989) to propose
that the Chilas Complex intruded during mantle diapirism at incipient stages of back-arc
spreading. This idea is supported by Hafnium (Hf) isotopes (εHf = 10.4), which indicate
that the Chilas Complex samples a mantle source, which is different from the one of the
lower units of the Kohistan arc such as the Southern metaplutonic Complex (εHf ≈ 14)
(Schaltegger et al., 2002). However, the relationships between the main gabbronorite and
the UMA associations is debated and forms a central part of this thesis: Khan and co-
workers (1985; 1989) interpreted the UMA associations as cumulates derived from picritic
melts emplaced into a crystallizing magma chamber. A similar model is proposed by Niida
et al. (1996) interpreting the UMA as a ultramafic crystal-mush intrusive into a
crystallizing gabbroic magma chamber. Kubo et al. (1996), however, argued that the UMA
are older than the gabbronorite. In contrast to the cumulate origin of the ultramafites, the
UMA dunites have recently been interpreted as interaction and reaction products between
host peridotites and percolating melts (Burg et al., 1998). Based on facing directions of
graded layers, the Chilas Complex has been interpreted as a several tens of kilometers high
antiform with a near-vertical axial plane (Coward et al., 1982; Coward et al., 1986). Burg
et al. (1998) however, noted that the main gabbronorite displays a magmatic fabric and that
the axial plane of this alleged fold runs through the outcrops of UMA associations; they
concluded that the facing directions do not reflect crustal-scale folding but instead
oppositely facing margins of smaller intrusive diapirs. Consequently, the UMA bodies
have been interpreted as apices of intra-arc mantle diapirs intruding into the extending
Island arc.

1.3.3.6 Southern Amphibolites

The Southern Amphibolites can be divided into the Southern metavolcanic Complex lying
on top of the Southern metaplutonic Complex.
 21

1.3.3.6.1 Southern metavolcanic Complex

The Southern metavolcanic Complex is dominantly composed of banded, fine grained,
amphibolites with intercalated calc-silicates and alternating quartz-feldspar rich bands and
subordinate metadiorites and -gabbros. High-Ti MORB type amphibolites and remnants of
pillow structures within the complex are interpreted as remnants of the Tethys crust on
which the Kohistan arc was constructed (Khan et al., 1993).

1.3.3.6.2 Southern metaplutonic Complex

The southern metaplutonic Complex is made up of variously metamorphosed, calc-alkaline
laccoliths of granite, metadiorite, and -gabbro. Garnet-granulites exposed at the southern
border of this complex records high pressure granulite-facies conditions (Ringuette et al.,
1999; Yamamoto, 1993). An intrusive sheet of granite close to the contact with the
overlying Southern Amphibolites has a zircon U-Pb age of 97.1 ± 0.2 Ma (Schaltegger et
al., 2002). A diorite within the Southern Metaplutonic Complex yielded an U-Pb zircon
age of 91.8 ± 1.4 Ma and the so-called Sarangar Gabbro, exposed in the southern part of
the complex, yielded an U-Pb zircon age of 98.9 ± 0.4 Ma (Schaltegger et al., 2002). The
Hf isotopic compositions (εHf ≈ 14) for the dated units indicate a MORB-type mantle
reservoir. Two-pyroxene granulites at the southern, bottom border of the Complex yielded
pressure-temperature estimates of 0.7-1.1 GPa and 700 to 900 °C (Yoshino et al., 1998)
and a Sm-Nd age of 118 ± 12 Ma (Yamamoto and Nakamura, 2000). Garnet-
clinopyroxene-plagioclase assemblages from the same outcrop recorded significantly
higher pressures (1.2 to 2.0 GPa (Jan and Howie, 1981; Ringuette et al., 1999; Yamamoto,
1993; Yamamoto and Yoshino, 1998; Yoshino et al., 1998) and a younger Sm-Nd age
(91.0 ± 6.3 Ma (Yamamoto and Nakamura, 2000) in accordance with the Sm-Nd age of 95
± 2.7 Ma measured by Anczkiewicz and Vance (2000). Pressure increase likely reflects
thickening through magmatic accretion within the arc (Burg et al., 1998; Ringuette et al.,
1999; Yamamoto and Nakamura, 2000; Yoshino and Okudaira, 2004).

1.3.3.7 Jijal Complex

The >3 km thick Jijal Complex is one of few outcrops of so-called layered cumulates
occurring in the immediate hanging wall of the Indus Suture. It is composed of dunites,
lherzolites, websterites, pyroxenites, garnetites and hornblendites. It occurs at the base of
the arc sequence and has been interpreted in terms of a fossil mantle-crust transition at the
 22

root of the island arc (Burg et al., 1998). Layered peridotite, mainly composed of dunite
and (garnet-) lherzolite at the base of the Complex are overlain by websterite. Cr-diopside-
bearing dykes intrude the partially layered peridotite sequence. The upper part of the
Complex is composed of garnetite and hornblendite. Slivers of hornblendite are frequently
found within the garnet-granulite of the base of the Southern Metaplutonic Complex.
Garnet-hornblendite, record pressures of 1.2-1.4 GPa (Jan and Howie, 1981) and Sm-Nd
isotope yielded an age of 83 ± 10 Ma (Yamamoto and Nakamura, 2000).

1.3.4 Indian plate

In the Northwestern Himalaya, the Indian Plate gneisses are separated from the Kohistan
Arc Complex by the Indus Suture Zone. The rocks of the northwestern part of the Indian
plate include a Precambrian basement unconformably overlain by a thick sequence of
Paleozoic to Mesozoic sediments, intruded by various plutons (Tahirkheli, 1979). During
Himalayan collision, the Indian paleo-margin was dominantly metamorphosed in
amphibolite-facies conditions. Early Tertiary, coesite-bearing eclogites exposed in the
upper Kaghan Valley indicate subduction of Indian continental rocks down to 90-100 km
(O'Brien et al., 2001). Metamorphic rocks were subsequently imbricated by late north-
dipping thrusts, resulting in an overall tectonic inversion of the metamorphic grade
(Treloar et al., 1989a).

1.4 A brief history of time: A summary of the tectono-metamorphic evolution of the

Kohistan arc.
The Kohistan arc was initiated above an intra-oceanic subduction zone in the equatorial
area of the Tethys. The time of initiation of subduction is unclear but 150 Ma U-Pb zircon
ages of Matum-Das tonalite within the Kohistan Batholith (Schaltegger et al., 2004)
indicate that arc magmatism existed in late Jurassic/early Cretaceous times. High pressure
(1.2-2.0 GPa Ringuette et al., 1999; Yamamoto, 1993) granulite-facies assemblages within
the Southern Metaplutonic and Jijal Complexes yield Cretaceous Sm-Nd ages (118 ± 12,
95 ± 2.7, 91.0 ± 6.3, 83 ± 10 Ma (Anczkiewicz and Vance, 2000; Yamamoto and
Nakamura, 2000). Since pressure increase is attributed to reflect crustal thickening (Burg et
al., 1998; Ringuette et al., 1999; Yamamoto and Nakamura, 2000; Yoshino and Okudaira,
2004), the available data indicates that arc magmatism had formed a ca. 40-60 (?) km thick
crust within less than 40 Ma. Assuming an initially 6 km thick oceanic crust on which the
Kohistan arc was build, a 150-200 km wide arc (the present day distance form the Indus to
 23

the Kohistan-Karakoram Suture zone) and simplifying the shape of the arc to a simple
triangular geometry, the observed thickening of the arc corresponds to a magma production
rate of 82-110 km3/km per Ma. This figure is comparable to crust production rate estimated
for the Izu-Bonin arc (80 km3/km per Ma.), the Mariana arc (70 km3/km per Ma.), the
Tonga and Vanuata arc (60-80 km3/km per Ma) (Dimalanta et al., 2002; Taira et al., 1998),
Aleutian (55-82 km3/km per Ma) (Holbrook et al., 1999) and the New Hebrides (87-95
km3/km per Ma) (Greene et al., 1994). However, it contrasts the recently questioned
average arc production rate of 23-33 km3/km per Ma (Reymer and Schubert, 1984).
The assumed time of suturing of Kohistan to Asia is between 104 and 85 Ma (Petterson
and Windley, 1985; Treloar et al., 1996). The intrusion of the Chilas Complex at about 85
Ma (Schaltegger et al., 2002; Zeitler, 1985) during intra-arc rifting (Burg et al., 1998; Khan
et al., 1989) is temporally associated with isothermal decompression of the lower crustal
garnet granulites from ~2.0 GPa to 0.33 GPa (Ringuette et al., 1999; Yamamoto, 1993) and
subsequent cooling below ~500 °C dated by 39Ar-40Ar hornblende data at around ~80 Ma
(Treloar et al., 1989b; Wartho et al., 1996).
During the Andean-margin stage of the Kohistan Arc Complex, magmatism continued
producing the main part of the Kohistan Batholith (Petterson and Windley, 1985) and the
Dir-Utror and Shamran volcanites (Petterson and Treloar, 2004).

This summary of previous work sets the rationale of this thesis: the tectono-magmatic
history of the arc seems to be dominated by two main events: 1) the intrusion of the Chilas
Complex and 2) suturing of Kohistan to Asia. The two points were treaded separately in
this thesis:
1) The intrusion of the Chilas Complex possibly had some influence on the lower arc crust
reflected in the temporally linked isothermal decompression of the lower arc. However, its
intrusion mechanism and the relationship between the UMA and the surrounding main
gabbronorite are insufficiently understood. Testing the various hypotheses for the UMA
origin, determining the mantle source and emplacement mechanisms for the ultramafic
units, documenting the interactions between the mafic (gabbronoritic) and ultramafic units
were the principal research goals on the Chilas Complex. Results have general implications
for the petrogenesis of large mafic-ultramafic complexes, which is at present weakly
understood. chapter 2 documents the geology of the Chilas Complex and adjoining units in
the Chilas region; chapter 3 addresses the structures within the Chilas Complex; chapter 4
 24

considers geochemical and isotopic aspects of these rocks and chapter 5 describes the
petrology and mineral major and trace element chemistry of the Chilas Complex rocks.
2) The age of suturing against Asia directly determines the intra-oceanic or Andean
margin setting of a large part (the stage 2 plutons) of the Kohistan Batholith. The origin of
the Kohistan Batholith is of importance, because if it has formed during the Andean-
margin-type stage of the arc, as presently assumed, the corresponding rocks must be
excluded from bulk chemical composition estimates of the intra-oceanic arc crust. chapter
6 presents new geochronological constraints on several plutons of the Kohistan Batholith.
The appendix presents a geological map of the Kohistan Arc Complex drawn after remote
sensing interpretations of Landsat 7 ETM + pictures.
 25

Reference

Anczkiewicz, R., and D. Vance, 2000, Isotopic constraints on the evolution of

metamorphic conditions in the Jijal-Patan Complex and Kamila Belt of the
Kohistan Arc, Pakistan Himalaya, in M. A. Khan, J. Treloar Peter, P. Searle
Michael, and M. Q. Jan, eds., Tectonics of the Nanga Parbat syntaxis and the
western Himalaya., Geological Society of London. London, United Kingdom.
2000.
Asrarullah, Z. Ahmad, and G. Abbas, 1979, Ophiolites in Pakistan; an introduction, in F.
Abul, and K. A. DeJong, eds., Geodynamics of Pakistan.: Quetta, Pakistan, Geol.
Surv. Pakistan, p. 181-192.
Bard, J. P., 1983, Metamorphism of an obducted island arc; example of the Kohistan
Sequence (Pakistan) in the Himalayan collided range: Earth and Planetary Science
Letters, v. 65, p. 133-144.
Bignold, S. M., and P. J. Treloar, 2003, Northward subduction of the Indian Plate beneath
the Kohistan island arc, Pakistan Himalaya; new evidence from isotopic data:
Journal of the Geological Society of London, v. 160, p. 377-384.
Burg, J. P., J. L. Bodinier, S. Chaudhry, S. Hussain, and H. Dawood, 1998, Infra-arc
mantle-crust transition and intra-arc mantle diapirs in the Kohistan Complex
(Pakistani Himalaya); petro-structural evidence: Terra Nova. The European
Journal of Geosciences, v. 10, p. 74-80.
Butt, K. A., M. N. Chaudhry, and M. Ashraf, 1980, An interpretation of petrotectonic
assemblage west of western Himalayan syntaxis in Dir District and adjoining
areas in northern Pakistan, in R. A. K. Tahirkheli, M. Q. Jan, and M. Majid, eds.,
Proceedings of the Internatioanl Committee on Geodynamics, Group 6 meeting.:
Geological Bulletin, University of Peshawar, v. 13: Peshawar, Pakistan,
University of Peshawar, Department of Geology, p. 79-86.
Chaudry, M. N., M. Ghazantar, and M. Ashraf, 1983, A plate tectonic model for nortwest
Himalayas: Kahsmir Kournal of Geology, v. 1, p. 109-112.
Coward, M. P., R. W. H. Butler, K. M. Asif, and R. J. Knipe, 1987, The tectonic history of
Kohistan and its implications for Himalayan structure: Journal of the Geological
Society of London, v. 144, p. 377-391.
 26

Coward, M. P., M. Q. Jan, D. Rex, J. Tarney, M. F. Thirlwall, and B. F. Windley, 1982,

Structural evolution of a crustal section in the western Himalaya: Nature
(London), v. 295, p. 22-24.
Coward, M. P., B. F. Windley, R. D. Broughton, I. W. Luff, M. G. Petterson, C. J. Pudsey,
D. C. Rex, and K. M. Asif, 1986, Collision tectonics in the NW Himalayas, in M.
P. Coward, and C. Ries Alison, eds., Collision tectonics.: Geological Society
Special Publications, v. 19: London, United Kingdom, Geological Society of
London, p. 203-219.
Danishwar, S., R. J. Stern, and M. A. Khan, 2001, Field relations and structural constraints
for the Teru volcanic formation, northern Kohistan Terrane, Pakistani Himalayas:
Journal of Asian Earth Sciences, v. 19, p. 683-695.
Debon, F., P. Le Fort, D. Dautel, J. Sonet, and J. L. Zimmermann, 1987, Granites of
western Karakorum and northern Kohistan (Pakistan); a composite Mid-
Cretaceous to upper Cenozoic magmatism: Lithos, v. 20, p. 19-40.
Dimalanta, C., A. Taira, G. P. Yumul, Jr., H. Tokuyama, and K. Mochizuki, 2002, New
rates of western Pacific island arc magmatism from seismic and gravity data:
Earth and Planetary Science Letters, v. 202, p. 105-115.
Fraser, J., M. P. Searle , R. R. Parrish, and S. Noble, 1999, U-Pb geochronology on the
timing of metamorphism and magmatism in the Hunza Karakoram (abstract):
Terra Nostra, v. 99, p. 45-46.
Gaetani, M., 1997, The Karakorum Block in Central Asia, from Ordovician to Cretaceous:
Sedimentary Geology, v. 109, p. 339-359.
Gaetani, M., and E. Garzanti, 1991, Multicyclic history of the northern India continental
margin (northwestern Himalaya): AAPG Bulletin, v. 75, p. 1427-1446.
Gansser, A., 1964, The Alps and the Himalayas.
Greene, H. G., J. Y. Collot, M. A. Fisher, and A. J. Crawford, 1994, Neogene tectonic
evolution of the New Hebrides island arc; a review incorporating ODP drilling
results, Proceedings of the Ocean Drilling Program, Scientific Results, v. 134, p.
19-46.
Holbrook, W. S., D. Lizarralde, S. McGeary, N. Bangs, and J. Diebold, 1999, Structure
and composition of the Aleutian island arc and implications for continental crustal
growth: Geology (Boulder), v. 27, p. 31-34.
 27

Jan, M. Q., and R. A. Howie, 1980, Ortho- and clinopyroxenes from the pyroxene
granulites of Swat Kohistan, northern Pakistan: Mineralogical Magazine, v. 43, p.
715-726.
Jan, M. Q., and R. A. Howie, 1981, The mineralogy and geochemistry of the
metamorphosed basic and ultrabasic rocks of the Jijal Complex, Kohistan, NW
Pakistan: Journal of Petrology, v. 22, p. 85-126.
Jan, M. Q., and B. F. Windley, 1990, Chromian spinel-silicate chemistry in ultramafic
rocks of the Jijal Complex Northwest Pakistan: Journal of Petrology, v. 31, p.
667-715.
Jan, Q. M., M. U. K. Khattak, M. K. Parvez, and B. F. Windley, 1984, The Chilas
stratiform complex; Field and mineralogical aspects: Geological Bulletin,
University of Peshawar, v. 17, p. 163-169.
Johnson, B. D., C. M. Powell, and J. J. Veevers, 1976, Spreading history of the eastern
Indian Ocean and Greater India's northward flight from Antarctica and Australia:
Geological Society of America Bulletin, v. 87, p. 1560-1566.
Khan, M. A., M. Habib, and J. Q. M., 1985, Ultramafic and mafic rocks of the Thurley
Gah and their relationship to the Chilas Complex: Geological Bulletin, University
of Peshawar, v. 18, p. 83-102.
Khan, M. A., M. Q. Jan, and B. L. Weaver, 1993, Evolution of the lower arc crust in
Kohistan, N. Pakistan; temporal arc magmatism through early, mature and intra-
arc rift stages, in P. J. Treloar, and M. P. Searle, eds., Himalayan tectonics.:
Geological Society Special Publications, v. 74: London, United Kingdom,
Geological Society of London, p. 123-138.
Khan, M. A., M. Q. Jan, B. F. Windley, J. Tarney, and M. F. Thirlwall, 1989, The Chilas
mafic-ultramafic igneous complex; the root of the Kohistan island arc in the
Himalaya of northern Pakistan, in L. Malinconico Lawrence, Jr., and J. Lillie
Robert, eds., Tectonics of the western Himalayas.: Special Paper - Geological
Society of America, v. 232: Boulder, CO, United States, Geological Society of
America (GSA), p. 75-94.
Khan, M. A., R. J. Stern, R. F. Gribble, and B. F. Windley, 1997, Geochemical and
isotopic constraints on subduction polarity, magma sources and palaeogeography
of the Kohistan Arc, northern Pakistan: Journal of the Geological Society of
London, v. 154 Part 6, p. 935-946.
 28

Kubo, K., Y. Swada, Y. Takahashi, A. B. Kausar, Y. Seki, I. H. Khan, T. Khan, N. A.

Khan, and Y. Takahashi, 1996, The Chilas igneuous complex in the western
Himalayas of northern Pakistan: Proceedings of Geoscience Colloquium,
Geoscience Labatory; Geological Survey, Islamabad, v. 15, p. 63-68.
Le Fort, P., A. Michard, J. Sonet, and J. L. Zimmermann, 1983, Petrography, geochemistry
and geochronology of some samples from the Karakorum axial batholith, northern
Pakistan, in F. A. Shams, ed., Granites of Himalayas, Karakorum and Hindu
Kush.: Lahore, Pakistan, Inst. Geol., Punjab Univ., p. 377-387.
Le Pichon, X., M. Fournier, and L. Jolivet, 1992, Kinematics, topography, shortening, and
extrusion in the India-Eurasia collision: Tectonics, v. 11, p. 1085-1098.
Molnar, P., and P. Tapponnier, 1975, Cenozoic tectonics of Asia; effects of a continental
collision: Science, v. 189, p. 419-426.
Niida, K., A. B. Kausar, and S. R. Khan, 1996, Ultramafic crystal mush intrusion into the
crystallizing magma chamber: field evidence from the Chilas Complex, northern
Pakistan: Proceedings of Geoscience Colloquium, Geoscience Labatory;
Geological Survey, Islamabad, v. 15, p. 157-172.
O'Brien, P. J., N. Zotov, R. Law, M. A. Khan, and M. Q. Jan, 2001, Coesite in Himalayan
eclogite and implications for models of India-Asia collision: Geology (Boulder),
v. 29, p. 435-438.
Patriat, P., and J. Achache, 1984, India-Eurasia collision chronology has implications for
crustal shortening and driving mechanism of plates: Nature (London), v. 311, p.
615-621.
Petterson, M. G., M. B. Crawford, and B. F. Windley, 1993, Petrogenetic implications of
neodymium isotope data from the Kohistan Batholith, North Pakistan: Journal of
the Geological Society of London, v. 150 Part 1, p. 125-129.
Petterson, M. G., and P. J. Treloar, 2004, Volcanostratigraphy of arc volcanic sequences in
the Kohistan arc, North Pakistan: volcanism within island arc, back-arc-basin, and
intra-continental tectonic settings: Journal of Volcanology and Geothermal
Research, v. 130, p. 147-178.
Petterson, M. G., and B. F. Windley, 1985, Rb-Sr dating of the Kohistan arc-batholith in
the Trans-Himalaya of North Pakistan, and tectonic implications: Earth and
Planetary Science Letters, v. 74, p. 45-57.
 29

Petterson, M. G., and B. F. Windley, 1986, Petrological and geochemical evolution of the
Kohistan arc-batholith, Gilgit, N. Pakistan: Geological Bulletin, University of
Peshawar, v. 19, p. 121-149.
Petterson, M. G., and B. F. Windley, 1991, Changing source regions of magmas and
crustal growth in the Trans-Himalayas; evidence from the Chalt Volcanics and
Kohistan Batholith, Kohistan, northern Pakistan: Earth and Planetary Science
Letters, v. 102, p. 326-341.
Petterson, M. G., B. F. Windley, and I. W. Luff, 1991, The Chalt Volcanics, Kohistan, N
Pakistan; high-Mg tholeiitic and low-Mg calc-alkaline volcanism in a Cretaceous
island arc, in K. Sharma Kewal, ed., Geology and geodynamic evolution of the
Himalayan collision zone; Part 1.: Physics and Chemistry of the Earth, v. 17 Part
II: Oxford-New York-Toronto, International, Pergamon, p. 19-30.
Pudsey, C. J., M. P. Coward, I. W. Luff, R. M. Shackleton, B. F. Windley, and M. Q. Jan,
1985, Collision zone between the Kohistan Arc and the Asian Plate in NW
Pakistan: Transactions of the Royal Society of Edinburgh: Earth Sciences, v. 76,
p. 463-479.
Reymer, A., and G. Schubert, 1984, Phanerozoic addition rates to the continental crust and
crustal growth: Tectonics, v. 3, p. 63-77.
Ringuette, L., J. Martignole, and B. F. Windley, 1999, Magmatic crystallization, isobaric
cooling, and decompression of the garnet-bearing assemblages of the Jijal
Sequence (Kohistan Terrane, western Himalayas): Geology (Boulder), v. 27, p.
139-142.
Robertson, A. H. F., and A. S. Collins, 2002, Shyok suture zone, N Pakistan; late
Mesozoic-Tertiary evolution of a critical suture separating the oceanic Ladakh Arc
from the Asian continental margin, in C. Wang, C. J. L. Wilson, and M. P. Searle,
eds., 15th international Himalaya-Karakoram-Tibet workshop., Pergamon.
Oxford, United Kingdom. 2002.
Schaltegger, U., S. Heuberger, M. Frank, D. Fontignie, S. Sergeev, and J. P. Burg, 2004,
Crust-mantle interaction during Karakoram-Kohistan accretion (NW Pakistan).
Goldschmidt 2004.
Schaltegger, U., G. Zeilinger, M. Frank, and J. P. Burg, 2002, Multiple mantle sources
during island arc magmatism; U-Pb and Hf isotopic evidence from the Kohistan
arc complex, Pakistan: Terra Nova, v. 14, p. 461-468.
 30

Searle, M. P., M. A. Khan, J. E. Fraser, S. J. Gough, and J. M. Qasim, 1999, The tectonic
evolution of the Kohistan-Karakoram collision belt along the Karakoram Highway
transect, North Pakistan: Tectonics, v. 18, p. 929-949.
Searle, M. P., A. J. Rex, R. Tirrul, D. C. Rex, A. C. Barnicoat, and B. F. Windley, 1989,
Metamorphic, magmatic, and tectonic evolution of the central Karakoram in the
Biafo-Baltoro-Hushe regions of northern Pakistan, in L. Malinconico Lawrence,
Jr., and J. Lillie Robert, eds., Tectonics of the western Himalayas.: Special Paper -
Geological Society of America, v. 232: Boulder, CO, United States, Geological
Society of America (GSA), p. 47-73.
Searle, M. P., and R. Tirrul, 1991, Structural and thermal evolution of the Karakoram
crust: Journal of the Geological Society of London, v. 148 Part 1, p. 65-82.
Sullivan, M. A., B. F. Windley, A. D. Saunders, J. R. Haynes, and D. C. Rex, 1993, A
palaeogeographic reconstruction of the Dir Group; evidence for magmatic arc
migration within Kohistan, N. Pakistan, in P. J. Treloar, and M. P. Searle, eds.,
Himalayan tectonics.: Geological Society Special Publications, v. 74: London,
United Kingdom, Geological Society of London, p. 139-160.
Tahirkheli, R. A. K., 1979, Geology of Kohistan and adjoining Eurasia and Indio-Pakistan
continents, Pakistan: Geological Bulletin, University of Peshawar, v. 11, p. 1-30.
Tahirkheli, R. A. K., M. Mattauer, F. Proust, and P. Tapponnier, 1979, The India-Eurasia
suture zone in northern Pakistan; synthesis and interpretation of recent data at
plate scale, in F. Abul, and K. A. DeJong, eds., Geodynamics of Pakistan.: Quetta,
Pakistan, Geol. Surv. Pakistan, p. 125-130.
Taira, A., Saito S, Aoike K, Morita S, Tokuyama H, Suyehiro K, Takahashi N, Shinohara
M, Kiyokawa S, Naka J, and Klaus A, 1998, Nature and growth rate of the
Northern Izu-Bonin (Ogasawara) arc crust and their implications for continental
crust formation: The Island Arc, v. 7, p. 395-407.
Treloar, P. J., R. D. Broughton, M. P. Williams, M. P. Coward, and B. F. Windley, 1989a,
Deformation, metamorphism and imbrication of the Indian Plate, south of the
Main Mantle Thrust, North Pakistan, in A. C. Barnicoat, and J. Treloar Peter, eds.,
Himalayan metamorphism.: Journal of Metamorphic Geology, v. 7; 1: Oxford,
United Kingdom, Blackwell, p. 111-125.
Treloar, P. J., M. G. Petterson, M. Q. Jan, and M. A. Sullivan, 1996, A re-evaluation of the
stratigraphy and evolution of the Kohistan Arc sequence, Pakistan Himalaya;
 31

implications for magmatic and tectonic arc-building processes: Journal of the

Geological Society of London, v. 153 Part 5, p. 681-693.
Treloar, P. J., D. C. Rex, P. G. Guise, M. P. Coward, M. P. Searle, B. F. Windley, M. G.
Petterson, M. Q. Jan, and I. W. Luff, 1989b, K-Ar and Ar-Ar geochronology of
the Himalayan collision in NW Pakistan; constraints on the timing of suturing,
deformation, metamorphism and uplift: Tectonics, v. 8, p. 881-909.
Wartho, J. A., D. C. Rex, and P. G. Guise, 1996, Excess argon in amphiboles linked to
greenschist facies alteration in the Kamila amphibolite belt, Kohistan island arc
system, northern Pakistan; insights from (super 40) Ar/ (super 39) Ar step-heating
and acid leaching experiments: Geological Magazine, v. 133, p. 595-609.
Yamamoto, H., 1993, Contrasting metamorphic P-T-time paths of the Kohistan granulites
and tectonics of the western Himalayas: Journal of the Geological Society of
London, v. 150 Part 5, p. 843-856.
Yamamoto, H., and E. Nakamura, 2000, Timing of magmatic and metamorphic events in
the Jijal Complex of the Kohistan Arc deduced from Sm-Nd dating of mafic
granulites, in M. A. Khan, J. Treloar Peter, P. Searle Michael, and M. Q. Jan, eds.,
Tectonics of the Nanga Parbat syntaxis and the western Himalaya., Geological
Society of London. London, United Kingdom. 2000.
Yamamoto, H., and T. Yoshino, 1998, Superposition of replacements in the mafic
granulites of the Jijal Complex of the Kohistan Arc, northern Pakistan;
dehydration and rehydration within deep arc crust: Lithos, v. 43, p. 219-234.
Yoshino, T., and T. Okudaira, 2004, Crustal Growth by Magmatic Accretion Constrained
by Metamorphic P–T Paths and Thermal Models of the Kohistan Arc, NW
Himalayas: Journal of Petrology, v. in press.
Yoshino, T., H. Yamamoto, T. Okudaira, and M. Toriumi, 1998, Crustal thickening of the
lower crust of the Kohistan Arc (N. Pakistan) deduced from Al zoning in
clinopyroxene and plagioclase: Journal of Metamorphic Geology, v. 16, p. 729-
748.
Zeitler, P. K., 1985, Cooling history of the NW Himalaya, Pakistan: Tectonics, v. 4, p.
127-151.
Zeitler, P. K., C. P. Chamberlain, and H. A. Smith, 1993, Synchronous anatexis,
metamorphism, and rapid denudation at Nanga Parbat (Pakistan Himalaya):
Geology (Boulder), v. 21, p. 347-350.
 32

Zeitler, P. K., J. F. Sutter, I. S. Williams, R. E. Zartman, and R. A. K. Tahirkheli, 1989,

Geochronology and temperature history of the Nanga Parbat-Haramosh Massif,
Pakistan, in L. Malinconico Lawrence, Jr., and J. Lillie Robert, eds., Tectonics of
the western Himalayas.: Special Paper - Geological Society of America, v. 232:
Boulder, CO, United States, Geological Society of America (GSA), p. 1-22.
 33

2 Geology of the Chilas area

This chapter describes the various lithologies present within and adjacent to the Chilas
Complex and their respective relationships. A map of the area around the town of Chilas
documents the field observations. Cross-cutting relationships, in particular, were used to
infer relative ages between various lithologies and to identify the sequence of magmatic
intrusions. Focus is on the mutual relationship of the gabbronorite sequence with the
ultramafite.

2.1 Technique
Mapping and sampling was carried out during several field seasons (April 2001 to January
2004). Field work, originally planed for September-October 2001 and later, was cancelled
due to September, 11th 2001 events and the subsequent unstable political situation.
However, the total amount of field work in the Chilas area amounts to about 15 weeks. The
map presented is based on own field observations, interpretation of enhanced, multi-
spectral Landsat 7 ETM+ images and maps published by the Pakistan Geological survey
(Khan and Khan, 1998; Khan et al., 1999a; Khan et al., 1999b). Mapping was conducted in
the Indus valley and the accessible tributaries, namely from west to east the Khanbari,
Thor, Hodar, Giche, Butho, Thak, Khiner and the Gine valley.
 34

Figure 2.1: Geological map of the Chilas area based on own field observation, interpretation of
enhanced multi-spectral Landsat 7 ETM + pictures and maps published by the Geological survey of
Pakistan (Khan and Khan, 1998; Khan et al., 1999a; Khan et al., 1999b)
 35

2.2 Regional structures in the mapped area

The broad structure in map-view resembles a westward opened V with the Chilas Complex
in its center. The northern arm trends NW-SE and the southern arm trends nearly W-E and
swings in the vicinity of the Nanga Parbat syntaxis to a SE-NW trend following the Indus
Suture zone (Figure 2.1).
The gabbronorite is in general massive, with magmatic layering being usually of short
lateral extent (up to few tens of meters). A weak planar fabric is developed parallel to the
layering. Due to the nearly complete absence of intra-crystalline strain features (see
chapter 3) the fabric is interpreted as magmatic. High strain zones associated with high- or
low-temperature deformation features cross-cut the magmatic fabric.
The foliation anastomoses around the ultramafite bodies located within the gabbronorite
sequence and the diorite of the Chilas Complex. Few bodies are reported to occur within
the sheared amphibolites exposed north of the Chilas Complex (Khan et al., 1999a) but
have not been visited. The ultramafic and hornblenditic bodies have an elongated shape
with the long axis generally parallel to the regional foliation.

2.3 Lithology
The mapped lithologies are described from north to south along the geological profile that
crosses nearly all rocks encountered on the map:

2.3.1 Kohistan Batholith

In the mapped area the Kohistan Batholith is composed of leucocratic intrusions mainly of
tonalitic composition. The rocks have a yellowish weathering colour and a slightly more
whitish colour on a fresh cut surface. The main minerals are feldspar, quartz, amphibole,
biotite, ±garnet. A magmatic fabric is defined by the preferred orientation of minerals and
xenoliths of an older gabbro-diorite. Fine grained, amphibole-bearing dykes cross-cut the
fabric indicating later intrusion into the tonalite (Figure 2.2).

2.3.2 Gabbro
Dark brownish weathering, coarse-grained (~0.5 cm) gabbro with a dark bluish-whitish
fresh cut surface is composed of clinopyroxene, plagioclase and hornblende (Figure 2.2).
The preferred orientation of minerals defines a magmatic fabric. Abundant xenoliths of
foliated, grayish, finer grained (~0.2 cm) diorite composed of plagioclase, hornblende,
 36

clinopyroxene and quartz are common. Garnet-bearing quartzo-feldspatic veins cross-cut

the magmatic fabric.

Figure 2.2: Field relationship between tonalite and intrusive fine grained amphibolite dykes (left).
Intrusion of the tonalite into the gabbro is indicated by diffuse tonalitic veins between larger gabbroic
blocks (right).

2.3.3 Sheared amphibolites

Dark brownish weathering, foliated, banded and fine grained (<0.1 cm) amphibole-rich
rocks composed of plagioclase, clinopyroxene and hornblende have a dark grayish-bluish
fresh cut surface. Mylonitic shear zones are developed with lenses of less deformed blocks
preserving a coarse grained (~0.5 cm) texture reflecting the dominantly meta-gabbroic to -
dioritic composition of these amphibolites. The shear zones have various orientations
warping around the larger boudins accommodating the shape of the preserved rock lenses.
The thickness of the shear zones vary from a few centimeters- to some meters-scale.
Bending of the foliation into the high strain zones indicates a bulk top-to-the-south sense of
shear.
 37

2.3.4 Metasediments
Dark brown to whitish metasediments (weathered colour) with a grey, fresh surface are
exposed along the contact between the Chilas main gabbronorite and the diorite and
between the diorite and sheared amphibolites. Meter-scale xenoliths of metasediments
within the main gabbronorite, in the diorite and in the sheared amphibolites are frequent.
The metasediments are composed of biotite-(garnet)-schists, paragneisses, quartz-feldspar-
rocks. Metatonalitic units are frequent. The mineralogy of the fine grained biotite-schists is
composed of: quartz-biotite-plagioclase-hornblende ± garnet with accessory amounts of
calcite and retrograde chlorite and epidote.
Coarse grained (~0.5 cm) paragneisses are composed of plagioclase-quartz-biotite-
hornblende ± garnet. Locally brownish-greenish garnet and epidote-bearing calc-silicate
layers are present. A distinctive tens-of-centimeter–scale compositional layering defined
by alternating layers of coarse grained paragneisses and whitish-yellowish quartz-feldspar
rocks occur. The metasediments are strongly folded and local differentiation into
melanocratic domains dominantly composed of biotite and garnet and leucocratic veins
and layers dominantly composed of quartz and feldspar indicate partial melting (Figure
2.3)
 38

Figure 2.3: Photomicrograph of metasediments. The picture on the right displays garnet-bearing calc-
silicate layers intercalated between biotite schist and quartz-feldspar layers. The picture on the left
shows biotite schist with numerous leucocratic patches indicating partial melting.

2.3.5 Chilas Complex

The Chilas Complex is composed of gabbronorite, gabbro, diorite and tonalite, the so-
called main gabbronorite sequence and of ultramafite rocks of the ultramafic-mafic-
anorthositic association (so-called UMA association of Khan et al., 1989). These two
sequences have been defined based on the different plagioclase chemistry of the two
(above cit.). In chapter 5, it will be shown that the mineral chemistry is in accordance with
a continuous differentiation sequence from the UMA sequence to the main gabbronorite
sequence. Accordingly, the original classification of Khan et al.(1989) is modified and it is
distinguished between olivine- and pyroxene-dominated ultramafite (dunite, lherzolite,
pyroxenite), hornblendite and the plagioclase- and pyroxene-dominated main gabbronorite
sequence (gabbronorite, gabbro, anorthosite, diorite, tonalite). In the field, the Chilas
Complex has been defined as mafic-ultramafic rocks without any pervasive deformational
fabric.
 39

2.3.5.1 Main Gabbronorite sequence

2.3.5.1.1 Diorite
Diorite boarders the gabbronorite to the north and to the south. The southern diorite is a
grayish weathering, quartz-bearing, medium to coarse grained (~0.5 cm), massive rock
composed of plagioclase, hornblende, quartz and minor pyroxene. A planar fabric parallel
to the axial plane of meter-scale isoclinal folds is locally exposed and the transition
between the diorite and the gabbronorite is gradational indicating a comagmatic origin for
the gabbronorite and the diorite.
The northern diorite is a slightly finer grained (~0.2-0.3 cm) homogeneous, massive,
grayish weathering body with the similar mineralogy than the southern one. However, the
planar fabric, rare magmatic layering excepted, is absent in large parts of the northern
diorite, which has a sharp intrusive contact with intercalated metsasediments. The detailed
contact relationship with the gabbronorite could not be established in the field, but a
sharply defined colour contrast in the Landsat 7 ETM+ pictures, coincident with identified
contact locality in the field, indicates a sharply defined contact.

2.3.5.1.2 Gabbronorite
Massive, coarse to medium grained (~ 0.3-0.5 cm), dark brownish-reddish on weathered
surface, gabbronorite to gabbro has a grayish, fresh cut surface. It is the largest unit of the
Chilas Complex. It has a rather homogeneous modal composition of 60 % plagioclase, 40
% pyroxene and accessory amounts of hornblende and opaque phases (magnetite and
illmenite). The pyroxene content varies between 30 % orthopyroxene and 10 %
clinopyroxene in the gabbronoritic units and these proportions are inverted in the gabbros.
Interstitial hornblende, riming pyroxene, occurs as a late stage phase locally retrogressed to
biotite. The transition between gabbronorite and gabbro is gradual on a meter- to tens-of-
meter-scale and both can be present within a single, otherwise homogeneous outcrop. The
modal amount of orthopyroxene increasing towards outcrops of ultramafite (Figure 2.4)
maybe associated with assimilation of ultramafite, frequently observed within the contact-
zone between gabbronorite and ultramafite (see further in this chapter). Accordingly, the
two lithologies have been mapped as a single unit called hereafter gabbronorite.
Subordinate layered olivine-bearing gabbronorite and rare anorthosite are present mainly
within the contact-zone between gabbronorite and ultramafite. Syn-depositional normal
 40

faulting in layered sequences and truncation of older layers by younger ones is frequently
observed (Figure 2.5).

Figure 2.4: Change of modal abundance of orthopyroxene defining the transition between
gabbronorite and gabbro. The modal amount of Orthopyroxene decreases with increasing distance
from an ultramafite outcrop Sample profile taken along KKH (35°31`50.4``/73°42`33.4``) west of the
Basheri body.

Figure 2.5: Photomicrographs of the field relation within the gabbronorite. A) normal faulting within
layered sequences developed in the lower layers whereas it is absent in the uppermost one indicating a
syn-depositional faulting. b) and c) Truncation of layers by younger ones is frequently observed, and
evidence for truncation is also found within the non-layered part of the gabbronorite sequence
 41

2.3.5.1.3 Tonalite
Massive, coarse grained (~0.3 cm) tonalite has a yellowish-whitish weathered colour. It is
composed of quartz, plagioclase, hornblende and biotite. Tonalite occurs in the
gabbronorite sequence; relationships are often obscure and could not be defined as
progressive or intrusive in the field.

2.3.5.2 Ultramafites
Ultramafites occur in several places within the gabbronorite sequence. Outcrop dimensions
range from tens of meters- to kilometer-scale. Lithologically, they are composed of
porphyroclastic lherzolite, dunite, lherzolite and pyroxenite, often spatially but not
exclusively associated with hornblendite. In the following paragraphs the different
lithologies present within ultramafite outcrops are described in more details.

2.3.5.2.1 Porphyroclastic lherzolite

Porphyroclastic lherzolite occurs as few tens of meters big, isolated bodies within dunite.
Outcrops have been found in the Indus valley near the confluence with the Thak River
(35°24`42.2``/74°08`12.4``) and on the northern bank of the Indus opposite of the Thurli
River confluence (35°29`25.9``/73°53`50.4). The rock is coarse (grain size about 1 cm)
with a dark reddish weathering colour and a dark greenish fresh broken surface. It is
composed of clinopyroxene, olivine, amphibole, orthopyroxene and minor amounts of Cr-
spinel and chlorite. Centimeter-size clinopyroxene porphyroclasts within an olivine-rich
matrix are characteristic.

2.3.5.2.2 Dunite
Dunite is the dominant ultramafite with a characteristic orange-reddish weathering surface
and a dark greenish fresh broken surface. It is homogeneous, with coarse grained olivine
(grain size < 2 cm) and subordinate amounts of millimetric Cr-spinel, clinopyroxene and
amphibole. No fabric can be discerned in the field.

2.3.5.2.3 Lherzolite
Coarse grained (<1 cm) lherzolite has a slightly darker and more reddish weathering colour
than the dunite but also a dark greenish, fresh broken surface. It is composed of
clinopyroxene, olivine, minor amounts of amphibole, orthopyroxene, spinel and locally
 42

plagioclase. Lherzolite is usually associated with dunite and the transition from lherzolite
to dunite is gradual, defined by the progressively decreasing amount of olivine down to
<5% so that the final lithology of this transition is an olivine-bearing pyroxenite.
Pyroxenite has a dark brown to greenish weathering colour and a brownish-greenish, fresh
cut surface. However, due to the presence of olivine and the transitional nature of the
contact (see paragraph 2.5) those pyroxenites are grouped and treated with the lherzolite.
The spatial extent of lherzolite in the field varies between a few centimeters to tens of
meters. Lherzolite systematically occurs between the gabbronorite and dunite. Contrary to
the porphyroclastic lherzolite, this lherzolite is void of the large centimeter scale
clinopyroxene porphyroclasts. Clinopyroxene present is interstitial between larger olivine
grains.

2.3.5.2.4 Pyroxenite
With a dark, brownish-greenish weathering with similar fresh colours, coarse grained (~0.5
cm) pyroxenite occurs either as patches in dunite and lherzolite or as dykes crosscutting the
gabbronorite and other ultramafite. It is composed of clinopyroxene, orthopyroxene and
subordinate amounts of amphibole, spinel, plagioclase or olivine. Spectacular outcrops of
pyroxenite within dunite are observed within the Khanbari Valley
(35°37`40.8``/73°53`33.5``). Contrasting with the transitional pyroxenite grouped within
the lherzolite above, this pyroxenite occurs as patches or layers with sharply defined
contacts to the surrounding rock.

2.3.5.3 Hornblendite
Hornblende-dominated pegmatite dykes composed of hornblende and plagioclase and
medium grained (<0.5cm), dark to black hornblende schlieren intruding into the
gabbronorite sequence are grouped within the term hornblendite. Hornblendites, often but
not exclusively, associated with ultramafites are found in the contact-zone between
ultramafite and gabbronorite sequence. The spatial extent of hornblendite can be
significanty larger than the associated ultramafite.
Extremely coarse grained (tens-of-centimeter-scale) hornblende-plagioclase pegmatites
with a characteristic dark black and bright white colour are present. The modal amount of
plagioclase varies between 0% and 70% but in general the pegmatites are hornblende
dominated. The size of single hornblende crystal within the pegmatites reaches up to 0.5 m
and skeletal grown amphiboles are frequent (Figure 2.6). The pegmatites are frequently
 43

zoned with a plagioclase-rich core and a hornblende-rich rim or vice versa. Hornblende is
often orthogonal to the dyke boundaries and repetitive intrusions of dykes into dykes are
common. Dunite xenoliths have been found in such pegmatite dykes spatially not
associated with any ultramafite (35°30`20.2``/74°00`04.2``). Irregular distributed
hornblende schlieren occur within the gabbronorite at the contact-zone of larger (up to
hundred of meters) pegmatite dykes.

Figure 2.6: Photomicrograph displaying the spectacular skeletal amphiboles of the hornblende
pegmatite dykes.

2.3.6 Fine grained amphibolites dykes

Fine grained (<0.1 cm) amphibolites dykes have a dark grayish-blackish weathered surface
and a blackish fresh cut surface. They are composed of amphibole and plagioclase and
minor amounts of illmentite and magnetite. They occur usually as meter wide dyke cutting
the magmatic fabric.

2.3.7 Thak amphibolite

Is a fine grained (<0.2 cm) unit with a dark weathering surface and dark grayish-green
colour on a fresh surface. The main minerals are amphibole, plagioclase, quartz, biotite and
secondary chlorite. Alignment of biotite and amphibole define a strongly developed planar
 44

fabric. However, amphibole displays no preferred alignment on the foliation surface.

Numerous centimeter- to meter-scale, yellowish-whitish quartzo-feldspatic veins, cross
cutt the foliation, displaying spectacular ptygmatitic folds with the fold axial plane parallel
to the foliation.

2.3.8 Thak Diorite

A homogeneous diorite with a bluish-whitish fresh surface and a dark yellowish-grayish
weathering colour builds a coarse grained (~0.5 cm) body. It is composed of plagioclase,
clinopyroxene, amphibole and quartz. It has a planar fabric marked by the alignment of
large amphibole, clinopyroxene and plagioclase crystals point to a magmatic origin of the
fabric.

2.3.9 Southern metavolcanic Complex

This complex includes dark brownish weathering, fine grained amphibolites with a
grayish-bluish colour on the fresh surface. It is strongly foliated but partly preserved pillow
structures, cherts, calc-silicates and agglomerate (35°17`48.6``/73°57`45.5``) indicate a
volcano sedimentary origin. However, coarse grained plagioclase, clinopyroxene and
amphibole bearing meta-diorites and meta-gabbros are intercalated. The whole Complex is
frequently intruded by leucocratic intrusives with a granitic to tonalitic composition.
Interpretation of Landsat 7 ETM+ pictures indicate that the mapped unit is the lateral of the
so-called Kamila (Southern) Amphibolites exposed in the Indus Valley. (Treloar et al.,
1996).

2.3.10 Southern Metaplutonic Complex

Is a thick pile of imbricated mafic calc-alkaline laccoliths spectacular exposed in the Indus
valley near Patan. It is composed of variously deformed (meta-) gabbro-diorites.

2.3.11 Greenshists
Fine grained dark greenish metapelitic rocks occur close to the Indus Suture Zone. In the
field the mineral assemblage has been determined as epidot, quartz, plagioclase and
chlorite indicating greenshistfacies metamorphic grade.
 45

2.4 Ultramafic bodies

In the following paragraph the location and the characteristic of the main ultramafic /
hornblendite bodies visited are briefly described:

2.4.1 Darel Body

Ultramafite have been seen in the upper part of the Darel Valley.

2.4.2 Basha Body

The Basha body is exposed within the Indus valley (Figure 2.7). It is mainly composed of
ultramafite associated with minor amounts of hornblendite. The Body is composed of a
main large body with smaller satellites. A larger second body has been mapped based on
interpretation of Landsat 7 ETM+ pictures. The bodies are elongated with the long axis
NE-SW. The contact with surrounding gabbronorite is generally sub-vertical.

Figure 2.7: Detailed geological map of the Basha peridotite and a small hornblendite body opposite of
the Darel Body (not described in the text). The detail sample profile demonstrating the transitional
transition between gabbronorite and gabbro (Figure 2.4) was taken 100 m west of the Basha body.
 46

2.4.3 Khanbari body

The Khanbari body is exposed in the upper part of the Khanbari river. Detailed contact
relationships with the surrounding diorite have not been observed. It is dominantly
composed of ultramafite and minor amount of hornblendite. Numerous pyroxene patches
and dykes are present. Sub-horizontal sills of pyroxenite branch off from a vertically
oriented pegmatitic pyroxenite dyke (Figure 2.13a, b). This structure if interpreted as the
result of melt infiltration indicate a sub-vertical orientation of the body.

2.4.4 Thurli body

The Thurli body is located at the confluence of the Thurli and Indus Rivers. A prominent
steep SE down-dip plunging mineral lineation is developed at the ultramafite-gabbronorite
contact. The body has an elongated shape with the long axis running NE-SW parallel to the
regional foliation direction. The contact is regular and straight on a tens-of-meter-scale
where it is parallel oriented to the foliation. Where the contact is discordant with the
regional foliation trend, e.g. in the SE or NW termination of the Thurli body, tens-of-
meter-scale folds are developed. The axial plane is parallel to the regional foliation and the
fold axes are parallel to the mineral lineation (Figure 2.8). Fine grained amphibolites dykes
are found frequently throughout the body. Hornblendite, rare in the southern part of the
body, increases in amount in the northern termination. The contact along the long axis of
the body dips generally steeply to the SW. Accordingly, the shape of the body resembles a
steeply, slightly to the SW dipping dyke.
 47

Figure 2.8: Detailed geological map of the Thurli Body. The contact in the SW termination is folded
with the fold-axes plane parallel to the regional foliation and the fold-axes parallel to the mineral
lineation.

2.4.5 Giche body

The Giche body located within the Giche valley (Figure 2.9) is entirely composed of
hornblendite. On the north-western termination, it is bounded by a NE-SW trending and
steeply SW dipping fault zone. Discrete fault planes and pseudotachylites indicate brittle
deformation. Hornblende pegmatite crosscut the magmatic layering and a layer parallel
magmatic foliation. The magmatic layering however, is locally deformed into upright,
open, similar folds. A sub-vertical planar fabric parallel to the axial plane is sub-parallel to
orientation of hornblende pegmatite dykes.

2.4.6 Butho body

In the Butho Valley two bodies composed mainly of hornblendite and minor ultramafite
are present (Figure 2.9). The unfriendly population prevented us to visit the ultramafite.
The contact relationship of the western body with the surrounding gabbronorite sequence
 48

was not observed. The internal structure of the western body is dominated by various dyke-
like intrusions of hornblendite-pegmatites.
The western termination of the second eastern body is sub-vertical and sheared. A steeply,
south plunging lineation is developed. Brittle shear bands indicate downwards
displacement of the hornblendite with respect to the surrounding gabbronorite. Isoclinal
folds are developed within the hornblendite. The fabric, however, diminishes within tens-
of-meter from the contact and thereafter massive hornblende-pegmatites are exposed. The
northern contact is exposed in a small gully where it is obliterated by a late stage brittle
fault.

Figure 2.9: Detail geological map of the Giche body (in the NW corner) and the two bodies
outcropping in the Butho valley. Landsat 7 ETM + pictures indicate that the larger eastern body might
be linked to the body exposed in the Thak valley.

2.4.7 Upper Thak body

The upper Thak body, hereafter called Thak body, displays, where exposed the best contact
relationships and the freshest outcrops of all bodies visited. It is dominantly composed of
ultramafite. Since the peridotite-gabbronorite contact is largely covered by Quaternary
sediments the detailed structure of the body could not be fully established (Figure 2.10).
 49

The exposed part, however, has an elongated shape with the long axis running E-W. The
foliation in the gabbronorite directly in the vicinity of the body strikes N-S. The available
data indicate that the foliation encircles the body: In the Indus valley, in the extension of
the long axis of the body, the orientation remains constant. Towards the north in the
Thalpan area and in the south in the Thak valley the foliation strikes E-W. Accordingly,
the body resembles a steeply dipping slightly to west verging ellipsoid. Along the
gabbronorite-ultramafite contact irregular inter-fingering between gabbronorite and
peridotite at all scales is present (Figure 2.11).
Meter-sized calcite-dolomite-bearing pods within the gabbronorite and the peridotite
sequence are locally developed (Yoshino and Satish, 2001). Later stage shear zones with
top to the south sense of shear and brittle faults are common.

Figure 2.10: Detailed geological map of the Thak Body. The foliation in the vicinity of the contact is
generally orthogonal to the contact, which is strongly folded and inter-fingered. Limited foliation
measurements in the Thalpan area and south of the body indicates encircling of the body by the
foliation within a distance of few hundred meters. Black star indicates location of detailed map (Figure
2.11) and red star the location of carbonate bearing dykes.
 50

Figure 2.11: Detail sketch-map of an outcrop of the Thak body near the Indus river
(35°24`34,28``/74°8`35,66``) illustrating the interfingering relationship at various scale between the
ultramafite and the surrounding gabbronorite.

2.4.8 Lower Thak body

The lower Thak body appears to be a large mass of ultramafite encircled by hornblendite
(Figure 2.9). The exposure of the main body is in rather inaccessible steep mountainous
terrain. On Landsat 7 ETM + pictures the extent of the ultramafite can be mapped based on
a sharply defined colour contrast in the Landsat 7 ETM+ pictures, coincident with
identified contact locality in the field. Following the same procedure, the hornblendite
outcropping in the Butho Gah can be followed into the Thak Valley in accordance with
previously published maps (Khan, 1988; Khan and Khan, 1998). Along the river bank of
the Thak valley a tonalite body is exposed. The observed ultramafite-gabbronorite contact
is overprinted by a steeply dipping, E-W trending sinistral strike-slip shear zone developed
between tonalite and peridotite. A strong sub-horizontal oriented mineral lineation defined
by elongated amphibole and plagioclase aggregates occurs. The tectonic foliation itself is
boudinaged on a centimeter-scale and quartz-feldspar patches fill boudins necks. Peridotite
lenses are frequent within the high strain zone.
 51

2.4.9 Gine Body

The Gine Body is a 7 km long and 3 km wide irregular body dominantly composed of
spectacular hornblende pegmatites and minor hornblende schlieren intruding the
gabbronorite sequence. The contact itself is not exposed. Within the gabbronorite,
pyroxenite dykes are boudinaged and the boudin necks are filled with pegmatitic
hornblende-plagioclase patches. In the upper part of the Gine valley tens meter wide
exposures of peridotite have been found. The dominantly dunitic bodies, vertically
continuous over at least hundred of meters, are exposed in a steep cliff. The contact itself is
not exposed but the shapes of the bodies strongly resemble a dyke.

2.5 Contact relationship between ultramafite and gabbronorite

sequence
The basic field relationships between peridotite and gabbronorite have been established in
the well exposed Thak and Thurly bodies, and have been sufficiently observed in the other
bodies to tell that the same description apply to all. A summary of the contact relation ship
is given next.
The ultramafite bodies are concentrically but irregularly zoned with a massive dunitic core
and subsequent shells of lherzolite, plagioclase-bearing lherzolite and plagioclase-bearing
olivine-pyroxenite. Amphibole-bearing lherzolite with porphyroclastic clinopyroxene
appears as xenoliths within the dunitic core; it is therefore older and relictual. Dunitic
dykes cross cut the porphyroclastic lherzolite (Figure 2.12). Larger tens-of-centimeters
wide dykes have straight contacts (Figure 2.12 a) but smaller centimeter-sized dykes
display irregular contact relations with embayment of dunite into porphyroclastic lherzolite
(Figure 2.12b). Centimeter- to tens-of-centimeters-big xenoliths of the porphyroclastic
lherzolite are found in increased quantity within the contact-zone between massive dunite
and porphyroclastic lherzolite.
Numerous and massive amphibole-bearing pyroxenite patches and dykes are found within
the dunitic core of the Khanbari body, whereas in other outcrops those pyroxenites are rare
(Figure 2.13 a-d). The main part (~95%) of the ultramafite are composed of homogeneous
dunite with local centimeter-sized chromite veins or patches (Figure 2.14 a). The transition
between dunite and lherzolite is generally gradational with increasing modal amounts of
pyroxene and amphibole and decreasing amounts of olivine. However, sharp contacts
 52

between dunite and lherzolite exist. In the following paragraph the transition between
dunite and lherzolite is described from the dunite towards the surrounding gabbronorite:
Interstitial clinopyroxene between larger olivine-clasts occurs as centimeter-scale ovoidal,
vertically elongated patches within dunite, which coalesce where patches become
numerous, up to forming dyke-shaped alignment (Figure 2.14 b). Plagioclase appears along
triple junctions of large, centimeter-sized olivine grains displaying a reaction rim of
pyroxene-spinel symplectite at plagioclase-olivine contacts. Patchy spheroidal to
ellipsoidal clusters of this reaction texture have structures similar to the clinopyroxene
patches described above (Figure 2.14 c-e). Patches can be locally gabbronorite and a
continuous transition exists between those gabbronorite patches and patches defined by the
two-pyroxene-spinel reaction texture (Figure 2.14 e-g). Gabbronorite occurs also as
schlieren around larger ultramafite clasts (Figure 2.14 h) or as localized infiltration
disintegrating dunite along grain boundaries (Figure 2.15 a). In plagioclase-dominated
veins, this disintegration of the dunite can result in rocks with troctolitic mineral
composition (Figure 2.15 b). Plagioclase-bearing veins vanish along strike into planar
traces defined by two-pyroxene-spinel symplectites (Figure 2.15 c, d). Around larger
meter-sized gabbronorite patches, tens-of-centimeters thick reaction halos rich in pyroxene
document the transformation of dunite into lherzolite and angular dunite blocks (Figure
2.15 e). Transformation also occurs over tens-of-meters, indicated by fragmented dunite
blocks floating in lherzolite with various amounts of pyroxene (Figure 2.15 f, g). Dunite
blocks are surrounded and possibly protected from transformation by a layer of pyroxenite.
Irregularly shaped xenoliths of gabbronorite within the lherzolite are frequent (Figure 2.15
h). Taking into account the vertical attitude of both the mineral lineation and the foliation
in some of the gabbronorite xenoliths (Figure 2.15 i), some of the “xenoliths” in plane few
may be tube-shaped melt-channels in the vertical direction.
The gabbronorite-ultramafite contact is sharp but very irregular (Figure 2.16 a),
burgeoning, often inter-fingering (Figure 2.16 b). In the gabbronorite next to the
ultramafite, ultramafite xenoliths (Figure 2.16 a,d,e) are numerous. Disaggregating of
larger into smaller xenolith (Figure 2.15e) and “lava-lamb” like structures were ultramafite
blobs come of from ultramfite fingers within gabbronorite (Figure 2.16 d) indicate
assimilation of ultramafite. Additionally, coarse grained hornblendite, hornblende-anorthite
and K-feldspar bearing pegmatites occur within the contact-zone.
 53

2.5.1 Contact relationships of the hornblendite

The large hornblendite bodies (Gine, Giche, Butho) have slightly different contact
relationships with the gabbronorite. Meter thick hornblende pegmatites crosscut the
magmatic fabric and are therefore later stage intrusions. Unfortunately the contacts of the
main bodies with the surrounding gabbronorite are often obliterated by late stage brittle
faulting (Figure 2.17) and good exposures are rare. However, ductile folding and melt-
filled boudin necks implies that some deformation occurred at high temperature (Figure
2.17). The parallel orientation of an axial plane foliation to the strike of hornblende
pegmatites described above in the Giche body may imply that at least some of the ductile
deformation was induced by the emplacement of the hornblendite dykes. However,
additional fieldwork is necessary before definite statement can be made.
 54

Figure 2.12: Photographs of the field relationships between dunite and porphyroclastic lherzolite. a)
Tens-of-centimeter wide dykes often display a sharp contact with surrounding porphyroclastic
lherzolite. b) Smaller scale dykes however, often have irregular contacts with frequent embayment of
the dunite into the porphyroclastic lherzolite. Branches of the dunitic dykes surround larger clasts. c)
Branches of dunitic dykes can die off within short distance. d) The dunitic dykes become frequent and
only relictual porphyroclastic lherzolite xenoliths float within.
 55

Figure 2.13: Photographs of the field relationships of massive pyroxenite within the peridotite
outcropping in the Khanbari Valley. a) Subhorizontal branches of pyroxenite veins branches of from a
massive pegmatitic pyroxenite dyke. b) Lineout of picture (a). c) Patches of pyroxenite within
peridotite (d) pyroxenite dyke crosscutting peridotite.
 56

Figure 2.14: Photographs of the field relationships of gradual transformation of dunite into lherzolite.
a) The dunitic core is homogeneous with patches of Cr-spinel. b-e) The structures of the patchy
clinopyroxene (b) are equal to the structures of the two-pyroxene–spinel symplectite (c-e) and are
accordingly illustrated using the later due to better contrasts. Picture d and f are rotate 90° counter
clockwise for layout reasons. See text for discussion
 57

Figure 2.15: Photographs of the field relationships of gradual transformation of dunite into lherzolite.
Pictures c, d and i are rotate 90° counter clockwise for layout reasons. a-d) Photographs of
gabbronoritc veins intruding into the ultramafite resulting in disintegration of the structure of the
ultramafite. e) Reactive zones around gabbronorite patches transforming dunite into lherzolite. g-h)
Reactive zones with preserved dunitic fragments. The pyroxene content can be very irregularly
 58

developed within the lherzolite. h-i) Shape of gabbronorite xenoliths within lherzolite. j) Contract
between gabbronorite and lherzolite. See text for discussion

Figure 2.16: Photographs of the field relationships at the contact-zone between gabbronorite sequence
and ultramafite. The orientation of the pictures is indicated; in map view the arrow points towards
north. a-c) Irregular contact between gabbronorite and ultramafite with interfingering between the
both. d) “Lava-lamb” like structure of a ultramafite drop within a gabbronorite xenolith. The next
drop was about to be formed. e) Disintegrating ultramafic xenolith within the gabbronorite sequence.
 59

Figure 2.17: Photograph of the deformation fabric observed at the contact between hornblendite and
gabbronorite sequence. left) brittle shearbands indicating upwards movement of the gabbronorite in
respect to the hornblendite. right) Boudinage of orthopyroxene rich dyke in the vicinity of the contact
with the Gine body. The boudin necks are partly filled with pegmatitic plagioclase and hornblende
minerals indicating high temperature deformation.

2.5.2 Kinematics
Within the contact zone between gabbronorite and ultramafite kinematic criteria indicate
upwards directed magmatic flow in respect to ultramafite xenoliths. Additionally upwards
movement of the ultramafite in respect to the gabbronorite is indicated by local bending of
layered gabbronorite. In the following paragraph these criteria will be described:
The sub-vertical gabbronorite-ultramafite actual contact is concordant with the pronounced
foliation and lineation present in the gabbronorite sequence within the contact-zone (Figure
2.15j). Few meters away in the gabbronorite, this attitude changes gradually and the
gabbronorite is generally homogeneous, displaying a weak, planar magmatic fabric.
Numerous xenoliths present in the gabbronorite within the contact-zone disaggregated
indicating assimilation. Larger, meter- to tens-of-centimeters-big xenoliths break down
into centimeter-big ones. The small xenoliths, aligned along vertically-oriented tails are
systematically located above the “mother-xenoliths” (Figure 2.16 e). These tails are
 60

parallel to the foliation and lineation. The foliation below the “mother-xenolith” is
deflected around. Within the lee part of the xenolith the foliation is absent. The structure is
consistent with sub-vertical, upwards flow of the magma parental to the gabbronorite
sequence with respect to the xenolith. The deflection of the vertically oriented foliation
around xenolith and the absent of the foliation in the lee part of the xenolith indicates a
magmatic origin of the planar fabric due to magmatic flow.
Locally upwards deflection of graded, layered, olivine-bearing gabbronorite towards the
contact with ultramafite is observed. Layers get thinned towards and truncated by the
contact with the ultramafite (Figure 2.19). Instabilities, developed at the interface of two
layers, give important information about this deflection. The instabilities develop where
less dense plagioclase is overlain by denser olivine and pyroxene rich parts of the
following layer during crystallization. The inverted density gradient across the interface of
two layers (so-called Rayleigh-Taylor instability) can trigger density-driven mass transfer
with vertical directed upwards movement of less dense material in respect to the denser
(Woidt, 1978). These instabilities are sub-vertical oriented where the layers are sub-
horizontal. However, where the layers are sub-vertically oriented these instabilities are
dragged downwards. Accordingly, the instabilities must have formed before the layers
were dragged into the vertical orientation. The downward directed instabilities and the
vertically oriented layers are interpreted to be result of upwards directed shear of the
ultramafite with respect to the gabbronorite.
 61

Figure 2.18: Photograph illustrating details of picture Figure 2.15 e. An ultramafic xenolith is
assimilated and disintegrated. The smaller xenoliths form a vertically oriented tail above the “mother-
xenolith”. See text for detailed discussion.
.
 62

Figure 2.19: Sketch of the upwards bending and shearing of layered olivine-bearing gabbronorite at
the contact with ultramafite (35°19`20.9``/74°07`42``). Instabilities developed at the interfaces between
two layers are sub-vertical in the lower sub-horizontal part of the layers, whereas in the sub-vertical
part the instabilities are dragged downwards and layers are truncated by ultramafite, indicating
upwards shearing of the ultramafite in respect to the gabbronorite.

2.6 Discussion of the field relationship

The contact relationships are summarized in Figure 2.20.

2.6.1 Origin of the zonation

The observed zoning of the ultramafic bodies is defined by a decrease in olivine content
and increase of pyroxene content towards the surrounding gabbronorite. The increase of
pyroxene content is irregular and within a single lherzolite outcrop, dyke-shaped pyroxene-
richer domains can be identified. This observation contrasts with sharp defined layers
within ultramafite cumulate sequences (Irvine, 1974). Even so the succession from dunite-
lherzolite-pyroxenite is in accordance with the classical discontinuous differentiation
sequence (Bowen, 1922) the field observations are not in accordance with a ordinary
cumulate origin in terms of crystal settling in a magma chamber. Reactive halos around
 63

gabbronoritic dykes, dunite blocks within lherzolite and other observations described
above indicate lherzolite formation due to melt-rock interaction. Accordingly, the zoning is
interpreted as the result of an irregular centimeter to meter-scale reaction front developed
at the contact between ultramafite-and surrounding gabbronorite. As working hypothesis,
the reaction can be explained by the infiltration of an olivine-undersatured melt reacting
with olivine to form pyroxene according to the well-known peritectic reaction (Bowen and
Andersen, 1914):

ol+melt1 = pyx + melt2 (1)

The reaction results in localized or pervasive refertilisation of the dunite. The consequent
working hypothesis is that the melt infiltrating into the dunite and assimilating olivine is
parental to the surrounding gabbronorite as it has been described in the Big Jim Complex,
(Washington Cascades Kelemen and Ghiorso, 1986). However, in the case of the Chilas
Complex not all minerals within the reaction zone are due to reaction (1): for instance the
presence of plagioclase in triple junction represents a true liquidus phase.

2.6.2 Origin of the dunite core

Based on the crosscutting and replacive relationships described above the amphibole-
bearing porphyroclastic lherzolite is the oldest lithology. Dunite first occurs as dykes
within the porphyroclastic lherzolite. The absence of any discontinuity between the dunitic
dykes and the massive dunite core indicates that it is a continuous lithology and the
formation mechanism for both was probably similar. The numerous dunitic dykes can be
explained by three mechanisms: 1) The dunite dyke represent crystallized ultramafic
liquid. 2) Dunite dykes represent former melt-filled dilational fracture zones. The
ascending melt crystallized olivine (+spinel) filling the fracture zone with cumulative
dunite. 3) The olivine originated through melt-rock interaction during localized porous
flow in non-dilational conduits forming “replacive dunite” (Boudier and Nicolas, 1972;
Boudier and Nicolas, 1977).
Thermal subduction zones models predict maximum mantle-wedge temperatures beneath
the volcanic arc of ~1200°C (e.g. Peacock and Wang, 1999) significantly lower than the
extremely high melting temperature of Mg-rich olivine (~1890°C Bowen and Schairer,
1935) rendering explanation 1) unlikely. To distinguish between hypotheses 2) and 3) is
not straightforward. The dunitic dykes rarely exceed a width of tens of centimeters making
 64

them an ineffective transport mechanism due to the large surface area to volume ratio.
Magma moving along small fractures would freeze or at least cool more rapidly than
magma ascending in larger fractures. Accordingly the mineralogy should be different,
which is not observed. Additionally, the above described irregular contact relationships of
some dunitic dykes and other observations described above indicates replacement of the
porphyroclastic lherzolite by dunite making hypotheses 3) the most likely.
Based on structural and geochemical arguments “replacive dunite” have been interpreted
as channels of chemically isolated transport of melt (Kelemen et al., 1992; Kelemen et al.,
1995; Quick, 1982; Suhr, 1999). Accordingly, we interpret the origin of the dunite core as
due to infiltration of a silica-undersaturated melt into the porphyroclastic lherzolite due to
the opposite direction of reaction (1) (Kelemen, 1990):

pyx (+amph) +melt3 = ol + melt4 (2)

Field observations indicate the following sequence of events:

Whereas melt rock interaction (1) and (2) correspond to reaction (1) and (2) respectively.
The two reactions can be either due to the infiltration of different melts with different silica
activity or due to a single melt and changing intensive parameters (e.g. pressure and
temperature). The petro-chemical work that follows in this thesis aims to test this
hypothesis.

2.6.3 Rheological consideration

The contact relationships of dunite xenoliths in lherzolite are distinctly different from the
contact relationships between gabbronorite xenoliths in lherzolite. Dunite xenoliths within
lherzolite are sharply defined, angular blocks (Figure 2.15 e, f). The presence of melt-filled
tension gashes indicates a rigid behavior of the dunite with respect to the lherzolite.
Xenoliths of gabbronorite within lherzolite are rounded with irregular inter-fingering
contact (e.g. Figure 2.15 f, g; Figure 2.16 c) indicating similar viscosities. In the light of
the inferred reaction front due to melt infiltration into the dunite, the similar rheology of
gabbronorite and lherzolite indicate that both were likely a crystal-mush during intrusion
whereas the dunite behaved more solid.
 65

Coexisting crystal-mush also explains the mutual intrusive relationships observed between
gabbronorite and lherzolite.

2.6.4 Interpretation of the intrusive relationship

Magmatic flow indicators point towards vertical directed flow of the melt parental to the
gabbronorite (Figure 2.18). Additional, the upwards bending of layered gabbronorite
indicates uprising and invasion of the ultramafite into the semi-consolidated gabbronorite
sequence (Figure 2.19). Trapped melt present in the gabbronorite mush, in turn, intruded
and reacted with the dunite forming lherzolite crystal-mush. Such conditions could be
present in a melt extraction system were replacive dunite, acting as a melt-conduit is
mobilized due to high melt-rock ratio in the channel and intrudes into higher level mafic
sequences.

2.7 Conclusion
Field relations indicate intrusion of the ultramafic rocks under sub-solidus conditions into a
partially consolidated gabbronorite mush, which in turn has intruded and reacted with the
ultrabasic rocks. Infiltration of the melt parental to the gabbronorite reacted with the dunite
forming a tens-of-meter- to centimeter-scale reaction zone resulting in refertilisation of the
dunite. The dunite has a replacive origin resulting from melt-rock interaction between
amphibole bearing lherzolite and an unconstrained melt.
 66

Figure 2.20: Schematic illustration of contact relationship between gabbronorite and ultramafite.
Indicated are also the locations of key rock samples, which will be frequently mentioned in subsequent
chapters.
 67

References

Boudier, F., and A. Nicolas, 1972, Fusion partielle gabbroique dans la lherzolite de Lanzo
(Alpes piemontaises): Schweizerische Mineralogische und Petrographische
Mitteilungen = Bulletin Suisse de Mineralogie et Petrographie, v. 52, p. 39-56.
Boudier, F., and A. Nicolas, 1977, Structural controls on partial melting in the Lanzo
peridotites, in H. J. B. Dick, ed., Magma genesis 1977; proceedings of the
American Geophysical Union Chapman conference on partial melting in the
Earth's upper mantle.: Bulletin - Oregon, Department of Geology and Mineral
Industries, v. 96: Portland, OR, United States, Oregon Department of Geology and
Mineral Industries, p. 63-78.
Bowen, N. L., 1922, The reaction principle in petrogenesis: Journal of Geology, v. 30, p.
177-198.
Bowen, N. L., and O. Andersen, 1914, The binary system MgO-SiO (sub 2): American
Journal of Science, v. 37, p. 487-500.
Bowen, N. L., and J. F. Schairer, 1935, The system MgO-FeO-SiO (sub 2): American
Journal of Science, v. 29, p. 151-217.
Irvine, T. N., 1974, Petrology of the Duke Island ultramafic complex, southeastern Alaska:
Memoir - Geological Society of America, v. 138: Boulder, CO, United States,
Geological Society of America (GSA), 240 p.
Kelemen, P. B., 1990, Reaction between ultramafic rock and fractionating basaltic magma;
I, Phase relations, the origin of calc-alkaline magma series, and the formation of
discordant dunite: Journal of Petrology, v. 31, p. 51-98.
Kelemen, P. B., H. J. B. Dick, and J. E. Quick, 1992, Formation of harzburgite by
pervasive melt/ rock reaction in the upper mantle: Nature (London), v. 358, p.
635-641.
Kelemen, P. B., and M. S. Ghiorso, 1986, Assimilation of peridotite in zoned calc-alkaline
plutonic complex; evidence from the Big Jim Complex, Washington Cascades:
Contributions to Mineralogy and Petrology, v. 94, p. 12-28.
Kelemen, P. B., N. Shimizu, and V. J. M. Salters, 1995, Extraction of mid-ocean-ridge
basalt from the upwelling mantle by focused flow of melt in dunite channels:
Nature (London), v. 375, p. 747-753.
 68

Khan, M. A., 1988, Petrology and structure of the Chilas ultramafic-mafic complex,
Kohistan Arc, NW Himalayas, University of London, London.
Khan, M. A., M. Q. Jan, B. F. Windley, J. Tarney, and M. F. Thirlwall, 1989, The Chilas
mafic-ultramafic igneous complex; the root of the Kohistan island arc in the
Himalaya of northern Pakistan, in L. Malinconico Lawrence, Jr., and J. Lillie
Robert, eds., Tectonics of the western Himalayas.: Special Paper - Geological
Society of America, v. 232: Boulder, CO, United States, Geological Society of
America (GSA), p. 75-94.
Khan, N. A., and T. Khan, 1998, Geology of the Chilas Quadrangle (43-I/3) Diamir
District, Northern Areas, Pakistan: Geological Survey of Pakistan.
Khan, N. A., T. Khan, G. Mujtaba, H. Hussain, and R. Khan, 1999a, Geology of the Kiner
Gah Quadrangle (43-I/2) Diamir District, Northern Areas, Pakistan: Geological
Survey of Pakistan.
Khan, N. A., M. Latif, M. S. Bakht, and A. Fayaz, 1999b, Geology of the Gunar
Quardangle (43-I/7), Diamir District, Northern Areas, Pakistan: Geological
Survey of Pakistan.
Peacock, S. M., and K. Wang, 1999, Seismic consequences of warm versus cool
subduction metamorphism; examples from Southwest and Northeast Japan:
Science, v. 286, p. 937-939.
Quick, J. E., 1982, The origin and significance of large, tabular dunite bodies in the Trinity
Peridotite, northern California: Contributions to Mineralogy and Petrology, v. 78,
p. 413-422.
Suhr, G., 1999, Melt migration under oceanic ridges; inferences from reactive transport
modelling of upper mantle hosted dunites: Journal of Petrology, v. 40, p. 575-599.
Treloar, P. J., M. G. Petterson, M. Q. Jan, and M. A. Sullivan, 1996, A re-evaluation of the
stratigraphy and evolution of the Kohistan Arc sequence, Pakistan Himalaya;
implications for magmatic and tectonic arc-building processes: Journal of the
Geological Society of London, v. 153 Part 5, p. 681-693.
Woidt, W. D., 1978, Finite element calculations applied to salt dome analysis:
Tectonophysics, v. 50, p. 369-386.
Yoshino, T., and K. M. Satish, 2001, Origin of scapolite in deep-seated metagabbros of the
Kohistan Arc, NW Himalayas: Contributions to Mineralogy and Petrology, v. 140,
p. 511-531.
 69

3 Pre-collision tilt of crustal blocks in extended island

arcs: structural evidence in the Kohistan Arc

Jean-Pierre Burg, Oliver Jagoutz, Hamid Dawood1, Shahid S. Hussain1

Geologisches Institut, ETH and University Zurich, Sonneggstrasse, 5, CH 8092 Zurich,

Switzerland
1
Pakistan Museum of Natural History, Garden Avenue, Shakarparian, 44 000 Islamabad,
Pakistan
 70

This chapter is written in paper format. It describes structures within the gabbronorite
sequence indicating that the Chilas Complex, still in its original vertical orientation and
composed of multiple intrusions, intruded along a south dipping normal fault.

Abstract
Magmatic and deformational fabrics and dyke orientations in the Chilas gabbronorite of
the Kohistan Arc show that this reportedly huge intrusion is a sheeted-dyke-like complex
emplaced into generally southward-dipping listric faults that dominated arc extension
about 85 Ma ago. Vertical extensional veins of late hornblende pegmatite suggest that
original orientations have been preserved. We hypothesize that the northward dip of the
granulite and amphibolite facies rocks on the southern side of the Kohistan is inherited
from rotation of crustal blocks in the hanging wall of the listric faults.

Key words: Magmatic structures, island arc, extension, Kohistan

3.1 Introduction
The Kohistan terrane in NE Pakistan (Figure 3.1) is regarded as a fossil island arc obducted
between the collided Indian and Asian plates (Bard et al., 1980; Tahirkheli et al., 1979).
Owing to the admirable quality of exposures, the Kohistan offers an unrivalled opportunity
to investigate the structure of an island arc and related subduction processes (e.g. Bard,
1983b; Treloar et al., 1996). In particular, numerous time markers in form of intrusive
bodies make the Kohistan an exceptional place to study the significance of magmatic
structures in the deep crust of an arc system.
We present results from field-based studies and thin section observation within the
gabbronoritic-ultramafic Chilas Complex, away from the tilted western limb of the crustal-
scale Nanga Parbat antiform (Figure 3.1) (Butler et al., 1992). Emphasis is placed on flow-
related structures (Balk, 1959) and pegmatite veins. Flow-related foliation and lineation
suggest that the gabbronorite emplaced into a predominantly southwest-dipping
extensional zone becoming a dextral transtension zone to the west. We argue that this part
of the Kohistan has preserved a near-original orientation since the emplacement of sub-
 71

vertical pegmatite veins that cut the flow-related structures during late brittle stages of
magma intrusion. If this hypothesis is verified, the north-dipping attitude of the southern
Kohistan may reflect late-Cretaceous tilting of the southern (oceanward) parts of the
Kohistan Arc, an event which is theoretically possible on a south-dipping listric fault
system at a lithospheric-scale.

3.2 Kohistan Arc: Outline

The Kohistan Arc originated during the Cretaceous through north-dipping subduction in
the equatorial area of the Tethys Ocean (e.g. Yoshida et al., 1996). Three major units
depict the gross organisation of the arc complex. They are, from North to South (Figure
3.1):
1. The Chilas Complex is described as a >300 km long and up to 40 km thick
mafic-ultramafic lopolith (Bard, 1983b; Jan et al., 1984; Khan et al., 1993; Khan et al.,
1989). It intruded at about 85Ma (Schaltegger et al., 2002; Zeitler et al., 1980) and is
attributed to within-arc extension (Burg et al., 1998; Khan et al., 1989). Crosscutting
pegmatite dykes have a zircon U-Pb age of ca. 83 Ma (Schaltegger et al., 2002). Sm-Nd
dating indicates cooling of the main gabbronorite below 600°C at about 70 Ma (Ali et al.,
2002; Yamamoto and Nakamura, 1996). Our mapping identified a multiple intrusion
system, as previously suggested by these authors and (Jan and Howie, 1980). Individual
bodies have dyke-like stratiform shapes broadly parallel to the mesoscopic layering. The
Complex has intrusive contacts with both metasediments and metavolcanites, probably of
the Jaglot Group, and plutonic rocks possibly belonging to the early phases of the Kohistan
Batholith, to the north and the Southern Amphibolites to the south (Searle et al., 1999;
Treloar et al., 1996).
2. The Southern (Kamila) Amphibolites denominate a composite association of
heterogeneously strained amphibolitic rocks (Jan, 1988), among which some are 90-100
Ma old meta-diorites and -gabbros (Schaltegger et al., 2002). For sake of simplicity, we
ascribe lowermost crust-mantle rocks of Jijal (Burg et al., 1998) to this unit.
3. Kohistan Batholith area composed of multiple intrusion of composed
dominantly of upper crustal plutonic, volcanic and sedimentary rocks and their
metamorphosed equivalents (Petterson and Windley, 1985); (Pudsey et al., 1985).
The systematically N-dipping attitude of foliations in the Southern Amphibolites and the
bulk distribution of upper crustal rocks to the north of deeper crustal and mantle rocks has
led to the conclusion that the whole Kohistan Arc Complex has been tilted by about 30° to
 72

the north during southward collision and obduction onto the Indian continent (Bard, 1983b;
Coward et al., 1982; Coward et al., 1986; Tahirkheli et al., 1979).

Figure 3.1: Sketch map of the Kohistan Arc, mostly based on own mapping and interpretation of
multispectral Landsat7 ETM+ satellite images. The Chilas Complex is restricted to the gabbro-norites,
a lithological simplification that highlights the backbone location between the North Kohistan Region
and the Southern Amphibolites. Bounding diorites and tonalites traditionally ascribed to the Chilas
Complex are being mapped.

3.3 Magmatic structures in the Chilas Complex

3.3.1 Description
The main gabbronorite is composed of plagioclase, clinopyroxene, orthopyroxene,
accessory ilmenite, apatite and oxides and minor amphibole riming pyroxenes; the rock
displays a predominantly granoblastic texture (grain size 1 to 5 mm; Figure 3.2). Euhedral
to subhedral pyroxenes and plagioclase are locally preserved. Twinning within plagioclase
is very common and typically parallel sided. Lithological changes involve variable
proportions of these main constituent minerals, olivine in ultramafic lithologies and the
occurrence of quartz and biotite in dioritic and tonalitic differentiates.
Detailed petrography, unessential to this work, can be found in (Jan et al., 1984) and
mineral chemistry in (Khan et al., 1989).
The main gabbronorite displays a spectrum of fabrics that can be separated into primary
igneous vs secondary metamorphic/tectonic fabrics. Secondary fabrics include shear zones,
rare pseudotachylites and foliated regions that have undergone ductile, usually weak
 73

deformation easy to identify since it generally cuts across the primary textural and
structural elements. In thin section, evidence includes undulatory extinction of minerals,
mechanical twins in plagioclase, more or less intense bending of pyroxenes, plagioclase
and their alteration products and small intergranular aggregates of polygonal plagioclase
grains in the more strongly deformed rocks. Mineral associations indicate a wide range of
shear zones formed from subsolidus conditions to green-schist facies retrogression.
The meaning of the secondary fabric in relation to regional tectonics and/or magmatic
history is beyond the scope of this work. Related structural measurements, which
document solid-state strain, were excluded from this analysis. Also excluded are
measurements from the high-temperature, mylonitic southern border of the gabbronorite,
which bears down-dip lineations associated with top-to-the-south criteria of sense of shear.
Here we concentrate on orientation data of primary igneous fabrics, which were identified
from both field and microscopy observation. Primary fabrics represent flow-related
structures ranging from faint mineral preferred orientation in otherwise massive,
homogeneous rocks (Figure 3.2) to composition layers (Figure 3.3 a) particularly
prominent in the vicinity of lens-shaped, generally concordant ultrabasic bodies (Jan et al.,
1984). The absence of crystal-plasticity deformation features in conventional thin sections
(Figure 3.2) indicates that even pronounced layering with slump-folds and geopetal
structures (Khan et al., 1989; Sawada et al., 1994) represents planar concentration domains
inherited from magmatic processes. In places, interlayered lithologies show syn-
sedimentation faults, erosional contacts, cross-bedding and magmatic breccias that reflect
very dynamic, multiphase magmatic activity (Figure 3.3 a). Cross-cutting relationships
observed in such structures may explain map-scale discordance of foliation trajectories in
zones where neither lithological contrast, nor shear zone nor fault is apparent (Figure 3.3 d,
e). We emphasise that such “turbiditic-like” structures with “synsedimentary” faults and
slump folds are evidence for magmatic currents down non-horizontal slopes rather than
sub-vertical crystal settling under the influence of gravity on a horizontal magma-floor
(e.g. McBirney and Nicolas, 1997). Cross-cutting relationships between variable
lithologies indicate that the gabbronorite is a multiple magma complex accumulated
through repeated injections of magma pulses of similar composition. They further indicate
that tonalitic and dioritic magmas are younger than the main gabbronorite.
 74

Figure 3.2: Granular texture of the Chilas gabbro-norite. Thin section: left, transmitted light; right,
polarized light. Undeformed magmatic grains (note twins in feldspars, arrow, and the absence of
crystal-plasticity deformation features such as undulose extinction and subgrains) have a preferred
orientation (Sm) defining the magmatic foliation. pyx= pyroxene; fsp = feldspar. GPS: N35°26’09”;
E073°49’36.1”.

Locally abundant xenoliths are an important rock component aligned in the flow plane,
which, along with the mineral lineation, is interpreted to constitute the magmatic fabric
(Marre, 1982; Paterson et al., 1989). The xenoliths, which generally are darker and finer
grain than the host gabbronorite, may display an earlier but likely rotated fabric that was
not recorded. A strong evidence for xenolith alignment at sub-solidus stage is the
occurrence of boudinaged xenoliths and layers with plagioclase-rich veins, thus some melt,
filling the interboudin space (Figure 3.4 a,b).
 75

Figure 3.3: a) Normal faulting within layered sequences developed in the lower layers whereas it is
absent in the uppermost one indicating a syn-depositional origin. b & c) Truncation within layered
gabbronorite are easily recognized. d) Truncation in homogeneous Gabbronorite are difficult to
recognize in the field. The contact (arrow) between two subsequent intrusion is marked by xenolith (X)
cut of by the younger intrusion. The two intrusion have different orientation of the magmatic fabric
(Sm). e) Field map of the southern part of the Khiner valley close to the confluence with the Indus
River. Regional scale truncation is indicated by the trend of the foliation. The strike changes within
meters from ~E-W orientation towards N-S. The transition is within an otherwise homogeneous
gabbronorite without lithological contrast, shear zone or fault identified in between in the field.

The magmatic fabric locally evolves into, and becomes overprinted by a solid-state fabric
(Figure 3.4 c) marked by internal elongation of crystals, including brittle boudinage of
pyroxene and amphibole. This is particularly true near the borders of the body (D2 of Bard,
1983a), where solid-state deformation eventually yields coalescing shear zones with a
distinct mylonitic, hornblende-rich microstructure. Consistency of solid-state and
magmatic foliations and lineations suggests that both were produced during the last
crystallisation stages and flow increments of the gabbronorite, which further indicates that
 76

at least some of the “metamorphic/tectonic” fabrics pertain to the emplacement history of

the gabbronorite as much as to the regional tectonics. However, to avoid ambiguity,
measurements of solid-state fabrics were excluded from the data set discussed in this work.

Figure 3.4: a) Boudinaged compositional layering. Arrow: primary unconformity (GPS: N35°24’20.4”;
E074°01’10.4”) . b) Parallel xenoliths (x) lying in the magmatic foliation. Arrows: feldspar vein along a
normal fault indicating brittle behaviour of xenoliths and overall extension at a syn-magmatic, pre-full
crystallisation stage (GPS: N35°32’25.2”; E073°31’27.1”). c) Solid state deformation of the Chilas
gabbronorite (GPS: N35°29’38.8”; E073°52’52.5”). d) Imbricate hornblendite dykes with several tens-
cm long mineral fibres and within-dyke zoning indicating subhorizontal opening (GPS: N35°00’21.5”;
E074°00’07.6”). Such dyke swarms suggests formation in a zone of continuous but episodic rifting.

3.3.2 Interpretation
Magmatic structures develop during ascent of the magma under the influence of gravity,
internal pressures and regional deformation. They are thus critical for understanding the
emplacement setting of plutons, which in turn influences interpretations of the structures
and metamorphism in the surrounding rocks (Berger and Pitcher, 1970; Paterson et al.,
1989). Viscous flow, ductile shearing and, ultimately, brittle fracture document the
transition from ductile magma to brittle, crystallised rock. Pre-full crystallisation normal
faults (Figure 3.2 a), structurally and kinematically consistent with the fabrics of the
surrounding rock, are found throughout the Chilas Complex. The plutonic bodies that make
the Chilas Complex commonly have length >10 times longer than width, which suggests a
general sheeted-dyke-like imbrication concordant with generally steep contacts (Figure 3.4
c). Subparallel, <1m thick, hornblendite dykes (Figure 3.4 d) intrude each-other and have
developed up to hundreds meters thick and kilometres long packages with up to 100%
 77

dyke frequency in several places amid the gabbronorite, which further substantiates the
interpretation of sheeted-dyke-like imbrication of Chilas intrusions. These magmatic
features confirm that the Chilas Complex was emplaced in an extensional environment
(Burg et al., 1998; Khan et al., 1993).
The magmatic foliation is generally parallel to the trend of the Chilas Complex, its internal
layering and its contacts, and the lineation plunges steeply (71 towards 225). Two main
domains are distinguished from the orientation of planar flow markers and foliations
(Figure 3.5). The northern, Chilas-Kandiah domain dips dominantly SSW, which implies
that the Kohistan Batholith (our definition, i.e. the north-Kohistan region) is the footwall
and the Southern Amphibolites the hanging wall of the Chilas intrusion. The western,
Kandiah-Dir domain strikes SSW-NNE and is subvertical to WNW-dipping. The facts that
magmatic minerals are mostly undeformed (Figure 3.2) and that lineation directions are
similar in both domains (Figure 3.6) suggest that the arcuate map view of the Chilas
gabbronorite and other Kohistan units (Figure 3.1) is primary rather than the result of
folding of a consolidated linear body. If this hypothesis is correct, then the Chilas Complex
represents a sheeted-dyke region where magmas filled the spaces created between the
footwall Kohistan Batholith and the hanging wall Southern Amphibolite Belt.

Figure 3.5: Orientation diagrams of poles to magmatic foliations of the Chilas Complex. Contour
intervals at 2, 4, 6, 8 and 10 per unit area. Average pole for the Chilas-Kandiah region: Plunge = 32
towards 014.
 78

Figure 3.6: Mineral/magmatic lineation data from the Chilas Complex. Same contour intervals as .
Average pole orientation 71 towards 225.

The controversy we provoke centers on whether the internal magmatic fabrics reflect
emplacement of a horizontal lopolith at the contact interface between sediments of the
Kohistan Batholith and the Southern Amphibolites (Bard, 1983a; Jan et al., 1984; Khan et
al., 1993; Khan et al., 1989) later tilted to the measured, steep attitude. We challenge this
interpretation from three types of arguments. Firstly, tilt to near-vertical, as would be the
case with an upright, isoclinal fold of some tens of kilometres amplitude (Coward et al.,
1987), is mechanically impossible and would involve intense, vertical foliation along the
hinge zone. Such a foliation does not exist and the two theoretical limbs display lithologies
and metamorphic grades that cannot be correlated, which is inconsistent with upright
folding. Secondly, tilt of a thick crustal block around a horizontal axis is equally
implausible owing to the lack of evidence for a vertical crustal sequence: sediments and
volcanites are well documented on both sides of the Chilas Complex, as mentioned above.
Thirdly, metamorphic pressures to the south (the floor of the supposed lopolith) of the 30
km thick Chilas Complex should be about 1.0 GPa higher than those in the north (the
eventual roof of the lopolith). Sediments and metavolcanites of the Southern Amphibolites
are low to medium grade, which precludes a crustal pile of > 30 km above them, and
indicates that fault zones along the southern boundary of the gabbronorite had a limited
 79

throw. Sediments and metavolcanites to the north have also amphibolite facies
parageneses. Pressure estimates based on the plag-grt-bio-qz barometer (Hoisch, 1990)
applied to a metapelitic sample (C226) from the Khiner River, less than one kilometer to
the north of the boundary with the Chilas gabbronorite, indicates peak metamorphic
conditions of ~0.7 GPa and ~700°C. We conclude that the interpretation of a large,
originally horizontal lopolithic gabbronorite body is at odds with the geological
information.

3.4 Hornblende-plagioclase pegmatite veins

3.4.1 Description
Foliation-parallel and foliation and/or lineation-normal hornblende-plagioclase pegmatite
veins represent cross and longitudinal joints originated when the magma cooled
(Bergbauer and Martel, 1999; Marre, 1982). Those pegmatite veins indicate that brittle
deformation occurred in the presence of residual melt during intrusion of the main body.
They are cut by hornblende-bearing pegmatite dykes that bear no simple relationship to the
magmatic fabric but may also pertain to the latest crystallisation stages of the Chilas
Complex since they yielded 39Ar-40Ar ages of 80-90 Ma (D. Rex, 1985, in Coward et al.,
1987). Although Ar-Ar ages older than the 83-85 Ma emplacement age of the gabbronorite
are questionable, the 80 Ma age bracket places a bound to rapid cooling down to ~500°C
(Harrison, 1981). Using strike variations as gauge of the ratio of magmatic to tectonic
stresses (Baer et al., 1994; Delaney et al., 1986; Jolly and Sanderson, 1997) and the
strength of coarse norite (Hoek et al., 1998) we calculated a magma pressure of 0.2 to 0.3
GPa at a depth equivalent to 0.6 GPa. We will discuss swarms of thin (a few centimetres
up to less than 40 cm) and long (more than 50 m) pegmatite veins (Figure 3.7), which,
owing to their particularly high length/width ratio, are ascribed to post-crystallisation
brittle failure.
 80

Figure 3.7: Landscape (left) and outcrop (right) views of subvertical pegmatite veins. White arrow:
thin tip of one of these extensional veins, interpreted as standing in its near-original attitude and that
cut older magmatic joints (black arrow). Outcrop at, and view from GPS: N35°30’45.3”;
E073°23’30.1”.

3.4.2 Interpretation
Pegmatite veins of the latest generation have a near-vertical dip, which we interpret to be
their near-original attitude for the following argument (Figure 3.7):

Their remarkably planar shapes with very large length/width ratios and lack of vein-
parallel offsets of earlier structures indicate that those veins represent fluid-filled
extensional fractures. Such single, linear fractures tend to open perpendicular to the least
principal stress and dilate parallel to the maximum principal stress (Pollard and Johnston,
1973; Price and Cosgrove, 1990; Segall, 1984). The nearly vertical attitude of the thin
pegmatite veins thus suggests that the largest principal stress under which they formed has
a present-day vertical orientation. Is it coincidence after general rotation of the arc about a
generally subhorizontal axis? Vertical, long and narrow vein sets have a wide range of
different strikes; hence, they would require different rotation axes (horizontal axes are
parallel to the strike direction) which, by chance, have each the ad-hoc orientation to bring
individual vein sets to near-vertical dips (Figure 3.8). The regional continuity of host rock
structures (magmatic foliations in particular) excludes the existence of multiple rotation
axes. An important corollary is that those veins have not or little been rotated around
 81

horizontal axes since they formed; consequently the magmatic structures measured in the
host Chilas Complex, out of influence of the Nanga Parbat Syntaxis, are in a near-original
attitude, eventual rotations around vertical axes excepted. As a matter of fact, the
extensional nature of the Chilas Complex is theoretically consistent with a vertical greatest
principal stress (Price and Cosgrove, 1990), as suggested by the vertical vein sets.

Figure 3.8: Orientation diagrams of poles to veins within the Chilas Complex. a) Original orientation
as measured in the field. b) The measured poles rotated 30° around E-W horizontal axis to reconstruct
the supposed original orientation before tilting of the arc. Due to the different strike direction the sub-
vertical orientation can not be maintained with a single rotation axis. If rotation occurred after the
formation of the veins, the veins would have had a rather illogical arbitrary orientation before rotation
and are than rotated into a symmetric vertical orientation by chance. Alternatively, each set of dyke
orientation were rotated around a different axis to maintain the sub vertical orientations.

3.5 Discussion
The primarily arcuate trace of the Chilas Complex is reminiscent of the shape of listric
fault systems with a neighbouring transfer or transform fault (Hossack, 1984; Vendeville,
1991). Taking into account the size of the Chilas Complex, the magmatic structures
reported in this work document the listric shape of a within-arc extensional system on a
lithospheric scale, an event that we tentatively attribute to slab rollback (Figure 3.9, see
also Treloar et al., 1996). Tilting of the hanging wall is an important consequence of listric
geometries (Mauduit and Brun, 1998; McClay et al., 1991). On a bulk southward-dipping
listric fault, like the Chilas body, one expects antithetic rotation resulting in a bulk
northward dip of the hanging wall. We suggest that the general attitude of the Southern
Amphibolites is essentially inherited from rollover tilting, which took place when the
 82

Chilas Complex was emplaced at about 85 Ma. This working hypothesis is consistent with
the fact that the Chilas gabbronorite, which has arguably preserved its original dip, cuts the
north-dipping Southern Amphibolites on its southern boundary (Treloar et al., 1996). This
pre-India-Kohistan collision event, which could be the “yet unrecognized tectonic event”
(Ringuette et al., 1998) responsible for partial exhumation of the base of the Kohistan
39
terrane, would also provide an explanation for the regionally distributed Ar/40Ar
cooling/unroofing ages at 83-80 Ma, significantly older than the 60-50 Ma old India-Asia
collision (Treloar et al., 1996; Treloar et al., 1989). Pre-collision tilting also explains how
the basal Indus Suture climbs northwards, from mantle peridotites to upper crustal
sediments, up section of the Southern Amphibolites (Figure 3.1), while there is no obvious
reason for this contact to be a backthrust between the two northward dipping sutures of the
collisional system; the observed geometry reflects inherited inclination of the southern arc.
Accepting that the southern Kohistan Complex was tilted before collision and kept its bulk
orientation during later tectonics is further consistent with still flat-lying, 40-30 Ma old
(Teru) lava flows in northwestern Kohistan (Danishwar et al., 2001; Petterson and Treloar,
2004) and analogue modelling of arc-accretion, which do not involve much rotation during
collision (Boutelier et al., 2003).

Figure 3.9: Tectonic interpretation of the Kohistan Arc Complex in late Cretaceous times, modified
from Burg et al. (1998).

3.6 Conclusion
Magmas of the Chilas Complex exploited the spaces created during within-Kohistan-Arc
extension at about 85 Ma and were deformed in the process. An oceanward-dipping
rupture dominated extension and magma emplacement and compelled pre-collision tilt of
the hanging wall crustal block, now represented by the Southern Amphibolites. This study
illustrates three general points pertinent to present and past island arcs: (1) within-arc space
can be made available for the emplacement of magmas, so that the lower crust of an
extended arc appears as a large-scale sheeted-dyke complex; (2) rotation of large crustal
 83

blocks representing the hanging wall of lithospheric-scale listric fault zones may occur
during the arc history and is responsible for older-than-collision cooling/ exhumation ages
and (3) pre-collision arc-structures may be preserved in collision orogens but are unrelated
to continental collision processes per se.

Acknowledgements: The Swiss National Science Foundation supports our work (grants
20-49372.96 and 20-61465.00). P. Treloar provided critical and useful comments on an
earlier version of the paper. The Pakistan Museum of Natural History supports SH and HD.
 84

References

Ali, A., E. Nakamura, and H. Yamamoto, 2002, Sm-Nd mineral ages of pegmatite veins
and their host rocks from Swat area, Chilas Complex, northern Pakistan, Journal
of Asian Earth Sciences, p. 331-339.
Baer, G., M. Beyth, and Z. Reches, 1994, Dikes emplaced into fractured basement, Timna
Igneous Complex, Israel, Journal of Geophysical Research, p. 24039-24050.
Balk, R., 1959, Structural behavior of igneous rocks, Geological Society of America
Memoir, p. 177 p.
Bard, J.-P., 1983a, Metamorphic evolution of an obducted island arc: Example of the
Kohistan Sequence (Pakistan) in the Himalayan collided range, Geological
Bulletin, University of Peshawar, p. 105-184.
Bard, J.-P., 1983b, Metamorphism of an obducted island arc: Example of the Kohistan
sequence (Pakistan) in the Himalayan collided range, Earth and Planetary Science
Letters, p. 133-144.
Bard, J.-P., H. Maluski, P. Matte, and F. Proust, 1980, The Kohistan sequence: crust and
mantle of an obducted island arc, Geological Bulletin of the University of
Peshawar, p. 87-94.
Bergbauer, S., and S. J. Martel, 1999, Formation of joints in cooling plutons, Journal of
Structural Geology, p. 821-835.
Berger, A. R., and W. S. Pitcher, 1970, Structures in granitic rocks: a commentary and a
critique on granite tectonics, Proceedings of the Geologist's Association, London,
p. 441-461.
Boutelier, D., A. Chemenda, and J.-P. Burg, 2003, Subduction versus accretion of intra-
oceanic volcanic arcs: insight from thermo-mechanical analogue experiments,
Earth and Planetary Science Letters, p. 31-45.
Burg, J.-P., J.-L. Bodinier, M. N. Chaudhry, S. Hussain, and H. Dawood, 1998, Infra-arc
mantle-crust transition and intra-arc mantle diapirs in the Kohistan Complex
(Pakistani Himalaya): petro-structural evidence, Terra Nova, p. 74-80.
Butler, R. W. H., M. George, N. B. W. Harris, C. Jones, D. J. Prior, P. J. Treloar, and J.
Wheeler, 1992, Geology of the northern part of the Nanga Parbat massif, northern
Pakistan, and its implications for Himalayan tectonics, Journal of the Geological
Society of London, p. 557-567.
 85

Coward, M. P., R. W. H. Butler, M. A. Khan, and R. J. Knipe, 1987, The tectonic history
of Kohistan and its implications for Himalayan structure, Journal of the
Geological Society of London, p. 377-391.
Coward, M. P., M. Q. Jan, D. C. Rex, J. Tarney, M. Thirlwall, and B. F. Windley, 1982,
Geotectonic framework of the Himalaya of N. Pakistan, Journal of the Geological
Society of London, p. 299-308.
Coward, M. P., B. F. Windley, R. D. Broughton, I. W. Luff, M. G. Petterson, C. J. Pudsey,
D. C. Rex, and M. A. Khan, 1986, Collision tectonics in the NW Himalayas, in A.
C. Ries, ed., Collision tectonics, London, Geological Society Special Publication,
p. 203-219.
Danishwar, S., R. J. Stern, and M. A. Khan, 2001, Field relations and structural constraints
for the Teru volcanic formation, northern Kohistan Terrane, Pakistani Himalaya,
Journal of Asian Earth Sciences, p. 683-695.
Delaney, P. T., D. D. Pollard, J. I. Ziony, and E. H. McKee, 1986, Field relations between
dikes and joints: Emplacement processes and paleostress analysis, Journal of
Geophysical Research, p. 4920-4938.
Harrison, T. M., 1981, Diffusion of 40Ar in Hornblende.: Contributions to Mineralogy and
Petrology, v. 78, p. 324-331.
Hoek, E., P. Marinos, and M. Benissi, 1998, Applicability of the geological strength index
(GSI) calssification for very weak and sheared rock masses. The case of the
Athens Schist Formation, Bulletin of Engineering Geology and the Environment,
p. 151-160.
Hossack, J. R., 1984, The geometry of listric growths faults in the Devonian basins of
Sunnfjord, W Norway, Journal of the Geological Society of London, p. 629-637.
Hutton, D. H. W., and R. J. Reavy, 1992, Strike-slip tectonics and granite petrogenesis,
Tectonics, p. 960-967.
Jan, M. Q., 1988, Geochemistry of amphibolites from the southern part of the Kohistan arc,
N. Pakistan, Mineralogical Magazine, p. 147-159.
Jan, M. Q., and R. A. Howie, 1980, Ortho- and clinopyroxenes from the pyroxene
granulites of Swat Kohistan, northern Pakistan, Mineralogical Magazine, p. 715-
726.
Jan, Q. M., M. U. K. Khattak, P. M.K., and B. F. Windley, 1984, The Chilas Stratiform
Complex: field and mineralogical aspects, Geological Bulletin, University of
Peshawar, p. 153-169.
 86

Jolly, R. J. H., and D. J. Sanderson, 1997, A mohr circle construction for the opening of a
pre-existing fracture, Journal of Structural Geology, p. 887-892.
Khan, M. A., M. Q. Jan, and B. L. Weaver, 1993, Evolution of the lower arc crust in
Kohistan, N. Pakistan: temporal arc magmatism through early, mature and intra-
arc rift stages, in M. P. Searle, ed., Himalayan tectonics, London, Geological
Society Special Publication, p. 123-138.
Khan, M. A., M. Q. Jan, B. F. Windley, J. Tarney, and M. F. Thirlwall, 1989, The Chilas
mafic-ultramafic igneous complex: the root of the Kohistan Island Arc in the
Himalaya of northern Pakistan, Geological Society of America, Special Paper, p.
75-93.
Marre, J., 1982, Méthodes d'analyse structurale des granitoïdes, Manuels et Méthodes,
Orléans, Bureau de recherches géologiques et minières, p. 128.
Mauduit, T., and J.-P. Brun, 1998, Growth fault/rollover systems: Birth, growth, and
decay, Journal of Geophysical Research, p. 18119-18136.
McBirney, A. R., and A. Nicolas, 1997, The Skaergaard layered series. Part II. Magmatic
flow and dynamic layering, Journal of Petrology, p. 569-580.
McClay, K. R., D. A. Waltham, A. D. Scott, and A. Abousetta, 1991, Physical and seismic
modelling of listric normal fault geometries., Geological Society Special
Publication, p. 231-239.
Paterson, S. R., R. H. Vernon, and O. T. Tobisch, 1989, A review of criteria for the
identification of magmatic and tectonic foliations in granitoids, Journal of
Structural Geology, p. 349-363.
Petford, N., 1996, Dykes or diapirs?, Transactions of the Royal Society of Edinburgh:
Earth Sciences, p. 105–114.
Petterson, M. G., and P. J. Treloar, 2004, Volcanostratigraphy of arc volcanic sequences in
the Kohistan arc, North Pakistan: volcanism within island arc, back-arc-basin, and
intra-continental tectonic settings, Journal of Volcanology and Geothermal
Research, p. 147-178.
Pollard, D. D., and A. M. Johnston, 1973, Mechanics of growth of some laccolithic
intrusions in the Henry mountains, Utah, II : Bending and failure of overburden
layers and sill formation, Tectonophysics, p. 311-354.
Price, N. J., and J. W. Cosgrove, 1990, Analysis of geological structures, Cambridge,
Cambridge University Press, p. 502.
 87

Ringuette, L., J. Martignole, and B. F. Windley, 1998, Magmatic crystallization, isobaric

cooling, and decompression of the garnet-bearing assemblages of the Jijal
sequence (Kohistan terrane, western Himalayas), Geology, p. 139-142.
Sawada, Y., A. B. Kausar, K. Kubo, Y. Takahashi, and Y. Takahashi, 1994, Sedimentary
structures in the Chilas igneous complex of the Kohistan arc, northern Pakistan,
Earth Science, p. 33-38.
Schaltegger, U., G. Zeilinger, M. Frank, and J.-P. Burg, 2002, Multiple mantle sources
during island arc magmatism: U-Pb and Hf isotopic evidence from the Kohistan
arc complex, Pakistan, Terra Nova, p. 461-468.
Searle, M. P., A. M. Khan, J. E. Fraser, S. J. Gough, and Q. M. Jan, 1999, The tectonic
evolution of the Kohistan-Karakoram collision belt along the Karakoram Highway
transect, north Pakistan, Tectonics, p. 929-949.
Segall, P., 1984, Formation and growth of extensional fracture sets, Geological Society of
America Bulletin, p. 454-462.
Tahirkheli, R. A. K., M. Mattauer, F. Proust, and P. Tapponnier, 1979, The India Eurasia
Suture Zone in Northern Pakistan: Synthesis and interpretation of recent data at
plate scale, in K. A. De Jong, ed., Geodynamics of Pakistan, Quetta, Geological
Survey of Pakistan, p. 125-130.
Treloar, P. J., M. G. Petterson, M. Qasim Jan, and M. A. Sullivan, 1996, A re-evaluation of
the stratigraphy and evolution of the Kohistan arc sequence, Pakistan Himalaya:
implications for magmatic and tectonic arc-building processes, Journal of the
Geological Society of London, p. 681-693.
Treloar, P. J., D. C. Rex, P. G. Guise, M. P. Coward, S. M.P., B. F. Windley, M. G.
Petterson, M. Q. Jan, and I. W. Luff, 1989, K/Ar and Ar/Ar geochronology of the
Himalayan collision in NW Pakistan: constraints on the timing of suturing,
deformation, metamorphism and uplift, Tectonics, p. 881-909.
Vendeville, B., 1991, Mechanisms generating normal fault curvature: a review illustrated
by physical models, in B. Freeman, ed., The geometry of normal faults, London,
Geological Society, p. 241-249.
Yamamoto, H., and E. Nakamura, 1996, Sm-Nd dating of garnet granulites from the
Kohistan complex, northern Pakistan, Journal of the Geological Society of
London, p. 965-969.
Yoshida, M., H. Zaman, and M. N. Ahmad, 1996, Paleopositions of Kohistan Arc and
surrounding terranes since Cretaceous time: the paleomagnetic constraints,
 88

Proceedings of Geoscience Colloquium, Geoscience Laboratory Project,

Islamabad, Pakistan, p. 83-101.
Zeitler, P., R. A. K. Tahirkheli, C. Naeser, N. Johnson, and J. Lyons, 1980, Preliminary
fission track ages from the Swat valley, Northern Pakistan, Geological Bulletin,
University of Peshawar, p. 63-65.
 89

4 Petrology
This chapter describes petrological aspects of rocks of the Chilas Complex. An entire
thesis of Khan (1988) has been previously largely devoted on this subject. The chemistry
data of rock forming minerals within the Chilas Complex has been additionally presented
by several authors (Bard, 1983; Jan and Howie, 1980; Jan and Howie, 1982; Jan et al.,
1992; Khan et al., 1989). In this chapter complementing mineral analyses are presented.
The main focus is on the trace element composition of minerals from different units of the
Chilas Complex, which have not been analysed before.

4.1 Texture
Generally, peridotite and gabbronorite samples are extremely fresh and occasional
alteration is restricted along joints. All peridotites are garnet-free spinel-peridotite.
Amphibole-bearing porphyroclastic lherzolite, with an inequigranular - polygonal to
interlobate microstructure have complicated textural relationships with interlobate
clinopyroxene and relict orthopyroxene porphyroclasts showing exsolution lamellae of the
complementary pyroxene and Cr-spinel. Textural relationships indicate that amphibole
occurs as replacement of orthopyroxene, both of which are in turn replaced by olivine
(Figure 4.1 a). Granoblastic olivine has a bimodal grain size distribution whereby large
olivine grains show undulose extinction. Smaller recrystallized neoblastic olivine grains
have a granoblastic texture and show no feature of crystalline plasticity. Trails of small
olivine grains crosscut the larger ones and occur between pyroxene grains, rarely
penetrating into pyroxene grains (Figure 4.1 b & c); they form an intergranular network of
olivine between large pyroxene and amphibole porphyroclasts. The orientation of
exsolution trails within pyroxene can be preserved in olivine replacing pyroxene (Figure
4.1 d).
 90

Figure 4.1: Photomicrographs of texture within a porphyroclastic lherzolite (Sample C03-45) with
mineral abbreviations (Kretz, 1983) a) The embayment of amphibole into the orthopyroxene and the
embayment of olivine into amphibole and orthopyroxene indicates that amphibole replaces
orthopyroxene and subsequently both are replaced by olivine. b) Trails of small unstrained olivine
neoblast crosscutting larger olivine grains with undulose extinction. c) Trails of neoblastic olivine
grains along grain boundaries between porphyroclastic orthopyroxene and clinopyroxene d)
Orientation of exsolution trails of spinel in orthopyroxene are preserved in olivine grains embaying the
corresponding orthopyroxene indicating a replacive origin for the olivine.

Towards the contact, the modal amount of olivine increases and pyroxene decreases within
the porphyroclastic lherzolite. However, the actual contact between replacive dunite and
porphyroclastic lherzolite is sharp. Dunite has a granoblastic texture with olivine grains
that are virtually undeformed. Crosscutting trails of small olivine neoblasts within larger
porphyroclasts indicate multiple recrystallisation of the olivine (Figure 4.2 a). Accessory
opaque minerals, Cr-spinel, Cr-bearing magnetite and Fe-Ni sulfide are present as
inclusions within olivine or along grain-boundaries (Figure 4.2 b).
 91

Figure 4.2: Photomicrographs of texture within the dunite a) Trails of smaller olivine neoblasts
crosscutting a larger olivine grain (C35). b) Cr-rich spinels along grain boundaries of granoblastic
olivine grains (C174).

Towards the contact with the surrounding gabbronorite, the dunite gradually transforms
into lherzolite and pyroxenite, and plagioclase can locally be present. Minor pyroxene, and
to a lesser extent amphibole, occurs either as centimeter sized clusters along grain
boundaries of olivine crystals or as trails (Figure 4.3 a, b). With increasing pyroxene
content an intergranular network of partly exsolved porphyroblastic clinopyroxene grains
between larger olivine crystals is developed (Figure 4.3 c). In pyroxenite occurring close to
the contact with the surrounding gabbronorite, olivine grains occur as relicts between large
pyroxene crystals (Figure 4.3 d). First occurrence of plagioclase is along triple junctions of
larger olivine grains. Symplectitic coronas are systematically developed along olivine-
plagioclase contacts. Within these coronas, orthopyroxene nucleates close to the olivine
contact and pleonastic spinel form symplectites with clinopyroxene and amphibole towards
the plagioclase (Figure 4.3 e). In places, plagioclase can be completely replaced by
symplectites forming blobs or vermicular trails between olivine grains (Figure 4.3 f).
Accessory Cr-spinel, magnetite and Fe-Ni sulfide are present throughout the transition
zone. Close to the contact with the gabbronorite, orthopyroxene appears as large
(centimeter-scale) poikiloblast with abundant olivine (Figure 4.3 g), clinopyroxene and
plagioclase inclusions (Figure 4.3 h).
 92

Figure 4.3: Photomicrographs of textures for the gradual dunite-lherzolite-pyroxenite transition a)

vein like shaped trails of pyroxene grains (in this case opx) along grain boundaries of granoblastic
olivine. b) With increasing pyroxene content prominent veins are develop. c) With further increasing
pyroxene content larger porphyroblastic pyroxene can be present. d) Within pyroxenite close to the
actual contact with surrounding gabbronorite olivine grains (high birefringence colors) are only relicts
between large pyroxene crystals. e) Two-pyroxene-spinel symplectite mineral reaction developed
between plagioclase and olivine. f) Same symplectitic reaction without the presence of plagioclase,
 93

which was completely transformed. g) Large (centimeter-scale) poikilitic orthopyroxene with olivine
inclusion and h) with an hypidiomophic plagioclase inclusion

The gabbronorite displays a predominantly granoblastic fabric, composed of plagioclase,

clinopyroxene and orthopyroxene with an intergranular texture (grain size ~ 0.5 cm)
(Figure 4.4). Pyroxene is frequently rimed by amphibole. Partly hypidiomorphic to
idiomorphic pyroxenes and plagioclase are locally preserved. Twinning within plagioclase
is very common and typically continuous, with parallel sides. Undulose extinction is
generally absent in all minerals; rarely undulose quartz is observed. Layering, if present is
mainly defined by the modal amount of plagioclase and pyroxenes.

Figure 4.4: Photomicrographs of textures in the gabbronorite a) actual contact between ultramafite,
here mainly orthopyroxene on the left and gabbronorite on the right. b) Texture within gabbronorite
composed of plagioclase (white), orthopyroxene (reddish) and clinopyroxene (greenish). Pyroxenes are
frequently rimed by amphibole (green).

4.2 Mineral Chemistry of rocks within the Chilas Complex

4.2.1 Method
Major element chemistry of minerals was determined using an electro-microprobe (EMP)
(Cameca SX 50 and Jeol Superprobe) and trace elements have been analyzed with a laser-
ablation inductively coupled plasma mass spectrometer (LA-ICP-MS) both at ETH Zurich.
Operating parameters for the EMP include acceleration voltage of 15 kV, beam current of
20nA and a beam size of 1 µm. Measuring time were 60 sec for Ni and 20 sec for the
remaining elements. Operating conditions for the LA-ICP-MS are given elsewhere (Pettke
et al., 2000).
 94

Figure 4.5: Representative compositional profiles through the main rock forming minerals present in
the Chilas Complex. A) Profile through different neighboring plagioclase grains in gabbronorite and in
plagioclase-bearing lherzolite. Plagioclase from gabbronorite (C41) is relatively homogeneous and
partly a slight positive zonation with An poor rims is developed. Profound reverse zoning however, is
developed in plagioclase from ultramafite samples (C194). B) Olivine from ultramafite (C33) is
generally homogeneous. C) Cores of clinopyroxene (C169) are relatively homogeneous whereas the rim
displays slight zoning. The Al2O3 content can scatter significantly but systematic compositional zoning
 95

is present only in the outermost rim. D) Profile through poikilitic orthopyroxene (C194) with a
homogeneous composition but with Al2O3 slightly decreasing against clinopyroxene inclusions.

4.2.2 Results
The results are tabulated in Appendix IV and graphically presented below. In general
minerals are homogeneous in major elements chemistry but slight core to rim zoning can
be present (Figure 4.5). Profound reverse zoning is found in plagioclase present in
lherzolite and has been attributed due to diffusion exchange of Na between plagioclase and
amphibole (Khan et al., 1989).
In the following paragraphs the chemical characteristic of the most important investigated
minerals will be described:

4.2.3 Plagioclase
The anorthite (An) content of plagioclase indicates a continuous overlapping sequence
from An91 to An48 for the gabbronorite (including tonalite occasionally present within the
gabbronorite) and An98-An80 for plagioclase-bearing lherzolite (Figure 4.6). The
plagioclase chemistry of fine grained amphibolitic dykes have a much lower potassium
content than plagioclase from the gabbronorite with equal anorthite content.

1.00
main gabbronorite
plagiclase bearing lherzolite
ol-bearing gabbronorite
Khan main gabbronorite
0.90
Khan UMA
Hbl bearing dykes
UMA

0.80
Anorthite

0.70

0.60 main gabbronorite

0.50

0.40
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80
K2O [wt%]

Figure 4.6: Plagioclase composition of tonalite, gabbronorite, olivine-bearing gabbronorite and

plagioclase-bearing lherzolite compared with results from (Khan, 1988). The apparent gap invoked by
 96

Khan et al. (1989) in plagioclase composition between UMA (red bar) and gabbronorite (brown bar)
between An83 and An64 is not verified.
The trace elements concentrations have been determined on 6 to 9 mineral grains separated
from four different samples (two gabbronorite (C7 & C 48), an olivine-bearing
gabbronorite (C66) and one plagioclase-bearing lherzolite (C218). The averaged, primitive
mantle normalized (Sun and McDonough, 1989), trace elements patterns of all analyses are
very similar (Figure 4.7). The trace element concentrations increase with decreasing An-
content from plagioclase-bearing lherzolite (An96) to gabbronorite (An56) by approximately
a factor of 10. The incompatible elements are variably enriched (1 to 30 times primitive
mantle) and HREE are strongly depleted (0.01 to 0.1 times primitive mantle). Ba, Pb, Sr,
Eu display a positive anomaly with respect to light rare earth (LREE).

Figure 4.7: Primitive Mantle (Sun and McDonough, 1989) normalized average trace element
composition of plagioclase from gabbronorite (C7 & C48), olivine-bearing gabbronorite (C66) and a
plagioclase-bearing lherzolite (C218). The grey line indicates the detection limit of LA-ICP-MS. Errors
are given as 2σmean. Values at or below the detection limit have been omitted.

4.2.4 Olivine
Olivine is a common mineral within the ultramafic units and within olivine-bearing
gabbronorite. It is absent in the large mass of main gabbronorite. The gradual dunite-
lherzolite-pyroxenite transition is associated with a constant decrease in modal content of
olivine, associated with a decrease in olivine Mg# (Mg# = Mg/(Fe+Mg)) from 0.92 to
0.77. A good correlation is observed between the Mg# and the modal amount of pyroxene.
 97

The Mg# asymptotically approaches the olivine Mg# value measured in olivine-bearing
gabbronorite, exposed at the contact between ultramafite and gabbronorite, with decreasing
modal olivine (Figure 4.8).
The NiO content of olivine scatters significantly between 0 and 0.3 [wt%] similar to other
ultramafic complexes (e.g. Cabo Ortegal (Spain) Santos et al., 2002). Significant variations
are observed within single grains even if there is no systematic zoning. The NiO and MnO
contents for a given Mg# are systematically lower than average mantle olivine
concentration (Takahashi et al., 1987), which indicates a magmatic origin for the olivine. A
compositional gap exists between Mg# 0.87 and 0.85 (Figure 4.9). The Mg# in Mg-rich
olivine is positively correlated with MnO content and the spread in MnO concentrations is
relatively restricted. In the low Mg# olivine this correlation is weakly developed until Mg#
0.84. For olivine with lower Mg# values no correlation with MnO exists and MnO
concentrations are about two times more variable.

0.95

0.9

0.85
Mg#

0.8
Mg# in Ol-Gabbro

0.75

0.7

0.65
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Modal Ol/(Ol+Pyx)

Figure 4.8: Relationship between modal abundance of Olivine and Pyroxene versus the Mg# in olivine
in dunite (circles), lherzolite (triangles), plagioclase bearing lherzolite (open triangles) and olivine
bearing gabbronorite (red squares). With increasing pyroxene content the Mg# number decreases
approaching asymptotically the Mg# of olivine found in gabbronorite at the mafic - ultramafic contact.
 98

0.40

Porphyric lherzolite
0.35
lherzolite
dunite
0.30 ol-bearing gabbronorite
Ultramafic dykes

0.25
MnO [wt%]

0.20

0.15

0.10

0.05

0.00
0.760 0.780 0.800 0.820 0.840 0.860 0.880 0.900 0.920 0.940
Mg#

Figure 4.9: Olivine composition of ultramafite and olivine-bearing gabbronorite. A compositional gap
exists between Mg# 0.87 and 0.85. High Mg# olivine has a positive correlation with MnO content and
relatively restricted concentrations, whereas low Mg# olivine have no correlation with MnO and the
MnO concentrations are about two times more variable.

4.2.5 Clinopyroxene
Clinopyroxene is extremely rich in Mg with Mg# as high as 0.95. Ternary endmember
composition plots of orthopyroxene and clinopyroxene indicate only a moderate Fe
enrichment with differentiation (Khan et al., 1989). The Al2O3 content varies
systematically within clinopyroxene and can be separated in an igneous trend defined by
core compositions or unzoned clinopyroxene and a metamorphic trend defined by core to
rim zoning (Figure 4.10). Within the igneous trend the Al2O3 content of clinopyroxene
from samples devoid of plagioclase increases with decreasing Mg#. Opposite to that, the
Al2O3 content of clinopyroxene coexisting with plagioclase decreases with decreasing
Mg#. The increasing aluminum trend has been previously noticed in pyroxenes from the
Aleutian volcanic rocks (Kay and Kay, 1985), from lower crustal cumulates (Debari et al.,
1987), in the Tonsina ultramafic-mafic assemblage (Debari and Coleman, 1989) and
experimentally in studies on pyroxenites (Müntener et al., 2001). However, this enrichment
is not known from low-pressure cumulates like Skaergaard (Debari and Coleman, 1989).
The generally accepted explanation is that plagioclase crystallisation is suppressed by high
 99

total pressure and high water pressure (Yoder and Tilley, 1962). The crystallisation of Al-
poor phases like olivine and pyroxene enriches the corresponding melt in Al. Higher
pressure in Tonsina (0.95-1.1 GPa (Debari and Coleman, 1989) than in Chilas (0.6-0.7 GPa
see paragraph 4.3.1) explains the slightly higher absolute Al2O3 content observed in
Tonsina (8.85 wt%) compared to Chilas (below 6 wt% excluding one analyses of 7.28 wt
%.), since solubility of Al in pyroxene increases with increasing pressure (Gasparik, 1984;
Kushiro and Yoder, 1966).
The metamorphic trend observed in Figure 4.10 is due to a Tschermak exchange of
(Fe,Mg)SiAl-2 during cooling. In terms of major element chemistry the clinopyroxene from
porphyroclastic lherzolite are characterized by low Na, Al, Ti and Cr and high Mg# and are
similar to primitive clinopyroxene from the lherzolite.

Figure 4.10: Al2O3 versus Mg# for clinopyroxene from the Chilas Complex. The solid line represents
an igneous trend determined from core composition. Grey arrows represent metamorphic cooling
observed in zoned cpx. The igneous trend is characterized by an increasing Al2O3 content with
decreasing Mg# until plagioclase is present in the rocks. Due to the fractionation of plagioclase the
Al2O3 content of clinopyroxene decreases with decreasing Mg#. The metamorphic trend represents a
Tschermak (Fe,Mg)SiAl-2 exchange during cooling.

Trace element characteristics of clinopyroxene (Figure 4.11) have been determined from
three samples of the porphyroclastic lherzolite (C03-43, C03-44 and C03-45), a weakly
 100

infiltrated lherzolite/dunite with high Mg# olivine (C174), a plagioclase-bearing lherzolite

(C 218), an olivine-bearing gabbronorite (C66) and two gabbronorite (C7 & C 48).
Clinopyroxene from all samples is characterized by incompatible trace element depletion
in respect to middle and heavy rare earth elements (MREE and HREE) and a negative Nb
and Ta anomaly. The absolute trace element concentrations are a factor of 7 to 10 higher in
clinopyroxene from gabbronorite than from the other samples. Gabbronoritic samples (C7
& C48) display an increasing more pronounced negative Pb, Sr and Eu spikes with
increasing trace element concentration indicating equilibration with plagioclase. Primitive
clinopyroxene from plagioclase-bearing lherzolite (C218), however, have positive Pb and
Sr spike indicating that clinopyroxene crystallized before plagioclase and that
clinopyroxene did not equilibrate with plagioclase. Clinopyroxene, in contrast to
plagioclase, shows within a single sample an increase in REE element concentration by a
factor of 2 while the Mg# varies only slightly.

100
gabbronorite (C48)
lherzolite/dunite (C174)
gabbronorite (C7)
plag beraing lherzolite (C218)
ol-bearing gabbronorite
porphyric lherzolite (C03-43)
Porphyric lherzolite (C03-44)
porphyric lherzolite (C03-45)
10

0.1
La Ce Pb Pr Sr Nd Zr Hf Sm Eu Gd Tb Dy Ho Y Er Yb Lu

Figure 4.11: Abundance of trace elements in clinopyroxene normalized to primitive mantle (Sun and
McDonough, 1989). Given is the average value of 6 to11 mineral grains analysed per sample. Error is
given as 2σmean.
 101

4.2.6 Orthopyroxene
Orthopyroxene has enstatitic composition with Mg# ranging in the ultramafite from 0.84 to
0.72 and in the gabbronorite from 0.76 to 0.56. The relationship between Al2O3 and Mg#
varies like in clinopyroxene (Figure 4.12) in accordance with previous measurements from
deep-seated intrusions (Debari and Coleman, 1989). The Fe/Mg of orthopyroxene and
olivine is systematically higher compared to the Fe/Mg in clinopyroxene. This discrepancy
increases with increasing Fe content (Figure 4.13). Orthopyroxene and olivine however,
are close to unity in Fe/Mg (not shown). Accordingly, this discrepancy is mainly due to the
Mg-rich composition of clinopyroxene. In high-temperature peridotite, the Fe/Mg between
olivine, orthopyroxene and clinopyroxene are close to unity (Obata, 1980). The deviation
from the 1:1 line could indicate equilibration at subsequent lower temperature with
increasing differentiation (Müntener, 1997).
The trace elements of orthopyroxene (Figure 4.14) are characterized by steep LREE to
HREE pattern and positive anomalies of U, Th and Pb. Orthopyroxene from the olivine-
bearing gabbronorite display a slight negative anomaly of Zr and Hf, contrasting
orthopyroxene from gabbronorite display a positive Zr and Hf anomaly. Those
orthopyroxene are accordingly not in equilibrium with clinopyroxene and could indicate
growth of orthopyroxene at the expense of olivine.
 102

Figure 4.12: Al2O3 versus Mg# for orthopyroxene from the Chilas Complex. Symbols are the same as
in Figure 4.10. Similar to cpx an igneous and a metamorphic trend is observed, however the
metamorphic Tschermak exchange is weaker in opx than in clinopyroxene.

Figure 4.13: Plot of Fe/Mg in clinopyroxene against Fe/Mg in olivine and orthopyroxene. The 1:1 line
is given for comparison only. All data consistently plot over the 1:1 line. The misfit with unity increases
with increasing Fe content.
 103

Figure 4.14: Abundance of trace elements in orthopyroxene normalized to primitive mantle (Sun and
McDonough, 1989). Given is the average value of 5 to10 mineral grains analysed per sample. Errors
are given as 2σmean, the grey line is the detection limit.

4.2.7 Spinel
The spinel composition has been discussed in detail by (Jan et al., 1992). Within the
ultramafite spinels are mainly Cr-spinel, hercynite and Cr-bearing magnetite whereas in the
gabbronorite sequence spinels are mainly illmenite, ulvospinel and magnetite. A distinctive
data density minimum is developed in the spinel composition (Figure 4.15). Spinels from
dunite (with high Mg# in olivine) are partly offset from the solvus, shifted towards the Cr-
apex, approaching the trend defined by Cr-spinel from the Jijal complex (Jan and Windley,
1990). The spinels show a wide range in Mg# and Cr# whereas the former increases when
the later decreases. A data minimum is also present (Figure 4.16). Spinels from the high
Mg# dunitic samples display a Mg# enrichment trend with constant Cr#. A similar trend at
higher Cr# values is observed in spinel from the Jijal complex (Jan and Windley, 1990).
Within the gabbronorite, spinel constitutes less than 3 vol % of the rock. An apparent
continuous solid-solution between ilmenite and ulvospinel and between ulvospinel and
 104

magnetite is developed (Figure 4.17). Cr-bearing spinels are present within gabbronorite in
the vicinity of the ultramafite outcrops.
The data density minimum observed has been interpreted as a solvus in the spinel solid
solution due to decreasing temperature (Jan et al., 1992). The solvus is mainly defined by
spinels from reacted lherzolite, dunite and the few Cr-spinels found in gabbronoritic
samples. However, the rather patchy and irregular exsolution texture of spinels indicate
that exsolution is magmatic. If exsolution were associated with later cooling, the exsolution
pattern would follow the regular cubic crystal structure. Accordingly, we interpret spinels
to crystallize along or near a solvus intersecting the solidus at higher pressure.

Figure 4.15: Ternary plot (Fe3+-Cr-Al) illustrating spinel composition from the Chilas Complex
compared with data from the Jijal Complex (Jan and Windley, 1990)
 105

1.00

0.90

0.80

0.70

0.60

Cr# (Cr/(Cr+Al)
0.50

0.40

0.30

0.20
porphyroclastic lherzolite
dunite
0.10
lherzolite
Jijal
0.00
0.70 0.60 0.50 0.40 0.30 0.20 0.10 0.00
Mg# (Mg/(Mg+Fe2+)

Figure 4.16: Spinel compositions in the ultramafite compared to spinel composition in the Jijal
complex (Jan and Windley, 1990).

Figure 4.17: Spinel composition of the main gabbronorite. Solid squares are own analyses and opened
circles are from Kha, (1989). The analysed spinels form solid solution between ilmenite and ulvospinel
and between ulvospinel and magnetite.
 106

4.2.8 Amphibole
The textural relationship of amphibole, riming pyroxene within the gabbronorite, indicates
that amphibole represents a late magmatic phase or an early metamorphic overgrowth.
Within the lherzolite magmatic amphibole interstitial between large olivine grains gets
replaced by metamorphic amphibole (Figure 4.18). The composition of magmatic
amphibole is dominantly pargasitic-, tschermakitic- to magnesium-hornblende (Figure
4.19). The trace element characteristics have been determined from amphibole of the
porphyroclastic lherzolite (Figure 4.20). The trace element pattern of Mg-hornblende
present in porphyroclastic lherzolite is very similar to that of clinopyroxene indicating
equilibrium between the two minerals. Accordingly a magmatic origin of the magnesio-
hornblende is likely. The trace element pattern is characterized by LREE depletion and a
negative Hf and Zr anomaly. The lack of a negative Eu and Sr anomaly indicate the
amphibole did not equilibrate with plagioclase.

Figure 4.18: Backscatter image of Sample C03-15 illustrating replacement of cpx and Mg-hornblende
by tremolitic amphibole. Mg-Hornblende represents an early magmatic phases whereas Tremolite
replaces Mg-hornblende and cpx and is a later stage metamorphic overgrow.
 107

Figure 4.19: Magmatic amphibole composition within various units of the Chilas Complex.
Nomenclature after (Leake, 1978).

Figure 4.20: Averaged primitive mantle (Sun and McDonough, 1989) normalized trace element pattern
of amphibole from olivine-bearing gabbronorite (red) and porphyroclastic lherzolite (black).
 108

4.3 Discussion

4.3.1 P-T conditions during intrusion

The granoblastic texture and the partly exsolved pyroxene indicate slow cooling of the
gabbronorite. Two-pyroxene mineral thermometry (Wood and Banno, 1973; Wells, 1977)
indicates a temperature range of 850-920°C, regardless of lithological unit, in accordance
with previously published estimates (Jan and Howie, 1981). These temperatures
correspond to granulite facies conditions. The solidus temperatures are possibly significant
higher. As magmatic amphibole is partly present, the temperature of crystallisation can be
constrained to less than 1100°C (Allen et al., 1975) and are estimated in the range of 1000
to 1100°C. Pressure estimates within gabbronoritic mineral-assemblages are not
straightforward. The clinopyroxene barometer (Nimis and Ulmer, 1998) yields pressures
between 1.0 and 1.3 GPa which probably are slightly to high since garnet should form in
the magmatic assemblage at such high pressure (Müntener et al., 2000); that is not the case
in the Chilas Complex. Geothermobarometry, using the plag-grt-bio-qz barometer (Hoisch,
1990) and Fe-Mg grt-bio exchange thermometer, applied to a metapelitic sample (C226)
sampled within the Khiner Valley, less than one kilometer to the north of the boundary
with the Chilas gabbronorite, indicates peak metamorphic conditions of ~0.7 GPa and
~700°C. A pressure above 0.6 GPa is also implied by the two-pyroxene-spinel symplectites
developed between olivine and plagioclase, which occur above ~ 0.6-0.7 GPa in ultramafic
rocks (Kushiro and Yoder, 1966). Accordingly the pressure during intrusion of the Chilas
Complex can be constrained around 0.6-0.7 GPa.

4.3.2 A composite Chilas Complex?

It is well established that plagioclase forms a continuous solid solution sequence between
anorthite and albite during crystal fractionation processes. Based on the analytical gap in
plagioclase mineral chemistry between the UMA sequence (An98-83) and the main
gabbronorite (An64-40) (Figure 4.6) the UMA were interpreted as being issued from a
different, more mafic magma batch than the gabbronorite (Jan et al., 1984; Khan et al.,
1989). However, our results do not support the presence of a gap and show a continuous
trend of K2O versus An content (Figure 4.6). The primitive mantle (Sun and McDonough,
1989) normalized trace element patterns of plagioclase from plagioclase-bearing lherzolite,
olivine-bearing gabbronorite and gabbronorite are parallel. The absolute trace element
 109

concentration increases with decreasing anorthite content and is consistent with

differentiation of a common magma body for the entire sequence. Continuous
differentiation trends are also observed in the major element chemistry of olivine (Figure
4.9), clinopyroxene (Figure 4.12) and orthopyroxene (Figure 4.13). The positive Sr and Pb
anomalies in primitive clinopyroxene and the lack of a negative Eu anomaly indicates that
the clinopyroxene fractionated before plagioclase. These results are therefore consistent
with a crystallisation sequence of ol (+sp) – ol+cpx (+sp +amph) – ol-cpx-opx (+sp) – ol-
cpx-opx (+sp) – cpx-opx-plag (+sp) – cpx-opx-plag-qtz (+sp) from one differentiating
magma for both the so-called UMA and the gabbronorite sequences. The fractionation of
clinopyroxene before plagioclase, the presence of magmatic amphibole- and the
continuous fractionation of spinel indicate that the Chilas Complex is derived from a
hydrous parental melt (Müntener et al., 2001). In the light of the new data, we therefore
infer that the entire mafic-ultramafic sequence is derived from a common source. However,
the plagioclase chemistry of late stage, meter-scale fine grained amphibolite dykes,
scattered within the Chilas Complex, is different, which suggests that these dykes were
probably derived from a different, later magmatic event.
As described in chapter 2 the field relations between the ultramafites are not consistent
with a simple cumulate sequence from dunite to gabbronorite. In particular, the fact that
the porphyroclastic lherzolite is replaced by dunite cannot be explained in terms of a single
differentiation sequence. The origin of the dunitic core of the ultramafite is discussed in
chapter 5.

4.3.2.1 Mineral trace element modelling

To obtain information about the degree of fractionation of analyzed gabbronorite samples
the mineral trace element data has been modeled. The applied model follows the method of
Hermann et al. (2001) which is based on the model of Langmuir (1989). The rationale
behind this approach is that zoning of the minerals has been obliterated within deep crustal
intrusions because of slow cooling under granulite facies conditions. Accordingly, if
interstitial liquid was present during crystallisation, it completely equilibrated with the
cumulus assemblage. The basis of the model is to model REE patterns of mineral phases
fractionated from a primitive starting liquid. The unknown is the amount of interstitial
liquid (L) and the degree of fractionation (F) of a particular. These two parameters are
varied to reproduce the measured REE pattern of mineral phases.
 110

The parental liquid composition of the gabbronorite has been calculated in equilibrium
with the mean clinopyroxene composition of a primitive olivine-bearing gabbronorite
(C66) using KD`s of Hart and Dunn (1993) with an interpolated value for Eu (Ionov et al.,
2002). In contrast to the main part of gabbronorite, this sample was collected from an
outcrop with magmatic layering typical for cumulate gabbros. The pyroxenes display no
negative Eu anomaly indicating a “pure” cumulate origin. Accordingly, the calculated
liquid composition should be close to the original liquid composition parental to the
gabbronorite sequence. Starting from this liquid composition, the REE pattern of
clinopyroxene, plagioclase and orthopyroxene of gabbronorite samples have been modeled
using the equations presented in (Hermann et al., 2001). The model can only be applied to
relatively primitive samples devoid of significant amounts of apatite and Ti-pargasite,
influencing the REE budget (Hermann et al., 2001). Accordingly, two homogeneous
gabbronorite samples (C48, C7) have been chosen to evaluate the degree of fractionation
and to approximate the amount of trapped liquid equilibrated with the cumulus mineral
assemblage.

Figure 4.21: Primitive mantle normalized (Sun and McDonough, 1989) results of the trace element
modeling following the approach of (Hermann et al., 2001). The mean modal composition of the
gabbronorite has been approximated as plag 60%, opx 30% and cpx 10% based on an average value
of grid counting and modal calculation based on whole rock analyses. Squares =clinopyroxene,
triangles = plag, and circles opx. The composition of plag and opx has been averaged and assumed to
 111

be constant, whereas the spread in cpx can be explained by equilibration with different amount of
interstitial liquid. The upper diagram displays the measured values. In the lower diagram the open
symbols are the results of the model, solid symbols are the measured values of the sample. A) The
results for sample C48 can be reproduced by 60% of differentiation. The spread in cpx composition
indicates equilibration with different amount of interstitial liquid. The solid line represents 2% of
interstitial liquid whereas the dashed line was calculated with 7% B) Results for sample C7 indicate
65% of differentiation and an average of 6,5% of interstitial liquid. The spread in cpx composition can
be modeled with interstitial liquid between 5-8% (not shown).

4.3.2.2 Results of mineral trace element modelling

The results of the mineral trace element modelling are presented in Figure 4.21. The results
indicate that the amount of trapped liquid was small. Mineral REE pattern from sample C
48 can be reproduced by 60% of differentiation. The variation of trace element
concentration in clinopyroxene can be reproduced varying the interstitial liquid between 2
and 7%. Mineral trace element patterns of sample C7 can be reproduced with 65% of
differentiation and 5 to 8% of interstitial liquid to account for the slight variation in
clinopyroxene composition. The model indicates that both samples represents cumulate
dominated rocks and the variation in clinopyroxene trace element composition indicates
equilibration with various amounts of interstitial liquid.

4.3.3 Water content

It is widely accepted that crystal fractionation in a hydrous system differs significantly
from an anhydrous system. The presence of magmatic amphibole, the high anorthite
content of plagioclase in gabbronorite samples and the crystallisation of cpx before
plagioclase are indications that water played a vital role in the formation of the Chilas
Complex. Fractionation of magmatic amphibole requires at least 4 wt% H2O (Cawthorn
and OHara, 1976). Since amphibole occurs after the fractionation of olivine the initial
water content could have been lower. Thermodynamic modelling indicating that 5-6 wt%
H2O are required to explain some observed calcic plagioclase composition (Marsh et al.,
1990). Accordingly, the water content in the parental liquid can be very roughly
approximated between 2 to 7 wt% H2O. Similar results are also gained using a plagioclase-
melt hygrometer (Housh and Luhr, 1991)
 112

4.3.4 Parental liquid composition

The high Mg# of olivine found within dunitic to lherzolitic ultramafite (Mg# = 0.9)
indicates that the liquid forming the dunite had a high Mg#melt= 0.73 (assuming Kdol(Fe-Mg)
=0.3 (Roeder and Emslie, 1970)) in equilibrium with mantle olivine. The usual approach to
constrain the major element liquid composition by simply adding back the inferred
fractionated component to match the appropriate Mg# is not followed because the most
primitive liquid composition is already relative evolved (Mg#melt = 0.54) and adding
significant amounts of crystallized phases easily leads to false composition (OHara and
Herzberg, 2002). However, the REE element concentration can be calculated in
equilibrium with the most primitive clinopyroxenes present in the dunite (C174) (using the
above cited Kd values). The calculated melt is extremely primitive and slightly LREE
enriched (Figure 4.22). The flat MREE to HREE and the high Mg#melt are similar to basalt
from the Izu-Bonnin arc (Arculus et al., 1992). Non-modal batch melting trace element
modeling (Shaw, 1970) using a depleted source as starting material (Grove et al., 2002)
indicates that trace element characteristics of the HREE are best reproduced by ~20% of
partial melting in the spinel field indicating a shallow melting region (< 1.5 GPa). The
MREE and LREE are progressively less well reproduced indicating an increasing influence
of a slab derived fluid component for those elements (McCulloch and Gamble, 1991). The
trace element concentration of the liquid can be calculated assuming mass balance:

Cmelt= Cmantle(1-Xf)+XfCf (2)

where Cmelt, Cmantle and Cf is the concentration of the element of interest in the parental
melt, in the mantle derived melt and in the fluid component, respectively. Xf is the weight
fraction of fluid added. The fluid composition has been calculated assuming 2 and 7 wt%
of water (Figure 4.22).
The REE concentration of the calculated fluid component is comparable of the fluid
component derived for the Mt. Shasta (Grove et al., 2002) and the Mariana Through
magmas (Stolper and Newman, 1994). The REE concentration are however, significantly
higher than a fluid inferred to be in equilibrium with an eclogitic slab or mantle wedge
(Ayers, 1998). The inferred fluid composition is most similar to a low percentage melt
(2%-5%) derived from the eclogitisied basaltic part of the slab (Grove et al., 2002).
 113

In conclusion, the high Mg# of the primitive melt and the olivine dominated fractionation
sequence indicate that the Chilas Complex is derived from a primitive, hydrous, shallow-
mantle derived melt with a slab derived fluid –rich component. A similar melt in major
element composition is thought to be parental to the Dras volcanite in the Ladakh arc
(Dietrich et al., 1983).

Figure 4.22: Primitive Mantle normalized spidergram comparing the Chilas parental melt
composition, the result of 20% non-modal batch melting (ol 40%, opx 20%, cpx 30%, amph 9%, sp
1%) of a previously slightly depleted source (Grove et al., 2002) and the fluid rich component. The
fluid component (hatched field) is compared to a fluid component inferred for the Mt. Shasta volcanite
(grey field) (Grove et al., 2002), a fluid in equilibrium with the wedge, an eclogitisised slab (Ayers,
1998) and with a fluid involved in the Mariana magma formation (Stolper and Newman, 1994). The
Chilas fluid component is most similar to a low percentage melt derived from the basaltic part of a
subducted slab. It displays a negative Gd anomaly which is interpreted as an artifact due to a slight
negative anomaly present in the Chilas parental melt which is however, within error and so not
significant. However, due to the low percentage of added fluid this “unreal” anomaly is strongly
enhanced. The interpreted true pattern is indicated by the hatched lines.

4.4 Conclusion
Mineral trace and major element data indicate that the Chilas Complex intruded the
Kohistan arc sequence around 0.6-0.7 GPa. The continuous trend in plagioclase
composition between the ultramafite and the gabbronorite indicate that both sequence stem
from a common magma body. Differences in mineral trace element concentration in
gabbronorite can be explained by in-situ crystallization with various amount of interstitial
 114

liquid. Modeling of the REE composition of the parental melt of the Chilas Complex
indicates that the melt derived from a previously depleted upper mantle by relatively high-
percentage of partially melting (~20%). The absent of any garnet signature in the REE
composition indicate that melting occurred within the spinel stability field in the upper
mantel (≤ 1.5 GPa). A external component, necessary to explain the REE composition, is
most similar to a melt derived by low percentage partial melting (2-5%) of an eclogitised
basaltic layer of the subducted slab.
In the next chapter whole rock and isotope data are presented and compared with the
calculated REE concentration of the parental magma.
 115

Reference

Allen, J. C., A. L. Boettcher, and G. Marland, 1975, Amphiboles in andesite and basalt; I,
Stability as a function of P-T-f (sub 02): American Mineralogist, v. 60, p. 1069-
1085.
Arculus, R. J., J. A. Pearce, B. J. Murton, and d. L. S. R. van, 1992, Igneous stratigraphy
and major-element geochemistry of Holes 786A and 786B, in P. Fryer, A. Pearce
Julian, B. Stokking Laura, R. Ali Jason, R. Arculus, L. Ballotti Dean, M. Burke
Margaret, G. Ciampo, A. Haggerty Janet, B. Haston Roger, D. Heling, A. Hobart
Michael, T. Ishii, E. Johnson Lynn, Y. Lagabrielle, W. McCoy Floyd, H.
Maekawa, S. Marlow Michael, J. Milner Greg, J. Mottl Michael, J. Murton
Bramley, P. Phipps Stephen, A. Rigsby Catherine, L. Saboda Kristine, B. Stabell,
S. van der Laan, Y. Xu, H. Dearmont Lona, K. Mazzullo Elsa, J. Stewart
Norman, and R. Winkler William, eds., Proceedings of the Ocean Drilling
Program, Bonin/ Mariana region; covering Leg 125 of the cruises of the Drilling
Vessel JOIDES Resolution, Apra Harbor, Guam, to Tokyo, Japan, sites 778-786,
15 February 1989-17 April 1989.: Proceedings of the Ocean Drilling Program,
Scientific Results, v. 125: College Station, TX, United States, Texas A & M
University, Ocean Drilling Program, p. 143-169.
Ayers, J., 1998, Trace element modeling of aqueous fluid; peridotite interaction in the
mantle wedge of subduction zones: Contributions to Mineralogy and Petrology,
v. 132, p. 390-404.
Bard, J. P., 1983, Metamorphic evolution of an obducted island arc; Example of the
Kohistan sequence (Pakistan) in the Himalayan collided range: Geological
Bulletin, University of Peshawar, v. 13, p. 105-184.
Cawthorn, R. G., and M. J. OHara, 1976, Amphibole fractionation in calc-alkaline magma
genesis: American Journal of Science, v. 276, p. 309-329.
Debari, S., S. M. Kay, and R. W. Kay, 1987, Ultramafic xenoliths from Adagdak Volcano,
Adak, Aleutian Islands, Alaska; deformed igneous cumulates from the Moho of
an island arc: Journal of Geology, v. 95, p. 329-341.
Debari, S. M., and R. G. Coleman, 1989, Examination of the deep levels of an island arc;
evidence from the Tonsina ultramafic-mafic assemblage, Tonsina, Alaska, in A.
Page Robert, ed., Special section on northern Chugach Mountains-southern
 116

Copper River basin segment of the Alaskan Transect; Part 1.: Journal of
Geophysical Research, B, Solid Earth and Planets, v. 94; 4: Washington, DC,
United States, American Geophysical Union, p. 4373-4391.
Dietrich, V. J., W. Frank, and K. Honegger, 1983, A Jurassic-Cretaceous island arc in the
Ladakh-Himalayas, in S. Aramaki, and I. Kushiro, eds., Arc volcanism.: Journal
of Volcanology and Geothermal Research, v. 18; 1-4: Amsterdam, Netherlands,
Elsevier, p. 405-433.
Gasparik, T., 1984, Two-pyroxene thermobarometry with new experimental data in the
system CaO-MgO-Al (sub 2) O (sub 3) -SiO (sub 2): Contributions to
Mineralogy and Petrology, v. 87, p. 87-97.
Grove, T. L., S. W. Parman, S. A. Bowrig, R. C. Price, and M. B. Baker, 2002, The role of
an H (sub 2) O-rich fluid component in the generation of primitive basaltic
andesites and andesites from the Mt. Shasta region, N California: Contributions
to Mineralogy and Petrology, v. 142, p. 375-396.
Hart, S. R., and T. Dunn, 1993, Experimental cpx/melt partitioning of 24 trace elements:
Contributions to Mineralogy and Petrology, v. 113, p. 1-8.
Hermann, J., O. Müntener, and D. Günther, 2001, Differentiation of Mafic Magma in a
Continental Crust-to-Mantle Transition Zone: J. Petrology, v. 42, p. 189-206.
Hoisch, T. D., 1990, Empirical calibration of six geobarometers for the mineral assemblage
quartz+muscovite+biotite+plagioclase+garnet: Contributions to Mineralogy and
Petrology, v. 104, p. 225-234.
Housh, T. B., and J. F. Luhr, 1991, Plagioclase-melt equilibria in hydrous systems:
American Mineralogist, v. 76, p. 477-492.
Ionov, D. A., J.-L. Bodinier, S. B. Mukasa, and A. Zanetti, 2002, Mechanisms and Sources
of Mantle Metasomatism: Major and Trace Element Compositions of Peridotite
Xenoliths from Spitsbergen in the Context of Numerical Modelling: J. Petrology,
v. 43, p. 2219-2259.
Jan, M. Q., and R. A. Howie, 1980, Ortho- and clinopyroxenes from the pyroxene
granulites of Swat Kohistan, northern Pakistan: Mineralogical Magazine, v. 43,
p. 715-726.
Jan, M. Q., and R. A. Howie, 1981, The mineralogy and geochemistry of the
metamorphosed basic and ultrabasic rocks of the Jijal Complex, Kohistan, NW
Pakistan: Journal of Petrology, v. 22, p. 85-126.
 117

Jan, M. Q., and R. A. Howie, 1982, Hornblendic amphiboles from basic and intermediate
rocks of Swat-Kohistan, Northwest Pakistan: American Mineralogist, v. 67, p.
1155-1178.
Jan, M. Q., M. A. Khan, and B. F. Windley, 1992, Exsolution in Al-Cr-Fe (super 3+) -rich
spinels from the Chilas mafic-ultramafic complex, Pakistan: American
Mineralogist, v. 77, p. 1074-1079.
Jan, M. Q., and B. F. Windley, 1990, Chromian spinel-silicate chemistry in ultramafic
rocks of the Jijal Complex Northwest Pakistan: Journal of Petrology, v. 31, p.
667-715.
Jan, Q. M., M. U. K. Khattak, M. K. Parvez, and B. F. Windley, 1984, The Chilas
stratiform complex; Field and mineralogical aspects: Geological Bulletin,
University of Peshawar, v. 17, p. 163-169.
Kay, S. M., and R. W. Kay, 1985, Aleutian tholeiitic and calc-alkaline magma series; 1,
The mafic phenocrysts: Contributions to Mineralogy and Petrology, v. 90, p.
276-290.
Khan, M. A., 1988, Petrology and structure of the Chilas ultramafic-mafic complex,
Kohistan Arc, NW Himalayas, Unpublished PhD Thesis, University of London,
London.
Khan, M. A., M. Q. Jan, B. F. Windley, J. Tarney, and M. F. Thirlwall, 1989, The Chilas
mafic-ultramafic igneous complex; the root of the Kohistan island arc in the
Himalaya of northern Pakistan, in L. Malinconico Lawrence, Jr., and J. Lillie
Robert, eds., Tectonics of the western Himalayas.: Special Paper - Geological
Society of America, v. 232: Boulder, CO, United States, Geological Society of
America (GSA), p. 75-94.
Kretz, R., 1983, Symbols for rock-forming minerals: American Mineralogist, v. 68, p. 277-
279.
Kushiro, I., and H. S. Yoder, Jr., 1966, Anorthite-forsterite and anorthite-enstatite reactions
and their bearing on the basal-eclogite transformation: Journal of Petrology, v. 7,
p. 337-362.
Langmuir, C. H., 1989, Geochemical consequences of in situ crystallization: Nature
(London), v. 340, p. 199-205.
Leake, B. E., 1978, Nomenclature of amphiboles: American Mineralogist, v. 63, p. 1023-
1052.
 118

Marsh, B. D., J. H. Fournelle, J. D. Myers, and I. M. Chou, 1990, On plagioclase

thermometry in island arc rocks; experiments and theory, in J. Spenser Ronald,
and I. M. Chou, eds., Fluid-mineral interactions; a tribute to H. P. Eugster.:
Special Publication - Geochemical Society, v. 2: University Park, PA, United
States, Geochemical Society, p. 65-83.
McCulloch, M. T., and J. A. Gamble, 1991, Geochemical and geodynamical constraints on
subduction zone magmatism: Earth and Planetary Science Letters, v. 102, p. 358-
374.
Müntener, O., 1997, The Malenco peridotites (Alps): Petrology and geochemistry of
subcontinental mantle and Jurassic exhumation during rifting, Swiss Ferderal
Institute of Technology Zurich, Zurich, 205 p.
Müntener, O., J. Hermann, and V. Trommsdorff, 2000, Cooling history and exhumation of
lower-crustal granulite and upper mantle (Malenco, eastern Central Alps):
Journal of Petrology, v. 41, p. 175-200.
Müntener, O., P. B. Kelemen, and T. L. Grove, 2001, The role of H (sub 2) O during
crystallization of primitive arc magmas under uppermost mantle conditions and
genesis of igneous pyroxenites; an experimental study: Contributions to
Mineralogy and Petrology, v. 141, p. 643-658.
Nimis, P., and P. Ulmer, 1998, Clinopyroxene geobarometry of magmatic rocks; Part 1, An
expanded structural geobarometer for anhydrous and hydrous, basic and
ultrabasic systems: Contributions to Mineralogy and Petrology, v. 133, p. 122-
135.
Obata, M., 1980, The Ronda peridotite; garnet-, spinel-, and plagioclase-lherzolite facies
and the P-T trajectories of a high-temperature mantle intrusion: Journal of
Petrology, v. 21, p. 533-572.
OHara, M. J., and C. Herzberg, 2002, Interpretation of trace element and isotope features
of basalts; relevance of field relations, petrology, major element data, phase
equilibria, and magma chamber modeling in basalt petrogenesis, in O. Mysen
Bjorn, ed., A special issue dedicated to Hatten S. Yoder, Jr., Pergamon. Oxford,
International. 2002.
Pettke, T., C. A. Heinrich, A. C. Ciocan, and D. Guenther, 2000, Quadrupole mass
spectrometry and optical emission spectroscopy: Detection capabilities and
representative sampling of short transient signals from laser-ablation: Journal of
analytical atomic spectrometry, v. 15, p. 1149-1155.
 119

Roeder, P. L., and R. F. Emslie, 1970, Olivine-liquid equilibrium: Contributions to

Mineralogy and Petrology, v. 29, p. 275-289.
Santos, J. F., U. Schaerer, I. J. I. Gil, and J. Girardeau, 2002, Genesis of pyroxenite-rich
peridotite at Cabo Ortegal (NW Spain); geochemical and Pb-Sr-Nd isotope data:
Journal of Petrology, v. 43, p. 17-43.
Shaw, D. M., 1970, Trace element fractionation during anatexis: Geochimica et
Cosmochimica Acta, v. 34, p. 237-243.
Stolper, E., and S. Newman, 1994, The role of water in the petrogenesis of Mariana
Trough magmas: Earth and Planetary Science Letters, v. 121, p. 293-325.
Sun, S. S., and W. F. McDonough, 1989, Chemical and isotopic systematics of oceanic
basalts; implications for mantle composition and processes, in A. D. Saunders,
and M. J. Norry, eds., Magmatism in the ocean basins.: Geological Society
Special Publications, v. 42: London, United Kingdom, Geological Society of
London, p. 313-345.
Takahashi, E., K. Uto, and J. G. Schilling, 1987, Primary magma compositions and Mg/ Fe
ratios of their mantle residues along Mid Atlantic Ridge 29 degrees N to 73
degrees N: Technical Report of ISEI, Series A, v. 9: Misasa, Japan, Institute for
Study of the Earth's Interior, Okayama University, 14 p.
Yoder, H. S., Jr., and C. E. Tilley, 1962, Origin of basalt magmas; an experimental study
of natural and synthetic rock systems: Journal of Petrology, v. 3, p. 342-529.
 120

5 Lower continental crust formation through focused

flow in km-scale melt conduits: The zoned ultramafic
bodies of the Chilas Complex in the Kohistan Island arc
(NW Pakistan)

O.Jagoutz1, O.Müntener2, J-P.Burg1, P.Ulmer1, E.Jagoutz,1 T.Pettke1

1
Department of Earth Sciences, ETH Zurich, Sonneggstr. 5, CH-8092 Zurich, Switzerland
2
Institute of Geological Sciences, University of Bern, Baltzerstr. 1, CH-3012 Bern,
Switzerland
3
Max-Planck Institute für Kosmochemie, Saarstrase 23, Mainz, Germany
 121

This chapter is written in paper format; accordingly it repeats some field relationship and
textural description previously presented in chapter 2. Additionally, whole rock chemistry,
Sm-Nd isotope and clinopyroxene trace element data from the ultramafite and the
gabbronorite sequence are presented and discussed.

5.1 Introduction
It is in generally believed that the continental crust is formed through mantle derived
melts, which are, at some stage, in equilibrium with a mantle assemblage. The paradox is
that the bulk continental crust composition is estimated to be andesitic (57-64 wt% SiO2
and 0.5 < Mg# (molar Mg/(Mg+ Fe) > 0.56 (Rudnick and Gao, 2003 and references
therein), whereas melts in equilibrium with mantle peridotite are generally basaltic with
Mg# > 0.7. Accordingly, any model for crust formation through melt in equilibrium with
mantle peridotite requires the loss of high Mg# cumulates. Fractionation models indicate
that between 15-35 % crystallization is necessary to lower the Mg# adequately (Conrad
and Kay, 1984). Numerous models have been proposed to overcome the paradox (for
detailed review see Kelemen, 1995; Rudnick, 1995) among which are two conflicting ones:
1) continental crust is basaltic rather than andesitic because cumulates form a major
component below the seismic MOHO (Kay and Kay, 1985a) and are underestimated by
bulk continental compositions based on seismic properties; and 2) andesites form through
crustal differentiation of basaltic or high Mg# andesitic magmas with the mafic, dense,
crystalline residue eventually delaminating and sinking back into the mantle (Kay and Kay,
1991). If, according to the first model, significant crystal fractionation occurs in the upper
mantle a mechanism is needed to chemically isolate the evolved melts from the upper
mantle. The fractionated melts, out of equilibrium with the mantle, must travel through the
overlying upper mantle without re-equilibration. Based on the lack of appropriate upper
mantle rocks brought to the surface in continental regions (Rudnick and Gao, 2003) and
the missing primitive cumulate rocks in the exposed Talkeetna paleo-Island arc crust
section (Kelemen et al., 2004), the delamination theory is presently favored.

5.1.1 Present day origin of continental crust

In general, present day continental crust formation is attributed to two distinct plate
tectonic settings: active continental margin and intraplate volcanism. Trace element
 122

similarities, e.g. enrichment of the light-rare earth elements (L-REE), the depletion of Nb,
Ti and enrichment of Pb with respect to the REE, between bulk crust estimates and present
day subduction zone volcanism, indicate that a significant proportion (65-90% (Rudnick,
1995) of the post-Archean (?) continental crust formed by processes similar to present day
subduction zone setting. Primary arc magmas are thought to have basaltic rather than
andesitic composition (Anderson, 1982; Crawford et al., 1987) even though primary high
Mg# andesites do exit (Kelemen, 1995). The bulk composition of island arcs is also
thought to be basaltic (DeBari and Sleep, 1991). Active island arcs allow only direct
surface observations. Lower-crustal arc processes can either be studied on lower-crustal or
upper-mantle xenoliths brought to the surface by the arc vulcanics (Ducea and Saleeby,
1998) or in accreted and uplifted paleo-island arcs. One of the best exposed arc section is
the Kohistan paleo-island arc, in NW Pakistan. We present results of zoned ultramafic
bodies from the Chilas Complex, a volumetrically important calc-alkaline intrusion
predominantly made of gabbronorite. Field observations, whole-rock and mineral major
and trace elements and Sm-Nd isotopic data of the main units within the Chilas Complex
are presented. The ultramafic bodies are interpreted as melt channels through which the
entire Chilas gabbronorite sequence was fed. The mafic sequence originates though
differentiation of a primary, high Mg# (>0.7), basaltic arc magma within km-scaled
isolated melt conduits. The estimated basaltic-andesitic bulk composition of the Chilas
gabbronorite sequence matches closely estimates of lower crustal composition. We
conclude that differentiation within the upper mantle is an important crust forming process
within the Chilas Complex.

5.2 Geological setting

The Kohistan terrane in NW Pakistan (Figure 5.1) is regarded as a fossil
Jurassic/Cretaceous island arc that was captured between the colliding Indian and Asian
plates and obducted during the Himalayan collision in the late Cretaceous/early Tertiary
(Bard, 1983; Tahirkheli, 1979). It is one of the few places in the world where the full
structure of an arc - from upper mantle in the south to the uppermost volcanic sequences in
the north - is preserved (Coward et al., 1986; Searle et al., 1999; Treloar et al., 1996). Due
to outstanding exposures and freshness of the rock, the Kohistan offers an unrivalled
opportunity to investigate island arc building processes. It is generally believed that the
terrain originated as an intraoceanic arc above a north dipping subduction zone somewhere
 123

in the equatorial area of the Tethys Ocean (Bignold and Treloar, 2003; Coward et al., 1987;
Khan et al., 1993). The fossil arc is bounded to the north and south by the Kohistan-
Karakorum and Indus suture zones, respectively. The Chilas Complex (Figure 5.1), which
includes zoned ultramafic bodies, separates the Southern Amphibolites Complex in the
south, representing the lower crust of the Island arc (Jan and Howie, 1981; Jan and
Windley, 1990; Ringuette et al., 1999) from the Gilgit domain, composed predominantly of
upper crustal plutonic, volcanic and sedimentary rocks and their metamorphosed
equivalents (Petterson and Windley, 1985; Pudsey et al., 1985).

Figure 5.1 Geological map of the Kohistan arc, NE Pakistan. The ultramafite (asterisks) have been
studied in the vicinity of the town Chilas in the eastern part of the Kohistan arc

5.2.1 Chilas Complex

The Chilas Complex is a 250 km long and up to 30 km wide calc-alkaline, mafic-
ultramafic intrusive body into the Kohistan arc. Khan et al. (1989) stated that 85% of the
Complex is composed of homogeneous gabbronorite, which they called main gabbronorite,
containing several km-sized bodies of ultramafic rocks with associated anorthosites,
olivine gabbros and gabbronorite (the so-called Ultramafc-Mafic-Anorthositic (UMA)
association Khan et al., 1989). The large volume of homogeneous gabbronorite led these
authors to the conclusion that the Chilas Complex intruded during an incipient intra-arc
rifting stage. Khan et al. (1989) interpreted the UMA as isolated pulses of picritic magma
injected late into the crystallizing magma chamber forming the gabbronorite, whereas
 124

Kubo (1996) argued that the UMA formed prior to the gabbronorite sequence. The
homogeneous and identical Pb, Nd and Sr isotopic composition of the UMA and the
gabbronorite points to a common source for the two rock associations (Khan et al., 1997).
However, the detailed contact relationships between the UMA and the gabbronorite
sequence are less documented. Based on the new results on mineral chemistry indicating a
continuous compositional sequence between UMA and gabbronorite (chapter 4), we do not
follow strictly the temporal separation of Khan et al. (1989), but do distinguish between the
various peridotites and pyroxenites and the gabbronorite sequence, which includes olivine-
bearing gabbronorite, gabbronorite, diorite and minor anorthosites.

5.3 Field and petrological observations

5.3.1 Field relations

The gabbronorite is generally homogeneous. Layering can be present but is usually of short
lateral extent (up to few tens of meters). Upward bending and shearing of layered olivine-
bearing gabbronorite at the contact with the ultramafics indicate upwards movement of the
ultramafics with respect to the gabbronorite units. The ultramafic bodies are zoned with a
massive dunitic core and subsequent shells of lherzolite, plagioclase-bearing lherzolite and
pyroxenite towards the contact with surrounding gabbronorite. Tens-of-meterscale bodies
of relict amphibole-bearing lherzolite, hereafter called porphyroclastic lherzolite, contains
centimeter-sized porphyroclastic clinopyroxene clasts and are enclosed within the dunitic
core. Dunitic dykes cross cut the porphyroclastic lherzolite and xenoliths of the lherzolite
float in the dunite (Figure 5.2). Field observations indicate that the zoning results from a
centimeter to tens of meter scale reaction rim developed at the contact between dunitic
lithologies and surrounding gabbroic units: The transition between the subsequent shells is
gradational from centimeter to tens of meter scale; the modal amount of pyroxene
increases and that of olivine decreases outwards from the dunitic core. Brecciated
fragments of dunite, displaying little or no reaction texture, float within the lherzolite.
Plagioclase appears along triple junctions displaying a reaction rim of pyroxene-spinel
symplectite at plagioclase-olivine contacts. Patchy concentric clusters of this reaction
texture are elongated in the third (vertical) dimension (Figure 5.2). Plagioclase-bearing
veins vanish along strike into a planar trace defined by pyroxene-spinel-symplectites.
Xenoliths of gabbronorite are frequently encountered within the lherzolite and pyroxenite
along the contact zone. Additionally, coarse-grained hornblendites and K-feldspar bearing
 125

pegmatites occur within the contact zone. The actual sharp, irregular contact between
gabbronorite and ultramafite indicates similar, sub-solidus rheological properties during
emplacement of the rocks (Figure 5.3). In the gabbronorite, areas rich in ultramafic
xenoliths are identified next to the contact.
In summary, field relationships indicate that the oldest rocks identified in the field are
porphyroclastic lherzolite, which were replaced by massive dunite, which in turn shows a
reactive contact zone with the surrounding gabbronorite resulting in the formation of
lherzolite and pyroxenite. The contact relationships are schematically illustrated in Figure
5.2 and have been interpreted to indicate emplacement of ultramafic rocks under sub-
solidus condition into partially consolidated gabbronorite, which in turn has intruded the
ultramafic rocks (Burg et al., 1998).

Figure 5.2 Block diagram of the contact relationship between zoned ultramafic and gabbronorite.
Indicated are key contact relationship supporting a replacive origin for the dunite. A reaction zone is
developed between ultramafite and gabbronorite. It is interpreted as the result of infiltration of a melt
parental to the gabbronorite into the dunite forming lherzolite. Indicated also are location for six
samples which, have been selected for LA-ICP-MS analysis and partly (C174 and C7) for Sm-Nd
isotopic investigation.
 126

5.3.2 Petrological description

Porphyroclastic lherzolite with clinopyroxene porphyroclasts has an inequigranular -
polygonal to interlobate fabric with coarse grained centimeter sized interlobate
clinopyroxene and orthopyroxene, which show exsolution lamellae of the complementary
pyroxene and Cr-spinel. Granoblastic olivine has a bimodal grain size distribution whereby
larger olivine grains show undulose extinction. Smaller recrystallized olivine grains have a
granoblastic texture. Trails of smaller olivine garins crosscut larger grains and occur
between pyroxene grains, rarely penetrating into the pyroxene grains indicating melt
infiltration. Pargasitic amphibole is either developed along the rim of seriate pyroxene
grains or as poikiloblastic grains with numerous pyroxene inclusions.
The dunite is composed of straight extincting, granoblastic olivine. Minor clinopyroxene
occurs in centimeter-sized clusters along grain boundaries of the olivine crystals.
Accessory opaque minerals are Cr-spinel and Cr bearing magnetite.
Within the contact zone from the dunitic core outwards, clinopyroxene and amphibole first
occur along grain boundaries forming an intergranular network between larger olivine
crystals. With increasing clinopyroxene content exsolved porphyroblastic clinopyroxene
grains appear. Hercynitic spinel, magnetite, and ilmentite are common accessory minerals.
With further increasing pyroxene content additionally orthopyroxene appears and contains
abundant olivine inclusion indicating that the orthopyroxene formed through a peritectic
reaction. The gabbronorite displays a predominantly magmatic fabric, composed of
plagioclase, clinopyroxene and orthopyroxene with a granular texture (grain size ~ 0.5cm).
A thin rim of pargasitic amphibole is developed around pyroxene and (Ti-) magnetite
occurs as an accessory minerals.
 127

Figure 5.3: Picture of the reaction zone between gabbronorite and dunite forming a centimeter scale
pyroxenite. The contact between pyroxenite and gabbronorite is sharp but irregular, indicating similar
sub-solidus rheology for both.

5.4 Geochemical data

5.4.1 Whole rock geochemistry

Whole rock samples of the porphyroclastic lherzolite, dunitic core, lherzolite and from the
gabbronorite sequence have been analysed (Actlabs, Canada). Major elements were
determined by Inductively Coupled Plasma Optical Emission Spectrometry (ICP-OES),
and trace elements by Inductively Coupled Plasma Mass Spectrometry (ICP-MS). Results
are illustrated in Figure 5.4 and Figure 5.5 and listed in Table 5.1.
Major elements of the porphyroclastic lherzolite show a variable CaO and low Al2O3
content but a relative homogeneous Mg# (0.85). The primitive mantle (Sun and
McDonough, 1989) normalized rare earth element (REE) pattern of the porphyroclastic
lherzolite is slightly convex upwards. The scatter in REE concentration ranges from 0.1-
1.2 times primitive mantle. The light REE are slightly depleted compared to the middle and
heavy REE (CeN/YbN< 0.8) whereas the middle REE are enriched compared to the heavy
REE (GdN/YbN > 1.3). Based on major elements characteristics, dunitic samples can be
grouped into those with a high Mg# (0.87) and those with a low Mg# (0.82) but with
 128

nearly equal SiO2 and CaO contents. The difference in Mg#s is mainly due to olivine
chemistry from Fo91 to Fo80. The REE concentrations, which are the most depleted ones
analysed in this study, do not mirror this grouping. The primitive mantle normalized REE
pattern is relatively straight with a slight enrichment of the heavy REE (GdN/YbN~ 0.17-
0.9). Sample C 123 shows a distinct enrichment of the light REE. Lherzolite from the
contact reaction zone displays a significant scatter in REE concentration ranging from
0.03-1 times primitive mantle. In general samples closer to the contact, (increasing
pyroxene content) display an increasingly convex upward REE pattern and increasing REE
concentrations. Two samples (C 210, C 211) show a significantly enriched heavy REE
pattern, due to unusually large amounts of orthopyroxene.
The selected gabbronorite samples were collected along a traverse ranging from the
contact-zone with the ultramafic bodies towards the southern margin of the Chilas
Complex. The major element chemistry, excluding an olivine-bearing gabbronorite (C66),
is relatively constant (Figure 5.4) whereas the trace element show a significant spread
(Figure 5.5). The REE content increases roughly ten times from primitive samples in the
vicinity of the ultramafic bodies, to evolved ones at the margin of the Chilas Complex
indicating an internal gradient in REE of the otherwise homogeneous gabbronorite. The
primitive mantle (Sun and McDonough, 1989) normalized REE pattern are convex
downwards with a modest enrichment of the light REE with respect to the middle REE
(CeN/GdN~ 1.1-2.4) and slightly depleted middle REE in respect to the heavy REE
(GdN/YbN~ 0.6-1.0). An olivine-gabbronorite sample from the vicinity (<1m) of the
ultramafic bodies has the lowest REE content and a strong positive Eu anomaly ((Eu/Eu*)N
= 5.6), indicating plagioclase accumulation. With increasing distance from the ultramafic
body the REE content increases roughly symmetrically towards the margin of the Chilas
Complex. Additionally, the Eu anomaly is less and less pronounced and samples from the
marginal zone of the Chilas Complex both to the north and the south display a smooth REE
pattern. Primitive samples have a strong positive Eu anomaly ((Eu/Eu*)N up to 2.19),
which is usually attributed to accumulation of plagioclase. With increasing REE content
the Eu anomaly becomes less pronounced and samples with the highest REE concentration
have a negligible Eu anomaly ((Eu/Eu*)N ~1.02-1.07). The size of the anomaly does not
correlate with the Sr concentration or the whole rock Mg# and correlates negative with the
potassium concentration suggesting that accumulation of plagioclase was not significant.
 129

Table 5.1: Whole rock major (in wt%) and trace element (in ppm) concentration of various units of the
Chilas Complex. “<” Symbols indicate values lower than the detection limit.
 130

Figure 5.4 Total Alkali vs Silica (TAS) diagram for rock classification of samples from the
gabbronoritie sequence of the Chilas Complex. Nomenclature of igneous rocks after Cox et al. (1979).
The solid line divides sub-alkalic and alkalic magma series (Miyashiro, 1978). The whole rocks show
only a limited spread in composition and are classified as sub-alkalic gabbros to diorites. Sample C 66
an olivine-bearing gabbronorite has extremely low values of silica and alkali and plots outside of the
classification schema of Cox et al. (1979).
 131

Figure 5.5: Primitive mantle normalized REE spectra A) from whole rock samples from
porphyroclastic lherzolite (solid line without symbols), dunites (dashed lines with open symboles,
absent of symbols indicated extrapolated values below or at the detection limit) and peridotites form
the contact zone (solid line with symbols) and B) gabbronorite whole rock samples.

5.4.2 Clinopyroxene chemistry

Whole rock samples were crushed and sieved to obtain a 200-500 µm fraction. Ten to 40
optically clear clinopyroxene crystals per sample were hand picked under an optical
 132

binocular and mounted into epoxy. Major elements were determined with a Cameca SX 50
at the ETH Zurich and the rare earth element characteristics for eight samples were
determined with a Laser ablation ICPMS (LA-ICPMS) conducted at the ETH Zurich
(Pettke T., 2000). The results are given in Table 5.2 and illustrated in primitive mantle
(McDonough and Sun, 1995) normalized spidergrams (Figure 5.6) Since the clinopyroxene
analyzed had orthopyroxene exsolution lamellae, the laser ablation spot size was adjusted
between 60-110 µm in order to obtain the primary pyroxene composition
Clinopyroxene from porphyroclastic lherzolite (Figure 5.6 A) show a weak core-to-rim
zoning in some major elements, especially in Al2O3, whereas zoning in trace elements is
insignificant. The REE concentration is generally low and normalized REE patterns show a
slight light REE depletion and convex upwards pattern (CeN/YbN = 0.43-0.76) and
enrichment of the middle compared to the heavy REE ((GdN/YbN = 1.2-1.8).
Of the dunitic samples, only C 174 had enough clinopyroxene to be analyzed. C 174 is a
primitive, high Mg# dunite optically unaffected from the contact reaction zone. We
focused on clinopyroxenes grains with high Mg# (0.92-0.93). The primitive mantle
normalized concentration displays a convex upward pattern. The ratio of middle to heavy
REE are close to equal whereas the light REE are modestly depleted (GdN/YbN~ 1 and
CeN/GdN~ 0.4-0.5). From the contact reaction zone clinopyroxenes from sample C 218 has
been analysed. C218 is a strongly reacted plagioclase-bearing lherzolite. The normalized
REE pattern display also a convex upwards REE pattern. The heavy REE are slightly
depleted in respect to the middle REE (GdN/YbN~ 1.3-1.6) and the light REE are modestly
depleted in respect to the middle REE (CeN/GdN~ 0.3).
An olivine-bearing gabbronorite directly from the contact with the ultramafics and two
gabbronorite samples, one within a meter of the contact (C 48) and the other approx 1 km
from the contact (C7), have been analysed. The REE concentrations of olivine-bearing
gabbronorite are low, and the light REE are depleted with respect to the middle and heavy
REE (CeN/GdN~ 0.3-0.55 and CeN/YbN~ 0.4-0.6); the middle REE are slightly enriched
compared to the heavy REE (GdN/YbN~ 1.2-1.47). The clinopyroxene with a low REE
concentration show a slight positive EU anomaly. The clinopyroxene from homogeneous
gabbronorite have significantly higher REE concentrations and the light REE are more
depleted in respect to the middle REE (CeN/GdN~ 0.3-0.4). The middle REE are only
slightly enriched relative to the heavy REE (GdN/YbN~ 1.2). With increasing REE content
the clinopyroxene display an increasingly more pronounced negative EU anomaly that also
correlates with decreasing Mg# indicationg fractionation of clinopyroxene prior to
 133

plagioclase. Experimental studies on similar bulk compositions (Muentener et al., 2001)

imply the presence of H2O at pressures of 0.5-0.8 GPa to crystallize clinopyroxene prior to
plagioclase.

Table 5.2: Clinopyroxene trace element concentration (ppm) of various units from within the Chilas
Complex. n= number of grains analysed
 134

Figure 5.6: Primitive mantle (Sun and McDonough, 1989) normalized spider diagram of clinopyroxene
from A) porphyroclastic lherzolite (C03-43 open squares; C03-44 filled squares; C03-45 open circle) B)
dunites (solid symbols) and peridotite form the contact zone (open symbols) and C) olivine-bearing
gabbronorite (open squares) and gabbronorite (C 48 solid circles and C7 open circles).
 135

5.4.3 Nd Isotope geochemistry

The Nd isotopic composition of two whole rock samples from the homogeneous
gabbronorite (C 48 and C7), the mineral separates (clinopyroxene and plagioclase) of C7
and the whole rock and clinopyroxenes of a dunite (C 174) have been determined. The
analyzed samples where crushed and sieved to obtain a 200-500 µm fraction. Enriched
mineral fractions were obtained through a Franz magnetic separator; those fractions were
extremely carefully hand picked under an optical binocular to obtain optical pure mineral
separate. Samples are leached in hydrochloric acid and washed several times in ultra pure
water. Sample size was between 50-100 mg. All Sm and Nd isotope data were obtained at
the Max Planck Institute für Chemie, Mainz following the procedures described elsewhere
(Jacob et al., 1994; Jagoutz, 1988; Thoni and Jagoutz, 1992). Blanks for the whole
chemical procedures are < 50pg for Sm and Nd. 143Nd/144Nd ratios are normalized against
the value of 146Nd/144Nd = 0.7219. Nd isotopic ratios reported here are relative to the value
of 0.511880 ± 15 for the LaJolla Nd standard. Uncertainty, including all sources of error of
Sm/Nd ratios is ±0.5%. The Sm-Nd data is listed in Table 5.3 and illustrated in an
isochrone diagram (Figure 5.7). The Nd isotopic compositions (143Nd/144Nd = 0.5129 -
147
0.5130) is typical of MORB. In a Sm/144Nd versus 143
Nd/144Nd diagram, whole rocks
and mineral separates define a linear array, corresponding to an age of 104 ± 27 Ma and
indicating isotopic equilibrium between the analysed samples. The Sm-Nd age dates the
time the two different rock units were separated from a common source, e.g. the time of
initiation of magmatic differentiation. The apparent age of 104 ± 27 Ma is consistent
within errors with previously published single grain U-Pb zircons ages of 85 Ma for the
gabbronorite (Schaltegger et al., 2002; Zeitler, 1985).

Table 5.3: Sm, Nd concentration and Nd Isotopic ratios for samples from the Chilas Complex.
 136

147
Figure 5.7: A Sm/144Nd versus 143
Nd/144Nd diagram of mineral separates (clinopyroxene and
plagioclase) and whole rocks from a dunite sample (C174) and two gabbronorite samples (C7 and
C48). Interpreting the linear relationship between the data-points as an isochrone corresponds to an
age of 104 ±27 Ma

5.5 Discussion

5.5.1 Crystallisation mechanism of the main gabbronorite

The occurrence of vast amounts of relatively homogeneous gabbronorite characterised by a
rather homogeneous modal abundance of pyroxene and plagioclase is astonishing. Several
points of evidence indicate a magmatic evolution. In the field different rock types (olivine-
bearing gabbronorite, gabbronorite, diorite and tonalite) have been recognized. The whole-
rock Mg# decrease from 0.79 to 0.48 and are associated with an increase of the REE by a
factor of ~50. These are typical features of so-called in-situ crystallisation. In-situ
crystallisation is thought to occur in a solidification zone, a transition zone between
consolidated cumulates and the magma chamber. Minerals nucleate and grow in-situ and
equilibrate to a different degree with interstitial liquid (McBirney and Noyes, 1979). In
contrast to the cumulate pile where interstitial liquid is effectively trapped, the
permeability in the solidification zone is sufficiently large to allow that a certain amount of
interstitial liquid can be expelled back into the main magma. A geochemical consequence
of such a model is that the behavior of incompatible trace elements is decoupled from that
of the major elements (Langmuir, 1989). The major element composition of the
gabbronorite is close to the liquid composition from which it crystallized and, accordingly,
the fractionated mineral mode remains constant, whereas the incompatible element
 137

concentration increases. In the following section results of trace element modelling are
presented to show that in-situ crystallisation is an important process in the Chilas
gabbronorite.
The whole rock composition has been modelled applying the equation for in-situ
crystallisation given by (Langmuir, 1989):

(1)

were C0, is the initial concentration of an element in the original liquid and CL is the
concentration in the remaining liquid, M0 and ML are the initial and remaining mass of the
magma chamber, respectively, and f is the mean amount of liquid returning into the magma
chamber from the solidification front (f=1 means all liquid returns from the solidification
front and the model is pure fractional crystallisation). E denotes the ratio of the
concentration of an element between the liquid returning from the solidification zone and
the magma chamber. Assuming complete equilibration between crystals and liquid in the
solidification front E ≈ 1/(D(1-f)+f) where D is the bulk solid partition coefficient for the
element of interest.

10
fractional crystallization trend f= 1 f= 0.75

f= 0.1
starting liquid composition
1
0.50 0.48
0.61

in-situ crystallization liquid

Ba/Sr

0.57 0.51
0.51
0.65 0.52
0.65
0.66 0.54
0.52
0.79
0.1

in-situ crystallization corresponding solid

0.01
10 100 1'000
Ba [ppm]

Figure 5.8: Results of the in-situ crystallisation modelling based on the model of (Langmuir, 1989).
Shown are own whole rock analyses of gabbronorite (squares) and unpublished ones (Khan, 1990)
 138

(small squares) compared with whole rock composition of the Braccia gabbro (circles) (Hermann et al.,
2001) and the Fiambala gabbro (crosses) (DeBari, 1994). The bulk solid partition coefficient used are
KD(Ba)= 0.045 KD(Sr)= 2.2 and were chosen to explain the observed spread of Ba concentrations. The
initial liquid composition is calculated in equilibrium with plagioclase in the olivine bearing
gabbronorite. KD are calculated using average An89 and 1100°C and the semi-empirical formulation of
Blundy and Wood (1991). Concentrations are Ba =113 ppm and Sr= 428 ppm. The corresponding solid
composition has been calculated assuming constant element concentration in the system. The evolution
of the liquid is shown for three cases: f= 1 is the pure fractional crystallisation trend, f= 0.75 and f= 0.1.
Notice that f = 1 indicates that 100 % of the liquid within the solidification zone returns into the
magma chamber, hence 0% of liquid is trapped in cumulates. The symbols of the modelled trend for f=
0.1 represent 10% of differentiation. The red line indicates the evolution of the liquid composition and
the black line the corresponding solid calculated assuming mass balance. Own samples have been
labeled with the corresponding Mg#. See text for further discussion.

5.5.1.1 Results
The whole rock composition have been modelled with 90% and 25% of trapped liquid (f=
0.1 and f= 0.75 respectively) (Figure 5.8). The trend of the whole rock data are well
reproduced assuming a high percentage of trapped liquid i.e. close to equilibrium
crystallisation and differs significantly from a fractional crystallisation trend (f=1).
However, the range of Ba content predicted by the model from fractionating a single
parental magma is strongly influenced by the bulk partition coefficient of Ba, which is not
well constrained in the case of Chilas. The bulk partition coefficient of Sr, and Ba are
dominated by the partition coefficient of plagioclase. The partition coefficient of Ba and Sr
between plagioclase and coexisting melt vary significantly as a function of the An content
(Bindeman and Davis, 2000; Blundy and Wood, 1991) and, at least Sr partitioning, is
pressure dependent (Vander et al., 2000). On the other hand, using inter-mineral
partitioning to calculate mineral-melt partition coefficients is strongly dependent on the
assumption that the minerals record near-solidus values and that the measured values are
not affected by fluid-/melt-derived micro-inclusions trapped in all minerals that might
particular affect Ba and Sr concentrations in plagioclase. Consequently, we varied the bulk
partition coefficient used in the model to cover the entire range of Ba composition
observed within genetically associated units within the Chilas Complex (30 to 316 ppm).
If the major part of the Chilas Complex evolved through fractionation of a single parental
magma, as indicated by the plagioclase composition (see chapter 4) Ba must have been
strongly incompatible with an averaged whole rock bulk partition coefficient <0.05
 139

implying an average plagioclase-melt partition coefficient of ~0.07. A similar KD value of

0.10-0.12 is obtained using a semi-empirical relation ship for plagioclase KD (Blundy and
Wood, 1991) assuming an average plagioclase composition of An75 and equilibration
temperatures of 1000-1100°C.
The Ba content of the whole rock samples, with Ba concentration and Ba/Sr ratios less than
the initial liquid composition (so-called solid composition), increase slightly with
decreasing whole rock Mg#. In contrast, the Ba concentrations in samples with Ba
concentrations and Ba/Sr ratios greater than the initial liquid composition (so-called liquid
composition) increase considerably with decreasing Mg# (Figure 5.9). The Mg# for the
solid composition evolve from 0.79 to 0.54 whereas Ba concentration increases by a factor
of ~3 from 30 to 100 ppm, whereas the liquid composition show a more restricted range of
Mg# from 0.52-0.48 with a relatively stronger increase in Ba content. The solid-
compositions are cumulate dominated with variable amounts of trapped liquid, whereas the
liquid-compositions are liquid dominated and therefore display buffered major element
compositions. The cross-over point of these two different trends marks the initial liquid
composition (Figure 5.9)

Figure 5.9: Two different trend observed in whole rock composition (symbols as in ). The Mg# of the
cumulate dominated trend decreases rapidly with increasing Ba content, whereas the Mg# in the liquid
is buffered and decreases only slightly with strongly increasing Ba concentration. The crossing point of
the two trends roughly corresponds to the initial melt composition parental to the Chilas gabbronorite.
 140

5.5.1.2 Implication of the modelling

The models give first order information of the nature of different gabbroic samples and
help to differentiate between cumulate- or liquid-dominated samples. In detail the models
have shown that:
(i) The large spread of Ba compositions observed in whole rocks can be explained by in-
situ crystallisation of a single parental magma composition. (ii) Cumulate-dominated
whole rock compositions are characterized by a rapid decrease of Mg# associated with a
slight increase of Ba concentration. (iii) The Mg# of liquid-dominate whole rock
compositions is buffered by the liquid composition whereas the Ba concentration increases
strongly.
In Figure 5.8 the trend of the Chilas gabbronorite is compared with the Braccia gabbro
(Hermann et al., 2001) and the Fiambala gabbro (DeBari, 1994). Both gabbro follow the
same trend as observed for the Chilas gabbronorite sequence. For the Braccia gabbro, in-
situ crystallization has been proposed indicating that in-situ crystallization of lower-crustal
intrusion could represent a common dominant process.

5.5.2 Origin of the dunitic core in the ultramafite bodies

Based on the crosscutting and replacive relationships described above the amphibole-
bearing porphyroclastic lherzolite are the oldest lithologies (Figure 5.2). Dunite first occurs
as dykes within the porphyroclastic lherzolite. The absence of any discontinuity between
the dunitic dykes and the massive dunite core indicates that it is a continuous lithology and
the formation mechanism for both was probably similar. The numerous dunitic dykes can
be explained by three mechanisms: 1) The dunite dyke represent crystallized ultramafic
liquid. 2) Dunite dykes represent former melt-filled dilational fracture zones. The
ascending melt crystallized olivine (+spinel) filling the fracture zone with cumulative
dunite. 3) The olivine originated through melt-rock interaction during localized porous
flow in non-dilational conduits forming “replacive dunite” (Boudier and Nicolas, 1972;
Boudier and Nicolas, 1977).
Thermal subduction zones models predict maximum mantle-wedge temperatures beneath
the volcanic arc of ~1200°C (e.g. Peacock and Wang, 1999) significantly lower than the
extremely high melting temperature of Mg-rich olivine (~1890°C Bowen and Schairer,
1935) rendering explanation 1) unlikely. The dunitic dykes rarely exceed a width of tens of
centimeters making them an ineffective transport mechanisim due to the large surface area
to volume ratio. Magma moving along small fractures would freeze or at least cool more
 141

rapidly than magma ascending in larger fractures. Accordingly if the dunite formed in
dilatational fractures the mineralogy should be different, which is not observed.
Additionally, irregular contact relationships of some dunitic dykes indicate replacement of
the porphyroclastic lherzolite by dunite making hypotheses 3) the most likely.
Based on structural and geochemical arguments “replacive dunite” have been interpreted
as channels of chemically isolated transport of melt (Kelemen et al., 1992; Kelemen et al.,
1995; Quick, 1982; Suhr, 1999). The REE pattern of a melt calculated to be in equilibrium
with the clinopyroxene in the primitive dunite sample (C174) corresponds to the REE
pattern of the whole rock samples of the gabbronorite (Figure 5.10 A). The Sm-Nd data
indicate isotopic equilibrium between the gabbronorite and the dunites (Figure 5.7).
Consequently, we interpret the origin of the dunite core as the result of infiltration of a
silica-undersaturated melt parental to the gabbronorite sequence into the porphyroclastic
lherzolite forming replacive dunite. Thin section observations indicate that the reaction
involved clinopyroxene and amphibole dissolution and precipitated olivine similar to the
reaction described by Kelemen (1990):

melt1 + x cpx + y amph = melt2 + z ol. (1)

Depending on the change in melt mass, and to lesser extent, to the energy budget of the
reaction, the melt flux can be facilitated or inhibited. Energy consideration is mainly
relevant if the melt-rock ratio is small. In such a case an endothermic reaction can freeze
the melt. In the case of the Chilas Complex we anticipate large melt masses and therefore
the change in melt mass is more important. An increase in melt mass results in an increase
of porosity and enhancing the melt flux. In order to estimate whether the reaction
facilitated the melt flux, we quantified the Ma/Mc (mass assimilated/ mass crystallized e.g.
mass (clinopyroxene + amphibole) dissolved over mass (olivine) precipitated). Due to the
enormous amount of gabbronorite in the Chilas Complex melt-rock ratios should be so
large that mass of melt1 ≈ melt2. Additionally, the modal amount of amphibole within the
porphyroclastic lherzolite is negligible compared to the amount of clinopyroxene, so that
Ma approx Mcpx. Precipitation of olivine affects mainly the Mg/Fe ratio of the whole rock.
The slope of the regression calculated from whole rock analyses of porphyroclastic
lherzolite and primitive dunite samples (Figure 5.10 B, see Figure caption for further
details) indicates Ma/Mc to be in the order of 4-5, falling in between the estimates of
Bowen (1922a) (Ma/Mc= 7) and Kelemen (1990) (Ma/Mc= 1.4). Due to the comparable
 142

heat of fusion of forsterite and diopside (Richet et al., 1993; Ziegler and Navrotsky, 1986)
we conclude the reaction is endothermic. The melt mass, however, increases. Accordingly,
the observed reaction could result in a self stimulating process, facilitating the melt flux
within the channel. It should however, be mentioned that the calculations are very sensitive
to the assumed Mg# of the precipitated olivine that is most probably not constant,
decreasing with increasing melt/rock ratios.

5.5.3 Origin of the zonation within the ultramafite bodies

The observed zoning of the ultramafic bodies is expressed by a decrease in olivine content
and an increase of pyroxene content towards the surrounding gabbronorite. The increase of
pyroxene content is irregular. Within a single lherzolite outcrop, dyke-shaped pyroxene-
rich domains can be identified. This observation contrasts with sharply defined layers
within ultramafite cumulate sequences (Irvine, 1974). Even so the succession from dunite
to lherzolite to pyroxenite is in accordance with the classical discontinuous differentiation
sequence (Bowen, 1922b), the field observations do not support an ordinary cumulate
origin resulting from crystal settling in a magma chamber. Reactive halos around
gabbronoritic dykes, dunite blocks within lherzolite and other observations described
above imply lherzolite formation due to melt-rock interaction. Accordingly, the zoning is
interpreted as the result of an irregular centimeter to meter-scale reaction front developed
at the contact between ultramafite-and surrounding gabbronorite. The reaction can be
explained by the infiltration of an olivine-undersatured and pyroxene oversaturated melt
reacting with olivine to form pyroxene according to the well-known peritectic reaction
(Bowen and Andersen, 1914):

Ol+melt1 = Pyx + melt2 (2)

The reaction results in localized or pervasive refertilization of the dunite. Whole rocks
analysis indicate different trends for rocks genetically linked by reaction (1) than for rocks
from the infiltration zone (Figure 5.10 C). The simple method presented in this study is
insufficient to characterize the second reaction since the presence of plagioclase on triple
junction (Burg et al., 1998) indicates that the infiltrating melt was already evolved and had,
beside pyroxene, plagioclase as a liquidus phase. Primarily based on structural arguments,
and the observed decrease of Mg#olivine with increasing pyroxene content (see chapter 4)
we interpret the zonation to result from the re-infitration of an olivine-undersaturated
 143

interstitial melt present in the semi-crystallized gabbronorite sequence into the dunitic melt
channels forming replacive lherzolite and pyroxenite (Bodiner et al., 2004; Kelemen and
Ghiorso, 1986).

5.5.4 The origin of the ultramafite bodies

Geochemical evidence indicates that the replacive dunite and the replacive lherzolite are
formed from a common magma genetically linked. High pressure phase equilibrium of
basaltic liquids show that liquids initially in equilibrium with olivine and pyroxene are
saturated only in olivine at lower pressure and will consume pyroxene and crystallize
olivine (e.g. Kelemen, 1995). Fractionation of olivine with decreasing temperature can
finally results in olivine-undersaturated melts. Consequently, a possible scenario could be
equilibrium crystallization in a magma chamber where assimilation of earlier fractionated
phases occurs. However, this process cannot easily explain the origin of the replacive
dunite and is not in accordance with the result of whole rock modeling presented above.
Additionally, kinematic criteria described previously (chapter 2) indicate vertical upwards
movement of the ultramafite with respect to the gabbronorite sequence. Since field
relations between the ultramafic bodies and the surrounding gabbronorite sequence
indicate intrusion of the ultramafic rocks under sub-solidus conditions into partially
consolidated gabbronorite mush, which in turn have contemporaneously intruded and
reacted with the ultrabasic rocks the uprising cannot be explained by a later event.
The intrusion of the ultramafite can be explained by buoyancy driven up-rise due to high
melt fraction constrained within the melt channels. However, even with high melt-rock
ratio it is unrealistic that the density of the melt channels is lower than the surrounding
gabbronorite. The crucial parameter is the depth-integrated mean density difference of the
melt-channel must be lower than that of the surrounding wall rock to mobilize the
ultramafite (e.g. Biot and Ode, 1965). If the ultramafite are emplaced due to density driven
forces, the melt channels must reach into dense ultramafite rock. i.e. into the upper mantle.
It has been shown numerically that density driven diapirs can grow through and above the
neutral buoyancy level (Poliakov et al., 1996). We speculated that the ultramafite are
surface expression of vertically continuous melt channels reaching into the upper mantle,
possibly already down to the top of the melting region. Melt channels with high melt-rock
ratios were emplaced into the gabbronorite sequence due to buoyancy. Accordingly, the
dyke-like shape of the ultramafite closely resembles the accurate structure of the melt
channel at depth.
 144

Figure 5.10: A) Primitive mantle (Sun and McDonough, 1989) normalized spider-diagram of whole
rock data for homogeneous gabbronorite solid symbols. Open boxes are the mean values of a melt
calculated in equilibrium with primitive (Mg# =0.92-0.94) clinopyroxene (using KD values of (Hart and
Dunn, 1993) with an extrapolated value for Eu from (Ionov et al., 2002)) extracted from a primitive
dunitic sample (C 174). The red field indicates the spread of the six clinopyroxene analyzed. B) Same
samples in molar Fe/Mn vs Mg/Mn plot. The slope of the reaction from amphibolite lherzolite to
replacive dunite is 0.22. Precipitation of only olivine with Mg# 0.8 would result in a slope of 0.25
 145

wereas clinopyroxene or amphibole dissolution alone in a slope of -0.1. The current model implies that
1 mole of Olivine precipitation is associated with redissolution of 3 moles of cpx/amph. C) Molar
Ca/Mn vs Mg/Mn plot for whole rock analysis from Amphibole bearing peridotite (squares), Dunites
and lherzolite (diamonds) and gabbronoritic (triangles) samples. Three different trends can be
identified. 1.) Amphibole bearing lherzolite loose mainly Ca and gain slightly Mg when transformed to
Dunite indicating disolution of clinopyroxene/amphibole and precipitation of olivine (see text for
details). 2.) Dunite transformation to lherzolites show two different trends corresponding to the
presence or absence of plagioclase in the sample, but they approach the filed of gabbro-gabbronorite
samples confirming the field observed re-infiltration of melt parental to the gabbronorite into the
dunitic units. The observed trend in whole rocks results from both reaction of melt + ol = pyx and
trapped interstitial liquids ultimatively forming the characteristic zonation of the zoned ultramafic
bodies.

5.5.5 Basaltic or andesitic flux out of the mantle?

A basic assumption of most models of arc growth is that the melt passing from the mantle
wedge into the overlying arc crust is close to Fe-Mg exchange equilibrium with mantle
peridotites (Mg#melt=0.70 Kelemen et al., 2004), i.e. fractional crystallization within the
upper mantle is of minor importance. We have argued that the zoned ultramafic bodies of
the Kohistan are melt channels feeding the gabbronorite intrusion. The bulk composition of
the gabbronorite sequence should equal the flux out of the melt channels since the Chilas
Complex mainly crystallized in-situ. In order to obtain an estimate of the parental magma
composition we compare the results of two different approaches: 1) We calculated a melt
in equilibrium with clinopyroxene present in the most primitive cumulate (C66); and 2)
The bulk composition should be located between that of the evolved cumulate-dominated
rocks (e.g. C 48 and C7) and the primitive liquid-dominated rocks (e.g. C138 and C 220),
assuming that no significant amount of highly evolved melt escaped into higher levels. In a
first step we calculated the arithmetic mean REE composition of the gabbronorite whole
rock samples (excluding the extremely primitive C66). This "arbitrary" average
composition displays a positive Eu anomaly and is thus dominated by cumulus-dominated
samples. To correct this bias, we added an average interstitial liquid composition,
calculated in equilibrium with an average of 17 clinopyroxene analysed from cumulate
gabbronorite (C48 and C7). Mixing calculations indicate that 5% of such an interstitial
liquid is required to eliminate the Eu anomaly. Both calculated average gabbronorite
compositions are nearly equal and are illustrated in Figure 5.11.
 146

Figure 5.11: Primitive mantle normalized spidergram illustrating the calculated parental magma
composition of the Chilas gabbronorite sequence. The parental liquid calculated in equilibrium with
primitive clinopyroxene (using the above cited set of KD`s) from olivine-bearing gabbronorite (C66) is
indicated as red boxes. The Black line is the parental liquid calculated assuming that the parental
liquid is located in between the liquid- and cumulate-dominated composition corrected for trapped
interstitial liquid.

The REE concentrations of these average compositions are very similar and lie between
sample C 134 and C 135. We assume that those two samples are close proxies for the bulk
composition of the gabbronorite for both trace and major elements. The calculated parental
bulk composition is a basaltic-andesite (SiO2 52-54wt%; Na2O +K2O ~4 wt% and Mg#~
0.52-0.54). For the major elements a similar result is obtained by calculating the average
composition of 87 published and unpublished Chilas gabbronorite whole rock analyses
(Khan, 1988; Mikoshiba et al., 1999).
In Figure 5.12 an average composition of the liquid-dominated rocks and the calculated
bulk composition are compared with lower crust estimates of Taylor and McLennan (1985)
and Rudnick and Gao (2003). The similarity between the lower crustal estimates and the
liquid dominated rocks, and the calculated bulk composition is striking although the
agreement seems to be slightly better for the liquid dominated rock composition. However,
the incompatible elements (U, Th, Rb, Ta) of the calculated bulk and to a lesser extent of
the liquid dominated rocks deviate significantly from lower crustal estimates. This
mismatch implies selective escape of a few highly incompatible elements through evolved
melts or a fluid phase derived form the Chilas intrusion. In conclusion the flux out of the
 147

melt channels is in the case of the Chilas Complex essentially andesitic and very close to
estimated lower crustal compositions. The major element and trace element characteristic
(low SiO2, high TiO2 and low NaO2, DyN/YbN< 1.2 and low Sr/Nd (~ 30-60)) for both the
approximated bulk composition and the liquid compositions indicate that the gabbronorite
is more likely derived by differentiation of a primitive basaltic melt than from a primitive
andesite (Kelemen et al., 2004).

Figure 5.12: Comparison of crystallized liquid (above) and the calculated bulk gabbronorite
composition (below) with lower crust estimates of Taylor and McLennan (1985) and Rudnick and Gao
(2003) (triangles and diamonds respectively) see text for details.

5.5.6 A model for the origin of the lower continental crust

The tectonic position of the Chilas Complex is related to the initial stage of back arc
spreading in an extending island arc due to mantle diapirism (Burg et al., 1998; Khan et al.,
 148

1993). Accordingly, the bulk composition reflects a lower crustal intrusion rather than a
mid crustal intrusion although the intrusion depth is only around 0.6-0.7 GPa. We
speculate that the melt channels are vertically continuous reaching into the upper mantle.
Consequently, we propose, in accordance with the model of Kay and Kay (1985b), that in
the case of the Chilas Complex primitive basaltic liquids derived from the mantle wedge
fractionate olivine, spinel and pyroxene as they ascend through isolated melt-channels
through the upper mantle forming the replacive dunite. The presence of low Mg#whole-rock
dunite samples indicate that the Mg#melt of the melt is lowered from > 0.70 to ~0.50-0.60
through differentiation. The change in seismic properties from mantle values (Vp ~ 8 km s-
1
) to lower crust values (Vp < 7.3 km s-1) would mark the beginning of plagioclase
fractionation. The bulk flux through the seismic MOHO is thus basaltic-andesitic to
andesitic whereas the bulk flux through the petrological Moho is basaltic. Accordingly, the
bulk continental crust estimates based on seismic properties is andesitic, whereas the bulk
continental crust estimates based on petrology is basaltic because it includes ultramafic
cumulates within the upper mantle. The world wide association of zoned ultramafic bodies,
of which some have been previously interpret as feeder channels (Murray, 1972), with
lower crustal mafic intrusion within a present or inferred subduction zone setting
(Fershtater et al., 1997; Kelemen and Ghiorso, 1986; Noble and Taylor Jr, 1960; Snoke et
al., 1981; Thayer, 1960; Tistl et al., 1994) points to a common process of lower crust
formation. However, we want to stress the point that not all ultramafic bodies are
necessarily feeder channels, especially those with sharply defined layered sequences
represent ultramafic cumulates (Irvine, 1974). The presented model does not require a
priori that delamination of the ultramafite bodies must occur to explain the composition of
the arc crust. Although it might nevertheless occur because Fe-richer cumulates are denser
than surrounding mantle (Müntener et al., 2001).

5.6 Conclusion
We have presented field observations, mineral trace element and whole rock chemistry.
The dunitic core of the ultramafite bodies present within the Chilas Complex are
interpreted as replacive dunite channels. Geochemical data indicate that the melt parental
to the gabbronorite sequence formed this dunite. Since the ultramafite intrude the
gabbronorite sequence they have been interpreted as melt channels feeding the Chilas
Complex. We speculated that the melt channels reached into the upper mantle. It has been
shown that the melt passing out of the melt channels has an andesitic composition whereas
 149

the parental magma was a primitive basalt liquid. Consequently fractionation of mantle
derived can take place within isolated conduits. This process could be a significant crust
forming process explaining the absence of high Mg# cumulates in the lower crust of
exposed island arcs.
 150

References

Anderson, A. T., 1982, Parental basalts in subduction zones: implications for continental
evolution: Journal of Geophysical Research, v. 87, p. 7047-7060.
Bard, J. P., 1983, Metamorphism of an obducted island arc; example of the Kohistan
Sequence (Pakistan) in the Himalayan collided range: Earth and Planetary Science
Letters, v. 65, p. 133-144.
Bignold, S. M., and P. J. Treloar, 2003, Northward subduction of the Indian Plate beneath
the Kohistan island arc, Pakistan Himalaya; new evidence from isotopic data:
Journal of the Geological Society of London, v. 160, p. 377-384.
Bindeman, I. N., and A. M. Davis, 2000, Trace element partitioning between plagioclase
and melt; investigation of dopant influence on partition behavior: Geochimica et
Cosmochimica Acta, v. 64, p. 2863-2878.
Biot, M. A., and H. Ode, 1965, Theory of gravity instability with variable overburden and
compaction: Geophysics, v. 30, p. 213-227.
Blundy, J. D., and B. J. Wood, 1991, Crystal-chemical controls on the partitioning of Sr
and Ba between plagioclase feldspar, silicate melts, and hydrothermal solutions:
Geochimica et Cosmochimica Acta, v. 55, p. 193-209.
Bodiner, J.-L., M. A. Menzies, N. Shimizu, F. A. Frey, and E. McPherson, 2004, Silicate,
Hydrous and Carbonate Metasomatism at Lherz, France: Contemporaneous
Derivatives of Silicate Melt-Harzburgite Reaction: J. Petrology, v. 45, p. 299-320.
Boudier, F., and A. Nicolas, 1972, Fusion partielle gabbroique dans la lherzolite de Lanzo
(Alpes piemontaises): Schweizerische Mineralogische und Petrographische
Mitteilungen = Bulletin Suisse de Mineralogie et Petrographie, v. 52, p. 39-56.
Boudier, F., and A. Nicolas, 1977, Structural controls on partial melting in the Lanzo
peridotites, in H. J. B. Dick, ed., Magma genesis 1977; proceedings of the
American Geophysical Union Chapman conference on partial melting in the
Earth's upper mantle.: Bulletin - Oregon, Department of Geology and Mineral
Industries, v. 96: Portland, OR, United States, Oregon Department of Geology and
Mineral Industries, p. 63-78.
Bowen, N. L., 1922a, The behavior of inclusions in igneous magmas: Journal of Geology,
v. 30, p. 513-570.
 151

Bowen, N. L., 1922b, The reaction principle in petrogenesis: Journal of Geology, v. 30, p.
177-198.
Bowen, N. L., and O. Andersen, 1914, The binary system MgO-SiO (sub 2): American
Journal of Science, v. 37, p. 487-500.
Bowen, N. L., and J. F. Schairer, 1935, The system MgO-FeO-SiO (sub 2): American
Journal of Science, v. 29, p. 151-217.
Burg, J. P., J. L. Bodinier, S. Chaudhry, S. Hussain, and H. Dawood, 1998, Infra-arc
mantle-crust transition and intra-arc mantle diapirs in the Kohistan Complex
(Pakistani Himalaya); petro-structural evidence: Terra Nova. The European
Journal of Geosciences, v. 10, p. 74-80.
Conrad, W. K., and R. W. Kay, 1984, Ultramafic and mafic inclusions from Adak Island;
crystallization history, and implications for the nature of primary magmas and
crustal evolution in the Aleutian Arc: Journal of Petrology, v. 25, p. 88-125.
Coward, M. P., R. W. H. Butler, K. M. Asif, and R. J. Knipe, 1987, The tectonic history of
Kohistan and its implications for Himalayan structure: Journal of the Geological
Society of London, v. 144, p. 377-391.
Coward, M. P., B. F. Windley, R. D. Broughton, I. W. Luff, M. G. Petterson, C. J. Pudsey,
D. C. Rex, and K. M. Asif, 1986, Collision tectonics in the NW Himalayas, in M.
P. Coward, and C. Ries Alison, eds., Collision tectonics.: Geological Society
Special Publications, v. 19: London, United Kingdom, Geological Society of
London, p. 203-219.
Cox, K. G., J. D. Bell, and R. J. Pankhurst, 1979, The interpretation of igneous rocks:
London, Allen and Unwin, 450 p.
Crawford, A. J., T. J. Falloon, and S. Eggins, 1987, The origin of island arc high-alumina
basalts: Contributions to Mineralogy and Petrology, v. 97, p. 417-430.
DeBari, S. M., 1994, Petrogenesis of the Fiambala gabbroic intrusion, northwestern
Argentina, a deep crustal syntectonic pluton in a continental magmatic arc:
Journal of Petrology, v. 35, p. 679-713.
DeBari, S. M., and N. H. Sleep, 1991, High-Mg, low-Al bulk composition of the Talkeetna
island arc, Alaska; implications for primary magmas and the nature of arc crust:
Geological Society of America Bulletin, v. 103, p. 37-47.
Ducea, M., and J. B. Saleeby, 1998, Crustal recycling beneath continental arcs; silica-rich
glass inclusions in ultramafic xenoliths from the Sierra Nevada, California: Earth
and Planetary Science Letters, v. 156, p. 101-116.
 152

Fershtater, G. B., P. Montero, N. S. Borodina, E. V. Pushkarev, V. N. Smirnov, and F. Bea,

1997, Uralian magmatism; an overview, in A. Perez Estaun, D. Brown, and D.
Gee, eds., Europrobe's Uralides Project.: Tectonophysics, v. 276; 1-4: Amsterdam,
Netherlands, Elsevier, p. 87-102.
Hart, S. R., and T. Dunn, 1993, Experimental cpx/melt partitioning of 24 trace elements:
Contributions to Mineralogy and Petrology, v. 113, p. 1-8.
Hermann, J., O. Müntener, and D. Günther, 2001, Differentiation of Mafic Magma in a
Continental Crust-to-Mantle Transition Zone: J. Petrology, v. 42, p. 189-206.
Ionov, D. A., J.-L. Bodinier, S. B. Mukasa, and A. Zanetti, 2002, Mechanisms and Sources
of Mantle Metasomatism: Major and Trace Element Compositions of Peridotite
Xenoliths from Spitsbergen in the Context of Numerical Modelling: J. Petrology,
v. 43, p. 2219-2259.
Irvine, T. N., 1974, Petrology of the Duke Island ultramafic complex, southeastern Alaska:
Memoir - Geological Society of America, v. 138: Boulder, CO, United States,
Geological Society of America (GSA), 240 p.
Jacob, D., E. Jagoutz, D. Lowry, D. Mattey, and G. Kudrjavtseva, 1994, Diamondiferous
eclogites from Siberia; remnants of Archean ocean crust: Geochimica et
Cosmochimica Acta, v. 58, p. 5191-5207.
Jagoutz, E., 1988, Nd and Sr systematics in an eclogite xenolith from Tanzania; evidence
for frozen mineral equilibria in the continental lithosphere: Geochimica et
Cosmochimica Acta, v. 52, p. 1285-1293.
Jan, M. Q., and R. A. Howie, 1981, The mineralogy and geochemistry of the
metamorphosed basic and ultrabasic rocks of the Jijal Complex, Kohistan, NW
Pakistan: Journal of Petrology, v. 22, p. 85-126.
Jan, M. Q., and B. F. Windley, 1990, Chromian spinel-silicate chemistry in ultramafic
rocks of the Jijal Complex Northwest Pakistan: Journal of Petrology, v. 31, p.
667-715.
Kay, R. W., and S. M. Kay, 1991, Creation and destruction of lower continental crust, in B.
Stoeckhert, and K. H. Wedepohl, eds., Crustal dynamics; pathways and records.:
International Journal of Earth Sciences, v. 80; 2: Berlin, Federal Republic of
Germany, Springer International, p. 259-278.
Kay, S. M., and R. W. Kay, 1985a, Aleutian tholeiitic and calc-alkaline magma series; 1,
The mafic phenocrysts: Contributions to Mineralogy and Petrology, v. 90, p. 276-
290.
 153

Kay, S. M., and R. W. Kay, 1985b, Role of crystal cumulates and the oceanic crust in the
formation of the lower crust of the Aleutian arc: Geology, v. 13, p. 461-464.
Kelemen, P., K. Hanghoj, and A. Greene, 2004, One view of the geochemistry of
subduction-related magmatic arcs, with an emphasis on primitive andesite and
lower crust., in R. L. Rudnick, ed., The Crust: Treatise on Geochemistry, v. 3:
Oxford, Elsevier-Pergamon, p. 593-659.
Kelemen, P. B., 1990, Reaction between ultramafic rock and fractionating basaltic magma;
I, Phase relations, the origin of calc-alkaline magma series, and the formation of
discordant dunite: Journal of Petrology, v. 31, p. 51-98.
Kelemen, P. B., 1995, Genesis of high Mg andesites and the continental crust:
Contributions to Mineralogy and Petrology, v. 120, p. 1-19.
Kelemen, P. B., H. J. B. Dick, and J. E. Quick, 1992, Formation of harzburgite by
pervasive melt/ rock reaction in the upper mantle: Nature (London), v. 358, p.
635-641.
Kelemen, P. B., and M. S. Ghiorso, 1986, Assimilation of peridotite in zoned calc-alkaline
plutonic complex; evidence from the Big Jim Complex, Washington Cascades:
Contributions to Mineralogy and Petrology, v. 94, p. 12-28.
Kelemen, P. B., N. Shimizu, and V. J. M. Salters, 1995, Extraction of mid-ocean-ridge
basalt from the upwelling mantle by focused flow of melt in dunite channels:
Nature (London), v. 375, p. 747-753.
Khan, M. A., 1988, Petrology and structure of the Chilas ultramafic-mafic complex,
Kohistan Arc, NW Himalayas, University of London, London.
Khan, M. A., M. Q. Jan, and B. L. Weaver, 1993, Evolution of the lower arc crust in
Kohistan, N. Pakistan; temporal arc magmatism through early, mature and intra-
arc rift stages, in P. J. Treloar, and M. P. Searle, eds., Himalayan tectonics.:
Geological Society Special Publications, v. 74: London, United Kingdom,
Geological Society of London, p. 123-138.
Khan, M. A., M. Q. Jan, B. F. Windley, J. Tarney, and M. F. Thirlwall, 1989, The Chilas
mafic-ultramafic igneous complex; the root of the Kohistan island arc in the
Himalaya of northern Pakistan, in L. Malinconico Lawrence, Jr., and J. Lillie
Robert, eds., Tectonics of the western Himalayas.: Special Paper - Geological
Society of America, v. 232: Boulder, CO, United States, Geological Society of
America (GSA), p. 75-94.
 154

Khan, M. A., R. J. Stern, R. F. Gribble, and B. F. Windley, 1997, Geochemical and

isotopic constraints on subduction polarity, magma sources and palaeogeography
of the Kohistan Arc, northern Pakistan: Journal of the Geological Society of
London, v. 154 Part 6, p. 935-946.
Kubo, K., Y. Swada, Y. Takahashi, A. B. Kausar, Y. Seki, I. H. Khan, T. Khan, N. A.
Khan, and Y. Takahashi, 1996, The Chilas igneuous complex in the western
Himalayas of northern Pakistan: Proceedings of Geoscience Colloquium,
Geoscience Labatory; Geological Survey, Islamabad, v. 15, p. 63-68.
Langmuir, C. H., 1989, Geochemical consequences of in situ crystallization: Nature
(London), v. 340, p. 199-205.
McBirney, A. R., and R. M. Noyes, 1979, Crystallization and layering of the Skaergaard
Intrusion: Journal of Petrology, v. 20, p. 487-554.
McDonough, W. F., and S. S. Sun, 1995, The composition of the Earth, in W. F.
McDonough, N. T. Arndt, and S. Shirey, eds., Chemical evolution of the mantle.:
Chemical Geology, v. 120; 3-4: Amsterdam, Netherlands, Elsevier, p. 223-253.
Mikoshiba, M. U., Y. Takahashi, Y. Takahashi, A. B. Kausar, T. Khan, K. Kubo, and T.
Shirahase, 1999, Rb-Sr isotopic study of the Chilas igneous complex, Kohistan,
northern Pakistan, in A. Macfarlane, B. Sorkhabi Rasoul, and J. Quade, eds.,
Himalaya and Tibet; mountain roots to mountain tops.: Special Paper - Geological
Society of America, v. 328: Boulder, CO, United States, Geological Society of
America (GSA), p. 47-57.
Miyashiro, A., 1978, Nature of alkalic volcanic rock series: Contributions to Mineralogy
and Petrology, v. 66, p. 91-104.
Muentener, O., P. B. Kelemen, and T. L. Grove, 2001, The role of H (sub 2) O during
crystallization of primitive arc magmas under uppermost mantle conditions and
genesis of igneous pyroxenites; an experimental study: Contributions to
Mineralogy and Petrology, v. 141, p. 643-658.
Müntener, O., P. B. Kelemen, and T. L. Grove, 2001, The role of H (sub 2) O during
crystallization of primitive arc magmas under uppermost mantle conditions and
genesis of igneous pyroxenites; an experimental study: Contributions to
Mineralogy and Petrology, v. 141, p. 643-658.
Murray, C. G., 1972, Zoned Ultramafic Complexes of the Alaskan Type; Feeder Pipes of
Andesitic Volcanoes, Studies in Earth and Space Sciences.: Memoir - Geological
 155

Society of America, v. 132: Boulder, CO, United States, Geological Society of

America (GSA), p. 313-335.
Noble, J. A., and H. P. Taylor Jr, 1960, Correlation of the ultramafic complexes of south
eastern Alaska with those of other parts of North America and the world: Report
of the ... Session - International Geological Congress, v. Part 13, p. 188-197.
Peacock, S. M., and K. Wang, 1999, Seismic consequences of warm versus cool
subduction metamorphism; examples from Southwest and Northeast Japan:
Science, v. 286, p. 937-939.
Petterson, M. G., and B. F. Windley, 1985, Rb-Sr dating of the Kohistan arc-batholith in
the Trans-Himalaya of North Pakistan, and tectonic implications: Earth and
Planetary Science Letters, v. 74, p. 45-57.
Pettke T., C. A. H., A.C. Ciocan, D. Guenther,, 2000, Quadrupole mass spectrometry and
optical emission spectroscopy: Detection capabilities and representative sampling
of short transient signals from laser-ablation: Journal of analytical atomic
spectrometry, v. 15, p. 1149-1155.
Poliakov, A. N. B., Y. Y. Podladchnikov, E. C. Dawson, and C. J. Talbot, 1996, Salt
diapirism with simultaneous brittle faulting and viscous flow, in G. I. Alsop, D. J.
Blundell, and I. Davison, eds., Salt tectonics.: Geological Society Special
Publications, v. 100: London, United Kingdom, Geological Society of London, p.
291-302.
Pudsey, C. J., M. P. Coward, I. W. Luff, R. M. Shackleton, B. F. Windley, and M. Q. Jan,
1985, Collision zone between the Kohistan Arc and the Asian Plate in NW
Pakistan: Transactions of the Royal Society of Edinburgh: Earth Sciences, v. 76,
p. 463-479.
Quick, J. E., 1982, The origin and significance of large, tabular dunite bodies in the Trinity
Peridotite, northern California: Contributions to Mineralogy and Petrology, v. 78,
p. 413-422.
Richet, P., F. Leclerc, and L. Benoist, 1993, Melting of forsterite and spinel, with
implications for the glass transition of Mg (sub 2) SiO (sub 4) liquid: Geophysical
Research Letters, v. 20, p. 1675-1678.
Ringuette, L., J. Martignole, and B. F. Windley, 1999, Magmatic crystallization, isobaric
cooling, and decompression of the garnet-bearing assemblages of the Jijal
Sequence (Kohistan Terrane, western Himalayas): Geology (Boulder), v. 27, p.
139-142.
 156

Rudnick, R. L., 1995, Making continental crust: Nature, v. 378, p. 571-578.

Rudnick, R. L., and S. Gao, 2003, The Composition of the Continental Crust, in R. L.
Rudnick, ed., The Crust: Treatise on Geochemistry, v. 3: Oxford, Elsevier, p. 1-
64.
Schaltegger, U., G. Zeilinger, M. Frank, and J. P. Burg, 2002, Multiple mantle sources
during island arc magmatism; U-Pb and Hf isotopic evidence from the Kohistan
arc complex, Pakistan: Terra Nova, v. 14, p. 461-468.
Searle, M. P., M. A. Khan, J. E. Fraser, S. J. Gough, and J. M. Qasim, 1999, The tectonic
evolution of the Kohistan-Karakoram collision belt along the Karakoram Highway
transect, North Pakistan: Tectonics, v. 18, p. 929-949.
Snoke, A. W., J. E. Quick, and H. R. Bowman, 1981, Bear Mountain igneous complex,
Klamath Mountains, California; an ultrabasic to silicic calc-alkaline suite: Journal
of Petrology, v. 22, p. 501-552.
Suhr, G., 1999, Melt migration under oceanic ridges; inferences from reactive transport
modelling of upper mantle hosted dunites: Journal of Petrology, v. 40, p. 575-599.
Sun, S. S., and W. F. McDonough, 1989, Chemical and isotopic systematics of oceanic
basalts; implications for mantle composition and processes, in A. D. Saunders,
and M. J. Norry, eds., Magmatism in the ocean basins.: Geological Society
Special Publications, v. 42: London, United Kingdom, Geological Society of
London, p. 313-345.
Tahirkheli, R. A. K., 1979, Geology of Kohistan and adjoining Eurasia and Indio-Pakistan
continents, Pakistan: Geological Bulletin, University of Peshawar, v. 11, p. 1-30.
Taylor, S. R., and S. M. McLennan, 1985, The Continental crust: Its Composition and
Evolution: Oxford, Blackwell.
Thayer, T. P., 1960, Some critical differences between alpine-type and stratiform
peridotite-gabbro complexes: Report of the ... Session - International Geological
Congress, v. Part 13, p. 247-259.
Thoni, M., and E. Jagoutz, 1992, Some new aspects of dating eclogites in orogenic belts;
Sm-Nd, Rb-Sr, and Pb-Pb isotopic results from the Austroalpine Saualpe and
Koralpe type-locality (Carinthia/ Styria, southeastern Austria): Geochimica et
Cosmochimica Acta, v. 56, p. 347-368.
Tistl, M., K. P. Burgath, A. Hoehndorf, H. Kreuzer, R. Munoz, and R. Salinas, 1994,
Origin and emplacement of Tertiary ultramafic complexes in Northwest
 157

Colombia; evidence from geochemistry and K-Ar, Sm-Nd and Rb-Sr isotopes:
Earth and Planetary Science Letters, v. 126, p. 41-59.
Treloar, P. J., M. G. Petterson, M. Q. Jan, and M. A. Sullivan, 1996, A re-evaluation of the
stratigraphy and evolution of the Kohistan Arc sequence, Pakistan Himalaya;
implications for magmatic and tectonic arc-building processes: Journal of the
Geological Society of London, v. 153 Part 5, p. 681-693.
Vander, A. J., J. Longhi, and J. C. Duchesne, 2000, The effect of pressure on D (sub Sr)
(plag/ melt) and D (sub Cr) (opx/ melt); implications for anorthosite petrogenesis:
Earth and Planetary Science Letters, v. 178, p. 303-314.
Zeitler, P. K., 1985, Cooling history of the NW Himalaya, Pakistan: Tectonics, v. 4, p.
127-151.
Ziegler, D., and A. Navrotsky, 1986, Direct measurement of the enthalpy of fusion of
diopside: Geochimica et Cosmochimica Acta, v. 50, p. 2461-2466.
 158

6 The plutonic crust of an island arcs: History from LA-

ICPMS U-Pb zircon ages of intrusive units from the
Kohistan paleo-island arc (NW Pakistan).

O. Jagoutz1, J.P. Burg1, T. Iizuka2, T. Hirata2, S Maruyama2, S. Hussain3, H. Dawood3,

N.M. Chaudhry4

1
Department of Earth Sciences, ETH Zurich, Sonneggstr. 5, CH-8092 Zurich, Switzerland
2
Department of Earth and Planetary Sciences, Tokyo Institute of Technology, Tokyo 152, Japan
3
Pakistan Museum of Natural History, Garden Avenue, Shakarparian, 44 000 Islamabad, Pakistan
4
University of the Punjab, Quaid-e-Azam Campus 54590 Lahore, Pakistan
 159

This chapter is written in paper format. U-Pb laser ablation ICP-MS zircon ages from
plutonic samples from the Kohistan Batholith are presented and discussed.

Abstract
U-Pb zircons ages from plutonitic samples of the SW Kohistan paleo-island arc using laser
ablation ICP-MS indicate continuous magmatism from at least 78 Ma to approx 44 Ma.
Relative dating of plutons according to the presence or the absence of a penetrative fabric
is inconsistent with the U-Pb ages since undeformed rocks may be older than deformed
ones. Intrusive relationships indicate that some of the so-called Dir-Utror meta-volcanite
are older than 71 Ma. Geochemical data suggest that these metavolcanites are, at least in
part, the effusive equivalents of the Chilas gabbronorite. This information supports the
conclusion that the Kohistan arc has been rifted. Accumulation of volcanite and sediments
occurred in a related basin that possibly existed until collision of India with Kohistan.

6.1 Introduction
Intra-oceanic subduction, building of an island arc and subsequent collision are considered
to be a major mechanism for the development of the continental crust (Kelemen, 1995).
However, complete sections of an island arc are rare and the best preserved is probably the
Kohistan paleo-arc, in NW Pakistan. Geological understanding of the evolution of the
Kohistan arc is therefore essential to infer and comprehend crust differentation taking place
in active island arcs. It is generally assumed that the Kohistan arc evolved through three
main stages 1) as an intra-oceanic island arc separated from Eurasia by a backarc basin of
unknown size, 2) as an active continental margin after accretion with Eurasia and 3) as the
leading edge of a continent-continent collision between India and Eurasia (Petterson and
Windley, 1985; Searle et al., 1999; Treloar et al., 1996). Active island arcs often evolve
from juvenile through intra-arc rifting to mature stages (e.g. Tamaki, 1985). The lack of
geochronological information hampers the attribution of different tectono-magmatic units
to the different stages of the Kohistan arc. This information is essential to clarify the
geological evolution of the arc and the related arc building mechanisms. This information
is particularly essential to approximate the bulk composition of the intra oceanic Kohistan
arc at various stages of its growth. In this paper we present U-Pb zircon ages form several
intrusive units from the central SW-Kohistan arc.
 160

6.2 Geological Setting

6.2.1 The Kohistan arc

The Kohistan arc, located in NE Pakistan (Figure 6.1) is bounded to the north from the
Karakoram micro plate by the Kohistan-Karakoram Suture and in the south from the
Indian Plate by the Indus Suture Zone. The Kohistan is generally regarded as a fossil
Jurassic/Cretaceous island arc that was wedged between the Indian and Asian plates during
the Himalayan collision (Bard, 1983; Tahirkheli et al., 1979). The intraoceanic arc
originated through a northward dipping subduction somewhere in the equatorial area of the
Tethys Ocean in the vicinity of the Eurasian continent (Bignold and Treloar, 2003; Coward
et al., 1987; Khan et al., 1993). The entire section of the full structure of an arc ranging
from oceanic upper mantle to the uppermost volcanic sequences is supposedly now
exposed (Coward et al., 1986; Searle et al., 1999; Treloar et al., 1996).

Figure 6.1: Geological map of the Kohistan arc. The box indicates the location of map Sample C-01-77
is from the western part of the Kohistan arc, near the town of Chilas and indicated separately. The
location of the remaining samples are given in Figure 6.2. The location of Figure 6.8 in the northern
part of the Kohistan arc is additionally indicated

The Jjial- Kalam complex, representing the lower crust of the island arc (Jan and Howie,
1981; Jan and Windley, 1990; Ringuette et al., 1999) is exposed in the southern part of the
Kohistan. It is separated by the Chilas Complex, interpreted to have intruded during
 161

incipient back-arc rifting (Burg et al., 1998; Khan et al., 1989; Khan et al., 1997) from the
Gilgit Domain in the north, which is composed dominantly of upper crustal plutonic,
volcanic and sedimentary rocks and their metamorphosed equivalents (Petterson and
Windley, 1985); (Pudsey et al., 1985). The Gilgit Domain is in turn subdivided into the so-
called Kohistan Batholith with subordinate volcano-sedimentary units (south and centre)
and volcano-sedimentary series (Chalt, Shamran and Utror volcanites and Yasin
sediments) found in the north (Tahirkheli et al., 1979). In the vicinity of the town Dir
(Figure 6.2), plutons of the Batholith intrude meta-sediments and so-called Dir or Utror
volcanite, which are associated with the Barul Banda slate formation (Sullivan et al.,
1993). Hornblende of an andesitic lava flow yielded an 40Ar-39Ar age of 55 ± 2 Ma years
(Treloar et al., 1986) and fossils indicates a Late Palaeocene age for the Baraul Banda Slate
Fm (Sullivan et al., 1993). Our work focused on dating plutonic bodies with clear
relationships to the volcanites and sediments.

6.2.2 Previous geochronological constraints

Based on Rb-Sr whole rock data, Ar-Ar hornblende dating and the presence or absence of a
penetrative fabric within the plutonic units, the evolution of the Kohistan Batholith has
been subdivided into three stages (George et al., 1993; Petterson and Windley, 1985;
Treloar et al., 1989): The first stage (104-85 Ma) is composed of a bi-modal sequence of
foliated gabbro-diorites and trondhjemites. Stage 2 consists of undeformed gabbros,
diorites and granites that intruded between 85 and 40 Ma with a general basic to acid trend.
Sr and Nd contents supportively indicate an increasing crustal component for the younger
stage 2 plutons (Petterson et al., 1993; Petterson and Windley, 1991). Stage 3 refers to
swarms of circa 30 Ma old leucogranitic dykes.
Intra arc deformation features as in stage 1 plutons are attributed to the collision of the
Kohistan arc to the Karakoram terrain to the north (Coward et al., 1986; Petterson and
Windley, 1985; Treloar et al., 1996). Based on this hypothesis, suturing is inferred to have
occurred between the age of deformed and undeformed plutons, between 104 and 85 Ma
(Treloar et al., 1996). Stage 1 plutons would have intruded during the intra-oceanic stage
of the Kohistan arc and undeformed stage 2 plutons would have intruded after the
Kohistan-Karakoram collision, in an Andean type margin. Plutons that have intruded the
40
Utror volcanites have been attributed to stage 2 based on an Ar-39Ar hornblende age of
48 ± 1 Ma (Treloar et al., 1989), whereas dioritic plutons to the south of the Barul Banda
slates were attributed to stage 1 (Sullivan et al., 1993). Treloar et al. (1996) noted that the
 162

geochronological information contains a gap in magmatic activity within the Dir area.
Therefore, above authors proposed time-depending and iterative extensional and
compression within the Kohistan arc. Major magmatic hiatuses are due to compression
around 104-85 Ma and between 75-55 Ma. Our geochronological work aimed at testing
this hypothesis.

Figure 6.2: Detailed map of the Dir-Kalam are based on own field observations, interpretation of
multispectral Landsat 7 ETM + satellite images and published maps (Butt et al., 1980; Chaudhry et al.,
1987; Jan and Mian, 1971; Sullivan et al., 1993). Previously plutonic units south of the Barul-Banda
Slate Formation have been interpreted as stage 1, those in the north of the volcanites as stage 2
(Sullivan et al., 1993).

6.3 Analytical Method

Where possible 70-100, hand picked, long prismatic, idiomorphic zircons where mounted
on epoxy resin, polished and analysed with an inductively coupled plasma mass
spectrometer (ICP-MS) combined with a laser ablation system (LA) established at the
Tokyo Institute of Technology, Japan. We used a ThermoElectron VG PlasmaQuad 2
 163

quadripole based ICPMS equipped with the S-option interface (Hirata and Nesbitt, 1995)
coupled with a MicroLas production (Göttingen, Germany) GeoLas 200CQ laser ablation
system. This system utilizes Lambda Physik (Göttingen, Germany) COMPex 102 ArF
excimer laser as a 193 nm DUV (deep ultraviolet) light source. The repetition rate of the
laser system was 6 Hz, the beam diameter has been 16 µm and the source pulse energy
140mJ. NIST 610 SRM was used as standard material. As an additional internal standard a
sample from an aplitic dyke collected near the Kohistan-Karakoram suture zone (01B22)
and previously dated by conventional U-Pb zircon thermal ion mass spectrometry and
showed various inherited zircons with different ages (47 Ma, 275 Ma, 582 Ma und 782 Ma
S. Heuberger personal. com.), was used to evaluate the precision and accuracy of the
unknown samples. Zircons where examined by cathodoluminescence (CL) to check for
heterogeneity and core-rim relationships.
Analysing followed the standard lab measuring procedure: measuring five times the
gasblank (background) and five times the NIST 610 standard. Thereafter ten analyses on
zircons were done and the procedure was repeated from the beginning. The Laser was fired
five to six seconds before measurement in the mass-spectrometer was activated to “clean”
the analysed spot. Detailed lab procedure and discussion of the precision and accuracy of
the method used are published elsewhere (Izuka and Hirata, 2003).
Whole rock major element composition has been determined using a Rigaku RINT 2000
powder X-ray diffractometer conducted at the Tokyo Institute of Technology, Japan.

6.3.1 Sample description

C-01-77 (35°33´003´´N; 74°09´129´´E)
Foliated, coarse grained granite from southern limit of the Kohistan arc in the eastern part
of Kohistan, near the town of Chilas (Khiner valley). The tonalite has many volcanic
xenoliths and is crosscut by amphibolitic dykes.

MR-02-3 (35°33´205´´N; 72°33´451´´E)

Coarse grained granite intruding the Utror volcanics. Towards the contact abundant
xenoliths of volcanic material are found within the granite. In the contact zone apophyses
of granite into volcanics are common, indicating an intrusive contact.

Mn-02-04 (35°34´646´´N; 72°45´599´´E)

 164

Same unit as sample MR-02-3 sampled from a neighboring valley approximately 10 km

distance. Coarse grained granite with abundant xenoliths of volcanics indicating an
intrusive relationship with Utror volcanics..

BO-02-13 (35°22´579´´N; 72°03´257´´E)

Medium to coarse grained quartz diorite with xenoliths of volcano-sedimentary units.
Local shearing is observed. The quartz diorites are crosscut by tonalities.

RB-02-16 (35°33´712´´N; 72°12´666´´E)

Whitish alkali granite with brownish-pinkish microcline mega crystals and 15-20 % of
biotite and amphibole.

DR-02-18 (35°14´497´´N; 71°51´641´´E)

Coarse grained diorite, frequent apophysen into the volcanics indicate an intrusive
relationship with the volcanics.

BR-02-19 (35°07´687´´N; 72°56´229´´E)

Coarse grained granodiorite. Apophysen of granodioritic material into volcano-sediments
indicate an intrusive contact. The granodiorite itself is frequently intruded by pegmatite.

6.4 Results

6.4.1 Cathodoluminescence
In general zircons from all but two samples (01B22 and C-01-77) show homogeneous
growth patterns and partly thin overgrowths that revealed no age difference and are not
further discussed here (Figure 6.3 A). In contrast old zircons from sample 01B22 show
recrystallisation features whereas young zircons from the same sample show no evidence
of recrystallisation (Figure 6.3 C,D). The majority of zircons from sample C-01-77 show
evidence for (slight) recrystallisation (Figure 6.3); however, if an age difference exists
between the two populations, it is within error of the method.
 165

Figure 6.3: Zircon Cathodoluminescent images from different samples. Circular objects are the laser
ablation pits after analyses and the resulting age is given. A) recrystallized Zircon from sample 01B22;
there is no detectable age difference between different areas of the Zircon and the whole Zircon is
recrystallized; B) young Zirkon from the same sample showing a light core and a darker rim
displaying an oscillatory growing pattern. Apparently there is no detectable age difference between
core and rim, but notice that the shot of the LA-ICPMS

6.4.2 Zircon chronology

The results are given in Table 1. Whole rock compositions are illustrated in Figure 6.4 and
geochronological result are graphically presented in conventional concordia (Figure 6.5),
frequency plots (Figure 6.6) and in Figure 6.7. A general problem of dating young zircons
207 207
is the low concentration of Pb. The large uncertainties on Pb are mirrored by the
207
scatter of the data in conventional concordia plots parallel to the Pb/235U axes. Ages
207
calculated in this study involving Pb (207Pb/206Pb and 207
Pb/235U) are geologically
meaningless and give in general unrealistic numbers (sometimes “future” ages and ages
older than 4.5 Ga). They are therefore rejected. The U-P ages from zircons from sample
 166

01B22 determined by conventional TIMS are well reproduced. The youngest population of
zircons yields an age of 44.8 ± 3.4 Ma in accordance with the conventional U-Pb TIMS
age of 47.4 ± 0.5 Ma, which is interpreted as the intrusion age of the aplitic dyke. Older
206
ages of inherited, discordant zircons are less well reproduced since Pb/238U ages of
discordant zircons tend to be too young. Based on this result we are confident that young,
homogeneous zircons plotting on the Concordia reflect intrusion ages. The analytical
results on older inherited zircons are not entirely relevant to this work and are ignored. All
206
ages presented in this study are Pb/238U ages and the error is given as 2σmean. Zircons
204
from sample BO-02-13, which have a higher Pb content than background, have been
corrected by present day common lead mantle values (Zartman and Doe, 1981) since the
normal distribution within the frequency plot improved (Figure 6.6). However, the
correction does not affect the mean age. For all other samples common lead correction had
no effect (e.g. see RB-02-16 in Figure 6.6 ). Ages given in Table 4 are interpreted as
intrusion ages.
Table 4: Results of Major Element analyses and U-Pb LA-ICPMS zircon analysis from samples of the
central zone from the Kohistan arc error on ages is given as 2σ (mean)
Sample Nr 01B22 C-01-77 MR-02-03 MN-02-04 BO-02-13 RB-02-16 DR-02-18 BR-02-19
Granodiorit
Rocktyp Aplite Tonalite Granit Granite Diorite Adamellite Diorite
e

SiO2 - 75.7 73.3 70.7 70.8 69.5 56.9 69.1

TiO2 - 0.2 0.3 0.2 0.5 0.4 0.7 0.4

Al2O3 - 12.4 14.5 15.1 14.8 16.5 18.3 16.5

Fe2O3 - 1.8 2.3 3.3 4.8 2.3 7.8 3.7

MnO - 0 0.1 0.1 0.1 0 0.1 0.1

MgO - 0.4 0.2 0.5 0.9 0.7 3.0 0.9

CaO - 1.5 1.4 2.9 3.9 2.5 6.6 4.1

Na2O - 3.44 5.4 3.6 4.3 3.7 3.2 4.2

K2O - 3.77 2.7 3.2 0.7 4.9 2.7 1.2

P2O5 - 0.04 0.1 0.1 0.1 0.1 0.2 0.2

Total - 99.3 99.8 99.8 101.1 100.2 99.6 100.3

207
Pb/235U - 0.015035 0.01118 0.01129 0.00955 0.00886 0.00665 0.01165

44.8±3.
Age (in Ma) 67.6±1.2 71.7±2.1 72.4±1.1 61.3±1.5 56.9±1.5 42.7±0.5 74.7±1.7
4

Sample C-01-77, a tonalite from the southern margin of the Kohistan Batholith, gives an
age of 67.6 ± 1.2 Ma. The pinkish, roughly 200-300 µm large, partly recrystallised zircons
 167

show a slightly spreaded normal distribution within the frequency plot (Figure 6.6). Two
maxima may be present, one at 64 Ma and the other at 71 Ma, suggesting two magmatic
intrusion times. This would explain the frequent recrystallised grains. However, the two
maxima are not clearly resolved and vanish within error. All other dated samples show a
Gaussian normal distribution. Colorless 100-200 µm large zircons from the granite MN-
02-04 yield an age of 72.4 ± 1.1 Ma within error, of the same age as sample MR-02-03
(71.7 ± 2.1 Ma), which was collected from the same granite body (Figure 6.2),
approximately 10 km further west. Granite RB-02-16 yields an age of 56.9 ± 1.5 Ma,
which is consistent with remote sensing mapping that indicates it is intrusive in the
surrounding diorite BO-02-13 (Figure 6.2) dated at 61.3 ± 1.5 Ma. A quartz-diorite body
south of Dir (BR-02-19) yields an intrusion age of 74.7 ± 1.7 Ma. A diorite (Dr-02-18) has
the youngest intrusion age of 42.7 ± 0.5. All samples show sporadically inherited, older
zircons. The inherited zircons are generally younger than the inferred initiation of the
Kohistan arc magmatism (150 Ma Schaltegger et al., 2004) except for zircons inherited in
sample BR-02-19 which are as old as 249 Ma and sample DR-02-18 which has 407 Ma old
inherited zircons.

Figure 6.4: Total Alkali vs Silica (TAS) diagram for rock classification. Nomenclature of igneous rocks
after (Cox et al., 1979).
 168

Figure 6.5 Conventional concordia plot of the analysed zircons, ages are calculated from the 206Pb/235U
ratios and errors are given as 2σmean..
 169

Figure 6.6 Frequency histogram of the analysed Zircon. Black bars indicate non common lead
corrected values and white indicate corrected ones. The correction has no effect on the mean age and
has a slight positive effect on the normal distribution of sample BO-02-13 whereas the normal
distribution of sample Rb-02-16 is unaffected by the correction.

6.5 Discussion
New dating in the Dir area indicates continuous intrusive activity from ~75 Ma to at least
~56 Ma possibly until ~42 Ma (Figure 6.7) similar to results from the Ladakh Batholith
 170

(Weinberg and Dunlap, 2000). The presence of old (407 Ma) inherited zircons in sample
DR-02-18 likely indicates the presence of Indian crust beneath the Kohistan arc Complex
at ~42 Ma. The presence of up to 249 Ma old zircons of sample BR-02-19 is not well
understood but could steam from the older oceanic lithosphere.

Figure 6.7: Plot illustrating the continuous spectrum of the ages

Geographically, there is no significant age difference between intrusive units in the eastern
(C-01-77) and western parts (all other samples) of the southern Kohistan Batholith. The
age difference between presumed stage 1 (BR-02-19 (74.7 Ma)) and stage 2 (MR-02-03
and MN-02-03 (~72 Ma)) plutons (Sullivan et al., 1993), as defined from their deformation
fabric, is small and both are younger than the undeformed 85 Ma old (Schaltegger et al.,
2002; Zeitler, 1985) Chilas gabbronorite believed to post-date docking of the Kohistan Arc
against the Karakoram Asia (Treloar et al., 1996). The relative dating between stage 1 and
stage 2 plutons was based on the assumptions that (1) deformation affecting the northern
part of the arc is all related to Kohistan-Karakoram suturing and (2) that a clear age
difference between the deformed and undeformed plutons exists. Our geochronological
results do not support this interpretation. There is no age gap between foliated and non-
foliated intrusions and fabrics affecting rocks of the Kohistan Batholith cannot all be
evidence of widespread intra-arc deformation due to Kohistan-Karakoram suturing before
85 Ma (Coward, 1983; Coward et al., 1986; Petterson and Windley, 1985)). This
 171

assumption invoked the northward increase of strain intensity and the coplanarity of
flattening fabrics within the Kohistan with those in the suture (Treloar et al., 1996).
However, deformation and metamorphism are extremely low in the Yasin sediments,
against the suture zone (Petterson and Windley, 1985), and remote sensing on the Landsat
ETM+ images readily shows that planar fabrics strike oblique to the Kohistan-Karakoram
Suture in the NE Kohistan arc (Figure 6.8). The inferred strain gradient and accordance of
fabrics are therefore at odd with geological observation. In fact, in most regions of the
Kohistan Batholith that we have structurally studied, like in the Dir area, (i) strain is
heterogeneously distributed, ii) magmatic foliations within plutonic rocks tend to be in
continuity with main foliations in the country, often hornfelsed rocks, (iii) folds have
complex trends concordant with plutonic margins and (iv) lineations, including those
within anastomosing networks of ductile shear zones, are usually plunging steeply. We
speculate that deformation, as zircon recrystallisations, are essentially related to magmatic
emplacement of imbricate and nested intrusions during the build up of the island arc and
does not exclusively constrain the age of the overprinting Kohistan-Karakoram Suture.

Figure 6.8: Landsat 7 ETM + pictures (combination of bands 745) of the NE Kohistan. Indicating that
trend of structures are influenced by the trend of the Kohistan-Karakoram Suture zone only within a
few Kilometers of the Suture zone. Bright blue color represents snow or ice fields, white fields are
eliminated shadows.
 172

The ~72 Ma intrusion age of a granite (MN-02-04 & MR-02-03) intrusive into Utror
volcanites contradicts the assumed late Eocene age of these rocks. Sullivan et al. (1993),
39
referring to an Ar-40Ar hornblende age of 55 Ma (Treloar et al., 1989), interpreted the
Utror volcanites as late Palaeocene/Eocene lavas of a continental margin correlated with
the so-called Shamran volcanites found further in the north. However, Shah and Shervais
(1999) challenged this interpretation and noted that the majority of the Utror volcanites are
geochemically rather primitive. Trace and major element characteristics of the Chilas
gabbronorite (chapter 5) are very similar to those of certain Dir and Utror volcanites (group
1 of Shah and Shervais, 1999) (). Consistently, with the > 72 Ma age of these volcanites,
we therefore interpret them as the effusive equivalent of the 85 Ma old Chilas
gabbronorite. The Chilas Complex has been interpreted to result from splitting of the arc
during incipient back arc rifting (Khan et al., 1989). As a corollary, the Utror volcanites
have been emplaced within an intra-arc rift basin. Late Palaeocene fauna in the Baraul
Banda Slate Formation (Sullivan et al., 1993) indicate that accumulation within this basin
possibly took place until initial collision of Kohistan with the Indian Plate, around 55 Ma
(Rowley, 1996, 1998), which may be reflected in the 55 Ma 39Ar-40Ar hornblende cooling
age of an andesitic lava flow (Treloar et al., 1989).

6.6 Implication for the development of the Kohistan arc

The time span from initiation of island arc magmatism in the Kohistan arc until 85 Ma
rifting is ill constrained but available age data indicate that this periode is at least in the
range of 70-80 Ma lasting from 150-85 Ma. U-Pb zircon ages on the so-called Matum Das
pluton, North of Gilgit, indicate that magmatism started as early as 150 Ma years ago
(Schaltegger et al., 2004). Sm-Nd mineral and whole rock ages together with pressure
temperature condition recorded in the Jijal-Kalam complex indicates that significant crustal
thickening occurred within 30-40 Ma from 118 to 91 or 83 Ma: Two-pyroxene granulites
within the Jijal-Kalam complex reveal pressure-temperature estimates of 0.7-1.1 GPa and
700 to 900 °C (Yoshino et al., 1998) and yield an Sm-Nd age of 118 ± 12 Ma (Yamamoto
and Nakamura, 2000). Whereas garnet-clinopyroxene-plagioclase assemblages indicating
significantly higher pressure of up to 1.2 to 2.0 GPa (Jan and Howie, 1981) (Yamamoto,
1993) (Yamamoto and Yoshino, 1998) (Yoshino et al., 1998) (Ringuette et al., 1999) have
a younger Sm-Nd age of 91.0 ± 6.3 Ma (Yamamoto and Nakamura, 2000). Since both
assemblages have been dated within the same outcrop (Yamamoto and Nakamura, 2000)
 173

they represent different events and can not be attributed to difference in regional cooling.
Garnet-hornblendite, recording similar pressure of 1.2-1.4 GPa (Jan and Howie, 1981)
have within error the same Sm-Nd age of 83 ± 10 Ma (Yamamoto and Nakamura, 2000).
Pressure increase reflects arc thickening through magmatic accretion (Burg et al., 1998;
Ringuette et al., 1999); (Yamamoto and Nakamura, 2000); (Yoshino and Okudaira, 2004).
Based on the jadeite content in clinopyroxene (Ringuette et al., 1999) proposed a switch
from near isobaric cooling from ~750 °C, 1.5-1.8 GPa to near isothermal decompression at
~550 °C possibly down to 0.33GPa. An Ar-Ar hornblende age of 76 ± 4 Ma (Wartho et al.,
1996) indicates that regional cooling of the Jijal-Kalam complex below ~500°C and
occurred shortly after intrusion of the Chilas gabbronorite around 85 Ma. Since the
volcanite associated with the Chilas Complex are located north of the Complex, rifting
appears to be asymmetric along a south dipping normal fault. Magmatism, similar to the
development of a axial rift in the southern Havre Trough (Fujiwara et al., 2001), focused
along the thinned crust resulting in the voluminous intrusion of two-pyroxene gabbro of
the Chilas Complex at pressure around 0.6-0.8 GPa (Jan and Howie, 1980) and deposition
of sediments and volcanics within an intra-arc basin further to the north. At present, no
evidence indicates that the intra Kohistan rift evolved into an active spreading back-arc
ridge.
Asymmetric arc wide rifting along a south dipping shear zone resulted in significant
lateral grows and decompression of the lower part of the arc (Fig). Accordingly, the
Kohistan arc Complex did not grow continuously from a thin immature arc towards a thick
mature one, producing steadily higher pressure crystal cumulate sequences with time.
Extension and associated lateral grow produced lower pressure crystal cumulates even
within a mature arc. Since the Chilas Complex is the largest intrusion complex within the
Kohistan arc and one of the largest in the world, this mechanism significantly contributes
to island arc crust formation.

6.7 Conclusion
Structural mapping and geochronology in the SW Kohistan Batholith, between Dir and
Kalam, indicates that plutonism was continuous from at least 78 Ma to 56 Ma. Given that
pluton emplacement of foliated bodies postdates the presumed collision between Kohistan
and Karakoram, a large part of the deformation in the Kohistan Batholith has taken place
during arc magmatism. Intrusive relationships indicate that at least parts of the Utror
volcanites are older than 71 Ma. Their geochemical characteristics point to co-magmatism
 174

with the 85 Ma Chilas gabbronorite. We conclude that the Utror volcanites were deposited
into a late-Cretaceous intra-arc rift, in which sediments were deposited up to the late-
Paleocene.
 175

Figure 6.9 : A) Primitive mantle (Sun and McDonough, 1989) normalized trace element pattern of
Chilas gabbronorite (solid symbols) compared to the so-called group 1 volcanite from the Dir area
(Shah and Shervais, 1999). The REE pattern of both groups is sub-parallel. The Dir volcanites have
higher concentration and a negative Eu anomaly whereas the Chilas gabbronorite have lower
concentrations and a complementary positive anomaly. B) The Ce/Yb ratio is similar for both groups.
C) All Dir volcanite of Shah and Shervais (1999) compared to the in-situ crystallization model
presented in chapter 5. Some volcanite (circles) follow the trend defined by the Chilas gabbronorite
(big squares own data and small squares from Khan (1988)). Some volcanite follow the trend whereas
other clearly deviate from, indicating that not all volcanite in the Dir area are necessarily co-magmatic
with the Chilas gabbronorite.
 176

Reference

Bard, J. P., 1983, Metamorphism of an obducted island arc; example of the Kohistan
Sequence (Pakistan) in the Himalayan collided range: Earth and Planetary Science
Letters, v. 65, p. 133-144.
Bignold, S. M., and P. J. Treloar, 2003, Northward subduction of the Indian Plate beneath
the Kohistan island arc, Pakistan Himalaya; new evidence from isotopic data:
Journal of the Geological Society of London, v. 160, p. 377-384.
Burg, J. P., J. L. Bodinier, S. Chaudhry, S. Hussain, and H. Dawood, 1998, Infra-arc
mantle-crust transition and intra-arc mantle diapirs in the Kohistan Complex
(Pakistani Himalaya); petro-structural evidence: Terra Nova. The European
Journal of Geosciences, v. 10, p. 74-80.
Butt, K. A., M. N. Chaudhry, and M. Ashraf, 1980, An interpretation of petrotectonic
assemblage west of western Himalayan syntaxis in Dir District and adjoining
areas in northern Pakistan, in R. A. K. Tahirkheli, M. Q. Jan, and M. Majid, eds.,
Proceedings of the Internatioanl Committee on Geodynamics, Group 6 meeting.:
Geological Bulletin, University of Peshawar, v. 13: Peshawar, Pakistan,
University of Peshawar, Department of Geology, p. 79-86.
Chaudhry, M. N., M. S. Hussain, and N. Quamae, 1987, Geology and Petrography of
Barwal-Dir_Bibor area (Toposheet No. 38 M/16) Dir district, NWF Pakistan:
Geological Bulletin, University of Punjab, v. 22, p. 143-152.
Coward, M. P., 1983, Thrust tectonics, thin skinned or thick skinned, and the continuation
of thrusts to deep in the crust, in Anonymous, ed., Balanced cross-sections and
their geological significance; a memorial to David Elliott.: Journal of Structural
Geology, v. 5; 2: Oxford-New York, International, Pergamon, p. 113-123.
Coward, M. P., R. W. H. Butler, K. M. Asif, and R. J. Knipe, 1987, The tectonic history of
Kohistan and its implications for Himalayan structure: Journal of the Geological
Society of London, v. 144, p. 377-391.
Coward, M. P., B. F. Windley, R. D. Broughton, I. W. Luff, M. G. Petterson, C. J. Pudsey,
D. C. Rex, and K. M. Asif, 1986, Collision tectonics in the NW Himalayas, in M.
P. Coward, and C. Ries Alison, eds., Collision tectonics.: Geological Society
Special Publications, v. 19: London, United Kingdom, Geological Society of
London, p. 203-219.
 177

Cox, K. G., J. D. Bell, and R. J. Pankhurst, 1979, The interpretation of igneous rocks:
London, Allen and Unwin, 450 p.
Fujiwara, T., T. Yamazaki, and M. Joshima, 2001, Bathymetry and magnetic anomalies in
the Havre Trough and southern Lau Basin; from rifting to spreading in back-arc
basins: Earth and Planetary Science Letters, v. 185, p. 253-264.
George, M. T., N. B. W. Harris, and R. W. H. Butler, 1993, The tectonic implications of
contrasting granite magmatism between the Kohistan island arc and the Nanga
Parbat-Haramosh Massif, Pakistan Himalaya, in P. J. Treloar, and M. P. Searle,
eds., Himalayan tectonics.: Geological Society Special Publications, v. 74:
London, United Kingdom, Geological Society of London, p. 173-191.
Hirata, T., and R. W. Nesbitt, 1995, U-Pb isotope geochronology of zircon; evaluation of
the laser probe-inductively coupled plasma mass spectrometry technique:
Geochimica et Cosmochimica Acta, v. 59, p. 2491-2500.
Jan, M. Q., and R. A. Howie, 1980, Ortho- and clinopyroxenes from the pyroxene
granulites of Swat Kohistan, northern Pakistan: Mineralogical Magazine, v. 43, p.
715-726.
Jan, M. Q., and R. A. Howie, 1981, The mineralogy and geochemistry of the
metamorphosed basic and ultrabasic rocks of the Jijal Complex, Kohistan, NW
Pakistan: Journal of Petrology, v. 22, p. 85-126.
Jan, M. Q., and B. F. Windley, 1990, Chromian spinel-silicate chemistry in ultramafic
rocks of the Jijal Complex Northwest Pakistan: Journal of Petrology, v. 31, p.
667-715.
Jan, Q. M., and I. Mian, 1971, Preliminary geology and petrography of Swat Kohistan:
Geological Bulletin, University of Peshawar, v. 6, p. 1-32.
Kelemen, P. B., 1995, Genesis of high Mg andesites and the continental crust:
Contributions to Mineralogy and Petrology, v. 120, p. 1-19.
Khan, M. A., 1988, Petrology and structure of the Chilas ultramafic-mafic complex,
Kohistan Arc, NW Himalayas, University of London, London.
Khan, M. A., M. Q. Jan, and B. L. Weaver, 1993, Evolution of the lower arc crust in
Kohistan, N. Pakistan; temporal arc magmatism through early, mature and intra-
arc rift stages, in P. J. Treloar, and M. P. Searle, eds., Himalayan tectonics.:
Geological Society Special Publications, v. 74: London, United Kingdom,
Geological Society of London, p. 123-138.
 178

Khan, M. A., M. Q. Jan, B. F. Windley, J. Tarney, and M. F. Thirlwall, 1989, The Chilas
mafic-ultramafic igneous complex; the root of the Kohistan island arc in the
Himalaya of northern Pakistan, in L. Malinconico Lawrence, Jr., and J. Lillie
Robert, eds., Tectonics of the western Himalayas.: Special Paper - Geological
Society of America, v. 232: Boulder, CO, United States, Geological Society of
America (GSA), p. 75-94.
Khan, M. A., R. J. Stern, R. F. Gribble, and B. F. Windley, 1997, Geochemical and
isotopic constraints on subduction polarity, magma sources and palaeogeography
of the Kohistan Arc, northern Pakistan: Journal of the Geological Society of
London, v. 154 Part 6, p. 935-946.
Petterson, M. G., M. B. Crawford, and B. F. Windley, 1993, Petrogenetic implications of
neodymium isotope data from the Kohistan Batholith, North Pakistan: Journal of
the Geological Society of London, v. 150 Part 1, p. 125-129.
Petterson, M. G., and B. F. Windley, 1985, Rb-Sr dating of the Kohistan arc-batholith in
the Trans-Himalaya of North Pakistan, and tectonic implications: Earth and
Planetary Science Letters, v. 74, p. 45-57.
Petterson, M. G., and B. F. Windley, 1991, Changing source regions of magmas and
crustal growth in the Trans-Himalayas; evidence from the Chalt Volcanics and
Kohistan Batholith, Kohistan, northern Pakistan: Earth and Planetary Science
Letters, v. 102, p. 326-341.
Pudsey, C. J., M. P. Coward, I. W. Luff, R. M. Shackleton, B. F. Windley, and M. Q. Jan,
1985, Collision zone between the Kohistan Arc and the Asian Plate in NW
Pakistan: Transactions of the Royal Society of Edinburgh: Earth Sciences, v. 76,
p. 463-479.
Ringuette, L., J. Martignole, and B. F. Windley, 1999, Magmatic crystallization, isobaric
cooling, and decompression of the garnet-bearing assemblages of the Jijal
Sequence (Kohistan Terrane, western Himalayas): Geology (Boulder), v. 27, p.
139-142.
Rowley, D. B., 1996, Age of initiation of collision between India and Asia; a review of
stratigraphic data: Earth and Planetary Science Letters, v. 145, p. 1-13.
Rowley, D. B., 1998, Minimum age of initiation of collision between India and Asia north
of Everest based on the subsidence history of the Zhepure Mountain section:
Journal of Geology, v. 106, p. 229-235.
 179

Schaltegger, U., S. Heuberger, M. Frank, D. Fontignie, S. Sergeev, and J. P. Burg, 2004,

Crust-mantle interaction during Karakoram-Kohistan accretion (NW Pakistan).
Goldschmidt 2004.
Schaltegger, U., G. Zeilinger, M. Frank, and J. P. Burg, 2002, Multiple mantle sources
during island arc magmatism; U-Pb and Hf isotopic evidence from the Kohistan
arc complex, Pakistan: Terra Nova, v. 14, p. 461-468.
Searle, M. P., M. A. Khan, J. E. Fraser, S. J. Gough, and J. M. Qasim, 1999, The tectonic
evolution of the Kohistan-Karakoram collision belt along the Karakoram Highway
transect, North Pakistan: Tectonics, v. 18, p. 929-949.
Shah, M. T., and J. W. Shervais, 1999, The Dir-Utror metavolcanic sequence, Kohistan arc
terrane, northern Pakistan: Journal of Asian Earth Sciences, v. 17, p. 459-475.
Sullivan, M. A., B. F. Windley, A. D. Saunders, J. R. Haynes, and D. C. Rex, 1993, A
palaeogeographic reconstruction of the Dir Group; evidence for magmatic arc
migration within Kohistan, N. Pakistan, in P. J. Treloar, and M. P. Searle, eds.,
Himalayan tectonics.: Geological Society Special Publications, v. 74: London,
United Kingdom, Geological Society of London, p. 139-160.
Sun, S. S., and W. F. McDonough, 1989, Chemical and isotopic systematics of oceanic
basalts; implications for mantle composition and processes, in A. D. Saunders,
and M. J. Norry, eds., Magmatism in the ocean basins.: Geological Society
Special Publications, v. 42: London, United Kingdom, Geological Society of
London, p. 313-345.
Tahirkheli, R. A. K., M. Mattauer, F. Proust, and P. Tapponnier, 1979, The India-Eurasia
suture zone in northern Pakistan; synthesis and interpretation of recent data at
plate scale, in F. Abul, and K. A. DeJong, eds., Geodynamics of Pakistan.: Quetta,
Pakistan, Geol. Surv. Pakistan, p. 125-130.
Treloar, P. J., M. G. Petterson, M. Q. Jan, and M. A. Sullivan, 1996, A re-evaluation of the
stratigraphy and evolution of the Kohistan Arc sequence, Pakistan Himalaya;
implications for magmatic and tectonic arc-building processes: Journal of the
Geological Society of London, v. 153 Part 5, p. 681-693.
Treloar, P. J., D. C. Rex, P. G. Guise, M. P. Coward, M. P. Searle, B. F. Windley, M. G.
Petterson, M. Q. Jan, and I. W. Luff, 1989, K-Ar and Ar-Ar geochronology of the
Himalayan collision in NW Pakistan; constraints on the timing of suturing,
deformation, metamorphism and uplift: Tectonics, v. 8, p. 881-909.
 180

Wartho, J. A., D. C. Rex, and P. G. Guise, 1996, Excess argon in amphiboles linked to
greenschist facies alteration in the Kamila amphibolite belt, Kohistan island arc
system, northern Pakistan; insights from (super 40) Ar/ (super 39) Ar step-heating
and acid leaching experiments: Geological Magazine, v. 133, p. 595-609.
Weinberg, R. F., and W. J. Dunlap, 2000, Growth and deformation of the Ladakh
Batholith, Northwest Himalayas; implications for timing of continental collision
and origin of calc-alkaline batholiths: Journal of Geology, v. 108, p. 303-320.
Yamamoto, H., 1993, Contrasting metamorphic P-T-time paths of the Kohistan granulites
and tectonics of the western Himalayas: Journal of the Geological Society of
London, v. 150 Part 5, p. 843-856.
Yamamoto, H., and E. Nakamura, 2000, Timing of magmatic and metamorphic events in
the Jijal Complex of the Kohistan Arc deduced from Sm-Nd dating of mafic
granulites, in M. A. Khan, J. Treloar Peter, P. Searle Michael, and M. Q. Jan, eds.,
Tectonics of the Nanga Parbat syntaxis and the western Himalaya., Geological
Society of London. London, United Kingdom. 2000.
Yamamoto, H., and T. Yoshino, 1998, Superposition of replacements in the mafic
granulites of the Jijal Complex of the Kohistan Arc, northern Pakistan;
dehydration and rehydration within deep arc crust: Lithos, v. 43, p. 219-234.
Yoshino, T., and T. Okudaira, 2004, Crustal Growth by Magmatic Accretion Constrained
by Metamorphic P–T Paths and Thermal Models of the Kohistan Arc, NW
Himalayas: Journal of Petrology, v. in press.
Yoshino, T., H. Yamamoto, T. Okudaira, and M. Toriumi, 1998, Crustal thickening of the
lower crust of the Kohistan Arc (N. Pakistan) deduced from Al zoning in
clinopyroxene and plagioclase: Journal of Metamorphic Geology, v. 16, p. 729-
748.
Zartman, R. E., and B. R. Doe, 1981, Plumbotectonics; the model, in R. E. Zartman, and S.
R. Taylor, eds., Evolution of the upper mantle.: Tectonophysics, v. 75; 1-2:
Amsterdam, Netherlands, Elsevier, p. 135-162.
Zeitler, P. K., 1985, Cooling history of the NW Himalaya, Pakistan: Tectonics, v. 4, p.
127-151.
 181

7 Summary and Conclusions

7.1 Introduction
Integrated field, petrological and geochemical analyses were carried out to answer three
main questions: 1) What are the mutual relationships between the ultramafic units of the
Chilas Complex and the surrounding large mass of gabbronorite? 2) Which effects had this
large intrusion on the tectono-magmatic evolution of the Kohistan arc? 3) What is the time
relationship between accretionary / obduction events and deformed plutonites of the
Kohistan Batholith?
In the following section, I will summaries the results of this thesis and try to answer these
questions. I will propose a tectonic model for the evolution of the Kohistan arc and finally
bring this thesis to an end by addressing some open questions and propose future research.

7.2 Relationships between the ultramafic units and the

surrounding gabbronorite
The continuous trend in plagioclase compositions from the ultramafite to the gabbronorite
indicates that both lithologies stem from a common magma. Structures within the dunite
indicate a replacive nature of the dunitic core due to melt rock interaction. A centimeter- to
meter-scale reaction zone is developed between the ultramafite and the gabbronorite. The
Nd isotopic equilibrium between ultramafite and gabbronorite and the similar REE
concentration pattern of gabbronorite whole rock samples and melt calculated in
equilibrium with clinopyroxene separated from dunite samples, indicate that the replacive
dunite and the gabbronorite formed through a common magma.
Modeling of the REE composition of the parental melt of the Chilas Complex indicates
that the melt derived from a previously depleted upper mantle by relatively high-
percentage of partial melting (~20%). The absent of any garnet signature in the REE
composition indicate that melting occurred within the spinel stability field in the upper
mantel (≤ 1.5 GPa). The presence of magmatic amphibole and anorthite-rich plagioclase
indicate the presence of water in the system. Using the inferred constrain about the origin
 182

of the parental magma, hydrous melting experiments (Ulmer, 2001) can be used to
qualitatively approximate the parental composition as an olivine tholeiite (Figure 7.1).

Figure 7.1: Molecular normative projection from clinopyroxene onto the basalt tetrahedron of hydrous
fluid undersaturated melting experiments (Ulmer, 2001). The hatched area indicate the likely primary
melt composition of the Chilas Complex based on constrain of the melting depth. The primary melt
composition of the Chilas Complex is most likely an olivine tholeiite. However, the field slightly touches
the quartz normative compositions.

Decreasing pressure leads to an increase of the stability field of olivine relative to

pyroxene. Accordingly, liquids at high pressure initially in equilibrium with olivine and
pyroxene are saturated only in olivine at lower pressure and will consume pyroxene and
crystallize olivine (for a summary of experimental data see Kelemen, 1995). The olivine
tholeiite while ascended started reacting with a porphyroclastic lherzolite wallrock and
formed replacive dunite. The presence of dunite with high and low Mg#whole rock within the
ultramafite indicates that the melt evolved during ascend explaining also the presence of
 183

amphibole-bearing pyroxenite patches within the Kahnbari body. Accordingly, the dunite
acted as melt channels through which the entire Chilas gabbronorite sequence was fed. The
fact that the melt channels presumably feeding the magma parental to the surrounding
gabbronorite sequence are now exposed in the gabbronorite is puzzling. Kinematic criteria
indicate vertical upwards movement of the ultramafite in respect to the gabbronorite
sequence. Additionally, field relations within the Chilas Complex between the ultramafic
bodies and the surrounding gabbronorite sequence indicate intrusion of the ultramafic
rocks under sub-solidus conditions into partially consolidated gabbronorite mush, which in
turn has intruded and reacted with the ultramafic rocks. Melt parental to the gabbronorite
infiltrated and reacted with the dunite forming a tens-of-meter- to centimeter-scale reaction
zone resulting in refertilization of the dunite and zonation of the ultramafic complex.
A possible explanation for this apperant contradiction is that the melt fraction within the
melt channel was high enough to lower the depth-integrated density of the melt channels
and mobilize the ultramafite which intruded the semi-consolidated gabbronorite sequence.
However, this model implies that the melt channels are vertically continuous reaching into
the upper mantle and possibly down to the melting region. Accordingly, the dyke-like
shape of the ultramafite closely resembles the accurate structure of the melt channel in the
depth (Figure 7.2).
The melt channels provide a chemically isolated environment in which the melt stemming
from the melting region ascended through the shallow upper mantle fractionating olivine,
pyroxene amphibole and spinel without necessary reequilibrating with mantle assemblage.
This model can explain the missing high Mg# cumulates in expose Island arc sections
(Kelemen et al., 2004).
The zonation of the ultramafite bodies results from the intrusion of the mobilized
ultramafite into the semi-solidified gabbronorite sequence. Due to fractionation and
cooling the magma evolved to an olivine undersaturated composition. As soon as the
ultramafite intruded the gabbronorite sequence, trapped liquid present within the semi-
crystallized gabbronorite infiltrated and reacted with the ultramafite. A possible PT path
and the associated phase relationships are illustrated in Figure 7.3.
 184

Figure 7.2: Schematic illustration of the shape of the melt channels.

Figure 7.3: A not impossible P-T trajectory for the development of the Chilas. Phase reactions are
based on hydrous (2,8% wt%) equilibrium fractionation experiments using a picrobasalt from the
Adamello (Pulmer unpublished). Due to adiabatic decompression the melt fractionating olivine is
undersaturated in pyroxene. During this stage the melt will form the replacive dunite. Due to cooling
the melt will crystallize olivine, spinel and subsequently clinopyroxene and plagioclase. Finally, in this
case around 1000°C, the melt becomes olivine undersaturated. If melt at these conditions infiltrates
 185

and reacts with the uprising ultramafite it will consume olivine to produce pyroxene forming the
observed zonation

7.3 Effects of the intrusion of the Chilas Complex on the tectono-

magmatic evolution of the Kohistan arc
The intrusion of the calc-alkaline Chilas Complex (Khan et al., 1989) at about 85 Ma
(Schaltegger et al., 2002; Zeitler, 1985) during intra-arc rifting (Burg et al., 1998; Khan et
al., 1989) is temporally associated with isothermal decompression of the lower crustal
garnet granulites from ~1.5-1.8 GPa to 0.33 GPa (Ringuette et al., 1999; Yamamoto, 1993)
and subsequent cooling below ~500 °C (dated by 39Ar-40Ar hornblende data at around
~80 Ma Treloar et al., 1989; Wartho et al., 1996). Even so the exact pressures might be
questioned; the available data indicate significant decompression contemporaneous with
the intrusion of the Chilas Complex. Syn-magmatic normal faults within the Chilas
Complex and amphibolite facies normal faults described in the Southern (Kamila)
Amphibolites (Treloar et al., 1996) indicate an extensional setting during intrusion. Based
on intrusive relationships and U-Pb zircon ages, it has been shown in chapter 6 that the
Dir-Utror volcanites, originally interpreted as late Palaeocene/Eocene in age (Sullivan et
al., 1993), are actually older than ~72 Ma. Additionally, whole rock trace element
characteristics of some of the volcanites are strikingly similar to the Chilas gabbronorite
sequence. Accordingly, the Dir volcanites may represent effusive equivalents of the
gabbronorite sequence. The melting region of the melt parental to the Chilas Complex has
been interpreted as <1.5 GPa based on the absence of a geochemical garnet signature in the
melt. The thickness of the Kohistan arc prior to the intrusion of the Chilas Complex is
inferred as reaching pressures of 1.5-1.8 GPa at its basis (Ringuette et al., 1999;
Yamamoto, 1993). Therefore, prior to the intrusion of the Chilas Complex, the arc crust
must have been thinned and the underlying mantle uplifted. Decompression melting of a
previously metasomatised mantle is also consistent with the inferred ≈ 20% of partial
melting.
Those new arguments supports previous interpretations that the Chilas Complex intruded
during an intra-arc rifting stage associated with mantle diapirism (Burg et al., 1998; Khan
et al., 1989). Related the Dir-Utror volcanite are deposited within an intra-arc rift basin.
Based on structural arguments (chapter 3) this intra-arc rift has been interpreted as being
asymmetric with the Chilas Complex intruding along a south-dipping listric normal fault.
 186

Vertical extensional veins of late hornblende pegmatite suggest that the original orientation
of the Chilas Complex has essentially been preserved. The general N-dip attitude of the
Southern Amphibolites is attributed to antithetic rollover tilting, which took place when the
Chilas Complex was emplaced at about 85 Ma. This scenario is consistent with the fact
that the Chilas gabbronorite cuts the north-dipping Southern Amphibolites on its southern
boundary (Treloar et al., 1996).
39
The hypothesis provides an explanation for the regionally distributed Ar/40Ar
cooling/unroofing ages at 83-80 Ma, significantly older than the 60-50 Ma old India-Asia
collision (Treloar et al., 1996; Treloar et al., 1989). The reasons for the intra-arc extension
can only be speculated. An important linked question is the timing of the closure of the
Kohistan-Karakoram Suture zone, in the North. Accordingly, I will first comment on time
relationship between accretionary / obduction events and deformed plutonites of the
Kohistan Batholith, and afterwards I present a possible tectono-magmatic model for the
evolution of the Kohistan arc.

7.4 Time relationship between accretionary / obduction events

and deformed plutonites of the Kohistan Batholith
The initial collision of the Indian plate with the Kohistan arc is reasonably well constrained
based on a drop in drift velocity around 52 Ma documented in paleamagnetic data (Patriat
and Achache, 1984). The closure of the Karakoram-Kohistan Suture zone however, is still
debated. Intra arc deformation features as in stage 1 plutons are attributed to the collision
of the Kohistan arc with the Karakoram terrain to the north, i.e. the closure of the
Kohistan-Karakoram Suture zone (Coward et al., 1986; Petterson and Windley, 1985;
Treloar et al., 1996). Based on this hypothesis, suturing is inferred to have between 104
and 85 Ma (Treloar et al., 1996).
However, the new geochronological results presented in chapter 6 conflicts with this
interpretation since deformed plutons are younger than undeformed ones. Additionally, the
Yasin sediments, against the suture zone are only slightly deformed (Petterson and
Windley, 1985). Consequently, fabrics affecting rocks of the Kohistan Batholith cannot all
be evidence of widespread intra-arc deformation due to Kohistan-Karakoram suturing
before 85 Ma (Coward, 1983; Coward et al., 1986; Petterson and Windley, 1985). In this
thesis it has been speculated that deformation is essentially related to magmatic
 187

emplacement of imbricate and nested intrusions during the build up of the island arc and
does not exclusively constrain the age of the overprinting Kohistan-Karakoram Suture.
Therefore, the time of closure of the Kohistan-Karakoram suture is unconstrained.
However, several line of evidence summarized above indicate that the Chilas Complex
intruded during extension of the Kohistan arc. If collision of the Kohistan arc with the
Karakoram occurred before the intrusion of the Chilas Complex extension must have
occurred in a supposedly overall compressional system.
Contrasting if intrusion of the Chilas Complex predates the collision of the Kohistan Arc
with the Karakoram, extension occurred within a subduction zone setting. Extension in the
overriding plate in subduction zones setting is well documented, and finally may result in
back-arc basin formation (e.g. Karig, 1971). A number of possible models have been
suggested to explain extension (see Mantovani et al. (2001) for detailed review) including
among others: retreat of the subduction zone (“slab rollback” e.g. Molnar and Atwater,
1978), “corner flow” in the mantle induced by the subducted slab (e.g. Toksoz and Bird,
1977) or the so called “sea-anchor model” were extension is triggered by landward motion
of the overriding plate opposed by the resistance force of the underlying mantle (Scholz
and Campos, 1995). About the responsible mechanism in the case of the Kohistan arc can
only be speculated.
Based on the fact that extension is well documented within the overriding plate in the
following model it is assumed that the intrusion of the Chilas Complex predates the
collision of Kohistan with Asia. However, as I mentioned before the timing of collision
with Eurasia is essentially unconstrained.

7.5 Tectono-magmatic evolution of the Kohistan arc

A tectono-magmatic model illustrating the inferred evolution of the Kohistan arc is given
in Figure 7.4.
The Kohistan arc was initiated above an intra-oceanic subduction zone in the equatorial
area of the Tethys (Figure 7.4 A). The time of initiation of subduction is unclear but 150
Ma U-Pb zircon ages of Matum-Das tonalite within the Kohistan Batholith (Schaltegger et
al., 2004) indicate that arc magmatism existed in late Jurassic/early Cretaceous times. High
pressure (1.2-2.0 GPa Ringuette et al., 1999; Yamamoto, 1993) granulite-facies
assemblages within the Southern Metaplutonic and Jijal Complexes yield Cretaceous Sm-
Nd ages (118 ± 12, 95 ± 2.7, 91.0 ± 6.3, 83 ± 10 Ma (Anczkiewicz and Vance, 2000;
Yamamoto and Nakamura, 2000). Since pressure increase is attributed to reflect crustal
 188

thickening (Burg et al., 1998; Ringuette et al., 1999; Yamamoto and Nakamura, 2000;
Yoshino and Okudaira, 2004) continuous subduction and associated arc magmatism
formed a ca. 40-60 (?) km thick crust within less than 40 Ma during a build up stage of the
arc (Figure 7.4 B).
Around 85 Ma the arc extended along a south dipping normal shear zone (Figure 7.4 C).
The N-dip attitude of the Southern Amphibolites resulted from antithetic rollover tilting of
the hanging wall during normal faulting. Associated mantel diapirism resulted in high-
percentage of partial melting (~20%) of a previously depleted, metasomatized mantle. As a
corollary, the Dir and Utror volcanites have been emplaced within an intra-arc rift basin.
Late Palaeocene fauna in the Baraul Banda Slate Formation (Sullivan et al., 1993) indicate
that accumulation within this basin possibly took place until initial collision of Kohistan
with the Indian Plate, around 55 Ma (Rowley, 1996, 1998) (Figure 7.4 D). Subduction of
the Indian plate following the collision resulted in ultra high pressure (UHP)
metamorphism of Indian plate gneisses (O'Brien et al., 2001) and partial melting of Indian
plate lithologies can be responsible for the ~45 Ma old magmatism. Late Eocene
extensional shearing resulted in extrusion of the UHP assemblage (Treloar et al., 2003).
Extension continued along the Indus suture zone (Burg et al., 1996; Vince and Treloar,
1995) (Figure 7.4 E).

7.6 Remaining open questions and outlook for future work

Open questions within the Kohistan arc are related to the origin, nature and timing of the
Kohistan-Karakoram suture zone. Additionally, linked with the question of the closure of
the Kohistan-Karakoram suture zone is the nature and origin of the so-called Kohistan
Batholith. The origin of the deformation in intrusive plutons is still poorly understood.
Detailed geological maps of the Kohistan arc sequence and profound understanding of the
tectono-metamorphic evolution is essential to be finally able to approximate the bulk
composition of the arc.
Zoned ultramafic complexes similar to those described in this thesis are described world
wide. Therefore focused flow in isolated melt channels might be a common crust forming
process. However, the observed ultramafite in the Chilas Complex are only the surface
expression of such interpreted melt extraction systems. If those “melt channels” are
vertically continuous, evidence should be found in exposed sub-arc mantle sections
 189

Figure 7.4: Illustration of a model for the tectono-magmatic evolution of the Kohistan arc during
northward directed subduction of the Tethyan ocean and subsequent collision with Eurasia and India.
From the present data the timing of collision of Kohistan with Eurasia is unconstrained. Accordingly,
two possible scenarios have been drawn in C1 and C2 depending whether the collision occurred before
or after the intrusion of the Chilas Complex.
 190

References:

Anczkiewicz, R., and D. Vance, 2000, Isotopic constraints on the evolution of

metamorphic conditions in the Jijal-Patan Complex and Kamila Belt of the
Kohistan Arc, Pakistan Himalaya, in M. A. Khan, J. Treloar Peter, P. Searle
Michael, and M. Q. Jan, eds., Tectonics of the Nanga Parbat syntaxis and the
western Himalaya., Geological Society of London. London, United Kingdom.
2000.
Burg, J. P., J. L. Bodinier, S. Chaudhry, S. Hussain, and H. Dawood, 1998, Infra-arc
mantle-crust transition and intra-arc mantle diapirs in the Kohistan Complex
(Pakistani Himalaya); petro-structural evidence: Terra Nova. The European
Journal of Geosciences, v. 10, p. 74-80.
Burg, J. P., M. N. Chaudhry, M. Ghazanfar, R. Anczkiewicz, and D. Spencer, 1996,
Structural evidence for back sliding of the Kohistan Arc in the collisional system
of Northwest Pakistan: Geology (Boulder), v. 24, p. 739-742.
Coward, M. P., 1983, Thrust tectonics, thin skinned or thick skinned, and the continuation
of thrusts to deep in the crust, in Anonymous, ed., Balanced cross-sections and
their geological significance; a memorial to David Elliott.: Journal of Structural
Geology, v. 5; 2: Oxford-New York, International, Pergamon, p. 113-123.
Coward, M. P., B. F. Windley, R. D. Broughton, I. W. Luff, M. G. Petterson, C. J. Pudsey,
D. C. Rex, and K. M. Asif, 1986, Collision tectonics in the NW Himalayas, in M.
P. Coward, and C. Ries Alison, eds., Collision tectonics.: Geological Society
Special Publications, v. 19: London, United Kingdom, Geological Society of
London, p. 203-219.
Karig, D. E., 1971, Origin and development of marginal basins in the western Pacific:
Journal of Geophysical Research, v. 76, p. 2542-2561.
Kelemen, P., K. Hanghoj, and A. Greene, 2004, One view of the geochemistry of
subduction-related magmatic arcs, with an emphasis on primitive andesite and
lower crust., in R. L. Rudnick, ed., The Crust: Treatise on Geochemistry, v. 3:
Oxford, Elsevier-Pergamon, p. 593-659.
Kelemen, P. B., 1995, Genesis of high Mg andesites and the continental crust:
Contributions to Mineralogy and Petrology, v. 120, p. 1-19.
 191

Khan, M. A., M. Q. Jan, B. F. Windley, J. Tarney, and M. F. Thirlwall, 1989, The Chilas
mafic-ultramafic igneous complex; the root of the Kohistan island arc in the
Himalaya of northern Pakistan, in L. Malinconico Lawrence, Jr., and J. Lillie
Robert, eds., Tectonics of the western Himalayas.: Special Paper - Geological
Society of America, v. 232: Boulder, CO, United States, Geological Society of
America (GSA), p. 75-94.
Mantovani, E., M. Viti, D. Babbucci, C. Tamburelli, and D. Albarello, 2001, Back arc
extension; which driving mechanism?: Journal of the Virtual Explorer (online), v.
3, p. 17-45.
Molnar, P., and T. Atwater, 1978, Interarc spreading and Cordilleran tectonics as alternates
related to the age of subducted oceanic lithosphere: Earth and Planetary Science
Letters, v. 41, p. 330-340.
O'Brien, P. J., N. Zotov, R. Law, M. A. Khan, and M. Q. Jan, 2001, Coesite in Himalayan
eclogite and implications for models of India-Asia collision: Geology (Boulder),
v. 29, p. 435-438.
Patriat, P., and J. Achache, 1984, India-Eurasia collision chronology has implications for
crustal shortening and driving mechanism of plates: Nature (London), v. 311, p.
615-621.
Petterson, M. G., and B. F. Windley, 1985, Rb-Sr dating of the Kohistan arc-batholith in
the Trans-Himalaya of North Pakistan, and tectonic implications: Earth and
Planetary Science Letters, v. 74, p. 45-57.
Ringuette, L., J. Martignole, and B. F. Windley, 1999, Magmatic crystallization, isobaric
cooling, and decompression of the garnet-bearing assemblages of the Jijal
Sequence (Kohistan Terrane, western Himalayas): Geology (Boulder), v. 27, p.
139-142.
Rowley, D. B., 1996, Age of initiation of collision between India and Asia; a review of
stratigraphic data: Earth and Planetary Science Letters, v. 145, p. 1-13.
Rowley, D. B., 1998, Minimum age of initiation of collision between India and Asia north
of Everest based on the subsidence history of the Zhepure Mountain section:
Journal of Geology, v. 106, p. 229-235.
Schaltegger, U., S. Heuberger, M. Frank, D. Fontignie, S. Sergeev, and J. P. Burg, 2004,
Crust-mantle interaction during Karakoram-Kohistan accretion (NW Pakistan).
Goldschmidt 2004.
 192

Schaltegger, U., G. Zeilinger, M. Frank, and J. P. Burg, 2002, Multiple mantle sources
during island arc magmatism; U-Pb and Hf isotopic evidence from the Kohistan
arc complex, Pakistan: Terra Nova, v. 14, p. 461-468.
Scholz, C. H., and J. Campos, 1995, On the mechanism of seismic decoupling and back arc
spreading at subduction zones: Journal of Geophysical Research, B, Solid Earth
and Planets, v. 100, p. 22,103-22,115.
Sullivan, M. A., B. F. Windley, A. D. Saunders, J. R. Haynes, and D. C. Rex, 1993, A
palaeogeographic reconstruction of the Dir Group; evidence for magmatic arc
migration within Kohistan, N. Pakistan, in P. J. Treloar, and M. P. Searle, eds.,
Himalayan tectonics.: Geological Society Special Publications, v. 74: London,
United Kingdom, Geological Society of London, p. 139-160.
Toksoz, M. N., and P. Bird, 1977, Formation and evolution of marginal basins and
continental plateaus, in M. Talwani, and W. C. Pitman, III, eds., Island arcs, deep
sea trenches and back-arc basins.: Maurice Ewing Series, v. 1: Washington, DC,
United States, American Geophysical Union, p. 379-393.
Treloar, P. J., B. P. J. O, R. R. Parrish, and M. A. Khan, 2003, Exhumation of early
Tertiary, coesite-bearing eclogites from the Pakistan Himalaya: Journal of the
Geological Society of London, v. 160, p. 367-376.
Treloar, P. J., M. G. Petterson, M. Q. Jan, and M. A. Sullivan, 1996, A re-evaluation of the
stratigraphy and evolution of the Kohistan Arc sequence, Pakistan Himalaya;
implications for magmatic and tectonic arc-building processes: Journal of the
Geological Society of London, v. 153 Part 5, p. 681-693.
Treloar, P. J., D. C. Rex, P. G. Guise, M. P. Coward, M. P. Searle, B. F. Windley, M. G.
Petterson, M. Q. Jan, and I. W. Luff, 1989, K-Ar and Ar-Ar geochronology of the
Himalayan collision in NW Pakistan; constraints on the timing of suturing,
deformation, metamorphism and uplift: Tectonics, v. 8, p. 881-909.
Ulmer, P., 2001, Partial melting in the mantle wedge; the role of H (sub 2) O in the genesis
of mantle-derived "arc-related" magmas, in C. Rubie David, and D. van der Hilst
Rob, eds., Processes and consequences of deep subduction., Elsevier. Amsterdam,
Netherlands. 2001.
Vince, K. J., and P. J. Treloar, 1995, Miocene, north vergent extensional displacements
along the Main Mantle Thrust, NW Himalaya, Pakistan: Journal of the Geological
Society, v. 153, p. 677-680.
 193

Wartho, J. A., D. C. Rex, and P. G. Guise, 1996, Excess argon in amphiboles linked to
greenschist facies alteration in the Kamila amphibolite belt, Kohistan island arc
system, northern Pakistan; insights from (super 40) Ar/ (super 39) Ar step-heating
and acid leaching experiments: Geological Magazine, v. 133, p. 595-609.
Yamamoto, H., 1993, Contrasting metamorphic P-T-time paths of the Kohistan granulites
and tectonics of the western Himalayas: Journal of the Geological Society of
London, v. 150 Part 5, p. 843-856.
Yamamoto, H., and E. Nakamura, 2000, Timing of magmatic and metamorphic events in
the Jijal Complex of the Kohistan Arc deduced from Sm-Nd dating of mafic
granulites, in M. A. Khan, J. Treloar Peter, P. Searle Michael, and M. Q. Jan, eds.,
Tectonics of the Nanga Parbat syntaxis and the western Himalaya., Geological
Society of London. London, United Kingdom. 2000.
Yoshino, T., and T. Okudaira, 2004, Crustal Growth by Magmatic Accretion Constrained
by Metamorphic P–T Paths and Thermal Models of the Kohistan Arc, NW
Himalayas: Journal of Petrology, v. in press.
Zeitler, P. K., 1985, Cooling history of the NW Himalaya, Pakistan: Tectonics, v. 4, p.
127-151.
 194

8 Appendix IV

Mineral major and trace element composition of samples from the Chilas Complex

Table 1: Major element composition of Olivine

Table 2: Major element composition of Clinopyroxene

Table 3: Major element composition of Orthopyroxene

Table 4: Major element composition of Amphibole

Table 5: Major element composition Plagioclase

Table 6: Major element composition Spinel

Table 7: Trace element composition of Olivine

Table 8: Trace element composition of Clinopyroxene

Table 9: Trace element composition of Orthopyroxene

Table 10: Trace element composition of Amphibole

Table 11: Trace element composition of Plagioclase

 195

Table 5: Averaged Olivine composition from samples of the Chilas Complex.

 196

Table 1: continued.
 197

Table 2: Averaged Clinopyroxene composition from samples of the Chilas Complex.

 198

Table 2: continued.
 199

Table 3: Averaged Orthopyroxene composition from samples of the Chilas Complex.

 200

Table 3: continued.
 201

Table 4: Averaged Amphibole composition from samples of the Chilas Complex.

 202

Table 4: continued.
 203

Table 5: Averaged plagioclase composition from samples of the Chilas Complex.

 204

Table 5: continued
 205

Table 6: Averaged trace element composition of Clinopyroxene

 206

Table 7: Averaged trace element composition of Orthopyroxene

* measured by TIMS
 207

Table 8: Averaged trace element composition of Amphibole

 208

Table 9: Averaged trace element composition of Plagioclase

 209

List of Figures

Figure 1.1: False color Landsat ETM+ Picture of NE Pakistan. The box indicates the
extents of the geological map in . 13
Figure 1.2: Mid-Cretaceous (100-80 Ma) reconstruction of palaeopositions of the Kohistan
(Koh) and Ladakh (Lad) arc complexes, the Indian, African and E-Gondwana
continental plates (Zaman and Torii, 1999) 14
Figure 1.3: Index-map of the Western Himalaya, with the major tectono-stratigraphic
blocks and sutures. Hatched area = Karakoram block (after Gaetani, 1997) 16
Figure 1.4: Geology of Kohistan based on interpretation of Landsat ETM + pictures. 17
Figure 2.5: Geological map of the Chilas area based on own field observation,
interpretation of enhanced multi-spectral Landsat 7 ETM + pictures and maps
published by the Geological survey of Pakistan (Khan and Khan, 1998; Khan et
al., 1999a; Khan et al., 1999b) 34
Figure 2.6: Field relationship between tonalite and intrusive fine grained amphibolite dykes
(left). Intrusion of the tonalite into the gabbro is indicated by diffuse tonalitic
veins between larger gabbroic blocks (right). 36
Figure 2.7: Photomicrograph of metasediments. The picture on the right displays garnet-
bearing calc-silicate layers intercalated between biotite schist and quartz-
feldspar layers. The picture on the left shows biotite schist with numerous
leucocratic patches indicating partial melting. 38
Figure 2.8: Change of modal abundance of orthopyroxene defining the transition between
gabbronorite and gabbro. The modal amount of Orthopyroxene decreases with
increasing distance from an ultramafite outcrop Sample profile taken along
KKH (35°31`50.4``/73°42`33.4``) west of the Basheri body. 40
Figure 2.9: Photomicrographs of the field relation within the gabbronorite. A) normal
faulting within layered sequences developed in the lower layers whereas it is
absent in the uppermost one indicating a syn-depositional faulting. b) and c)
Truncation of layers by younger ones is frequently observed, and evidence for
truncation is also found within the non-layered part of the gabbronorite
sequence 40
Figure 2.10: Photomicrograph displaying the spectacular skeletal amphiboles of the
hornblende pegmatite dykes. 43
 210

Figure 2.11: Detailed geological map of the Basha peridotite and a small hornblendite body
opposite of the Darel Body (not described in the text). The detail sample profile
demonstrating the transitional transition between gabbronorite and gabbro
(Figure 2.8) was taken 100 m west of the Basha body. 45
Figure 2.12: Detailed geological map of the Thurli Body. The contact in the SW
termination is folded with the fold-axes plane parallel to the regional foliation
and the fold-axes parallel to the mineral lineation. 47
Figure 2.13: Detail geological map of the Giche body (in the NW corner) and the two
bodies outcropping in the Butho valley. Landsat 7 ETM + pictures indicate that
the larger eastern body might be linked to the body exposed in the Thak valley.
48
Figure 2.14: Detailed geological map of the Thak Body. The foliation in the vicinity of the
contact is generally orthogonal to the contact, which is strongly folded and
inter-fingered. Limited foliation measurements in the Thalpan area and south of
the body indicates encircling of the body by the foliation within a distance of
few hundred meters. Black star indicates location of detailed map (Figure 2.15)
and red star the location of carbonate bearing dykes. 49
Figure 2.15: Detail sketch-map of an outcrop of the Thak body near the Indus river
(35°24`34,28``/74°8`35,66``) illustrating the interfingering relationship at
various scale between the ultramafite and the surrounding gabbronorite. 50
Figure 2.16: Photographs of the field relationships between dunite and porphyroclastic
lherzolite. a) Tens-of-centimeter wide dykes often display a sharp contact with
surrounding porphyroclastic lherzolite. b) Smaller scale dykes however, often
have irregular contacts with frequent embayment of the dunite into the
porphyroclastic lherzolite. Branches of the dunitic dykes surround larger clasts.
c) Branches of dunitic dykes can die off within short distance. d) The dunitic
dykes become frequent and only relictual porphyroclastic lherzolite xenoliths
float within. 54
Figure 2.17: Photographs of the field relationships of massive pyroxenite within the
peridotite outcropping in the Khanbari Valley. a) Subhorizontal branches of
pyroxenite veins branches of from a massive pegmatitic pyroxenite dyke. b)
Lineout of picture (a). c) Patches of pyroxenite within peridotite (d) pyroxenite
dyke crosscutting peridotite. 55
 211

Figure 2.18: Photographs of the field relationships of gradual transformation of dunite into
lherzolite. a) The dunitic core is homogeneous with patches of Cr-spinel. b-e)
The structures of the patchy clinopyroxene (b) are equal to the structures of the
two-pyroxene–spinel symplectite (c-e) and are accordingly illustrated using the
later due to better contrasts. Picture d and f are rotate 90° counter clockwise for
layout reasons. See text for discussion 56
Figure 2.19: Photographs of the field relationships of gradual transformation of dunite into
lherzolite. Pictures c, d and i are rotate 90° counter clockwise for layout
reasons. a-d) Photographs of gabbronoritc veins intruding into the ultramafite
resulting in disintegration of the structure of the ultramafite. e) Reactive zones
around gabbronorite patches transforming dunite into lherzolite. g-h) Reactive
zones with preserved dunitic fragments. The pyroxene content can be very
irregularly developed within the lherzolite. h-i) Shape of gabbronorite xenoliths
within lherzolite. j) Contract between gabbronorite and lherzolite. See text for
discussion 58
Figure 2.20: Photographs of the field relationships at the contact-zone between
gabbronorite sequence and ultramafite. The orientation of the pictures is
indicated; in map view the arrow points towards north. a-c) Irregular contact
between gabbronorite and ultramafite with interfingering between the both. d)
“Lava-lamb” like structure of a ultramafite drop within a gabbronorite xenolith.
The next drop was about to be formed. e) Disintegrating ultramafic xenolith
within the gabbronorite sequence. 58
Figure 2.21: Photograph of the deformation fabric observed at the contact between
hornblendite and gabbronorite sequence. left) brittle shearbands indicating
upwards movement of the gabbronorite in respect to the hornblendite. right)
Boudinage of orthopyroxene rich dyke in the vicinity of the contact with the
Gine body. The boudin necks are partly filled with pegmatitic plagioclase and
hornblende minerals indicating high temperature deformation. 59
Figure 2.22: Photograph illustrating details of picture Figure 2.19 e. An ultramafic xenolith
is assimilated and disintegrated. The smaller xenoliths form a vertically
oriented tail above the “mother-xenolith”. See text for detailed discussion. 61
Figure 2.23: Sketch of the upwards bending and shearing of layered olivine-bearing
gabbronorite at the contact with ultramafite (35°19`20.9``/74°07`42``).
Instabilities developed at the interfaces between two layers are sub-vertical in
 212

the lower sub-horizontal part of the layers, whereas in the sub-vertical part the
instabilities are dragged downwards and layers are truncated by ultramafite,
indicating upwards shearing of the ultramafite in respect to the gabbronorite. 62
Figure 2.24: Schematic illustration of contact relationship between gabbronorite and
ultramafite. Indicated are also the locations of key rock samples, which will be
frequently mentioned in subsequent chapters. 66
Figure 3.1: Sketch map of the Kohistan Arc, mostly based on own mapping and
interpretation of multispectral Landsat7 ETM+ satellite images. The Chilas
Complex is restricted to the gabbro-norites, a lithological simplification that
highlights the backbone location between the North Kohistan Region and the
Southern Amphibolites. Bounding diorites and tonalites traditionally ascribed to
the Chilas Complex are being mapped. 72
Figure 3.2: Granular texture of the Chilas gabbro-norite. Thin section: left, transmitted
light; right, polarized light. Undeformed magmatic grains (note twins in
feldspars, arrow, and the absence of crystal-plasticity deformation features such
as undulose extinction and subgrains) have a preferred orientation (Sm)
defining the magmatic foliation. pyx= pyroxene; fsp = feldspar. GPS:
N35°26’09”; E073°49’36.1”. 74
Figure 3.3: a) Normal faulting within layered sequences developed in the lower layers
whereas it is absent in the uppermost one indicating a syn-depositional origin. b
& c) Truncation within layered gabbronorite are easily recognized. d)
Truncation in homogeneous Gabbronorite are difficult to recognize in the field.
The contact (arrow) between two subsequent intrusion is marked by xenolith
(X) cut of by the younger intrusion. The two intrusion have different orientation
of the magmatic fabric (Sm). e) Field map of the southern part of the Khiner
valley close to the confluence with the Indus River. Regional scale truncation is
indicated by the trend of the foliation. The strike changes within meters from
~E-W orientation towards N-S. The transition is within an otherwise
homogeneous gabbronorite without lithological contrast, shear zone or fault
identified in between in the field. 75
Figure 3.4: a) Boudinaged compositional layering. Arrow: primary unconformity (GPS:
N35°24’20.4”; E074°01’10.4”) . b) Parallel xenoliths (x) lying in the magmatic
foliation. Arrows: feldspar vein along a normal fault indicating brittle behaviour
of xenoliths and overall extension at a syn-magmatic, pre-full crystallisation
 213

stage (GPS: N35°32’25.2”; E073°31’27.1”). c) Solid state deformation of the

Chilas gabbronorite (GPS: N35°29’38.8”; E073°52’52.5”). d) Imbricate
hornblendite dykes with several tens-cm long mineral fibres and within-dyke
zoning indicating subhorizontal opening (GPS: N35°00’21.5”; E074°00’07.6”).
Such dyke swarms suggests formation in a zone of continuous but episodic
rifting. 76
Figure 3.5: Orientation diagrams of poles to magmatic foliations of the Chilas Complex.
Contour intervals at 2, 4, 6, 8 and 10 per unit area. Average pole for the Chilas-
Kandiah region: Plunge = 32 towards 014. 77
Figure 3.6: Mineral/magmatic lineation data from the Chilas Complex. Same contour
intervals as . Average pole orientation 71 towards 225. 78
Figure 3.7: Landscape (left) and outcrop (right) views of subvertical pegmatite veins.
White arrow: thin tip of one of these extensional veins, interpreted as standing
in its near-original attitude and that cut older magmatic joints (black arrow).
Outcrop at, and view from GPS: N35°30’45.3”; E073°23’30.1”. 80
Figure 3.8: Orientation diagrams of poles to veins within the Chilas Complex. a) Original
orientation as measured in the field. b) The measured poles rotated 30° around
E-W horizontal axis to reconstruct the supposed original orientation before
tilting of the arc. Due to the different strike direction the sub-vertical orientation
can not be maintained with a single rotation axis. If rotation occurred after the
formation of the veins, the veins would have had a rather illogical arbitrary
orientation before rotation and are than rotated into a symmetric vertical
orientation by chance. Alternatively, each set of dyke orientation were rotated
around a different axis to maintain the sub vertical orientations. 81
Figure 3.9: Tectonic interpretation of the Kohistan Arc Complex in late Cretaceous times,
modified from Burg et al. (1998). 82
Figure 4.1: Photomicrographs of texture within a porphyroclastic lherzolite (Sample C03-
45) with mineral abbreviations (Kretz, 1983) a) The embayment of amphibole
into the orthopyroxene and the embayment of olivine into amphibole and
orthopyroxene indicates that amphibole replaces orthopyroxene and
subsequently both are replaced by olivine. b) Trails of small unstrained olivine
neoblast crosscutting larger olivine grains with undulose extinction. c) Trails of
neoblastic olivine grains along grain boundaries between porphyroclastic
orthopyroxene and clinopyroxene d) Orientation of exsolution trails of spinel in
 214

orthopyroxene are preserved in olivine grains embaying the corresponding

orthopyroxene indicating a replacive origin for the olivine. 90
Figure 4.2: Photomicrographs of texture within the dunite a) Trails of smaller olivine
neoblasts crosscutting a larger olivine grain (C35). b) Cr-rich spinels along
grain boundaries of granoblastic olivine grains (C174). 91
Figure 4.3: Photomicrographs of textures for the gradual dunite-lherzolite-pyroxenite
transition a) vein like shaped trails of pyroxene grains (in this case opx) along
grain boundaries of granoblastic olivine. b) With increasing pyroxene content
prominent veins are develop. c) With further increasing pyroxene content larger
porphyroblastic pyroxene can be present. d) Within pyroxenite close to the
actual contact with surrounding gabbronorite olivine grains (high birefringence
colors) are only relicts between large pyroxene crystals. e) Two-pyroxene-
spinel symplectite mineral reaction developed between plagioclase and olivine.
f) Same symplectitic reaction without the presence of plagioclase, which was
completely transformed. g) Large (centimeter-scale) poikilitic orthopyroxene
with olivine inclusion and h) with an hypidiomophic plagioclase inclusion 92
Figure 4.4: Photomicrographs of textures in the gabbronorite a) actual contact between
ultramafite, here mainly orthopyroxene on the left and gabbronorite on the
right. b) Texture within gabbronorite composed of plagioclase (white),
orthopyroxene (reddish) and clinopyroxene (greenish). Pyroxenes are
frequently rimed by amphibole (green). 93
Figure 4.5: Representative compositional profiles through the main rock forming minerals
present in the Chilas Complex. A) Profile through different neighboring
plagioclase grains in gabbronorite and in plagioclase-bearing lherzolite.
Plagioclase from gabbronorite (C41) is relatively homogeneous and partly a
slight positive zonation with An poor rims is developed. Profound reverse
zoning however, is developed in plagioclase from ultramafite samples (C194).
B) Olivine from ultramafite (C33) is generally homogeneous. C) Cores of
clinopyroxene (C169) are relatively homogeneous whereas the rim displays
slight zoning. The Al2O3 content can scatter significantly but systematic
compositional zoning is present only in the outermost rim. D) Profile through
poikilitic orthopyroxene (C194) with a homogeneous composition but with
Al2O3 slightly decreasing against clinopyroxene inclusions. 94
 215

Figure 4.6: Plagioclase composition of tonalite, gabbronorite, olivine-bearing gabbronorite

and plagioclase-bearing lherzolite compared with results from (Khan, 1988).
The apparent gap invoked by Khan et al. (1989) in plagioclase composition
between UMA (red bar) and gabbronorite (brown bar) between An83 and An64
is not verified. 95
Figure 4.7: Primitive Mantle (Sun and McDonough, 1989) normalized average trace
element composition of plagioclase from gabbronorite (C7 & C48), olivine-
bearing gabbronorite (C66) and a plagioclase-bearing lherzolite (C218). The
grey line indicates the detection limit of LA-ICP-MS. Errors are given as
2σmean. Values at or below the detection limit have been omitted. 96
Figure 4.8: Relationship between modal abundance of Olivine and Pyroxene versus the
Mg# in olivine in dunite (circles), lherzolite (triangles), plagioclase bearing
lherzolite (open triangles) and olivine bearing gabbronorite (red squares). With
increasing pyroxene content the Mg# number decreases approaching
asymptotically the Mg# of olivine found in gabbronorite at the mafic -
ultramafic contact. 97
Figure 4.9: Olivine composition of ultramafite and olivine-bearing gabbronorite. A
compositional gap exists between Mg# 0.87 and 0.85. High Mg# olivine has a
positive correlation with MnO content and relatively restricted concentrations,
whereas low Mg# olivine have no correlation with MnO and the MnO
concentrations are about two times more variable. 98
Figure 4.10: Al2O3 versus Mg# for clinopyroxene from the Chilas Complex. The solid line
represents an igneous trend determined from core composition. Grey arrows
represent metamorphic cooling observed in zoned cpx. The igneous trend is
characterized by an increasing Al2O3 content with decreasing Mg# until
plagioclase is present in the rocks. Due to the fractionation of plagioclase the
Al2O3 content of clinopyroxene decreases with decreasing Mg#. The
metamorphic trend represents a Tschermak (Fe,Mg)SiAl-2 exchange during
cooling. 99
Figure 4.11: Abundance of trace elements in clinopyroxene normalized to primitive mantle
(Sun and McDonough, 1989). Given is the average value of 6 to11 mineral
grains analysed per sample. Error is given as 2σmean. 100
Figure 4.12: Al2O3 versus Mg# for orthopyroxene from the Chilas Complex. Symbols are
the same as in . Similar to cpx an igneous and a metamorphic trend is observed,
 216

however the metamorphic Tschermak exchange is weaker in opx than in

clinopyroxene. 102
Figure 4.13: Plot of Fe/Mg in clinopyroxene against Fe/Mg in olivine and orthopyroxene.
The 1:1 line is given for comparison only. All data consistently plot over the
1:1 line. The misfit with unity increases with increasing Fe content. 102
Figure 4.14: Abundance of trace elements in orthopyroxene normalized to primitive mantle
(Sun and McDonough, 1989). Given is the average value of 5 to10 mineral
grains analysed per sample. Errors are given as 2σmean, the grey line is the
detection limit. 103
Figure 4.15: Ternary plot (Fe3+-Cr-Al) illustrating spinel composition from the Chilas
Complex compared with data from the Jijal Complex (Jan and Windley, 1990)
104
Figure 4.16: Spinel compositions in the ultramafite compared to spinel composition in the
Jijal complex (Jan and Windley, 1990). 105
Figure 4.17: Spinel composition of the main gabbronorite. Solid squares are own analyses
and opened circles are from Kha, (1989). The analysed spinels form solid
solution between ilmenite and ulvospinel and between ulvospinel and
magnetite. 105
Figure 4.18: Backscatter image of Sample C03-15 illustrating replacement of cpx and Mg-
hornblende by tremolitic amphibole. Mg-Hornblende represents an early
magmatic phases whereas Tremolite replaces Mg-hornblende and cpx and is a
later stage metamorphic overgrow. 106
Figure 4.19: Magmatic amphibole composition within various units of the Chilas Complex.
Nomenclature after (Leake, 1978). 107
Figure 4.20: Averaged primitive mantle (Sun and McDonough, 1989) normalized trace
element pattern of amphibole from olivine-bearing gabbronorite (red) and
porphyroclastic lherzolite (black). 107
Figure 4.21: Primitive mantle normalized (Sun and McDonough, 1989) results of the trace
element modeling following the approach of (Hermann et al., 2001). The mean
modal composition of the gabbronorite has been approximated as plag 60%,
opx 30% and cpx 10% based on an average value of grid counting and modal
calculation based on whole rock analyses. Squares =clinopyroxene, triangles =
plag, and circles opx. The composition of plag and opx has been averaged and
assumed to be constant, whereas the spread in cpx can be explained by
 217

equilibration with different amount of interstitial liquid. The upper diagram

displays the measured values. In the lower diagram the open symbols are the
results of the model, solid symbols are the measured values of the sample. A)
The results for sample C48 can be reproduced by 60% of differentiation. The
spread in cpx composition indicates equilibration with different amount of
interstitial liquid. The solid line represents 2% of interstitial liquid whereas the
dashed line was calculated with 7% B) Results for sample C7 indicate 65% of
differentiation and an average of 6,5% of interstitial liquid. The spread in cpx
composition can be modeled with interstitial liquid between 5-8% (not shown).
110
Figure 4.22: Primitive Mantle normalized spidergram comparing the Chilas parental melt
composition, the result of 20% non-modal batch melting (ol 40%, opx 20%, cpx
30%, amph 9%, sp 1%) of a previously slightly depleted source (Grove et al.,
2002) and the fluid rich component. The fluid component (hatched field) is
compared to a fluid component inferred for the Mt. Shasta volcanite (grey field)
(Grove et al., 2002), a fluid in equilibrium with the wedge, an eclogitisised slab
(Ayers, 1998) and with a fluid involved in the Mariana magma formation
(Stolper and Newman, 1994). The Chilas fluid component is most similar to a
low percentage melt derived from the basaltic part of a subducted slab. It
displays a negative Gd anomaly which is interpreted as an artifact due to a
slight negative anomaly present in the Chilas parental melt which is however,
within error and so not significant. However, due to the low percentage of
added fluid this “unreal” anomaly is strongly enhanced. The interpreted true
pattern is indicated by the hatched lines. 113
Figure 5.1 Geological map of the Kohistan arc, NE Pakistan. The ultramafite (asterisks)
have been studied in the vicinity of the town Chilas in the eastern part of the
Kohistan arc 123
Figure 5.2 Block diagram of the contact relationship between zoned ultramafic and
gabbronorite. Indicated are key contact relationship supporting a replacive
origin for the dunite. A reaction zone is developed between ultramafite and
gabbronorite. It is interpreted as the result of infiltration of a melt parental to
the gabbronorite into the dunite forming lherzolite. Indicated also are location
for six samples which, have been selected for LA-ICP-MS analysis and partly
(C174 and C7) for Sm-Nd isotopic investigation. 125
 218

Figure 5.3: Picture of the reaction zone between gabbronorite and dunite forming a
centimeter scale pyroxenite. The contact between pyroxenite and gabbronorite
is sharp but irregular, indicating similar sub-solidus rheology for both. 127
Figure 5.4 Total Alkali vs Silica (TAS) diagram for rock classification of samples from the
gabbronoritie sequence of the Chilas Complex. Nomenclature of igneous rocks
after Cox et al. (1979). The solid line divides sub-alkalic and alkalic magma
series (Miyashiro, 1978). The whole rocks show only a limited spread in
composition and are classified as sub-alkalic gabbros to diorites. Sample C 66
an olivine-bearing gabbronorite has extremely low values of silica and alkali
and plots outside of the classification schema of Cox et al. (1979). 130
Figure 5.5: Primitive mantle normalized REE spectra A) from whole rock samples from
porphyroclastic lherzolite (solid line without symbols), dunites (dashed lines
with open symboles, absent of symbols indicated extrapolated values below or
at the detection limit) and peridotites form the contact zone (solid line with
symbols) and B) gabbronorite whole rock samples. 131
Figure 5.6: Primitive mantle (Sun and McDonough, 1989) normalized spider diagram of
clinopyroxene from A) porphyroclastic lherzolite (C03-43 open squares; C03-
44 filled squares; C03-45 open circle) B) dunites (solid symbols) and peridotite
form the contact zone (open symbols) and C) olivine-bearing gabbronorite
(open squares) and gabbronorite (C 48 solid circles and C7 open circles). 134
Figure 5.7: A 147Sm/144Nd versus 143Nd/144Nd diagram of mineral separates (clinopyroxene
and plagioclase) and whole rocks from a dunite sample (C174) and two
gabbronorite samples (C7 and C48). Interpreting the linear relationship between
the data-points as an isochrone corresponds to an age of 104 ±27 Ma 136
Figure 5.8: Results of the in-situ crystallisation modelling based on the model of
(Langmuir, 1989). Shown are own whole rock analyses of gabbronorite
(squares) and unpublished ones (Khan, 1990) (small squares) compared with
whole rock composition of the Braccia gabbro (circles) (Hermann et al., 2001)
and the Fiambala gabbro (crosses) (DeBari, 1994). The bulk solid partition
coefficient used are KD(Ba)= 0.045 KD(Sr)= 2.2 and were chosen to explain the
observed spread of Ba concentrations. The initial liquid composition is
calculated in equilibrium with plagioclase in the olivine bearing gabbronorite.
KD are calculated using average An89 and 1100°C and the semi-empirical
formulation of Blundy and Wood (1991). Concentrations are Ba =113 ppm and
 219

Sr= 428 ppm. The corresponding solid composition has been calculated
assuming constant element concentration in the system. The evolution of the
liquid is shown for three cases: f= 1 is the pure fractional crystallisation trend,
f= 0.75 and f= 0.1. Notice that f = 1 indicates that 100 % of the liquid within the
solidification zone returns into the magma chamber, hence 0% of liquid is
trapped in cumulates. The symbols of the modelled trend for f= 0.1 represent
10% of differentiation. The red line indicates the evolution of the liquid
composition and the black line the corresponding solid calculated assuming
mass balance. Own samples have been labeled with the corresponding Mg#.
See text for further discussion. 137
Figure 5.9: Two different trend observed in whole rock composition (symbols as in ). The
Mg# of the cumulate dominated trend decreases rapidly with increasing Ba
content, whereas the Mg# in the liquid is buffered and decreases only slightly
with strongly increasing Ba concentration. The crossing point of the two trends
roughly corresponds to the initial melt composition parental to the Chilas
gabbronorite. 139
Figure 5.10: A) Primitive mantle (Sun and McDonough, 1989) normalized spider-diagram
of whole rock data for homogeneous gabbronorite solid symbols. Open boxes
are the mean values of a melt calculated in equilibrium with primitive (Mg#
=0.92-0.94) clinopyroxene (using KD values of (Hart and Dunn, 1993) with an
extrapolated value for Eu from (Ionov et al., 2002)) extracted from a primitive
dunitic sample (C 174). The red field indicates the spread of the six
clinopyroxene analyzed. B) Same samples in molar Fe/Mn vs Mg/Mn plot. The
slope of the reaction from amphibolite lherzolite to replacive dunite is 0.22.
Precipitation of only olivine with Mg# 0.8 would result in a slope of 0.25
wereas clinopyroxene or amphibole dissolution alone in a slope of -0.1. The
current model implies that 1 mole of Olivine precipitation is associated with
redissolution of 3 moles of cpx/amph. C) Molar Ca/Mn vs Mg/Mn plot for
whole rock analysis from Amphibole bearing peridotite (squares), Dunites and
lherzolite (diamonds) and gabbronoritic (triangles) samples. Three different
trends can be identified. 1.) Amphibole bearing lherzolite loose mainly Ca and
gain slightly Mg when transformed to Dunite indicating disolution of
clinopyroxene/amphibole and precipitation of olivine (see text for details). 2.)
Dunite transformation to lherzolites show two different trends corresponding to
 220

the presence or absence of plagioclase in the sample, but they approach the filed
of gabbro-gabbronorite samples confirming the field observed re-infiltration of
melt parental to the gabbronorite into the dunitic units. The observed trend in
whole rocks results from both reaction of melt + ol = pyx and trapped
interstitial liquids ultimatively forming the characteristic zonation of the zoned
ultramafic bodies. 144
Figure 5.11: Primitive mantle normalized spidergram illustrating the calculated parental
magma composition of the Chilas gabbronorite sequence. The parental liquid
calculated in equilibrium with primitive clinopyroxene (using the above cited
set of KD`s) from olivine-bearing gabbronorite (C66) is indicated as red boxes.
The Black line is the parental liquid calculated assuming that the parental liquid
is located in between the liquid- and cumulate-dominated composition
corrected for trapped interstitial liquid. 146
Figure 5.12: Comparison of crystallized liquid (above) and the calculated bulk
gabbronorite composition (below) with lower crust estimates of Taylor and
McLennan (1985) and Rudnick and Gao (2003) (triangles and diamonds
respectively) see text for details. 147
Figure 6.1: Geological map of the Kohistan arc. The box indicates the location of map
Sample C-01-77 is from the western part of the Kohistan arc, near the town of
Chilas and indicated separately. The location of the remaining samples are
given in Figure 6.2. The location of Figure 6.8 in the northern part of the
Kohistan arc is additionally indicated 160
Figure 6.2: Detailed map of the Dir-Kalam are based on own field observations,
interpretation of multispectral Landsat 7 ETM + satellite images and published
maps (Butt et al., 1980; Chaudhry et al., 1987; Jan and Mian, 1971; Sullivan et
al., 1993). Previously plutonic units south of the Barul-Banda Slate Formation
have been interpreted as stage 1, those in the north of the volcanites as stage 2
(Sullivan et al., 1993). 162
Figure 6.3: Zircon Cathodoluminescent images from different samples. Circular objects are
the laser ablation pits after analyses and the resulting age is given. A)
recrystallized Zircon from sample 01B22; there is no detectable age difference
between different areas of the Zircon and the whole Zircon is recrystallized; B)
young Zirkon from the same sample showing a light core and a darker rim
 221

displaying an oscillatory growing pattern. Apparently there is no detectable age

difference between core and rim, but notice that the shot of the LA-ICPMS 165
Figure 6.4: Total Alkali vs Silica (TAS) diagram for rock classification. Nomenclature of
igneous rocks after (Cox et al., 1979). 167
Figure 6.5 Conventional concordia plot of the analysed zircons, ages are calculated from
the 206Pb/235U ratios and errors are given as 2σmean.. 168
Figure 6.6 Frequency histogram of the analysed Zircon. Black bars indicate non common
lead corrected values and white indicate corrected ones. The correction has no
effect on the mean age and has a slight positive effect on the normal distribution
of sample BO-02-13 whereas the normal distribution of sample Rb-02-16 is
unaffected by the correction. 169
Figure 6.7: Plot illustrating the continuous spectrum of the ages 170
Figure 6.8: Landsat 7 ETM + pictures (combination of bands 745) of the NE Kohistan.
Indicating that trend of structures are influenced by the trend of the Kohistan-
Karakoram Suture zone only within a few Kilometers of the Suture zone. Bright
blue color represents snow or ice fields, white fields are eliminated shadows.
171
Figure 6.9 : A) Primitive mantle (Sun and McDonough, 1989) normalized trace element
pattern of Chilas gabbronorite (solid symbols) compared to the so-called group
1 volcanite from the Dir area (Shah and Shervais, 1999). The REE pattern of
both groups is sub-parallel. The Dir volcanites have higher concentration and a
negative Eu anomaly whereas the Chilas gabbronorite have lower
concentrations and a complementary positive anomaly. B) The Ce/Yb ratio is
similar for both groups. C) All Dir volcanite of Shah and Shervais (1999)
compared to the in-situ crystallization model presented in chapter 5. Some
volcanite (circles) follow the trend defined by the Chilas gabbronorite (big
squares own data and small squares from Khan (1988)). Some volcanite follow
the trend whereas other clearly deviate from, indicating that not all volcanite in
the Dir area are necessarily co-magmatic with the Chilas gabbronorite. 174
Figure 7.1: Molecular normative projection from clinopyroxene onto the basalt tetrahedron
of hydrous fluid undersaturated melting experiments (Ulmer, 2001). The
hatched area indicate the likely primary melt composition of the Chilas
Complex based on constrain of the melting depth. The primary melt
 222

composition of the Chilas Complex is most likely an olivine tholeiite. However,

the field slightly touches the quartz normative compositions. 182
Figure 7.2: Schematic illustration of the shape of the melt channels. 184
Figure 7.3: A not impossible P-T trajectory for the development of the Chilas. Phase
reaction are based on hydrous (2,8% wt%) equilibrium fractionation
experiments using a picrobasalt from the Adamello (Pulmer unpublished). Due
to adiabatic decompression the melt fractionating olivine is undersaturated in
pyroxene. During this stage the melt will form the relacive dunite. Upon
cooling it will continue crystallizing olivine, spinel and subsequently
clinopyroxene. Finally, in this case around 1000°C, the melt becomes olivine
saturated. If melt at these conditions infiltrates and reacts with the uprising
ultramafite it will consume olivine to produce pyroxene forming the observed
zonation 184
Figure 7.4: Illustration of a model for the tectono-magmatic evolution of the Kohistan arc
during northward directed subduction of the Tethyan ocean and subsequent
collision with Eurasia and India. From the present data the timing of collision of
Kohistan with Eurasia is unconstrained. Accordingly, two possible scenarios
have been drawn in C1 and C2 depending whether the collision occurred before
or after the intrusion of the Chilas Complex. 190
 223

Curriculum Vitae

PERSONAL DETAILS
Address: Fabrikstr. 21, 8005 Zürich, Switzerland
Date of Birth: 02/04/1971
Place of Birth: Mainz , Germany
Nationality: Austrian
Status: Single

Academic Career:

- October 2000 –Dez 2004: PhD-studies at the Geological Institute ETH

Zurich.
- 18.9.2000: Diploma in Geology at the Johannes Gutenberg Universität Mainz. Final
grade: sehr gut (1.26)
- May 1999 until July 2000: Diplomathesis within an Erasmus programs at the ETH
Zürich entitled: “Description of the deformation structures in the hanging and foot wall of
the Glarus thrust.” Final Grade: excellent (1.0). Supervised by Dr. B. den Brok and Prof. L.
Baumgartner.
- 1997: Scholarship of the Johannes Gutenberg University for young researchers
(Förderungsstipendium für Nachwuchskräfte)
- 1996: Vordiplom Geology final grade: 1.6
- 1995: Beginning of Study of Geology
- 1995: First Part of Vordiplom in Chemistry (Anorganische Chemie, Physik,
Mathematik)
- 1993/94: Study of Chemistry Diploma at the Johannes Gutenberg Universität Mainz