The Origin of Adakites in the Garibaldi Volcanic Complex, Southwestern British Columbia, Canada

The Origin of Adakites in the Garibaldi Volcanic Complex, southwestern British Columbia, Canada

A Thesis

Submitted to the Faculty of Graduate Studies and Research

In Partial Fulfillment of the Requirements

For the Degree of

Master of Science

Geology

University of Regina

Julie Anne Fillmore

Regina, Saskatchewan

November 2014

UNIVERSITY OF REGINA

FACULTY OF GRADUATE STUDIES AND RESEARCH

SUPERVISORY AND EXAMINING COMMITTEE

Julie Anne Fillmore, candidate for the degree of Master of Science in Geology, has presented a thesis titled, The Origin of Adakites in the Garibaldi Volcanic Complex, Southwestern British Columbia, Canada, in an oral examination held on August 22, 2014. The following committee members have found the thesis acceptable in form and content, and that the candidate demonstrated satisfactory knowledge of the subject material.

External Examiner: Dr. Martin Beech, Campion College

Supervisor: Dr. Ian M. Coulson, Department of Geology

Committee Member: Dr. Tsilavo Raharimahefa, Department of Geology

Chair of Defense: Dr. Josef Buttigieg, Department of Biology

Abstract

The Garibaldi Volcanic Complex (GVC) is located in southwestern British

Columbia, Canada. It comprises two volcanic fields: the Garibaldi Lake Volcanic

Field (GLVF) in the north and the Mount Garibaldi Volcanic Field (MGVF) in the south. Petrographical and geochemical studies on volcanic rocks collected from the GVC have determined that they exhibit adakitic characteristics; these intermediate rocks range from basaltic andesite to dacite represented mainly by lava flows, domes and minor pyroclastic material. All the lavas exhibit evidence of magma mixing, which include sieve textured crystals, dehydration reaction textures, differently sized phenocryst populations, xenocrysts and xenoliths. The geochemistry of the GVC magmas exhibit several adakitic indicators which

# include high Sr/Y (> 40), Mg (~ 51), Al2O3 (> 15 wt. %), low K2O/Na2O (~ 0.3), low Yb (< 1.9 ppm) and fractionated rare earth element (REE) compositions

(La/Yb(N)~ 10), which have not been identified in previous studies. Adakites are the product of subducted slab partial melts within the garnet stability field, subdivided into low silica adakites (LSA; < 60 wt. % SiO2) and high silica adakites (HSA; > 60 wt. % SiO2) reflecting differing source regions, where HSA are primary slab melts and LSA are partial melts of mantle wedge peridotite that has been previously modified by slab-derived magmas. Identification of adakitic rocks in tectonic regimes unrelated to subduction have led to the argument that the distinctive high Sr/Y and La/Yb ratios are not unique and cannot be used as an indicator of slab partial melting. Basalts in adakite suites are often enriched iii in high field strength elements (HFSE; Nb in particular) and are classified as niobium enriched basalts (NEB); NEB is argued to originate from mantle wedge peridotite that has been previously metasomatised by slab partial melts. Trace element modelling illustrates that the GVC adakites can be generated by partial melting of subducted ocean crust. Incompatible and compatible element ratios and relative crustal thickness beneath the GVC preclude AFC processes or high pressure partial melting or crystallisation of basalt as the source of the adakite signature in the GVC rocks and suggests that the GVC adakites likely result from slab partial melts.

Acknowledgements

I would like to acknowledge everyone who provided funding and support for this project. I thank my supervisor, Dr. Ian Coulson, who provided funding through his Natural Sciences and Engineering Research Council (Discovery Grant) and facilitated microprobe analysis of selected samples, as well as much support and consultation. The University of Regina provided financial support through the

Faculty of Graduate Studies and Research scholarships, the Teaching

Assistantship program and through the Department of Geology Teaching

Assistantship program. I would especially like to thank Dr. Michael Clynne, former editor in chief of the Bulletin of Volcanology for extremely helpful comments and discussion on my paper published last year in the Bulletin and strongly influenced this study. Dr. William Minarik and Glenna Keating, who provided whole rock geochemical analysis of the samples studied at McGill

University. Finally, I would like to thank my family and all my friends for their moral support and putting up with my frustrations.

Most of all, I have to thank my wonderful and loving husband, without whom this project could not have been possible.

Table of Contents

Abstract ii Acknowledgements iv Table of Contents v List of Figures viii List of Appendices x List of Acronyms xii 1. Introduction 1 1.1 Previous Work 2 1.2 Adakites in the GVC 6 2. Regional Geology 7 2.1 Eruptive History of the GVC 9 2.2 History and definition of adakites 12 2.3 Adakite – tonalite-trondhjemite-granodiorite (TTG) association 15 2.4 Non-slab melt models for adakite genesis 18 3. Petrography 20 3.1 GLVF 20 3.1.1 Cheakamus Valley Basalts (CVB) 20 3.1.2 Helm Creek Basaltic Andesite (HCBA) 23 3.1.3 Desolation Valley Basaltic Andesite (DVBA) 27 3.1.4 Barrier andesite 29 3.1.5 Black Tusk 30 3.2 MGVF 33 3.2.1 Ring Creek andesite 33 vi

3.2.1.1 Proximal Ring Creek andesite 33 3.2.1.2 Distal Ring Creek andesite 35 3.2.2 Columnar Peak dacite 36 3.2.3 Paul Ridge andesite 37 4. Geochemistry 39 4.1 Analytical techniques 40 4.1.1 Whole Rock Geochemistry 40 4.1.2 Mineral Geochemistry 42 4.2 Results 43 4.2.1 Basalts and Basaltic Andesites 43 4.2.2 Adakites 48 5. Discussion 54 5.1 Adakite Geochemistry of the GVC 56 5.1.1 La and Cr 56 5.1.2 LSA versus HSA 57 5.1.3 Magma Mixing in the GVC 58 5.1.4 Interaction with Mantle Peridotite 70 5.1.5 Isotope Geochemistry 71 5.2 Possible models for Adakite Genesis 73 5.2.1 Partial melting of basaltic lower crust 73 5.2.2 High pressure fractionation/AFC of basaltic magma 77 5.2.3 Partial melting of subducted ocean crust 80 5.2.3.1 Association with niobium-enriched basalts (NEB) 84 5.2.3.2 Ni content in olivine 93 5.2.3.3 Regional tectonic regime for southwestern British Columbia 96 vii

5.3 Petrogenetic model for Adakite Genesis beneath the GVC 98 6. Conclusions 104 References 107 Appendix A – Whole Rock Geochemistry 127 Appendix B – Normative Mineralogy 130 Appendix C – Mineral Geochemistry 132 Appendix D – Trace element data for modelling of slab partial melts 142

viii

List of Figures

Fig. 1: Location Map of the GVC 3 Fig. 2: Geology of the GVC (modified from Green 1977) with sample locations 4 Fig. 3: Field Photographs 10 Fig. 4: Thin section photomicrographs of basalts and basaltic andesites 21 Fig. 5a, b: Mineral composition of mafic minerals (CVB and HCBA) 24 Fig. 5c, d: Mineral composition of mafic minerals (DVBA and PR) 25 Fig. 6a-f: Thin section photomicrographs of adakites (BF, BT, URC) 31 Fig. 6g-l: Thin section photomicrographs of adakites (LRC, CP, PR) 32 Fig. 7: Plot of total alkalis vs. silica 45 Fig. 8: Major element Harker diagrams 46 Fig. 9: Trace element Harker diagrams 47 Fig. 10: Primitive mantle normalised multi-element diagrams 49 Fig. 11: Chondrite normalised multi-element diagrams 50 Fig. 12: Adakite-normal arc rock differentiation plots (Sr/Y vs. Y, La/Yb vs. Yb) 53

Fig. 13: K/Rb vs. SiO2/MgO and Sr-K/Rb-[(SiO2/MgO)*100] plots 55 Fig. 14: LSA-HSA differentiation plots 59 Fig. 15: SEM photomicrographs of the GVC adakites 62 Fig. 16: Incompatible-compatible element ratio plots for the GVC adakites 68 Fig. 17: Adakitic indices plots for GVC and lower crustal melts 76 Fig. 18: Modelled slab partial melts compared to observed GVC adakite compositions 82 Fig. 19: Binary mixing plots for the CVB and HCBA 88 ix

Fig. 20: Simple three component mixing model for the CVB 90 Fig. 21: Incompatible-compatible element ratio plots for the CVB and HCBA 91 Fig. 22: NiO vs. Fo content in GVC olivines 95 Fig. 23: Interpreted petrogenetic model for adakite genesis in the GVC 99

List of Appendices Appendix A – Whole Rock Geochemistry 127 Table A1: Major and minor element composition of investigated 128 samples from the GVC Table A2: Trace and rare earth element composition of 129 investigated samples from the GVC Appendix B – Normative Mineralogy 130 Table B1: Normative mineralogy of representative samples 131 from the GVC Appendix C – Mineral Geochemistry 132 Table C1: Electron microprobe compositions of olivine from the 133 GVC Table C2: Electron microprobe compositions of clinopyroxene 136 from the GVC Table C3: Electron microprobe compositions of orthopyroxene 137 from the GVC Table C4: Electron microprobe compositions of plagioclase from 138 the GVC Table C5: Electron microprobe compositions of oxides from the 140 GVC Appendix D – Trace element modelling of slab partial melts 142 Table D1: Partition coefficients for elements used in slab partial 143 melt model Table D2: Modelled compositions of slab melts at various 143 melt fractions (F); residue=70% Cpx, 10% Gt, 20% Hbl Table D3: Modelled compositions of slab melts at various melt 143 fractions (F); residue=70% Cpx, 20% Gt, 9.5% Hbl, 0.5% Rut Table D4: Modelled compositions of slab melts at various melt 143 fractions (F); residue=70 % Cpx, 25 % Gt, 4.5 % Hbl, 0.5% Rut Table D5: Estimated starting composition of Juan de Fuca (JdF) 143 MORB xi

Fig. D1: Failed slab partial melt models for various residual 144 compositions

xii

List of Acronyms ADR Andesite-dacite-rhyolite AFC Assimilation-fractional crystallisation BF Barrier flow BT Black Tusk Cay Mt. Cayley CP Columnar Peak dacite CVB Cheakamus Valley Basalt DVBA Desolation Valley Basaltic Andesite GLVF Garibaldi Lake Volcanic Field GVB Garibaldi Volcanic Belt GVC Garibaldi Volcanic Complex HCBA Helm Creek Basaltic Andesite HFSE High field strength element(s) HREE Heavy rare earth element(s) HSA High silica adakite LC Lower crust LILE Large ion lithophile element(s) LOI Loss on ignition LRC Lower Ring Creek andesite LREE Light rare earth element(s) LSA Low silica adakite MGVF Mt. Garibaldi Volcanic Field MREE Middle rare earth element(s) NEB Niobium-enriched basalt (N)MORB (Normal) mid-ocean ridge basalt xiii

OIB Ocean island basalt PR Paul Ridge andesite REE rare earth element(s) RC Ring Creek andesite TTG tonalite-trondhjemite-granodiorite URC Upper Ring Creek andesite

1. Introduction

Partial melting of hydrated mantle wedge is a favoured mechanism in the generation of arc-type magmas, and previous studies of magmatism within the

Garibaldi Volcanic Complex (GVC) have supported this model (Green 1977,

1981, 1990, 2006; Green and Harry 1999, Harry and Green 1999). The distinctive chemistry of the intermediate to felsic rocks of the GVC, however, suggests that mantle wedge partial melts are not the only possible source of magmatism in the GVC. The GVC rocks exhibit geochemical characteristics that identify them as adakites, with high Sr/Y (> 40) and low Yb (< 1.9 ppm). This signature is attributed to partial melting of subducted basaltic ocean crust, at a depth where garnet is stable but plagioclase is not (Defant and Drummond

1990), and generates felsic to intermediate melts. Adakite magmas are unique in that they are defined almost solely on geochemistry (Richards and Kerrich

2007); the mineralogy of adakitic rocks and tectonic environments in which these have been identified can be quite variable. However, in several environments where adakites are found in association with basaltic rocks, basalt (and basaltic andesite) rocks are commonly found to be unusually enriched in high field strength elements (HFSE) and Nb in particular (Sajona et al. 1996). These mafic rocks have been termed niobium enriched basalt (NEB) and are argued to originate from mantle wedge peridotite altered by the interaction with ascending slab partial melts. This may suggest that a petrogenetic link exists between adakite and NEB however, further studies of this relationship have found that 2 they are not mutually exclusive and there are suites where adakite is present but

NEB are not (e.g. Chile and Argentina; Stern and Kilian 1996) and vice versa. As a result, the definition of ‘adakite’ and whether or not it truly represents a distinct magmatic process has been the subject of much discussion.

1.1 Previous Work

The GVC was first studied by Burwash (1914a), who described the relative timing between volcanism therein and the effects of glaciation occurring in the area that includes Mt. Garibaldi and Black Tusk. Mathews (1951, 1952), followed up in this work and expanded upon the Quaternary geology and glacial geology.

More detailed studies were completed later by Mathews (1957, 1958), where the rock types of the GVC and Mt. Cayley were characterised in terms of rock petrography and major element chemistry, in addition to a description of the

Quaternary glacial geology and stratigraphy of both the Mt. Garibaldi Volcanic

Field (MGVF) and The Table (Figs. 1 & 2). Work in the GVC ceased until 1970’s when a Ph.D. study by Fiesinger further investigated the petrography of rocks collected from the GVC, Mt. Cayley and the Quesnel Highlands, as well as their major element chemistry (Fiesinger 1975). A B.Sc. study by Sivertz (1976) was the first to characterise the rocks of Opal Cone and the Ring Creek flow as dacite through a combination of petrographic study and whole rock major and limited trace element chemistry (Fig. 2). Sivertz (1976) proposed that the source of the Ring Creek rocks was hydrous partial melting of quartz eclogite. The subsequent Ph.D studies by Green (1977), however, suggested that the basalt 3

Figure 1: Location map of the GVC in southwestern British Columbia, with plate boundaries highlighted and relative plate motions of the Juan de Fuca and Explorer plates indicated. Map modified after Hickson et al. (1999) and Madsen et al. (2006).

Figure 2: Geology of the GLVF and the MGVF in the GVC. Sample locations are shown. Map modified after Green (1977) and Green et al. (1988). Note: stratigraphy of the GLVF and MGVF is younging upwards independent of each other. 5 and andesite rocks of the Garibaldi Lake Volcanic Field (GLVF) were the result of multistage fractionation and partial melting of mantle wedge peridotite, based on petrography, whole rock geochemistry, microprobe and isotopic data, and geothermometry. The GLVF has been studied extensively, primarily by Green

(1981, 1990, 2006; Green et al. 1988), where the volcanology and eruptive sequence of each centre has been described in depth. Despite these many investigations, to date, there are no whole rock REE compositional data for the

MGVF and only one published study of a few samples from the GLVF (Green and Henderson 1984)1. Recent studies of the GVC have focused primarily on the basaltic lavas and how their chemistry (particularly the LILE, HFSE and REE) is affected by the thermal structure of the Juan de Fuca Plate (Green and Harry

1999, Harry and Green 1999, Green and Sinha 2005, Green 2006). These investigations have found that the subduction component (i.e., LILE, LREE contents) decreases from the volcanoes of Glacier Peak in southern

Washington, northwards to north of Mt. Meager in central British Columbia and that this is due to the decrease in age of the Juan de Fuca Plate further north, and the resulting earlier onset of dehydration reactions within the slab (Green and Sinha 2005, Green 2006). As older studies (e.g., Green 1977, Sivertz 1976) have already described the mineralogy and major element chemistry of the intermediate and felsic rocks of the GVC, very little new research has been done on them. This is likely the reason that adakitic rocks have not been recognised in

1Current research at the University of British Columbia at the time of completion of this thesis is revisiting previously collected samples of the intermediate rocks in the GVC for geochemical and isotopic analysis. 6 the GVC until the study by Fillmore and Coulson (2013) and also by extension, that the basalts in the GVC have not been recognised as NEB. There have been some new isotopic studies on the basalts of the Cascades range (which includes the Cheakamus Valley Basalts and the Helm Creek Basalt; Martindale et al.

2014) as well as on select evolved rocks from the GVC and Mt. Cayley

(Martindale et al. 2014). The NEB characteristics of the GVC basalts were also identified in these works (although not explicitly identified as such), but this new isotope data has not been evaluated in the context of the slab melt model presented in this thesis and are not discussed further.

1.2 Adakites in the GVC

Geochemical and petrological attributes of andesite and dacite rocks that comprise the GVC combined with associated field relationships (i.e., NEB) strongly suggest that they are the result of slab melting, under the definition put forth by Defant and Drummond (1990). Interactions between these melts and the overlying mantle wedge are also evident in their trace element compositions, a refinement made to the slab melt model by Martin et al. (2005). This study of the

GVC provides a unique opportunity to evaluate the following: (1) that partial melting of subducted oceanic crust (i.e., subducted slab) is a process that is more common in modern environments than previously thought; (2) magmas produced by such a process can in fact migrate through the mantle wedge and crust to be erupted at surface with the slab melt signature intact, and (3) the presence of NEB and adakite in the same suite of rocks can suggest a 7 petrogenetic link and that these support a slab melt origin for adakitic volcanism in the GVC.

The objectives of this study are: (1) to provide the first complete geochemical characterisation of the intermediate and felsic rocks of the MGVF and the GLVF,

(2) be the first work to characterise the intermediate and felsic rocks in the GVC as adakite, and basalt and basaltic andesite as NEB, (3) evaluate the various models of adakite genesis in the context of the GVC, (4) introduce a potential link between NEB and LSA as parent and daughter magmas, and (5) suggest that the source for NEB in the GVC is mantle peridotite metasomatised by partial melts of subducted oceanic lithosphere, which in turn, suggests a petrogenetic link between the NEB and adakite rocks of the GVC.

2. Regional Geology

The Garibaldi Volcanic Belt (GVB) extends from the Canada-U.S.A. border northwards into British Columbia for approximately 140 km (Fig. 1; Sherrod and

Smith 1990). The GVC lies within the southern portion of this belt between the towns of Whistler and Squamish, and comprises two fields; the GLVF in the north and the MGVF in the south (Fig. 2). The volcanic rocks from the GLVF and the MGVF range in composition from basalt to dacite and have been previously interpreted to be the result of hydrous melting of the mantle wedge above the

Juan de Fuca Plate, which is subducting beneath the North American Plate, and subsequent fractionation at various depths during ascent (Green 1977, 1981, 8

1990; Green and Harry 1999, Green and Sinha 2005, Green et al. 1988).

Basaltic volcanism in the GVC is thought to be related to the decreased volatile content of the Juan de Fuca Plate and results in lower degrees of melting under higher pressures and temperatures (Green 2006). The decreased slab flux is attributed to both a northward decrease in plate movement and plate age

(Riddihough 1981, 1984; Green 1990, Wilson 2002). The younger, more buoyant

Explorer Plate separated from the Juan de Fuca Plate at 4 Ma (Wilson 2002,

Audet et al. 2008), and this deviation may relate to an increase in Quaternary volcanism in the GVC.

Subduction of the Explorer Plate beneath North America is slower than that of the Juan de Fuca Plate and it has been suggested that the Explorer Plate is undergoing capture by the North American Plate (Audet et al. 2008). The difference in subduction rates has caused a region of extension and slab thinning along the Nootka Fault zone (Fig. 1), a transform fault that fractured in response to an interval of ridge propagation and reorientation (Riddihough 1984,

Wilson 1988, Madsen et al. 2006) and created the Explorer and Juan de Fuca plates. Independent movement of the Explorer Plate northward relative to the east-northeasterly movement of the Juan de Fuca Plate suggests that the separation encompasses both the subducted portions of the plates in addition to the non-subducted oceanic parts (Madsen et al. 2006). This segmentation coupled with the relative subduction vectors of the plates has resulted in a change in mantle flow. Recent studies (Madsen et al. 2006, Audet et al. 2008) 9 have determined and modelled upwelling of mantle material along this boundary and suggest the resulting structure has the potential to create a slab window.

Several workers (e.g., Breitsprecher et al. 2003, Thorkelson and Breitsprecher

2005, Ickert et al. 2009) have found a link between the formation of slab windows or slab tears and adakitic volcanism, whereby asthenospheric upwelling provides the heat flux necessary to melt the edges of slabs and in doing so, facilitate the generation of adakite magmas.

2.1 Eruptive History of the GVC

The GLVF and the MGVF sit unconformably on the Coast Mountain Crystalline

Complex, which is a series of metamorphosed quartz diorite and granodiorite plutons of Cretaceous age (Rusmore and Woodsworth 1991). Timing and eruptive products from both fields have been summarised by previous authors

(Green 1977, 1981, 1990; Green et al. 1988) and are briefly outlined below.

Quaternary volcanic centres in the GLVF include Black Tusk, Cinder Cone,

Clinker Peak, Mt. Price and The Table, as well as the Cheakamus Valley Basalts which were erupted from an unknown centre (Fig. 2). The oldest activity in the

GLVF was at Black Tusk and Mt. Price with episodic volcanism beginning at 1.3

Ma (Fig. 3a). The rocks of Black Tusk are hornblende andesite and orthopyroxene andesite flows. The oldest rocks of Mt. Price are a series of hornblende andesite and andesite flows followed by the formation of the hornblende-biotite andesite satellite cone along Garibaldi Lake. Volcanism ended in the Mt. Price area with the eruption of the Barrier and Culliton Creek 10

Figure 3: Field photographs. (a) Black Tusk; person for scale (red circle), (b) Barrier andesite (yellow) filling valley (red); inset=porphyritic texture of the Barrier andesite, (c) the Cinder Cone (red), looking east, (d) The Table, looking south, (e) columnar jointing in the CVB, near Daisy Lake; inset=vesicular texture in the CVB, (f) columnar jointing in the Columnar Peak dacite, (g) Atwell Peak (back left) and Columnar Peak (front), (h) Opal Cone (red) and the Ring Creek flow (yellow). 11 andesite flows from Clinker Peak at 100 ka (Fig. 3b). Activity at Cinder Cone began post 100 ka (Fig. 3c) with the formation of a tuff ring and the eruption of the basaltic andesite of Desolation Valley, followed by the Helm Creek composite flow. The Table formed approximately at 100 ka, when hornblende andesite magma erupted beneath the Cordilleran Ice Sheet and melted its way upwards to form a steep, flat topped tuya (Fig. 3d; Mathews 1951, Green 1981).

The olivine-bearing Cheakamus Valley Basalts were erupted from an unknown vent post 100 ka and eruptions continued episodically to approximately 34 ka

(Fig. 3e; Green et al. 1988).

The MGVF is comprised of Mt. Garibaldi and its subsidiary vents, Dalton Dome and Atwell Peak, Opal Cone and the Ring Creek andesite flow, Columnar Peak and the andesite flows of Paul Ridge. Recent activity in this field began at approximately 700 ka with the eruption of hornblende andesite flows atop pre 1.3

Ma hornblende-andesite and basaltic andesite along Paul Ridge. Increased volcanism occurred between 260 ka and 220 ka with the eruption of hornblende orthopyroxene dacite at Columnar Peak (Fig. 3f), followed by hornblende orthopyroxene dacite flows and minor pyroclastic material from Mt. Garibaldi

(Green et al. 1988). The composite dacite cone of Dalton Dome formed later but before the belt was overridden by glacial ice. Post 100 ka, dacitic pyroclastic flows were erupted from Atwell Peak atop the glacial ice as well as additional dacite flows from Dalton Dome. These flows and the west flank of Atwell Peak collapsed following glacial retreat (Fig. 3g; Mathews 1952, 1958; Green 1990). 12

The most recent volcanism in the GVC was the eruption of the Ring Creek andesite flow from Opal Cone between 10.7 ka and 9.3 ka (Fig. 3h; Brooks and

Friele 1992), which extends for some 17.5 km south around Paul Ridge and then west towards Squamish River (Fig. 2).

2.2 History and definition of adakites

Adakites are a group of intermediate to felsic igneous rocks formed in subduction zones involving relatively young, hot oceanic lithosphere (≤ 25 Ma).

Adakite is named after a suite of magnesian andesite rocks first described by

Kay (1978) from Adak Island in the Aleutian Arc. These rocks are believed to be the result of partial melting of subducted ocean crust, creating typically sodic

‘slab melts’. Based on the work done by Kay (1978), Defant and Drummond

(1990) described the mineralogy and geochemistry of adakites more fully, as well as field relationships and suggesting a petrogenetic model for adakite genesis. Defant and Drummond (1990) argued that because adakites are most commonly found in subduction zones that are young and hot, the subducted slab still retains much of its heat and melts at shallower levels than older and colder slabs, rather than dehydrate (or a combination of both). The partial melts remove most or all of the plagioclase from the slab, resulting in the high Sr contents of adakites as Sr strongly partitions into plagioclase (Rollinson 1993) and leaves garnet as a residual phase, which strongly partitions Y and HREE. This process generates the distinctive high Sr/Y and La/Yb ratios of adakite magmas. Defant and Drummond (1990) developed the definition of adakite as volcanic or 13 intrusive arc rocks that contain ≥ 56 wt. % SiO2, ≥ 15 wt. % Al2O3, usually < 3 wt.

% MgO (but can be up to 6 wt. %), ≤ 18 ppm Y, ≤ 1.9 ppm Yb and Sr in excess of 400 ppm. The mineralogy of adakites as described by Defant and Drummond

(1990) includes plagioclase and amphibole as the most common phases, with clinopyroxene and orthopyroxene occurring frequently with some biotite and opaques as well.

Following the work by Defant and Drummond (1990), adakites have been identified and described at other “hot” subduction zones including: Panama

(Defant et al. 1991), Mt. St. Helens, U.S.A. (Dawes et al. 1994), southern Chile and Argentina (Stern and Kilian 1996), and Ecuador (Samaniego et al. 2002), suggesting that slab melting may be a more common process than once thought. A study by Martin et al. (2005) expanded on the adakite definition by analysing a geochemical database of > 340 samples (previously compiled by

Martin and Moyen 2003) and found that two groups were present, a low silica adakite (LSA) and a high silica adakite (HSA). LSA contain < 60 wt. % SiO2, higher MgO (4 - 9 wt. %), CaO + Na2O > 10 wt. %, Sr > 1000 ppm, generally higher LREE and are Rb poor. HSA, by contrast, have SiO2 contents > 60 wt. %, lower MgO (0.5 - 4 wt. %), CaO + Na2O < 11 wt. %, Sr < 1100 ppm and generally lower HFSE and HREE than for LSA. Both groups display a low

# K2O/Na2O ratio (~ 0.42), high Mg (~ 51) and high Ni and Cr contents (24 and 36 ppm, respectively). Mineralogically, LSA differs from HSA in that LSA contains pyroxene phenocrysts (Martin et al. 2005). Recognition of these distinct groups 14 led to the idea of different sources for magma genesis to explain the geochemical characteristics of each; Martin et al. (2005) stated that HSA magmas were direct partial melts of subducted basaltic crust that were subsequently altered by interaction with mantle peridotite, as the melts migrated upwards through the mantle wedge. This interaction increased the Mg#, Ni and

Cr contents, but left the adakitic signature (high Sr/Y, La/Yb, etc.) relatively unchanged. Hence, the source of LSA magmas is not subducting crust, but the overlying mantle wedge that has been metasomatised by the passage of previous HSA melts (Martin et al. 2005). Partial melting of this altered mantle wedge gives LSA their higher MgO, TiO2 and lower SiO2 contents than for HSA magmas. Similarly, the higher Nb and Ta contents of LSA are transferred to the mantle wedge by the HSA melts; melt is a more effective method of transferring these elements out of the subducted slab than aqueous fluids (Tatsumi et al.

1986, Martin et al. 2005) and results in the higher amounts of these elements in

LSA versus HSA (on average).

Partial melting of basalt to generate adakite is supported by experimental work

(e.g., Rapp et al. 1999) and evidenced in natural rocks from subduction zones

(e.g., Schiano et al. 1995) and, as studies into magma genesis continued, an ever widening group of different models were suggested. Consequently, a wide range of magma compositions from different tectonic environments have been classified as adakites. Models that explain the formation of adakite falls into two main groups: genesis from slab partial melts and magmas generated by various 15 methods that can reproduce the distinctive adakite chemistry (which are not necessarily subduction related). Further ambiguity arises when a suite of rocks are classified as adakites (which may not follow the parameters outlined by

Defant and Drummond 1990) and are subsequently found not to be related to slab melting. This leads to the conclusion that adakites are not slab melts and has provided the means of pooling many different rock types under the one name. Defant and Drummond (1990) first introduced the term ‘adakite’ and the criterion by which they are defined. Hence the GVC adakites are evaluated based upon this original definition and the adakite subdivisions of Martin et al.

(2005).

2.3 Adakite – tonalite-trondhjemite-granodiorite (TTG) association

Tonalite-trondhjemite-granodiorite (TTG) sequences comprise a large component of Archaean greenstone belts and were first introduced by Jahn et al.

(1981). TTG are sodic plutonic rocks with several geochemical characteristics that are very similar to adakites including high La/YbN (> 15), 20 < Sr/Y < 200, high Ni and Cr (14 and 29 ppm respectively), low K2O/Na2O (0.3-0.6) and low Yb

(< 2 ppm; Martin et al. 2005, Moyen and Martin 2012). The high La/Yb and Sr/Y ratios of TTG magmas were interpreted to originate from a source that was plagioclase poor but garnet rich, and partial melting of hydrated metabasalt (with residual garnet) became the generally accepted model for TTG petrogenesis

(Moyen and Martin 2012). The identification of the similar high La/Yb and Sr/Y values in adakites led to studies comparing them with TTG rocks; the 16 observation that adakite magmas are generally rare and occur in predominantly hotter subduction zones (as per the original definition by Defant and Drummond

1990) suggested that perhaps subduction was occurring in the Archaean and the source of TTG magmas was partial melts of subducted basaltic crust (Moyen and Martin 2012). The prevalence of TTG rocks compared to modern adakites is attributed to the higher geothermal gradient in the Archaean, where slab melting was likely easier and more common than in most present day subduction environments which are cooler.

Martin et al. (2005) found a relationship between Archaean TTG’s and sanukitoids and the HSA and LSA groups. TTG prior to approximately 3.3 Ga

# are not adakitic and exhibit much higher SiO2 and lower Mg , MgO and Sr contents. Post 3.3 Ga TTG, however, are geochemically very similar to modern

HSA magmas with depleted HREE, higher Mg# (up to 65), Ni and Cr contents up to 70 and 200 ppm respectively, and Sr up to 1200 ppm. LSA magmas have no

TTG analogue, but exhibit similar geochemical traits to sanukitoids, first introduced by Shirey and Hanson (1984). Sanukitoids show high LILE and fractionated REE with low Y and Yb concentrations, which are similar to TTG and adakite but have lower silica (< 60 wt. %), higher Mg# (>> 62), MgO (> 6 wt.

%) and K2O (up to 4.5 wt. %), characteristic of modern LSA magmas (Martin et al. 2005). The progression of non-adakite-like TTG in the Paleoarchaean to the

LSA-like sanukitoids in the Late Archaean was interpreted by Martin et al. (2005) to reflect a petrogenetic shift that changed the source region. Pre 3.3 Ga, the 17 geothermal gradient was such that basalt melting occurred at a much shallower depth and oceanic crust was much more buoyant, resulting in a very low subduction angle that precluded the formation of a mantle wedge for the TTG magmas to interact with. Post 3.3 Ga, the Earth (and ocean crust) had cooled sufficiently to facilitate steeper subduction and basalt melting at greater depths; this allowed for a mantle wedge to form and subsequent TTG melts interacted

# with the wedge to generate the lower SiO2 and higher Mg chemistry akin to modern HSA. In the Neoarchaean, further cooling had reduced the efficiency of slab melting and thereafter the majority of melts produced were consumed by slab melt/peridotite reactions (Rapp et al. 1999, Martin et al. 2005), prior to emplacement in the crust. Partial melting of this slab melt metasomatised mantle peridotite resulted in magmas that had much lower silica and higher LILE, MgO,

Ni and Cr than Mesoarchaean TTG, very similar to LSA lavas (Martin et al.

2005).

The driving force behind comparing adakites and TTG is that the geochemical similarities suggest magma genesis by a similar mechanism. The relative abundance of TTG in the Archaean with the rare occurrence of true adakites (as defined by Defant and Drummond 1990) that are only found in hot subduction zones led to the interpretation that partial melting of subducted basaltic crust was the method by which both magma types were formed. A review paper by

Moyen and Martin (2012) on TTG research noted some key points when considering adakites as analogous to TTG, the most important of which is that 18

TTG are plutonic rocks and adakites are volcanic. As such, they likely should not be considered equivalent because different processes can change the geochemistry of volcanic rocks (e.g., differentiation/fractionation, magma mixing) as compared to plutons (e.g., assimilation, H2O), despite having a similar source region (Moyen and Martin 2012). In the literature, TTG have had similar problems to adakites with the term being misused, where almost any somewhat sodic Archaean pluton is identified as ‘TTG’, many of which exhibit few of the original geochemical characteristics that define TTG (Jahn et al. 1981, Moyen and Martin 2012). Hence, caution is needed when classifying a suite as either

‘TTG’ or ‘adakite’.

2.4 Non-slab melt models for adakite genesis

In tectonic regimes where there is no subduction or in subduction zones where the slab is too cold to facilitate melting, the adakite signature needs to be explained by a model that does not require slab partial melts. Several models have been suggested to reproduce the adakite chemistry, the most common of which include partial melting of mafic lower continental crust (Guo et al. 2006,

Huang and He 2010), fractional crystallisation of basaltic magma containing garnet (Macpherson et al. 2006, Coldwell et al. 2011) and high pressure assimilation, fractional crystallisation (AFC) processes of mantle-derived melts and magma mixing (Castillo et al. 1999, Chiaradia et al. 2009). Potassic, continental type adakites (K- and I-type adakites) have also been suggested, originating from crustal thickening (e.g., Xiao et al. 2007). The commonality of all 19 of these models is that the source region is one where garnet is stable but plagioclase is not. This implies that ‘adakite’ is not a petrogenetic term at all and that rocks exhibiting an adakitic signature are simply the result of specific temperature and pressure requirements being met in the source region, which can occur in multiple tectonic environments.

In the study by Defant and Drummond (1990), basalt and basaltic andesite rocks were rarely found in association with adakite. When basalt and adakite were present in the same suite, the basalts were usually quite enriched in large ion lithophile elements (LILE), approaching a shoshonitic composition. Further studies into the nature of the mafic rocks associated with adakite showed that contrary to the conclusions of Defant and Drummond (1990), the basalts were enriched in high field strength elements (HFSE) rather than LILE. NEB were first described by Sajona et al. (1993, 1994, 1996) and are defined as arc basalts that contain 7 ppm < Nb << 20 ppm, as well as enrichments in other HFSE including higher TiO2 contents (~ 1 - 2 wt. %) and low LILE/HFSE and

LREE/HFSE ratios (> 2 for both in normal arc rocks, Rollinson 1993). These enrichments result in a much smaller, if not absent, Nb-Ta anomaly on primitive mantle normalised multi-element plots; contrary to the steep negative anomaly that is characteristic of normal subduction related magmas. NEB are commonly found in adakitic suites and it was proposed by Sajona et al. (1993, 1996) that the source of NEB was mantle wedge peridotite that had been previously metasomatised by slab partial melts, suggesting a petrogenetic link between 20 adakite and NEB and reinforcing the slab melt model for adakite genesis.

However, other studies have shown that NEB and adakite are not always found in association with each other and that one can be present without the other

(e.g., Southern Volcanic Zone (SVZ), South America; Stern and Kilian 1996).

This has led to some authors suggesting that it is only due to favourable tectonic conditions that NEB and adakite erupt together and that there is no genetic link between them (Macpherson et al. 2010), as there are other ways to generate

NEB independent of slab melt contribution to the mantle wedge (e.g. mixing of

OIB and MORB components in the mantle, Castillo et al. 2007).

3. Petrography

3.1 GLVF

3.1.1 Cheakamus Valley Basalts (CVB)

Five samples were collected from the Cheakamus Dam phase of the

Cheakamus Valley Basalts (CVB) around the western edge of Daisy Lake (Fig.

2). The sampled flow is porphyritic, with 10 to 15 % total phenocryst content. The most abundant phenocryst is plagioclase (~ 10 %), followed by approximately equal amounts of olivine and augite (4 - 5 % each); phenocrysts of orthopyroxene occur rarely (1 %). Plagioclase occurs as tabular to equant crystals up to 3 mm in size. The plagioclase crystals are subhedral and infrequently form glomeroporphyritic aggregates with augite (Fig. 4a). Several textures are exhibited in plagioclase that includes sieve textured cores, 21

Figure 4: Thin section photomicrographs of the GLVF mafic rocks. (a) Ophitic texture in the CVB. (b) Plagioclase oikocryst with included augite crystals in the CVB. (c) Zoned olivine phenocrysts in the CVB. (d) Olivine phenocrysts altering to red iddingsite and flow banding in the HCBA. (e) Concentrically zoned augite phenocrysts with a sieve textured core in the HCBA. (f) Zoned augite phenocrysts and rare pipe vesicles in the HCBA. (g) Strongly resorbed augite (upper left corner) and orthopyroxene (lower left center) with primary, sub-ophitic orthopyroxene (center) in the DVBA. (h) Twinned and zoned augite phenocrysts in the DVBA. All photomicrographs in cross-polarized light. 22 resorption of grain margins, rings of melt inclusions and rare poikilitic crystals

(Fig. 4b). Augite phenocrysts are generally smaller than plagioclase, with a tabular habit of up to 2 mm in size. These grains are anhedral to subhedral, occurring both as relatively unaltered phenocrysts and as strongly embayed and resorbed crystals. Ophitic and subophitic are common textures between the augite and plagioclase phenocrysts, with several smaller plagioclase laths included within the augites (Fig. 4a). Olivine phenocrysts are of a similar size to augite, approximately 1 - 2 mm. The crystals are anhedral to subhedral and some exhibit poikilitic textures with plagioclase (Fig. 4b). Olivine also rarely forms glomeroporphyritic aggregates with plagioclase and augite. Rare, weak normal zoning is present in some olivine phenocrysts, that is best observed near grain margins (Fig. 4c); some crystals also exhibit alteration to reddish orange iddingsite along grain boundaries. Orthopyroxene phenocrysts are ~ 1 mm in size and equant to prismatic in form. They include both anhedral to euhedral varieties that occur as individual crystals as well as aggregates with plagioclase and augite. Orthopyroxene locally exhibits ophitic and subophitic textures, and some grains show evidence of resorption at their crystal margins (Fig. 4a, bottom right). The groundmass to the CVB samples is weakly vesiculated (< 1

%), chiefly comprising microcrystalline plagioclase, augite and orthopyroxene, forming an intergranular texture. Lesser amounts of olivine and opaque oxides are also present. Where present, vesicles are spherical to slightly elongated, ~

0.5 to 1 mm in diameter. 23

Mineral compositions for olivine, clinopyroxene, plagioclase and oxide phases, as determined by electron microprobe, are presented in Tables C1, C2, C4, and

C5 (Appendix C). Forsterite contents show moderate variation, ranging from Fo78 to Fo68 (Table C1). Where present, zoning is weakly normal with only a 1-2 % change in Fo (e.g., grain analysis 21-4, Table C1). Figure 5a illustrates typical rim (open circles) and core (filled circles) compositions in olivine. Olivine phenocrysts host appreciable Ni (up to 0.14 oxide wt. %). A single grain of clinopyroxene was analysed; the phenocryst contains ~ 1 wt. % TiO2 (Table C2) and is slightly more Mg-rich (En45 – Wo41). On the pyroxene quadrilateral, the crystal (blue square) plots as a diopsidic augite (Fig. 5a). The determined compositions of plagioclase phenocrysts are typically more felsic than would be expected for such a primitive basalt (An63-65). Opaque oxides are chromian magnetite, with Cr2O3 up to 23 wt. % (Table C5). These grains also contain appreciable TiO2 (7.7 - 8.5 wt. % TiO2), Al2O3 (6.2 - 7.5 wt. %) and MgO (up to

5.5 wt. %).

3.1.2 Helm Creek Basaltic Andesite (HCBA)

Two samples (10JF013, 10JF024) were collected from the centre portion of the

HCBA flow (Fig. 2). This flow is porphyritic with ~ 10 to 15 % phenocrysts of roughly equal amounts of olivine and augite. Olivine crystals are anhedral and are up to 1 mm in size. Several, but not all, of the olivine phenocrysts are altered with pervasive reddish orange iddingsite occurring along fractures and at grain margins (Fig. 4d). Fine grained opaque phases also occur along the grain 24

Figure 5: Mineral compositions as determined by electron microprobe for (a) CVB and (b) HCBA. Aug=augite, Di=diopside, Hd=hedenbergite. See text for explanation. 25

Figure 5 cont’d: Mineral compositions as determined by electron microprobe for (c) Desolation Valley and (d) Paul Ridge andesite. Abbreviations as in Fig. 5 a & b. See text for explanation. 26 margins. Some olivine crystals also exhibit resorption and are embayed; opaque inclusions are common. Rare, small glomeroporphyritic aggregates with augite are present, up to 2.5 mm in size. Augite phenocrysts are slightly larger, ~ 2 mm in size with rare larger phenocrysts up to 4 mm. The augite is euhedral to subhedral and equant to prismatic in form. Simple twins are present in most crystals as well as various patterns of zonation that include concentric zoning

(most common) and sector zoning. The larger augite crystals commonly exhibit a distinct core mantled by a compositionally distinct, but uniform, rim and more rarely have sieve textured cores (Fig. 4e). Augite also occurs as aggregates that locally hosts inclusions of unaltered olivine. The groundmass to the Helm Creek flow is trachytic, with several domains exhibiting different orientations. Strong flow banding around the phenocrysts is also present. It is also weakly vesicular, comprising mainly acicular plagioclase with lesser amounts of fine grained augite, opaque oxide phases and brown glass. Vesicles (~ 1 %) are commonly rounded to irregularly shaped, some 1 - 1.5 mm in diameter. Rare pipe vesicles are also present, approximately 2 - 3 mm in length (Fig. 4f).

Mineral compositions for olivine, clinopyroxene, plagioclase and oxide phases as determined by electron microprobe are presented in Tables C1, C2, C4 and C5

(Appendix C). The HCBA olivine phenocrysts are quite unique in that they are near primitive (Fo86-89; Fig. 5b) despite being hosted in a basaltic andesite flow.

They show very little to no compositional zoning from core to rim and most fresh crystals appear to be in equilibrium with the surrounding melt. Clinopyroxene is 27 again titaniferous, with up to 1.3 wt. % TiO2 (Table C2). Several crystals are reversely zoned, with SiO2 decreasing from core to rim, but the En component remains relatively unchanged or exhibits normal zoning. Clinopyroxene phenocrysts are slightly more felsic than groundmass crystals (SiO2 of ~ 50 wt.

% versus ~ 48 wt. %; Table C2). Cr values span a wide range among the analysed crystals, from below detection limits to as high as 0.9 wt. % Cr2O3. The core compositions of the clinopyroxenes are augitic and trend towards diopside

(Fig. 5b). Groundmass clinopyroxene (filled diamonds) straddle the diopside- augite boundary. Plagioclase compositions also exhibit a wide range, from An6 up to An56 (Table C4). BaO exhibits a general positive correlation with silica

(Table C4). Oxides are predominantly magnetite with a variable ulvöspinel component (TiO2 up to 12 wt. %; Table C5). Cr contents are well below 0.5 wt.

%, with the exception of one chromian magnetite crystal that contains over 9 wt.

% Cr2O3 (24-6, Table C5). There is appreciable Al2O3 in the magnetite grains, as well as MgO (up to 5.5 and up to 14.5 wt. % respectively, Table C5); however, these do not always correlate with each other.

3.1.3 Desolation Valley basaltic andesite (DVBA)

One sample (10JF025) was collected from the flow terminus of the DVBA (Fig.

2). This flow is porphyritic, with ~ 20 % total phenocrysts, the most abundant of which is plagioclase (~ 10 %), followed by olivine (~ 5 %) and roughly equal amounts of orthopyroxene and augite (totaling ~ 5 %). Plagioclase crystals are tabular to bladed, ~ 2 mm in size; it occurs both as discrete grains that display 28 oscillatory zoning, and as glomeroporphyritic aggregates with orthopyroxene and augite. Some phenocrysts also exhibit resorption along grain margins. Olivine occurs as anhedral grains of ~ 1 mm in diameter. The phenocrysts are strongly resorbed and embayed, with some weak alteration to iddingsite along fractures.

Orthopyroxene forms equant to prismatic crystals of up to 2 mm. The phenocrysts occur in two populations: strongly resorbed and embayed crystals, and rare euhedral crystals. Euhedral orthopyroxene occur rarely as subophitic grains with plagioclase (Fig. 4g) and on occasion, display weak oscillatory zoning (Fig. 4f), whereas resorbed orthopyroxene commonly host melt inclusions. Augite phenocrysts are typically smaller than plagioclase and orthopyroxene, up to 1 mm in size. Augite may form equant subhedral to anhedral crystals that are rarely sector zoned, and more commonly resorbed and embayed along grain margins. Few of the augite crystals appear pristine; those that are exhibit simple twins and a prismatic habit. The groundmass of the

DVBA has no evident fabric but weak flow banding is observed locally around phenocrysts (Fig. 4g). The groundmass is microcrystalline with microlites of bladed to acicular plagioclase and anhedral orthopyroxene with lesser amounts of interstitial opaque oxides.

Mineral compositions for olivine, clinopyroxene, orthopyroxene, plagioclase and oxide phases as determined by electron microprobe are presented in Tables C1-

C5 (Appendix C). The forsterite content of the olivine phenocrysts span a small range, from Fo73 to Fo79; the rim and core compositions are slightly normally 29 zoned and overlap with core compositions (Table C1, Fig. 5c). SiO2 contents decrease slightly from core to rim in the olivine grains, as does the NiO; the CaO and MnO values increase towards the rim. Olivine lacks any significant TiO2 or

Cr2O3 and the phenocrysts are relatively inclusion free. Clinopyroxene shows slight normal zoning as well, the rim composition is more Ca and Fe rich than the core (Table C2) and when plotted on the pyroxene quadrilateral the core is slightly richer in the augite component than the rim (Fig. 5c). TiO2 also increases towards the rim; sector zoned crystals differ in their Ti contents. Cr2O3 decreases significantly in the rim while MnO and Na2O both increase. Both phenocrystic and groundmass orthopyroxene have been analysed; phenocrysts are more mafic than the groundmass crystals with enstatite contents of En73, compared to

En68 in the groundmass (Table C3, Fig. 5c). The groundmass orthopyroxene is also considerably richer in Al2O3 and Na2O, though CaO, MnO and NiO do not differ significantly between phenocrysts and groundmass crystals. Plagioclase compositions cover a relatively wide range of compositions, ranging from An65 to

An42 (Table C4). Oxide phases are mainly magnetite with a significant ulvöspinel component, up to 9.7 wt. % TiO2 (Table C5). These are also chromian, ranging from 0.95 - 2.3 wt. % Cr2O3. Al2O3 and MgO contents are similar for all phenocrysts, between 2 - 3 wt. %.

3.1.4 Barrier andesite

Four samples (09JF004, 09JF005, 09JF006, 09JF012) were collected from the

Barrier andesite lava flow along the northern shore of Garibaldi Lake (Fig. 2). 30

This flow is porphyritic, with 10 to 15 % phenocrysts. Plagioclase is the most abundant (~ 10 %), followed by approximately equal amounts of hornblende and biotite (2 - 3 % each). Quartz phenocrysts occur rarely (< 1 %). Plagioclase phenocrysts (up to 2 mm) form subhedral to anhedral equant to tabular crystals that display complex zonation. Several grains contain concentric trails of melt inclusions along their outer margins; sieve textures are also present. Hornblende phenocrysts are smaller than plagioclase, ~ 1 mm in size and prismatic to bladed in habit. The crystal edges are diffuse and commonly show alteration to fine grained opaque minerals, possibly indicative of disequilibrium. Hornblende also occurs as rare glomeroporphyritic aggregates (Fig. 6a). Biotite is the largest phenocryst phase; up to 4 mm in size. The phenocrysts are equant, subhedral and strongly pleochroic. The crystal margins are extensively altered and/or heavily corroded; resulting in a rim opaque phases (Fig. 6b). Quartz phenocrysts are anhedral, equant and ~ 1 mm in diameter. Where these occur, quartz crystals are surrounded by a reaction rim of fine-grained, tabular, augite.

The groundmass of the Barrier andesite is comprised principally of brown glass that comprises roughly equal amounts of crystallites and microlites of plagioclase and rare hornblende. Rounded and irregularly shaped vesicles (~ 2

%) also occur in the groundmass. Flow banding is locally present around the larger phenocrysts.

3.1.5 Black Tusk

Figure 6: Thin section photomicrographs of the GLVF and the MGVF adakite rocks. (a) Strongly altered biotite crystal with a rim of oxide phases in the Barrier andesite, cross-polarized light. (b) Glomeroporphyritic hornblende in the Barrier andesite, plane light. (c) Strongly resorbed orthopyroxene phenocryst in the Black Tusk andesite, cross-polarized light. (d) Coarser grained crystal clot in the Black Tusk andesite, cross-polarized light. (e) Biotite xenocryst from the proximal Ring Creek andesite, cross-polarized light. (f) Hornblende phenocrysts from the proximal Ring Creek andesite; note the dark fine grained reaction rims along the crystal margins, plane light.

Figure 6 cont’d: Photomicrographs of the MGVF adakite rocks. (g) Mafic to intermediate xenolith comprised of plagioclase and orthopyroxene in the distal Ring Creek andesite, plane light. (h) Quartz xenocryst with a reaction rim of fine grained augite; weak flow banding is also present, cross-polarized light. (i) General texture observed (cross-polarized light) and (j) concentrically zoned hornblende phenocrysts present within the Columnar Peak dacite, plane light. (k) Plagioclase rich nature of the Paul Ridge andesite, with rounded olivine (lower right), cross-polarized light. (l) Olivine (second order blue) being replaced by fibrous orthopyroxene (yellow) in the Paul Ridge andesite, cross-polarized light.

Two samples (10JF015, 10JF016) of orthopyroxene andesite were collected from the west bluff of Black Tusk (Fig. 2). The andesite is essentially aphyric, with rare (~ 1 %) phenocrysts of plagioclase and orthopyroxene. The plagioclase phenocrysts are ~ 1 - 2 mm in size and are elongate to acicular in form. The crystals are subhedral and fractured but are relatively clear. Orthopyroxene phenocrysts (~ 1 - 2 mm) are equant, strongly fractured and in some examples, exhibit resorption and disaggregation into the surrounding melt (Fig. 6c). Rare sieve textured grains also occur as well as slightly coarser grained crystal clots

(Fig. 6d). The groundmass of the Black Tusk andesite comprises approximately equal amounts of acicular, plagioclase microlites and brown glass. Strong flow banding predominates in this unit.

3.2 MGVF

3.2.1 Ring Creek andesite

Four samples (09JF007, 09JF008, 10JF022, and 10JF023) were collected from the Ring Creek andesite; two taken proximal to Opal Cone (09JF007, 09JF008) and two taken approximately 2 km from the flow terminus (10JF022, 10JF023; see Fig. 2). The mineralogy of the proximal Ring Creek andesite is different to that of the distal portion of the flow (first noted by Sivertz 1976) and hence will be described separately.

3.2.1.1 Proximal Ring Creek andesite 34

The andesite is porphyritic; main phenocrysts are plagioclase (15 %), followed by hornblende (10 %) and augite (2 %). Quartz occurs in trace amounts (0.1 %).

Biotite (3 %) is present as an alteration product of hornblende and as rare large xenocrysts (Fig. 6e). Plagioclase occurs in two size populations, the larger phenocrysts are ~ 2 mm in size and the smaller less than 1 mm. The majority of the plagioclase crystals are subhedral, equant to tabular; more rarely these form glomeroporphyritic aggregates. Several features are exhibited in plagioclase that includes sieve textures, resorption of grain margins and in some of the larger crystals, seritization. However, an equal proportion of plagioclase phenocrysts are inclusion free and ‘pristine’. Hornblende phenocrysts are second to plagioclase in abundance and range in size from less than 0.5 mm up to 1 mm.

The majority of the crystals exhibit various disequilibrium textures including fibrous cores of clinopyroxene and alteration to opaque oxides along the crystal rims (Fig. 6f). Biotite occurs as subhedral xenocrysts up to 1 mm in size with rare crystals of up to 3 mm. The edges of biotite crystals are diffuse and poorly defined with rims showing alteration to a mass of fine grained opaque minerals.

The larger biotite xenocrysts are heavily embayed and have sieve textured cores. With few exceptions these xenocrysts display extensive replacement by opaque oxide phases; resorption of grain boundaries is also common. Augite occurs as prismatic to equant crystals up to 3.5 mm in size. The margins of the phenocrysts are resorbed and contain abundant inclusions of apatite and oxides.

The crystals are greenish brown and not distinctly pleochroic. Only a few quartz 35 crystals have been identified in this part of the flow. The phenocrysts are anhedral and less than 0.5 mm in size and exhibit resorption along the grain margins. The groundmass of the proximal Ring Creek andesite is approximately equal parts crystallites and brown glass. Plagioclase, augite altered hornblende and oxide phases comprise the majority of the crystallites. Local flow banding is evident around the larger phenocrysts.

3.2.1.2 Distal Ring Creek andesite

The andesite is porphyritic, slightly less than the proximal portion with ~ 20 % total phenocrysts. The mineralogy of the distal part of the flow differs from the proximal portion in that the only phenocryst phases present are plagioclase and augite. Xenoliths of mafic-intermediate cumulate inclusions that host orthopyroxene occur rarely. The inclusions are heavily corroded and partially melted (Fig. 6g). Plagioclase, again occurring in two size populations, is the most abundant phenocryst (15 %), followed by augite (5 %). The larger plagioclase crystals are up to 3.5 mm in size, the smaller approximately 1 mm; all crystals are equant to tabular and subhedral, and display complex zonation and various degrees of resorption. Sieve textured crystals are less common than the plagioclase in the proximal portion of the flow but occur mainly in the larger grains that also contain concentric trails of melt inclusions along their margins.

Augite phenocrysts are smaller than plagioclase, commonly less than 1 mm in size. The crystals are equant and subhedral to euhedral; some grains display simple twinning as well as glomeroporphyritic aggregates with plagioclase. Rare, 36 altered orthopyroxene crystals, likely derived from the mafic-intermediate cumulate xenoliths, up to 2 mm in size are present but have been almost completely altered to chlorite and opaque oxides. Rare quartz is also present occurring as anhedral crystals that appear in disequilibrium with the surrounding melt, in exhibiting reaction rims of fine-grained, radiating clusters of augite (Fig.

6h). This is in contrast to the proximal portion of the flow where the quartz appears to be primary. The groundmass is predominantly crystallites of plagioclase with lesser amounts of augite and brown glass. The groundmass displays local, weakly developed flow foliation.

3.2.2 Columnar Peak dacite

Four samples from the orthopyroxene-hornblende dacite of Columnar Peak

(09JF009, 09JF010, 10JF019 and 10JF020) were examined as part of this study

(Fig. 2). This unit is described as a series of flows by Green (1977). The dacite is porphyritic with 10 to 15 % phenocrysts (Fig. 6i). Plagioclase is the most abundant phase (~ 7 %), followed by hornblende (~ 6 %) and orthopyroxene (< 1

%). Plagioclase occurs in two size populations; the larger crystals are up to 4.5 mm in size and the smaller crystals are ~ 1.5 mm. The phenocrysts are subhedral to anhedral and are complexly zoned. Several disequilibrium textures are observed in the large plagioclase phenocrysts including sieve textured cores, resorption along crystal edges and several generations of concentric melt inclusions. These textures are less common in the smaller plagioclase crystals.

Some of the larger grains exhibit fresh, clear overgrowths on the outer edges. 37

Hornblende is considerably smaller than plagioclase, up to 1 mm in size and is subhedral to euhedral and equant to prismatic. The crystals are in disequilibrium with the surrounding melt; the phenocrysts are commonly embayed with some grains exhibiting sieve textured cores. All the crystals have resorbed margins through disequilibria reaction rims comprised of opaque oxide phases. Weak concentric zoning is also observed in some of the larger crystals (Fig. 6j).

Orthopyroxene crystals are quite small, less than 1 mm in size. The phenocrysts are yellowish, equant to prismatic and mainly euhedral in form. The majority of the orthopyroxene grains are relatively fresh, with occasional large melt inclusions. Some crystals occur as aggregates or are perhaps fragments of a single crystal separated by glass. The groundmass is comprised of roughly equal amounts of brown glass and crystallites of plagioclase and opaque oxides.

Flow foliation is observed locally.

3.2.3 Paul Ridge andesite

Three samples of the Paul Ridge andesite (09JF011, 10JF017, 10JF018) were collected in this study (Fig. 2). The andesite is described as a series of hornblende andesite flows by Green (1977); these are porphyritic, with ~ 20 % phenocrysts (Fig. 6k). Plagioclase phenocrysts are the most common (10 %), followed by orthopyroxene and olivine (5 % each, marginally higher amounts in

09JF011) with trace amounts of xenocrystic quartz (< 0.5 %). The plagioclase crystals are tabular and can also form aggregates of up to 3 mm in size. The dominant crystal size is approximately 2 mm. Plagioclase is subhedral to 38 euhedral and is complexly zoned with reverse and oscillatory zoning. Some of the phenocrysts exhibit uneven extinction, suggestive of strain deformation. The crystals display varying degrees of sieve texture in the cores and resorption along the edges. Some of the phenocrysts have clear cores with concentric rings of melt inclusions near their margins. Orthopyroxene phenocrysts are markedly less abundant than plagioclase, occurring mainly in the glassy groundmass with some larger crystals. The relatively pristine orthopyroxene phenocrysts are up to

1 mm in size, green, broken and exhibits reverse zoning. Orthopyroxene also commonly forms reaction rims on olivine and as an interstitial phase to crystal aggregates. The majority of olivine crystals occur in a state of incipient replacement by orthopyroxene (Fig. 6l); unaltered olivine is rare. These replacement textures are slightly more common in 09JF011 than the other samples. Larger olivine crystals occur occasionally in aggregates with plagioclase; the olivine is anhedral and usually exhibits alteration to brown iddingsite along crystal edges and fractures. Embayments are also common in the crystals. A few quartz xenocrysts have been identified; the quartz is anhedral, usually exhibiting weakly uneven extinction. The quartz xenocrysts exhibit reaction rims and are surrounded by small, outwardly radiating augite crystals. The groundmass is mainly brown glass with microlites of plagioclase and opaques.

The compositions of the main mineral phases from two samples of the Paul

Ridge andesite (10JF017 and 10JF018) were determined by electron 39 microprobe and include olivine, orthopyroxene, plagioclase and oxides; these values are listed in Tables C1, C2, C4 and C5 (Appendix C). Olivine compositions between the samples are quite different; phenocrysts from

10JF018 are slightly normally zoned with more mafic olivine cores (~ Fo75) than

10JF017 (~ Fo65; Fig. 5d), but SiO2 contents of 10JF017 are lower than that of

10JF018 (Table C1). Orthopyroxene compositions between both samples are generally similar with cores of En64-68, but 10JF018 exhibits strong reverse zoning with rim compositions of up to En76 (Table C2, Fig. 5d). The Mg-Fe contents of the olivine and orthopyroxene crystals in all samples probed span very similar ranges (Fig. 5d). Reverse zoning is also observed for the TiO2 and

Cr2O3 contents in orthopyroxene; SiO2 remains relatively constant from core to rim. Plagioclase also varies widely between samples; anorthite contents of

10JF018 are higher and span a much narrower range than the phenocrysts in

10JF017 (An66-72 versus An46-66 respectively, Table C4). With the exception of one crystal (grain analysis 17-3, Table C4), SiO2 values of 10JF017 are only slightly higher than 10JF018 (~ 51 wt. % versus ~ 50 wt. %). Oxide phases in

10JF018 are magnetite and ulvöspinel (Table C5), whereas 10JF017 contains mainly magnetite. The magnetite grains in 10JF017 are more enriched in Cr than

10JF018 and have higher average MgO and Al2O3.

4. Geochemistry 40

4.1 Analytical techniques

4.1.1 Whole Rock Geochemistry

Nine samples from the MGVF and nine from the GLVF have been analysed for whole-rock major, trace and rare earth element (REE) concentrations. All samples were crushed in a jaw crusher and then powdered to less than 74 µm in a tungsten mill at the Dept. of Geology, University of Regina. All elements were measured by the Trace Element Analysis Laboratories at McGill University,

Montréal, Canada. Data are presented in Tables A1 and A2 (Appendix A). Major elements and some trace elements, which include Ba, Ce, Cr, Cu, Ni, V, Sc and

Zn were determined on a Philips PW 2440 X-ray fluorescence spectrometer instrument, with samples prepared as fused glass discs. LOI was determined by incrementally heating 4 g of rock powder in air to 400°C for 60 minutes, 800°C for another 60 minutes, then to 1000°C for 180 minutes. The product was weighed after cooling in a desiccator to calculate LOI. Other trace elements (Ga,

Nb, Rb, Sr, Th, U, Y, and Zr) were analysed on pressed powder pellets. Two internal standards of known composition were sent with our samples to test the accuracy and reproducibility of the results. The accuracy for silica is within 0.5

%. For other major elements it is within 1 % of the element analysed. For trace elements as well, the accuracy is within 1 %, as determined from replicate analyses of internal standards. The limiting factor for accuracy is the degree of scatter of analyses from which the consensus values are determined

(Govindaraju et al. 1994, p. 256). Instrument precision is within 0.12 % relative. 41

This is the percent relative variation obtained when the same sample is analysed repeatedly for the same element. Overall precision for glass discs and pressed pellets is within 0.5 % relative. This is determined by repeatedly analysing two glass discs or pressed pellets prepared from two different aliquots removed from the same sample powder vial during the same day and used to prepare a fused disc or pressed pellet by an experienced operator using routine procedures.

REE were measured by inductively coupled plasma-mass spectrometry (ICP-

MS) using a borate fusion decomposition method (after Panteeva et al. 2003).

The analyses of the REE were performed on solutions using a

PerkinElmer/SciEx Elan 6100 DRCplus ICP-MS. Samples (0.4 g, corrected for

LOI) were fused using a lithium metaborate mixture and then dissolved into nitric acid and diluted. Standards and calibration solutions were prepared from fusion blanks. Oxide corrections on the middle and heavy REE were made offline using oxide production rates determined daily from single REE standard solutions.

Rock-sample detection limits (based upon three times the background standard deviation) are 10 ppb for La through Pr, and 5 ppb for Nd through Lu. A set of three internal laboratory reference materials were fused and run with each batch of samples to evaluate long term precision. Precision was additionally evaluated through repeat measurements of samples, including repeat fusions and dilution; it is better than 3 % RSD in all cases. Accuracy was evaluated using a series of six standard reference materials (SRMs) that span the sample concentration range, prepared using the same procedure as the samples. Our determinations 42 agree with the accepted values for these SRMs with discrepancies of less than 5

%. Ferrous (FeO) iron was determined using an ammonium metavanadate titration.

4.1.2 Mineral Geochemistry

The major mineral phases in two samples from the MGVF and four samples from the GLVF were analysed utilising a Cameca SX-50 electron microprobe housed in the Dept. of Earth, Atmospheric and Ocean Sciences, University of

British Columbia. Approximately 3 to 5 individual crystals were analysed of each mineral phase in each thin section. Probed minerals included olivine, clinopyroxene (where present), orthopyroxene, plagioclase and opaque oxides

(Tables C1-C5, Appendix C). Where zonation was present, core and rim compositions were determined; complexly zoned crystals of olivine and pyroxene were probed multiple times from core to rim or rim to rim.

Compositions of suitable microphenocrysts in the crystalline groundmass were also determined, where they occurred. The operating conditions were: accelerating voltage 20 kV; beam current 20 nA for all minerals; normally, the beam was kept focused, giving a spot diameter in the region of 1 μm for most minerals and counting time of 30 seconds over each spot. A rastered beam covering ~ 12 μm × 12.5 μm was used for the analysis of plagioclase which may be susceptible to decay under the electron beam. Larger scale rastering was not acceptable due to spectrometer defocussing. Sodium was analysed for first followed by other major elements to minimise artefacts due to decay. Mineral 43 standards were probed at regular intervals, approximately every 10 analyses.

Diopside standards were used for the olivine and pyroxene analyses, anorthite for the plagioclase analyses and magnesium spinel for the opaque mineral analyses. Additional compositional data for the investigated mineral phases were determined by scanning electron microscopy-energy dispersive- spectrometry (SEM-EDS) analysis, utilising a Jeol JSM-6360 SEM and Noran system 7 EDS system, housed at the Electron Microbeam Facility, University of

Regina. Small scale textures and mineral-melt interactions were identified and photographed for samples throughout the GVC, and X-ray element dispersal maps were generated for crystals of interest. Mineral and glass compositions were determined but due to the semi-qualitative nature of the EDS system, this data is not discussed in the text. Backscattered scanning electron (BSE) images of element distributions and zoning are included (see Section 5.1.3).

4.2 Results

4.2.1 Basalts and Basaltic Andesites

Three samples of the CVB and one sample from the HCBA were analysed as part of this study. The geochemistry of the GLVF is well documented in several other studies (Green 1977, 1981, 1990, 2006; Green and Henderson 1984) and select data from these studies have been included for comparison in this work.

The CVB have determined compositions transitional between the alkaline and sub-alkaline fields and exhibit a narrow range of total alkali contents (Na2O + 44

K2O (wt. %) = 3.91 - 4.81; Fig. 7). The HCBA samples are predominantly alkaline in character with a wide range of alkali and silica contents (4.14 - 7.13 wt. % Na2O + K2O) that reflect two distinct geochemical trends. The most mafic samples plot along the alkaline-sub-alkaline boundary, closer to normal arc basalts while the remainder fall well inside the alkaline field and range from trachybasalt-hawaiite to basaltic trachyandesite-mugearite (Fig. 7). The CVB are characterised by decreasing MgO, Fe2O3, CaO, TiO2, Ni and Cr with increasing

SiO2; whereas K2O, Na2O, P2O5 and V increase with SiO2 (Figs. 8, 9). Al2O3 does not vary significantly with increasing silica, as does Sr. The two distinct trends identified in the HCBA data are more evident in the Harker diagrams shown in Fig. 8. The mafic samples generally have the same characteristics observed in the CVB samples; however, the more evolved samples exhibit different geochemical trends. For example, Fe2O3 contents remain relatively constant with increasing silica, whereas MgO, CaO, TiO2, K2O, Na2O and Cr all decrease. P2O5 shows a strong negative correlation with increasing SiO2, while

Al2O3, Ni and Rb positively correlate with SiO2. V contents in the HCBA are scattered and show no correlation with SiO2. Despite a normal decrease in MgO

# contents for both the CVB and HCBA, the Mg s increase with SiO2 in both groups (Figs. 8, 9).

Primitive mantle normalised spider diagrams show that the HCBA samples have stronger LILE and LREE enrichment than the CVB, as well as a prominent positive Sr anomaly and Nb-Ta and Ti depletions. The CVB exhibit only a small 45

Figure 7: Total alkalis vs. SiO2 diagram with IUGS rock designations for the GVC rocks (after Le Bas and Streckeisen 1991). Dashed line represents the alkaline-subalkaline boundary from Macdonald (1968). Key: PR=Paul Ridge andesite, RC=Ring Creek andesite, CP=Columnar Peak dacite, BT=Black Tusk andesite, BF=Barrier andesite, Table=Table andesite, Cay=Mt. Cayley, CVB=Cheakamus Valley Basalts, HCBA=Helm Creek Basaltic Andesite. Open symbols(Hist)=literature data, Mt. Cayley data from Kelman et al. (2001). 46

Figure 8: Harker diagrams illustrating variations in major element chemistry with increasing SiO2 (wt. % oxide). All data have been recalculated to 100 % on an anhydrous basis. Symbols as in Fig. 7. 47

Figure 9: Harker diagrams illustrating variations in trace element chemistry and Mg# with increasing silica for the GVC rocks. All trace elements are in ppm, silica is reported in wt. % oxide. Symbols as in Fig. 7.

Nb-Ta depletion compared to the HCBA, despite both groups having very similar

Nb concentrations (~ 8.1 - 8.5 ppm). The CVB are also enriched in Sr, less so than the HCBA, and have no Ti depletion (Fig. 10a, b). The Nb concentrations in both the CVB and the HCBA classify them as Nb-enriched basalts (NEB), first described by Sajona et al. (1993, 1994, 1996) as basalts and basaltic andesites with Nb values of 7 ppm < Nb << 20 ppm. The CVB and HCBA both exhibit additional geochemical characteristics of NEB, including higher TiO2 (~ 1 - 2 wt.

%) and little to no Zr depletion. NEB also have low LILE/HFSE and LREE/HFSE ratios; Rb/NbPM and La/NbPM are > 2 in typical arc basalts. The CVB have

Rb/NbPM and La/NbPM well below that of arc basalts (0.85 and 1.1 respectively); the HCBA has low Rb/NbPM (1.7) but higher La/NbPM than most NEB (~ 3). The

MREE and HREE contents of both the CVB and HCBA are very similar, including a small Y depletion in both groups. These characteristics are also evident in chondrite-normalised spider diagrams where HCBA rocks exhibit stronger enrichment in LREE than for the CVB samples, but similar MREE and

HREE contents and behaviour. This results in a fractionated and slightly concave upward pattern for the HCBA versus a less enriched, more linear trend for CVB (Fig. 11a, b).

4.2.2 Adakites

Nine MGVF samples were analysed; 3 samples from Paul Ridge, 2 from

Columnar Peak dacite and 4 samples of Ring Creek (2 each from the proximal and distal portions of this andesite flow). An additional 4 samples from the 49

Figure 10: Primitive mantle normalised spider diagrams for GVC samples analysed in this study (a & c) and for previously published data (b & d). Data in (b) derive from Green (2006), data in (d) derive from Green and Henderson (1984). Primitive mantle normalising values are from Sun and McDonough (1989). Symbols as in Fig. 7. 50

Figure 11: Chondrite normalised REE spider diagrams for GVC samples analysed in this study (a & c) and for previously published data (b & d). Data in (b) and (d) derive from sources listed in Fig. 10. Symbols as in Fig. 7. Chondrite normalising values from Sun and McDonough (1989). 51

Barrier andesite in the GLVF and 1 sample from Black Tusk were analysed (see

Tables A1 & A2, Appendix A). Data from Mt. Cayley (Kelman et al. 2001), located north of the GVC, also exhibits adakitic characteristics and has been included for comparison. The investigated MGVF and GLVF rocks are subalkaline in character ranging in composition from basaltic andesite to dacite, with the bulk of samples falling in the andesite field (Fig. 7). All the GVC samples define similar trends as individual centres in several major and trace element variation diagrams (Figs. 8, 9). The volcanic products from each centre are characterised by decreasing Fe2O3, CaO, MgO, TiO2, and V with increasing

SiO2. Al2O3 decreases only slightly as SiO2 increases, with alumina concentrations spanning a range of about 1.5 wt. %. Ni, Cr, and Na2O do not vary significantly with increasing SiO2, forming relatively flat trends. K2O and Rb contents increase with SiO2, while Sr values increase to approximately 60 wt. %

# SiO2 and then decrease sharply. The Mg s of all the GVC rocks exhibit a narrow range (46 to 52), except for the Paul Ridge andesite, where one sample has Mg# of 62. Primitive mantle normalised multi-element spider diagrams show that all centres have LILE enrichment and Nb-Ti negative anomalies, typical of subduction related rocks (Fig. 10c, d). The Paul Ridge andesite rocks exhibit the strongest depletion in Th and other incompatible elements and the highest

MREE-HREE abundances of all the GVC rocks. One Paul Ridge sample

(09JF011; see Tables A1 & A2, Appendix A) has higher MgO, TiO2, Ni and Cr values than the others. Chondrite normalised REE spider diagrams (Fig. 11c, d) 52 for the GVC volcanic rocks display enrichment of LREE over HREE and lack any significant Eu anomalies. The andesite rocks of Black Tusk display the lowest

LREE/HREE fractionation of all the GVC volcanic rocks with La/YbN averaging ~

5.9. The Ring Creek andesite samples exhibit the highest fractionation with

La/YbN ranging from 8.5 to 9.6. All of the GVC rocks have similar MREE/HREE fractionation with ratios of 1.5 to 1.9. On a Sr/Y versus Y diagram, all of the GVC rocks plot within the adakite field, but fall outside of this field on the La/Yb versus

Yb diagram, having values typical of normal arc-rocks (Fig. 12a, b). The HCBA has a silica content that is too low to be considered adakitic (Table A1, Appendix

A) but still exhibits several other adakitic indicators (see section 5.2.3.1) and has been included in the adakite plots. The accepted minimum value for La/Yb as an adakitic indicator ranges from as low as 8 (Drummond and Defant 1990) up to

20 and greater (Castillo et al. 1999, Richards and Kerrich 2007). The La/Yb values for the GVC have a range of ~ 9 (Paul Ridge andesite) to 13.4 (Ring

Creek andesite), plotting in the lower end of the dashed adakite field. When plotted against SiO2, both Sr/Y and La/Yb appears to increase for Paul Ridge andesite rocks in contrast to the other GVC centres, for which Sr/Y decreases with SiO2, but La/Yb increases (Fig. 12c, d). Other adakitic indices (Sr, Na2O,

Al2O3) do not show definitive trends with fractionation indices (SiO2, Ni, Cr), suggesting that the chemistry is not controlled exclusively by fractionation processes (Chiaradia 2009). The GVC samples may also be divided into the low-silica and high-silica (LSA and HSA) groups on the basis of geochemical 53

Figure 12: Plots of adakitic indices (a & b) for the GVC rocks and variations in these indices with increasing silica (c & d). Adakite and normal arc-rock fields derive from Castillo (2006) and Richards and Kerrich (2007). Trace elements are in ppm, silica in wt. % oxide. Symbols as in Fig. 7. ADR=andesite-dacite-rhyolite

54 characteristics outlined for adakitic rocks by Martin et al. (2005). On a K/Rb versus SiO2/MgO diagram, the Paul Ridge and Black Tusk andesite rocks exhibit high K/Rb relative to SiO2/MgO and form a sub-vertical trend (Fig. 13a) indicative of LSA; historical data for Table andesite samples also follow this trend. The

Ring Creek andesite and the Columnar Peak dacite have lower K/Rb values and plot sub-horizontally. The Barrier andesite samples appear transitional, plotting at the intersection between the LSA and HSA fields. This may relate to the Rb contents of Barrier andesite samples, which are slightly elevated. The LSA and

HSA groupings are still evident, though not as well defined, in the Sr-K/Rb-

(SiO2/MgO)*100 ternary diagram (Fig. 13b). Here, only the Columnar Peak dacite and the Ring Creek andesite are distinctly HSA and the Table andesite rocks are clearly LSA.

5. Discussion

Compositions determined for the GVC rocks analysed as part of this study and those of previously published data conform to virtually all of the adakitic geochemical traits put forward by Defant and Drummond (1990) and Martin et al.

(2005). The GVC rocks analysed in this study are characterised by high Sr/Y (46

# - 98), Al2O3 (16.9 - 18.8 wt. %), Mg (~ 51), low Y (≤ 17.1 ppm), and low Yb (≤

1.9; Tables A1 & A2, Appendix A). Some of the GVC adakites can be further divided into HSA and LSA groups. For example, while the Ring Creek andesite and Columnar Peak dacite units more closely represent HSA, with > 60 wt. %

SiO2, lower MgO (1.8 - 2.5 wt. %) and < 1100 ppm Sr (750 - 1078), the Paul 55

Figure 13: Discriminant diagrams for the LSA and HSA groups. (a) K/Rb vs. SiO2/MgO plot illustrating the distinction between LSA and HSA groups within the investigated GVC rocks. LSA plots higher in K/Rb while HSA defines a sub- horizontal trend. (b) Sr-K/Rb-{(SiO2/MgO)*100} ternary diagram distinguishing between LSA and HSA for the GVC adakites. LSA and HSA fields modified from Martin et al. (2005). Symbols as in Fig. 7. 56

Ridge andesite generally resembles LSA, with < 60 wt. % SiO2, higher MgO (3.4

- 4.9 wt. %) and low Rb (8.4 - 13.1 ppm). Mineralogically, the Paul Ridge andesite fits with LSA, but the Ring Creek andesite and the Columnar Peak dacite differ from the HSA definition in that they both contain pyroxene phenocrysts, which is a characteristic of LSA. The Black Tusk andesite samples

(and the Barrier andesite to a lesser degree) straddle the boundary between

LSA and HSA, with SiO2 contents ~ 60 wt. %. The Table andesite is the only unit that exhibits all the geochemical traits characteristic of LSA (Fig. 13a, b), as outlined by Martin et al. (2005). However, there are some variations in major and trace element contents which illustrate a more typical arc-like magma composition for the GVC as a whole and these are discussed below.

5.1 Adakitic Geochemistry of the GVC

5.1.1 La and Cr

The two key geochemical indicators for the identification of adakitic magmas are

Sr/Y ≥ 40 and La/Yb ≥ 20 (Defant and Drummond 1990). Determined Sr/Y ratios for the GVC rocks investigated herein cover a wide range (38 - 109; Fig. 12,

Table A2; Appendix A), but the majority of values fall between 75 and 90. Some of the previously published data (Black Tusk and the Barrier flow, in particular) straddle the boundary between adakite and normal arc-rocks, which appears to be related to elevated Y contents reported in these earlier published data. All of the GVC rocks have La/Yb ratios that lie in the field of normal arc-magmas and 57 not in the adakite field (Fig. 12). This is a function of the low La in the GVC relative to Yb content, which are typical of adakites (< 1.9 ppm). La concentrations in the GVC range from 10.7 to 18.3 ppm, lower than the average for typical intermediate rocks (~ 20 - 30; GERM, earthref.org). However, La values for adakite from other localities are in fact quite variable, from ~ 30 ppm

(Cook Island, Stern and Kilian 1996) to as low as 10 - 12 ppm (Pinchincha volcano and Tonga Trench; Bourdon et al. 2002, Falloon et al. 2008, respectively). Similarly, the Cr concentrations from all the centres in the GVC are found to be lower than that proposed by Martin et al. (2005) (~ 36 ppm), though

Cr contents of ≥ 30 ppm may be considered adakitic (Martin 1999, Richards and

Kerrich 2007), and these values are not significantly different from that of the average Cr content for the GVC adakites (~ 27 - 32 ppm).

5.1.2 LSA versus HSA

The majority of the GVC adakite units exhibit geochemical traits of both the LSA and HSA groups (except for The Table andesite). In general these units may be subdivided based upon their SiO2 (and to a lesser extent, MgO) contents. This may be the result of overlap of the LSA and HSA groups in the criterion outlined by Martin et al. (2005) (e.g. CaO + Na2O < 11 wt. %, Sr < 1100 ppm for HSA,

CaO + Na2O > 10 wt. % and Sr > 1000 ppm for LSA). However, the trace element data for the GVC also do not distinguish HSA from LSA. On a primitive mantle normalised spider diagram (Fig. 10a, b), LSA differs from HSA in lower

Rb and higher Nb values, with a stronger positive Sr anomaly and no Ti anomaly 58

(Martin et al. 2005). The rocks with a stronger LSA character (Paul Ridge and

Black Tusk andesite rocks) have lower Rb, but also lower Nb and similar, if not lower, Sr to HSA and a negative Ti anomaly. Similarly, the Ring Creek andesite and the Columnar Peak dacite, which are predominantly HSA, have higher Nb and Sr (on average) and this is a characteristic typical of LSA. The Barrier andesite samples have Rb contents that fall between those of the Paul Ridge and Columnar Peak rocks, but also the highest Nb contents of all the rocks of the GVC (up to 8 ppm). Martin et al. (2005) used a series of binary plots to help distinguish LSA from HSA (e.g., MgO vs. SiO2, Nb vs. SiO2 and Sr vs. CaO +

Na2O; Fig. 14). Differentiation between the two groups based upon these element pairings is not, however, possible for the GVC adakite rocks as the current data lack the extreme values of the adakite dataset used by Martin et al.

(2005) (e.g., Sr up to 3000 ppm, Nb > 20 ppm, Rb up to 150 ppm, etc.).

The tendency for the GVC rocks to display geochemical traits of both the LSA and HSA groups (except for the Table andesite) in addition to the other variations in major and trace element chemistry may suggest one of two things:

(1) the current dataset is insufficient to permit an adequate assessment of adakite affinity for the investigated GVC rocks, or (2) that the GVC rocks have been modified by other processes, that could have altered their adakitic signature.

5.1.3 Magma Mixing in the GVC 59

Figure 14: (a) MgO vs. SiO2, (b) Nb vs. SiO2, and (c) CaO+Na2O discriminant diagrams illustrating the variability of HSA and LSA compositions in the GVC dataset. LSA and HSA fields are modified from Martin et al. (2005). Major elements in wt. % oxide, Nb and Sr are in ppm. Symbols as in Fig. 7. 60

All the GVC adakite magmas show petrographic evidence of high-level magma mixing (Fig. 6) with the Paul Ridge andesite exhibiting the strongest. The geochemical variation that suggest a more typical arc-type composition are most likely the result of mixing between a pre-existing HSA magma and an intruding non-adakitic magma (or magmas) of broadly similar composition. Mixing of HSA magmas with normal, intermediate arc type magmas can decrease the adakitic character of the GVC lavas relative to the average adakite. In the case of the

Paul Ridge andesite, multiple periods of mixing are observed across the samples studied. For one sample (09JF011), the intruding magma is suggested to be of a more basaltic composition. The occurrence of olivine and orthopyroxene with quartz, and clear disequilibrium textures observed between these phases and that of their surrounding groundmass (see Fig. 6) support the interaction of compositionally distinct magmas in this sample, as has been documented at other volcanoes in the GVB (cf. Mt. Meager, Hickson et al. 1999).

The other Paul Ridge samples also exhibit disequilibrium textures, but not so markedly. There are no quartz xenocrysts observed in either 10JF017 or

10JF018, and while olivine crystals undergoing replacement by fibrous clinopyroxene are observed, these are somewhat less abundant than in

09JF011; large, relatively unaltered olivine is also present. The geochemistry of

09JF011 is also distinct from the other samples; 09JF011 has a much higher content of SiO2, MgO, Ni and Cr, but lower in Fe2O3 than determined for two other Paul Ridge andesite samples. Similarly, 09JF011 has a stronger adakitic 61 affinity, with higher Mg#, Sr/Y and La/Yb values (Tables A1 & A2; Appendix A).

This suggests that perhaps this sample did not undergo the same degree of mixing prior to eruption, than for the other Paul Ridge andesite samples, which may have been more extensively mixed. Green (1977) noted that some minor pyroclastic materials were present at Paul Ridge. It is possible that the 09JF011 sample taken from the northern part of Paul Ridge may represent a pyroclastic component that has a stronger basaltic character. Microprobe analyses of mineral phases from the Paul Ridge andesite (Tables C1-C5, Appendix C) and

SEM imaging supports these mixing relationships and that multiple flows of distinct composition are present within the Paul Ridge centre. For example, the compositions of phenocrystic olivine cores from sample 10JF018 are more primitive than those determined for olivine found in sample 10JF017 (Fo75 and

Fo65, respectively). Similarly, orthopyroxene phenocrysts present 10JF018 display strong reverse zoning towards their crystal margins, with rim compositions exhibiting a ~ 10 % increase in the enstatite component as compared to core compositions (Table C3; Fig. 15a, b). Plagioclase phenocrysts also record similar trends, although their compositions are less pronounced

(e.g., plagioclase found in sample 10JF017 is less calcic than that from sample

10JF018; An61-66 versus An66-72). However, despite the more primitive compositions identified in olivine, orthopyroxene and plagioclase phenocrysts, the whole rock composition of sample 10JF018 is surprisingly more felsic than that of sample 10JF017 (SiO2 is higher and Fe2O3 and TiO2 lower than for 62

Figure 15: Back scattered electron images (BSEI) photomicrographs of the GVC adakites illustrating magma mixing processes. (a) reverse zoning in orthopyroxene from the Paul Ridge andesite (red line), (b) X-ray element map of orthopyroxene in (a) showing the distribution of Mg in the crystal; note the more intense yellow colour representing an increase in Mg towards the crystal rim which then falls again as the outer rim is approached, highlighting the effects of fractionation on top of the magma mixing as the crystal grew. (c) plagioclase crystal from the Ring Creek andesite illustrating several generations of partial melting and crystallisation; lighter grey shades indicate higher Ca content, darker grey represents higher Na. (d) strongly resorbed original orthopyroxene (Px) crystal (light grey core) surrounded by new orthopyroxene (darker mantle) crystallising from new magma pulse in the Black Tusk andesite; light grey=higher Fe, dark grey=higher Mg. (e) quartz xenocryst (Qtz) surrounded by a reaction rim of augite from the Barrier andesite. (f) hornblende phenocrysts (Hbl) from Columnar Peak with dehydration reaction rims (resulting from the crystal being removed from its stability field) that is surrounded by a rim of fresh hornblende crystallising from the surrounding melt. Plag=plagioclase, Gl=glass, TiMt=titanomagnetite. 63

10JF017; Table A1). The identification of the distinct chemical differences between the Paul Ridge andesite samples may suggest a significant change in either the source region and/or magmatic processes on ascent. The 09JF011 sample maintains a strong mafic character (high MgO, Ni, Cr and Mg#) despite having the highest SiO2 content of all the Paul Ridge samples (Tables A1 & A2,

Appendix A) as well as incomplete mixing between very compositionally distinct magmas, whereas mixing in 10JF017 and 10JF018 appears to be somewhat more cryptic. A possible interpretation could be that 09JF011 originated as an

HSA magma and underwent more extensive interaction with mantle peridotite, while 10JF017 and 10JF018 represent true LSA magmas. The potential pyroclastic nature of 09JF011 suggests that it may be older than the other andesite samples, which are likely to be flows, and this relative timing coupled with the geochemistry and petrography indicates a possible transition from an

HSA source region to LSA source region. It must be noted that while different sources could explain the geochemical differences between the Paul Ridge andesite samples, all of these still exhibit traits of both the LSA and HSA groups.

For the Ring Creek andesite samples, there is no evidence of interaction with a basaltic intruding magma and mixing relationships are less clear, suggesting that the mixing components were perhaps of broadly similar composition or that they were more effectively mixed. The mineralogy of the Ring Creek andesite varies from an augite-hornblende-biotite-bearing assemblage in the proximal portion of the flow to augite only-bearing in the distal portion. The proximal andesite also 64 appears to contain primary phenocrystic quartz, whereas quartz crystals in the distal Ring Creek andesite show disequilibrium with the surrounding melt and hence, is difficult to explain through a simple mixing process. However, Sivertz

(1976) in mapping the Ring Creek andesite and Opal Cone also noted this difference in mineralogy and concluded that the hydrous mineral assemblage in the proximal Ring Creek andesite was identical to the mineralogy of Opal Cone itself. It is possible, therefore, that the proximal Ring Creek andesite entrained material from Opal Cone during eruption and inherited then hornblende-biotite- quartz mineralogy.

Mixing can result from convective overturn initiated by intrusion of hotter magma in to sub-volcanic magma chamber from depth (Hickson et al. 1999), but there is no evidence of mixing between compositionally distinct magmas in the Ring

Creek andesite. Magma chamber overturn between magmas of a similar composition can be caused if the intruding magma has a high upward momentum (Turner and Campbell 1986), or was significantly hotter. This is evidenced in Figure 15c, which shows several generations of compositional zoning with a distinct partial melting zone in a plagioclase phenocryst by a hotter intruding melt. The Ring Creek andesite flow is unusually extensive for an intermediate to felsic composition, having flowed for ~ 17 to 18 km from its vent

(Sivertz 1976; Green 1977; Brooks and Friele 1992). It represents predominantly non-explosive eruptive volcanism with no evidence to support a coeval pyroclastic component associated with this flow. Sivertz (1976) described lapilli 65 and block fragments of dacitic composition, but this material is only found within

Opal Cone. The large extent of the Ring Creek andesite flow and the lack of any significant pyroclastic material might suggest that intrusion of a large volume of relatively fast moving magma (the proximal andesite) of similar composition into pre-existing, cooling magma (the distal andesite) within the magma chamber, created turbulence that facilitated entrainment of wall-rock material from Opal

Cone as the eruption proceeded. The intermediate cumulate xenoliths observed in the distal part of this andesite flow (but not found in the proximal andesite in this study) were likely incorporated from the roof or walls of the magma chamber, during rapid eruption. It is not known to what extent the hornblende- biotite-quartz mineralogy is present in the Ring Creek andesite further south and access to the central portion of the flow is very limited (Sivertz 1976), making it difficult to resolve further.

In the Black Tusk centre, mixing relationships in the andesite samples are not as clear as for other centres within the GVC. The lava is essentially aphyric and the few phenocrysts present are strongly resorbed. This suggests that an intruding magma was likely significantly hotter than the pre-existing magma, resulting in strong thermal disequilibrium that caused near total resorption of earlier formed phenocrysts, thus obscuring any potential melt-crystal relationships. Additionally, the intruding magma appears to be a very similar composition to the resident magma. This may indicate that the volume of the intruding magma was large or rapidly intruded, in order to superheat the system and allow for the 66 disaggregation of all the phenocrysts. SEM imaging of orthopyroxene phenocrysts show rounded, strongly resorbed Mg-rich cores surrounded by a reversely zoned, euhedral mantle (Fig. 15d), illustrating the progressive intrusion of a much hotter magma.

Mixing in the Barrier andesite was likely between that of a dacitic pre-existing magma and an andesitic intruding magma. This is evidenced by reaction rims of augite on quartz phenocrysts (Fig. 15e) and strongly embayed biotite crystals.

Hornblende appears to be in equilibrium with the intruding andesitic lava; there are few disequilibrium features in the phenocrysts except for rims of opaques along the grain margins and these may be the result of hornblende being removed from its stability field during ascent.

At Columnar Peak, both the pre-existing and intruding magmas appear to be compositionally alike. There are disequilibrium features in the plagioclase and hornblende phenocrysts, but the intruding magma generally contains the same mineralogy, with the exception of minor orthopyroxene (< 1 %). These features indicate that despite the similar composition, the relative temperatures of both magmas were significantly different such that the larger, pre-existing phenocrysts were strongly resorbed and embayed relative to the phenocrysts from the intruding magma. Figure 15f shows fresh regrowth rims on pre-existing, embayed hornblende crystals, further evidence of interaction between magmas of similar composition. 67

Petrographic evidence for magma mixing is also supported in identified geochemical traits and trends for the investigated GVC rocks. Schiano et al.

(2010) used trace element modelling, on a database of 700 rocks from the

Ecuadorian Andes, to illustrate that mixing was the major control on their evolution. By plotting ratios of compatible and incompatible elements, fractional crystallisation, partial melting and mixing processes can be distinguished from each other (Allègre and Minster 1978, Schiano et al. 2010). By plotting an incompatible element versus the ratio of that incompatible element and a compatible element (e.g., Rb versus Rb/V), mixing and fractional crystallisation will form a curved trend, whereas partial melting results in a more linear trend

(Fig. 16a). In the case of GVC centres, the Black Tusk and Paul Ridge andesites show evidence of clear mixing or fractional crystallisation processes, while the

Ring Creek and Barrier andesites only show a slight curved trend, perhaps reflecting the degree/efficiency of mixing (i.e., the Paul Ridge and Black Tusk andesites show arguably stronger evidence of magma mixing while the Ring

Creek and Barrier andesites appear to be more effectively mixed). The limited dataset from the Columnar Peak dacite precludes any interpretation; however, when combined with other data for the GVC as a whole, this also illustrates a curved array. To isolate the effects of mixing from fractional crystallisation, a companion diagram is needed, where the incompatible/compatible element ratio is plotted against 1/compatible element (e.g., 1/V versus Rb/V, Fig. 16b). On this companion plot, mixing creates a linear trend, whereas partial melting and 68

Figure 16: Incompatible/compatible element ratio plots for (a) Rb vs. Rb/V and (b) 1/V vs. Rb/V distinguishing mixing from both partial melting and fractional crystallisation (FC). Insets are modified from Schiano et al. (2010). In (a), mixing and/or fractional crystallisation is represented by a curved trend. In (b), mixing is isolated and is illustrated by a straight line; this supports the dominance of magma mixing in controlling the chemistry of the GVC adakites. Symbols as in Fig. 7.

69 fractional crystallisation will both appear as curves; all of the GVC samples plot as linear trends. Figure 16b further illustrates that a mixing relationship reflects only magma mixing and not mixing of sources. Partial melting of a heterogeneous source would significantly modify the incompatible/compatible element ratio, whereas ratios of incompatible elements alone would not cause a change (Langmuir et al. 1978, Schiano et al. 2010). These plots and data suggest that the mixing relations observed in the GVC occurred in the magma chamber (or reservoirs) beneath each centre, after segregation from their solid source and argues against significant fractional crystallisation processes dominating the chemistry of the GVC adakites, as has been suggested for other adakite magmas (e.g. Castillo et al. 1999).

The mixing relationships in the GVC rocks are significant for adakite genesis for several reasons: (1) despite mixing with arguably non-adakitic magmas (in some cases extensive mixing), the majority of the adakitic characteristics are preserved in the GVC rocks (only Cr and La are affected); coherent, linear trends of major and trace elements with increasing SiO2 (Figs. 8, 9) also argue that both magmas are similar in composition, (2) these mixing processes may explain the fact that the GVC rocks exhibit geochemical characteristics of both the HSA and LSA groups, and (3) with no evidence of interaction with mafic magmas (except for the Paul Ridge andesite), the adakitic Ni, Cr and Mg# of the

GVC rocks likely reflect interaction with mantle peridotite during ascent through the mantle wedge (Rapp et al. 1999, Martin et al. 2005). 70

5.1.4 Interaction with Mantle Peridotite

The Ni, Cr and Mg# values for the GVC adakite rocks are generally higher than for normal andesite or dacite (~ 20, ~ 25 ppm and ~ 42, respectively) and, due to the lack of evidence of a basaltic mixing component for the majority of the GVC

(with the exception of the Paul Ridge andesite), these values perhaps reflect the interaction of HSA magmas with mantle peridotite during ascent from their source region. Increasing the Mg#, Ni and Cr concentrations in slab melts through the assimilation of peridotite (as described for HSA by Martin et al.

2005), however, is an unlikely process for the GVC. Typical mantle peridotite contains ~ 3200 ppm Cr, ~ 2300 ppm Ni and ~ 42 wt. % MgO (Sigurdsson et al.

2000). To inherit the values observed in the investigated GVC adakitic rocks, assimilation of peridotite would be strongly limited, as assimilation of > 1 % peridotite would result in Ni and Cr values higher than is observed in the GVC.

Partial melting of peridotite is an unlikely method to modify ascending adakitic magmas as peridotite partial melts would be basaltic (e.g. McKenzie and Bickle

1988), and mixing of these melts would be an inefficient way to increase the mafic content of HSA magmas while maintaining higher SiO2. Furthermore, there is no petrographic evidence for mixing with basaltic magmas observed in the

GVC (with the exception of one sample of the Paul Ridge andesite). A more likely process would be zone refining, whereby the ascending HSA magma gains

Ni, Cr and MgO by diffusion. This process enriched the adakite magma in mafic components at the expense of SiO2, but by and large preserves the incompatible 71 element ratios of the slab melt; the diffusion rates of incompatible elements (e.g.,

REE, Y) would be too slow to significantly modify the ascending magma (Wilson

1989) and obscure the slab melt signature.

5.1.5 Isotope Geochemistry

Limited 87Sr/86Sr and 87Rb/86Sr data from previous studies (Green 1977, 1990) of the GVC rocks show a smaller range of 87Sr/86Sr for intermediate compositions

(0.7031 - 0.7035; Green 1990) than for 87Rb/86Sr (0.025 - 0.120; Green 1990), which appears to relate to the increasingly felsic character of the rocks analysed, i.e. the rhyodacite and rhyolites from Mt. Garibaldi contain the highest ratios of

87Rb/86Sr. Green (1990) stated that the MGVF rocks contain a significant Rb-rich crustal component over the GLVF, which exhibit lower 87Rb/86Sr, and attributed this to be the result of AFC processes combined with contamination from crustal xenoliths and mixing with anatectic melts during ascent. While mixing of melts is present in the MGVF, there is little evidence for the incorporation of crustal xenoliths in any rocks examined as part of this study. Rare xenoliths are present in the Ring Creek andesite, but are mafic to intermediate in composition and would not significantly modify Rb values; the mineral phases present in the xenoliths are predominantly pyroxene and plagioclase, neither of which preferentially concentrates Rb (KdRb << 1 for orthopyroxene, clinopyroxene and plagioclase, Rollinson 1993). Lower Rb in the GLVF (4 - 18 ppm) was suggested by Green (1990) to reflect a depletion of LILE in the source region and less crustal interaction than the MGVF. For the MGVF rocks in this study, the Rb 72 concentrations are comparable to that determined by Green (1990) for the GLVF

(average 14.5 ppm). Stern and Kilian (1996) noted that the effects of crustal interactions were present in the adakite rocks from the Austral Volcanic Zone

(AVZ) and these decreased southward in the belt as the angle of subduction became more orthogonal. This resulted in negligible interaction of the Cook

Island adakites with crustal material and hence, their Sr isotopic compositions more closely resemble MORB (and by extension, slab partial melt) values.

87Sr/86Sr data from the Cook Island adakite samples approximates to 0.7028

(Stern and Kilian 1996), which is quite similar to the 87Sr/86Sr ratios determined for the GVC adakites (Green 1977, 1990) as well as average 87Sr/86Sr values for

MORB (~ 0.703; GERM, eartheref.org) and 87Sr/86Sr values for Juan de Fuca

MORB (~ 0.7025 – 0.7028; Hegner and Tatsumoto 1989, Chadwick et al. 2005).

The Cook Island adakite rocks are arguably the best known representation of potentially pristine slab melt, and Stern and Kilian (1996) based this on several geochemical traits that suggested a basaltic source, in addition to isotopic data.

The Rb/Sr ratios of the andesite rocks from Cook Island are extremely low, lower than typical MORB (and by extension, slab partial melt) values (~ 0.002). The

Rb/Sr concentrations in the GVC rocks (~ 0.01) are not as low as those found in the Cook Island andesites, but are comparable to MORB (0.006 - 0.009; Wilson

1989). Therefore, the possibility exists that: (1) the isotopic and element ratios of the GVC are inherited from a basaltic (MORB) source, and (2) they also argue against any significant contamination by crustal material. 73

5.2 Possible models for Adakite Genesis

Adakites were originally defined as felsic partial melts of subducted basaltic crust, leaving a residuum that was garnet rich but plagioclase poor and giving rise to the distinctive high Sr/Y and La/Yb ratios (Defant and Drummond 1990).

As studies continued, high Sr/Y and La/Yb ratios in arc-magmas were found to not necessarily be indicative of slab melts as different processes in subduction zone environments were shown to produce these high values; the most common of which are partial melting of mafic lower crust (e.g. Wang et al. 2005,

Macpherson et al. 2006) and high pressure AFC processes +/- magma mixing

(e.g. Castillo et al 1999, Chiaradia 2009). This has cast doubt on the use of high

Sr/Y and La/Yb as unique characteristics of slab partial melts. The following section evaluates the most common processes to generate adakite compositions in the context of the GVC to determine the most likely scenario for the origins of the GVC adakites.

5.2.1 Partial melting of basaltic lower crust

It has been shown that adakitic magmas can be generated from partial melting of thickened lower continental crust (Guo et al. 2006, Huang and He 2010) or remelting of underplated basaltic magma (Macpherson et al. 2006). A condition of these models is that magma genesis, whole or in part, must occur in the garnet stability field to generate the low HREE and Y concentrations in the resulting adakite melts. Garnet is stable in the lower crust as garnet amphibolite or 74 eclogite facies rocks, generally at ≥ 40 km depth (Richards and Kerrich 2007).

LITHOPROBE seismic transect studies from British Columbia, in the vicinity of the GVC, have shown that the approximate thickness of the crust is ~ 34 km

(Perry et al. 2002), too thin for garnet to be a stable phase at the base of the crust. Although some studies have found that garnet can be stable at depths as shallow at ~ 30 - 35 km (Garrido et al. 2006, Macpherson et al. 2006), this stability is dependent on the magmas being water saturated (Rooney et al.

2011). The pressure at the base of the GVC crust is approximately 1 GPa, hence for garnet to be stable at this depth, H2O contents must exceed 4 - 5 wt.

% (Alonso-Perez et al. 2009). In the experiment conducted by Alonzo-Perez et al. (2009), their liquids were corundum-normative and the fractionation of large amounts of hornblende (> 50 %) played an important role in generating an adakitic signature. In the case of the GVC, all adakite magmas are water undersaturated, which precludes garnet stability in the crust beneath the GVC.

The absence or sparsity of primary hydrous phases in the GVC adakite samples supports this interpretation. All of the GVC rocks are also mainly quartz- normative (Table B1, Appendix B) and there is little evidence of fractionation of significant amounts of hornblende (see Section 5.2.2). Moreover, studies of the thermal structure of the Juan de Fuca subduction system (Harry and Green

1999, Green and Harry 1999, Green 2006) have concluded that the system becomes increasingly anhydrous from southern California northwards to southwestern British Columbia, thereby reducing the amount of available H2O. 75

Gomez-Tuena et al. (2007) argue that partial melts of mafic lower crust would result in felsic magmas that have Mg#s less than 50; all the GVC adakites have average Mg#s greater than 50 (Table A1, Appendix A). Mafic lower crust partial melts would also not exhibit the elevated Ni and Cr contents typical of adakite

(Martin et al. 2005, Richards and Kerrich 2007). Fig. 17 shows adakitic indicator diagrams for the GVC rocks plotted with the average composition of lower crust partial melts from a study by Borg and Clynne (1998) in the Mt. Lassen volcanic field in the Southern Cascades. The inverted triangles represent the average composition of low silica partial melts leaving a plagioclase rich residue (filled triangle, LC 1) and the average composition of high silica partial melts leaving a hornblende rich residue (open triangle, LC 2). The two partial melts are interpreted to result from lower crustal anatexis initiated by underplating of hydrous basalt (hornblende residuum) versus anhydrous basalt (plagioclase residuum, Borg and Clynne 1998). The geochemical characteristics of the lower crustal melts are very different to those of the GVC adakites; they are both more felsic than the GVC rocks, with lower Mg# (43 and 46, Fig. 17a) as well as Ni, Cr,

Sr and Sr/Y values that are significantly lower than typical adakites (Fig. 17b-e).

The chemistry is mainly controlled by the modelled residual mineralogy of the crustal partial melts, which is made up of primarily plagioclase (40 – 50 wt. %) and clinopyroxene (20 – 25 wt. %) with varying amounts of hornblende (up to 3 wt. % for LC 1, 25 wt. % for LC 2), magnetite and orthopyroxene (9 – 15 wt. %;

Borg and Clynne 1998). The high residual plagioclase in both models lowers the 76

Figure 17: Adakite indicator plots comparing the GVC adakites to modelled lower crustal partial melts from Borg and Clynne (1998). Both partial melts plot well away from the GVC rocks in all diagrams and argues against partial melting of mafic lower crust as the source for the GVC adakites; see text for explanation. LC 1=low silica partial melt with plagioclase-rich residue, LC 2=high silica partial melt with a hornblende-rich residue. Silica is in wt. % oxide, trace elements in ppm. Mg# calculated as outlined in Table A1. Symbols as in Fig. 7. 77

Sr concentration to that well outside of the adakite field (Fig. 17d, e), while the high residual hornblende in LC 2 yields similar Y and Yb to the GVC rocks but

# lowers the SiO2, Mg , Ni and Cr (Fig. 17). The source of the GVC adakite magma is plagioclase-poor and high amounts of residual hornblende is unlikely because high water contents are required (> 5 % H2O; Borg and Clynne 1998) and this is inconsistent with the erupted GVC products. The crust beneath Mt.

Lassen is marginally thicker than the crust beneath the GVC (~ 38 km versus ~

34 km respectively, Guffanti et al. 1990, Perry et al. 2002) an the absence of garnet as a residual phase in the crustal melt models likely also precludes garnet stability at the base of the GVC crust. As the HREE depletion in the GVC rocks is most likely due to garnet (as compared to the crustal melts of Mt. Lassen where HREE depletion is controlled by hornblende), the GVC adakites must be sourced from the slab where garnet is stable to fractionate the HREE. Thus, partial melting of basaltic lower crust does not adequately explain the presence of adakite within the GVC, nor their compositions.

5.2.2 High pressure fractionation/AFC of basaltic magma

Fractionation of basaltic magma to generate adakite is the most common model for adakite genesis independent of slab partial melting (e.g., Castillo et al. 1999,

Chiaradia 2009, Chiaradia et al. 2009, Coldwell et al. 2011). Primitive magmas derived from partial melting of the mantle wedge may ascend to the base of continental lithosphere, where they may be halted because of differences in density. The elevated lithostatic pressures at this juncture may be sufficient to 78 form garnet, imparting low HREE and Y in melts that may form in this region and result in evolved magmas which differentiate into adakitic compositions

(Macpherson et al. 2006). Variations in this model also include multiple fractionations of mantle-derived basaltic magma and mixing of magmas originating in deep crustal hot zones (Castillo et al. 1999). A key factor in these processes is the necessity for garnet stability in the evolving magmas, for which there is no evidence beneath the GVC (see Section 5.2.1). Adakite melts that are generated from high pressure fractionation show correlations with differentiation indices (i.e. decreasing Fe2O3, MgO, Al2O3 with increasing SiO2) as well as adakitic indices (positive correlations between Sr/Y, La/Yb, Ni, Cr with increasing SiO2) and exhibit a wide range of SiO2 contents. In the GVC, some normal differentiation trends are observed (MgO, CaO, Fe2O3, etc., Fig. 8), but several major and trace element contents from each centre have either inverse correlations with silica or remain relatively constant as SiO2 increases (e.g. TiO2,

# Na2O, Ni, Cr, Mg ; Figs. 8, 9). Additionally, the GVC adakites exhibit a relatively small range of silica contents (~ 8 wt. %) compared to other adakite suites (~ 17 wt. %, Macpherson et al. 2006). Fractionation of hornblende has been argued as a proxy for garnet in controlling HREE and Y concentrations in adakite melts

(Richards and Kerrich 2007, Dessimoz et al. 2012). As basaltic magma crystallises, the HREE and Y are generally incompatible in the major mafic silicate phases (olivine, orthopyroxene) and these elements increase with increasing SiO2 up to ~ 56 wt. % where hornblende may crystallise. As the result 79 of hornblende fractionation, Y and HREE can inversely correlate with SiO2, illustrating a crystallisation path from normal arc rocks into the adakite field

(Richards and Kerrich 2007, Moyen 2009).

The CVB are the only mafic rocks in the GVC and previous studies have found that they do not represent the parent magma to the GVC adakites (Green 1977,

1990; Green and Henderson 1984). There is an absence of data between the

SiO2 contents of the CVB (48 - 51 wt. %; Table A1, Appendix A; Figs. 7, 8, 9) and the GVC adakites (> 56 wt. %) and as such, a trend illustrating hornblende fractionation (if for illustrative purposes only) cannot be effectively demonstrated.

However, Y and Yb contents of the CVB are relatively constant over the, albeit narrow, range of SiO2 concentrations (i.e. shows no increase with increasing

SiO2, which would be expected if these were controlled by fractionation). Y and

HREE generally increase with SiO2 for the GVC adakites (Table A2, Appendix

A); hornblende crystallisation would decrease these values as the magma evolves to more felsic compositions. While hornblende can preferentially fractionate HREE, the MREE have higher compatibilities (Rollinson 1993). This can be illustrated in chondrite normalised spider diagrams where the effect of hornblende can be identified by a listric or spoon shaped profile; a negative correlation of MREE/HREE (Dy/Yb) with SiO2 also indicates interaction of hornblende (Rollinson 1993, Richards and Kerrich 2007). The majority of the

GVC adakites have a smooth negative pattern on chondrite normalised spider diagrams (Fig. 11), suggesting that small amounts of hornblende may be 80 present, but does not rule out the effect of residual garnet in the source region and the MREE/HREE ratios remain relatively constant with increasing silica (not shown). The Columnar Peak dacite does show a slight listric pattern which may result from small amounts of hornblende phenocrysts (see Section 5.1.3).

Significant hornblende crystallisation would also lower the Mg content of the resulting melts and hence lower the Mg#, which is not observed in the GVC adakite rocks. Lastly, experimental studies on high pressure fractionation of basalts showed that the resulting melts would be corundum normative (e.g.

Müntener et al. 2001; see previous section). This is in contrast to the GVC adakites, which are all predominantly quartz normative (Table B1, Appendix B).

5.2.3 Partial melting of subducted ocean crust

The failure of the previously discussed models to generate the distinctive, adakitic chemistry of the GVC rocks indicates that a slab partial melt scenario is a perhaps a more likely process for generating adakite within the GVC. The majority of geochemical traits as outlined by Defant and Drummond (1990) and

Martin et al. (2005) are observed in the GVC rocks; the few that are not (La, Cr,

LSA versus HSA) can be explained away in magma mixing (see Section 5.1.3).

This slab partial melt hypothesis is further explored in the context of the GVC using trace element batch melting models, to determine if the GVC adakites can be generated from a MORB source. Model concentrations of trace elements Rb,

Ba, Th, Nb, La, Ce, Sr, Nd, Sm, Zr, Eu, Dy, Y, Yb and Lu were determined using an a restite composition of predominantly clinopyroxene, garnet and hornblende 81

+/- rutile. Based on the estimated depth and pressure of the Juan de Fuca slab beneath the GVC (~ 60 km, 1.5-1.6 GPa; Audet et al. 2008), the starting composition is assumed to be amphibolite. Partial melting experiments on amphibolite (Wolf and Wyllie 1993, Rapp 1995, Rapp and Watson 1995) indicate that garnet pyroxenite or eclogite is a suitable residuum for the approximated conditions beneath the GVC. Experiments conducted by Rapp et al. 1999 show that a Ti-bearing phase is ubiquitous in the residue and hence, rutile is included in the model. The modal proportions of each mineral phase were varied

(sometimes significantly, from ~ 10 % up to 75 %) to test the effects on the modelled composition compared to that of the GVC adakite samples, at incremental melt percentages of up to 25 % partial melt. Mineral/melt partition coefficients for silicic melts at temperatures and pressures comparable to the

Juan de Fuca subduction system, sources and values used are listed in

Appendix D. The starting composition for N-MORB representing the subducted slab is derived from Karsten et al. (1990) and Chadwick et al. (2005); enriched, depleted and transitional basalt magmas have been erupted from the Juan de

Fuca Ridge system (Cousens et al. 1995) and a slightly enriched composition

(compared to average Pacific MORB) has been used in the modelling. A few model compositions can approximate the GVC adakites (see Fig. D1, Appendix

D), but the model that best reproduces the GVC adakites is a 15 to 20 % partial melt of N-MORB leaving a residuum of ~ 70 % clinopyroxene, ~ 20 % garnet, ~

9.5 % hornblende with 0.5 % rutile (Fig. 18a). A second model where garnet is 82

Figure 18: Results of best fitting simple batch melting models (15 % melt fraction) to the GVC adakites. The residues modelled fit the range of the GVC adakites reasonably well; the model composition in (b) is preferred based on previously published experimental data on partial melting of basalt. N-MORB normalising values are estimated from Karsten et al. (1990), Cousins et al. (1995) and Chadwick et al. (2005). Cpx=clinopyroxene, Gt=garnet, Hbl=hornblende, Ru=rutile. See text for explanation. 83 increased to 25 % and hornblende is decreased to 4.5 % is also a good fit, slightly less so than the former (Fig. 18b). Changes in the modal percentage of clinopyroxene does not significantly change the modelled composition relative to the GVC rocks; however, small to moderate increases in the amount of residual garnet (> 25 %) results in lower HREE and Y than is observed. The presence of hornblende (+/- rutile) is needed to moderate the HFSE and MREE contents; a clinopyroxene and garnet only residue results in Nb, Zr and Eu (as well as Ce) values that are too high. A similar model approximating the GVC chemistry can be obtained with a residue of 60 % clinopyroxene, 15 % garnet and 25 % hornblende (Fig. 18c); Zr, Sm and Eu values are closer to the observed concentrations in the GVC rocks, but also result in lower Sr and Nd and higher

Nb and an overall poorer fit to the GVC data. Ickert (2006) tested the adakite slab melt model for the Princeton Group adakites in south-central British

Columbia and noted that partial melting of amphibolite could not generate the sodic, tonalitic melts that are consistent with adakites. This is the case only if sufficient hornblende remains in the residue; removal of most or all residual hornblende will result in partial melts that are adakitic (Rapp 1995). Additionally,

Ickert (2006) used the most primitive andesites to model, whereas primitive slab melts are dacitic in composition. Thus, the model with 25 % garnet and 4.5 % hornblende is the most likely residuum beneath the GVC (Fig. 18b). The modelled composition fits the observed data quite well; discrepancies (slightly lower Sr and slightly higher HFSE) could be influenced by magma mixing 84 processes (see section 5.1.3) and/or minor differentiation. Alternatively, the partition coefficients used in the batch melting model (Barth et al. 2002) could cause the elevated Zr and Eu anomalies as the determined Kd’s for these elements are generally lower than other studies (e.g. Klein et al. 1999, Green et al. 2000). The overall success of the model in replicating the chemistry of the evolved GVC rocks suggests the following: (1) the model emphasizes the petrographical and geochemical observations that the magma mixing components are broadly similar in composition and adakitic (or near-adakitic),

(2) the mixing relationships do not significantly affect the model chemistry, even the more mobile LILE, and most importantly (3) the trace element chemistry of the GVC adakites can be generated from partial melting of N-MORB and the good fit of the model composition to the observed composition suggests that the trace and REE values of the GVC adakites may represent true slab melt values.

In addition to the trace element modelling and geochemical support for a slab partial melt model, field relationships and the tectonic environment in southwestern British Columbia also suggest that the GVC adakites originated from slab melts and are discussed below.

5.2.3.1 Association with niobium-enriched basalts (NEB)

NEB were first identified and defined by Sajona et al. (1993, 1996) from the

Zamboanga arc in the Philippines. These basalts are characterised by enrichment in HFSE as compared to ‘normal’ subduction related magmas, which 85 generally exhibit strong depletion in HFSE. Further, NEB contain relatively high

TiO2 (1 - 2 wt. %) and low LILE/HSFE and LREE/HSFE ratios (Sajona et al.

1993, 1996). These enrichments result in positive or weakly negative Nb anomalies on mantle normalised multi-element plots. NEB and adakite are often

(but not always) associated with each other and this has led to the suggestion that there is a petrogenetic link between them (Sajona et al. 1996, Aguillón-

Robles et al. 2001), though some studies argue against this hypothesis (Castillo et al. 2007, Hastie et al. 2011). Sajona et al. (1996) propose that NEB are generated by partial melting of the mantle wedge that had been previously metasomatised by slab melts. Interaction with the mantle wedge resulted in the slab melts being consumed generating a metasomatic mineral assemblage that included primarily orthopyroxene and amphibole. This mineral assemblage scavenges HFSE from subsequent slab partial melts (likely by a process of zone refining) as they ascend through the altered wedge. As subduction continues, the metasomatic assemblage is pulled further down into the mantle where it undergoes partial melting and generates a HFSE-enriched NEB basaltic melt

(Sajona et al. 1993, 1996). Some NEB are erupted at surface (e.g., Mexico,

Philippines) as primitive basalts while some likely differentiate and/or fractionate at depth and erupt as more evolved compositions (discussed below). Other studies suggest that the HFSE enrichment originates from mixing between a

MORB source and an OIB source (Castillo et al. 2007, Macpherson et al. 2010), or partial melting of lower crust comprised of accreted oceanic terrains and/or 86 entrainment of asthenospheric material by convection into the mantle wedge

(Borg et al. 1997).

The CVB exhibit several geochemical traits that identify them as NEB, including higher Nb (~ 8.2 ppm), high TiO2 (~ 1.5 wt. %) and low LILE/HFSE and

LREE/HFSE ratios (e.g., Rb/NbMN = 0.85, La/NbMN = 1.1; both ratios greater than

2 in normal arc rocks). On a primitive mantle normalised multi-element plot, the

CVB samples display weak Nb anomalies and no Ti anomaly, as well as, lower total LILE than is typical of arc rocks (e.g., ~ 6 ppm Rb versus ~ 20 ppm; Fig. 8).

Partial melting of accreted oceanic terrains to generate the CVB NEB is thought unlikely; the basement rocks below the GVC are quartz diorite and diorite intrusions of Cretaceous age (Souther 1970, Green 1977, Rusmore and

Woodsworth 1991), that generally host silica contents of greater than 60 wt. %

SiO2 (Mathews 1958). Partial melting of such a composition could not generate primitive basalts. Similarly, asthenospheric entrainment into the mantle wedge is also deemed an unlikely process; a large slab window is present to the north of the GVC that extends from the Alaska-Yukon border to Mount Itcha and the

Anahim-Wells Gray volcanic field (Thorkelson et al. 2011). Asthenospheric material would generally flow in a northward direction into the slab window rather than east-west into the mantle wedge beneath the GVC. A slab window has also been suggested to be present between the Juan de Fuca and Explorer plates, along the Nootka Fault Zone (Audet et al. 2008, Thorkelson et al. 2011), and asthenospheric upwelling through this window (or around the eastward edge of 87 the Juan de Fuca Plate) may be the source of the alkaline volcanism found east of the GVC in south-central British Columbia (e.g., Princeton Group volcanics,

Ickert et al. 2009, Thorkelson et al. 2011). However, these studies have suggested that the slab window between the Juan de Fuca and Explorer plates is in an early stage of formation and that it has not propagated westward enough to cause significant alkaline volcanism in the GVC.

Mixing of MORB and OIB components has been argued as the source of several

NEB occurrences (Macpherson et al. 2006, 2010; Castillo 2008, Castillo et al.

2007) and is the most common model for generating NEB independent of slab- melt contribution. In the case of the CVB, a simple mixing model of MORB and

OIB cannot successfully explain the geochemistry. Figure 19 shows a simple binary mixing plot between OIB and MORB using various pairs of trace elements. Figure 19a shows that mixing of 10 % OIB into MORB could potentially generate the CVB, but the data does not fit well and even less so for the HCBA. This could be the influence of fluid interactions, as Ba is not as incompatible as Nb, but the data does not appear to support an OIB component.

Figure 19b uses Zr instead of Ba, which is generally thought to behave as an immobile element in aqueous fluids. The CVB data correlates well with a 10 %

OIB component; the HCBA shows less scatter but still trends towards more

MORB-like values. Based on these element pairs, it appears that the CVB can originate from N-MORB mixed with 10 % OIB. However, further testing of this simple mixing model with additional trace elements (e.g., HREE) fails to support 88

Figure 19: Simple binary mixing models between MORB and OIB components for (a) Ba vs. Nb and (b) Zr vs. Nb. In (a), both the CVB and HCBA rocks plot off the mixing trend. The high Ba for the HCBA may be the result of fluid influence. In (b), the data plots much closer to the mixing line and suggests that a mix of 90 % MORB and 10 % OIB can generate the CVB rocks (and to a lesser extent, the HCBA). MORB composition derives from Sun and McDonough (1989), OIB from Borg et al. (1997). Symbols as in Fig. 7.

OIB-MORB mixing. Figure 20 is a primitive mantle normalised multi-element diagram comparing the CVB chemistry to a simple mixing model of 10 % OIB and 90 % MORB (the data for the HCBA are not included as they represent more evolved compositions). Some of the elevated HFSE values (Zr, Ti) can be generated from mixing approximately 10 % OIB and 90 % MORB, however, this composition results in HREE that are much higher than that of the CVB, and

LILE concentrations that are significantly lower with a strong negative Pb anomaly. The addition of a bulk sediment component (~ 10 %) to the OIB-MORB mixture increases LILE closer to CVB values, but does not lower the HREE and requires 15 % OIB. Additionally, this three component model magma would result in the 87Sr/86Sr isotopic ratios to be much higher (~ 0.7044; Plank and

Langmuir 1998) than published data for the CVB (0.7032; Green 1977, 1990).

Lastly, similar plots of incompatible and compatible element ratios (as discussed at the end of Section 5.1.3) illustrate that the main control on the chemistry of the

CVB is partial melting (Fig. 21) and not fractional crystallisation or magma mixing. For the CVB to be generated by OIB and MORB mixing, this would require that the solid source be a heterogeneous mixture of OIB and MORB components that subsequently underwent partial melting (Schiano et al. 2010).

Determining if the source was efficiently mixed cannot be effectively performed with these plots but as discussed above, a heterogeneous mixture cannot reproduce the CVB data. Thus, for the CVB NEB, the most effective way to

Figure 20: Primitive mantle normalised mixing models of MORB, OIB and bulk sediment components as compared to the CVB. Mixing between 10 % OIB and 90 % MORB replicates the HFSE of the CVB but poorly matches the majority of the data. Addition of 15 % bulk sediment and increasing the OIB component to 15 % results in a marginally better fit but overall fails to reproduce the CVB chemistry. N-MORB composition from Sun and McDonough (1989), symbols as in Fig. 7.

Figure 21: Incompatible/compatible element ratio plots for (a) Ba vs. Ba/Ni, (b) 1/Ni vs. Ba/Ni, (c) La vs. La/Cr and (d) 1/Cr vs. La/Cr distinguishing mixing from both partial melting and fractional crystallisation (FC) processes. The geochemistry of the CVB is controlled mainly by fractional crystallisation, whereas the HCBA defines a distinct trend from both the CVB and the primitive Helm Creek basalts. Inset as in Fig. 16, symbols as in Fig. 7. Trace elements expressed in ppm.

92 increase the HFSE without significant modification to LILE is partial melting of slab melt-metasomatised mantle wedge.

In the case of the Helm Creek flow, the processes influencing the geochemistry are more complex. The HCBA was originally mapped by Green (1977) as a composite flow; the lava exhibits a range of silica contents (49-55 wt. % SiO2).

Primitive alkali olivine basalt is also described and occurs within the Cinder Cone

(Green 1977, 1990, 2006). Figure 11 (a and c) is a chondrite normalised plot showing the REE values for the CVB, the HCBA from this study and the REE values from primitive and evolved basalts taken from Green and Henderson

(1984) and Green (2006). The more evolved Helm Creek rocks exhibit higher

HFSE and lower HREE than the more primitive basalt; these concentrations should remain relatively unchanged through differentiation from basalt to basaltic andesite and the observed values may suggest that the source region for the more evolved Helm Creek flow is different than that of the alkali olivine basalts.

Figure 11c also illustrates that the HREE contents of the HCBA are very similar to that of the CVB. It is possible that the HCBA may have had a NEB parent that either did not erupt at surface or experienced differentiation and/or fractional crystallisation prior to eruption. The differences in the primitive basalts and the

HCBA are further illustrated in Fig. 21 as both groups form diverging trends. In

Fig. 21a, the HCBA show a positive curved relationship while the primitive rocks form a linear negative trend. The companion plot shows the inverse of this relationship illustrating that mixing is not the main control on the evolved 93 compositions but may affect the chemistry of the primitive Helm Creek basalts.

Ultimately, this shows that the source regions of the primitive and evolved Helm

Creek rocks are likely different.

These observations outlined above are significant to the argument about whether or not NEB and adakite are genetically related; the HCBA exhibits several adakitic traits such as high Sr (> 1400 ppm), low Y and Yb (15.4 and 1.6 ppm respectively), high Ni, Cr and Mg# (113 and 135 ppm,~ 55; Tables A1 & A2,

Appendix A). The only condition that precludes the HCBA from being classified as adakite is that the silica content is too low (< 56 wt. %). Martin et al. (2005) likened LSA to NEB based on their elevated Nb contents (up to 10 ppm), and the source region described by Martin et al. (2005) for LSA is identical to that proposed by Sajona et al. (1996) for NEB. Both models require a process that enriches the mantle wedge in Nb; the elevated values in NEB (> 7 ppm; adakite average of 6 ppm) cannot be obtained by fluid flux from the dehydrating slab because these fluids are not effectively able to concentrate Nb (Tatsumi et al.

1986, Martin et al. 2005). A slab partial-melt can concentrate and transfer Nb as it will preferentially go into the melt phase due to its incompatibility. Hence, based on this data, it seems likely that NEB could represent a potential parental magma to LSA and/or NEB are more extensive partial melts of slab melt- metasomatised mantle wedge; both of which suggest that NEB and adakites are related in the GVC.

5.2.3.2 Ni content in olivine 94

Several recent works (Sobolev et al. 2005, 2007; Straub et al. 2008) have recognized anomalously high nickel concentrations in olivine phenocrysts that are inconsistent with magma generation in equilibrium with a peridotite source, first noted in the Hawaiian basalts. This was explained by magma generation from a pyroxenite source; partial melts from recycled oceanic crust (i.e. slab melts) at depth infiltrate and interact with mantle peridotite, which consumes olivine and creates solid pyroxenite. The pyroxenite melts at lower pressures and the resulting liquids are highly enriched in Ni, as pyroxene has a lower partition coefficient than olivine for Ni (~ 1 versus >> 1 for olivine, Herzberg

2011). Olivine phenocrysts that crystallise from this modified source reflect this elevated Ni (Sobolev et al. 2005, 2007). A pyroxenite source region can be distinguished from peridotite on a NiO versus forsterite content diagram where olivines from a pyroxenite source plot at a higher NiO value for a given Fo

(Sobolev et al. 2005). The significance of this in the context of the GVC is that both LSA and NEB are essentially generated from a pyroxenite source (Sajona et al. 1996, Martin et al. 2005). Figure 22 is a NiO versus Fo plot for olivines from the GVC with fields outlining peridotite and pyroxenite sources (fields modified from Sobolev et al. 2005 and Gao et al. 2008). The HCBA is clearly distinct from the other flows; it has the most primitive olivines and forms a smooth, sub-vertical negative trend into the pyroxenite field. The olivine rim compositions (open symbols) are considerably poorer in Ni than the cores and reflect disequilibrium conditions between the phenocrysts and the surrounding 95

Figure 22: NiO (wt. %) vs. Fo (mole %) diagram of olivines in the GVC. Fields are modified from Sobolev et al. (2005) and Gao et al. (2008) and differentiate between a pyroxenite source and a peridotite source. The HCBA and DVBA both steeply trend into the pyroxenite field, whereas the CVB and PR have exhibit shallower trends towards the overlap between the pyroxenite and peridotite fields and suggest the involvement of slab melts in the source region for all the lavas. Filled symbols=core compositions, open symbols=rim compositions. PR, CVB, DVBA and HCBA acronyms as in Fig. 7.

96 melt. The olivines from the Desolation Valley flow are much lower in their forsterite content but strongly trend into the pyroxenite field as well (Fig. 22). The

Paul Ridge andesite and the CVB have a similar range of forsterite contents, but the CVB have slightly less Ni; both datasets plot mainly outside the peridotite and pyroxenite fields. The Fo values are not primitive for either the CVB or the

Paul Ridge andesite and the potential source region cannot be adequately determined but the data for both flows trend towards the area of overlap between both fields (Fig. 22). Similarly, the Ni contents are not unusually high, however studies on the Mexican Volcanic Belt (MVB; Straub et al. 2008, 2011) found that lower Ni and lower Mg olivines originated from a mixed pyroxenite/peridotite source. Ultimately, the data trends potentially provide further support for a slab melt component in the generation of the GVC NEB. It should be noted that further microprobe analyses of olivine may or may not corroborate these interpretations.

5.2.3.3 Regional tectonic regime for southwestern British Columbia

Studies on the thermal structure of the Juan de Fuca Plate (Green and Harry

1999, Harry and Green 1999, Green and Sinha 2005, Green 2006) have shown that the Plate (and magmas generated above it) becomes increasingly anhydrous further northwards. The geochemistry of erupted magmas also changes further northwards; magmatism at Mt. Cayley, approximately 20 km north of the GVC, exhibits an even stronger adakitic character, which has also not been previously recognized. Sr/Y ratios for the Mt. Cayley adakites are as 97 high as 215, well above that of the GVC (Kelman et al. 2001). Average Mg# (~

53), Cr (~ 59) and Sr (~ 1123) are also higher than the GVC adakite rocks, with lower Y (≤ 15 ppm) and lower range of K2O/Na2O (0.21 - 0.42; Kelman et al.

2001). The Mt. Cayley adakite magmas are predominantly HSA, and no basaltic volcanism is associated with the suite (Kelman et al. 2001). Figures 10, 13 and

14 compare the adakitic characteristics of Mt. Cayley to that of the GVC adakites. Mt. Meager, which is approximately 60 km north of the GVC, represents another change in magma composition from the GVC to that of an alkaline affinity. The rocks of Mt. Meager include trachydacite, trachyandesite, basalt and trachybasalt (Stasiuk et al. 1994). An apparent transition northward from adakitic calc-alkaline in the GVC to strongly adakitic calc-alkaline at Mt.

Cayley to alkaline volcanism at Mt. Meager appears to be linked to the extension and thinning of the slab between the Juan de Fuca and Explorer plates, along the Nootka Fault Zone (Audet et al. 2008). Recent studies (Madsen et al. 2006,

Audet et al. 2008) have shown that the Explorer Plate is likely being captured by the North American Plate and is no longer subducting, whereas the Juan de

Fuca Plate subducts at a relatively slow rate (~ 4 cm/yr; Wilson 2002). This has caused an approximate 40 km of stretched slab between the Explorer and Juan de Fuca plates (Audet et al. 2008). Asthenospheric upwelling into this stretched section would increase the heat flow and could facilitate partial melting; this could potentially generate the alkaline character of the Mt. Meager centre, as the volcano is situated on the Nootka Fault Zone (Thorkelson et al. 2011). Further 98 southward, the heat decreases and the alkaline signature changes to a strong adakitic character (Mt. Cayley rocks), as the slab melt chemistry dominates.

Beneath the GVC, the slab is very young (~ 16 Ma; Green 2006) and retains sufficient heat to melt, though the mantle wedge is more hydrous than further north, which results in both adakitic and some normal arc-type melt-generation.

These normal (but likely near-adakitic) arc magmas mix with the adakitic partial melts of the slab in the magma reservoirs to generate the hybrid GVC rocks (see section 5.1.3). This along arc variation in magma geochemistry suggests that partial melting of the subducted slab is not only occurring beneath the GVC, but in fact may become an increasingly dominant magma generating process further northwards.

5.3 Petrogenetic model for adakite genesis beneath the GVC

The model proposed herein illustrates that the most likely method to explain the geochemical and mineralogical variations in the rocks of the GVC is partial melting of subducted ocean crust, followed by magma mixing in upper level magma chambers, and subsequent melting of slab melt-metasomatised mantle wedge; Figure 23 illustrates the interpreted sequence of volcanic events generating the adakites and NEB of the GVC. Based on the stratigraphy and relative timing of eruption of the rocks in the GVC, there appears to be multiple periods of both slab melting and melting of altered mantle wedge. Partial melts of the subducting Juan de Fuca Plate ascended through the mantle wedge and were consumed through interactions with mantle mineralogy, forming 99

Figure 23 (1 of 3): Simplified petrogenetic model for adakite and NEB genesis in the GVC. Estimated depths in km, acronyms as in Fig. 7. *veins and partial melts may extend further upwards than is depicted in this schematic. NOTE: model only includes volcanic centers studied in this work and does not represent all eruptive episodes in the presented time scale. 100

Figure 23 (2 of 3): Simplified petrogenetic model for adakite and NEB genesis in the GVC. Estimated depths in km, acronyms as in Fig. 7. *veins and partial melts may extend further upwards than is depicted in this schematic. NOTE: model only includes volcanic centers studied in this work and does not represent all eruptive episodes in the presented time scale. 101

Figure 23 (3 of 3): Simplified petrogenetic model for adakite and NEB genesis in the GVC. Estimated depths in km, acronyms as in Fig. 7. *veins and partial melts may extend further upwards than is depicted in this schematic. **slab window is north of the study area and not shown in this figure. NOTE: model only includes volcanic centers studied in this work and does not represent all eruptive episodes in the presented time scale. 102 orthopyroxene and hornblende (+/- phlogopite) at the expense of olivine and clinopyroxene and creating a metasomatic assemblage (i.e. pyroxenitic source).

This assemblage created a veined, metasomatised mantle wedge; these veins of metasomatic minerals created pathways that allowed subsequent slab melts to ascend into the magma chambers beneath the GVC with their HSA chemistry relatively intact. As the altered mantle wedge was subducted further via slab pull, the metasomatic mineral assemblage present in the veins melted to generate the LSA magmas. These magmas ascended and in doing so mixed with the pre- existing HSA already present in the GVC magma chambers reservoirs, such as those occurring beneath Black Tusk at 1.3 Ma - which shows petrographic evidence for the superheated mixing (see Section 5.1.3) of pre-existing HSA magma and the much hotter intruding LSA magma, giving rise to the mixed geochemical signature between LSA and HSA for this centre (Fig. 23, 1.3 Ma,

GLVF). The Paul Ridge andesite in the MGVF erupted next at approximately

700 ka whereby the intrusion of a hotter LSA magma caused a pyroclastic eruption, possibly resulting from a temperature-driven convective overturn in the magma chamber, and producing the distinct geochemical and petrographical variations in sample 09JF011 (Fig. 23, 1.3 Ma and 700 ka, MGVF) compared to the subsequent eruptions at Paul Ridge. A hiatus in volcanism in the GVC (700 ka to 260 ka) allowed for a second cycle of slab melts to be consumed and metasomatise the mantle wedge (Fig. 23, pg. 2, MGVF), which provided pathways for the eruption of HSA at Columnar Peak, beginning at 260 ka. This 103 newly metasomatised wedge subducts and partially melts, creating the largely

LSA Table and Barrier andesites which were approximately coeval eruptions at ~

100 ka. The DVBA also erupted at 100 ka; geochemical data is sparse for this flow but the few published analyses indicate that it exhibits some characteristics of NEB (~ 9 ppm Nb, ~ 118 ppm Zr, 0.94 wt. % TiO2,; Green 1981) and is also close to an adakitic composition (Sr > 750 ppm, ~ 23 ppm Y, 55.9 wt. % SiO2).

The HCBA, which erupted at 40 ka, also shows some geochemical traits of NEB

(8.5 ppm Nb, 126 ppm Zr, 0.95 wt. % TiO2; Tables A1 & A2, Appendix A) but exhibits a far stronger adakite signature than is present in Desolation Valley flow rocks (Sr > 1400 ppm, Sr/Y = 92, La/Yb > 15; Table A2, Appendix A). Alkali olivine basalts occur within the Cinder Cone volcano as the last eruptive products; this progressive change in chemistry from near adakiticstrongly adakiticalkaline over time may suggest a change in the magma source region.

The DVBA likely reflects a source where the slab has insufficient garnet to generate an adakite signature; the resulting slab melts had higher, non-adakitic concentrations of HREE and Y, which were transferred to the mantle wedge as the slab melts interacted with the mantle wedge and were consumed creating a metasomatic assemblage. The slab melts, however, were able to transfer HFSE from the slab and imparted some geochemical characteristics of NEB.

Stratigraphically, the Barrier and Black Tusk andesites (which are adakitic) are approximately coeval with the Desolation Valley flow and hence it is possible that the DVBA was an LSA magma that mixed with normal arc magmas prior to 104 eruption (Fig. 23, 100 ka, GLVF). This is supported by the observation that

Clinker Peak, Black Tusk and Cinder Cone are not temporally far apart and it is unlikely that the residual slab composition has significantly changed between centres. The adakite signature is stronger in the HCBA, likely because the slab became richer in garnet and as subduction continued, the mantle wedge was altered by subsequent slab melts that were strongly adakitic, resulting in the higher Sr/Y and La/Yb ratios. The source of the alkali basalts is speculative, but it may reflect the start of a westward progression of alkaline volcanism; volcanism that exhibits an intraplate character has been identified in south- central British Columbia and believed to have started around 47 Ma and continued to as recent as a few hundred years ago (Ickert et al. 2009,

Thorkelson et al. 2011 and references therein). The episodic and relatively voluminous outpouring of the mildly alkaline CVB from a centre further north suggests an increase in heat into the system, possibly due to the upwelling of asthenospheric material into the Nootka Fault Zone, which melted out a large portion of the altered mantle wedge and imparting the NEB characteristics to the

CVB. This increased heat flow also caused significant melting of newly subducted ocean crust and resulted in a third cycle of HSA generation with the eruption of the extensive Ring Creek andesite flow at 10 ka (Fig. 23, pg. 3, 10 ka, MGVF).

6. Conclusions 105

The identification of adakites introduces new mechanisms of magma genesis beneath the GVC which have not been previously considered. Evaluation of the main models for generating adakites coupled with simple batch melting modelling have shown that the most likely scenario for the generation of the

GVC adakites is partial melting of the subducting Juan de Fuca slab. The mixed nature of the GVC rocks precludes classification of most of the adakites as LSA or HSA, but despite mixing with arguably non-adakitic magmas the slab melt signature is still evident. The presence of NEB associated with the GVC adakite strengthens the slab melt model, as trace element modelling of the CVB and

HCBA argues against mixing between MORB and OIB components to generate the GVC NEB. Higher Ni contents in olivine phenocrysts may also support a slab melt component to the GVC NEB; however a larger dataset is needed to further evaluate this. The observation that the HCBA (as well as the Desolation Valley flow to a degree) exhibits geochemical traits of both NEB and adakite suggests that some LSA magmas could be generated from differentiation/fractionation of primary NEB magmas, especially as the proposed source region for both NEB and LSA is mantle peridotite that has been altered by slab melts (Sajona et al.

1996, Martin et al. 2005). It must be noted that while the data presented in this study does support: (1) a slab melt model for generating adakites in the GVC, (2) the generation of NEB from slab melt metasomatised mantle wedge, and (3)

LSA magmas could be generated from primary NEB magmas, more extensive sampling and analysis is needed to more rigorously test these hypotheses and 106 the smaller dataset carries with it a degree of speculation as to magma origins.

Additionally, while future studies may challenge this interpretation for the presence of adakite (and NEB) in the GVC, this work introduces new perspectives and ideas that are still worthwhile to evaluate in future studies.

107

References

Alonso-Perez, R., Müntener, O., Ulmer, P. (2009). Igneous garnet and amphibole fractionation in the roots of island arcs: experimental constraints on andesitic liquids. Contributions to Mineralogy and Petrology, vol. 157, p. 541-

558.

Aguillón-Robles, A., Calmus, T., Benoit, M., Bellon, H., Maury, R. C., Cotten, J.,

Bourgois, J., Michaud, F. (2001). Late Miocene adakites and Nb-enriched basalts from Vizcaino Peninsula, Mexico: Indicators of East Pacific Rise subduction below southern Baja California? Geology, vol. 29, p. 531-534.

Allègre, C. J., Minster, J. F. (1978). Quantitative models of trace element behaviour in magmatic processes. Earth and Planetary Science Letters, vol. 38, p. 1-25.

Audet, P., Bostock, M. G., Mercier, J. P., Cassidy, J. F. (2008). Morphology of the Explorer-Juan de Fuca slab edge in northern Cascadia: imaging plate capture at a ridge-trench-transform triple junction. Geology, vol. 36, p. 895-898.

Barth, M. G., Foley, S. F., Horn, I. (2002). Partial melting in Archean subduction zones: constraints from experimentally determined trace element partition coefficients between eclogitic minerals and tonalitic melts under upper mantle conditions. Precambrian Research, vol. 113, p. 323-340. 108

Borg, L. E., Clynne, M. A., Bullen, T. D. (1997). The variable role of slab-derived fluids in the generation of a suite of primitive calc-alkaline lavas from the southernmost Cascades, California. The Canadian Mineralogist, vol. 35, p. 425-

452.

Borg, L. E., Clynne, M. A. (1998). The petrogenesis of felsic calc-alkaline magmas from the southernmost Cascades, California: Origin by partial melting of basaltic lower crust. Journal of Petrology, vol. 39, p. 1197-1222.

Bourdon, E., Eissen, J. P., Gutscher, M. A., Monzier, M., Samaniego, P., Robin,

C., Bollinger, C., Cotton, J. (2002). Slab melting and slab metasomatism in the

Northern Andean Volcanic Zone: adakites and high-Mg andesites from

Pichincha volcano (Ecuador). Bulletin de la Société Géologique de France, vol.

173, p. 195-206.

Breitsprecher, K., Thorkelson, D. J., Groome, W. G., Dostal, J. (2003).

Geochemical confirmation of the Kula-Farallon slab window beneath the Pacific

Northwest in Eocene time. Geology, vol. 31, p. 351-354.

Brooks, G. R., Friele, P. A. (1992). Bracketing ages for the formation of the Ring

Creek lava flow, Mount Garibaldi volcanic field, southwestern British Columbia.

Canadian Journal of Earth Sciences, vol. 29, p. 2425-2428.

Burwash, E. M. (1914). Pleistocene Vulcanism of the Coast Range of British

Columbia. Journal of Geology, vol. 22, p. 260-267. 109

Castillo, P. R. (2006). An overview of adakite petrogenesis. Chinese Science

Bulletin, vol. 51, p. 257-268.

Castillo, P. R. (2008). Origin of the adakite-high-Nb basalt association and its implications for postsubduction magmatism in Baja California, Mexico.

Geological Society of America Bulletin, vol. 120, p. 451-462.

Castillo, P. R., Janney, P. E., Solidum, R. U. (1999). Petrology and geochemistry of Camiguin Island, southern Philippines: insights to the source of adakites and other lavas in a complex arc setting. Contributions to Mineralogy and Petrology, vol. 134, p. 33-51.

Castillo, P. R., Rigby, S. J., Solidum, R. U. (2007). Origin of high field strength element enrichment in volcanic arcs: Geochemical evidence from the Sulu Arc, southern Philippines. Lithos, vol. 97, p. 271-288.

Chadwick J., Perfit, M., Ridley, I., Jonasson, I., Kamenov, G., Chadwick, W.,

Embley, R., le Roux, P., Smith, M. (2005). Magmatic effects of the Cobb hot spot on the Juan de Fuca Ridge. Journal of Geophysical Research, vol. 110, B03101.

Chiaradia, M. (2009). Adakite-like magmas from fractional crystallization and melting – assimilation of mafic lower crust (Eocene Macuchi arc, western

Cordillera, Ecuador). Chemical Geology, vol. 265, p. 468-487. 110

Chiaradia, M., Müntener, O., Beate, B., Fontignie, D. (2009). Adakite-like volcanism of Ecuador: lower crust magmatic evolution and recycling.

Contributions to Mineralogy and Petrology, vol. 158, p. 563-588.

Coldwell, B., Adam, J., Rushmer, T., Macpherson, C. G. (2011). Evolution of the

East Philippine Arc: experimental constraints on magmatic phase relations and adakitic melt formation. Contributions to Mineralogy and Petrology, vol. 162, p.

835-848.

Cousens, B. L., Allan, J. F., Leybourne, M. I., Chase, R. L., Wagoner, N. V.

(1995). Mixing of magmas from enriched and depleted mantle sources in the northeast Pacific: West Valley segment, Juan de Fuca Ridge. Contributions to

Mineralogy and Petrology, vol. 120, p. 337-357.

Dawes, R. L., Green, N. L., Defant, M. J., Drummond, M. S. (1994). Mount St.

Helens: Potential example of the partial melting of the subducted lithosphere in a volcanic arc: Comments and Reply. Geology, vol. 22, p. 187-190.

Defant, M. J., Drummond, M. S. (1990). Derivation of some modern arc magmas by melting of young subducted lithosphere. Nature, vol. 367, p. 662-665.

Defant, M. J., Richerson, P. M., De Boer, J. Z., Stewart, R. H., Maury, R. C.,

Bellon, H., Drummond, M. S., Feigenson, M. D., Jackson, T. E. (1991). Dacite

Genesis via both Slab Melting and Differentiation: Petrogenesis of La Yeguada

Volcanic Complex, Panama. Journal of Petrology, vol. 32, p. 1101-1142. 111

Dessimoz, M., Müntener, O., Ulmer, P. (2012). A case for hornblende dominated fractionation of aarc magmas: the Chelan Complex (Washington Cascades).

Contributions to Mineralogy and Petrology, vol. 163, p. 567-589.

Drummond, M. S., Defant, M. J. (1990). A model for trondhjemite-tonalite-dacite genesis and crustal growth via slab melting: Archean to modern comparisons.

Journal of Geophysical Research, vol. 95, p. 21503-21521.

Falloon, T. J., Danyushevsky, L. V., Crawford, A. J., Meffre, S., Woodhead, J.

D., Bloomer, S. H. (2008). Boninites and adakites from the Northern Termination of the Tonga Trench: implications for adakite petrogenesis. Journal of Petrology, vol. 49, p. 697-715.

Fiesinger, D. W. (1975). Petrology of the Quaternary volcanic centres in the

Quesnel Highlands and Garibaldi Provincial Park areas, British Columbia. PhD

Dissertation, University of Calgary, 132p.

Fillmore, J., Coulson, I. M. (2013). Petrological and geochemical constraints on the origin of adakites in the Garibaldi Volcanic Complex, southwestern British

Columbia, Canada. Bulletin of Volcanology, vol. 75, p. 730-753.

Foley, S. F., Barth, M. G., Jenner, G. A. (2000). Rutile/melt partition coefficients for trace elements and an assessment of the influence of rutile on the trace element characteristics of subduction zone magmas. Geochimica and

Cosmochimica Acta, vol. 64, p. 933-938. 112

Gao, S., Rudnick, R. L., Xu, W., Yuan, H., Liu, Y., Walker, R. J., Puchtel, I. S.,

Liu, X., Huang, H., Wang, X., Yang, J. (2008). Recycling deep cratonic lithosphere and generation of intraplate magmatism in the North China Craton.

Earth and Planetary Science Letters, vol. 270, p. 41-53.

Garrido, C. J., Bodinier, J. L., Burg, J. P., Zeilinger, G., Hussain, S., Dawood, H.,

Nawaz Chaudhry, M., Gervilla, F. (2006). Petrogenesis of mafic garnet granulite in the lower crust of the Kohistan Paleo-arc Complex (Northern Pakistan): implications for intra-crustal differentiation of island arcs and generation of continental crust. Journal of Petrology, vol. 47, p. 1873-1914.

Geochemical Earth Reference Model (GERM) Reservoir Database (2013). earthref.org/GERMRD/datamodel/ oceanic crust, N-MORB, primitive mantle, dacite, andesite, basalts. Accessed: Feb 1-4, 2013.

Gómez-Tuena, A., Langmuir, C. H., Goldstein, S. L., Straub, S. M., Ortega-

Gutiérrez, F. (2007). Geochemical Evidence for Slab Melting in the Trans-

Mexican Volcanic Belt. Journal of Petrology, vol. 4, p. 537-562.

Govindaraju, K., Potts, P. J., Webb, P. C., Watson, J. S. (1994). Report on Whin

Sill dolerite WS-E from England and Pitscurrie microgabbro PM-S from Scotland: assessment by one hundred and four international laboratories. Geostandards

Newsletter, vol. 18, p. 211-300. 113

Green, N.L. (1977). Multistage andesite genesis in the Garibaldi Lake area, southwestern British Columbia. PhD Dissertation, University of British Columbia,

265p.

Green, N. L. (1981). Geology and petrology of Quaternary volcanic rocks,

Garibaldi Lake area, southwestern British Columbia. Geological Society of

America Bulletin, vol. 92, p. 697-702.

Green, N. L. (1990). Late Cenozoic volcanism in the Mount Garibaldi and

Garibaldi Lake volcanic fields, Garibaldi volcanic belt, southwestern British

Columbia. Geoscience Canada, vol. 17, p. 171-174.

Green, N. L. (2006). Influence of slab thermal structure on basalt source regions and melting conditions: REE and HFSE constraints from the Garibaldi volcanic belt, northern Cascadia subduction system. Lithos, vol. 87, p. 23-49.

Green, N. L., Henderson, P. (1984). Rare earth element concentrations in

Quaternary volcanic rocks of southwestern British Columbia. Canadian Journal of Earth Sciences, vol. 21, p. 731-736.

Green, N. L., Armstrong, R. L., Harakal, J. E., Souther, J. G., Read, P. B. (1988).

Eruptive history and K-Ar geochronology of the late Cenozoic Garibaldi volcanic belt, southwestern British Columbia. Geological Science of America Bulletin, vol.

100, p. 563-579. 114

Green, N. L., Harry, D. L. (1999). On the relationship between subducted slab age and arc basalt petrogenesis, Cascadia subduction system, North America.

Earth and Planetary Science Letters, vol. 171, p. 367-381.

Green, T. H., Blundy, J. D., Adam, J., Yaxley, G. M. (2000). SIMS determination of trace element partition coefficients between garnet, clinopyroxene and hydrous basaltic liquids at 2-7.5 GPa and 1080-1200°C. Lithos, vol. 53, p. 165-

187.

Green, N. L., Sinha, A. K. (2005). Consequences of varied slab age and thermal structure on enrichment processes in the sub-arc mantle of the northern

Cascadia subduction system. Journal of Volcanology and Geothermal Research, vol. 140, 107-132.

Guffanti, M., Clynne, M. A., Smith, J. G., Muffler, L. J. P., Bullen, T. D. (1990).

Late Cenozoic volcanism, subduction and extension in the Lassen region of

California, southern Cascade range. Journal of Geophysical Research, vol. 95, p. 19453-19464.

Guo, Z., Wilson, M., Liu, J. (2006). Post-collisional adakites in south Tibet:

Products of partial melting of subduction-modified lower crust. Lithos, vol. 96, p.

205-224.

Harry, D. L., Green, N. L. (1999). Slab dehydration and basalt petrogenesis in subduction systems involving very young oceanic lithosphere. Chemical

Geology, vol. 160, p. 309-333. 115

Hastie, A. R., Mitchell, S. F., Kerr, A. C., Minifie, M. J., Millar, I. L. (2011).

Geochemistry of rare high-Nb basalt lavas: Are they derived from a mantle wedge metasomatized by slab melts? Geochimica and Cosmochimica Acta, vol.

75, p. 5049-5072.

Hegner, E., Tatsumoto, M. (1989). Pb, Sr, and Nd Isotopes in Seamount Basalts

From the Juan de Fuca Ridge and Kodiak-Bowie Seamount Chain, Northeast

Pacific. Journal of Geophysical Research, vol. 94, p. 17839-17846.

Herzberg, C. (2011). Identification of source lithology in the Hawaiian and

Canary Islands: implications for origins. Journal of Petrology, vol. 52, p. 113-146.

Hickson, C. J., Russell, J. K., Stasiuk, M. V. (1999). Volcanology of the 2350

B.P. eruption of Mount Meager volcanic complex, British Columbia, Canada: implications for hazards from eruptions in topographically complex terrain.

Bulletin of Volcanology, vol. 60, p. 489-507.

Huang, F., He, Y. (2010). Partial melting of the dry mafic continental crust: implications for petrogenesis of C-type adakites. Chinese Science Bulletin, vol.

55, p. 2428-2439.

Ickert, R. B. (2006). Adakitic Volcanism in Southern BC during the early Eocene:

Isotopic and geochemical constraints from the Princeton Group. MSc thesis,

University of Alberta, 116p. 116

Ickert, R. B., Thorkelson, D. J., Marshall, D. D., Ullrich, T. D. (2009). Eocene adakitic volcanism in southern British Columbia: remelting of arc basalt above a slab window. Tectonophysics, vol. 464, p. 164-185.

Jahn, B., Glickson, A. Y., Peucat, J.-J., Hickman, A. H. (1981). REE geochemistry and isotopic data of Archaean silicic volcanics and granitoids from the Pilbara block, western Australia: implications for early crustal evolution.

Geochimica and Cosmochimica Acta, vol. 45, p. 1633-1652.

Karsten, J. L., Delaney, J. R., Rhodes, J. M., Liias, R. A. (1990). Spatial and

Temporal Evolution of Magmatic Systems Beneath the Endeavour Segment,

Juan de Fuca Ridge: Tectonic and Petrologic Constrains. Journal of Geophysical

Research, vol. 95, p. 19235-19256.

Kay, R. W. (1978). Aleutian magnesian andesites: melts from subducted Pacific

Ocean crust. Journal of Volcanology and Geothermal Research, vol. 4, p. 117-

132.

Kelman, M. C., Russell, J. K., Hickson, C. J. (2001). Preliminary petrography and chemistry of the Mount Cayley volcanic field, British Columbia. Geological

Survey of Canada, Current Research 2001-A11, 12p.

Klein, M., Stosch, H.-G., Seck, H. A., Shimizu, N. (1999). Experimental partitioning of high field strength and rare earth elements between clinopyroxene and garnet in andesitic to tonalitic systems. Geochimica and Cosmochimica

Acta, vol. 64, p. 99-115. 117

Langmuir, C. H., Vocke, R. D. Jr., Hanson, G. N., Hart, S. R. (1978). A general mixing equation with applications to Icelandic basalts. Earth and Planetary

Science Letters, vol. 37, p. 380-392.

Le Bas, M. J., Streckeisen, A. L. (1991). The IUGS systematics of igneous rocks.

Journal of the Geological Society, London, vol. 148, p. 825-833.

Macdonald, G. A. (1968). Composition and origin of Hawaiian lavas. In: Coats,

R. R., Hay, R. L., Anderson, C. A. (eds). Studies in volcanology: a memoir in honor of Howel Williams, vol. 116, Geological Science of America Memoir, p.

477-522.

Macpherson, C. G., Dreher, S. T., Thirlwall, M. F. (2006). Adakites without slab melting: high pressure differentiation of island arc magma, Mindanao, the

Philippines. Earth and Planetary Science Letters, vol. 243, p. 581-593.

Macpherson, C. G., Chiang, K. K., Hall, R., Nowell, G. M., Castillo, P. R.,

Thirlwall, M. F. (2010). Plio-Pleistocene intra-plate magmatism from the southern

Sulu Arc, Semporna peninsula, Sabah, Borneo: Implications for high-Nb basalt in subduction zones. Journal of Volcanology and Geothermal Research, vol.

190, p. 25-38.

Madsen, J. K., Thorkelson, D. J., Friedman, R. M., Marshall, D. D. (2006).

Cenozoic to recent plate configurations in the Pacific basin: ridge subduction and slab window magmatism in western North America. Geosphere, vol. 2, p. 11-34. 118

Martin, H (1999). The adakitic magmas: modern analogues of Archaean granitoids. Lithos, vol. 46, p. 411-429.

Martin, H., Moyen, J. F. (2003). Secular changes in TTG composition: comparison with modern adakite. EGS-AGU-EUG Joint Meeting, Nice, Apr

VGP7-1 FR2O-001.

Martin, H., Smithies, R. H., Rapp, R., Moyen, J. F., Champion, D. (2005). An overview of adakite, tonalite-trondhjemite-granodiorite (TTG) and sanukitoid: relationships and some implications for crustal evolution. Lithos, vol. 79, p. 1-24.

Martindale, M., Mullen, E., Weis, D. (2014). Geochemical constraints on the origin of evolved magmas in the Northern Cascades. Goldschmidt 2014, June 9-

13; Sacramento, California; p. 1604.

Mathews, W. H. (1951). The Table, a flat topped volcano in southern British

Columbia. American Journal of Science, vol. 249, p. 830-841.

Mathews, W. H. (1952). Mount Garibaldi, a supraglacial volcano in southwestern

British Columbia. American Journal of Science, vol. 250, p. 81-103.

Mathews, W. H. (1957). Petrology of Quaternary Volcanics of the Mount

Garibaldi Map-Area, Southwestern British Columbia. American Journal of

Science, vol. 255, p. 400-415. 119

Mathews, W. H. (1958). Geology of the Mount Garibaldi map-area, southwestern

British Columbia, Canada. Geological Science of America Bulletin, vol. 69, p.

161-198.

McKenzie, D., Bickle, M. J. (1988). The Volume and Composition of Melt

Generated by Extension of the Lithosphere. Journal of Petrology, vol. 29, p. 625-

679.

Moyen, J. F. (2009). High Sr/Y and La/Yb ratios: the meaning of the “adakitic signature”. Lithos, vol. 112, p. 556-574.

Moyen, J. F., Martin, H. (2012). Forty years of TTG research. Lithos, vol. 148, p.

312-336.

Müntener, O., Kelemen, P. B., Grove, T. L. (2001). The role of H2O during crystallization of primitive arc magmas under uppermost mantle conditions and genesis of igneous pyroxenites: an experimental study. Contributions to

Mineralogy and Petrology, vol. 141, p. 643-685.

Panteeva, S. V., Gladkochoub, D. P., Donskaya, T. V., Markova, V. V.,

Sandimirova, G. P. (2003). Determination of 24 trace elements in felsic rocks by inductively coupled plasma mass spectrometry after lithium metaborate fusion.

Spectrochimica Acta, part B, vol. 58, p. 341-350. 120

Perry, H. K. C., Eaton, D. W. S., Forte, A. M. (2002). LITH5.0: a revised crustal model for Canada based on lithoprobe results. Geophysical Journal

International, vol. 150, p. 285-294.

Plank, T., Langmuir, C. H. (1998). The chemical composition of subducting sediment and its consequences for the crust and mantle. Chemical Geology, vol.

145, p. 325-394.

Rapp, R. P. (1995). Amphibole-out phase boundary in partially melted metabasalt, its control over liquid fraction and composition and source permeability. Journal of Geophysical Research, vol. 100, no. B8, p. 15601-

15610.

Rapp, R. P., Watson, E. B. (1995). Dehydration Melting of Metabasalt at 8-32 kbar: Implications for Continental Growth and Crust-Mantle Recycling. Journal of

Petrology, vol. 36, p. 891-931.

Rapp, R. P., Shimizu, N., Norman, M. D., Applegate, G. S. (1999). Reaction between slab-derived melts and peridotite in the mantle wedge: experimental constraints at 3.8 GPa. Chemical Geology, vol. 160, p. 335-356.

Richards, J. P., Kerrich, R. (2007). Special paper: adakite-like rocks: their diverse origins and questionable role in metallogenesis. Economic Geology, vol.

102, p.537-576. 121

Riddihough, R. P. (1981). One hundred million years of plate tectonics in western Canada. Geoscience Canada, vol. 9, p. 28-34.

Riddihough, R. P. (1984). Recent movements of the Juan de Fuca Plate system.

Journal of Geophysical Research, vol. 89 (B8), p. 6980-6994.

Rollinson, H. R. (1993). Using Geochemical Data: Evaluation, Presentation,

Interpretation. Longman Scientific and Technical, London. 352p.

Rooney, T. O., Franceschi, P., Hall, C. M. (2011). Water-saturated magmas in the Panama Canal region: a precursor to adakite-like magma generation?

Contributions to Mineralogy and Petrology, vol. 161, p. 373-388.

Rusmore, M. E., Woodsworth, G. J. (1991). Coast Plutonic Complex: A mid-

Cretaceous contractional orogeny. Geology, vol. 19, p. 941-944.

Sajona, F. G., Maury, R. C., Bellon, H., Cotten, J., Defant, M., Pubellier, M.

(1993). Initiation of subduction and the generation of slab melts in the western and eastern Mindanao, Philippines. Geology, vol. 21, p. 1007-1010.

Sajona, F. G., Bellon, H., Maury, R. C., Pubellier, M., Cotten, J., Rangin, C.

(1994). Magmatic response to abrupt changes in tectonic setting: Pliocene-

Quaternary calc-alkaline lavas and Nb-enriched basalts of Leyte and Mindanao

(Philippines). Tectonophysics, vol. 237, p. 47-72.

Sajona, F. G., Maury, R. C., Bellon, H., Cotten, J., Defant, M. (1996). High Field

Strength Element Enrichment of Pliocene-Pleistocene Island Arc Basalts, 122

Zamboanga Peninsula, Western Mindanao (Philippines). Journal of Petrology, vol. 37, p. 693-726.

Samaniego, P., Martin, H., Robin, C., Monzier, M. (2002). Transition from calc- alkalic to adakitic magmatism at Cayambe volcano, Ecuador: Insights into slab melts and mantle wedge interactions. Geology, vol. 30, p. 967-970.

Schiano, P., Clocchiatti, R., Shimizu, N., Maury, R. C., Jochum, K. P., Hofmann,

A. W. (1995). Hydrous silica-rich melts in the sub-arc mantle and their relationship with erupted arc lavas. Nature, vol. 377, p. 595-600.

Schiano, P., Monzier, M., Eissen, J. P., Martin, H., Koga, K. T. (2010). Simple mixing as the major control of the evolution of volcanic suites in the Ecuadorian

Andes. Contributions to Mineralogy and Petrology, vol. 160, p. 297-312.

Sherrod, D. R., Smith, J. G. (1990). Quaternary extrusion rates of the Cascade

Range, northwestern United States and southern British Columbia. Journal of

Geophysical Research, vol. 95, p. 19465-19474.

Shirey, S. B., Hanson, G. N. (1984). Mantle-derived Archaean monzodiorites and trachyandesites. Nature, vol. 310, p. 222-224.

Sigurdsson, H, Houghton, B, McNutt, S. R., Rymer, H., Stix, J. (eds) (2000).

Encyclopedia of volcanoes. Academic Press, California, 1417p. 123

Sivertz, G. W. G. (1976). Geology, petrology and petrogenesis of Opal Cone and the Ring Creek lava flow, southern Garibaldi, British Columbia. B.Sc. Thesis,

University of British Columbia, 79p.

Sobolev, A. V., Hofmann, A. W., Sobolev, S. V., Nikogosian, I. K. (2005). An olivine-free mantle source of Hawaiian shield basalts. Nature, vol. 434, p. 590-

597.

Sobolev, A. V., Hofmann, A. W., Kuzmin, D. V., Yaxley, G. M., Arndt, N. T.,

Chung, S., Danyushevsky, L. V., Elliott, T., Frey, F. A., Garcia, M. O., Gurenko,

A. A., Kamenetsky, V. S., Kerr, A. C., Krivolutskaya, N. A., Matvienkov, V. V.,

Nikogosian, I. K., Rocholl, A., Sigurdsson, I. A., Sushchevskaya, N. M., Teklay,

M. (2007). The Amount of Recycled Crust in Sources of Mantle-Derived Melts.

Science, vol. 316, p. 412-417.

Souther, J. G. (1970). Volcanism and its relationship to recent crustal movements in the Canadian Cordillera. Canadian Journal of Earth Sciences, vol.

7, p. 553-568.

Stasiuk, M. V., Russell, J. K., Hickson, C. J. (1994). Influence of magma chemistry on eruption behaviour from the distribution and nature of the 2400

Y.B.P. eruption products of Mount Meager, British Columbia. Geological Survey of Canada Open File 2843, 38p.

Straub, S. M., LaGatta, A. B., Martin-Del Pozzo, A. L., Langmuir, C. H. (2008).

Evidence from high-Ni olivines for a hybridized peridotite/pyroxenite source for 124 orogenic andesites from the central Mexican Volcanic Belt. Geochemistry,

Geophysics, Geosystems, vol. 9, 33p.

Straub, S. M., Gomez-Tuena, A., Stuart, F. M., Zellmer, G. F., Espinasa-Perena,

R., Cai, Y., Iizuka, Y. (2011). Formation of hybrid arc andesites beneath thick continental crust. Earth and Planetary Science Letters, vol. 303, p. 337-347.

Stern, C. R., Kilian, R. (1996). Role of subducted slab, mantle wedge and continental crust in the generation of adakites from the Austral Volcanic Zone.

Contributions to Mineralogy and Petrology, vol. 123, p. 263-281.

Sun, S. S., McDonough, W. I. (1989). Chemical and isotopic systematics of oceanic basalts: implications for mantle composition and processes. In:

Saunders, A. D., Norry, M. J. (eds). Magmatism in the ocean basins. Special

Publication 42, Geological Society of London, p. 313-345.

Tatsumi, Y., Hamilton, D. L., Nesbitt, R. W. (1986). Chemical characteristics of fluid phase from the subducted lithosphere: evidence from high-pressure experiments and natural rocks. Journal of Volcanology and Geothermal

Research, vol. 29, p. 293-309.

Thorkelson, D. J., Breitspretcher, K. (2005). Partial melting of slab window margins: genesis of adakitic and non-adakitic magmas. Lithos, vol. 79, p. 25-41. 125

Thorkelson, D. J., Madsen, J. K., Sluggett, C. L. (2011). Mantle flow through the

Northern Cordilleran slab window revealed by volcanic geochemistry. Geology, vol. 39, p. 267-270.

Turner, J. S., Campbell, I. H. (1986). Convection and mixing in magma chambers. Earth Science Reviews, vol. 23, p. 255-352.

Wang, Q., McDermott, F., Xu, J., Bellon, H., Zhu, Y. (2005). Cenozoic K-rich adakitic volcanic rocks in the Hohxil area, northern Tibet: Lower-crustal melting in an intracontinental setting. Geology, vol. 33, p. 465-468.

Wilson, D. S. (1988). Tectonic history of the Juan de Fuca ridge over the last 40 million years. Journal of Geophysical Research, vol. 93, p. 11863-11876.

Wilson, M. (1989). Igneous petrogenesis. HarperCollinsAcademic, London,

485p.

Wilson, D. S. (2002). The Juan de Fuca plate and slab: isochron structure and

Cenozoic plate motions. In: Kirby, S. H., Wang, K., Dunlop, S. G. (eds). The

Cascadia subduction zone and related subduction systems. US Geological

Survey Open File Report 02-328, p. 9-12.

Wolf, M. B., Wyllie, P. J (1993). Garnet Growth during Amphibolite Anatexis:

Implications of a Garnetiferous Restite. The Journal of Geology, vol. 101, p. 357-

373. 126

Xiao, L., Zhang, H. F., Clemens, J. D., Wang, Q. W., Kan, Z. Z., Wang, K. M., Ni,

P. Z., Liu, X. M. (2007). Late Triassic granitoids of the eastern margin of the

Tibetan Plateau: Geochronology, petrogenesis and implications for tectonic evolution. Lithos, vol. 96, p. 436-452.

127

Appendix A: Whole Rock Geochemistry

128

0.27

0.05

0.28

1.16

4.37

6.70

3.61

0.12

6.21

0.78

15.92

51.13

18.61

57.46

100.15

733.00

10JF018

0.19

0.21

0.33

0.89

4.61

6.39

3.42

0.14

7.25

0.96

99.91

16.46

45.92

18.51

56.31

881.00

10JF017

0.27

1.02

0.21

1.14

4.19

6.62

4.96

0.10

5.34

0.65

11.80

62.59

16.90

58.51

100.34

713.00

09JF011

0.37

0.77

0.16

1.64

4.46

4.52

1.95

0.09

3.49

0.41

99.73

33.11

50.11

17.19

64.56

706.00

09JF010

0.39

1.48

0.16

1.71

4.44

4.39

1.81

0.08

3.27

0.39

35.78

49.91

17.11

64.76

100.07

674.00

09JF009

0.34

0.63

0.24

1.49

4.39

5.65

2.40

0.08

4.19

0.55

26.07

50.74

17.52

62.57

100.27

668.45

Mt. Volcanic Garibaldi Field

10JF023

0.36

0.63

0.25

1.53

4.30

5.51

2.52

0.09

4.30

0.58

99.93

24.79

51.34

17.18

62.46

660.00

10JF022

0.32

3.15

0.24

1.37

4.29

5.28

2.24

0.08

3.90

0.57

99.97

26.94

50.81

17.97

60.34

653.00

09JF008

0.32

0.78

0.26

1.39

4.41

5.78

2.37

0.09

4.34

0.57

26.10

49.57

17.71

61.86

100.13

628.00

09JF007

0.28

0.10

0.24

1.25

4.47

6.09

3.05

0.10

5.28

0.63

99.85

19.53

50.97

18.37

59.58

788.00

10JF016

0.26

0.46

0.26

1.22

4.61

6.27

3.05

0.10

5.07

0.64

19.46

51.98

18.41

59.34

100.08

686.00

09JF012

0.25

4.69

0.28

1.19

4.69

6.36

2.88

0.10

5.16

0.67

20.61

50.11

18.80

59.35

100.26

619.00

09JF006

0.27

0.20

0.28

1.26

4.72

6.13

2.76

0.09

4.86

0.65

21.80

50.54

18.72

60.16

100.46

648.00

09JF005

0.26

0.37

0.25

1.19

4.62

6.22

2.92

0.09

5.00

0.61

20.46

51.23

18.52

59.74

100.20

678.00

09JF004

0.04

0.51

1.39

4.69

7.33

5.54

0.13

6.68

0.95

59.88

16.87

55.32

HCBA

100.37

Garibaldi Lake Volcanic Garibaldi Field

10JF013

d/l

0.24

0.48

3.46

8.84

7.94

0.17

1.50

CVB

55.05

11.67

15.64

49.12

100.47

09JF003

d/l

0.24

0.49

3.47

8.83

7.93

0.17

1.52

CVB

55.23

11.57

15.66

49.10

100.40

09JF002

d/l

0.25

0.47

3.46

8.90

7.84

0.17

1.53

CVB

54.88

11.60

15.74

49.22

100.59

09JF001

/MgO

O/Na

PR - Paul Ridge (basaltic PR - Paul Ridge andesite/andesite)

CVB Valley- Cheakamus Basalt, Peak (dacite); Basaltic CP HCBA Creek - (andesite); - Columnar Helm - Blackflow BT Creek Andesite; (andesite); - Ring RC flow BFTusk (andesite); - Barrier

Mg# =Mg# Fe = Fe xMg/(Mg+Fe) as molar Total FeO where 100,

Total iron reported as FeO reported iron Total

Total

SiO

K/Rb

Mg#

LOI

CaO

MgO

MnO

FeO

TiO

SiO Sample Table A1: Major and minor element composition of investigated samples from the Garibaldi Volcanic samplesthe Garibaldi Complex. from compositionelement A1:investigated minor and of Major Table 129

7.3

3.7

8.0

7.86

54.6

0.22

1.53

0.23

1.58

0.57

2.87

0.48

3.48

1.12

3.56

3.74

0.50

1.40

15.9

13.1

19.1

40.0

48.0

30.9

40.0

16.63

27.27

12.03

471.6

101.1

868.7

100.0

10JF018

6.6

3.8

8.4

6.24

43.9

0.28

1.85

0.28

1.98

0.68

3.42

0.58

4.00

1.35

4.02

3.78

0.30

0.80

91.4

19.1

20.0

53.0

66.0

30.0

17.8

53.0

15.0

17.49

26.54

11.54

466.1

838.7

10JF017

3.5

4.0

81.5

0.18

1.17

0.17

1.18

0.43

2.11

0.36

2.48

0.98

2.98

4.04

0.88

2.67

11.8

13.3

17.0

45.0

71.0

14.0

12.91

16.46

32.12

15.11

390.7

104.2

957.9

103.0

131.9

110.0

09JF011

d/l

4.7

4.2

78.0

0.16

1.01

0.14

0.98

0.34

1.73

0.30

1.98

0.76

2.42

3.30

1.05

2.63

10.1

19.5

16.3

17.0

51.0

30.0

22.9

63.0

12.72

12.90

26.83

12.85

558.8

100.5

784.8

09JF010

d/l

4.3

9.5

79.2

0.15

1.01

0.14

0.94

0.34

1.71

0.29

1.93

0.75

2.36

3.30

1.08

2.69

91.5

21.4

15.4

14.0

22.0

42.0

20.6

63.0

10.0

12.89

12.90

27.09

13.02

580.7

752.0

09JF009

8.2

4.2

81.5

0.18

1.19

0.18

1.27

0.45

2.30

0.41

2.94

1.08

3.65

4.73

0.80

2.10

12.6

18.7

25.0

50.0

26.0

16.8

81.0

11.0

14.20

19.46

36.37

16.90

549.4

111.7

1026.8

Mt. Volcanic Garibaldi Field

10JF023

7.3

4.5

76.7

0.19

1.28

0.19

1.27

0.46

2.39

0.42

3.05

1.07

3.83

4.87

0.80

2.10

13.2

19.4

18.4

26.0

59.0

30.0

18.0

84.0

10.0

13.38

20.11

37.25

17.13

561.9

119.7

1012.2

10JF022

d/l

3.0

4.7

98.6

0.17

1.11

0.16

1.13

0.39

2.07

0.36

2.40

0.99

3.36

4.39

0.68

1.88

10.9

18.1

18.0

23.0

49.0

30.0

32.4

86.0

13.73

17.82

33.40

15.24

469.5

115.7

1078.6

09JF008

4.6

4.4

89.9

0.18

1.19

0.17

1.20

0.43

2.24

0.39

2.72

1.03

3.31

4.25

0.70

1.89

11.9

18.5

17.6

30.0

41.0

28.0

32.2

85.0

11.0

12.50

17.48

32.56

14.87

478.0

117.9

1068.4

09JF007

6.5

3.3

8.58

75.8

0.19

1.25

0.19

1.33

0.46

2.37

0.41

2.80

1.00

3.09

3.27

0.55

1.34

91.5

12.0

13.2

18.9

28.0

75.0

56.0

20.8

97.6

11.0

14.32

24.00

10.73

436.3

909.8

10JF016

d/l

4.4

5.6

89.5

0.18

1.19

0.17

1.27

0.45

2.22

0.39

2.73

0.99

3.05

3.53

0.58

1.38

92.5

11.4

14.9

17.9

35.0

39.0

30.8

99.0

10.11

14.78

26.21

12.03

441.5

1023.5

09JF012

d/l

3.3

7.4

78.7

0.20

1.31

0.19

1.33

0.49

2.47

0.42

2.86

1.09

3.24

3.86

0.63

1.54

12.5

16.0

17.9

30.0

41.0

38.0

28.8

99.0

10.44

16.09

29.27

13.68

450.6

101.9

982.4

09JF006

2.8

7.3

73.4

0.20

1.29

0.19

1.32

0.47

2.48

0.42

2.96

1.06

3.19

3.77

0.65

1.61

13.1

16.2

17.8

34.0

37.0

34.0

28.9

96.0

11.0

10.43

15.79

28.80

13.46

477.0

101.8

957.8

09JF005

d/l

2.2

5.3

9.53

83.9

0.19

1.26

0.19

1.25

0.47

2.33

0.39

2.68

1.03

3.22

3.58

0.61

1.40

90.2

11.9

14.6

17.1

33.0

41.0

43.0

31.7

97.0

15.11

26.61

12.01

441.1

1000.3

09JF004

6.7

8.5

92.2

0.23

1.60

0.24

1.68

0.61

3.12

0.53

3.83

1.40

4.75

6.54

0.71

1.65

15.4

13.2

20.0

59.0

67.0

14.0

15.46

26.62

51.68

24.74

588.3

125.8

113.0

135.2

140.0

HCBA

1419.8

Garibaldi Lake Volcanic Garibaldi Field

10JF013

d/l

8.2

6.4

5.86

0.22

1.53

0.25

1.83

0.71

3.75

0.64

4.25

1.37

3.65

2.93

8.97

0.28

0.74

93.6

15.5

18.3

82.0

24.0

CVB

13.66

20.69

141.8

431.0

102.0

171.0

211.0

184.0

09JF003

d/l

8.1

6.1

5.64

0.22

1.59

0.24

1.82

0.70

3.71

0.63

4.33

1.37

3.77

2.94

8.95

0.26

0.76

93.8

16.8

18.0

81.0

96.0

16.0

CVB

13.81

20.83

142.5

433.8

171.0

215.1

184.0

09JF002

d/l

8.3

6.6

5.66

0.23

1.63

0.25

1.94

0.73

3.88

0.65

4.33

1.41

3.81

3.03

9.25

0.28

0.76

95.2

16.3

18.3

87.0

21.0

CVB

14.23

21.19

140.5

429.6

102.0

170.0

212.3

187.0

09JF001

CVB Valley- Cheakamus Basalt; Basaltic HCBA Creek - Helm Peak; Andesite; - Black CP BT PR - Columnar - Paul flow; flow; BFRidge Creek Tusk;- - Ring Barrier RC

d/l - below detection d/l limit- below

Elemental Elemental concentrations expressed in ppm

La/Yb

Sr/Y

Sc Sample Table A2: Trace and rare earth element composition of investigated samples from the Garibaldi Volcanic samplesthe Garibaldi Complex. from composition element investigated earth of rare and A2: Trace Table 130

Appendix B: Normative Mineralogy

131

0.01

0.65

3.80

1.48

3.33

7.22

7.46

99.95

11.29

64.71

10JF018

0.01

0.76

4.63

1.84

2.30

5.68

6.67

99.96

11.59

66.48

10JF017

0.01

0.49

3.09

1.25

6.45

7.09

8.59

99.94

13.13

59.84

09JF011

0.01

0.37

2.19

0.80

7.12

0.07

99.96

10.25

60.08

19.07

09JF010

0.01

0.37

1.65

0.76

7.37

0.20

99.98

10.79

59.44

19.39

MGVF

09JF009

0.03

0.56

3.58

1.04

5.57

2.62

9.30

99.93

60.92

16.31

10JF023

0.03

0.58

3.07

1.12

6.96

2.58

9.54

99.93

59.76

16.29

10JF022

0.03

0.58

3.07

1.10

7.13

0.19

8.83

99.95

63.46

15.56

09JF008

0.03

0.60

3.00

1.08

6.88

2.54

8.64

99.95

62.07

15.11

09JF007

0.01

0.56

1.83

1.20

2.29

7.81

9.54

99.97

12.30

64.43

10JF016

0.01

0.60

3.13

1.20

9.26

3.23

7.63

9.70

99.94

65.18

09JF012

0.03

0.65

2.09

1.27

2.87

7.40

8.48

99.96

10.81

66.36

09JF006

0.01

0.65

2.54

1.23

9.45

2.34

7.81

9.99

99.95

65.93

09JF005

0.01

0.58

2.70

1.16

9.75

2.84

7.45

99.95

65.46

10.00

GLVF

09JF004

0.03

1.18

6.08

1.80

9.66

8.72

2.15

99.94

10.07

60.25

HCBA

10JF013

0.01

0.56

2.09

2.85

4.03

2.98

CVB

99.96

18.29

13.84

55.31

09JF003

0.01

0.56

2.71

2.91

4.60

3.04

CVB

99.97

16.90

13.82

55.42

09JF002

0.01

0.58

2.71

2.92

5.30

2.92

CVB

99.95

16.20

13.78

55.53

09JF001

Values expressed as wt. %

Total

Ilm

Pl Qtz Table B1: Normative Mineralogy of representative samplesthe GVC from representative of Mineralogy B1: Normative Table 132

Appendix C: Mineral compositions for the GVC rocks

133

0.26

0.02

0.41

0.16

0.00

0.01

0.004

1.765

0.008

0.208

0.029

0.985

11.76

87.61

48.13

11.52

40.06

24-1 c 24-1

100.58

10JF024

0.04

0.39

0.37

0.01

0.02

0.04

21-6

0.011

1.365

0.009

0.607

0.016

0.001

0.991

31.03

67.99

99.05

33.89

27.58

36.68

71.2

0.07

0.01

0.37

0.32

0.06

0.02

0.03

21-5

0.009

1.441

0.008

0.516

0.049

0.001

0.974

27.94

36.88

25.80

37.16

100.73

0.07

0.00

0.39

0.34

0.03

0.01

0.010

1.425

0.009

0.530

0.051

0.001

0.974

28.71

70.39

36.27

26.37

36.94

21-4 r 21-4

100.43

0.10

0.29

0.24

0.02

0.06

0.007

1.472

0.006

0.487

0.054

0.002

0.972

26.68

72.68

101.3

38.12

24.95

37.53

21-4 c 21-4

0.10

0.29

0.26

0.01

0.03

0.01

21-3

0.007

1.491

0.006

0.473

0.044

0.001

0.977

25.57

73.76

38.50

23.79

37.62

100.62

0.12

0.05

0.29

0.19

0.02

0.05

0.02

21-2

0.005

1.539

0.006

0.431

0.034

0.002

0.982

23.08

76.35

40.03

21.57

38.06

100.39

77.9

0.14

0.03

0.26

0.23

0.06

0.03

21-1

0.006

1.571

0.006

0.395

0.042

0.001

0.978

21.52

100.6

41.19

20.43

38.23

10JF021

0.29

0.43

0.16

0.01

0.02

0.01

13-4

0.004

1.791

0.009

0.173

0.047

0.976

10.85

88.51

49.07

10.72

39.88

100.58

0.13

0.03

0.48

0.18

0.02

0.01

0.005

1.804

0.010

0.156

0.051

0.974

10.19

89.09

99.94

49.32

10.06

39.70

13-3 r 13-3

0.18

0.02

0.50

0.19

0.02

0.005

1.800

0.010

0.164

0.040

0.980

10.11

89.13

49.50

10.01

40.15

13-3 c 13-3

100.61

0.11

0.49

0.18

0.01

0.005

1.797

0.010

0.169

0.038

0.981

10.22

89.04

49.41

10.11

40.21

13-2 r 13-2

100.52

calculated from charge balance of analyzed FeO of balance charge calculated from

+3 +3

89.2

0.15

0.47

0.17

9.94

0.02

0.004

1.793

0.010

0.181

0.022

0.001

0.988

10.09

49.28

40.49

13-2 c 13-2

100.55

9.94

0.08

0.03

0.41

0.18

9.79

0.00

0.02

0.005

1.810

0.009

0.152

0.049

0.975

89.41

99.62

49.42

39.70

13-1 r 13-1

0.25

0.04

0.47

0.14

0.01

0.004

1.792

0.010

0.179

0.032

0.984

10.46

88.88

49.23

10.33

40.30

13-1 c 13-1

100.78

10JF013

GLVF

10JF021 - CVB, 10JF024 - HCBA (whole rock geochemistry- HCBA (whole - CVB, not analysed) 10JF024 10JF021

c - core composition; Valley(basaltic r - rim - Desolation composition;rockandesite; flow whole chemistry 10JF025 not analysed)

Cation values expressed per 4 oxygens expressed per values Cation

Endmember compositionsEndmember in mole %; Fe

Total

NiO

MnO

CaO

MgO

FeO

TiO SiO Table C1: Electron microprobe compositions of olivine from the Garibaldi Volcanic the Garibaldi Complex from compositionsC1: Electron olivine microprobe of Table 134

0.28

0.01

0.38

0.09

0.02

0.002

1.608

0.008

0.354

0.054

0.001

0.973

20.14

79.33

99.68

42.01

19.01

37.88

25-5 c 25-5

22.2

0.12

0.03

0.43

0.16

0.02

0.01

0.004

1.564

0.009

0.395

0.055

0.001

0.972

77.12

40.76

20.92

37.77

25-4 r 25-4

100.22

0.27

0.02

0.35

0.12

0.02

0.00

0.003

1.611

0.008

0.358

0.039

0.001

0.980

19.69

79.78

42.46

18.68

38.50

25-4 c 25-4

100.44

23.3

0.13

0.01

0.48

0.19

0.02

0.03

0.005

1.543

0.010

0.409

0.064

0.001

0.967

75.94

40.21

21.98

37.55

25-3 r 25-3

100.63

0.23

0.42

0.10

0.04

0.02

0.01

0.003

1.578

0.009

0.388

0.044

0.001

0.977

21.36

78.06

41.45

20.22

38.27

25-3 c 25-3

100.75

0.12

0.01

0.47

0.16

0.04

0.01

0.005

1.556

0.010

0.399

0.059

0.001

0.970

22.61

76.67

40.48

21.28

37.61

25-2 r 25-2

100.19

0.28

0.44

0.12

0.02

0.01

0.003

1.614

0.009

0.352

0.039

0.001

0.980

19.47

79.91

42.67

18.53

38.62

25-2 c 25-2

100.69

0.08

0.02

0.52

0.21

0.03

0.01

25-1

0.006

1.499

0.011

0.456

0.056

0.001

0.971

25.24

73.91

38.53

23.45

37.22

100.08

10JF025

0.09

0.02

0.52

0.16

0.01

0.004

1.787

0.011

0.189

0.018

0.991

10.29

88.97

49.46

10.19

40.86

24-5 r 24-5

101.33

0.23

0.00

0.42

0.14

0.02

0.004

1.781

0.009

0.192

0.029

0.001

0.985

10.96

88.44

49.00

10.82

40.41

24-5 c 24-5

101.07

0.39

0.02

0.40

0.14

0.02

24-4

0.004

1.783

0.008

0.184

0.041

0.001

0.979

11.13

88.28

48.58

10.92

39.76

101.07

0.14

0.00

0.46

0.19

0.005

1.797

0.010

0.172

0.034

0.983

10.21

89.07

49.29

10.07

40.21

24-3 r 24-3

100.37

calculated from charge balance of analyzed FeO of balance charge calculated from

+3 +3

0.27

0.03

0.37

0.17

0.00

0.03

0.004

1.774

0.008

0.192

0.043

0.001

0.978

11.66

87.75

48.56

11.50

39.91

24-3 c 24-3

100.85

0.13

0.48

0.16

0.03

0.01

0.004

1.792

0.010

0.172

0.043

0.001

0.978

10.64

88.66

49.01

10.48

39.86

24-2 r 24-2

100.19

0.36

0.25

0.15

0.01

0.00

0.004

1.748

0.005

0.223

0.038

0.981

12.94

86.59

47.36

12.62

39.60

24-2 c 24-2

100.35

0.11

0.03

0.48

0.15

0.02

0.004

1.764

0.010

0.208

0.028

0.001

0.986

11.74

87.57

100.6

48.17

11.51

40.13

24-1 r 24-1

10JF021 - CVB, 10JF024 - HCBA (whole rock geochemistry- HCBA (whole - CVB, not analysed) 10JF024 10JF021

c - core composition; Valley(basaltic r - rim - Desolation composition;rockandesite; flow whole chemistry 10JF025 not analysed)

Cation values expressed per 4 oxygens expressed per values Cation Endmember compositionsEndmember in mole %; Fe Table C1 cont'd C1 Table 135

0.13

0.34

0.11

0.03

0.00

0.02

18-4

0.003

1.546

0.008

0.424

0.039

0.001

0.980

22.89

76.59

40.24

21.44

38.03

100.35

23.3

0.13

0.03

0.31

0.12

0.01

0.02

0.003

1.542

0.007

0.424

0.048

0.976

76.21

40.30

21.96

38.00

18-3 r 18-3

100.88

0.17

0.04

0.33

0.11

0.02

0.03

0.003

1.550

0.007

0.419

0.040

0.001

0.979

22.75

76.75

40.84

21.58

38.44

18-3 c 18-3

101.58

0.18

0.02

0.38

0.13

0.04

0.02

0.004

1.517

0.008

0.446

0.048

0.001

0.975

24.42

74.97

39.31

22.82

37.66

18-2 r 18-2

100.57

0.14

0.33

0.11

0.02

0.03

0.003

1.551

0.007

0.413

0.049

0.001

0.974

22.86

76.63

40.34

21.45

37.77

18-2 c 18-2

100.21

0.10

0.02

0.40

0.10

0.02

0.003

1.492

0.009

0.480

0.032

0.001

0.984

25.41

74.02

38.70

23.68

38.03

18-1 r 18-1

101.07

0.10

0.34

0.13

0.02

0.03

0.01

0.003

1.553

0.007

0.422

0.026

0.001

0.986

22.27

77.19

100.9

40.75

20.96

38.56

18-1 c 18-1

0.08

0.41

0.10

0.03

0.003

1.518

0.009

0.453

0.034

0.001

0.983

24.14

75.28

39.47

22.56

38.09

18-1 r 18-1

100.74

10JF018

0.12

0.01

0.53

0.16

0.02

0.04

17-9

0.004

1.358

0.012

0.600

0.050

0.001

0.974

32.11

67.07

34.07

29.08

36.42

100.46

0.09

0.45

0.10

0.05

0.03

17-8

0.003

1.455

0.010

0.505

0.054

0.001

0.971

27.61

71.89

37.50

25.67

37.34

101.28

0.08

0.02

0.71

0.17

0.02

0.01

17-7

0.005

1.350

0.016

0.604

0.048

0.001

0.975

32.25

66.71

33.86

29.18

36.46

100.52

0.05

0.00

0.68

0.19

0.02

0.01

17-4

0.006

1.313

0.016

0.650

0.031

0.001

0.984

33.79

65.17

32.62

30.15

36.44

100.18

calculated from charge balance of analyzed FeO of balance charge calculated from

+3 +3

0.10

0.03

0.59

0.22

0.02

0.01

17-2

0.006

1.376

0.013

0.573

0.062

0.968

31.29

67.73

34.64

28.53

36.35

100.51

10JF017

MGVF

0.12

0.05

0.58

0.22

0.03

0.00

0.006

1.514

0.013

0.445

0.043

0.001

0.978

24.16

74.91

39.03

22.44

37.59

25-6 r 25-6

100.05

0.05

0.06

0.53

0.17

0.04

0.01

0.005

1.519

0.012

0.439

0.049

0.001

0.975

24.14

75.06

39.57

22.68

37.85

25-6 c 25-6

100.96

0.10

0.40

0.20

0.02

0.01

0.005

1.550

0.009

0.409

0.055

0.001

0.972

22.87

76.44

40.67

21.69

38.03

25-5 r 25-5

101.11

10JF021 - CVB, 10JF024 - HCBA (whole rock geochemistry- HCBA (whole - CVB, not analysed) 10JF024 10JF021

c - core composition; Valley(basaltic r - rim - Desolation composition;rockandesite; flow whole chemistry 10JF025 not analysed)

Cation values expressed per 4 oxygens expressed per values Cation Endmember compositionsEndmember in mole %; Fe Table C1 cont'd C1 Table 136

0.02

0.36

0.22

7.69

0.17

2.99

0.72

0.026

0.827

0.868

0.007

0.142

0.099

0.005

0.132

1.875

42.72

12.41

44.86

98.62

20.66

15.59

50.21

25-1 c 25-1

10JF025

0.02

0.81

0.21

6.62

3.28

1.17

0.058

0.839

0.875

0.007

0.038

0.166

0.142

0.033

1.842

43.73

10.63

45.63

99.24

21.23

15.92

49.97

24-9 gm 24-9

0.02

0.74

0.18

6.57

3.25

1.22

0.054

0.904

0.849

0.006

0.209

0.145

0.035

1.799

46.07

10.62

43.31

96.62

22.23

15.02

47.41

24-8 r 24-8

0.02

0.03

0.42

0.20

6.65

0.02

2.75

0.71

0.031

0.876

0.851

0.006

0.109

0.101

0.123

1.883

45.21

10.86

43.92

97.96

21.61

15.81

49.78

24-8 c 24-8

0.89

0.12

1.87

0.00

0.35

0.20

6.51

0.03

2.76

0.68

0.025

0.866

0.006

0.077

0.125

0.001

0.019

44.24

10.31

45.45

98.92

21.82

16.11

50.46

24-7 r 24-7

0.43

0.16

6.64

3.05

0.70

0.031

0.868

0.869

0.005

0.082

0.124

0.134

0.019

1.867

44.68

10.63

44.69

98.66

21.80

15.67

50.21

24-7 c 24-7

0.84

0.80

0.24

6.96

0.01

3.50

1.37

0.058

0.864

0.007

0.032

0.184

0.153

0.038

1.822

43.75

11.27

44.98

98.51

21.08

15.58

48.97

24-6 gm 24-6

0.05

0.02

0.73

0.20

6.54

3.28

1.32

0.052

0.852

0.865

0.006

0.153

0.143

0.037

1.841

44.38

10.56

45.06

98.89

21.46

15.66

49.68

24-5 gm 24-5

0.09

0.01

9.84

0.05

0.49

0.10

6.07

0.34

3.19

0.77

24-4

0.035

0.862

0.864

0.003

0.098

0.139

0.021

1.876

45.02

45.15

98.92

21.69

15.63

50.59

0.79

0.09

40.6

0.53

0.23

7.03

2.06

0.78

24-3

0.038

0.939

0.007

0.105

0.113

0.022

1.896

11.18

48.23

98.83

19.94

17.02

51.24

44.1

0.02

0.43

0.22

7.56

0.04

3.91

0.89

24-2

0.031

0.848

0.841

0.007

0.109

0.125

0.001

0.171

0.025

1.842

12.18

43.72

99.34

21.36

15.22

49.69

42.2

0.00

0.37

0.17

6.81

0.02

2.32

0.50

24-1

0.027

0.827

0.922

0.005

0.097

0.113

0.001

0.101

0.014

1.892

10.75

47.05

20.89

16.74

51.17

10JF024

0.04

0.39

0.15

8.28

0.70

2.61

1.01

21-1

0.028

0.789

0.866

0.212

0.046

0.021

0.115

0.028

1.895

41.21

13.51

45.27

99.26

19.72

15.57

50.78

10JF021

7.46

0.06

0.53

0.12

4.61

0.91

2.55

0.45

13-6

0.038

0.854

0.917

0.004

0.055

0.088

0.027

0.111

0.013

1.893

44.65

47.89

98.41

21.51

16.59

51.07

41.7

0.51

0.09

7.40

0.00

5.09

1.23

0.037

0.883

0.797

0.003

0.067

0.165

0.225

0.035

1.789

46.17

12.13

98.34

21.99

14.27

47.75

13-5 r 13-5

0.02

0.43

0.09

6.44

0.01

2.77

0.65

0.031

0.864

0.887

0.003

0.178

0.123

0.122

0.018

1.874

44.26

10.29

45.45

98.31

21.63

15.96

50.30

13-4 c 13-4

0.89

0.38

0.09

6.57

2.65

0.67

0.028

0.848

0.003

0.115

0.089

0.116

0.019

1.892

43.65

10.51

45.83

98.98

21.30

16.08

50.94

13-3 r 13-3

calculated from charge balance of analyzed FeO of balance charge calculated from

+3 +3

0.40

0.15

7.45

0.08

4.62

1.18

0.029

0.875

0.806

0.005

0.106

0.126

0.002

0.202

0.033

1.816

45.74

12.12

42.14

99.24

21.96

14.54

48.86

13-3 c 13-3

44.7

9.99

45.3

0.05

0.37

0.16

6.34

0.02

2.62

0.62

13-2

0.026

0.876

0.887

0.005

0.075

0.121

0.001

0.114

0.017

1.878

99.26

22.12

16.11

50.85

0.82

0.03

0.04

0.40

0.11

7.01

0.12

4.36

1.08

13-1

0.029

0.879

0.004

0.093

0.126

0.004

0.192

1.823

45.82

11.41

42.76

98.69

21.98

14.74

48.84

10JF013

GLVF

c - composition;core r - rim composition; crystal gm - groundmass composition

Cation values expressed per 6 oxygens expressed per values Cation

Endmember compositionsEndmember in mole %; Fe

Total

NiO

MnO

CaO

MgO

FeO

TiO SiO Table C2: Electron microprobe compositions of clinopyroxene from the Garibaldi Volcanic the compositionsGaribaldi C2: Complex Electron from clinopyroxene microprobe of Table 137

2.95

0.02

0.40

1.52

0.00

2.15

0.32

0.001

0.059

1.417

0.012

0.412

0.104

0.091

0.009

1.894

25.92

71.13

26.32

17.10

52.45

18-6 r 18-6

100.28

2.12

0.04

0.81

1.08

0.65

0.15

0.042

1.379

0.025

0.496

0.087

0.028

0.004

1.938

29.09

68.79

99.98

25.26

19.04

52.94

18-6 c 18-6

2.58

0.02

0.07

0.35

1.33

0.07

2.49

0.27

0.005

0.051

1.467

0.011

0.346

0.127

0.002

0.106

0.007

1.878

23.75

73.67

99.72

27.31

15.69

52.12

18-5 r 18-5

2.44

68.1

0.02

0.00

0.77

1.24

0.73

0.24

0.049

1.357

0.024

0.522

0.065

0.032

0.007

1.945

29.46

99.78

24.76

19.09

52.91

18-5 c 18-5

2.83

23.5

0.02

0.40

1.46

0.06

2.02

0.26

0.002

0.056

1.464

0.012

0.376

0.091

0.002

0.086

0.007

1.904

73.67

99.84

27.26

15.50

52.86

18-4 r 18-4

1.63

0.03

0.02

0.70

0.67

0.49

0.07

0.001

0.027

1.318

0.022

0.568

0.106

0.021

0.002

1.935

33.52

64.85

23.96

21.85

52.43

18-4 c 18-4

100.21

1.63

0.00

0.69

0.82

0.57

0.07

0.033

1.297

0.022

0.603

0.067

0.025

0.002

1.952

33.52

64.85

99.88

23.46

21.61

52.65

18-3 c 18-3

2.37

0.09

0.02

0.33

1.20

0.22

3.33

0.22

0.002

0.046

1.503

0.010

0.316

0.101

0.006

0.140

0.006

1.871

21.19

76.45

28.25

13.95

52.44

18-2 r 18-2

100.06

2.39

0.06

0.02

0.37

1.23

0.03

2.65

0.20

0.001

0.047

1.460

0.011

0.378

0.091

0.001

0.112

0.006

1.893

23.74

73.87

27.27

15.62

52.73

18-2 c 18-2

100.18

0.11

2.44

22.5

0.03

0.01

0.37

1.26

0.02

2.26

0.24

0.001

0.049

1.497

0.011

0.339

0.001

0.096

0.007

1.891

75.07

99.91

28.01

14.96

52.73

18-1 r 18-1

1.7

0.73

0.86

0.51

0.09

0.034

1.301

0.023

0.597

0.068

0.022

0.003

1.952

33.22

65.08

100.1

23.61

21.48

52.81

18-1 c 18-1

10JF018

2.47

0.00

0.67

1.25

0.04

0.83

0.18

0.049

1.296

0.021

0.572

0.086

0.001

0.036

0.005

1.933

32.82

64.71

99.75

23.45

21.20

52.13

17-1 r 17-1

calculated from charge balance of analyzed FeO of balance charge calculated from

+3 +3

2.7

99.3

0.08

0.65

1.34

0.11

1.25

0.19

0.006

0.054

1.285

0.021

0.557

0.091

0.003

0.055

0.005

1.923

32.63

64.67

23.17

20.84

51.68

17-1 c 17-1

10JF017

MGVF

5.52

97.5

0.03

0.71

0.50

2.47

3.15

0.40

0.052

0.099

1.223

0.016

0.458

0.139

0.011

2.003

25.74

68.75

21.96

14.65

53.63

25-2 gm 25-2

3.92

0.04

0.01

0.53

2.03

0.02

0.92

0.36

25-1

0.001

0.079

1.473

0.016

0.361

0.100

0.001

0.039

0.010

1.920

22.93

73.15

99.31

27.23

15.22

52.95

10JF025

GLVF

c - composition;core r - rim composition; crystal gm - groundmass composition

Cation values expressed per 6 oxygens expressed per values Cation

Endmember compositionsEndmember in mole %; Fe

Total

NiO

MnO

CaO

MgO

FeO

TiO SiO Table C3: Electron microprobe compositions of orthopyroxene from the Garibaldi Volcanic the Garibaldi Complex from compositionsC3: Electron orthopyroxene microprobe of Table 138

2.23

0.02

0.38

5.94

9.11

0.11

1.49

25-2

0.022

0.527

0.447

0.008

0.057

1.390

2.550

52.92

44.85

98.61

25.79

55.76

1.97

0.05

0.34

6.33

8.73

0.12

1.25

25-1

0.020

0.556

0.001

0.424

0.008

0.047

1.380

2.563

55.61

42.42

99.12

25.81

56.49

10JF025

0.20

4.46

7.16

2.27

0.01

0.81

24-5

0.257

0.628

0.003

0.110

0.001

0.031

1.112

2.859

25.83

63.12

11.05

99.01

20.87

63.24

1.9

0.12

0.32

4.54

0.01

0.06

0.94

24-4

0.019

0.405

0.002

0.558

0.004

0.001

0.036

1.552

2.424

41.26

56.84

98.57

11.31

28.61

52.66

2.34

0.07

0.39

4.69

0.05

0.94

24-3

0.023

0.417

0.001

0.544

0.003

0.036

1.528

2.447

42.39

55.27

98.74

11.07

28.23

53.30

9.53

0.17

1.58

6.38

6.30

0.04

1.37

24-2

0.092

0.564

0.003

0.308

0.003

0.052

1.312

2.667

58.53

31.93

98.74

24.42

58.49

8.25

0.27

4.05

7.73

0.00

1.69

0.00

1.08

24-1

0.232

0.674

0.005

0.082

0.041

1.073

2.894

23.51

68.24

99.35

20.22

64.31

10JF024

0.76

0.01

0.13

3.87

0.00

0.13

0.52

21-2

0.007

0.342

0.650

0.009

0.020

1.595

2.377

34.22

65.03

99.84

13.31

29.70

52.18

0.71

0.02

0.12

3.97

0.13

0.55

21-1

0.007

0.352

0.628

0.009

0.021

1.589

2.394

35.64

63.65

99.55

12.83

29.52

52.41

10JF021

4.2

0.18

0.70

5.77

0.04

8.56

0.03

1.38

13-5

0.040

0.508

0.003

0.416

0.002

0.001

0.052

1.395

2.582

52.65

43.16

99.70

26.10

56.94

61.6

6.59

0.18

5.43

6.92

0.02

1.34

0.01

0.86

13-4

0.313

0.607

0.003

0.065

0.001

0.033

1.065

2.912

31.81

99.08

19.97

64.34

0.14

2.49

6.97

0.02

4.92

0.01

1.46

13-3

0.143

0.607

0.002

0.237

0.001

0.055

1.222

2.732

14.46

61.53

99.88

23.07

60.79

10.3

62.9

26.8

0.32

1.75

7.03

0.01

5.42

0.03

1.12

13-2

0.101

0.616

0.006

0.263

0.002

0.001

0.042

1.257

2.713

99.21

23.57

59.97

8.8

0.14

1.47

6.36

0.01

6.64

0.01

1.22

13-1

0.085

0.558

0.002

0.322

0.046

1.319

2.666

57.84

33.37

99.41

24.70

58.86

10JF013

GLVF

Cation values expressed per 8 oxygens expressed per values Cation

Endmember compositionsEndmember in mole %

Total

BaO

MnO

CaO

MgO

FeO

Al SiO Table C4: Electron microprobe compositions of plagioclase from the Garibaldi Volcanic compositions the C4: Garibaldi Electron Complex plagioclase from microprobe of Table 139

0.48

0.06

0.08

3.18

0.09

0.53

18-5

0.005

0.283

0.001

0.696

0.006

0.020

1.673

2.315

28.74

70.78

99.50

14.17

30.94

50.45

0.52

0.09

3.70

0.01

0.07

0.64

18-4

0.005

0.327

0.669

0.005

0.025

1.650

2.319

32.65

66.83

99.81

13.71

30.71

50.87

0.06

0.02

0.10

3.52

0.01

0.08

0.53

18-3

0.006

0.311

0.688

0.006

0.020

1.642

2.326

31.16

68.78

14.10

30.58

51.05

100.00

0.32

0.04

0.06

3.08

0.01

0.07

0.58

18-2

0.003

0.275

0.001

0.723

0.005

0.022

1.689

2.282

27.42

72.25

99.36

14.69

31.18

49.65

0.49

0.00

0.08

3.13

0.08

0.63

18-1

0.005

0.276

0.721

0.005

0.024

1.677

2.290

27.57

71.93

14.79

31.28

50.33

100.33

10JF018

0.41

0.07

3.71

0.01

0.07

0.72

17-7

0.004

0.329

0.664

0.005

0.028

1.642

2.328

33.01

66.58

99.38

13.54

30.42

50.84

0.52

0.06

0.09

4.05

0.00

0.06

0.69

17-6

0.005

0.359

0.001

0.640

0.004

0.027

1.618

2.345

35.74

63.74

99.36

13.07

30.03

51.29

0.53

63.9

0.09

4.00

0.03

0.06

0.78

17-5

0.005

0.353

0.635

0.004

0.001

0.030

1.624

2.348

35.58

99.74

13.00

30.24

51.54

0.76

0.02

0.13

4.21

0.02

0.05

0.73

17-4

0.007

0.371

0.617

0.004

0.001

0.028

1.588

2.383

37.28

61.96

99.86

12.66

29.63

52.41

1.96

0.12

0.34

5.91

9.58

0.01

0.31

17-3

0.019

0.514

0.002

0.461

0.012

1.450

2.542

51.72

46.33

27.42

56.64

100.32

0.7

0.06

0.12

4.28

0.07

0.71

17-2

0.007

0.379

0.001

0.612

0.004

0.027

1.593

2.376

37.99

61.31

99.28

12.50

29.56

51.98

0.98

0.11

0.17

4.11

0.04

0.38

17-1

0.010

0.362

0.002

0.619

0.003

0.014

1.605

2.385

36.55

62.46

99.91

12.71

29.96

52.44

10JF017

MGVF

0.82

0.14

3.74

0.07

0.79

25-5

0.008

0.332

0.651

0.005

0.030

1.627

2.347

33.52

65.65

99.58

13.29

30.19

51.34

0.82

39.8

0.14

4.47

0.12

0.79

25-4

0.008

0.397

0.593

0.008

0.030

1.565

2.398

59.38

98.76

12.07

28.93

52.25

0.64

35.4

0.03

0.11

3.98

0.12

0.70

25-3

0.007

0.353

0.638

0.008

0.027

1.606

2.360

63.95

99.25

13.01

29.76

51.53

Cation values expressed per 8 oxygens expressed per values Cation Endmember compositionsEndmember in mole % Table C4 cont'd C4 Table 140

0.10

1.09

0.07

6.81

0.10

2.18

7.67

0.05

24-5

0.024

0.267

0.021

2.921

4.558

0.022

0.739

1.660

0.015

97.68

18.93

60.68

13.146

0.05

0.20

0.11

2.15

0.08

0.99

0.07

24-4

0.012

0.052

0.036

0.968

6.612

0.020

0.351

2.544

0.020

95.69

26.18

54.67

11.20

12.425

0.12

1.04

0.11

6.99

0.34

1.41

6.87

0.06

24-3

0.027

0.253

0.033

2.987

4.514

0.077

0.475

1.482

0.016

98.72

18.83

62.96

13.576

0.18

0.64

0.28

6.35

0.37

1.79

3.86

0.01

24-2

0.041

0.156

0.087

2.723

4.887

0.085

0.606

0.836

0.003

99.64

20.31

65.85

14.261

99.5

0.19

0.75

0.17

8.34

0.61

1.14

2.04

0.01

24-1

0.044

0.181

0.053

3.567

4.101

0.138

0.387

0.441

0.003

17.09

69.11

14.920

10JF024

0.07

0.38

0.06

4.15

6.29

8.56

0.06

21-2

0.015

0.090

0.018

1.739

5.909

6.760

4.877

2.081

1.808

0.017

98.67

25.16

31.98

21.96

0.11

0.27

0.02

5.87

7.55

7.78

0.05

21-1

0.025

0.063

0.007

2.422

5.279

6.295

5.198

2.465

1.619

0.014

98.41

22.80

30.21

23.75

10JF021

0.19

1.11

0.05

6.00

0.29

3.45

1.57

0.05

13-7

0.044

0.269

0.015

2.563

5.064

0.065

1.166

0.339

0.014

21.12

66.41

14.329

100.23

0.14

0.33

0.11

6.23

0.78

5.73

4.11

0.04

13-6

0.032

0.078

0.033

2.605

5.142

0.173

1.893

0.866

0.012

21.92

60.83

12.837

100.22

0.02

0.20

0.08

1.98

0.08

0.95

0.01

13-5

0.006

0.052

0.025

0.888

6.680

0.019

0.336

2.795

0.002

95.92

26.56

53.68

12.36

12.149

0.04

0.42

0.05

3.70

0.11

0.65

0.03

13-4

0.009

0.107

0.017

1.646

5.935

0.027

0.230

2.282

0.010

96.34

23.79

57.36

10.17

12.878

calculated from charge balance of analyzed FeO of balance charge calculated from

+3 +3

0.36

2.43

0.10

6.89

0.21

5.53

1.54

0.06

13-3

0.077

0.549

0.029

5.767

1.538

0.044

1.738

0.309

0.016

14.51

68.82

13.812

100.45

0.15

0.58

0.16

3.13

0.18

3.76

1.77

13-2

0.036

0.148

0.052

1.410

6.304

0.043

1.338

0.403

96.68

24.93

62.03

14.115

0.26

1.42

0.09

3.88

4.14

0.04

13-1

0.057

0.325

0.025

5.221

2.265

0.019

1.239

0.843

0.011

99.91

12.93

10.00

67.07

13.674

10JF013

GLVF

18-2a & b - single 18-2a crystal ilmenite component with exsolved

Cation values expressed per 4 oxygens expressed per values Cation

Endmember compositionsEndmember in mole %; Fe

Total

NiO

MnO

CaO

MgO

FeO

TiO SiO Table C5: Electron microprobe compositionsVolcanicC5: Electron the oxides microprobe Garibaldi Complex of from Table 141

0.05

0.26

0.10

1.03

0.14

0.76

0.11

18-4

0.012

0.065

0.033

0.448

7.112

0.033

0.264

2.617

0.032

99.59

29.03

56.22

11.88

12.391

0.04

0.52

0.08

2.60

5.13

18-3

0.008

0.112

0.023

0.982

5.674

3.986

8.312

1.299

99.69

26.78

20.91

43.62

0.59

0.15

1.78

0.04

5.68

0.126

0.041

0.674

5.950

3.910

0.007

8.223

1.443

99.74

28.01

20.46

43.04

18-2b

0.08

0.41

0.07

0.87

0.17

0.83

9.05

0.35

0.019

0.103

0.021

0.385

7.207

0.040

0.289

2.021

0.104

99.17

29.05

58.29

18-2a

13.014

0.11

0.44

0.02

1.79

0.25

2.44

9.28

0.05

18-1

0.026

0.109

0.005

0.777

6.828

0.058

0.839

2.034

0.015

99.61

28.02

57.20

12.542

10JF018

0.04

0.47

0.02

1.59

0.52

0.96

0.05

17-4

0.009

0.115

0.005

0.676

6.810

0.117

0.325

3.061

0.016

28.46

54.40

14.22

11.713

100.73

0.08

0.39

0.02

2.76

1.69

3.16

0.06

17-3

0.017

0.093

0.006

1.172

6.429

0.382

1.061

2.248

0.017

26.94

54.61

10.48

11.725

100.17

0.15

0.37

0.02

2.61

0.97

4.56

8.57

0.07

17-2

0.033

0.090

0.006

1.119

6.518

0.220

1.545

1.851

0.019

99.51

27.14

55.06

11.898

0.12

0.38

0.02

2.65

0.88

4.39

8.39

0.05

17-1

0.027

0.092

0.005

1.131

6.517

0.199

1.484

1.809

0.013

99.82

27.17

55.78

12.039

10JF017

MGVF

0.04

0.43

3.06

2.36

2.52

8.78

0.06

25-4

0.008

0.105

1.307

6.341

0.536

0.853

1.896

0.018

26.43

56.60

12.219

100.28

0.02

0.51

0.02

2.65

0.95

2.23

9.71

0.05

25-3

0.005

0.126

0.006

1.136

6.463

0.217

0.757

2.103

0.015

26.83

57.10

12.378

100.07

calculated from charge balance of analyzed FeO of balance charge calculated from

+3 +3

0.08

0.52

0.04

2.04

1.05

2.30

9.05

0.13

25-2

0.020

0.127

0.011

0.879

6.713

0.239

0.781

1.965

0.037

27.81

57.45

12.478

100.46

99.8

0.11

0.46

0.07

2.80

1.24

2.11

9.62

0.09

25-1

0.025

0.112

0.021

1.203

6.375

0.283

0.718

2.089

0.026

26.41

56.89

12.355

10JF025

0.17

0.87

0.09

7.54

0.09

1.94

6.76

0.05

24-7

0.038

0.210

0.026

3.204

4.340

0.021

0.652

1.449

0.015

98.59

18.20

62.90

13.497

0.45

0.49

0.09

7.00

9.21

2.31

2.02

0.05

24-6

0.105

0.120

0.027

3.008

4.684

2.098

0.784

0.439

0.014

98.91

19.43

57.86

12.552

18-2a & b - single 18-2a crystal ilmenite component with exsolved

Cation values expressed per 4 oxygens expressed per values Cation Endmember compositionsEndmember in mole %; Fe Table C5 cont'd C5 Table 142

Appendix D: Trace Element Modelling Data

143

2.1

0.44

0.12

0.11

1.75

0.115

0.106

0.135

0.129

0.124

0.119

0.173

0.167

0.161

0.155

0.016

2.59

1.21

1.18

1.14

1.65

1.42

0.847

0.813

0.782

0.752

0.948

0.912

0.879

0.848

1.257

0.0159

2.5

8.7

8.89

13.5

12.3

1.39

9.964

9.578

9.221

10.74

10.34

9.973

9.629

13.07

12.68

0.459

1.4

6.2

Dy*

4.55

1.76

1.93

2.23

2.17

2.11

3.08

0.01

1.948

1.881

1.818

1.998

1.867

1.807

2.306

0.82

1.48

1.55

1.64

1.39

1.817

1.967

2.143

2.354

1.534

1.623

1.722

1.834

1.417

0.004

0.232

0.327

1.4

3.79

0.25

0.614

136.27

118.85

123.88

129.37

135.36

117.43

122.25

127.48

133.19

144.322

153.385

163.663

2.8

2.1

2.4

3.08

0.59

3.694

3.769

3.847

3.929

3.204

3.233

3.263

3.293

3.032

3.048

3.064

0.712

8.41

2.79

0.44

14.61

15.31

16.09

16.95

10.84

11.03

11.24

11.46

9.311

9.373

9.437

9.501

0.684

0.131

0.3

121

0.048

0.073

475.75

582.49

750.97

435.56

520.06

645.24

849.79

415.37

489.85

596.89

763.77

0.0303

1056.59

9.72

0.05

44.96

0.899

0.296

0.053

35.715

42.884

53.652

71.641

28.592

32.541

37.756

25.845

28.835

32.607

37.515

0.5

3.61

0.237

0.056

0.021

14.184

17.363

22.378

31.466

12.387

14.607

17.796

22.766

11.532

13.365

15.891

19.595

1.3

102

4.568

4.722

4.887

5.064

3.944

4.024

4.106

4.192

7.322

8.046

8.929

10.03

0.003

0.002

0.31

0.15

0.54

1.438

1.861

2.637

4.524

1.366

1.735

2.378

3.778

1.337

1.686

2.281

3.527

0.014

0.001

0.1

0.02

0.005

244.94

323.85

477.79

910.63

229.91

296.62

417.86

706.73

225.45

288.78

401.61

659.11

0.0096

0.04

52.26

0.019

0.007

0.004

14.546

19.154

28.033

14.354

18.802

27.243

49.437

14.242

18.598

26.793

47.897

Hbl

Rut

F=5

Cpx

F=20

F=15

F=10

F=20

F=15

F=10

F=20

F=15

F=10

JdF NMORB JdF

*value from Sun and McDonough (1989); composition et Chadwick estimatedKarsten (1989); al. and etet Cousens from (2005). al.al.McDonough Sun and (1990), (1995) from *value

Table D5: Estimated MORB Fuca Table de starting(JdF) compositionJuan of

Mineral acronymsMineral as in Fig. 18

Table D4: Modelled compositions of slab partial melts at various melt fractions (F) for a residue of 70% compositionsCpx, of D4: Modelled a 25% slabresidue Gt,meltmelts partial of for fractions (F) Table 4.5 %at various Hbl, 0.5% Ru

Table D3: Modelled compositions of slab partial melts at various melt fractions (F) for a residue of 70% compositionsCpx, of D3: Modelled a 20% slabresidue Gt,meltmelts partial of for fractions (F) Table 9.5% at Hbl, various 0.5% Ru

Table D2: Modelled compositions of slab partial melts at various melt fractions (F) for a residue of 70% compositionsCpx, of D2: Modelled a 10% slabresidue Gt,meltmelts partial of for fractions (F) Table 20% Hbl at various Kd’s derive from Rollinson (1993), Foley etGERM Barth et and (1993), al. Rollinson al. (2000), (2002), from Kd’s derive Table D1: Partition coefficients for elements D1: Partition used in slabcoefficients for meltpartial Table model 144

Figure D1: Failed slab partial melt models of the GVC adakites. Mineral acronyms as in Fig. 18. Grey shaded area represents the range of the GVC adakites, the black line shows the best fit model to the GVC rocks. Normalising values for N-MORB are from Sun and McDonough (1989). 145

Figure D1 cont’d: Failed slab partial melt models of the GVC adakites. Mineral acronyms as in Fig. 18. Grey shaded area represents the range of the GVC adakites, the black line shows the best fit model to the GVC rocks. Normalising values for N-MORB are from Sun and McDonough (1989). 146