Paleomagnetism, Geochemistry, and U-Pb Geochronology of Proterozoic Mafic Intrusions

in the High Arctic: Relevance to the Nares Strait Problem

by

Steven Walter Denyszyn

A thesis submitted in conformity with the requirements for the degree of

Doctor of Philosophy

Graduate Department of Geology, University of Toronto

© Steven Walter Denyszyn 2008 Library and Bibliotheque et 1*1 Archives Canada Archives Canada Published Heritage Direction du Branch Patrimoine de I'edition

395 Wellington Street 395, rue Wellington Ottawa ON K1A0N4 Ottawa ON K1A0N4 Canada Canada

Your file Votre reference ISBN: 978-0-494-39837-1 Our file Notre reference ISBN: 978-0-494-39837-1

NOTICE: AVIS: The author has granted a non­ L'auteur a accorde une licence non exclusive exclusive license allowing Library permettant a la Bibliotheque et Archives and Archives Canada to reproduce, Canada de reproduire, publier, archiver, publish, archive, preserve, conserve, sauvegarder, conserver, transmettre au public communicate to the public by par telecommunication ou par Plntemet, prefer, telecommunication or on the Internet, distribuer et vendre des theses partout dans loan, distribute and sell theses le monde, a des fins commerciales ou autres, worldwide, for commercial or non­ sur support microforme, papier, electronique commercial purposes, in microform, et/ou autres formats. paper, electronic and/or any other formats.

The author retains copyright L'auteur conserve la propriete du droit d'auteur ownership and moral rights in et des droits moraux qui protege cette these. this thesis. Neither the thesis Ni la these ni des extraits substantiels de nor substantial extracts from it celle-ci ne doivent etre imprimes ou autrement may be printed or otherwise reproduits sans son autorisation. reproduced without the author's permission.

In compliance with the Canadian Conformement a la loi canadienne Privacy Act some supporting sur la protection de la vie privee, forms may have been removed quelques formulaires secondaires from this thesis. ont ete enleves de cette these.

While these forms may be included Bien que ces formulaires in the document page count, aient inclus dans la pagination, their removal does not represent il n'y aura aucun contenu manquant. any loss of content from the thesis. Canada Paleomagnetism, Geochemistry, and U-Pb Geochronology of Proterozoic Mafic

Intrusions in the High Arctic: Relevance to the Nares Strait Problem

Steven Walter Denyszyn

To fulfill the requirements for the degree of Doctor of Philosophy, 2008

Department of Geology, University of Toronto

Abstract

The paleomagnetism, geochemistry, and geochronology of mafic intrusions in Arctic

Canada and Greenland were investigated with the primary purpose of resolving the Nares

Strait Problem, a controversy regarding the location of a plate boundary between Greenland and North America and the relative displacement between the two plates. E/W-trending dykes in Arctic Canada and northwest Greenland have an age of 721 ±2 Ma and are associated with the Franklin magmatic event. Their geochemistry is comparable and the mean paleopole for the Canadian dykes (5.8°N, 188°E, N=12, A95=9.9°) is broadly similar to that of the Greenlandic dykes (8.8°N, 178.7°E, N=10, A95=7.2°) indicating that they are of the same swarm, but that of the Canadian dykes is offset from that of the Greenland dykes by a direction and magnitude consistent with a -200 km displacement along a fault beneath

Nares Strait, in accordance with other lines of evidence such as dyke distribution and age boundaries in the bedrock The paleopole from the Canadian dykes is significantly different

(p=0.05) from that of Franklin rocks elsewhere, suggesting rapid plate motion over the duration of magmatism. Also associated with Franklin magmatism are the N/S-trending

Clarence Head dyke swarm, dated at 715±1 Ma, and the Thule sills, dated for the first time,

ii at 712±2 Ma. Two Clarence Head dykes have been chemically remagnetized, likely as a result of fluids expelled by the Ellesmerian Orogeny.

Three dykes of the Melville Bugt dyke swarm were sampled, one dated at 1622±4

Ma. The measured pole (6.1°N, 267.7°E, dp=2.36°, dm=3.76°) indicates that Laurentia and

Baltica drifted separately at this time. Though the Melville Bugt swarm's extension was not found in Canada, a possible candidate, a dyke at Dundas Harbour, was dated at 1337±2 Ma, a previously unknown age of dyke emplacement in North America, with a VGP of 12.1°S,

250.8°E, dp=6.7°, dm=13.4°.

This study includes the first analysis of the effects of alpha recoil in baddeleyite, which could have a strong influence on analyses of very small crystals. The effect is apparently not as significant in baddeleyite as in zircon, as the higher density of baddeleyite's crystal lattice may restrict recoil distances

iii Acknowledgments

My sincere gratitude is extended to my supervising committee of Henry Halls, Don

Davis, and Russ Pysklywec, whose support, advice, and availability made a direct and significant contribution to the development and ultimate success of this research. Henry in particular is thanked for initiating this project as part of the Nares Strait Geocruise in 2001, and for giving me the opportunity to pursue this study.

Thanks also to my family, and to friends at the Geology department at the University of Toronto where I have been for less than ten years for moral (and technical) support.

Especially noteworthy in this regard are Sarah Hirschorn, Halan Wang, Chris Charles, Chris

White, AH Marin, and the many and various noodle-merchants of Toronto's Chinatown.

I would also like to acknowledge the staff of the Jack Satterly Geochronology Lab at

U. of T., particularly Mike Hamilton and Kim Kwok, for their generosity with their time in teaching me the ropes of geochronology, for assisting in the lab, and for providing the sort of camaraderie that made it a joy to fiddle with microscopic grains of baddeleyite.

Funding for this project was provided by the Natural Sciences and Engineering

Research Council and the Northern Scientific Training Programme. The Canadian Polar

Continental Shelf Project provided efficient and valuable logistical support for fieldwork in the High Arctic, without which, it seems, no work would get done in Canada's north. Ray

Mercredi of Grise Fiord, Ellesmere Island and Hans Jensen of Qaanaaq, Greenland are thanked for sharing their knowledge of how things get done up there.

Finally, thanks to Monica. Your patience and support and more patience have meant so much. This thesis is dedicated to you.

iv Table of Contents

Abstract ii Acknowledgments iv Table of Contents v

1. Introduction 1 The Nares Strait Problem 1 Plate kinematic history of the Arctic region 6 Geology and tectonic history of the Arctic region 14 The Eurekan Orogeny 24 Proposed plate boundaries 26 Mafic intrusions 29 The Melville Bugt dyke swarm 30 Proterozoic diabase dykes of unknown affinity 31 The Franklin igneous event 33

2. Methodology 38 Paleo-/Rock Magnetism 38 Petrography and Geochemistry 45 Tectonic reconstruction 47 Geochronology 48

3. Results 55 Geochronology 56 Franklin dyke swarm 57 Clarence Head Swarm (N/S dykes) 58 Melville Bugt swarm 63 Dundas Harbour dyke 63 Geochemistry and Petrography 67 Franklin dyke swarm 67 Clarence Head Swarm (N/S dykes) 77 Melville Bugt swarm 77 Dundas Harbour dyke 77 Paleo-/Rock Magnetism 80 Franklin dyke swarm 80 Clarence Head Swarm (N/S dykes) 97 Melville Bugt swarm 102 Dundas Harbour dyke 106

v 4. Discussion 110 Nares Strait Problem 110 Correlation of Neoproterozoic mafic dykes Across Nares Strait 113 Does the Kap Leiper dyke extend into Canada? 118 Reconstruction of the Nares Strait region 128 Geochronology of baddeleyite in the Franklin intrusions 133 Emplacement of orthogonal dykes and the remagnetization of the "Clarence Head" dykes 137 New Ages of Proterozoic Arctic Magmatism and its effect on the Apparent Polar Wander Path of Laurentia 142 Differences in paleopole measurements across the Franklin intrusions 148

5. Conclusions 160

6. References 164

Appendix A; Paleomagnetic data by specimen 180

Appendix B: Geochemical data 193

VI 1

1. Introduction

The Nares Strait Problem

Controversy about the existence of a major sinistral fault lying between Greenland and

Ellesmere Island (the "Nares Strait problem" of Dawes and Kerr 1982a) originated with early reconstructions of the two landmasses by Taylor in 1910 and Wegener in 1915 (Figure 1.1).

Taylor (1910) proposed a sinistral fault with an offset of about 300 km to explain the generation of Tertiary mountain chains by land drifting away from polar regions. Later, Wegener (e.g.,

1929) incorporated these ideas into his theory of continental drift and the fault thereafter became known as the "Wegener Fault". The problem arises because, subsequently, several workers have compared the stratigraphy of the Proterozoic and Paleozoic sedimentary rocks on either side of the Nares Strait (Figure 1.2) and concluded that no more than 70 km of lateral movement is permissible (Dawes & Kerr 1982b and references therein; Harrison 2006). However, this conclusion has been questioned and alternative solutions more in harmony with plate tectonic expectations have been presented (e.g., Miall 1983; Johnson & Srivastava 1982). In support of the plate tectonic model, geological and geophysical studies (particularly of marine magnetic anomalies) indicate that oceanic crust underlies the Labrador Sea and that the opening of the

Labrador Sea and Baffin Bay occurred between ~85 and 56 My ago (Srivastava et al 1981;

Oakey 1994; Chalmers & Laursen 1995), with Greenland moving in a NNE direction with respect to Canada. The Nares Strait Problem remains a significant controversy in plate tectonics in which all the "classic" elements indicating plate motion are present, but the onshore geology has been interpreted as telling a very different story.

The primary objectives of this thesis involve studying a combination of distinctive characteristics - paleomagnetism, petrography, geochemistry and U-Pb geochronology - to Figure 1.1: Speculative reconstructions of Greenland and North America by a) Taylor (1910) and b) Wegener (1915). In a), numbers represent distances in kilometres. Modified after De Paor et al. (1989). 3 determine if the E/W-trending diabase dyke swarm of the Thule region of Greenland can be correlated as an offset continuation of a set of dykes with similar trend in southeast Ellesmere

Island and eastern Devon Island (see Figure 1.2). The dykes in Greenland do not appear to continue across the Nares Strait into northern or central Ellesmere Island, and those found on

Ellesmere and Devon Islands do not seem to continue into western Greenland (Frisch 1984a,b,c;

Dawes 1991; Dawes & Garde 2004). If proven, this correlation would indicate a sinistral offset of approximately 200 km, the lateral distance required by plate tectonic reconstructions, and provide direct evidence that will help resolve what has been a fundamental challenge to a plate tectonic model for the region for decades. Scenarios have been proposed by numerous authors

(as described below), designed to find other ways to allow movement between the plates without involving Nares Strait. However, if a correlation of Neoproterozoic diabase dykes on either side of Nares Strait is demonstrated, it will also indicate that a significant amount of displacement along a fault must lie along the strait itself, or at least along the southern part of the strait.

Ideally, the identification of an individual dyke or group of dykes on both sides of Nares Strait by a distinctive geochemical or paleomagnetic signature would provide a precise marker unit for the accurate measurement of any displacement between North America and Greenland.

Estimates of the magnitude of offset originally based on matching shorelines (e.g. Taylor 1910) are ca. 330 km, a value that is too high considering the actual extent of the continental shelves of the two continents. In central Baffin Bay, seismic refraction studies (Reid & Jackson 1997) and gravity profiles (Geoffroy et al. 2001) indicate that the actual width of oceanic crust (or thinned, serpentinized continental crust) is approximately 260 km, which is a maximum limit on the degree of motion between Greenland and North America. If half of this spreading is accommodated by displacements to the west within the Canadian Arctic Archipelago or dextral 4 motion on the opposite side of the Baffin spreading ridge (Harrison 2006), then there could be as little as 150 km of sinistral displacement along the Wegener fault. When the areas of peak density of the two sets of E/W-trending dykes on either side of Nares Strait (central Devon Island in Canada, southern Inglefield Bay in Greenland) (Figure 1.2) are brought into alignment, then an offset of ca. 220 km is indicated, though there may be unseen dykes at the margins of the swarm (e.g., under Lancaster Sound) and therefore only the tracing of an individual dyke across

Nares Strait will provide a precise measure of the degree of offset.

This research was stimulated by the Nares Strait Geocruise of 2001, an expedition co­ ordinated by the Canadian and German geological surveys with a principal objective being an examination, primarily using seismic and aeromagnetic methods, of the evidence for and against the existence of the Wegener fault. As part of this expedition, a small collection of samples from dykes and sills were taken from the area, and the good quality of the paleomagnetic results from that collection raised the possibility of a larger-scale study that could test the possibility of a

-200 km offset directly, with the additional application of geochemical and geochronological methods.

Precisely-dated rocks with well-defined paleomagnetic poles are scarce in the Proterozoic record, yet are critical for reconstructing the past positions of continents. Diabase dyke swarms are particularly useful for this purpose (e.g., Buchan & Halls 1990); their shape and orientation make them physical piercing points; their mafic mineralogy makes them a reliable retainer of magnetic remanence; individual dykes can be traced over long distances by their geochemistry, which can remain uniform over hundreds of kilometres (e.g., Kalsbeek & Taylor 1986); and 80°W

... = Average strike = Carboniferous - Paleogene (Sverdrup Basin) of gneissosity = Lower Paleozoic - Devonian (Franklinian Basin) / = Diabase Dyke = Mesoproterozoic (Nauyat volcanics) a7io«ta= u-Pbage of = Mesoproterozoic (Thule/Borden Basins) basement units = Paleoproterozoic (Shield) E3= Archean (Shield) H= Archean (Pear/a)

Figure 1.2: Simplified geological map of the Nares Strait region of Arctic Canada and Greenland. After Frisch (1984a,b,c), Okulitch (1991), Dawes (1991). U-Pb (zircon) ages are from Frisch & Hunt (1988); Jackson (2000); Nutman et al. (in press). entire giant dyke swarms can be emplaced over only a few million years (e.g., the Mackenzie swarm, LeCheminant & Heaman 1989), which makes them virtual snapshots in the rock record.

In the Nares Strait region, only the dykes of the Thule swarm of Greenland and the E/W- trending dykes of Devon Island and Ellesmere Island, hereafter referred to as the Devon Island dyke swarm (presumed to be of the greater Franklin dyke swarm), are likely candidates to be found to cross the Strait, and the fact that their trend cuts Nares Strait at a high angle makes them useful marker units. The Melville Bugt swarm has also been conjectured to cross from

Greenland into Canada based on the trend of two dykes on Ellesmere Island (Nielsen 1990), but this relationship has yet to be established, and the fact that these dykes intersect Nares Strait at a lower angle makes them less desirable as piercing points. However, if the Melville Bugt dykes are indeed found on the Canadian side, their correlation would be an important corroboration of the results obtained from the Thule and Devon Island dykes. Dykes on Devon Island and

Ellesmere Island with a northwesterly to northerly trend (Frisch 1984a,b,c) may represent this continuation of the Melville Bugt swarm, and this possibility will be investigated.

Plate kinematic history of the Arctic region

The first stage of motion between the Greenland and North America plates began at approximately 95 Ma, with the rifting that would lead to the formation of the Labrador Sea

(Roest & Srivastava 1989, Geoffroy et al. 2001). This rifting was accompanied by northeastward motion of Greenland relative to North America, with sinistral strike-slip motion along Nares Strait (Figure 1.3). Labrador Sea rifting and subsequent seafioor spreading resulted from extension of the Atlantic Ocean spreading ridge system, and marine magnetic anomalies in 7

the oceanic crust under Labrador Sea indicate spreading was active between the Late Cretaceous

and the Oligocene, with contemporaneous spreading in the North Atlantic opening the

Norwegian and Greenland Seas with a ridge-ridge-ridge triple junction to the south of Greenland

(Talwani & Eldholm 1977, Srivastava et al. 1981). Roest & Srivastava (1989) proposed that the

first development of oceanic crust (shaded areas on Figure 1.3) occurred during magnetochron

33 (79.7-74.5 Ma), ending at chron 13 (33 Ma) (Figure 1.4). Chalmers & Pulvertaft (2001)

revised this interpretation of marine magnetic anomalies to conclude that the first seafloor

spreading occurred at chron 27 (61-3-60.9 Ma), and that the stretching of the continental crust is

greater in extent and took longer than previously thought. Sections of the crust underlying the

Labrador Sea and Baffin Bay, previously thought to be composed of oceanic crust, are currently

interpreted as stretched continental lithosphere, composed of "transitional crust" (serpentinized peridotite and heavily altered continental crust) based on aeromagnetic and seismic surveys

(Chalmers & Pulvertaft 2001; Geoffroy et al. 2001; Skaarup et al. 2006). Davis Strait, the

narrow waterway connecting Labrador Sea and Baffin Bay (DS, LS, and BB in Figure 1.5 respectively), represents a bend in the Labrador Sea/Baffin Bay spreading axis that may be

accommodated by an extremely oblique spreading axis with a trend of 010°, or a series of small transform faults (the "Ungava Fracture Zone") (Chalmers & Laursen 1995; Geoffroy et al.

2001 ;Skaarup et al. 2006). Baffin Bay itself has been shown to overlie Tertiary oceanic crust

(Jackson et al 1979; Jackson et al. 1992), which terminates to the west in a sharp paleoshelf that

has been interpreted as a transition between a continent-ocean boundary and a continent-

continent boundary just to the east of the coastlines of Devon Island and southeast Ellesmere

Island (Jackson et al. 1992, Reid & Jackson 1997). The spreading that occurred in the p *Ti a *~k. a- <§ OJ I-! 100 Ma o o L>

> * »-+> r*-' o o oH cc t/3 2 13 o w O Co3* cr W 3 O 8P 3 5" O >-t o' Ct> tr O 3 E^ ^—*,I—* , s0 * ^O Q*t 0^o0 j^

^•—^ •^ 30 Ma o 50 Ma a o D* o > TJ 3 N >-t II po * oa w o ft ft 3 >fl P P «-• £ ^" o N o o CO Js" C\ u< ISlB'j= Active seafloor spreading / = Compression = Relative motion of JsB AGreenlan d plate U»1 ^z = Strike-slip motion ';= Eurekan and Spitzbergen 500 km o Fold Belts •It: = Oceanic Crust sp Age Magnetic Geologic (Ma) Polarity Chron Age Jligocen e ell 30 cl2 c13 35 c16 Cl7 clfci 40 m

otee n c19 srt i « c20 45 ^^^ Q ^^^^^^^^^^^^^^H Q^, | (V 3

c24 55 Paleocen e c25 60 c26 c27 c28 65 c29 c30 70 c31 w—mammezi 75 m^m a

oeta c ^ 90 ^^H QC ^^^^|c34N 0 c(A

Figure 1.4: Magnetic polarity time scale. After Ogg (1995). 10

Labrador Sea and Baffin Bay between the Cretaceous and the Paleocene therefore indicates a net northward movement of Greenland of approximately 250 km, though magnetic anomalies and fracture patterns on the seafloor suggest that this motion occurred in, broadly, two phases.

Between the beginning of rifting at 95 Ma to approximately 55 Ma (chron 24r), Greenland moved in a NNE-ward direction relative to North America with a sinistral offset of

approximately 250 km (Oakey & Damaske 2006, Chalmers & Pulvertaft 2001, Roest &

Srivastava 1989), with an Euler pole of rotation indicated by small circles described by transform faults in both Baffin Bay and the Labrador Sea of 73°N, 79°W (Geoffroy et al. 2001). After 55

Ma, in response to seafloor spreading in the Atlantic Ocean, the direction of motion of Greenland shifted to NW, driving the Greenland plate into Ellesmere Island, with approximately 250 km of oblique (transpressive) convergence (Oakey & Damaske 2006), with an Euler pole at 50°N,

125°W (Geoffroy et al. 2001). These values have been refined to account for the more complex polyphase movement of Greenland based on fracture zone geometry. Oakey and Chalmers (in

Oakey 2006) and Roest and Srivastava (1989) calculated the Euler poles at different stages of motion of Greenland:

Chron (age) Lat (°N) Long(°E) Angle

21N (47.9-46.2 Ma) 53.2° -112.9° -1.65°

24N (53.3-52.3 Ma) 53.2° -112.9° -3.75

25N (56.4-55.9 Ma) 24.5° -137.3° -3.12°

26N (57.9-57.5 Ma) 26.5° -143.7° -3.40°

27N (61.3-60.9 Ma) 27.9° -149.3° -3.72°

Table 1.1: Euler poles for the relative motion of Greenland relative to Canada during the opening of the Labrador Sea. From Oakey (2006) and Roest & Srivastava (1989). 11

These rotation poles are linked by "stage poles" (Oakey 2006) in order to define a continuous motion of Greenland. The West Spitsbergen foldbelt of northern Greenland also indicates a dramatic change in direction of Greenland relative to surrounding plates at chron 24 (ca. 52 Ma), the time of North Atlantic spreading, from a N/S-directed compressional phase while still attached to Eurasia (i.e., pre-North Atlantic spreading) to strike-slip dominated transpression after chron 24 (Saalmann & Thiedig 2002; Piepjohn & Van Gosen 2001).

A current model (Srivastava et al. 1982, Oakey 1994, Piepjohn et al. 1998, Tessensohn &

Piepjohn 1998) that perhaps best explains schematically the structural and tectonic regimes in the area indicates three distinct phases of movement (Figure 1.3):

1) Between 85 Ma and 56 Ma, seafloor spreading in the Labrador Sea/Baffin Bay system caused northeasterly movement of Greenland and Eurasia approximately 175 km relative to

Canada, with strike-slip motion along Nares Strait, causing basement arch uplift on Ellesmere

Island;

2) Between 56 and 35 Ma, seafloor spreading was contemporaneous in the Labrador Sea and the northern Atlantic Ocean, the two spreading systems being linked by a ridge-ridge-ridge triple junction south of Greenland, causing northward escape movement of the Greenland plate, and the formation of the Eurekan and West Spitsbergen foldbelts. During this period, there was as much as 175 km of transpressive convergence (at an angle of 60° to 67° counterclockwise relative to the direction of initial rifting) with Ellesmere Island, the timing of which is consistent with the age of compressive deformation on Ellesmere.

3) By 35 Ma, seafloor spreading west of Greenland had ceased, ending the northward drift of Greenland, though transtensional opening of the Fram Strait between Greenland and

Svalbard was still occurring. 12

To the north, in the present-day Arctic Ocean region (Figure 1.5), the Canada oceanic basin (CB

in Figure 1.5) opened as Alaska rifted off the west coast of the Arctic archipelago and rotated to

its current position between ca. 155-135 Ma (Grantz et al. 1998). Prior to the time of the

Labrador Sea rifting, however, the Eurasia basin (EB) had yet to open and the Greenland plate's

northward movement caused its collision with not only Ellesmere Island (EI) and Svalbard (SV), but with the nascent Eurasia basin as well. This caused crustal shortening directly north of

Greenland and possible subduction of the Gakkel Ridge (GR) under Morris Jesup Rise (MJ) and

the Yermak Plateau (YP). This has been indicated by a prominent linear anomaly orthogonal to

the Gakkel Ridge's spreading centre revealed by both gravity and aeromagnetic surveys

(Brozena et al. 2003). A result of this opening is that prior to anomaly 25 time (ca. 56 Ma), the

Lomonosov Ridge (LR) was part of the Eurasian plate; subsequently, it moved with the North

American plate as it rifted from the Barents margin (Brozena et al. 2003). This implies

simultaneous relative motion of Greenland to both Ellesmere Island and the Lomonosov Ridge at that time, with transform motion continuing from the Nares Strait to the Gakkel Ridge

(Srivastava 1985). The relationship between the Lomonosov Ridge and the North American plate is supported by the ca. 150 km of oceanic crust between the Lomonosov Ridge and the

Eurasian plate created before anomaly 24 time, compared with 90 km in the Norwegian Sea.

Given the Eurasian basin's proximity to the North Atlantic/Norwegian Sea system's pole of

rotation, it follows that less seafloor would be produced in the Eurasian basin than the

Norwegian Sea, not more. Therefore, the Eurasia basin opening predates the North Atlantic and has been linked to the earlier spreading in the Labrador Sea/Baffin Bay system instead (Brozena

et al. 2003). Greenland acted as a triangular indenter as it moved northward into the east/west-

opening Eurasian basin (Torsvik et al. 2001; Brozena et al. 2003) (Figure 1.3). 13

Figure 1.5: Physiographic map of the Arctic region (modified after Jakobsson et al. 2000). AR = Alpha Ridge; BB = Baffin Bay; BI = Baffin Island; CB = Canada Basin; CI = Coburg Island; DI = Devon Island; DS = Davis Strait; EB = Eurasia Basin; EI = Ellesmere Island; GL = Greenland; GR = Gakkel Ridge; JS = Jones Sound; KB = Kane Basin; KC = Kennedy Channel; LA = Lancaster Sound; LR = Lomonosov Ridge; LS = Labrador Sea; MB = Melville Bugt; MJ = Morris Jesup Rise; NS = Norwegian Sea; SS = Smith Sound; SV = Svalbard; YP = Yermak Plateau. 14

Based on marine magnetic anomalies and the application of a fixed hotspot reference frame to paleomagnetic data, all three continents involved moved together in the late Cretaceous to the NNE, followed by a shift to the NW at approximately 80 Ma (Torsvik et al. 2001). This corresponds to the time of seafloor spreading in the Labrador Sea (Roest & Srivastava 1989), or at least of extensive rifting between Greenland and North America (Chalmers & Pulvertaft

2001). Between 55 and 45 Ma, however, the Eurasia plate shows a marked shift to a NE-ward drift direction, while North America and Greenland continue moving to the NW, corresponding to initial breakup of the North Atlantic (Torsvik et al. 2001). This is supported by the occurrence of basin infill and high sedimentation rates in the Atlantic at that time (Mosar et al. 2002).

Various reconstructions have been proposed of the kinematic history of the

Greenland/North America plates between 100 and 50 Ma, and specifically, of where the plate boundary between them lies. To properly evaluate these models, the geology of the region must first be examined. The correlation of geological units that pre-date the proposed drift of

Greenland is an essential part of any correlation between the Greenland and North American plates.

Geology and tectonic history of the Arctic Region

Tectonostratigraphic correlations of the Archean through Mesoproterozoic terranes that make up the bedrock of Greenland, Arctic Canada and northeastern Canada (e.g., St-Onge et al.

2005 and references therein) indicate that they were one continent at that time (Figures 1.2, 1.6).

The Rae craton, which extends from central Baffin Island into the Melville Bugt region of

Greenland, was formed by the late Archean, as indicated by ages obtained from the granitic to tonalitic gneisses of the Thule mixed gneiss complex (2.9 Ga Sm-Nd TDM model age), the Kap 15

60°W

Meosproterozoic Tertiary basalts ^- mafic dyke swarms Lower Paleozoic - Devonian (Franklinian Basin) (Gardar, Nain, Mesoproterozoic (Thule/Borden Basins) Harp, Nutak) Paleoproterozoic (supracrustal units) Paleoproterozoic (Cumberland/Proven batholith) Archean (Superior/Rae/North Atlantic cratons) Ketilidian/Makkovik Belt ca. 1.85 Ga Paleoproterozoic suture (Nagssugtoqidian/Torngat)

Figure 1.6: Simplified geological map of the Baffin Bay/Labrador Sea region. After Frisch (1984a), Dawes (1991), Dawes & Garde (2004), St-Onge et al. (2005), Buchan & Ernst (2004).. Paleoproterozoic suture from Connelly et al. (2006) and St-Onge et al. (2005). 16

York meta-igneous complex (ca. 2.7 Ga Rb-Sr) of northwest Greenland (Dawes 1991, 2006),

and the quartzofeldspathic to pelitic gneiss, schist, amphibolite and quartzite of the Lauge Koch

Kyst complex near Melville Bay, dated to 2.7-2.6 Ga by Rb-Sr and U-Pb methods (Dawes 1991,

2006; Dawes et al. 1988) (Figure 1.2). These are correlated with felsic and mafic metavolcanics,

amphibolites and granitic to intermediate intrusions of the Mary River Group and associated

rocks on northern Baffin Island, dated at ca. 2.9-2.7 Ga (U-Pb zircon, Jackson 2000) (Figure

1.6).

To the north, the metasedimentary rocks of the "Ellesmere-Devon terrane", that portion

of the shield rocks underlying Devon Island and southeast Ellesmere (Frisch 1988), comprise interpreted continental margin sequence of carbonates, sandstones and shales (with minor volcanics) that have undergone granulite-facies metamorphism to quartzofeldspathic to pelitic gneiss and quartzite with marble and migmatite (Frisch 1988). These have since been overlain by Mesoproterozoic Thule basin sandstones and lower Paleozoic Franklinian basin carbonates, and intruded by what are now meta-plutonic rocks ranging in composition from quartz norite to tonalite to peraluminous granite (Frisch 1988). The meta-plutonic rocks on Ellesmere have been dated by the U-Pb (zircon) method at 1960±5 and 1912±2 Ma, with additional ages of ca. 1930

Ma interpreted as representing the high-temperature metamorphism. Gneisses from the southern coast of Devon Island have a Neoarchean age, dated broadly at 2518 +56/-33 Ma and 2428 +36/-

31 Ma (U-Pb zircon; Frisch 1988; Frisch & Hunt 1988). The units on Ellesmere can be

correlated with the granulite-grade metamorphic terrane of the Inglefield Mobile Belt of

Greenland (Frisch & Dawes 1982; Dawes 1991). The rocks of this belt have been subdivided

(Dawes 2004, 2006) into the Etah meta-sedimentary rocks and gneisses (the Etah Group, Figure

1.2), which is intruded by the Etah meta-igneous complex comprising felsic gneisses, 17

monzogranite, diorite, syenite and less commonly gabbro. Dawes (2004) reports U-Pb detrital

zircon ages of the Etah Group of ca. 1980-1950 Ma, and the intrusion of the meta-igneous

complex has been dated at ca. 1950-1915 Ma, with later granulites being produced by partial

melting at ca. 1785-1740 Ma. Recent work (Nutman et al., in press), has corroborated these age

relationships, with U-Pb SHRIMP ages on zircon in the Etah Group of 1950-1900 Ma.

Paleoproterozoic units in central Baffin Island have also been matched with similar-aged

rocks in western Greenland. The correlative Piling and Hoare Bay Groups (Figure 1.6) cross the

width of Baffin Island, and have been dated by detrital zircon studies. Lowermost strata of

continental-margin clastic and carbonate platform rocks have a maximum age of 2160 Ma;

overlying rift-related sedimentary units and mafic igneous activity have been dated at 1980 Ma;

and above that, schist, shale and foredeep turbidites have a maximum age of 1915 Ma (St-Onge

et al. 2005 and references therein). These are intruded by granitic rocks with ages ca. 1.90-1.85

Ga (Jackson et al. 1990). On the west coast of Greenland, the Nagssugtoqidian and Rinkian belts, now considered to represent the southern and northern parts respectively of a single

orogenic complex (Connelly et al. 2006), contain Paleoproterozoic units that can be linked with

those across Baffin Bay. Gneisses of tonalitic to granodioritic composition have been dated to

ca. 2.71-2.57 Ga, and therefore may be correlated with the Rae craton (Kalsbeek 1986). This basement is overlain by the Karrat Group, comprised of shallow marine metasedimentary rocks

(quartzite, pelitic gneiss and schist) and marble, which is in turn overlain by black shale and a

thick flysch turbidite deposit (Grocott & Pulvertaft 1990). On the basis of detrital zircon

geochronology, this turbidite is constrained to have been deposited after 1949±11 Ma; principal

detrital zircon ages peak between 2100 and 1950 Ma (Kalsbeek et al. 1998). Based on

similarities between sedimentary and volcanic units, stratigraphic sequences and thicknesses, and 18 detrital zircon ages, St-Onge et al. (2005) correlate the Karrat Group and related rocks in the

Disko Bugt area of Greenland with the Piling and Hoare Bay Groups of Baffin Island (Figure

1.6), and interpret their tectonic setting as a crustal suture of one or more smaller plates, again indicating that Greenland and North America were together at that time.

The basement rocks on both sides of Nares Strait and Baffin Bay are unconformably overlain by

Mesoproterozoic sequences of siliciclastic rocks with some carbonates and basaltic intrusive and extrusive units. On north Baffin Island, these deposits comprise the Borden Basin, the Bylot

Basin, and, to the west, the Fury and Hecla Basin (Jackson & Ianelli 1981; Jackson 2000)

(Figure 1.2., Figure 1.6). On the east coast of Ellesmere Island and northwest Greenland, they form the Thule Basin (Frisch 1988; Dawes 1991, 1997, 2006; Dawes et al. 1982), made up of the sediments of the Bylot and Thule Supergroups, respectively. These flat-lying packages of peritidal- to subtidal-shelf to slope deposits (Jackson & Ianelli 1981; Narbonne & James 1996;

Dawes 1997) are not well-constrained in terms of age. The basal basaltic extrusive rocks of the

Bylot Supergroup, the Nauyat volcanics, are found in the southwest of the Borden basin and have been genetically linked based on their paleomagnetism to the Mackenzie intrusions (Fahrig et al. 1981). The Mackenzie intrusions have been dated elsewhere by the U-Pb method on baddeleyite at 1267±2 Ma (LeCheminant & Heaman 1989), thus providing an approximate age for initial sedimentation in the Borden Basin. The entire sedimentary package is cut by intrusions of the Franklin magmatic event (Fahrig et al. 1971; Christie & Fahrig 1983; Pehrsson

& Buchan 1999), dated at -720 Ma (Heaman et al. 1992; Pehrsson & Buchan 1999). These relationships loosely bracket the age of Borden sedimentation and may indicate continued deposition into the Neoproterozoic, though the paleomagnetic poles obtained from the basal

Nauyat Formation and the arenaceous sandstones of the upper Strathcona Sound Formation (both 19 as identified in the Borden Basin) are similar to each other as well as to that of the Logan Sills, dated at 1109+4/-2 (U-Pb baddeleyite, Davis & Sutcliffe 1985; Fahrig et al. 1981), suggesting a more restricted Mesoproterozoic age (Kah et al. 1999). In northwest Greenland, the age of the

Thule Supergroup is likewise constrained only by the sills and flows within its lowermost strata

(related to the Mackenzie event, dated by U-Pb on zircon and baddeleyite to 1268 ± 2 Ma

(LeCheminant & Heaman 1991)) and those that cut its sediments (the Franklin dykes), and studies of acritarch biostratigraphy have only been able to date the units to "the Mesoproterozoic and/or Neoproterozoic" (Dawes 2006). The Thule and Bylot Supergroups have been linked stratigraphically (Jackson & Ianelli 1981; Hofmann & Jackson 1996) and in terms of their 813C profiles (Kah et al 1999). Both of the Borden and Thule basins have been attributed to locally

NNE/SSW-oriented rifting, based on horst and graben structures preserved in the basement

(Jackson & Ianelli 1981; Dawes 1997). Jackson & Ianelli (1981) conclude that the Thule and

Borden basins are part of a regionally-extensive paleo-continental shelf that extended from as far west as the Cordillera of Yukon Territory more or less continuously through Victoria Island,

Somerset Island, Baffin Island, and northwest Greenland, a result of the opening of an ocean to the northwest (the Poseidon Ocean of Fahrig 1985) at approximately 1250 Ma, akin to the present-day opening of the Atlantic Ocean and the passive margin sequence along the east coast of North America (Figure 1.7). This implies that the present-day pattern of outcrop may be an artifact of preservation, and that the sedimentary rocks of the Thule and Borden basins should not be seen as distinct units only correlative locally, but as remnants of an extensive system of rift-related sedimentation. This rifting has also been linked to the emplacement of the

Mackenzie dykes and flows, which include the Nauyat volcanics (Fahrig 1985), which support the rifting hypothesis. Sediments of the Strathcona Sound Formation of the Borden Basin are 20

found on the south coast of Devon Island, separated from Baffin Island by Lancaster Sound

(Frisch 1984a) (Figure 1.2), which raises the possibility that Mesoproterozoic sediments underlie

Lancaster Sound, and that they may continue further under the flat-lying Paleozoic carbonate platform to the north, extending the northward limit of the Borden Basin and supporting the

reconstruction of Jackson & Ianelli (1981) (Figure 1.7).

The correlation described here of Archean through Mesoproterozoic units has been

interpreted by St-Onge et al. (2005) as evidence for a north-to-south sequence of accretion and

collision events, with the growth of the Rae craton by accretion of smaller terranes in the

Archean and Paleoproterozoic, before the Superior craton of North America collided with the

Rae craton in the Trans-Hudson orogeny. This process of growth of the Laurentia landmass requires the closure of the Baffin Bay/Labrador Sea waterway in order to maintain the continuity

of these units. Relevant to the Nares Strait problem is the fact that in the present-day plate

configuration, the regional strike of the basement units changes sharply across Nares Strait. On

southeast Ellesmere Island, the basement rocks are of Paleoproterozoic age with a dominant north-south structural trend; in the adjacent Thule/Inglefield Bay region of Greenland the trend is

WNW/ESE in the Archean basement, as well as in the Paleoproterozoic rocks to the north. On

Devon Island to the south (but adjacent to the Thule area if Labrador Sea and Baffin Bay are

closed, repositioning Greenland to the south), it is also WNW/ESE and Archean (Miall 1983;

Frisch 1984a,b,c; Dawes 2006) (Figure 1.2).

On Ellesmere Island and northern Greenland, the crystalline units are covered by a thick

succession of clastic and carbonate rocks that make up the Franklinian passive margin sequence

dated biostratigraphically as Cambrian to Silurian in age, though sedimentation may have begun

as early as the Neoproterozoic on what has been interpreted as a platform-to-basin transition J = Craton HI = Shelf Zone (Carbonate Platform) ^M = Rifted Basin /•— = Active (Rift) Faults *^^=s Paleocurrent Direction = = Franklin dyke swarm (schematic)

Figure 1.7: Schematic illustration of proposed ca. 1250-1200 Ma opening of "proto-Arctic ocean", after Jackson & Ianelli (1991), showing the extent and depositional setting of the clastic sequences of the Borden basin, Thule basin, and related Mesoproterozoic deposits. WNW paleocurrent directions are found in the sediments of both the Borden and Thule basins. "Shelf zone" indicates Mesoproterozoic carbonates deposited on the shelf of the proto-Arctic ("Poseidon") ocean, which was closed at ca. 1000 Ma. Red lines denote schematic location of the E/W-trending Thule and Devon Island mafic dyke swarms, dated to 720 Ma, for reference. along the northern margin of North America (Frisch & Trettin 1991; Trettin et al. 1991; Dewing

et al. 2004). This passive margin extends from the Mackenzie Mountains of mainland northwestern Canada, across the Arctic Archipelago, and into northern Greenland (Dawes et al.

2000; Dewing et al. 2004). The ocean adjacent to this margin closed in the latest Devonian to

Early Carboniferous, when the exotic terrane presently making up the northwestern quarter of

Ellesmere Island (known as Pearya) (Figure 1.2) accreted to the North American plate, causing the extensive compressional event called the Ellesmerian Orogeny (Trettin 1991a). In

Greenland, this orogen is known as the North Greenland fold belt (Soper & Higgins 1990).

Though subsequent tectonic activity has wholly or partially overprinted Ellesmerian structures, the deformation style is of thin-skinned thrust sheets that verge to the northwest generally, but locally have a N/S or E/W trend. Pearya was likely transported as multiple terranes, and their

accretion to the deep-water Franklinian basin caused intense deformation and likely involved faulting in the basement rocks as far away as northern Greenland (Soper & Higgins 1990; Trettin

1991a). East-west compression resulting from this collision likely caused the uplift of regions with northerly basement trends in the late Silurian-Early Devonian (Boothia Uplift) and Early

Devonian (Inglefield Uplift) (Trettin 1991a and references therein). Farther away from the site of collision, flat-lying and undeformed Franklinian carbonates (generally mid-Cambrian in age) unconformably overlie the basement rocks (including Proterozoic sedimentary rocks, where present) on south and east Ellesmere Island and on Devon Island (Frisch 1988).

Unconformably overlying the internal part of the Franklinian basin of northwestern

Ellesmere Island are deposits and minor volcanics of the Sverdrup basin, a successor basin that

formed as a result of post-orogenie extensional collapse and rifting in the mid-Carboniferous to

Early Permian. Predominantly clastic sedimentation continued with post-rift subsidence in the 23 late Permian through early Cretaceous (Patchett et al. 2004; Harrison et al. 2006). In the early

Cretaceous, uplift occurred throughout the basin, with renewed rifting and related basaltic volcanism. A final transgression in the mid-Cretaceous due to thermal subsidence caused the deposition of thick marine clastic sediments in the Sverdrup basin followed by bituminous

shales, while widespread uplift in the late Cretaceous terminated this period of sedimentation

(Trettin 1991a; Patchett et al. 2004). This inversion occurred in the Cretaceous during the rifting of the Greenland plate from North America, resulting in a rift zone that may link Labrador Sea,

Baffin Bay, Nares Strait, and the Gakkel Ridge (Weber & Sweeney 1990) (see Figure 1.3).

Associated with this rifting are the sedimentary units of the Cretaceous-Paleocene Baffin

sequence of sandstone, mudstone, conglomerate and coal, with minor basaltic flows and tuff, and the Pal eocene-Eocene Eureka Sound sequence of fluviodeltaic sandstone, mudstone, volcaniclastic sandstone, conglomerate and breccia (Trettin 1991a; Harrison et al. 2006). These units are extensive on northern and central Ellesmere Island, and are correlative over a large geographic area there (Harrison et al. 1999), though they are absent on adjacent areas of

Greenland (Dawes 2004, 2006).

The Quaternary geology of Arctic Canada and Greenland is complex, but in general reflects the extensive glaciation that has affected (and continues to affect) nearly the entire region

(e.g., Harrison et al. 1999; Dawes 2004, 2006). The present-day physiographic features of the

Arctic Islands, the many fiords, channels and other waterways (including Nares Strait), have been attributed to the scouring of Neogene sediments by valley glaciers (Trettin 1991b; Harrison

et al. 1999), though England (1987) describes the pattern of the inter-island channels as being

controlled by tectonic features such as local- to regional-scale faults currently parallel to the

major coastlines and fiords of Devon and Ellesmere Island (Frisch 1984a,c) because the large fiords of Ellesmere Island are too deep and long for fluvial or glacial processes to have caused them, and the channels postdate and crosscut the glacial drainage systems that have been proposed to have caused them (England 1987).

The Eurekan Orogeny

The Eurekan orogeny had a strong influence on the geology of the Nares Strait region, and as its formation has been linked to the collision of the Greenland plate with the North

American plate, its effects must be examined as part of any reconstruction of the area. The

Canadian Arctic region and Greenland were affected in the Eocene by both the North Atlantic and the Baffin Bay/Labrador Sea spreading systems, with two "megashears" developing on either side of Greenland: the Wegener Fault to the west, and the De Geer Fault Zone (DGFZ in

Figure 1.3) between Greenland and Svalbard in the east (Tessensohn & Piepjohn 1998) (Figure

1.3). This involved multiple phases of relative movement of Greenland and Ellesmere Island, ranging from initial transtension through oblique transpression and near-orthogonal convergence, resulting in a series of N/S-trending faults within Ellesmere Island (Miall 1985). To the east, the

West Spitzbergen foldbelt is considered the Eurekan orogeny's eastern extension (Saalmann &

Thiedig 2002; Piepjohn & Van Gosen 2001; Lepvrier 2000; Tessensohn & Piepjohn 1998).

The Eurekan Orogeny ended sedimentation in the Sverdrup basin, reactivating older compressive structures from the Ellesmerian Orogeny while extensive N/S-oriented normal faulting occurred throughout southern Ellesmere Island and nearby regions (Okulitch et al.

1990). The structural style of the orogeny is dominated by thrust-faulting, with thin thrust sheets along flat-lying detachments and up-sequence ramps, some with klippen exposed at the front of the sheet (Osadetz et al. 1982; Piepjohn et al. 1998). SE-vergent fold structures on the scale of 25 hundreds of kilometres also occur on Ellesmere Island (Piepjohn et al. 1998). The Eurekan

Orogeny is unusually broad for a fold-and-thrust belt, and while re-activation of Ellesmerian faults during the Eurekan makes it difficult to assign a given fault in the area to a specific event, up to 100 km of contraction is permissible within the Eurekan system (De Paor et al. 1989), with some estimates up to 170 km (e.g., Miall 1985). The West Spitzbergen foldbelt is also notably broad, with a width of about 150 km (Saalmann & Thiedig 2002), and crustal shortening across that belt has been estimated at 80 km (Manby & Lyberis 1996).

A significant gravity low indicates tectonic thickening by detachment and stacking of structural "slices" near the zone of maximum compression on Ellesmere Island. More distally, folding that may have involved the entire thickness of the crustal lithosphere is indicated by gravity modelling to the west of Ellesmere Island, along the west coast of Axel Heiberg Island

(Stephenson & Ricketts 1990). A gravity low exists over northern Nares Strait, occurring slightly inland, in the zone of severe deformation of the Eurekan orogeny, but is not present south of Bache Peninsula (Figure 1.2). This has been interpreted as indicating thicker crust in

Airy isostatic equilibrium (Sobczak et al. 1990). On the whole, the gravity studies over the

Nares Strait region illustrate significant variations in crustal thickness in the area, not unlike the pattern of anomalies associated with continental collision zones such as the Alps, the Himalayas and the Appalachians (Jackson & Koppen 1985).

As a result of the Eocene collision between Ellesmere Island and Greenland, large-scale block rotations are indicated within Ellesmere Island by paleomagnetic results, but the structural complexity of the area creates substantial uncertainty in the analysis, with the amount of rotation estimated by some workers to be between 10° and 30° counterclockwise. Rotations may represent a means to accommodate strain within the Arctic islands, and the direction of the 26 rotation is consistent with the expected effects of the convergence (Jackson & Halls 1988;

Wynne etal. 1983).

The contrasting structural styles and trends that occur in the Paleozoic rocks on either side of Nares Strait record fundamentally different patterns of mid-Paleozoic and earlier tectonic events that, taken together, indicate that Nares Strait is a real dividing line that Jackson &

Koppen (1985) interpret as characteristic of regions that are linked by "cryptic suture zones", where subduction once occurred but has left little or no trace. Furthermore, the well-developed fold-and-thrust structures, combined with the lack of syntectonic magmatism or metamorphism, and the absence of any indication of the past presence of oceanic crust anywhere in the region, all combine to indicate that the Eurekan orogeny on Ellesmere Island is an expression of intraplate deformation rather than the delineation of a plate margin on Ellesmere (Tessensohn &

Piepjohn 1998), in contrast with other models (Oakey 2006; Harrison 2006) that use the Eurekan frontal thrust as the indicator of the location of the Greenland/North America plate boundary.

One of the markers used by researchers who believe that little to no movement has occurred across Nares Strait is the southern limit of Ellesmerian/Eurekan deformation. This marker, however, curves towards parallelism with the Strait as it crosses it. The Eurekan deformation boundary, therefore, cannot be used to define displacement (or lack thereof) along

Nares Strait (e.g., Higgins & Soper 1989, Okulitch et al. 1990).

Proposed Plate Boundaries

In contrast to the general consensus of the 1982 Nares Strait volume (Dawes & Kerr

1982b), more recent evidence makes it difficult to argue against the motion of Greenland relative to Canada, initiated by spreading in Labrador Sea and Baffin Bay. The most significant question 27

that still remains concerns the location of the boundary between the two plates. There is little

debate that its northern extent cuts Judge Daly Promontory (see Figure 1.2) on the northeastern

edge of Ellesmere Island as a closely-grouped series of N-trending faults (Harrison 2006 and

references therein), but to the south, after the faults disappear offshore, no actual observation of a

significant fault or fault system has been made. Based on what has been interpreted as continuity

of the Thule basin under Nares Strait according to detailed stratigraphic correlations by Dawes

(1997, 2006), Harrison (2006) proposed that the plate boundary follows the southern limit of

Eurekan deformation into Ellesmere Island, curving westward from its northward trend to the

north and intersecting the coastline of Ellesmere north of Bache Peninsula (Figure 1.8). Oakey

& Damaske (2006) agree with this interpretation based on what they claim are aeromagnetic

anomalies of dykes that cross from Kap Leiper on northwest Greenland across Kane Basin and

into Ellesmere Island, just north of Pirn Island, with no more than 10 km of sinistral movement between Ellesmere and Greenland. These interpretations, therefore, preclude

there being any major fault in the waterway of Nares Strait. Furthermore, observations of facing

sections across Smith Sound of successions of Thule Basin rocks lead to correlations across

Nares Strait down to member level, and individual sills on either side have been interpreted as being one and the same (Dawes 2007). However, others (e.g., Denyszyn et al. 2006; Neben et al.

2006) maintain that the preferred location for a major sinistral fault or fault system is in Nares

Strait. The location of the plate boundary between Greenland and North America is, in essence,

the Nares Strait Problem and will be examined further in this study. 28

Chron 24N

svtrtiw

Figure 1.8: Reconstruction at Chron 24N (~53 Ma) (modified after Oakey & Damaske 2006, used with permission) showing alternate placement of the Greenland/North American plate boundary within Ellesmere Island. The arrow denotes the direction of movement of Greenland relative to Canada at this time, with the position of the dots representing the change in position at subsequent chrons (small numbers). The frontal thrust of the Eurekan Orogen is marked, as is the proposed alternate position of the "Wegener Fault" in Jones Sound. 29

Mafic intrusions

Mafic dyke swarms are useful "piercing points" in paleogeographic reconstructions as they tend to be emplaced within a relatively short time span, have a relatively simple and well- defined geometry, and can cover large areas that may have subsequently drifted away from each other (e.g., Buchan & Ernst 2006). The rocks of various ages and types in the Canadian Arctic and Greenland have been cut by numerous phases of mafic intrusions (Figure 1.9) (Fahrig et al.

1971; Frisch 1984a,b,c; Kalsbeek & Taylor 1986; Brown et al. 1987; Nielsen 1990; Dawes 1998;

Hald & Tegner 2000; Dawes & Garde 2004; Buchan & Ernst 2004 and references therein). The southern peninsula of Greenland is intruded by the ca. 2.2 Ga Iggavik and MD dykes, and the southwest coast is intruded by the ca. 2.0 Ga Kangamiut dykes . Mesoproterozoic mafic dykes in southwest Greenland and Labrador (Figure 1.6) can also be used to correlate across the

Labrador Sea, and further constrain positions at that time. The Gardar dykes (subdivided into

BD0, BDl, BD2, and BD3) of southwest Greenland have also been correlated with a northeast- trending dyke swarm, the 1273 ± 1 Ma Harp swarm (Cadman et al. 1993, Buchan et al. 1996) of central Labrador. While the BDl and BD2 dykes have yet to be precisely dated, the BD0 dykes are dated by U-Pb (baddeleyite) between 1279-1284 Ma (reported in Upton et al. 2003). East- trending Nain LP dykes dated at 1280-1277 Ma, as well as 1268 Ma Nutak dykes are found north of Nain (Buchan et al. 1996). The Nain LP dykes have been correlated with ESE-trending, ca.

1280 Ma BD0 dykes of southwest Greenland by their age, paleomagnetism and geochemistry, and are brought into alignment by closing the Labrador Sea.

The coastal areas of Baffin Bay, as well as the western half of Ellesmere Island and the north and east of Greenland host the High Arctic Large Igneous Province, a series of dykes, sills and volcanics that formed as a result of Cretaceous tectonic movement of Greenland (Tarduno 30

1998). This event is believed to have been caused by the movement of North America and

Greenland over what would eventually become the Iceland mantle plume, with both the 120-90

Ma magmatism and the Greenland/North America rifting itself being attributed to the plume's impingement upon the crust (Harrison et al. 1999; Maher 2001; Lawver et al. 2002).

The Melville Bugt dyke swarm

The Disko Island region of Greenland is cut by the ca. 1630 Ma (Hamilton et al.

2004) Melville Bugt diabase dyke swarm, which parallels the western coastline of Greenland to as far north as the Thule region of northwest Greenland. Its areal extent is difficult to establish as at its northwest and southeast limits the swarm disappears under the Inland Ice of Greenland, and while it may continue under the Inland Ice to the east, it has not been identified on Baffin

Island or under the waters of Baffin Bay to the west (Nielsen 1990). However, it is certainly a voluminous magmatic event, made up of dykes up to 500 m wide but more commonly 50-100 m wide, and at least one dyke can be traced continuously for ~400 km up the western coast of

Greenland (Kalsbeek & Taylor 1986). Dykes with a similar trend have been identified on

Ellesmere Island south of Bache Peninsula (Nielsen 1990; Frisch 1988) and the identification and verification of any likely candidates for the Melville Bugt swarm in Canada is an important part of this study.

A single Melville Bugt dyke to the south, near Disko Island, has been dated at 1629±1

Ma (Hamilton et al. 2004), and the paleomagnetic result from that study (VGP of 38°N, 115°E, dp = 9°, dm = 15°) is the only previously published result from the Melville Bugt swarm. This swarm represents a critical point on the apparent polar wander path for the Laurentia paleocontinent, as its pole would fill a time gap between the "key poles" (Buchan & Halls 1990) 31 of the 1740 Ma Cleaver dykes (Irving et al. 2004) and the 1590 Ma Western Channel Diabase

(Irving et al. 1972; Hamilton & Buchan 2007). As the paleomagnetic record for any continent in the Paleoproterozoic is poor at best, to establish a more reliable paleopole for the Melville Bugt dykes is an objective of this study. This pole would be important for testing Laurentia/Baltica reconstructions for the Paleo- and Mesoproterozoic, since a key pole exists at this time in Baltica

(the Sipoo Subjotnian quartz porphyry dykes of Finland (Mertanen & Pesonen 1995)), and none known otherwise for Laurentia.

Proterozoic diabase dykes of unknown affinity

There are several mafic dykes and dyke swarms not associated with either the Franklin or Melville Bugt events that have been identified over the course of geological mapping in the High Arctic region, and isotopic dating has identified some as being of

Proterozoic age. There is a set of NE-trending diabase dykes at the eastern end of Inglefield Bay called the Nigarfivik dykes by Buchan & Ernst (2004) (Figure 1.9). One has been observed to cut a Melville Bugt dyke and has been dated by the K/Ar method (whole rock) to 1313 ± 39 Ma

(Dawes & Rex 1986). Along the south coast of Devon Island, west of Dundas Harbour, a N/S- trending dyke has also been dated using the K/Ar (whole rock) method to 1336 ±15 Ma (Frisch

1988). There is, therefore, evidence for a Mesoproterozoic magmatic event in the High Arctic, though it does not appear to be widespread. It will require more precise (U/Pb) geochronology to verify these results. 00000000O* wwwwwwwuuO II inII II II nII II II II iII iII Figure 1.9: Map of the eastern Canadian Arctic and Greenland showing locations, schematic trends, and ages of various mafic dyke swarms in the area. References in text. 33

There are dykes mapped on the southwest coast of Greenland, near Disko Bay with an

E/W trend (Garde 1994) that also have no apparent association with the Melville Bugt swarm

which is also found in the area. These have been named the Ataa Sund swarm by Buchan &

Ernst (2004), and while very little is known about them, their trend and presumed Proterozoic

age (Buchan & Ernst 2006) indicate that upon the reconstruction of Roest & Srivastava (1989) in

which western Greenland is located adjacent to eastern Baffin Island, the Ataa Sund dykes may

represent an eastern continuation of the Franklin swarm into Greenland (Figure 1.9).

The Franklin igneous event

Eastern Ellesmere Island hosts the ca. 720 Franklin dykes and the undated

Clarence Head swarm on the southeastern part of Ellesmere Island. The Franklin dykes and sills

are also found to the west and south throughout the southern part of the Arctic archipelago, on

Victoria Island, Somerset Island, Boothia Peninsula, Melville Peninsula and Baffin Island. Also

found on Baffin Island are the undated Strathcona Sound dykes, which are thought to be

associated with the Franklin swarm but have been observed to crosscut Franklin dykes (Christie

& Fahrig 1983). With respect to the Nares Strait problem, the Franklin and Melville Bugt

swarms are both found in the area in question, and the Franklin swarm intersects Nares Strait at a high angle, making these dykes particularly suitable candidates for piercing points that can be

matched from one continent to the other.

The Franklin igneous event emplaced sills and dykes across more than 2500 km of lateral

distance in the Canadian Arctic and Greenland (Figure 1.10), making it one of the largest dyke

swarms on Earth. It is comprised of large (>20m in width) diabase dykes in mainland Canada,

Baffin Island and Greenland (Fahrig 1985; Frisch 1984a,b,c; Dawes 1991). Also included are the 34

Victoria Island sills and the Natkusiak volcanics of Victoria Island, the Coronation sills of mainland Northwest Territories, the Lasard River dykes, and the Thule sills of Greenland

(Shellnutt et al. 2004; Ernst et al. 2004; Buchan & Ernst 2004; Dawes 1991; Frisch 1984a,b,c;

Fahrig et al. 1971). Also proposed to be associated with the Franklin event are the Tree River

dykes of the Northwest Territories of Canada (Buchan & Ernst 2006). The dyke swarm as a whole has a fanning pattern, radiating from a focus located more or less north of Victoria Island

(Figure 1.10).

The Franklin magmatic event fills an important time gap in the Proterozoic of North

America, which is currently represented by only a handful of widely-spaced paleopoles with both high-precision U-Pb ages and well-characterized primary magnetizations (Buchan et al. 2000).

The currently-accepted age of 723 +4/-2 Ma (Heaman et al. 1992) and paleomagnetic pole

8.0°N, 163°E, A95=4.0° (Buchan et al. 2000) have been used extensively in Neoproterozoic plate reconstructions (e.g., Buchan et al. 2001). The Franklin pole is particularly important as the

Franklin magmatic event is postulated to be related to a former mantle plume north of Victoria

Island (Figure 1.10) that has been tied to the incipient breakup of the Neoproterozoic

supercontinent Rodinia (Shellnutt et al. 2004; Heaman et al. 1992). Sturtian global glaciation is proposed to have occurred during this time period, so accurate knowledge of plate reconstruction

from paleomagnetism may prove important for testing the "Snowball Earth" hypothesis

(Hoffman et al. 1998).

To date, several authors have reported ages to varying degrees of precision for the

Franklin dykes, a summary of which is presented here: 35

Reference Age Range (Ma) Error Range (Ma) Method Used Notes Fahrig 1973 535-841 27-124 K-Ar Franklin dykes on Baffin Fahrig 1973 675 (average) Paleopole Franklin dykes on Baffin Pehrsson & Buchan 689-758 16-104 U-Pb "Borden" dykes on 1999 Baffin, associated with Franklin Dawes & Rex 1985 645-729 13-50 K-Ar Sills/Dykes (chilled margins) from Thule area Heamanetal. 1992 714-724 2 U-Pb Franklin dykes and extrusives, Victoria Is. andNWT Ernst et al. 2004 708 4 Ar-Ar Lasard River dykes and sills Frisch 1988 625-791 9-60 K-Ar Dykes on Devon, Ellesmere Escher& Watt 1976 676 25 K-Ar Thule Group dyke Dawes et al. 1982 627 25 K-Ar Kap Leiper dyke Table 1.2: Summary of geochronological studies on the Franklin magmatic event.

While it is currently the working assumption that the intrusions being correlated here are of the

Franklin swarm (e.g., Dawes 1997), it will require precise geochronology to confirm this. As dykes of the Melville Bugt swarm have not been identified in Canada, the Franklin swarm represents the best possibility of successfully locating the plate boundary between North

America and Greenland.

The geochronology of mafic intrusions has advanced significantly since the development of dating techniques using baddeleyite (ZrC^), a U-bearing phase not uncommon in silica- undersaturated rocks (e.g., Krogh et al. 1985; Heaman & LeCheminant 1993). Baddeleyite is present in mafic rocks where silica activity in the melt was low and zircon is absent. As the

Franklin dykes of Baffin Island and Victoria Island have been reported elsewhere as olivine- bearing (Heaman et al. 1992; Jefferson et al. 1994), as have the Melville Bugt dykes (Nielsen

1990), it is the expected mineral to be extracted for U-Pb geochronological analyses of the dykes. 36

In summary, the primary objective of this research is a multidisciplinary comparison of mafic dyke swarms in the area in order to confirm or deny the existence of the Wegener fault between Ellesmere Island and Greenland, and to constrain the range of motion between the two plates (if any). More generally, paleomagnetic and geochronological data from various mafic intrusions in the High Arctic will contribute to the understanding and refinement of the Laurentia apparent polar wander path, and to a more complete knowledge of Arctic geology. 37

Figure 1.10: Distribution of intrusions or lava flows associated with the Franklin magmatic event (modified after Fahrig (1985), with dyke locations also from Frisch (1984a,b,c) and Dawes (1991)). CL = Clarence Head dykes; CS = Coronation Sills; DI = Devon Island dykes; LR = Lasard River dykes; NB = Natkusiak basalts (extrusive); SS = Strathcona Sound dykes; TD = Thule dykes; TR = Tree River dykes. Star indicates probable location of mantle plume source (Heaman et al. 1992, Buchan & Ernst 2006).Dykes are shown in black, sills in green, volcanics in blue. 38

2. Methodology

Paleo/Rock Magnetism

Paleomagnetism is the study of the Earth's magnetic field through time as recorded by the magnetization of minerals. It is a powerful tool in plate tectonic reconstructions, but is reliant upon two fundamental assumptions (Butler 1992): that the primary magnetization acquired by these minerals as they cool below their Curie temperature in an ambient geomagnetic field is in fact parallel to that field; and that the geomagnetic field, when averaged over time, closely resembles a geocentric axial dipole, meaning that it is dominantly dipolar in nature and the geomagnetic pole (again, when averaged over time) coincides with the geographic pole. When a ferromagnetic mineral, one that is capable of acquiring a spontaneous magnetization, cools below its Curie temperature, the spins of valence electrons within the lattice structure of the mineral align - and if a magnetic field is imposed upon it at this point, that alignment is directed into parallelism with the applied field. Above the Curie temperature, which is unique to each mineral, thermal agitation prohibits the acquisition of a spontaneous magnetization (McElhinny & McFadden 2000). Magnetizations acquired in this manner are termed thermoremanent magnetization (TRM), and as this magnetization (in unmetamorphosed rocks) is acquired at the time of formation of the rock, it is this magnetization, known as the primary magnetization, that is sought in this study.

An important factor in the ability of a mineral to accurately retain this primary magnetization is grain size. Magnetic domains are, in essence, zones of uniform magnetization within a crystal. In order to minimize magnetostatic energy, a crystal may be divided into several domains with opposing magnetization vectors. Larger grains of a ferromagnetic mineral, therefore, tend to be made up of several domains, and are termed multi-domain (MD). If a 90°IV 80°N 70°W 80°^

9o°iv 80°W

^^| = Franklin dyke swarm (ca. 720 Ma) ^B = Clarence Head dyke swarm (ca. 715 Ma) I 1 = Thule sills (712 Ma) B^[ = Dundas Harbour dyke (1337 Ma) ^^| = Melville Bugt dyke swarm (ca. 1625 Ma)

Figure 2.1: Map (polar stereographic projection) of diabase dykes sampled for this study, and their approximate ages and trends. Red symbols indicate dykes of the Franklin dyke swarm; Purple indicates dykes of the Clarence Head swarm; Orange denotes the Dundas Harbour dyke; Green indicates dykes of the Melville Bugt swarm. The orientation of the symbols indicates the trend of the dyke at the location where it was sampled. Yellow denotes a Thule sill. crystal is small enough, the energy required to generate domain boundaries is greater than the

energy saved by having a multi-domain grain. In this case, the crystal is made up of just one

domain and is termed single-domain (SD). A SD grain has a magnetization equal to its

saturation magnetization (the maximum induced magnetic moment that can be acquired in an

applied field). MD grains have magnetizations significantly less. Therefore, MD grains are more weakly magnetized than SD grains, and therefore more susceptible to later remagnetization. The maximum grain size for a single-domain crystal of magnetite, the most

common ferromagnetic mineral in mafic rocks, is about 1 micron (Butler 1992; Dunlop &

Ozdemir 2000). As most grains in a rock are greater in size than that, they are susceptible to being affected by later remagnetization. The ability of a crystal to be remagnetized is described by its coercivity, the applied field required to reset the crystal's magnetization. MD grains tend to have low coercivities, SD grains have high coercivities (McElhinny & McFadden 2000). MD grains, therefore, are liable to acquire a viscous remanent magnetization (VRM), which is a remanent magnetization that is gradually acquired over prolonged exposure to weak magnetic

fields, such as the geomagnetic field. Another way in which a rock can be remagnetized is by heating. Larger grains have lower blocking temperatures (the temperature above which the magnetization can be reset), and so even at temperatures well below the Curie temperature a partial TRM (PTRM) can be acquired. These effects can result in the larger grains of magnetite

in a rock having a secondary magnetization that can to varying degrees overprint the primary

TRM of the rock. These secondary magnetizations (if present) must be removed in the laboratory (if possible) before analysis if the remanence direction is to be included in a paleomagnetic pole. Without removal of secondary magnetizations, what is being observed in 41 the rock is a natural remanent magnetization (NRM), the vector sum of all the components of magnetization acquired by the rock, of which the primary TRM is often just one (Butler 1992).

The methods used for "cleaning" these secondary magnetizations from the sample are well-established (e.g., Collinson 1983; Butler 1992; McElhinny & McFadden 2000). Two common methods are used: alternating-field (AF) demagnetization and thermal demagnetization.

AF demagnetization involves stepwise increasing fields of selected alternating field along three orthogonal axes. The magnetization is remeasured after each step and the changes in both direction and intensity are recorded and analyzed, with the routine carried out until the sample is completely demagnetized, or until the limit of the demagnetizer is reached. Using the thermal demagnetization method, the sample is heated incrementally to higher temperatures, cooled in a field-free space after each heating, and the magnetization is again measured and analyzed until the intensity of the sample's magnetization drops to zero, typically at or around the Curie temperature.

The principle behind progressive stepwise demagnetization of the rock is that as increasing fields or temperatures are applied, those grains with low coercivities or unblocking temperatures - those that carry secondary components - have their contribution to the magnetization of the rock reduced to zero, leaving only the highest coercivity/unblocking temperature magnetizations which are the most resistant to remagnetization and therefore most likely (though not always) primary. Thermal demagnetization requires stepwise heating of the sample to remove any PTRM, and is necessary for demagnetizations of minerals such as hematite or pyrrhotite that have coercivities too high for them to be demagnetized using AF techniques. Measurement of the sample is then carried out using a spinner magnetometer. The specimen is placed in a specific orientation in a holder, and is rotated mechanically. The 42 magnetometer detects the oscillating magnetic field produced by the rotating magnetic moment of the rock. As the rotation frequency of the spinner is known, and as the amplitude of the current produced is proportional to the intensity of magnetization, the specimen is then spun in six orthogonal orientations (i.e., +X, -X, +Y, -Y, +Z, -Z) to determine the intensity and direction of the magnetic vector (Collinson 1983; Butler 1992).

A total of 390 samples were collected, either as field-drilled oriented cores or as oriented blocks, from 24 E-W trending dykes believed to be of late Neoproterozoic age (19 in Canada and

14 in Greenland), two Neoproterozoic sills in Greenland, five N/S-trending dykes of unknown affinity in Canada, and three NW-trending dykes in Greenland believed to be of Paleoproterozoic age (Figure 2.1). The samples were oriented by sun compass at all but ten sites (DH, CM, HM,

CW, PH, OR, SG2, NS, BE and HF), when weather conditions made sighting the sun impossible.

In these cases, multiple checks for the reliability of the magnetic compasses were made

(typically involving comparing the bearing given by the compass at various distances away from the dyke) before accepting a reading as satisfactory. Sample collection emphasized chilled margins of dykes in the expectation that the smaller grain size of magnetite in the margin samples would preserve high-coercivity remanence. As described above, single-domain magnetite carries a high-coercivity remanence that is the least likely to be replaced by viscous components (e.g., Butler 1992). When possible, a transect across the dyke's width was carried out, collecting 8 to 12 samples evenly spaced from one margin to the other. This was not always possible, however, if the margins or even significant portions of the dyke width were not physically accessible. When the host rock consists of Archean or Paleoproterozoic gneisses, as in Canada and occasionally in Greenland, the diabase dykes weather more than the host rock, and the chilled margins were frequently buried under rubble and scree. When the host rock is 43

sedimentary (i.e., the Thule sediments of Greenland), the dykes prove more resistant to

weathering than the host rock and margins tend to be much more readily accessible along the

resulting ridges.

The first season of sample collection was carried out by Henry C. Halls in 2001 on the

Nares Strait Geocruise (samples PH, TB, NU1, NU2, PK, KL and OR) and the remaining sites

were collected by HCH and SWD from 2002 to 2005. All sites were located in terms of their

latitude and longitude by a hand-held Garmin Etrex global positioning system (GPS) unit or by

onboard helicopter GPS system. Block samples were subsequently drilled in the laboratory

using a drill press, and all drill cores were sliced into cylindrical specimens 2.45 cm in both

length and diameter. At least one specimen from each sample was subjected to detailed AF

demagnetization, using a Schonstedt GSD-1 single-axis demagnetizer, in order to remove stray

or Present Earth Field (PEF) components. Field increments averaged 0.5 mT up to maximum

fields of 95 mT. After each demagnetization step, the direction and intensity of remanent

magnetization were measured using a modified Digico spinner magnetometer with repeatability

of magnetization intensities down to ~10 ~3 Am"1. Measurement procedure included an

averaging algorithm to reduce the dependence of results on the last sample axis to be

demagnetized in the single-axis demagnetizer. Thermal demagnetization was also used for

selected specimens, using a Schonstedt TSD-1 thermal demagnetizer. Stepwise heating was

conducted according to a measurement routine that involved measurement of the sample's

remanent magnetization at 20°C (NRM), then at increasing temperatures until the intensity of the

magnetization of the sample was too low to continue meaningful measurements. This routine was designed to isolate the direction of magnetization carried by magnetite (pure magnetite has a

Curie temperature of 580°C) at a maximum level of detail while being able to note the presence 44 and nature of low-temperature viscous remanent magnetization or partial thermoremanent magnetization (e.g., McElhinny & McFadden 2000). The orientation of the samples while in the cooling chamber was routinely changed in order to counteract, or at least identify, the presence of any trace ambient magnetic field within the chamber.

The paleomagnetic data were plotted on stereonets and vector diagrams (Zijderveld 1967) and then analyzed using Principal Component Analysis (Kirschvink 1980) that included a search routine to find all linear segments on vector diagrams consisting of three or more points and that pass a minimum acceptance criteria given by a goodness of fit parameter, the Maximum Angle of Deviation (Kirschvink 1980), which was set at < 10°. Linear segments that pass through the origin indicate "stable end-points", where only a single remanence component remains, with its magnetization intensity dropping with progressive demagnetization steps but maintaining a constant direction. Linear segments that do not pass through the origin are also important to identify as they define the directions of components that are being removed in the demagnetization procedure. For the suite of samples collected at each site, mean values of declination, inclination, paleolatitude, paleolongitude, Fisher precision parameter k (Fisher 1953) and semi-angle of 95% confidence for the mean (0195) were calculated, as well as the associated virtual geomagnetic pole (VGP) position. Where multiple specimens were measured from a single sample, their magnetization directions were averaged before inclusion into the site mean.

Plotting and calculations were carried out using the software package of Super-IAPD by Trond

Torsvik (NGU, Trondheim). The mean directions of the paleomagnetic data for each site were then converted to paleomagnetic poles in order to correct for the significant geographic spread in the sample locations. In order to present this location-corrected information in terms of paleomagnetic declination and inclination and to illustrate the presence of reversals and potential 45 outliers, the poles for each site (from both Canada and Greenland) were then re-converted to

declination and inclination values for a common, roughly central site at 77°N, 70°W in order to

eliminate the influence of sample location (specifically latitude) on the declination and inclination measured.

The dykes are close to vertical, with dips not less than 80°, with well-preserved chilled margins and in Greenland cut sub-horizontal Thule Group sediments and are overlain by flat- lying Paleozoic strata. Therefore tectonic activity that might remagnetize the rocks, or rotate magnetization directions, is likely to be minimal.

Susceptibility versus temperature data were obtained from at least one sample from each dyke using a Sapphire Instruments susceptibility meter to determine the magnetic mineralogy of the samples. This is important because it enables the quick and accurate identification of minerals in the sample that are carriers of magnetization. For example, the presence of pyrrhotite with a curie temperature of 320°C could indicate a secondary magnetization.

Measurements were made on one specimen from each sample collected across the Kap Leiper dyke (KL in Figure 2.1) in order to model the linear total field anomaly discovered by an aeromagnetic survey of the Kane Basin (Oakey & Damaske 2006), and additional measurements were made on specimens from dykes NS and CL that yielded atypical results. During measurement, the specimens were heated from 300 to 900 K and susceptibility measurements automatically recorded at 5-degree intervals.

Petrography and Geochemistry

For the purpose of comparing dyke swarms, the use of transmitted-light microscopy provides a rapid first-order assessment of a rock's petrology, as dykes from the same 46

swarm might be expected to share common mineralogy and textural relationships between

minerals. It also allows a quick way to check for the effects of alteration. Fluids flowing

through the rock following emplacement and crystallization can alter the existing mineral

assemblage and precipitate new minerals such as magnetite or pyrrhotite, which can affect the

rock's magnetization by adding a component of chemical remanent magnetization (CRM).

Transmitted-light petrography can often determine whether this has taken place in a rock by

identification of minerals characteristic of alteration such as chlorite, epidote, or clays, the

presence of which may indicate that the presence of a secondary magnetization component is

likely.

Geochemical analysis is a more in-depth and quantitative measure of the petrogenetic

connections between the dyke swarms. X-ray fluorescence (XRF) is a widely-used analytical technique for elements with atomic numbers greater than that of oxygen (i.e., >8). The sample, in the form of pressed powder pellet, is irradiated with a beam of X-rays, and the elements in the

sample are excited by the absorption of the beam and upon relaxation emit fluorescent X-rays in wavelengths characteristic of the element and with intensity proportional to the amount of that

element (e.g., Skoog et al. 1997). Similarity in the geochemical signatures of the dyke swarms

or of individual dykes would further strengthen a correlation. Distinctive compositions, particularly as seen in incompatible or non-mobile elements, will be especially useful, though

anomalous abundances of other elements may also exist, leading to a possible tectonic or petrological association with other intrusions in the area. Elements of particular interest include

Ti, Zr, Nb and Y, as these are readily and quickly analyzed to a high level of accuracy by XRF,

and are generally immobile during weathering and metamorphism up to and including greenschist facies (Jensen 1976, Pearce 1996). Individual dykes with characteristic geochemical 47

signatures may be identified, as geochemical changes can be negligible for hundreds of

kilometres along a dyke's length, though it may cut a diverse range of country rock compositions

(Kalsbeek & Taylor 1986). Furthermore, dykes may have varying geochemical characteristics

across a single swarm (Bossi et al. 1993). Therefore, any distinctive geochemical pattern in the

lateral distribution of the dyke swarm will be useful as a correlative tool, as it will enable the

relative location within the swarm of individual dykes to be established.

Thin sections were prepared from 25 samples, including at least one from each site

shown in Figure 2.1. They were examined under a transmitted-light microscope using both

plane-polarized and cross-polarized light.

Geochemical analyses were carried out on samples collected within 30 cm (if possible) of

dyke chilled margins, in order to avoid the effects of fractionation within a dyke. These margin

samples were checked for any signs of alteration. The samples were crushed in a jaw crusher, powdered in an alumina swing mill, and analyzed at the University of Toronto, using X-ray

fluorescence methods on a Philips 2404 spectrometer for major and minor elements, and neutron

activation analysis (irradiated at the Royal Military College, Kingston) for trace and rare earth

elements. Eight of these samples, were sent to SGS Mineral Laboratories for major and minor

element geochemistry using XRF to provide an inter-laboratory check. Fourteen samples were

separately sent to SGS Mineral Laboratories for ICP-MS measurement of rare-earth elements.

Tectonic Reconstruction

In order to generate plate-tectonic reconstructions of the North American paleocontinent at the time of formation of the various dyke swarms, as well as to plot apparent polar wander paths (APWPs), the software program GMAP (Torsvik & Smethurst 1999) was 48 used. Datasets from a number of sources were entered into the program manually to create

APWPs and to correlate the data from this study with existing data.

U-Pb Geochronology

U-Pb geochronology is an effective method for obtaining reliable ages of Earth materials. Zircon and baddeleyite are well-known as reliable geochronometers, as they contain almost no lead in their crystal structures upon formation, and decay constants of the two most common isotopes of uranium, U and U, are well-established (Jaffeyetal. 1971). The U-Pb system is also a uniquely powerful tool as two separate decay schemes can be used, with common parent and daughter nuclides, which allows geochronological information to be obtained even in some cases from disturbed systems (e.g., Dickin 2005). When a mineral has remained a closed system to U and Pb, its age as determined by the decay schemes of both 235U and U will be consistent, or concordant, and will lie upon a "concordia curve" in

(206Pb/238U)/(207Pb/235U) space, the curve denoting the points where the ratios of the measured isotopes agree in age. Should the U/Pb system have been disturbed, the analysis will plot off the concordia curve, and be termed discordant. Meaningful age information can still be obtained from data disturbed by a single event. In fact, given sufficient data, the original crystallization age can be determined as well as the age of the event that caused the disturbance in the system.

To determine absolute concentrations, isotope dilution methods are used (Krogh 1973). A solution (the "spike") with a known isotopic composition of, in this case, Z"U and zuTb is added to the sample prior to dissolution and eventual measurement in the mass spectrometer. As the analysis measures ratios of isotopes (as opposed to quantities), only by adding a known amount of spike to a known amount of sample can concentrations be established. One potential problem with this technique is that knowledge of the concentration of U is dependent upon accurate

knowledge of the weight of the sample. When single baddeleyite crystals are being analyzed,

which can each weigh less than 0.1 \xg, the exact weight may not be well-defined, though the age

information is unaffected as it relies upon the isotopic ratios, not their absolute abundances.

Baddeleyite (ZrC-2, monoclinic, usually thin, striated blades in shades of brown) and

zircon (ZrSiC<4, tetragonal, usually prismatic and colourless) may both be found in mafic dykes,

and both were sought for U-Pb analysis in this study. Whether baddeleyite or zircon is found is

dependent upon the silica activity of the rock - silica-saturated magmas will precipitate zircon,

silica-undersaturated systems will precipitate baddeleyite. Therefore, not only the initial

concentration of zirconium but the degree of fractionation that develops during crystallization

determines which, if any, of these phases are present in the rock (LeCheminant & Heaman

1993). However, local silica activity within the crystallizing magma (as opposed to bulk

composition) may determine whether baddeleyite or zircon crystallize, as many baddeleyite- bearing rocks also contain zircon, even within the Franklin dyke swarm itself (Heaman et al.

1992), and baddeleyite has been observed in apparent equilibrium with quartz (LeCheminant &

Heaman 1993). Although zircon can often be found as inherited grains in mafic dykes as it can be entrained from wall rocks, demonstrated instances of xenocrystic baddeleyite are unknown

and in all cases should retain high temperature closure ages in diabase or gabbro. Furthermore, baddeleyite is unlikely to be a xenocryst as it is rare even in mafic rocks in which it has

crystallized, which are themselves uncommon host rocks for intrusions into continental crust.

Baddeleyite is, therefore, often a more direct and reliable indicator of a dyke's age (Krogh et al.

1985). Petrography and geochemistry should be able to determine which phase is most likely to be present, though the likelihood of encountering one of these minerals in a thin section is small. For U-Pb geochronological analyses at the University of Toronto's Jack Satterly

Geochronology Laboratory (JSGL), 3- to 10-kg blocks were taken from the coarsest-grained parts of most dykes, typically the centre, in the hopes of finding the highest possible concentration and largest grains of the mineral baddeleyite in the dyke. Zircon would not be expected in silica-undersaturated mafic rocks such as diabase (Heaman & LeCheminant 1993), though mineral separation methods that isolate baddeleyite would also isolate any zircon present in the rock, which would be identified in the final separation processes. At every stage of handling of the rock, the resulting heavy-mineral concentrates, and the baddeleyite grains themselves, constant care was taken to minimize any chance of sample contamination through meticulous cleaning practices. Initially, baddeleyite grains were separated from the rock using

"traditional" mineral separation methods, i.e., those used to separate zircon. The rock was first crushed into cm-scale pieces in a jaw crusher. These pieces were then pulverized in a Bico disc mill, comminuted to an ideal size fraction of <200 microns. Passing the resulting powder over a

Wilfley water-shaking table eliminated the clay-sized fraction, and also concentrated the heaviest minerals. This concentrate was then sieved to screen out particles too large to pass through a 70 mesh screen (approximately 210 microns) that were insufficiently pulverized in the disc mill and which may represent aggregates of minerals rather than single grains. The sample was then poured in "free fall" past a strong electromagnet, which pulled off ferromagnetic minerals

(almost entirely magnetite and iron metal filings from the disc mill). A heavy-liquid separation using Bromoform in a separatory funnel was carried out on the non-magnetic portion of the sample, removing minerals with specific gravities below 2.9 such as feldspar and quartz. The collected sample was then passed through a Frantz Isodynamic Separator™, which isolates different fractions of paramagnetic minerals based upon their magnetic susceptibilities as 51 increasing currents are applied to the instrument's electromagnets. The non-magnetic fraction was then put through another heavy liquid separation step, using methylene iodide, which floats off apatite and other minerals with specific gravity less than 3.3. The heavy fraction was then again passed through the Frantz separator, using the maximum magnetic field strength and altering only the slope and tilt of the track along which the grains travel to optimize the isolation of the non-magnetic, heaviest mineral grains from the rock. The non-magnetic fraction was then hand-picked under a microscope to isolate the best-quality baddeleyite grains to be analyzed, and to identify and isolate zircon, if present. This procedure, however, was found to be relatively ineffective in extracting baddeleyite. Often samples from ten different dykes, only three (CG,

CL and OF) yielded any baddeleyite grains in the final microscopic examination.

To increase possible baddeleyite yield, the number of mineral-separation steps was minimized in order to reduce the opportunity for baddeleyite loss. Sample CH was crushed and milled as above, but the powder was then placed into a 1L beaker filled with water. The fraction of the sample that remained suspended in water after repeated stirring and ample time for settling was decanted off into a separate vessel. A handheld magnet was used to extract magnetic material (primarily magnetite) from the remainder. The sample was then further subdivided into heavy and light fractions using methylene iodide and the heavy fraction was picked through by eye. This yielded several fresh baddeleyite crystals and fragments, in contrast to an earlier attempt using the conventional procedures.

Subsequent samples were similarly crushed and milled, but the heavy minerals were separated using only the Wilfley water-shaking table, following a technique modified after Soderlund and

Johansson (2002) that involved concentration of fine, heavy minerals on the table, removal of magnetic material using a hand-held magnet and picking baddeleyite from the residue. The 52 sample was kept under liquid at all stages of the separation process. The powder from the disc mill was made into a slurry by adding water and a small amount of dish soap to act as a surfactant and to prevent the small, flat baddeleyite grains from "riding" the surface tension of the water as it passed over the ridges of the Wilfley table. The sample is conventionally added to a container on the table under a water spout with a high flow rate. The container was removed, the spout plugged, and one of the tubes that distributes water from a series of evenly-spaced holes to cover the rest of the table in water was replaced with an extended version to supply a low flow of water covering the entire table. The tubes were also angled backwards, directing the streams of water to bounce off the back wall of the table before flowing over it in order to eliminate the presence of eddies. The concentrate was removed from the table using a pipette, and placed in a small beaker containing ethanol to prevent oxidation of iron during storage. This residue was then examined under a microscope in a Petri dish under ethanol (Figure 2.2) and baddeleyite grains were picked using a small pipette and removed to a dry small Petri dish and photographed. Only a few baddeleyite blades and fragments, most typically under 80 microns in length and weighing less than a microgram, were recovered from each sample, though the recovery rate increased to 100% - every sample that underwent this procedure yielded sufficient amounts of baddeleyite for analysis, though the grains tended to be much smaller than those from the "traditional" method, which were commonly over 100 microns in length. This may be because while baddeleyite is apparently ubiquitous in all diabase of this study (every sample from dykes representing four separate magmatic events - the three from this study and a

Mesoproterozoic dyke from India yielded baddeleyite), the traditional method is only capable of isolating grains above a certain size, and is prone to losing the smaller grains in any one of the 53

- « .T^ 4

>-* c ;* •v •'"v ». «* ** - **/f-il nit*.

* * * **"

. .*•. V .

: "*, t: ' ^

Figure 2.2: a) Residue removed directly from the surface of the Wilfley table, using the technique modified from Soderlund & Johanssen (2002), that is then picked through in order to find baddeleyite. b) A single baddeleyite in the centre of the photo after other minerals have been cleared away. Actual width of both photos is approximately 1 cm. 54

many processing steps involved. In particular, the steps involving the Frantz Separator seem to be the most likely to lose track of baddeleyite, as when not under liquid, the crystals' small size and flat habit make them prone to being attracted by static electricity to other grains.

Dissolution, isotope dilution and sample loading methods were as described in Krogh

(1973), using a Pb- U spike and miniature bombs. No chemical separation procedures were required. The baddeleyite samples were then analyzed by a VG354 thermal ionization mass spectrometer using a Daly collector in pulse counting mode. The mass discrimination correction for this detector was constant over the period of the analytical work at 0.07%/AMU. Thermal mass discrimination corrections were 0.10%/AMU. Dead time of the measuring system was about 20 nsec and was monitored using the SRM982 standard.

Ages were obtained from E/W-trending dykes on Ellesmere (CG) and Devon (BG) Islands as well as from the centre of the swarm in Inglefield Bay of Greenland (QA), as well as a Thule sill (CA). N/S-trending dykes from were sampled from Ellesmere Island (CH, CL and CV), as well as a NW-trending Melville Bugt dyke at the east end of Inglefield Bay (OF). A dyke from the south coast of Devon Island with a paleomagnetic direction not unlike that characteristic of the Melville Bugt swarm (DH) was analyzed as well. Dykes were selected that have similar geochemical and paleomagnetic signatures to the other dykes shown in Figure 2.1, leading to confidence that the ages obtained for these dykes are representative of their respective dyke swarms. 55

3. Results

Paleomagnetic, geochemical and geochronological results reveal the presence of four distinct dyke swarms in the area (Figure 2.1). The vast majority of E/W-trending dykes are associated with the Neoproterozoic Franklin magmatic event, as are the N/S-trending Clarence

Head dykes. NW-trending dykes in northwest Greenland are a part of the Paleoproterozoic

Melville Bugt swarm. A single E/W-trending dyke on the south coast of Devon Island (here called the Dundas Harbour dyke), thought to be of the Melville Bugt swarm, has been determined to be of an altogether different magmatic event in the Mesoproterozoic.

The quality of the data obtained, particularly the paleomagnetic results, is chiefly limited by the number of sites collected. Over the course of five field seasons (2001 on the Nares Strait

Geocruise by Henry Halls, 2002-2005 by HCH and SWD), a maximum number of dykes were visited each year, controlled by such diverse factors as helicopter availability, the presence of the sun needed to orient the sample, the helicopter's ability to access an outcrop, and the field team's ability to safely sample a given dyke. Melville Bugt dykes (JP2 and QT in particular) were accessible only by small boat in iceberg-rich waters. Given these factors, and the particular remoteness of the region studied, nearly all the dykes mapped on the Canadian side (Frisch

1984a,b,c) were sampled (while avoiding dyke outcrop that could be an along-strike continuation of a dyke previously sampled), as were those in the Inglefield Bay area of Greenland (Dawes

1991). Due to logistical constraints, dykes in the Steensby Land area, between the PW dyke and the TB dyke (Figure 2.1) were impossible to sample. When attempting to resolve ca. 200 km of differential motion, the number of samples critically affects the size of statistical errors, and therefore the strength of proposed reconstructions. Therefore, being able to trace an individual 56

dyke across by its distinctive geochemical signature, or by a unique paleomagnetic direction

(such as the recording of a geomagnetic excursion, a typical aspect of secular variation), would be the most definitive test of the degree of relative motion between Greenland and North

America.

Geochronology

The uranium and lead isotopic results for each analysis (Table 3.1) were corrected

for common lead contamination using the isotopic composition of laboratory blank (see caption

to Table 3.1). The small size of the fractions (typically less than 0.5 micrograms) was necessary because of the low yield of small crystals (as described in Methodology) and in order to analyze the best quality crystals, i.e., those largest in size and without signs of alteration. In young

907

samples, radiogenic Pb has accumulated in a geologically short time from a parent isotope

(235U) that has been more rapidly depleted throughout Earth history than has the 206Pb/238U

system. Therefore measured amounts of 207Pb in young baddeleyites or zircons will be relatively 907 907 low, analytical precision will be poor, and the proportion of total Pb that will be common Pb 9^8 90A will be commensurately high. However, precise ages can still be determined using the ZJ0U-ZUDPb

system, which is only possible with low (normally sub-picogram) laboratory blanks. The use of

only one system renders assessment of secondary Pb loss effects difficult to quantify, though these tend to be much less pronounced in baddeleyite than in zircon, as it is more chemically

inert (Krogh et al. 1985). In these cases, as in this study, reproducibility of Pb/ U ages on

individual young fractions provides a first-order test of accuracy. Averages of 206Pb/238U ages

and linear regressions were calculated and plotted using programs of Davis (1982) and Ludwig

(2003). U decay constants are from Jaffey et al. (1971) and all errors are reported as 2-sigma. 57

Franklin dyke swarm

The Cadogan Glacier (CG) dyke (Figure 3.1a) was analyzed using four single baddeleyite crystals, and two fractions containing two crystals each. Grains were typically fresh, dark brown blade-like fragments, ranging from 50-150 um in the largest dimension. However, some grains were very dark and may have been somewhat altered. Furthermore, this sample suffered from having unusually high blanks (up to 9.9 pg of common Pb) due to the difficulty in handling the tiny baddeleyite crystals. Nevertheless, concordant Pb/ U ages were obtained, though some data scatter outside of error. The oldest data give an average ^°PbrJ0U age of 721 ±2 Ma (Table

3.1) while the other fractions define a resolvable range down to 714 ± 2 Ma. Possible reasons for this discrepancy are considered in the Discussion.

Baddeleyite grains from the Qaanaaq (QA) dyke in Greenland were rare, but fresh and relatively large (40-100 um long). They tended to take the form of long, thin needles and fragments, the smallest of which were combined into fractions of two grains each. Three out of four fractions of one or two grains each gave precise overlapping data with an average 206Pb/238U age of 721±4 Ma (Figure 3.1b).

Data from the Belcher Glacier (BG) dyke are discordant (Figure 3.1c). The average

^'Pb/zuoPb age is 726 ± 24 Ma from four fractions of 3 or 4 small baddeleyites each. Most grains were small (<50 um long), thin blade pale brown in colour, and many appeared to have a

"frosted" appearance, interpreted to be a fine coating of zircon. Some very small (<30 um) skeletal crystals of what may have been zircon were observed in the final concentrate after mineral separation, but these were too small and rare to analyze. The three most discordant fractions show calculated Th/U ratios well in excess of those that are typical of baddeleyite, which normally discriminates strongly against Th relative to U. It is likely that the grains of 58 baddeleyite in this sample were coated with a thin layer of high-uranium, radiation-damaged zircon, which may have suffered low-temperature Pb loss. This indicates changing silica activity in the magma during crystallization, as geochemical fractionation increases the relative amount of silica with cooling.

The Thule sill (Site GF) was analyzed using three fractions of three or four baddeleyites each. The sample yielded several very small (30-80 um) thin, light brown blades with visible striations, and so multiple grains were analyzed in each fraction. Precise overlapping data have an average 206Pb/238U age of 712±2 Ma (Figure 3.2). The sill's geochemistry (Appendix B) and paleomagnetism (Table 3.2) clearly indicate the sill's genetic association with the E/W-trending dykes in the area, but its age is distinctly younger.

Clarence Head Swarm (N/S dykes)

The N/S-trending dykes of the Clarence Head swarm (Figure 3.3) appear to be significantly younger than the Franklin dykes. The dyke at Cape Faraday (CV) was dated with three analyses on fractions of 2 to 7 crystals each. The baddeleyite recovered from this sample were pale, very thin, very fresh blades between 20-50 um long. The data overlap within error, and yield an average 206Pb/238U age of 713±3 Ma (Figure 3.3a).

The Craig Harbour (CH) dyke yielded several large (>100 um) grains when use of modified mineral separation techniques were employed, after those used to isolate zircon had failed (see Methodology, above). The honey-brown striated blades and fragments were thin and flat, and were analyzed as single grains. A 206Pb/238U age of 713±2 Ma (Figure 3.3b) was obtained based on two overlapping analyses of single baddeleyite crystals. 59

Table 3.1

dmod 2O6pb/204pb 207pb/2206pb % Sample Fraction Wmod Th/U cPb Wu 2o "Pb/^U 2a "Tb/^U 2o 2c Rho disc (mg) 0«n) (pg) (measured) age age cone. 1. CG-7, Cadogan Glacier diabase, Ellesmere Island 1 1 fresh badd.ftag 0.0001 36.4 0.001 4.9 314 0.118385 0.000 1.02526 0.034 721.3 2.1 701.8 66.9 -2.9 0.837 2 1 fresh badd. frag. 0.0001 34.0 0.003 9.9 121 0.117032 0.001 1.01536 0.098 713.5 5.1 705.6 196.7 -1.2 0.953 3 2 large dk. irreg blade frags. 0.0001 5.9 0.005 5.3 86 0.116287 0.001 1.01982 0.146 709.2 7.5 728.5 296.3 2.8 0.967 4 2 dk bm badd. frags. 0.002 14.3 0.030 1.0 606 0.117100 0.000 1.02284 0.018 713.9 2.4 720.0 33.3 0.9 0.542 5 1 dk. brn. badd. fiag. 0.0002* 13.3* 0.005 0.2 2179 0.118128 0.000 1.04055 0.006 719.8 2.1 737.9 9.9 2.6 0.634 6 1 dk. bm. badd. frag. 0.0001 * 13.0* 0.007 0.2 1402 0.118934 0.001 1.04893 0.009 724.4 2.9 740.5 15.0 2.3 0.606 2. QA-2, Qaanaaq diabase, Greenland 1 1 large badd. frag. 0.0003 3.0 0.028 1.7 128 0.119016 0.001 1.12123 0.111 724.9 4.7 843.9 193.1 18.5 O.805 5.9 2 1 badd needle, bm. 0.0001 0.199 1.2 217 0.118201 0.001 1.02948 0.052 720.2 3.8 713.9 99.8 -0.9 0.721 3.7 3 2 badd. fragments, brn. 0.0001 0.006 3.2 45 0.116653 0.003 0.99875 0.374 711.3 19.4 677.4 854.8 -5.3 0.949 6.8 4 2 badd. fragments, brn. 0.0001 0.071 1.4 154 0.118206 0.001 1.02320 0.075 720.2 4.9 700.8 147.0 -2.9 0.787 3. BG-6, Belcher Glacier diabase, Devon Island 1 3 euh. frags, medbrn. 0.0002 n/a 0.556 0.7 449 0.106825 0.000 0.93861 0.023 654.3 2.2 732.5 46.9 11.2 0.618 2 3 badd. frags; mod. fresh 0.0002 n/a 1.153 2.9 90 0.084667 0.001 0.76765 0.101 523.9 5.4 798.7 267.1 35.8 0.960 3 4 badd frags. 0.0002 n/a 0.524 0.8 674 0.102523 0.000 0.89604 0.014 629.2 1.5 721.3 29.9 13.4 0.594 4 4 small dk brn badds. 0.0001 n/a 0.064 0.8 114 0.116064 0.002 1.04461 0.106 707.9 8.9 767.7 205.8 8.1 0.668 4. CV-1, Cape Faraday diabase, Ellesmere Island 1 2 badd. frags. 0.001 8.5 0.001 4.0 41 0.117480 0.004 1.04898 0.436 716.0 22.2 766.5 950.4 7.0 0.967 16.9 2 2 badd. frags. 0.001 0.005 1.4 209 0.116944 0.001 1.01791 0.052 712.9 3.4 712.6 103.0 -0.1 0.796 4.4 3 7 fresh, pale blade frags. 0.0001 0.002 3.1 93 0.116868 0.001 1.03334 0.135 712.5 7.3 745.8 266.6 4.7 0.922 5. CH-1, Craig Harbour diabase, Ellesmere Island 1 1 badd. cracked 0.0007 7.7 0.018 0.6 1428 0.117045 0.001 1.02586 0.009 713.5 3.3 727.3 16.1 2.0 0.488 4.4 2 1 square badd. frag. 0.0003 0.002 0.6 471 0.116925 0.001 1.02461 0.033 712.8 3.3 726.8 64.2 2.0 0.487 6. CL-2, Clarence Head diabase, Ellesmere Island 41.4 1 lbadd. 0.0001 0.014 1.8 907 0.117157 0.000 1.02170 0.012 714.2 1.5 716.6 22.0 0.4 0.559 2 lbadd. 0.0009 26.2 0.002 1.0 621 0.117575 0.000 1.02640 0.019 716.6 2.0 718.8 35.9 0.3 0.591 59.0 3 1 badd., triangular 0.0001 0.031 0.3 4024 0.117536 0.000 1.02466 0.004 716.4 1.5 715.9 6.2 -0.1 0.690 4 1 badd. block 0.0002* 18.3* 0.022 0.3 5500 0.117718 0.000 1.03054 0.003 717.4 1.4 724.8 4.3 1.1 0.781 5 1 badd. blocky 0.0001* 26.7* 0.007 0.3 1524 0.116978 0.000 1.01925 0.008 713.1 2.7 714.8 14.2 0.2 0.602 6 1 long badd. frag. 0.0001* 16.8* 0.003 0.2 3228 0.117579 0.000 1.03090 0.005 716.6 1.7 728.0 6.8 1.7 0.698 7. GF-1, Granville Fjord diabase sill, Greenland 1 3 fresh medbm badd frags 0.0002 n/a 0.129 1.3 275 0.116844 0.000 1.01598 0.040 712.4 2.8 710.4 77.3 -0.3 0.748 2 3 fresh badd frags, mod brn 0.0002 n/a 0.049 0.8 179 0.116834 0.001 1.04240 0.063 712.3 5.3 764.9 119.0 7.3 0.656 3 4 small badd. frags 0.0001 n/a 0.015 3.2 50 0.117748 0.003 1.09664 0.335 717.6 18.4 854.7 644.7 16.9 0.876 8. OF-2, Olrik Fjord diabase, Greenland n/a 1 1 badd. short prism 0.0006 0.006 4.8 2247 0.287647 0.001 3.96246 0.016 1629.8 3.3 1622.4 5.2 -0.5 0.736 2 1 badd, stubby-prismatic 0.0009 n/a 0.004 2.0 3453 0.286828 0.001 3.94930 0.016 1625.7 4.1 1621.5 4.4 -0.3 0.804 3 1 badd., long prismatic 0.0006 n/a 0.001 1.9 12609 0.285193 0.001 3.93790 0.036 1617.5 2.5 1626.8 16.5 0.6 0.239 8. DH-1, Dundas Harbour diabase, Devon Island n/a 1 1 badd. Frag, mod. Bm. 0.0001 0.023 4.0 576 0.230502 0.001 2.74771 0.036 1337.1 2.9 1348.4 22.8 0.9 0.693 2 3 fragments badd., It. bm. 0.0001 n/a 0.001 4.3 96 0.232247 0.002 2.78809 0.256 1346.3 12.9 1362.0 163.4 1.3 0.901 3 1 badd., thin long needle 0.0001 n/a 0.002 2.8 98 0.231890 0.003 2.79259 0.252 1344.4 14.3 1368.1 159.9 1.9 0.828 4 2 fragments badd, brn. 0.0001 n/a 0.082 2.5 467 0.230353 0.001 2.75586 0.045 1336.4 3.6 1355.4 27.9 1.6 0.680

Table 3.1: U-Pb isotopic data for baddeleyite from mafic dykes in Canada and Greenland. Note: Errors are given at 2a. Sample locations are given in Table 1. badd. = baddeleyite; brn. = brown; frag. = fragment. Disc. = % discordance for the given age. cPb is total measured common Pb assuming the isotopic composition of laboratory blank: 206/204 = 18.221; 207/204 = 15.612; 208/204 = 39.360 (errors of 2%). Rho cone. = error correlation coefficient. Th/U calculated from radiogenic ZU0Pbr°Pb ratio and zu'Pbr"DPb age assuming concordance. dmod = thickness based on images normal to [100] face and assuming 300ppm U. Wmod = weight based on images normal to [100] face and assuming 300ppm U. % disc. = Percent discordance. * denotes weights/thicknesses determined by actual volume measurements based on SEM imaging. data-point error ellipses are 2s

721 ± 2 Ma too 3 fractions

data-point error ellipses are 2s

721.4 ±4 Ma 4 fractions, 1-2 baddoloyitoc •acn MSWD = 1,3, probability • 27%

0.4 0.6 0.8 1.0 1.2 1.4 1.6

207pb/235u

data-point error ellipses are 2s

Figure 3.1: Geochronological results for E/W-trending Franklin dykes: a) the Cadogan Glacier dyke of * Ellesmere Island, a. b) the Qaanaaq dyke of Greenland, and c) the Belcher Glacier dyke of Devon

Intercepts at Island. 0 ± 5 & 726 ± 24 Ma MSWD = 0.22

207pb/233u 62

data-point error ellipses are 2s

820 y^ 0.135- GF 760xT 0.125 - 1 740/^ 3 3 a 8 0.115<: a. 660 •o 8 0.105- 620tT 712 ± 2 Ma 0.095 • 3 fractions, 3-4 baddeleyites each MSWD = 0.16, probability 85%

0.085 • , 1 > (—— i 1 , 1 , 1 , 1 0.7 0.8 0.9 1.0 1.1 1.2 1.3 207Pb/235U

Figure 3.2: Geochronological results for the Thule sill at Granville Fjord, Greenland. 63

The baddeleyites from the Clarence Head (CL) dyke, one of the two N/S-trending dykes

with an anomalous paleomagnetic direction, were successfully obtained using the mineral

separation techniques typically used for zircon. Several dark brown, large (-100 (am), blocky

grains were obtained, as were thin blade fragments. Six single-crystal fractions yield

overlapping data with a slightly older 206pb/238U age of 716±1 Ma (Figure 3.3c). These results

indicate that the Clarence Head swarm is younger than the E/W-trending dykes of the Franklin

swarm (Table 3.1).

Melville Bugt swarm

Geochronological results (Figure 3.4) indicate a moderate degree of discordance, though

the baddeleyites were relatively large (up to ~200 urn long), having been isolated using

"conventional" mineral separation techniques. The grains tended to be dark brown, long, striated

blades, though one with a more blocky morphology was analyzed. Three fractions of one

baddeleyite each yield an average Pb- Pb upper concordia intercept age (in which the lower

intercept is anchored at 0±5 Ma) of 1622±3 Ma. This 207Pb/206Pb age is slightly younger than

Hamilton et al. (2004)'s reported age of 1629±1 Ma.

Dundas Harbour dyke

As part of the survey of Melville Bugt dykes, the extension of that swarm was sought on

the Canadian side of Nares Strait. Dykes mapped with a WNW to N trend (Frisch 1984a,b,c)

were selectively sampled (specifically, dykes CM, HM, BE, NS, CL, DH, and CG), and one

dyke (DH) yielded a paleomagnetic direction not unlike that characteristic of the Melville Bugt 64

data-point error ellipses are 2s a) 713 ±3 Ma ~\ 3 Fraction!, 2-7 baddojoyitos In oach '780 MSWD = 0.04, probability = TO*

a 8

0.105

207Pb/235U

data-point error ellipses are 2s

b) ' f~ 713 ±2 Ma 1 790 / 2 Fractions, 1 baddoloyito oach | , MSWD = 0.08 probability = 77% J 770 /

7S0/

$ 0.122

^ 0.118 • ™<^±2^2

• 690,/' / 0.110-- CH

650 / H ' 1 >- 1—•— 0.95 1.05 207„. ,235 pb/"aU

data-point error ellipses are 2s c) 716 ± 1 Ma 6 Fractions, 1 baddoloyito oach , MSWD = 3.1, probability = 10%

Figure 3.3: Geochronological results for N/S-trending Clarence Head dykes of , Ellesmere Island: a) the Cape Faraday dyke, b) the Craig Harbour dyke, and c) the Clarence Head dyke.

0.94 0.90 0.9S 1.00 1.02 1.04 1.06 1.08 1.10 207Pb/23SU 65

data-point error ellipses are 2s / s 0.293 OF 0.291 / /l650

/ /l640 0.289 - 1. D CO n lA 1630 <^ 0.287 a. 1620 ^2 © 0.285 1610,// ^3 0.283 - 1600./ / Intercepts at 0.281 / o Ma & 1622 ± 3 Ma rr 3 fractions, 1 baddeleyite each / / MSWD = 0.20 , 0.279 / . A • 3.75 3.85 3.95 4.05 4.15 207Pb/~ni_ ,235DU,

Figure 3.4: Geochronological results for the Olrik Fjord dyke of Greenland, part of the Melville Bugt swarm. 66

data-point error ellipses are 2s

1337+2 Ma 4 fractions, 1 -3 baddeleyites each MSWD = 1.07, probability =36%

0.212 -I ' 1 ' 2.6 2.7 2.8 2.9 207-,rpbr". ,235u.

Figure 3.5: Geochronological results for the Dundas Harbour dyke of south Devon Island. dykes. However, it yielded an unexpected U-Pb age. As there is just one site representing a magmatic event of unknown volume or geographic extent, interpretations its relationship with other dykes in the area must be considered preliminary.

Baddeleyite grains extracted from this sample ranged in length from 20-90 urn, and were thin, brown, striated blade fragments. The Dundas Harbour dyke gives a Pb/ U age of

1337±2 Ma (Figure 3.5) based on four overlapping analyses of one to three baddeleyite crystals per fraction. This represents the first firm evidence for Mesoproterozoic magmatism in the High

Arctic, though Dawes (2006) reports similar Rb/Sr ages (approximately 1.5-1.3 Ma) for the

Nigarfivik dykes (Buchan & Ernst 2004) that may be age-equivalent, and Frisch (1988) reports a whole-rock K-Ar age of 1335±15 Ma for a N-trending dyke on the south coast of Devon Island, about 100 km west of Dundas Harbour.

Geochemistry and Petrography

Franklin dyke swarm

Geochemical data are chiefly used here for the purposes of correlation. Should the dykes on either side of Nares Strait share similar geochemical characteristics, it will provide further evidence that the dykes of Devon Island/Ellesmere Island and Greenland are of the same swarm. Petrographic observations (Figure 3.6) are of a uniform mineralogy in all of the specimens from east-west trending dykes. They tend to be plagioclase-pyroxene subophitic diabase, which contain 50%-60% lath-shaped plagioclase; 25% subhedral clinopyroxene, rarely as inverted pigeonite; 7%-8% orthopyroxene; 5%-10% subhedral biotite; and 7%-10% euhedral to subhedral opaques. Titanite is rare and olivine is absent. Myrmekite and micrographic 68

2.5 mm K- •W Figure 3.6: Photomicrographs of dykes from Canada (left column) and Greenland (right column) under cross-polarized light, illustrating their similarity. 69

1.5 mm M- •W

Figure 3.7: Photomicrographs under plane light showing pleochroic haloes in amphibole, biotite or chlorite. These haloes believed to be caused by radioactive minerals and therefore the samples containing them were the first candidates for U-Pb dating. 70 intergrowths occur commonly. Grain sizes of plagioclase and pyroxene from interior samples range from 0.5 to 1.5 mm.

Hydrous alteration occurs variably in most specimens, ranging from mild sericitization of plagioclase (e.g., NU2-7-1 in Figure 3.6) to extensive sericitization and near-complete conversion of biotite and/or pyroxene to chlorite (e.g., BG7-6). Calcite has been observed in one specimen (PHI-1-1). Most samples contain pleochroic haloes, visible in amphibole, biotite and sometimes chlorite (Figure 3.7). While the radioactive core is generally too small for petrographic identification, baddeleyite was surmised in the centre of larger haloes. The recognition of potential U-bearing minerals served as a guide to selecting which dykes were the best candidates for U-Pb age dating.

Specimens closest to dyke margins were selected for geochemical analysis to limit fractionation effects. For some dykes, multiple specimens were analyzed to determine if there was, in fact, a difference in geochemistry from the margin to the interior of a dyke. At the major-/minor-element level, there does not seem to be any difference (Appendix B).

Nevertheless, when multiple specimens were analyzed, the specimen from the dyke's margin was included in the comparisons and plots when looking at the dyke swarms as a whole.

Geochemically, the Thule and Devon Island dykes are similar. The four figures (Figures

3.8-3.11) illustrate different aspects of the dykes' comparative geochemistry. A Jensen diagram

(Jensen 1976) (Figure 3.8) is a common plot for mafic intrusive rocks, and the division between tholeiitic and calc-alkaline is predominantly a function of (Fe+Ti) content versus Mg. The TAS

(total alkalis/silica) diagram (Figure 3.10) is the IUGS-recommended plot illustrating alkalinity

(Na20+K.20) versus increasing degree of fractionation (increasing SiCh). However, the elements used in these plots are all mobile under low-temperature conditions, and may not be diagnostic of 71 primary igneous crystallization conditions (Pearce 1996). The Meschede plot (Meschede 1986) of incompatible elements uses high field strength elements (Nb, Zr, Y) that are immobile even under upper-greenschist-fades conditions of metamorphism. The Jensen diagram indicates that the Thule and Devon Island dykes are of moderately tholeiitic composition, though trending to the calc-alkaline field (Figure 3.8). A plot of incompatible trace elements shows similar characteristics of the two swarms as well (Figure 3.9). A more striking feature is the high TiC>2 content of the Thule and Devon Island swarms, typically about 4-5 wt% (Appendix B). This may have shifted data points toward the middle of the "tholeiitic" field on the Jensen plot when otherwise they would be classified as more alkaline (Figure 3.10). There is more scatter in the major-element plots (Figures 3.8, 3.10), but as these elements (Na, K, Fe, Mg) are considerably more mobile than incompatible elements (e.g., Pearce 1996), the high field strength elements

(Figure 3.9) are more diagnostic of tectonic setting, and are more reliable in terms of establishing a correlation between the dykes. Loss on ignition, which is often a measure of the degree of hydrous alteration, is generally low, between 1 and 2 wt%.

A selection of samples was sent to SGS Laboratories for rare-earth analysis using inductively-coupled plasma mass spectrometry (ICP-MS). The purpose of this analysis was to investigate the possibility of a correlation of rare-earth element (REE) pattern with age (Halls et al., in press). While the results (Figure 3.11) support the geochemical correlation of the two swarms, as well as the significant difference of the "outliers": GF, BG and KL, no connection between the position of a dyke's REE plot and its age is apparent. BG appears to have a more primitive composition; the Thule sill GF has a prominent negative Eu anomaly, characteristic of prior removal of plagioclase from the melt and prior fractionation; and KL has the most elevated

REE pattern, consistent with its apparent enrichment in incompatible elements seen elsewhere 72

(see Appendix B). These dykes therefore represent geochemical outliers that, should they be located on the opposite side of Nares Strait, might be used as piercing points to precisely constrain the degree of offset. However, in this study, no such complement was found.

The anomalously high K, low Sr, and lower Fe content of the Kap Leiper dyke (Figs. 3.8, 3.10,

3.11) may be indicative of more extensive alteration and/or crustal contamination in this intrusion, though less-mobile trace element ratios, considered less susceptible to alteration, are also anomalous suggesting a magmatic origin. The dyke appears to have a unique composition compared to all the other E-W trending dykes in Figure 2.1, but it is nevertheless part of the

Thule swarm as it has a similar trend and paleomagnetic direction (Table 3.2), and cuts

Proterozoic Thule Group sedimentary rocks but not overlying Cambrian strata. KL also has a slightly elevated REE pattern (Figure 3.11).

While the host rocks of the Thule and Devon Island swarms are different

(Mesoproterozoic sandstone in Greenland, Archean/Paleoproterozoic gneisses in Canada), crustal contamination does not seem to have affected the rocks. There is no evidence that chilled margins entrained material from the host rock based on observations at the outcrop and in thin section, and furthermore, the geochemistry of chilled-margin samples from either swarm are similar in their immobile element abundances (Appendix B, Figure 3.9), regardless of the type of host rock. 73

Fe+Ti A • Canadian Franklin • Greenlandic Franklin • Dundas Harbour •Melville Bugt

LG C

w 4 KL Thorites

Komatiites Calc-Alkaline

Figure 3.8: Jensen diagram (Jensen 1976) of all dyke swarms. Data from samples located within 30 cm of dyke chilled margins. Labelled are dykes with anomalous geochemistry (BG, KL) and paleomagnetism (LG, CG) that are the most likely candidates for tracing an individual dyke across Nares Strait. 74

2Nb A • Canadian Franklin +Greenlandic Franklin « Dundas Harbour • Melville Bbgt

WPT

Zr/4

Figure 3.9: Incompatible-element plot (Meschede 1986) of all dyke swarms. WPA = Within-Plate Alkaline; WPT = Within-Plate Tholeiite. 7 •KL 6.5 • 6

5.5 • • 5 • /•Canadian Franklin 4.5 • • / • Greenlandic Franklin • Dundas Harbour Alkaline 4 • J? • Melville Bugt 3.5 • 3 • BG 2.5 Tholeiitic

40 45 50 55 60

Si02

Figure 3.10: Alkalis vs. Si02 plot (division from Hald & Tegner 2000) of dykes sampled, showing samples from all dyke swarms. Normalized to: CI chondrite (McDonough & Sun 95)

1000

-a t. a •a c s 100 }

La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Yb Lu

Figure 3.11: Rare-earth element geochemistry for selected dykes. Red symbols denote E/W- trending dykes on Devon or Ellesmere Island; Blue denotes E/W-trending dykes from Greenland; Purple denotes N/S-trendingdykes on Devon or Ellesmere Island; Yellow denotes a Thule sill. 77

Clarence Head Swarm (N/S dykes)

The petrography of the N/S-trending Clarence Head dykes is almost identical to that of the E/W-trending Franklin dykes (Figure 3.12), though with more extensive alteration. Site CL in particular contains more chlorite and fine-grained opaque phase(s).

Geochemically, the Clarence Head dykes are indistinguishable from the Franklin dykes

(Figures 3.8, 3.9, 3.10, 3.11, Appendix B). Their incompatible-element ratios (and absolute abundances) are similar to those of the Franklin swarm.

Melville Bugt swarm

Geochemical results (Appendix B) are similar to those previously published in Nielsen

(1990), with the incompatible-element comparison (Figure 3.9) showing a trend distinct from the

Franklin and Dundas Harbour dykes while remaining within the "within-plate basalt" field.

Petrographically, the Melville Bugt dykes are also distinct from other swarms. Plagioclase laths typically exceed 2 cm in length, uncommonly exceeding 10 cm. Olivine is present (Figure 3.13), and alteration is minor.

Dundas Harbour dyke

Petrographically, the Dundas Harbour dyke is similar to the Franklin swarm, in that it is a subophitic diabase, with no olivine. Geochemically, however, it lacks the characteristically high

Ti abundances (Appendix B), though on an incompatible-element plot (Figure 3.9), it does not appear significantly different, owing to their similar tectonic setting 78

Figure 3.12: Photomicrographs (under cross-polarized light) of specimens from the Clarence Head swarm. Actual width of photos = 2.5 mm. Note the higher degree of alteration in dykes CL and "NS, those with the "anomalous" remanence direction. 79

Figure 3.13: Photomicrograph (under cross-polarized light) of a specimen from the Olrik Fjord dyke of the Melville Bugt swarm. Actual width of photo = 2.5 mm. Note the presence of olivine, absent from the other dykes. 80

("within-plate basalt"). Nevertheless, even at the major-element level, the different swarms of dykes can be distinguished geochemically

Paleo-/ Rock magnetism

Franklin dyke swarm

Paleomagnetic results (Table 3.2, Figures 3.14, 3.15) indicate a stable high coercivity

(Hc), high unblocking temperature (tut>) characteristic component of magnetization in the dykes that is revealed after removal of lower Hc and tUb components by AF and thermal demagnetization. The directions of this component for each site are given in Table 3.2 along with their corresponding virtual geomagnetic poles (VGPs). The more common paleomagnetic direction, with shallow inclination and westerly declination, is hereafter referred to as having an

"N magnetization" while the direction antipodal to this one is "R". There are very few sites with an R-magnetization, and while the scatter in these sites is great, they are considered antipodal if their direction is antipodal within the range of those with an N direction. For example, though the declinations for R sites CV and SG2 are 46° apart, the magnetization direction of SG2 is directly opposite that of EG, and the magnetization direction of CV is directly opposite that of

LG. Only sites with number of specimens Ns>5, and 0195, the radius of the 95% confidence oval about the mean, <15° were included in the calculation of the mean paleopole. Two E/W- trending dykes on the Canadian side, SG2 and OR, have R-polarity directions, indicating that the period of dyke injection spanned at least one reversal of the Earth's magnetic field. The mean paleopole from the Devon Island and Ellesmere Island dykes lies at 5.8°N, 188°E, N = 12, A95 = 81

9.9°. On the Greenland side, two more dykes have R-polarity directions (NU1 and HE), and the mean paleopole for the dykes from Greenland lies at 8.8°N, 178.7°E, N = 10, A95 = 7.2° (Table

3.3, Figure 3.16).

A baked-contact test (Everitt & Clegg 1962) was performed on dyke and host rock samples from site GR, where the host rock included anorthosite, an often-reliable retainer of stable remanence. In order to determine whether the magnetization measured in the dyke is a primary TRM, samples are taken from the dyke, from the host rock within a half dyke width away, representing the zone where the temperature increase caused by the dyke's injection

"baked" the host rock and caused it to reset its magnetization according to the geomagnetic field, and host rocks farther away, where the temperature effects of the dyke's emplacement were not felt. If the samples all have the same magnetization direction, this constitutes evidence that a regional remagnetization event has occurred since the dyke's emplacement and therefore the dyke's magnetization is not primary, and the test is negative. If the distal host rock has a different magnetization direction than the dyke, and the proximal host rock has been overprinted by the dyke's magnetization direction (and there is no evidence for the host rock's component in the dyke samples), the test is considered positive and the dyke's magnetization is considered primary. The test at site GR was positive (Figures 3.17, 3.18) in terms of the change in paleomagnetic direction, though no samples yielded a hybrid magnetization which could have helped determine which magnetization is older. This hybrid magnetization would have been expressed by the presence of two components, with one (the younger) affecting the low- coercivity/low unblocking temperature grains of magnetite, and the primary magnetization isolated in the high-coercivity/high unblocking temperature grains, expressed as a linear segment going through the origin. The paleopole for the samples of unbaked host rock at Grise Fiord 82

(Pole 11, Table 3.3) is unlike any other pole from the area (see Table 3.3), and while it is somewhat close to that of the dykes with a Paleozoic overprint (see Clarence Head dykes, below), there is no evidence of the host rock's characteristic remanence in the dyke rocks (see

Figure 3.17, top row), and the host rock's pole does not fall on the Paleozoic apparent polar wander path for North America (Torsvik et al. 1996, see Figure 4.9). Therefore as the unbaked host remanence direction is older than that characteristic of the dyke, and the presence of reversals of magnetic polarity in other dykes of the Devon Island and Thule swarm, indicate that the baked contact test is positive and the characteristic shallow-inclination, westerly-directed magnetization is primary.

The data from the LG and CG dykes plot substantially to the south when compared to the other data (Figure 3.15), and in order to assess their effect on the pole, the mean pole for the

Canadian data was recalculated without those two dykes included (Pole 3, Table 3.3). The poles

(Poles 2 and 3, Table 3.3) are offset in the same sense as the offset when CG and LG are included in the mean (though with a smaller magnitude), and there is no compelling reason to exclude the data, as the CG and LG dykes are certainly of the Franklin swarm (the CG dyke is dated at 721 Ma), and they are located within the same rigid continental block of Ellesmere

Island that, even if considered separate from Devon Island (see Introduction), also hosts the GR dyke, which has the westerly-directed remanence direction. This more southerly declination is also found in samples from the CH and CV dykes of the Clarence Head swarm (see below), and in samples from the margin of the KL dyke of the Thule swarm (see Appendix). The CG dyke has been dated to 721 Ma, and the CH and CV dykes have been dated to 713 Ma (Table 3.1).

Therefore, this direction may be interpreted to indicate the presence of a younger component.

However, there is no evidence of removing a southerly-directed component during 83 demagnetization (see Figure 3.14 and Figure 3.22 below), dykes geographically close to those with this direction (e.g., GR near CH, and PK near KL (Figure 2.1)) do not show any evidence for the presence of a later, southerly-declination overprint. Therefore, this direction is considered primary, a result of secular variation, and the pole that includes CG and LG (Pole 1,

Table 3.3), is considered the pole to use for tectonic reconstructions. For the mean pole calculations, the interior samples of dyke KL were used. Compared to the interior samples, samples taken closer to the dyke's margins (KL-10, KL-11) tended to have low susceptibilities

(Table 3.3) and low intensities of magnetization of the characteristic remanence component

(averages of 590 mA/m for interior samples, 315 mA/ra for margin samples). This indicates a greater likelihood of single-domain magnetite being the carrier of the magnetization of the interior samples, and therefore less of a chance of their acquiring a secondary overprint. Another possibility for this difference in remanence directions is that the direction found in the margin samples may be older, and the direction of the interior samples may have been acquired during either prolonged cooling of the dyke or during the emplacement of a new pulse of magma within the dyke. However, there is no major difference in grain size from margin to interior that would indicate significantly slower cooling rates, and no field observations were made during sample collection at the dyke of evidence for later injections of magma.

Secular variation must be averaged out for both poles, or they can only be considered virtual geomagnetic poles and, essentially, spot readings of the Earth's magnetic field at the time of dyke cooling. More than ten dykes over a broad geographic area contribute to each pole, and emplacement of Franklin intrusions covers a time span over 5-10 million years. The presence of reversals recorded also indicates that the emplacement of the dykes extended well beyond the period of secular variation. After removing the within-site dispersion (e.g., Irving, 1964), the (X95 m rh oo ^. oo £ in CN sd'^j-'in'sdosos'inin.-icnvo' in r~ H ^ rt

© so -^j- o cn CN c— TJ- so in ^> Ttr~CNcncNTr©cncnooosT}- in in o o\ o Osoo—icNincN-iOsTrinc-^cNcN^© r- cn 1-1 —iciHHtN^vi'-1 r~ r> CN in TT

00 <—i oo ^~ •^ >n oo so O ON SO OS rr 0 . . . £ -0 <* ""! cn cn —'^•oc;0d<^ioo* ! ^oI2't' CN cnOsOs^oooooOe-; -Hfsroii(S'-

00CNi-^r-;in—JCNr^CNOinrOCNCNCN cn f- o m c— oo oo Os cn odi-iso'i-iOTrcNOsOscn'incnincnin -rfso-—iO—' in oo oo oo in ^. oo ooooc^t^oooooorn-

& cn as —; ^oorn"*ooot-;i-;w-)ino SO SO©OsSOCNOOt--©©OsSD© sq >n i-H oq i-H oldf-riooifiHMOr-oi^pi^M t— OO —I O .-H r~- m CN cn t-- in i—> oo so oo CN so soooooo\f^t-~oocNi-

so cn ,,-, CN © in -i „ *0 ^ T m -: oo CN CI « * 5 ^t t- ^ • CN oo ^ d ^ CN 2 £ <* £ oo" so °) Os CN yA CN cn Os s i 0> '—' i—' I ii—' t~~ —I r-t , -1 ,-H ^H H i m —i -i q o ^ s °. t~- ^H r-l V O so cn in >/-> oq r-^ >n •* 11m • • 1-1 Tf in 10. 9 15. 7 17. 3 -* £J 53 S f- t^ O CN 16. 3 15. 9 11. 9 10. 3 26. 4 -18. 5 26. 1 37. 7 -35. 6 -1. 6 9. 3 28. 9 32. 4 cn -29. 6 i i—i CN cn i i H —i ^H cn T

o in rn r- so so 00 00 -t so 00 m r~ os CN

osmoot^ost— t~-t~P--t— -i^minvo ooosoot-~^fr--^ocNinsoooin cn vo so rr- cn

a oooinrfo^oinino in o o o ^r o o in so o cn in o o o in © © in cNcn^rc~-cncncNincNCN^ in in cn CN in cn TT CN 12 0 so •—i —i cn in CN >—* CN CN SO >nininoininmoooooininin ominoooino _, _. 08 2 ooini-ir-sot--insooocN^osocNOO c-~o\osoooos©©-5-5 00 © 35 5 &0 CNCNcnCNCNCNCNCNCNCNCNCNCNcncn cNcNCNCNcncNcncn n C 29 0 cn 35 0 CN 35 5

CN © CN r-~ rj; i—i CN © SO © i> SO CN OS © © 'J- © in Tf i-H OS SO ^t_ in P-; in oo OS OS Os © y-Ji ^H »-H i—i £ Os r^ ^H OS cn ^t *o oo •* CN SO ©' V9 ©' ^r Os CN in ©' SO cn in cn od od © ^f in F-H »—i .—i i—* i—i i—* ^H -^ in OS © cn o © CN © Tf in o O •* CN m m •t in cn o cn m © o CN a" o o o o o o o •0 * o o o o o 0 o 24 o o o 0 o o o 0 O o o o o o o o o 5 on CN OS cn O i-l T—< CN in t~- O CN © OS r—( ^r O Os r—i CN © o oo o cn Os T—1 © so t~- 00 oo T-H •J 3 OO r- OO oo oo OO OO r- r-~ OO oo OO r^ OO oo 6 r- SO r-~ f~ r~ t-~ so f- f- SO f" r~ r- t^ r^ l-~ 00 ^H ^-, ^ Tf OS © OS O CN •* sq ,-, © ^r cn ^r ^q © © in r-~ cn Os Os cn cn m cn Os t-; cn CN T—1 1—*

tkli n OS •* in OS cn ©' »n in so 00 OS cn t< 00 OS od © 00 CN t~ CN so cn CN •* 1 cn CN od rf ^wz ' r—t ^ 1—1 r—t »-H a © CN cn ^J- CN Tf TT cn ^f cn © CN •* CN CN in cn CN •t in CN CN in cn % o^ r o o o o o o o O o o 0 0 0 o Ma o" * o 0 o o o o o o o o o 1o ^0r o 0 0c n ^ d in m SO >n in in in oo 00 •* in •* m T in u 00 SO t-~ t~- r~- r~ t- t-- so SO r~ t- r~ SO SO SO in •J « r- c- f- i> r~ t- r- c^ p- p~ r~- t> r~ t> r- ta r- t- r^ r^ t- r- r~ r- f- r~ r~ t- "S f- f~ r- t- f~

di e = g * * * CN * * * * * E(R ) anadia r F (Sill ) larenc e A (Sill ) £ CQ a Pi o Pi Pi O O o G2 * (R ) m ^ Q ^ < U1(R ) > hJ a t/3 U m PM o oq CQ CQ W J U u W< S PH u 00 Greenla n ^ H Z PH fc4PH a U U 35 V U o Z o m 85

Dundas Harbour dyke DH 73°31.1 82°23.4 270 29 8 26.2 -3.9 -12.1 250.8 23 13.4 MelvilH e Bugt dykes JP2 77°35.6 66°52.9 325 65 7 209.1 44.2 14.7 266.5 40 3.0 QT 77°28.7 66°25.5 310 100 8 195.9 28.8 3.1 278.1 41 8.8 OF 77° 10.0 66°20.0 330 200 7 13.1 -6.7 -9.2 280.1 25 10.5

Table 3.2: Summary of paleomagnetic results. W = intrusion width (m); N = number of samples; Str = strike direction; D = mean declination; I = mean inclination; k = Fisher precision parameter; a.95 = radius of 95% confidence circle about the mean. (R) denotes reversed-polarity dyke. * denotes samples oriented using only a magnetic compass. D' and V denote mean declination and inclination recalculated to a common site (75°N, 80°W). Dyke names are explained in full in the Appendix. 86

Pole Unit N P,a, (°) P.on (°E) A95(°) dp(°) dm(°) 1 All Canadian Franklin 12 5.8N 188.0 9.9 - 2 All Greenlandic Franklin 10 8.8N 178.7 7.2 - - 3 Pole 1 without CG, LG 10 6.9N 182.4 7.9 - - 4 Pole 2 without sills 7.9N 179.5 9.1 - - 5 JP2 14.7N 266.6 - 2.4 3.8 6 QT 3.IN 278.1 - 5.3 9.7 7 JP2 rotated 6.1N 267.7 - 2.4 3.8 8 QT rotated 1.4N 265.7 - 5.3 9.7 9 CL remagnetized 8.0S 296.5 - 4.0 7.9 10 Dundas Harbour 13.0S 235.3 - 6.7 13.4 11 Grise Fiord host rock 8 1.6N 277.1 8.4 - -

Table 3.3: Paleopoles from various regions/dyke swarms, dp, dm = major and minor (respectively) semi-axes of the 95% confidence error about the VGP of a dyke. JP2 rotated, QT rotated = VGPs of dykes JP2 and QT after rotation according to the Euler pole of Roest & Srivastava (1989) to correct for the drift of Greenland relative to North America. Dyke locations given in Figure 2.1. 87

a

3 20 0 r 1 Z •a>+ in • A)jSU3|U| % E » o CO1 >«0 C 1 1 0 a? o 3 i- 1 1 O E ( «o in IN T K' in (in •n 1 » u in V 0

CO m u• 0 s 40 0 20

Figure 3.14: Examples of stable end-points obtained from Franklin dykes on Devon and Ellesmere Islands. On equal-area stereoplots solid/open symbols are downward/upward- pointing magnetizations. On vector diagrams solid circles are projections of the tip of the magnetization vector on the horizontal plane, and open circles are projections on the vertical plane. Thermal data include an additional intensity decay plot. 88

Canada: D=274.3°, 1=11.3°, N=11, a95=13.8° Greenland: D=283.5°, 1 = 16.1°, N=9, a95=9.1

(Declination and inclination recalculated to a common site (75°N, 80°W)

Figure 3.15: Equal-area stereoplot of paleomagnetic data for the sites from Canada (red) and Greenland (blue), with declinations and inclinations recalculated to a common site. Closed circles = downward (positive) inclination; Open circles = upward (negative) inclination. + = mean of directions for each set of dykes. Directions are recalculated to a common geographic location (75 N, 80 W) to remove the effect of the difference in geographical location of the dykes. Therefore, based on the assumption of no relative motion, there is a difference (at the 76% confidence level) in their paleomagnetic directions. 89

.8°N,PkL=188.0°E

Figure 3.16: Mean paleopole locations for the sites from Canada (red) and Greenland (blue). Ovals are A95 (cone of 95% confidence about the mean). 90

8^ d d E E S

(%) A«SU9»U| (%) Xjisustui (%) Xtjsuatui

s u J! X "O "D O

0) .52 O in • O

Figure 3.17: Results of a paleomagnetic baked-contact test at site GR. Note similar magnetization directions for the dyke and proximal host rock using either AF or thermal demagnetization techniques, and the distinctly different direction of the magnetization in the more distal unbaked host rock. SCM/NCM is the southern/northern chilled margin. Other symbols as in Figure 3.14. 91

^ ^ "X ^ -L

Figure 3.18: Map showing location of samples and results for a baked-contact test at site GR. Circular symbols denote samples collected as drill cores; squares denote samples taken as oriented blocks. Red-filled symbols indicate samples with the dyke's characteristic remanence direction; blue denotes samples with the host rock's direction. White symbols denote samples that of too low intensity to yield a meaningful direction ("inconclusive"). 92

1.2 Mean K(0) = 51833.4 xlO6

0.8

0.6 Canada Dykes

0.4 BB6-4 ^-GR1-2 0.2 -•- SG2-5-3 o 300 400 500 600 700 800 900

1.4

300 400 500 600 700 800 900 Temperature [K]

Figure 3.19: Example plots of susceptibility (K) at temperature (T, degrees Kelvin), normalized to room temperature susceptibility (K0), for representative dykes from the Thule and Devon Island swarms. Note the uniform drop in susceptibility near 850K (580 C), indicative of low-Ti magnetite as the carrier of the magnetization. 93 for the Devon Island swarm is 7.6°, and that of the Thule swarm is 6.9°, so the contribution of within-site dispersion is not great. The 063, the measure of dispersion of one standard deviation about the mean, is 13.9° for the Devon Island swarm and 11.2° for the Thule swarm. As the angular dispersion of the site mean directions consistent with that of equatorial VGP dispersion is approximately 12° (Merrill & McElhinny 1983), secular variation is considered to have been averaged out for this study.

Figure 3.14 shows typical responses of specimens to AF and thermal demagnetization procedures. Most indicate a single component of magnetization, one that is stable and consistent until the specimen is completely demagnetized. Rarely found is a component with a steep down direction similar to the present Earth's field (PEF) and interpreted to be a viscous remanence component, as it is cleaned from the specimen after only a few low-intensity AF demagnetization steps. Temperature vs. susceptibility plots (Figure 3.19) show a sharp drop at about 580°C for all samples from E/W-trending dykes, indicating Ti-poor magnetite as the carrier of the characteristic magnetization in all dykes. Other magnetic phases (e.g., pyrrhotite and hematite, with Curie temperatures of ca. 300°C and 650°C respectively) that could add a chemical remanent magnetization appear to be absent.

An aeromagnetic survey of the Kane Basin (Oakey & Damaske 2006) includes two lines that cross the Kap Leiper dyke (site KL in Figure 2.1) roughly normal to its trend, data from two of which are shown in Figure 3.20. For the line closest to the sampling site (line 2) the anomaly was removed from the regional background by extrapolation of the background field by eye, and the resultant 2D anomaly fitted using Magix computer software. The susceptibility data and various parameters used in the model fitting are given in Appendix B. The results (Table 3.4,

Figure 3.21) show that a reasonable fit in amplitude between measured and calculated anomalies 94

(J.U) X|ISU3|U| p|9IJ 3II3UBDW

Figure 3.20: Two aeromagnetic profiles showing total field anomalies (arrowed) caused by the Kap Leiper dyke (courtesy G. Oakey and D. Damaske). Line 1 is 2 km to the west of Line 2 and is approximately 10 km west of site KL. 95

N

-•- Observed -*- Induced Only & Remanent + Induced

^^•^w \. "^,**,"^nnr»,

1 1 1000 2000 3000 4000 Distance (metres)

Figure 3.21: Observed and calculated total magnetic field anomalies over the Kap Leiper dyke. Modelled anomaly is Line 2 (see Figure 3.20). 96

KAP LEIPER DYKE AND AEROMAGNETIC SPECIFICATIONS

Dyke GPS Location: 78° 41' N; 70 ° 40' W; Dyke thickness: 54 m; Dip: 90°; Trend: 105 ° from longitudinal joints and magnetic anomaly; Field parameters using NOAA IGRF2000: Dec = -65° 43'; Inc = + 86°, F = 56324nT. Aeromagnetic flight line trend: 000°; Flight line elevation: 610 m

NRM AND SUSCEPTIBILITY (K) DATA: SAMPLE DEC[°] INC[°] J [xlO6 emu/cc] MEAN K [SI xlO"6]

KL1 307 -32 1049 48397 2 273 -46 966 57421 3 290 -52 1088 50308 4 278 -43 1141 63468 5 277 -14 1124 37111 6 301 -22 757 41485 7 275 -45 945 41025 8 304 -54 651 45293 9 279 -40 683 52781 10* 262 +45 1090 50616 11* 219 +7 314 36588

* Samples lie within 2 m of the southern chilled margin. They are either anomalous in direction or intensity compared to more interior samples and have been omitted from mean calculations of Susceptibility K and NRM direction and intensity J. Other samples are from a traverse across the dyke and are more than 2 m from any margin.

Remanent Magnetization Mean NRM direction: Dec= 287 °, Inc = -40 °, k= 23, (X95 = 11°, 063 = 12 °, N = 9 6 Mean NRM Intensity JR = 934 ±180 (st.dev.) x 10" emu/cc - 0.934 A/m along a direction Dec = 287°, Inc = - 40°. Conversion is 10 3 A/m = 10"6 emu/cc; AF demagnetized remanence direction is: Dec =283°, Inc =11°, k =74, (X95 =7° k= precision parameter, (X95 is the radius of the 95% confidence circle about the mean, 0 63 is the circular standard deviation, and N is the number of samples

Induced Magnetization Mean K = 48588 ± 7960 (st. dev.) x 10"6 SI; Field Strength Fi = 56324 x 10"9 T = 44.82 A/m. Conversion is 10"4 T = 103/4B A/m; Induced magnetization intensity Ji = KF 1 = 2.178 A/m, along a direction Dec = 294°, Inc = 86 °. Resultant Magnetization [Sum of Remanent and Induced]: Dec =288°, Inc = 61°, J =1.8 A/m

Table 3.4: Parameters used in magnetic modelling of Kap Leiper dyke. is obtained when remanent magnetization is ignored. When it is included, the theoretical anomaly becomes smaller but its shape becomes more like the observed one. The reason for the amplitude discrepancy is uncertain. The difference in the anomalies on lines 1 and 2 shows that the physical properties (such as width, bulk susceptibility, intensity of remanence, depth below surface, tilt) of the dyke vary along strike and therefore susceptibility measurements of a relatively small outcrop are unlikely to be representative of the larger volume of dyke that generates the anomaly. The important conclusion from this study is that the Kap Leiper dyke produces a significant aeromagnetic anomaly that is traceable across Smith Sound (Oakey &

Damaske 2004, 2006) and which therefore provides the most direct test for the existence of the

Wegener fault.

Clarence Head Swarm (N/S dykes)

Three dykes with a N/S trend produced identical paleomagnetic results to those with an E/W trend, one (CH) with an N direction, and two (BE, CV) with an R direction (Figure

3.22). Though their directions are about 50° apart in declination, they fall within the spread of antipodal declinations of the "N" set. Two of the Clarence Head dykes (CL and NS) have an anomalous direction compared with the "Franklin" direction of the other three (Figure 3.23).

Both of these characteristic components have typically high Hc, but CL and NS have distinctly lower tub, (ca. 320°C), though the specimens often maintain a high intensity of magnetization after AF cleaning to 1000 Oe.

Temperature vs. susceptibility plots (Figure 3.24) show a sharp drop at about 580°C for the sites with the "Franklin" paleomagnetic direction, indicating Ti-poor magnetite as the carrier of the characteristic magnetization in these dykes, but notable in Figure 3.24b is the presence of a 98 CH5-2, Craig Harbour dyke CV5-1, Cape Faraday dyke W, Up W, Up

100 50 10-30 mT

12.5 7.5-70 mT 25

100

Figure 3.22: Examples of stable paleomagnetic end-points from dykes from the Clarence Head swarm. Symbols as in Figure 3.14. 99

Atjniaiui %

o a

a> 00 1 3 z 5s 0) w C a> u O 20 0 .7 1 U .5-8 0 m T 20

Figure 3.23: Examples of stable paleomagnetic end-points from the Clarence Head dyke (CL), showing the "anomalous" paleomagnetic direction and the low unblocking temperature, high-coercivity component. Symbols as in Figure 3.14. a) 1.4 Mean K(0) = 42124.5 x 106

e* 0.8 • CH5-4 ^ 0.6 • NS2-3 0.4

0.2

300 400 500 600 700 800 900 Temperature [K] b)

Mean K(0) = 3628.9 x\06

300 400 500 600 700 800 900 Temperature [K]

Figure 3.24: Example plots of susceptibility (K) at temperature (T, in degrees Kelvin), normalized to room temperature susceptibility (KO), for a) representative dykes from the Clarence Head swarm that have the "typical Franklin" remanence direction, and b) site CL, a dyke in the Clarence Head swarm with an anomalous remanence direction. Note the lower bulk susceptibility compared to dykes with the Franklin remanence in b), and the asymmetric peak at ~300°C. 0) +- c 0) £ '5 0) I 0) S2 o m o oN in

E cs «i • in v> Z

Figure 3.25: Examples of stable paleomagnetic end-points from the Norton Shaw dyke (NS), showing a) the "Franklin" paleomagnetic direction and the unblocking temperature indicative of magnetite being the remanence carrier in samples from the fine-grained chilled margin, and b) the "anomalous", high-coercivity component in specimens from the coarser-grained interior. Symbols as in Figure 3.14 102 pronounced asymmetric peak in susceptibility for samples from site CL (with the "anomalous" remanence direction) in the heating run before a sharp decrease at 320°C. The bulk susceptibility (Ko) is also an order of magnitude lower in site CL than in those with magnetite as the principal remanence carrier. A small drop in susceptibility at 580°C, indicates the presence of magnetite, but the low bulk susceptibility value suggests that the amount of magnetite is small. All samples from dyke CL contain the remanence with the southerly declination, but none from dykes CH, BE, or CV carry it. The NS dyke shows a variation in magnetization direction from the margins of the dyke to the interior (Figure 3.25). The fine-grained specimens from the margin (Figure 3.25a) have a remanence direction characteristic of the Franklin "R" direction, with unblocking temperatures characteristic of magnetite (i.e., ca. 580°C), while coarser-grained specimens from the interior (Figure 3.25b) yield the more southerly direction, though thermal demagnetization did not reveal the presence of significant amounts of pyrrhotite in samples with the anomalous direction (NS4-2t). The significance of the difference in remanence directions between dykes CL and NS and the other Franklin dykes is reserved for the Discussion.

Melville Bugt dyke swarm

Three dykes of the Melville Bugt swarm were sampled from the swarm's northern limit, at the eastern end of Inglefield Bay (JP2, QT, and OF on Figure 2.1). Their remanence directions, characterized by components with high Hc and tub (Table 3.2, Figures 3.26, 3.27) record a reversal, with data from sites JP2 and QT having an intermediate-down inclination and a

SSW declination, and site OF having a shallow-up inclination and a NNE declination. As the scatter is much greater for site OF, the SSW-directed magnetization is considered the N direction, with site OF's NNE-directed magnetization considered "reversed". With only three 103

eU •o • oj,

o se a4> o E

(°r/r) Xijsusiui

o in W/U p

a 3 o o phin e Pear y Is . dyk 0) w u O o CO 10 u• o o o a. w

.

1 o in ^. © o mm a0.) 1 o in a> o OS c f™ o X a * a> Figure 3.26: Examples of stable V) paleomagnetic end-points obtained from O «* Melville Bugt dykes. N Symbols as in Figure 3.14. •A I Q. OF: D = 8.3°, I = -6.9°, N = 7, k = 24, a95 = 11.6° JP2: D = 209.1 °, I = 44.2°, N = 7, k = 401, a95 = 3.0 QT: D = 196.1°, I = 30.4°, N = 8, k = 44, a95 = 8.4°

Figure 3.27: Equal-area stereoplot of paleomagnetic data from the Melville Bugt dykes. Symbols as in Figure 3.15. Locations of the samples are described in Figure 2.1. 1.4

1.2 K(0)=23146.8x10

Figure 3.28: Plot of susceptibility (K) at temperature (T), normalized to room temperature susceptibility (KO) vs. T in degrees Kelvin, for a sample from the Olrik Fjord dyke of the Melville Bugt swarm. Note the drop in susceptibility at 850K (580 C), indicative of low-Ti magnetite as the carrier of the magnetization. 106

dykes measured, secular variation cannot be considered to have been averaged out, but the

presence of differently-magnetized younger dykes in the area precludes a large-scale

remagnetization event after the Neoproterozoic. Temperature versus susceptibility analyses

(Figure 3.28) indicate that Ti-poor magnetite is the remanence carrier for the Melville Bugt

dykes.

While most samples from the OF dyke yield stable endpoints (see Appendix A), two

samples, OF4 and OF5, never reach a stable endpoint during cleaning by either AF or thermal

demagnetization methods. While their remanence directions (see results for OF5, Figure 3.26)

trace a great-circle path, the results of a great-circle analysis (Halls 1976) on two samples would

be inconclusive. As the remanence directions from specimens from both OF4 and OF5 follow a

path from a steep-down direction to an intermediate-up direction, it is likely that a viscous

present-Earth's field component is being removed (see vector diagram, Figure 3.26).

The VGP for site JP2 is 8.9°N, 279.9°E, dp = 2.36°, dm = 3.76°, and when its pole (Pole

5, Table 3.3) is rotated about the Euler pole from Roest & Srivastava (1989) to correct for the

drift of Greenland, that pole (Pole 7, Table 3.3) is the one used here for tectonic reconstructions.

This paleomagnetic result is different from that obtained by Hamilton et al. (2004) for a Melville

Bugt dyke approximately 1200 km to the southeast near Disko Island (dated at 1629 ± 1 Ma, U-

Pb baddeleyite age), where they obtained a VGP of 38°N, 115°E, dp = 9°, dm = 15°.

Dundas Harbour dyke

A stable remanence direction is obtained for the Dundas Harbour dyke (Table 3.2, Figure

3.29), and the virtual geomagnetic pole (VGP) obtained lies at 12.1°S, 250.8°E, dp = 6.7°, dm =

13.4°. Since the dyke is parallel to and nearby the Franklin dyke at Cape Warrender (CW on 107

1 >> *V "D >• D O _Q o z w O •D C D Q

• IN X o

I

Figure 3.29: Examples of stable paleomagnetic end-points obtained from the Dundas Harbour dyke. Symbols as in Figure 3.14. 1.4 i2 K(0) = 23991.5x10

1 m

— 0.8 *o £ 0.6 • DH14-2

0.4

0.2

300 400 500 600 700 800 900 T[K]

Figure 3.30: Plot of susceptibility (K) at temperature (T), normalized to room temperature susceptibility (KO) vs. T in degrees Kelvin, for a sample from the Dundas Harbour dyke. Note the drop in susceptibility near 850K (580 C), indicative of low-Ti magnetite as the carrier of the magnetization. 109

Figure 2.1), and their paleomagnetic directions are different, a regional remagnetization event since the emplacement of the CW dyke can be ruled out. The Dundas Harbour dyke's magnetization cannot conclusively be shown to be primary, however, as a baked contact test was not possible due to the extremely weak magnetization of the granitic host rock. The dyke could have been remagnetized between its emplacement and that of the Franklin dykes. No evidence for chemical resetting of the magnetization exists, as the plot of magnetic susceptibility versus temperature (Figure 3.30) reveals magnetite to be the carrier of the characteristic magnetization. 110

4. Discussion

Nares Strait Problem

The paleomagnetic, geochemical, petrographic and geochronological results show little difference between the Thule and Devon Island dyke swarms, and therefore provide permissive evidence that they represent parts of a single swarm at least 600 km long. However, the 95% confidence ovals for the Devon Island and Thule pole positions are too large to resolve the question of whether the Canadian and Greenland dykes have been laterally offset (Figure

3.16). The test for statistical significance of the difference (from Butler 1992) is as follows: If the Devon Island dataset is made up of Ni directions which have a resultant vector Ri, and the

Thule dataset has N2 directions with resultant R2, the F statistic is given by

F = (N - 2)(Ri + R2 - R) / (N - Ri - R2) where N = Nj + N2 and R is the resultant of all N individual directions. The Devon Island dataset has Nj = 10, R, = 9.21, and the Thule data set has N2 = 8 and R2 = 7.76. For the combined dataset: N = 18, R = 16.88. This results in an F-statistic of 1.398. In order for the two datasets to be considered different at the 95% confidence level, an F-statistic of 3.29 would be required. With the current datasets, they can only be considered distinct at the 76% confidence level. The mean poles are offset with the direction and magnitude that would result from a ~300 km sinistral movement of Greenland relative to Canada (Figure 4.1, 4.2), a movement actually greater than allowed by the amount of seafloor spreading in Baffin Bay. While the uncertainty in the latitude resulting from the A95 value of 7.2° in the Greenland paleopole (indicated by the orange bar in Figure 4.2) combined with the longitudinal uncertainty in the reconstruction

(indicated by the direction of the orange arrow in Figure 4.2) do permit a reconstruction with zero offset, if there was no actual difference between the poles, they could be offset, within Ill

Figure 4.1: Pole trajectory showing the restoration of the Thule swarm (Greenland) paleopole atop the Devon Island swarm (North America) pole, assuming the offset between the poles results from the motion of Greenland as a plate relative to North America. 112

Figure 4.2: Reconstruction of North America and Greenland at ca. 720 Ma. North America is restored to the position indicated by the paleopole of the Canadian Franklin dykes. Greenland is restored by moving the pole from the Greenlandic Franklin dykes to correspond with that from Canada. Orange arrow denotes direction of possible movement of Greenland in this reconstruction given the longitudinal uncertainty of reconstructions based on paleopoles. Orange bars denote the latitudinal uncertainty given by the A95 limit of confidence about the mean of the Greenland paleopole. Red lines are a schematic representation of the Franklin dykes in the area. 113 error, in any direction. Instead, the poles are offset (with 76% confidence) in the sense required by plate tectonic reconstructions that place the Wegener fault under Nares Strait. This difference in the poles is the result of the difference in mean declination seen in the paleomagnetic datasets of the Thule and Devon Island dyke swarms (Figure 3.15). Various E/W-striking faults cross the northern margin of Devon Island (Frisch 1984a), which may have enabled north or south tilting of the basement rocks and the dykes; furthermore, Devon and Ellesmere Islands may have been tilted ca. 3° to the west as a result of rift flank uplift during Cretaceous-Paleogene time (Grist &

Zentilli 2005). However, the Franklinian (Paleozoic) carbonate rocks that overlie the basement are flat-lying and undeformed (Frisch 1988), and the Proterozoic dykes are near-vertical, with dips rarely exceeding 85°. Given the westerly declination of the paleomagnetic results (see

Figure 3.15), a tilt to the west would only affect the inclination (and therefore the paleolatitude), when the principal difference between the Thule and Devon Island datasets is in their declination.

Correlation ofNeoproterozoic mafic dykes across Nares Strait

An important issue to resolve is whether the apparent difference in dyke density on opposing sides of Nares Strait is real, or a result of dykes on Ellesmere Island being covered by ice. The amount of outcrop on Ellesmere Island is poor, perhaps as low as 20% (see Frisch

1984b,c), and the apparent absence of Thule dykes on Ellesmere may be a result of simply not seeing dykes that are covered by ice. However, when outcrop from within approximately 50 km of either side of a line of longitude is projected on to that line (Figure 4.3), it can be seen that an

E/W-trending dyke should outcrop somewhere along a virtually continuous north-south line were 114 that dyke present on Devon Island, Ellesmere Island, or Greenland. Therefore the discontinuity of the E/W-trending dykes of the High Arctic as seen on geological maps of the area (Frisch

1984a,b,c; Dawes 1991; Dawes & Garde 2004) appears to be real. If the dyke swarms have not been offset across Nares Strait, then the pattern of dyke emplacement must be primary. This would require that the dykes of the Devon Island swarm come to an end off the east coast of

Devon Island (as no dykes were observed to terminate on land in either direction along strike)

(see also Frisch 1984a,b,c) and that the dykes of the Thule swarm be emplaced vertically from a source under the waters of Nares Strait (as no dykes were observed to terminate on land) (see also Dawes 1991, 2004). The host rock's structural grain (as an indicator of maximum principal stress direction at the exposed crustal level) may seem to control dyke emplacement, as the strike of the gneissosity on Devon Island and in the Thule region of Greenland is roughly E/W and subparallel to the trend of the majority of the dykes in the area, while on southeast Ellesmere and

Coburg Islands, which host the majority of the N/S-trending Clarence Head dykes, the host rock gneissosity is more N/S as well. While the structural anisotropy of host rocks (including faults) has been shown to have an influence on dyke propagation direction (e.g., Jourdan et al 2006,

Baragar et al. 1996, Ernst et al. 1995), these effects tend to take the form of causing deflections in the trends of laterally propagating dykes rather than large-scale discontinuities. The dip of the gneissosity of Devon Island is near-vertical, as is that of the dykes, but in the Thule area of

Greenland the dip of the gnessic layering of the host rock rarely exceeds 45° in either direction and are cut at a high angle by the dykes (Dawes 1991). In the case of magma being emplaced vertically from a laterally-propagating dyke at depth and encountering a different stress regime, en-echelon dyke segments are expected to occur, and none were observed in the field nor were any such occurrences mapped (Frisch 1984a,b,c; Dawes 1991; Dawes 2004). In the case of en- 115 GREENLAND

81'00'N, 68WW 50km I

Humboldt Otacier

CANADA

7ri5'N,78WW 79t)2'N, BBWW

i%

78,00'N, 88'00'W

NU1 1 77*15'N. 78'00'W 7ri0'N, 68*00'W

Southern tip of Ellesmere 76'go'N, eroow

Figure 4.3: Diagram indicating amount of outcrop in Arctic Canada and Greenland when projected on to a line of longitude from 50 km 1 75-30'N, 68"00'W on either side. Inset map: Red lines mark the 75°15'N, SZWW longitudes upon which outcrop in adjacent areas (approx. 50 km to each side, areas covered are seen in pale red) is projected. Large figure: Green indicates where outcrop projects on the line; Black indicates where 74"25'N, BCTO'W no outcrop projects on the line; Red indicates Lancaster Sound where a dyke outcrop projects on the line. Sampled dykes are labelled in red. 116 echelon emplacement, the greatest variation in trend between the strike of the dyke at depth versus that of the deflected segment above is 14° at Ship Rock, New Mexico (Rickwood 1990), which is insufficient to account for the apparent offset in the E/W-trending dykes. Furthermore, both E/W-trending and N/S-trending dykes are found on Devon Island (E/W gneissosity),

Coburg Island (N/S gneissosity) and Ellesmere Island (N/S gneissosity). As throughout the survey area are found dykes that both nearly parallel the host rock's structural grain and cut it at a high angle, it is concluded that host rock anisotropy does not seem to control the emplacement direction of the dykes.

The identical ages obtained from dykes CG and QA and similar geochemical results for almost every Franklin dyke in this study regardless of location provide clear evidence that the

Greenlandic and Canadian dykes are parts of the same swarm. In previous cases where different dykes of the same magnetic polarity are dated within a single swarm, U-Pb ages generally differ by no more than ca. 5 Ma. Examples include the giant 1.27 Ga Mackenzie swarm (LeCheminant

& Heaman 1989) and the 2.17 Ga Biscotasing swarm (Buchan et al. 1993, Halls & Davis 2004).

Physical and chemical variations across the width of each swarm that can be correlated between

Canada and Greenland, or a single dyke or group of dykes with a unique geochemical signature, have not been found. Only the northern half of the Thule swarm has been sampled, so the data now available do not permit a rigorous geochemical comparison, although extensive sampling of the Devon Island swarm does not indicate any regional chemical variation across the breadth of the swarm. Longitudinal changes for hundreds of kilometres along a dyke have been shown to be negligible (e.g., Kalsbeek & Taylor 1986), and therefore a particular dyke or group of dykes of unusual or unique chemical composition (such as the Kap Leiper dyke, KL, of Greenland or 117 the Belcher Glacier dyke, BG, of Canada) may be traceable from Greenland into Canada. This possibility is discussed further below.

Geochemical analyses of Franklin dykes from Victoria Island and the northern mainland

Northwest Territories (Jefferson et al. 1994) yield TiCh values typically less than 2%. The higher

TiC>2 values in the Thule and Devon Island dykes are an indication of their more fractionated nature and perhaps a result of the greater distance from the plume head that was the source of the

Franklin swarm (see also Heaman et al. 1992). Alternatively, the high Ti02 content may reflect the geochemistry of the local underlying mantle, as unusually high TiC>2 is also a feature of

Tertiary basalts from both western and eastern Greenland (Hald & Tegner 2000).

Another goal of the study was to determine if a single dyke with a distinctive paleomagnetic direction could be identified on both sides of the proposed fault. As the magnetization of sampled from the margin of the KL dyke of the Thule swarm has a markedly more southerly declination (~260°) compared to the other dykes from that swarm, it was considered that this dyke might be such a distinctive marker. However, that southerly declination is seen in dykes of the Devon Island swarm across a significant latitudinal extent, in dykes LG, CG, and HM. If the KL dyke is matched with any of these, even though a match is precluded by differences in geochemistry (Appendix B, see Figure 4.5), the possibilities for reconstruction range from 15 km of relative offset (match with LG) to over 250 km (match with

HM). Therefore, it is not possible to use an individual dyke's paleomagnetism as a precise piercing point. 118

Does the Kap Leiper dyke extend into Canada?

The 50m-wide dyke at Kap Leiper (KL on Figure 2.1) was sampled as part of the Nares

Strait Geocruise in 2001, and its aeromagnetic anomaly modelled (Figures 3.20, 3.21) in the following year. At that time it became clear that the dyke held great promise to resolve the

Nares Strait debate if it could be followed aeromagnetically for the -100 km across Davis Strait into Ellesmere island. With this in mind a follow-up survey was carried out (Oakey & Damaske

2004, 2006), specifically targeting dykes of the Thule swarm. The results (Figure 4.4) suggested that the dyke did indeed cross the channel and it was concluded that the Nares Strait controversy was effectively over (Oakey & Damaske 2004). However, on closer inspection of the aeromagnetic data (Figure 4.4) the ca. 150 nT anomaly that can be truly ascribed to the dyke ends about 3 km off the Ellesmere shoreline (at the X in the top image, Figure 4.4, just east of the "Buchanan Bay East boundary" (Oakey 2006), BBE on the bottom image) and while three anomalies to the west have a similar WNW trend, they differ in being broader than that of the dyke and they are similar to other anomalies to both the north and south. In this region are mafic and felsic gneisses with measured susceptibilities that, particularly on Cocked Hat Island, are relatively high (Figure 4.4), especially for felsic units (e.g., Hunt et al. 1995). Furthermore,

Thule Group sediments extend from the coast of Greenland under Kane Basin, but the region of subdued, longer wavelength anomalies that indicate the presence of the Thule sedimentary cover ends approximately along a north-south line (the "BBE", Figure 4.4) that separates this region of reduced aeromagnetic relief from one to the west of high relief with a pronounced WNW grain, and that also passes through the end of the Kap Leiper anomaly (Oakey & Damaske 2006). To search for any landward continuation of the Kap Leiper dyke, a helicopter reconnaissance was carried out in 2005 of the Ellesmere shoreline, comprising Cocked Hat Island, along strike and to 119

Figure 4.4: Top: Shadow illuminated total field anomaly map of the Smith Sound area (modified after Oakey 2006, used with permission). The KL and LG dykes are marked, as are magnetic susceptibilities measured from samples taken as part of this study. Bottom: Enlargement of area shown in white box in top image. The western extension of the KL dyke is shown, with speculative extensions shown as white lines. BBE = Buchanan Bay East boundary defined by regional aeromagnetic patterns by Oakey (2006). N/S flight line spacing is 2 km; flight elevation 610 m. the west of the anomaly, and Pirn Island, located immediately south of Cocked Hat Island.

Despite nearly 100% gneiss outcrop, no dyke was seen. For 50 km west of Cocked Hat Island, at least five broad tracts of continuous outcrop run for 5-10 km across the expected position of the dyke and yet no dyke has been mapped (Frisch 1984c). If the Kap Leiper dyke continues into

Canada without significant displacement, the Leffert Glacier dyke (LG in Figure 2.1, Figure 4.4) appears to be the only possible candidate. It runs along the north side of the Leffert glacier and could represent a left-stepping segment of the KL dyke, with an offset of about 10 km. A comparison of chilled margin geochemistry and paleomagnetism of the LG and KL dykes is shown in Figure 4.5. Both dykes are of similar width, but the KL dyke differs geochemically from LG in having significantly different incompatible trace element (Nb, Zr, Y, Sr, Rb) ratios

(Figure 4.5) and rare-earth element pattern (Figure 3.11). Since individual dykes can be traced for hundreds of kilometres with basically no change in the chemistry of their margins (e.g.,

Kalsbeek & Talyor 1986), the observed differences between KL and LG are significant. One possibility that could account for the difference is crustal contamination because of the different host rocks into which the dykes are injected - Thule sediments for KL and Archean granites for

LG. Although KL contains more water (Appendix B), the variation in trace element ratios cannot be attributed to Thule contamination. The LG geochemistry (major-element, minor-element,

REE) from chilled margins is indistinguishable from that of almost every other dyke from

Ellesmere and Devon Islands (with the few exceptions noted above), as well as those in

Greenland (Figures 3.8-11), which are intruded into Thule sediments. Therefore the difference between the KL and LG geochemistry appears to rule out their being from the same dyke. The

LG and KL virtual geomagnetic poles are also different at the 95% confidence level (Figure 4.5), and this difference also contributes evidence against the correlation of KL and LG as this offset Figure 4.5: Comparison of geochemical and paleomagnetic results for the Kap Leiper dyke of Greenland and the Leffert Glacier, Cadogan Glacier, and Hope Monument dykes of Ellesmere Island, and Devon Island (see Figure 2.1 for locations), the most likely candidates for the western extension of the Kap Leiper dyke based on their paleomagnetism. 122 is in the same sense as that seen between the mean pole positions for the Devon Island dykes and those from Greenland (Figure 3.16), a difference that may reflect motion between Canada and

Greenland. The overall conclusion based on the geochemistry and paleomagnetism is that sites

KL and LG are not sampling the same dyke, and therefore plate reconstructions using this dyke as a linchpin to indicate that no offset exists across Nares Strait (Harrison 2006; Oakey 2006;

Dawes 2007) are not valid. The KL dyke ends about 3 km off the Ellesmere shoreline where, based on the Judge Daly Promontory faults (Harrison 2006) the Wegener fault would be placed - i.e., along the western side of Nares Strait. The dyke may have terminated naturally, which is unlikely as the direction of propagation of the Franklin dykes has been proposed to be from a source to the west (based on the radiating geometry of the Franklin swarm), and both E/W- trending and N/S-trending dykes are found in host rocks of various lithologies and structural trends (see Figure 1.2), which indicates that host rock type or structural grain does not control dyke emplacement in the area. Theoretically, an individual dyke of the typical scale of these

Franklin dykes (ca. 20 m thick) can propagate at a level of neutral buoyancy for over 2000 km before crystallizing, regardless of the type of host rock, assuming that sufficient magma pressure is maintained (Rubin 1995 and references therein). An alternate explanation, therefore, is that the KL dyke has been offset.

As the Thule/Franklin dykes are now correlated across Nares Strait, and as the paleomagnetic poles from dykes on either side of the Strait are offset in the direction and magnitude that allows sinistral motion of Greenland after the emplacement of the dykes, new evidence is provided for the degree of offset of the Greenland/North American plates, as well as constraining the location of the plate boundary between the two. 123

Another, less well-defined, marker across Nares Strait is the boundary in the crystalline basement rocks between Archean-age and Paleoproterozoic-age rocks. First, the transition between ca. 2.5 Ga Archean rocks to the south and ca. 2.0 Ga Paleoproterozoic rocks to the north occurs at about 78°N in the Thule area (though the exact boundary is covered by the Inland Ice) and at about 76°N on the Canadian side, somewhere under Jones Sound (Nutman et al. in press;

Dawes 2006; Dawes 2004; Frisch 1988) (Figure 1.2). Restoration of this roughly E/W-trending boundary into alignment would place Inglefield Bay directly across from Devon Island. Second, the trends of the gneissosity do not match when compared on either side in their present position:

E/W on Devon Island and Inglefield Bay, but N/S on Ellesmere Island across from Inglefield

Bay (Dawes & Garde 2004; Dawes 1991; Frisch 1984a,b,c) (Figure 1.2). However, restoration of the proposed sinistral motion between Greenland and North America would place regions with similar structural trend in closer alignment, though a mismatch would still exist between the

Paleoproterozoic rocks of Ellesmere Island (N/S gneissosity) and of Greenland (SW/NE north of

Inglefield Bay). Conversely, a reconstruction that places the Wegener Fault within Ellesmere

Island and through Jones Sound (Harrison 2006; Oakey 2006; Figure 1.8) would juxtapose the

Paleoproterozoic rocks with a roughly N/S gneissosity of southeast Ellesmere Island between the

Archean basement rocks with a roughly E/W gneissosity of Devon Island to the west, of

Inglefield Bay to the east, and of Baffin Island to the south (Figure 1.2).

In recently-published studies, the chief reasons for moving Greenland as a plate relative to North America (and therefore Ellesmere Island) are based on geological and geophysical observations to the southeast of the area of concern in this study. There is no doubt about the presence of oceanic crust under Labrador Sea and Baffin Bay, and of parallel marine magnetic anomalies indicating rift-related seafloor spreading during the Paleocene and Eocene (e.g., Oakey 2006; Skaarup et al. 2006; Geoffroy et al. 2001; Roest & Srivastava 1989; Srivastava et al. 1982). Furthermore, the correlation of geological units across the Labrador Sea and Baffin

Bay (e.g., St-Onge et al. 2005 and references therein; Buchan et al. 1996) make it evident that

Greenland and North America were immediately adjacent to one another prior to the opening of the Labrador Sea and Baffin Bay. There must, therefore, be a major fault to the northwest of the spreading system to accommodate the required movement of the Greenland plate; whether this fault is located in Nares Strait or not. There is little debate about the location of the boundary between the North American plate and the Greenland plate to the north of Kane Basin (Figure

4.6). The Judge Daly Promontory on the eastern coast of Ellesmere Island hosts several transcurrent fault strands, known collectively as the Judge Daly Fault System (Harrison 2006 and references therein). From onshore within the Judge Daly Promontory to offshore into the

Lincoln Sea, a linear trend of Tertiary-aged pull-apart basins extends NNE-ward, and have been partially overprinted by the Eocene transpression caused by the change in direction of Greenland upon opening of the Atlantic Ocean (see Introduction) (Damaske & Oakey 2006). Kennedy

Channel, the waterway between Ellesmere Island and Greenland north of Kane Basin, is underlain by a late Cretaceous-early Tertiary elongated basin that occupies much of the Channel, as has been indicated by reflection seismic profiles. This basin has been interpreted as a flower structure in which large-scale displacement occurred along the "stem" fault that extends into the basement rocks, and the "petal" faults bound the offshore basin (Jackson et al. 2006). This fault system has an eastern margin just west of Hans Island in Kennedy Channel, where both the reflection seismic survey of Jackson et al. (2006) and a refraction seismic study (Funck et al.

2006a) show a steeply-dipping fault marked also by a distinct change in lithology, from late Figure 4.6: Map showing proposed location of the Wegener Fault (red dashed line) and related faults according to seismic transects. JDFS = Judge Daly Fault System; NB = Northwater Basin; KB = Kiatak Basin; SB = Steensby Basin. TG = Thule Group outcrop. KL = Kap Leiper dyke, as located offshore of Greenland by aeromagnetic survey. Green line indicates approximate extent of seismic transect across southern Nares Strait (Funck et al. 2006). Silurian carbonates and mudrocks to the east, and late Cambrian to early Silurian carbonates to the west. Given the presence of two visible major sinistral faults in the area (the Judge Daly

Fault and the Archer Fiord Fault) (Harrison 2006), and the structural complexities imposed by both the Ellesmerian and Eurekan orogenies, compelling evidence exists that the plate boundary in this region takes the form of a system of parallel faults, with distributed sinistral motion. It is impossible, however, to determine the exact displacement here - the Eocene compressive event obscured many features, and a significant portion of the fault system is underwater.

Some workers (e.g., Dawes & Kerr 1982, Harrison 2006) have attempted to place absolute limits on displacement in the vicinity of Judge Daly Promontory based on correlating onshore sedimentary units. These correlations, however, focus on shallowly-dipping units, and tend to assume linear continuations of the facies changes, paleocontinental slopes, and leading edges of orogenie belts from one side of Nares Strait to the other, despite their irregular shapes elsewhere along their profiles. In particular, across Smith Sound the Thule Group sediments have been matched up to the member level, with individual sills on either side being cited as precise correlatives, and therefore evidence that the rocks of the Inglefield Bay area of

Greenland, the adjacent region of Ellesmere Island, and the seaway between them are a single, stable continental block through which no fault can exist (Dawes 1997, 2007). However, the units of the Thule basin and overlying rocks are flat-lying to gently folded where they drape over basement topography (Dawes 1997). Flat-lying units are poor in terms of their ability to be correlated, as their continuity makes them insensitive to lateral displacement, especially if units extend laterally to distances greater than the amount of proposed fault offset. Furthermore, the strike of the paleoshelf is to the NNE, almost parallel to Nares Strait (Hofmann & Jackson 1996;

Jackson & Ianelli 1981), with paleocurrent directions to the west on Ellesmere and the western 127 limit on Greenland, but locally to the southwest and south (Dawes 1997, Figure 1.7). As paleocurrent flow is generally orthogonal to the proposed Wegener fault, lateral facies changes parallel to the fault may be minimal. This weakens the strength of correlations of paleogeography and stratigraphy across where the fault may lie that lead to the conclusion of no intervening movement. The northern limit of the Thule basin is not defined offshore (Harrison

2006), and onshore exposures (Dawes 2006; Frisch 1984c) are scattered and patchy on the

Canadian side, and seem to disappear under the Eurekan fold-and-thrust belt (Harrison 2006, his figure 15). On the Greenland side, the basal strata are bounded by post-depositional faults to the north, south, and northeast (Dawes 1997). What is not yet clear, as the Borden and Fury and

Hecla Basins have not been as well-defined stratigraphically as the Thule Basin, is whether local dips in the strata are due to basin geometry, post-depositional deformation, or irregularities in the basement topography, and the duration of sedimentation also needs to be better defined. Another question with respect to the correlation of Thule basin units is of their lateral extent. The Thule basin has been correlated with the Mesoproterozoic basins of northern Baffin Island, the Borden

Basin and the Fury and Hecla Basin, stratigraphically (Hofmann & Jackson 1996; Narbonne &

James 1996; Jackson & Ianelli 1981) and in terms of their 813C profiles (Kah et al 1999), and to reconstruct the Inglefield Bay area to a position adjacent to Devon Island would connect these units closely. The Nauyat volcanics of the Borden Basin also have correlatives in sills and flows of Mackenzie age in the equivalent strata of the Thule Basin. Franklinian platform sediments are flat-lying when undeformed, and the zone of Eurekan deformation curves into parallelism with

Nares Strait just north of Bache Peninsula. This also inhibits their use as marker units.

Around Kane Basin, geological markers that might constrain the relative motion of the two plates are poorly-exposed (Harrison 2006). Outcrops of the Mesoproterozoic Thule basin are found on both sides of the Nares Strait at this latitude, but as no offshore seismic survey has been carried out in the area, the northern "zero-edge" of the Thule basin is presently unknown

(Harrison 2006). A reflection seismic survey carried out in the southern part of Nares Strait, where it meets Baffin Bay, reveals the presence of Thule sediments, as well as several younger basins interpreted to be late Cretaceous-Tertiary in age (Neben et al. 2006). One basin in particular, the Northwater Basin (NB in Figure 4.6), has a N/S trend and is bounded by a "fan­ like system" of steep, narrowly-spaced faults. This basin has been interpreted by Neben et al.

(2006) as a negative flower structure that follows the main trend of the postulated Wegener Fault in the area, and is of the same size and tectonic style as the basins found on the east coast of northern Ellesmere Island, i.e., in the vicinity of Judge Daly Promontory. Smaller basins adjacent to the Northwater Basin (Steensby Basin and Kiatak Basin, SB and KB on Figure 4.6) are interpreted as pull-apart basins, created by the northward movement of Greenland during earliest rifting of Baffin Bay, which were then deformed when Greenland changed direction with the opening of the Atlantic (Neben et al. 2006). Further potential constraints are discussed below. Unfortunately, a refraction seismic transect across the southern part of Nares Strait

(Funck et al. 2006b) has a 100 km-wide gap centred just off the east coast of Ellesmere Island

(Figure 4.6), preventing a definitive confirmation of the precise location of the Wegener fault in southern Nares Strait.

Reconstruction of the Nares Strait region

As mentioned in the Introduction, Harrison (2006) and Oakey (2006) place the plate boundary between North America and Greenland within central Ellesmere Island (Figure 1.8), with a significant sinistral fault - an alternate placement of the Wegener Fault, with a displacement of approximately 300 km - in Jones Sound, between Ellesmere and Devon Islands.

Spreading in Baffin Bay and Labrador Sea is taken up in various zones of deformation in the eastern Arctic archipelago. Crustal thinning in Lancaster Sound, between Devon Island and

Baffin Island, indicated by a gravity low may account for up to 40 km of extension between

Baffin Island and Devon Island (Oakey 2006). The Sverdrup Basin of western Ellesmere and

Axel Heiberg Islands has experienced N/S shortening of up to 100 km and rotation of approximately 10° counterclockwise (Wynne et al. 1983; Jackson & Halls 1988; De Paor et al.

1989; Okulitch et al. 1990) that could account for up to a further 100 km of intraplate E/W- directed deformation caused by the Eurekan orogeny, the result of Greenland's late-stage convergence with North America (De Paor et al. 1989). Restoration of this rotation and extension would place the eastern margin of Ellesmere Island closer to Greenland in a pre-drift reconstruction (Figure 4.7). Figure 4.7 shows a reconstruction with the above effects restored:

Lancaster Sound has been closed by approximately 40 km, and the "Ellesmere-Devon Terrane", the part of the Rae craton beneath Devon and southern Ellesmere Islands (e.g., St-Onge et al.

2005, Figure 1.6) moved slightly eastward, as this rigid, cratonic block may have acted to a small degree as an indenter as it was pushed westward during the collision of Greenland with North

America in the Tertiary. The Paleozoic Ellesmerian Orogeny produced a series of faults, some of which penetrated the entire thickness of the crust (Soper & Higgins 1990; Okulitch et al. 1990,

Trettin 1991a), and were reactivated during the Eurekan (De Paor et al. 1989; Piepjohn et al.

1998), thereby providing a setting for the uptake of deformation within Ellesmere Island.

Central/Eastern Ellesmere Island has been rotated back 10 degrees clockwise, and expanded in an E/W direction approximately 100 km to correct for compression during the Eurekan orogeny, resulting in an expanded Sverdrup Basin. To accommodate the oceanic crust under Baffin Bay 130 and Labrador Sea, and the various geological correlations across these waterways (see

Introduction), Greenland has been moved closer to Baffin Island in the amount indicated by their present continental shelves (Oakey 2006). The boundary has been maintained between

Greenland and North America in Nares Strait, in accordance with the correlations of the Franklin dyke swarm on either side. It is apparent that should the plate boundary lie within Ellesmere

Island (as Harrison (2006) and Oakey (2006) propose), the Wegener Fault must not only lie within Jones Sound, but be able to accommodate at least 300 km of sinistral strike-slip movement along a fault that passes between Coburg Island and Devon Island, a gap of no more than 25 kilometres. No evidence exists for a major, narrow fault in this area. Furthermore, Grist

& Zentilli (2005) report apatite fission-track data from the dykes at sea level along the eastern coast of Ellesmere Island, in the Clarence Head (the location of dyke CL, Figure 2.1) and Smith

Sound regions, that record a period of Late Cretaceous - Paleocene heating before cooling during the Late Paleocene. If Smith Sound is entirely within the Greenland plate, then it would be distantly removed from the tectonic activity during this time and therefore would not record the thermal effects of the plate motion, interpreted by Grist & Zentilli (2005) as being a result of the uplift of rift flanks along an active plate margin. Therefore, the existing evidence indicates that the boundary lies along the east coast of Ellesmere Island as shown in Figures 4.4 and 4.7.

The terrain, however, would have looked quite different at the time. Figure 4.6 maintains present-day physiography for ease in interpretation. Prior to rifting in Baffin Bay/Labrador Sea, today's highstanding basement gneisses of eastern Ellesmere, Devon, and Baffin Islands and northwestern Greenland would have been covered by extensive, predominantly clastic sediments of the Mesoproterozoic Thule Group as part of a Proterozoic shelf sequence (Jackson & Ianelli

1981; Kah et al. 1999). These sandstones would have been overlain by Lower Paleozoic 131

0 fej /«^S

^1

Figure 4.7: Reconstruction of pre-drift configuration of Greenland and North America, maintaining present-day physiography, showing the operations that were performed to create the reconstruction. 132 carbonate shelf strata and adjacent deep-water clastic sediments as part of a basin that would have extended across the Canadian Arctic islands and northwestern Greenland (e.g., Soper &

Higgins 1990), and atop these rocks would be overlain the sediments of a larger Sverdrup Basin than is preserved today. Indeed, the uplift of the Eurekan Orogen and multiple subsequent glaciations have reduced the western portion of the Arctic Archipelago from Banks Island to at least as far north as Ellef Ringnes Island, west of Ellesmere, to a peneplain (Harrison et all999).

The eastern Canadian Arctic and northwest Greenland have been affected by severe erosion caused by the uplift of the Ellesmerian and Eurekan Orogens, as well as several pulses of glaciation throughout the Quaternary (England 1987; Harrison et al. 1999). Thus the sedimentary cover was to varying degrees stripped away, from relatively minor erosion on south

Devon Island, to exposed Archean to Paleoproterozoic metamorphic basement in southeastern

Ellesmere and the Inglefield Land area (Frisch 1984a,b,c; Dawes 1991). Today, patches of these sedimentary units are preserved as remnants, with the original boundaries of the depositional environments destroyed. The Thule, Borden, and Fury and Hecla basins, therefore, are interpreted as being the remains of a once regionally extensive continental shelf depositional environment, rather than being discrete depocentres with well-preserved original boundaries.

An observation that must be accommodated in any reconstruction is the limit of ca. 100 km of left-lateral displacement placed on the faults that make up the Judge Daly Fault System in northern Nares Strait (Harrison 2006). While the amount of oceanic crust underlying Baffin Bay requires that the Greenland plate has moved at least 220 km northward relative to the North

American plate, Harrison (2006) notes that this degree of motion cannot be taken up entirely along the faults identified on northern Ellesmere Island. As described above, modelling of gravity anomalies within Ellesmere Island (Stephenson & Ricketts 1990) and Lancaster Sound 133

(Oakey 2006) indicates that the crust underlying Ellesmere Island and Devon Island has undergone considerable shortening. While the exact amount is not precisely known, approximately 140 km of the Baffin Bay displacement can be accommodated within the North

American plate, decreasing the amount of relative displacement required between Greenland and

Ellesmere Island. However, as only 40 km of this shortening is taken up within Lancaster Sound

(i.e., south of the zone of Eurekan deformation) a significant relative offset (ca. 200 km) is still required in the Smith Sound area, where the Franklin dykes are found.

Geochronology of baddelevite in the Franklin intrusions

The 723 +4/-2 Ma age on the Franklin swarm of Heaman et al (1992) is based on a regression of slightly discordant data with a lower intercept close to 0 Ma. It is therefore relatively insensitive to alpha recoil effects, which would produce only second order biases in zu'Pbr°Pb ages. The consistency of this age with the 721 Ma age presented here results from the

EAV trending dykes argues that the older fractions preserve accurate ages, although some analyses from samples CG-7 and BG are clearly biased downward, giving younger/discordant ages relative to other fractions from the same dyke. The 713-716 Ma results from three N/S trending dykes fall within the range of Pb/ U ages observed in the composite data set of

Heaman et al. (1992) but, due to the large 207Pb/235U errors from these small fractions, their data cannot be evaluated for discordance. Each of these individual dykes, where analyses were mostly done on fractions of 1 -2 grains so there is a minimal averaging effect, gives consistent overlapping 206Pb/238U ages. However, the possibility must be addressed that this relatively young age could be an artifact of alpha recoil. Subswarm Fraction #s used N Weighted average MSWD Upper intercept MSWD EVENT 206 238 Sample in calculations Pb/ U age (Ma) age (Ma) of regression FRANKLIN CG (E/W dyke) Devon Island 1,5,6 3 721 ±2 2.40 BG (E/W dyke) Devon Island 1,2,3,4 4 726 ± 24 0.22 QA (E/W dyke) Thule 1,2,3,4 4 721 ±4 1.30 GF (sill) Thule sills 1,2,3 3 712±2 0.16 CV (N/S dyke) Clarence Head 1,2,3 3 713 ±3 0.04 CH (N/S dyke) Clarence Head 1,2 2 713±2 0.08 CL (N/S dyke) Clarence Head 1,2,3,4,5,6 6 716±2 3.10 MELVILLE BUGT OF (NW dyke) 1,2,3 3 1622 ±3 0.20 DUNDAS HARBOUR DH (E/W dyke) 1,2,3,4 4 1337±2 1.07 Table 4.1: Summary of calculated/interpreted ages for each dyke analyzed in this study. Because of the small size of the baddeleyite crystals, possible effects of alpha recoil

(Kigoshi 1971, Fleischer 1982, Romer 2003) must be considered in these and previously published results. Decay of an alpha-emitting radionuclide causes the daughter nucleus to recoil through the crystal lattice. For zircon, recoil distances are thought to vary from about 20 nm to

30 nm but there are 8 alpha decays in the U system and 7 in the U system so the average total displacement for a radiogenic Pb isotope is estimated to be roughly 0.1 micron (Romer

2003, and included references). Any Pb loss due to this mechanism will directly affect the

206pb/238Tj age Thig ig likely tQ be minor (<0iio/0) for most zircon analyzed by TIMS, where the outer surface is removed by laboratory abrasion (Krogh 1982) and grain sizes are normally in excess of 50 microns. However, baddeleyite is not abraded and often takes the form of extremely thin tabular crystals, raising the possibility that significant recoil loss could have occurred normal to [100] faces. For recoil distances comparable to those of zircon (average alpha recoil track length of 0.025 microns) and assuming a quasi-infinite tabular geometry, the level of Pb loss would scale inversely with the thickness and should be about 0.45% for a thickness of 10 microns. Baddeleyite crystals dated in the early part of the present study, before the potential importance of alpha recoil effects was realized, were only photographed normal to the wide

[100] faces using an optical microscope and thicknesses were roughly estimated. Measurement of similar-size grains from the same samples indicates thicknesses that vary in the approximate range 6 to 20 microns. Therefore, ages too young by about 5 m.y. might be expected for the thinner grains if alpha-recoil effects are similar to those in zircon. Model thicknesses for these analyzed grains are given in Table 3.1 (as dmod), based on an assumed U concentration of 300 ppm, as U concentrations in most published analyses of large, accurately weighed baddeleyite fractions average about 300 ppm and do not vary much more than 50% about this value (e.g.,

Krogh et al. 1985). Calculated thicknesses based on this assumption show no consistent correlation with zuoPbrJ0U ages. As a further check on potential alpha recoil effects, grains from the E/W-trending CG-7 dyke and the N/S-trending CL-1 dyke were imaged at a high angle using a scanning electron microscope to determine precise dimensions, including thickness, before analysis (Figure 4.8). The results (analyses 5 and 6 from the CG-7 dyke and analyses 4, 5, and 6 from the CL-1 dyke) indicate that there is no such correlation for thicknesses varying over the range 10-20 microns. While the U concentrations for the accurately-measured grains from the

CG-7 dyke were close to the "model" value of 300 ppm U (CG7-5: 395 ppm, CG7-6: 371 ppm), those from the CL-2 dyke have significantly higher U concentrations (CL2-4: 1127 ppm, CL2-5:

623 ppm, CL2-6: 929 ppm), implying significant errors in the previous thickness estimations for these grains, though not necessarily for all of them. The lack of correlation between age and measured thickness argues that the N/S trending dykes form a distinct younger age group. An additional conclusion is that alpha recoil distances are either significantly smaller in baddeleyite than in zircon or that previous estimates based on recoil distances in zircon are too high. 136

Figure 4.8: Examples of SEM photomicrographs of single baddeleyite crystals illustrating their thicknesses relative to their larger dimensions. Shown are images of examples from: a) the Cadogan Glacier dyke; b) the Clarence Head dyke. 137

The Franklin intrusions were previously dated at several locations covering the known extent of the Franklin event, from Victoria Island in the west to the eastern tip of Baffin Island

(723+4A2 Ma, Heaman et al. 1992; 720±8 Ma, 716+4/-5 Ma, Pehrsson & Buchan 1999). Results presented here from the northeastern limit of the swarm range from 721 ± 2 Ma to 712 ± 2 Ma and are in approximate agreement with the age range of previous results. If these ages are accurate, they indicate that the large, high-density swarm of E/W-trending dykes on Devon

Island and Ellesmere Island is resolvably older than the geochemically similar N/S trending

"Clarence Head dykes". Previously published ages for the Franklin swarm were determined exclusively from dykes that are part of the radial pattern of the swarm, while the Clarence Head dykes are tangential to the focus of the swarm. Other sets of N/S-trending dykes, the Strathcona

Sound dykes on north Baffin Island and the Lasard River dykes of the mainland Northwest

Territories (Buchan & Ernst 2004), have not been dated although their paleomagnetism (Christie

& Fahrig 1983, Park 1981) indicate an affiliation with the Franklin event. The observation of the

Strathcona Sound swarm's crosscutting the radial Franklin dykes (Christie & Fahrig 1983;

Pehrsson & Buchan 1999) further supports the conclusion that the N-S dykes represent a distinct, younger event.

Emplacement of Orthogonal Dykes and the Remagnetization of the "Clarence Head" dykes

Examples of mafic dyke swarms tangential (or "circumferential") to the proposed plume source have been found in the caldera region of shield volcanoes (e.g., the Galapagos

Islands, Walker 1999). West and south of the main Neoproterozoic Franklin large igneous province, the SSE-trending "Lasard River" diabase dykes (LR in Figure 1.10) define a minor swarm, extending over a strike-length of roughly 300 km near the Arctic coast of the Northwest 138

Territories. On the basis of an Ar-Ar age of 708 ± 4 Ma (Ernst et al. 2004), and their paleomagnetic directions (Park 1981), these dykes (and minor sills) have been hypothesized to be related to the Franklin magmatic event (Buchan & Ernst 2006). Their age and orientation suggest that they could therefore be another set of dykes not only tangential to the Franklin plume but also younger than the radial swarm. The "Tree River" dykes (TR in Figure 1.10) have been proposed to be associated with the Franklin magmatic event as well, though they have not been dated (Ernst & Buchan 2004; Ernst & Buchan 2006), and also have a trend at a high angle to the radial Franklin swarm. Another mafic dyke swarm that may be tangentially emplaced relative to the proposed Franklin plume is the undated Strathcona Sound swarm (SS in

Figure 1.10). While their NNW orientation is not quite orthogonal to the trend of the dykes on

Baffin Island, which have a SE/NW trend, they are located in the NNW-trending Borden fault zone (Jackson et al. 1978), which may have influenced their orientation (e.g., Tokarski 1990).

While the more common radial pattern of dyke swarms is a result of the uplift of the lithosphere by the impinging mantle plume and injection of dykes from an overpressured magma reservoir, concentric zones of extension may form due to the depression of this source area as the plume reservoir is depleted (e.g., McKenzie et al. 1992). Magma could have been continuously or episodically transported along pre-existing conduits at a depth of neutral buoyancy (Lister &

Kerr 1990) before emplacement in the upper crust according to the prevailing stress regime.

Because the N/S-trending dykes are younger, it is possible that they were emplaced orthogonal to the E/W-trending dykes of the main Franklin swarm because of extension caused by the plume head depleting and collapsing.

Another explanation supported by the younger age of the N/S-trending dykes and their orientation orthogonal to that of the main swarm is a change in the regional deviatoric stress field 139 as a result of emplacement of the E/W-trending dykes to allow the emplacement of the N/S- trending dykes (radial to tangential; e.g., Feraud et al. 1985). If the regional horizontal stress field was relatively low and uniform such that the intrusion of the radial Franklin dykes changed the maximum principal stress direction by 90° (i.e., o\ became 03), this would have allowed the emplacement of the Clarence Head and possibly the Lasard River, Tree River, and Strathcona

Sound swarms. The absence of any significant change in composition supports the interpretation that the dykes, in either orientation, are being fed from the same source.

The most significant difference between the radial and tangential Franklin dykes is the anomalous remanence direction of dykes CL and NS, which also feature the prominent asymmetric peak in susceptibility vs. temperature plots (Figures 3.23, 3.24). Previous studies

(e.g., Halls et al. 2001) attribute the peak to the presence of the Fe9Sio form of pyrrhotite which exhibits ferrimagnetic properties in the vicinity of 200°C when antiferrimagnetic hexagonal pyrrhotite undergoes a crystallographic transformation to a ferrimagnetic phase, then at ~230°C, another transformation to antiferrimagnetic monoclinic pyrrhotite occurs (Dekker 1989). The lack of a significant increase in susceptibility at 320°C on the cooling run indicates that the pyrrhotite has been oxidized to magnetite, a transformation that typically occurs at temperatures higher than 500°C (McElhinny & McFadden 2000). Therefore, the anomalous remanence direction is interpreted as a chemical remanent magnetization (CRM) carried by pyrrhotite, which was precipitated by fluids expelled by the accretion of the Pearya terrane to the north of

Ellesmere Island in the Devonian (Trettin 1991; Figure 1.2; see Introduction), which was also responsible for the formation of the Polaris Mississippi-Valley-Type ore deposit on Little

Cornwallis Island, due west of Devon Island (Symons & Sangster 1992). This is supported by the position of the VGP of the CL dyke relative to the North American apparent polar wander path (Figure 4.9), where it fits reasonably well to the late Devonian-early Silurian part of the

APWP, corresponding to the time of accretion of Pearya. The N/S orientation of the dykes probably made them effective fluid conduits (Denyszyn et al. 2006a), and their petrography

(Figure 3.12) also indicates a much greater degree of alteration compared to other Franklin dykes

(Figure 3.6). Another line of evidence supporting the fluid-flow hypothesis is the observation of the difference in remanence directions measured in the samples from the interior of the NS dyke versus those from its chilled margin (Figure 3.25). Samples from the interior tend to have the more southerly direction characteristic of the CL dyke (declination ca. 170), where it is carried by pyrrhotite, while samples from the margin have a more easterly declination (ca. 135), characteristic of the Franklin "R" direction. Interior samples demagnetized thermally (NS4-2t) did not show a significant presence of pyrrhotite. The coarser-grained interior of the dyke would have permitted more fluid to flow through and allow the precipitation of pyrrhotite. The permeability of the interior may have been enhanced by the presence of dyke-parallel joints.

While shearing often occurs along dyke margins and therefore might make them channels for fluid flow, thin sections (Figure 3.12) do not show any evidence for shearing or fracturing that would increase permeability. This observation also places limits on the maximum temperature that the rocks have been heated since their formation in the Neoproterozoic. Apatite fission-track and (U-Th-Sm)/He thermochronological studies, capable of resolving thermal events through temperatures of 40-130°C, were carried out in the region (Grist & Zentilli 2005), revealing a spread of fission-track ages more or less continuous from the Late Carboniferous through the

Pal eocene, related to periods of erosion and clastic sedimentation at various times throughout the

Canadian Arctic Archipelago. This study shows that for southeast Ellesmere Island at least, at no time since the Devonian did temperatures of the rocks exceed 320°C. 141

Figure 4.9: Apparent polar wander path for North America (from Torsvik et al. 1996) and the VGP for site CL. 142

New Ages of Proterozoic Arctic Magmatism and its effect on the Apparent Polar Wander Path of Laurentia

The 207Pb/206Pb age of the Melville Bugt dyke at Olrik Fjord of 1622±4 Ma agrees

generally with that obtained by Hamilton et al. (2004) of 1629±1 Ma. Both ages agree with the

less-precise whole-rock Rb-Sr age of 1645±35 Ma reported by Kalsbeek & Taylor (1986). The

paleomagnetic results presented here differ from those of Hamilton et al. (2004), but their results

are from a single dyke compared to three in this study, and the fact that the results in this study

include a geomagnetic reversal suggest that the recorded magnetization of the three Melville

Bugt dykes from Inglefield Bay is primary. Furthermore, the Melville Bugt paleopole presented here (Figure 4.10) lies in close proximity to the 1740 Ma Cleaver Dykes pole (Irving et al. 2004) and the 1590 Ma Western Channel Diabase pole (Irving et al. 1972; Hamilton & Buchan 2007).

While coherence with the established APWP is not in itself validation of the pole, it is considered more reliable by the Melville Bugt pole's reasonable agreement with established results for the Laurentian craton. The agreement in ages between the results of this study and that of Hamilton et al. (2004) indicates emplacement of the Melville Bugt swarm's large

(~100m-wide) diabase dykes over at least 5-10 million years.

The age of 1337 Ma for the dyke at Dundas Harbour is a unique age of mafic dyke emplacement in North America, and therefore adds valuable information to a significant (>100

Ma) gap in the apparent polar wander path for North America in the Proterozoic (Figure 4.10).

This Mesoproterozoic age has been suggested elsewhere in the area: a N/S-trending dyke on the south coast of Devon Island, some 50 km to the west of Dundas Harbour, was dated at 1340 Ma by the K/Ar method (Frisch 1988), and the NE-trending Nigarfivik dykes of eastern Inglefield

Bay (Figure 1.9) have been surmised to have a Mesoproterozoic age (Buchan & Ernst 2004). 143

Though these dykes have different trends, their correlation should be attempted in order to determine if there is a previously unknown mafic dyke swarm in the Canadian Arctic and

Greenland. This would be useful for paleogeographic reconstructions such as that carried out in this study with the Franklin dykes, and also to better constrain a -1340 Ma paleopole for

Laurentia.

The U-Pb age of 721±2 Ma obtained for both the "Thule" dykes of northwest Greenland and the "Devon Island" dykes of Arctic Canada represents the first time these intrusions have been precisely dated and definitively correlated not only with each other, but with the Franklin dyke swarm as a whole. The age of 715 Ma of the N/S-trending "Clarence Head" dykes, and of

712 Ma for the "Thule sill" at Granville Fjord are the first ages ever obtained from these sets of intrusions, and indicate an extended period of Franklin magmatism, over 2000 km away from the proposed source (Heaman et al. 1992). Another significant finding is the identification of at least one magnetic reversal over the time period of emplacement of the Franklin swarm. The dykes dated to ca. 720 Ma (BG, CG, QA) have an N direction, dyke CV (dated to 713 Ma) has an R direction, and the CH (713 Ma) dyke and the 712 Ma GF sill both have the N magnetization direction. A geomagnetic reversal, therefore, occurred between the times of emplacement of the

CH and CV dykes, which are identical in age within 1 m.y. error. As there are both N- and R- magnetized dykes in both the E/W-trending swarm and the N/S-trending swarm, if all of the E/W dykes were emplaced before the N/S dykes, then this would mean that at least two reversals are recorded.

Combining the ages of the intrusions with their paleomagnetic results provides three new poles to add to the apparent polar wander path (APWP) of Laurentia in the Proterozoic (Figure

4.8, Buchan et al. 2000 and references therein), using only "key poles" (Buchan & Halls 1990), 144

Figure 4.10: Apparent polar wander path (APWP) of "key poles" for Superior (pre-1800 Ma) and Laurentia (post-1800 Ma). Paleopoles and VGPs obtained in this study are highlighted: Red = Franklin; Orange = Dundas Harbour; Green = Melville Bugt. Data from Buchan et al. (2000), Torsvik et al. (1996), and references therein. in which only paleopoles that are well-defined in both paleomagnetic direction and age are

accepted. Although the Dundas Harbour dyke only provides a virtual geomagnetic pole (VGP)

as it is a single dyke, it is included in the APWP because: a) its paleomagnetic direction is very

different from that of the Cape Warrender (CW) dyke located very nearby, suggesting that its

magnetization is primary; and b) it is the only paleomagnetic result available between the key

poles of the Mistastin complex at 1420 Ma and the Gronnedal-Ika complex at 1299 Ma (in

Buchan et al. 2000). The poles for the Melville Bugt and Franklin swarms are dated to periods

when supercontinents may have existed: the Mesoproterozoic supercontinent Columbia (also

known as Nuna), during the period -1800-1500 Ma and the Neoproterozoic supercontinent

Rodinia, between ~1100-700 Ma (Rogers & Santosh 2003). The emplacement of the Franklin mafic dyke swarm has been linked to the incipient breakup of Rodinia, with an as yet unidentified continental fragment rifting off the northern (in modern co-ordinates) margin of

Laurentia (Heaman et al. 1992). Even though these rocks were formed during periods when these supercontinents were extant, they indicate very rapid plate motions at these times, rather than after the supercontinents' breakup (Figure 4.10). A detailed discussion of this observation with respect to the Franklin dykes follows below.

Another observation that arises when these poles are added to the Laurentian APWP is the apparent "preferred locations" that paleopoles seem to return to periodically. One "cluster"

of poles appears at approximately 45°N, 165°E, and another appears at approximately 5°N,

150°E (Figure 4.10). For example, the pole for the Mackenzie intrusions at 1267 Ma and the pole presented here for the Franklin dykes overlap within error, despite being more than 500 million years apart in age and with a significant degree of plate motion in that interval.

Furthermore, the proposed mantle plume sources for the Franklin and Mackenzie swarms are 146

Figure 4.11: a) Reconstruction of Laurentia and Baltica at approximately 1630 Ma using dyke JP2 and the Subjotnian quartz porphyry dykes of Finland (1630 +/- 9 Ma) (Mertanen & Pesonen 1995). (VGP for the Subjotnian quartz porphyry: Plat = 4.0 N, Plon = 158.0 E, dp = 4.0 , dm = 4.0 ) Because of the lack of longitudinal control in paleomagnetic reconstructions, Baltica has been shown in a number of possible configurations relative to Laurentia (held fixed). b) Reconstruction in which the reversed-polarity option for Baltica is illustrated. 147

close to each other (Heaman et al. 1992; Baragar et al. 1996). This raises not only the possibility

that the North American plate returned to a geographical location it had been to 500 million

years previously in the same orientation, but that it might have been sampling a long-lived (>500

Ma) mantle plume that caused the emplacement of the two swarms, an observation that has been

proposed for Proterozoic dyke swarms on the southern margin of the Superior Province (Halls et

al., in press). Assuming that the plume hypothesis is correct for both the Franklin and

Mackenzie magmatic events, the alternative possibility is that the plate was at the same latitude,

with the same azimuthal orientation, 500 million years later and a mantle plume impinged in the

same place within the plate, possibly a result of lithospheric thinning caused by the previous plume. The only definitive way to test the consistency of this pattern is to greatly enhance the resolution of not only the North American plate's APWP, but also that of neighbouring paleocontinents to constrain the longitudinal motion of the plate(s).

Presently, the only rocks with a well-defined paleomagnetic pole at ca. 1628 Ma (the age

of emplacement of the Melville Bugt dykes) are in Baltica, the Sipoo Subjotnian quartz porphyry

dykes of Finland (Mertanen & Pesonen 1995). However, given the ambiguity of paleolongitude

and polarity inherent in paleomagnetic studies and the lack of a well-defined APWP for Baltica

for this time period that would permit a test of a shared history of motion of the two plates, a variety of reconstructions for the two plates is possible (Figure 4.11). The results indicate a

significant degree of motion between Laurentia and Baltica between 1630 Ma and 1265 Ma, the

age of the chronologically nearest key poles for which a reliable fit can be made (The Mackenzie in Laurentia and the Jotnian intrusions of Baltica, Buchan et al. 2001), and that Baltica and

Laurentia did not drift as a unit over at least the early part of this period. 148

Differences in Paleopole Measurements across the Franklin Intrusions

The paleopole presented here for the Devon Island dyke swarm (5.8°N, 188°E,

N = 12, A95 = 9.9°, datum E in Figure 4.12), and dated to 721 Ma, is significantly different from that previously published for the Franklin dyke swarm. As this study is the only one that includes

geochronological results, paleomagnetic results, and stability tests from the same dykes, this pole

could be considered a "key pole" (Buchan & Halls 1990; Buchan et al. 2000) for the Franklin intrusions. The N/S-trending "Clarence Head" dykes, dated to 716 Ma, do not yield a reliable paleopole, as dykes CL and NS suffer from magnetic overprinting (see above), dyke BE has an

(X95 value greater than 15°, and only three samples from site CV yielded stable end points (Table

3.2). Dyke CH is therefore the only N/S-trending dyke that gives what can be considered a reliable magnetic direction, but as it is from a single dyke, the resultant pole is a virtual geomagnetic pole, secular variation not being averaged out. Previous paleomagnetic analyses of the Franklin and related intrusions (Fahrig et al. 1971; Robertson & Baragar 1972; Christie &

Fahrig 1973; Fahrig & Schwarz 1973; Park 1981; Palmer et al. 1983) have revealed that the roughly E/W-trending, shallow paleomagnetic direction found here (allowing for the geographic extent of the various studies) is also found in the older results, some of which were incorporated into the "grand mean" of Buchan et al (2000) described above. However, many of these datasets were obtained through less reliable blanket AF demagnetization of specimens at low fields, and/or indicate significant remagnetization by a steep-down or steep-up directed component, attributed generally to Eocene tectonic activity (Fahrig & Schwarz 1973; Pehrsson & Buchan

1999). This could result in "streaking" of the data points along a great circle that includes the normal Franklin direction, a steep component, and a reversed Franklin direction. A conclusive test for the primary nature of the remanence direction is lacking in this early work and no sites 149

Figure 4.12: Locations of Franklin poles from: A, Christie & Fahrig 1983; B, Palmer et al. 1983; C, Fahrig & Schwarz 1973; D, Fahrig et al. 1971; E, this study, Devon Island swarm; F, this study, Thule swarm. Circles are A95 (95% confidence). 150 that give the Franklin direction have been precisely dated. However, two E/W-trending

"Borden" dykes with paleomagnetic directions strongly affected by the steep overprint (Fahrig &

Schwarz 1973) were dated by the U-Pb method by Pehrsson & Buchan (1999) and revealed to be of Franklin age (ca. 720 Ma).

The data for sites from previous studies that meet the acceptability criteria outlined above

(N>5, a95<150) are shown in Figure 4.12. Most include both normal and reverse magnetizations, denoted here by NN and NR respectively, which when isolated do not give significantly different results. The poles in Figure 4.12 (with locations marked on Figure 4.13) are:

Pole A: 9.2°N, 153.3°E, A95=4°, NN=4, NR=9 (Christie & Fahrig 1983, NW Baffin Island);

Pole B: 6.0°N, 159.0°E, A95=8°, NN=6, NR=8 (Palmer et al. 1983, Victoria Island);

Pole C: 6.0°N, 168.3°E, A95=7°, NR=6 (Fahrig & Schwarz 1973, E Baffin Island);

0 Pole D: 11.2°S, 163.6°E, A95=14 , NN=8, NR=5 (Fahrig et al. 1971, E Baffin Island);

Pole E: 5.8°N, 188.0°E, A95=10°, NN=11, NR=1 (this study, Devon and Ellesmere Islands);

Pole F: 8.8°N, 178.7°E, A9S= 7.2° NN=8, NR=2 (this study, northwest Greenland).

The streaking of the data towards a steep component mainly affects the inclination, which in terms of the paleopole, affects its latitude (i.e., moves the pole normal to the observed pole distribution). The data of Pole D are significantly streaked, which accounts for its southern position relative to the other poles. Pole D is therefore discarded from this analysis. The results

from this study, particularly those from Devon Island (Pole E) also include data with a

significant spread in inclination (Table 3.2, Figure 3.15). However, there is no evidence for

contamination by a steep overprint, as the results from specimens with either the steepest positive or negative inclinations (Figure 4.14) show linear decay of components through the 151 origin, indicating that the remanence direction that they record is a single component, uncontaminated by an overprint.

While Poles A, B, and C are compared with Poles E and F, they represent significantly different pole positions for rocks of the same magmatic event. They lie at the same latitude so the difference in their longitudes cannot be attributed to greater or lesser degrees of influence of the steep component. Various possibilities can be proposed to explain this difference. They are either "tectonic causes", such as rapid drift of the continents or post-emplacement movement of tectonic blocks that host the dykes, or "paleomagnetic causes", such as a failure to average out secular variation or the influence of a non-dipole field component on the magnetization of the rocks. These possibilities are discussed below:

1) There may have been E-W tectonic extension or block rotation since the emplacement of the dykes, though to reconstruct the landmasses such that their poles overlap would require moving

Devon Island, Ellesmere Island and eastern Baffin Island over 500 km to the west (or an equal amount of eastward motion of Victoria Island and western Baffin Island). While there was extension in that sense in the late Jurassic-early Cretaceous when the Canada Basin opened, rifting the north coast of Alaska off of the western margin of the Canadian Arctic archipelago

(Grantz et al. 1998), the amount of extension was not nearly great enough, nor is there any evidence for significant E/W displacement of Devon Island relative to Baffin Island. Lancaster

Sound, the waterway separating Baffin Island and Devon Island, has been interpreted to be a failed rift arm of the Baffin Bay spreading system based on a prominent gravity anomaly under its eastern limit (Oakey 2006) but the maximum 40 km of extension would result in a shift in the paleopole position of less than 0.1 ° in both latitude and longitude. However, this explanation is invoked for the difference in Poles E and F, as the restoration of the Greenland plate to its pre- Figure 4.13: Distribution of intrusions or lava flows associated with the Franklin magmatic event (modified after Fahrig (1985), with dyke locations also from Frisch (1984a,b,c) and Dawes (1991)). CL = Clarence Head dykes; CS = Coronation Sills; DI = Devon Island dykes; LR = Lasard River dykes; NB = Natkusiak basalts (extrusive); SS = Strathcona Sound dykes; TD = Thule dykes; TR = Tree River dykes. A, B, C, D, E, F = Locations of previous paleomagnetic studies described in text. 153

BB1 -2, Brae Bay dyke N,,Up

BG3-4, Belcher Glacier dyke N, Up

AF demag. 20-95 mT

100

200

Figure 4.14: Examples of stable paleomagnetic end-points obtained from the Franklin dykes with the steepest-up (BB) and steepest-down (BG) inclinations, showing linear decay to the origin. This indicates that a single component of magnetization is being removed, without a secondary steep overprint. Symbols as in Figure 3.14. 154 drift position closer to North America has the effect of moving the Thule dyke swarm's paleopole (Pole F) eastward such that it lies atop that of the Devon Island swarm (Pole E) (see discussion of "Nares Strait Problem" above). This also reinforces the validity of the easterly position of Pole E relative to the others. In order for the rotation of tectonic blocks to be the cause of the spread in paleopoles, large portions of the Canadian Shield (represented by locations where the Franklin dyke swarm was sampled) would have had to rotate about vertical and horizontal axes separate from each other. The reconstruction that requires the least relative movement of the blocks is to restore each of the poles to Pole C. To accomplish this, i.e., to assume that the spread in the poles is entirely attributable to rotation relative to central Baffin

Island, cratonic Devon and Ellesmere Islands (Pole E) would have to have been rotated about a vertical axis ~20° counterclockwise and tilted downward ~11° to the west; the Thule area of

Greenland (Pole F) would have been rotated ~10°counterclockwise and tilted downward -10° to the west; central Victoria Island (Pole B) rotated ~8° counterclockwise and tilted upwards to the east -5°; northern Baffin Island (Pole A) rotated -15° clockwise with no tilt; and eastern Baffin

Island (Pole D) must have rotated relative to central Baffin Island (ca. 10 km away) -8° clockwise and tilted downward to the west ~7°. While counterclockwise rotation has been interpreted in northern and central Ellesmere Island as a result of the collision of Greenland with the North American plate (Wynne et al. 1983; Jackson & Halls 1988), these observations have been made in the deformed Paleozoic and younger cover rocks, which are undeformed and flat- lying on Devon Island. Furthermore, fault systems younger than the age of emplacement of the

Franklin dykes are steep andNhave an E/W trend on Devon Island (Frisch 1988), and a SW/NE trend on Baffin Island, interpreted to have been caused by extension during the opening of Baffin

Bay and without significant lateral offset (Jackson 2000). In order to accept the rotation 155 hypothesis, there would have to be significant motion of at least three different portions of Baffin

Island along faults between the sampling locations of Poles A and C, and of Poles C and D since

723 Ma, as well as between southern Devon Island and northern Baffin Island. There is no evidence for these major faults (Frisch 1988, Jackson 2000, St-Onge et al. 2006).

2) North America might have drifted between the intrusion times of the studied rocks. The poles presented here (Poles E and F) are for E/W-trending dykes, emplaced at 721 Ma. However, this study and previously published geochronological studies show that Franklin magmatism likely extended for 5-10 million years, or even longer, as Dawes (2006) reports dykes with trends characteristic of the Franklin swarm cutting the sill complex that includes the Granville Fjord sill

(GF in Figure 2.1), dated in this study to 712 ± 2 Ma. This magmatism has been related to plume-associated breakup of the Neoproterozoic supercontinent Rodinia (e.g., Shelnutt et al.

2004). Gurnis & Torsvik (1994) suggest that the presence of thick (>300 km) roots under old continents such as Laurentia can result in augmentation of normal plate velocities by about 6 cm/yr when the driving force is a thermal upwelling generated in the lower mantle by insulation by a supercontinent, allowing the root to act as a "sail" and increasing plate velocity. Another effect of the deep continental root is that the plate can be "pulled" towards a mantle cold region created by extended subduction on the margins of a supercontinent, also enhancing its velocity.

Therefore, given the time span indicated by the geochronological studies of the Franklin event, the difference in the poles might be a trace of the paleocontinent Laurentia moving as Rodinia breaks up, though given the time period established to date, plate velocities could be on the order of 30 cm/yr. While the Franklin intrusions are among the best-studied worldwide in terms of their age, what is lacking is precise assignment of ages to the paleomagnetic poles measured. To date, only the 721 Ma poles (E and F) presented in this study from E-W trending dykes have 156 definite ages assigned to them. Precise ages for the dykes yielding the other poles could definitively show whether the poles reflect apparent polar wander or another process.

3) True polar wander (TPW) may have occurred over this period. This refers to a displacement of the global lithosphere relative to the rotation axis (paleomagnetic axis), and is inferred to have occurred when the paleomagnetic determination of the geographic north pole deviates from the

"actual" position of the geographic pole determined by a fixed hotspot reference frame, and/or when apparent rates of continental motion reach or exceed rates upwards of 30 cm/yr (Evans

1998; Torsvik et al. 2001). TPW would affect all continents equally, and so requires that the apparent polar wander paths for all continents be offset equally at times of proposed TPW

(McElhinny & McFadden 2000). This axis shift is considered a possible result of large-scale mass imbalances in the mantle that alter the moment of inertia of the planet (Evans 1998;

McElhinny & McFadden 2000). One way that this could be generated is by the upwelling of mantle material due to insulation (Anderson 1982), or through shielding of the mantle by subducting slabs (Evans 1998) under a supercontinent such as Rodinia, which may have existed between 1100 Ma and approximately 720 Ma. Rodinia's breakup (Heaman et al. 1992; Evans

1998; Buchan et al. 2001; Shelnutt et al. 2004) occurred after hundreds of millions of years of the supercontinent's existence. The mantle upwelling may cause a mass imbalance, resulting in a geoid that is increasingly prolate with a long axis that would approximate the centre of the supercontinent, thus inducing the rotational axis of the Earth to swing to a position normal to the long axis. This could cause the centre of the supercontinent to move relatively quickly to a position closer to the equator, producing paleomagnetic shifts of over 20° in less than 5 m.y.

(e.g., Evans 1998, Prevot et al. 2000), which is on the order of the pole difference found in the

Franklin intrusions, though the direction of that difference is impossible to establish without 157 precise ages assigned to each pole. The effect of this mass shift caused by Rodinia has been postulated as the cause of TPW during the time of Pangea due to Rodinia's "geoid legacy"

(Evans 1998), implying that the mass imbalance remained in effect for about 400 m.y. If

Neoproterozoic TPW is assumed, the driving force may be the "geoid legacy" of the proposed

Mesoproterozoic supercontinent Nuna (also known as Columbia) that is thought to have preceded it, about 400 m.y. before Rodinia's assembly (e.g., Rogers & Santosh 2002), and/or breakup of Rodinia 400 m.y. later. This hypothesis is compatible with the observation that

Laurentia, at the core of Rodinia, was at equatorial latitudes at the time of Rodinia's breakup

(Buchan et al. 2001; this study), as was the centre of Jurassic Pangea (Powell et al. 1993). While the Neoproterozoic paleomagnetic record for Laurentia is poor, the chronologically nearest pole is for the Hottah Gabbro and associated intrusions of western North America at 9°N, 136°E at

780 Ma (Buchan et al. 2001) (Figure 4.8), located to the west of Pole A. If poles A, B, C, and E define a unidirectional apparent polar wander path, this implies that pole A is the oldest, and E the youngest.

4) The fundamental assumption of paleomagnetism that the Earth's magnetic field in

Neoproterozoic time behaved as a geocentric axial dipole may prove untrue. Given that more than ten dykes were analyzed in each of the studies considered here, the continuous emplacement of Franklin intrusions over ca. 10 million years, and as the angular dispersion of the site mean directions is about 15°, consistent with that of equatorial VGP dispersion (Merrill & McElhinny

1983), secular variation in the Earth's magnetic field is considered to have been averaged out for each study. The effect of a non-dipole field on Phanerozoic rocks has been observed in other studies (e.g., Kent & Smethurst 1998; Torsvik et al. 2001) which indicate that non-dipole components can account for up to 10% of the Earth's field, and can offset the magnetic north 158 pole from geographic north by as much as 15° over short periods (Van der Voo & Torsvik 2001).

Rapid polar motions have also been tied to periods approaching magnetic reversals (e.g.,

Courtillot & Besse 1987), and the data presented here do include a record of at least one reversal, as both the older Franklin swarm and the younger Clarence Head swarm include both normal- and reverse-polarity magnetizations. These latter hypotheses, or combinations of them, to explain the difference in poles from the Franklin intrusions are difficult to evaluate given the paucity of reliable paleopoles in the Proterozoic compared to the much more complete rock record in the Phanerozoic, though the positions of the poles indicate that rapid, reversible (as the pole appears to return to previous positions after the high-velocity "swings" (Figure 4.8)), and non-random shifts in the geomagnetic pole or in the case of TPW, the lithosphere, would need to be invoked to consider these explanations plausible.

While other explanations cannot be conclusively dismissed, the simplest explanation for the difference in the paleomagnetic poles obtained from the Franklin intrusions is that they record the drift of Laurentia over approximately 10 m.y. as the supercontinent Rodinia breaks up.

If the poles A-C are interpreted as older than Pole E, based on the age of Heaman et al. (1992) of

723 +4/-2 Ma for Franklin intrusions of Victoria Island and Baffin Island, then Poles A, B, C, and E define a segment of Laurentia's apparent polar wander path (APWP), on a consistent track from the nearest key pole of the Hottah mafic magmatism at 780 Ma (Figure 4.8) through the

721 Ma Pole E. However, a more refined geochronology is required for verification. Pole A includes data from the NNW-trending Strathcona Sound dyke swarm (Figure 4.13), which has been observed to crosscut the NW-trending main Franklin swarm (Christie & Fahrig 1983). If, as indicated by their trend, the Strathcona Sound dykes are coeval with the Clarence Head dyke swarm, the APWP would then be in a westerly direction from Pole E to Pole A, towards the 780 159

Ma pole rather than away fromit , indicating a more complicated and possibly unrealistic history of plate motion. However, the trend of the Strathcona Sound dykes may be fault-controlled

(Christie & Fahrig 1983; Pehrsson & Buchan 1999; Jackson 2000), and as these dykes have not been dated, their age relationship (and that of their paleopole) remains speculative. Pehrsson &

Buchan (1999) report a U/Pb (baddeleyite) age for a NW-trending Franklin dyke on Baffin

Island of 716 +4/-5 Ma, which would suggest that the Strathcona Sound dykes are younger than this age (and therefore possibly related to the Clarence Head dykes of Ellesmere Island and

Devon Island), but this remains to be confirmed directly. The nearest age of Laurentia magmatism that postdates the Franklin is that of the Long Range dykes, dated to 615 Ma (U-Pb baddeleyite, Kamo & Gower 1994), which are more than 100 m.y. distant in age, and with a poorly-resolved paleomagnetic pole (e.g., Hodych et al. 2004). 5. Conclusions

The samples acquired for this study are valuable, as they are from a very remote location which limits the number of samples that can be collected, but due to this lack of material the quality of the data is necessarily also limited. The analysis of more samples from more rock units might have enhanced the strength of conclusions that can be drawn. Nevertheless, this study, an integration of paleomagnetic, geochemical, and geochronological methods applied to

Proterozoic dyke swarms about which very little was previously known, has resulted in significant findings and has opened new avenues for future work.

The Nares Strait Problem is not conclusively resolved, but the location of the Wegener

Fault is now constrained in the southern part of Nares Strait by the successful correlation of the apparently offset parts of the Franklin dyke swarm in the area. This correlation, the relative density of the sets of dykes, the correlation of the age boundaries in the bedrock of Arctic

Canada and Greenland, and the magnitude and direction of offset seen in the paleopoles of the

Franklin dykes, provide new lines of evidence in support of the presence of a sinistral fault with

~200km of displacement. New evidence is also provided for the placement of the plate boundary between the Greenland plate and the North American plate, along the east coast of

Ellesmere Island as seen in Figure 4.4. This correlation of dykes, and the establishment that the

Kap Leiper dyke cannot be considered to continue into Ellesmere Island, would preclude placing the boundary within Ellesmere, as some authors would suggest, keeping cratonic Ellesmere and

Devon Islands relatively fixed with respect to Greenland during the spreading of Baffin Bay and

Labrador Sea that started 95 million years ago. While there are outstanding issues in the development of an integrated and complete tectonic and geological framework for the High

Arctic region, this work and recently-published studies contribute to a fuller understanding of a 161 complex tectonic history. This history involves extensive intraplate deformation and in general, a rejection of the traditional view of strictly rigid, well-defined plates. Unresolved issues include the magnitude of Eurekan shortening, the original geometry of the Paleoproterozoic sedimentary basins, the precise boundaries of the Baffin Bay rift system at its northwestern extent near Devon

Island, and a precise accounting of how much strike-slip motion was taken up within the North

American plate. A more precise age for the Archean basement underlying Devon Island would also contribute to a possible correlation with the Archean basement of the Inglefield Bay area of

Greenland.

This study represents the first time that dykes on Devon and Ellesmere Island have been definitively assigned to the greater Franklin dyke swarm, and also reveals that the Franklin magmatic event was long-lived and generated intrusions over 2500 km away over at least 10 million years. Changing emplacement styles over time may reflect changes in prevailing stress fields, caused by either a new direction of extension created by the collapse of a depleting mantle plume head, or more locally, by release of horizontal compressive stress by earlier dyke injection. This is also the first time that intrusions orthogonal to the main swarm (Clarence Head swarm) have been precisely dated, and the first precise dating of a Thule sill. These new ages, when combined with the paleomagnetic data, record and date at least one reversal of the Earth's magnetic field in the Neoproterozoic, as there are N- and R-polarity dykes in both the older

Franklin swarm and the younger Clarence Head swarm. Some of the Clarence Head dykes have been chemically remagnetized by the precipitation of pyrrhotite during the Devonian Ellesmerian orogeny. The paleopole obtained for the Franklin dykes is distinctly different from that previously established as characteristic of the Franklin intrusions. By successfully showing that the magnetization measured is primary via a baked-contact test, and by obtaining precise U-Pb ages from the same dykes in the paleomagnetic study, the paleopole presented here should be

considered the key pole for the Franklin intrusions for use in plate reconstructions at 721 Ma.

This is an important time in Earth history as it represents the time of the breakup of Rodinia, and

the time of the proposed Sturtian "snowball Earth" global glaciation, a current controversy in

Earth sciences that can only be resolved by high-quality paleomagnetic and geochronological

data. The difference in the poles is most likely a result of the accelerated motion of the

Laurentian paleoplate during the breakup of Rodinia, though a higher-resolution geochronology

is required to prove this conclusion.

The Dundas Harbour dyke age of 1337 Ma represents a newly-established age of Arctic magmatism, one that might prove to be more extensive if the Nigarfivik dykes of northwest

Greenland and the proposed Mesoproterozoic dykes of southern Devon Island can be dated using the U-Pb method. Should their ages agree, this magmatic event will be an important target for paleomagnetic studies, filling a significant time gap in the Laurentian apparent polar wander path. Despite previous suggestions that it might have a major presence in Canada, the Melville

Bugt swarm has not been found in southern or central Ellesmere Island or Devon Island. The

swarm is here dated at 1622 Ma, within 5 Ma of another Melville Bugt dyke >1000 km to the

south, suggesting rapid emplacement of wide dykes over large distances. The paleomagnetic

direction differs from that previously published for the Melville Bugt dyke to the south, and

constrains paleocontinental reconstructions such as that of Laurentia and Baltica. This new pole

indicates that the two continents were not attached through the Paleoproterozoic, contrary to

recently-published reconstructions (Evans 2007). These new poles fill key gaps in the

Laurentian APWP, and have implications for rapid plate movement and supercontinental

configuration in the Proterozoic. The identification of the two "clusters" of poles in the 163 paleomagnetic record, with indications of relatively fast movement of the plate in intervening time periods and with connections to mantle plume longevity and how mantle plumes impact the lithosphere, leads to an interesting and previously unexplored avenue of research, one that requires a much more complete paleomagnetic record for not only Laurentia but adjacent paleocontinents in order to assess the validity of effects such as true polar wander.

This study also includes the first analysis of the effects of alpha recoil in baddeleyite.

While studies of recoil distances have been carried out previously for zircon, to apply these results to baddeleyite grains of the sizes commonly found in mafic dykes would require corrections to be made to a significant proportion of the published database. However, this study shows that the effect of alpha recoil is not as significant in baddeleyite as it is in zircon, likely due to the higher density of the baddeleyite crystal lattice (density of baddeleyite = 5.75 g/cc; density of zircon = 4.65 g/cc) permitting shorter recoil distances. A more thorough description of the effect should be carried out, using a wider range of grain sizes. This would, however, require the U-Pb analysis of single grains with thicknesses in the vicinity of 5 microns, which may in turn require the refinement of chemical and grain-handling techniques. 6. References

Abrahamsen N., Bengaard H-J., Friderichsen J. D., Van der Voo R. (1997): Paleomagnetism of three dyke swarms in Nansen Land, north Greenland (83° N). Geologie en Mijnbouw 76: 83-95.

Baragar W. R. A., Ernst R. E., Hulbert L., Peterson T. (1996): Longitudinal petrochemical variation in the Mackenzie dyke swarm, northwestern Canadian Shield. Journal of Petrology 37: 317-359.

Bossi J., Campal N., Civetti L., Demarchi G., Girardi V. A. V., Mazzuchelli M., Negrini L., Rivalati G., Fragoso Cesar A. R. S., Sinigoi S., Teixeira W., Piccirillo E. M., Malesini M. (1993) Early Proterozoic dike swarms from western Uruguay: Geochemistry, Sr-Nd isotopes and petrogenesis. Chemical Geology 106: 263-277.

Brown P. E., Parsons I., Becker S. M. (1987): Peralkaline volcanicity in the Arctic basin - the Kap Washington volcanics, petrology and palaeotectonics. Journal of the Geological Society (London) 144: 707-715.

Brozena J. M., Childers V. A., Lawver L. A., Gahagan L. M., Forsberg R., Faleide J. I., Eldholm O. (2003): New aerogeophysical study of the Eurasia Basin and Lomonosov Ridge: Implications for basin development. Geology 31: 825-828.

Buchan K. L., Ernst R. E. (2006): Giant dyke swarms and the reconstructions of the Canadian Arctic islands, Greenland, Svalbard, and Franz Josef Land. In: Hanski E., Mertanen S., Ramo T, Vuollo J. (eds.): Dyke swarms: Time markers of crustal evolution. A. A. Balkema, Rotterdam: 27-48.

Buchan K. L., Ernst R. E. (2004): Diabase dyke swarms and related rocks in Canada and adjacent regions. Scale 1:5000000. Geological Survey of Canada Map 2022A.

Buchan K. L., Ernst R. E., Hamilton M. A., Mertanen S., Pesonen L. J., Elming S-A. (2001): Rodinia: the evidence from integrated paleomagnetism and U-Pb geochronology. Precambrian Research 110: 9-32.

Buchan K. L., Mertanen S., Park R. G., Pesonen L. J., Elming S-A., Abrahamsen N., Bylund G. (2000): Comparing the drift of key paleomagnetic poles. Tectonophysics 319: 167-198.

Buchan K. L., Hodych J. P., Roddick J. C, Emslie R. F., Hamilton M. A. (1996): Paleomagnetism and U-Pb geochronology of Mesoproterozoic dykes of Labrador and correlations with dykes of southwest Greenland. In: Proterozoic Evolution on the North Atlantic Realm (compiled by C. F. Gower). COPENA-ECSOOT-IBTA conference Program and Abstracts: 37. 165

Buchan K. L., Mortensen J. K., Card K. D. (1993): Northeast-trending Early Proterozoic dykes of southern Superior Province: multiple episodes of emplacement recognized from integrated paleomagnetism and U-Pb geochronology. Canadian Journal of Earth Sciences 30: 1286-1296.

Buchan K. L., Halls H. C. (1990): Paleomagnetism of Proterozoic mafic dyke swarms of the Canadian Shield. In: Parker A. J., Rickwood P. C, Tucker D. H. (eds.) Mafic dyke swarms and emplacement mechanisms. A. A. Balkema, Rotterdam: 209-230.

Butler R F. (1992) Paleomagnetism: Magnetic domains to geologic terranes. Blackwell Scientific Publications. 319 pp.

Cadman A. C, Heaman L., Tarney J., Wardle R., Krogh T. E. (1993): U-Pb geochronology and geochemical variation within two Proterozoic mafic dyke swarms, Labrador. Canadian Journal of Earth Sciences 30: 1490-1504.

Chalmers J. A. (1991): New evidence on the structure of the Labrador Sea/Greenland continental margin. Journal of the Geological Society, London 148: 899-908.

Chalmers J. A., Laursen K. H. (1995): Labrador Sea: the extent of continental and oceanic crust and the timing of the onset of seafloor spreading. Marine and Petroleum Geology 12: 205-217.

Chalmers J. A., Pulvertaft T. C. R. (2001) Development of the continental margins of the Labrador Sea: a review. In Wilson R. C. L., Whitmarsh R. B., Taylor B., Froitzheim N. (eds.) Non-volcanic rifting of continental margins: A comparison of evidence from land and sea. Geological Society (London), Special Publication 187: 77-105.

Chian D., Keen C, Reid I., Louden K. E. (1995): Evolution of nonvolcanic rifted margins: New results from the conjugate margins of the Labrador Sea. Geology 23: 589-592.

Collinson D. W. (1983): Methods in Rock Magnetism and Palaeomagnetism: Techniques and Instrumentation. Chapman & Hall, London, 503 pp.

Connelly J. N., Thrane K., Krawiec A. W., Garde A. A. (2006): Linking the Palaeoproterozoic Nagssugtoqidian and Rinkian orogens through the Disko Bugt region of West Greenland. Journal of the Geological Society, London 162: 1-17.

Damaske D., Oakey G. N. (2006): Volcanogenic sandstones as aeromagnetic markers on Judge Daly Promontory and in Robeson Channel, northern Nares Strait. Polarforschung 74: 9-19.

Davis, D. W. (1982): Optimum linear regression and error estimation applied to U-Pb data. Canadian Journal of Earth Sciences 19: 2141-2149.

Davis D. W., Sutcliffe R. H. (1985): U-Pb ages from the Nipigon plate, and northern Lake Superior. Geological Society of America Bulletin 96: 1572-1579. 166

Dawes P. R. (2007): Intact Precambrian craton and Smith Sound (Nares Strait): Linchpin in paleogeographic reconstructions of the North Atlantic - Arctic. Geological Association of Canada/Mineralogical Association of Canada Joint Meeting, Yellowknife 2007 Program with Abstracts, p. 20.

Dawes P. R. (2006): Explanatory notes to the Geological map of Greenland, 1:500 000, Thule, Sheet 5 (1991). Geological Survey of Denmark and Greenland Map Series. 97pp.

Dawes P. R. (2004): Explanatory notes to the Geological map of Greenland, 1:500 000, Humboldt Gletscher, Sheet 6 (2004). Geological Survey of Denmark and Greenland Map Series. 48pp.

Dawes P. R. (1998): Geological Map of Greenland, 1:500 000, Thule, sheet 5. Copenhagen: Geological Survey of Denmark and Greenland.

Dawes P. R. (1997): The Proterozoic Thule Supergroup, Greenland and Canada: history, lithostratigraphy and development. Geology of Greenland Bulletin 174: 150 pp.

Dawes P. R. (1991): Geological compilation map of the Thule area, Greenland, scale 1: 500 000. Geological Survey of Greenland publication.

Dawes P. R., Garde A. A. (2004): Geological Map of Greenland, 1:500 000, Humboldt Gletscher, sheet 6. Copenhagen: Geological Survey of Denmark and Greenland.

Dawes P. R., Frisch T., Garde A. A., Ianelli T. R., Inoson J. R., Jensen S. M., Pirajno F., Sonderholm M., Stemmerik L., Stouge S., Thomassen B., van Gool J. M. (2000): Kane Basin 1999: mapping, strati graphic studies and economic assessment of Precambrian and lower Palaeozoic provinces in north-western Greenland. Geology of Greenland Survey Bulletin 186: 11-28.

Dawes P. R., Larsen O., Kalsbeek F. (1988): Archean and Proterozoic crust in North-West Greenland: evidence from Rb-Sr whole-rock age determinations. Canadian Journal of Earth Sciences 25: 1365-1373.

Dawes P. R., Rex D. C. (1986): Proterozoic basaltic magmatic periods in Northwest Greenland: Evidence from K-Ar ages. Gronlands Geologiske Undersagelse Rapport 130.

Dawes P. R., Frisch T., Christie K. L. (1982a): The Proterozoic Thule Basin of Greenland and Ellesmere Island: importance to the Nares Strait debate. In: Dawes P. R., Kerr J. W. (eds.) Nares Strait and the drift of Greenland: A conflict in plate tectonics. Meddelelser om Granland Geoscience 8: 89-104

Dawes P. R., Peel, J. S., Rex, D. C. (1982): The Kap Leiper basic dyke and the age of dolerites of Inglefield Land, North-West Greenland. Rapport Gronlands Geologiske Undersogelse 110: 14-19. 167

Dawes P. R., Kerr J. W. (eds.) (1982a): Nares Strait and the drift of Greenland: A conflict in plate tectonics. Meddelelser om Granland Geoscience 8. 392 pp.

Dawes P. R., Kerr J. W. (1982b): The case against major displacement along Nares Strait. In: Dawes P. R., Kerr J. W. (eds.) Nares Strait and the drift of Greenland: A conflict in plate tectonics. Meddelelser om Gronland Geoscience 8: 369-386.

Dekkers M. J. (1989): Magnetic properties of natural pyrrhotite II: High and low temperature behaviour of JK and TRM as a function of grain size. Physics of the Earth and Planetary Interiors 57:266-283.

Denyszyn S. W., Halls H. C, Davis D. W. (2006): A paleomagnetic, geochemical and U-Pb geochronological comparison of the Thule (Greenland) and Devon Island (Canada) dyke swarms and its relevance to the Nares Strait problem. Polarforschung 74: 63-75.

Denyszyn S. W., Halls H. C. (2004): Paleomagnetic, geochemical and U-Pb geochronological correlation of Proterozoic dykes between Greenland and Canada, and its bearing on the Nares Strait Problem. American Geophysical Union, Spring Meeting May 2004, Montreal, Canada. Abstract GP33B-05.

De Paor D. G., Bradley D. C, Eisenstadt G., Phillips S. M. (1989): The Arctic Eurekan orogen: A most unusual fold-and-thrust belt. Geological Society of America Bulletin 101: 952-967.

Dewing K., Harrison J. C, Pratt B. R., Mayr U. (2004): A probable late Neoproterozoic age for the Kennedy Channel and Ella Bay formations, northeastern Ellesmere Island and its implications for passive margin history of the Canadian Arctic. Canadian Journal of Earth Sciences 41: 1013-1025.

Dickin A. P. (2005): Radiogenic isotope geology (2nd ed.). Cambridge University Press: Cambridge. 492 pp.

Dunlop D. J., Ozdemir 6. (2001): Rock magnetism: Fundamentals and frontiers. Cambridge University Press: Cambridge. 573 pp.

England J. (1987): Glaciation and the evolution of the Canadian high arctic landscape. Geology 15:419-424.

Ernst R. E., Buchan K. L., Harris J. A., Milner M. W., La Prairie L. F. (2004): 40Ar-39Ar dating of the Lasard River mafic dykes, and implications for the focus of the 0.72 Ga Franklin Large Igneous Province of northern Canada. American Geophysical Union, Spring Meeting May 2004, Montreal, Canada. Program with Abstracts.

Ernst R. E., Buchan K. L. (2001) Large mafic magmatic events through time and links to mantle- plume heads, in Ernst R. E., Buchan K. L. (eds.) Mantle plumes: Their identification through time. GSA Special Paper 352: 483-575. 168

Ernst R. E., Head J. W., Parfitt E., Grosfils E., Wilson L. (1995) Giant radiating dyke swarms on Earth and Venus. Earth Science Reviews 39:1-58.

Escher A., Watt W. S. (eds.) (1976): Geology of Greenland. Geological Survey of Greenland, Copenhagen. 603 pp.

Estrada S., Hohndorf A., Henjes-Kunst F. (2001): Cretaceous-Tertiary volcanism in north Greenland: the Kap Washington Group. Polarforschung 69: 17-23.

Estrada S. (2000): Basaltic dykes in the Kap Washington and Frigg Fjord areas (north Greenland). Polarforschung 68: 19-23.

Evans, D. A. D. (2007): Growth of Laurentia from a paleomagnetic perspective: Critical review and update. Geological Association of Canada/Mineralogical Association of Canada Joint Meeting, Yellowknife 2007 Program with Abstracts, p. 26.

Everitt, C. W. F., Clegg, J. A. (1962): A field test of paleomagnetic stability. Geophysical Journal (London) 6: 312-319.

Fahrig W. F. (1985): The tectonic setting of continental mafic dyke swarms: failed arm and early passive margin. In: Mafic Dyke Swarms, H.C. Halls and W.H. Fahrig (editors), Geological Association of Canada Special Paper 34: 331-348.

Fahrig W. F., Christie K. W., Jones D. L. (1981): Paleomagnetism of the Bylot Basins: evidence for Mackenzie continental tectonics. In: Campbell, F. H. A. (ed.), Proterozoic Basins in Canada. Geological Survey of Canada Paper 81-10, pp. 303-312.

Fahrig W. F., Schwarz E. J. (1973): Additional paleomagnetic data on the Baffin diabase dykes and a revised Franklin pole. Canadian Journal of Earth Sciences 10: 576-581.

Fahrig W. F., Irving E., Jackson G. D. (1971): Paleomagnetism of the Franklin diabases. Canadian Journal of Earth Sciences 8: 455-467.

Fisher R. (1953): Dispersion on a sphere. Proceedings of the Royal Society of London, Series A 217: 295-305.

Frisch T., Dawes P. R. (1982): The Precambrian shield of northernmost Baffin Bay: correlations across Nares Strait. In: Dawes P. R., Kerr J. W. (eds.) Nares Strait and the drift of Greenland: A conflict in plate tectonics. Meddelelser om Granland Geoscience 8: 79-88.

Frisch T., Hunt P. (1988): U-Pb zircon and monazite ages from the Precambrian Shield of Ellesmere and Devon Islands, Arctic Archipelago. GSC Paper 88-2: 79-88.

Frisch T., Trettin H. P. (1991): Precambrian successions in the northernmost part of the Canadian Shield. In: Trettin H. P. (ed.) Geology of the Innuitian Orogen and Arctic Platform of Canada and Greenland. Geological Survey of Canada, Geology of Canada 3: 103 -108. 169

Frisch T. (1984a): Geology, Devon Ice Cap, District of Franklin, Northwest Territories. Geological Survey of Canada, map 1574A, scale 1: 250 000.

Frisch T. (1984b): Geology, Mackinson Inlet, District of Franklin, Northwest Territories. Geological Survey of Canada, map 1573 A, scale 1: 250 000.

Frisch T. (1984c): Geology, Prince of Wales Mountains, District of Franklin, Northwest Territories. Geological Survey of Canada, map 1572A, scale 1:250 000.

Frisch T. (1988): Reconnaissance geology of the Precambrian shield of Ellesmere, Devon and Coburg Islands, Canadian Arctic Archipelago. GSC Memoir 409,102 pp.

Funck T., Dehler S. A., Jackson H. R., Salisbury M. H., Reid I. D. (2006a): A refraction seismic image of the sediments in Kennedy Channel, northern Nares Strait. Polarforschung 74: 41-50.

Funck T., Jackson H. R., Dehler S. A., Reid I. D. (2006b): A refraction seismic transect from Greenland to Ellesmere Island, Canada: The crustal structure in southern Nares Strait. Polarforschung 74: 97-112.

Garde A. A. (1994): Precambrian geology between Qarajaq Isfjord and Jakobshavn Isfjord, West Greenland. Grenlands Geologiske Unders0gelse, scale 1:500 000.

Geoffroy L., Callot J-P., Scaillet S., Skuce A., Gelard J. P., Ravilly M., Angelier J., Bonin B., Cayet C, Perrot K., Lepvrier C. (2001): Southeast Baffin volcanic margin and the North American-Greenland plate separation. Tectonics 20: 566-584.

Grantz A., Clark D. L., Phillips R. L., Srivastava S. P. (2001): Phanerozoic stratigraphy of Northwind Ridge, magnetic anomalies in the Canada basin, and the geometry and timing of rifting in the Amerasia basin, Arctic Ocean. GSA Bulletin 110: 801-820.

Grist A. M., Zentilli M. (2005): The thermal history of the Nares Strait, Kane Basin and Smith Sound region in Canada and Greenland: Constraints from apatite fission track and (U-Th-Sm)/He dating. Canadian Journal of Earth Sciences 40: 1547-1569.

Grocott J., Pulvertaft T. C. R. (1990): The early Proterozoic Rinkian belt of central West Greenland. In: Lewry J. F., Stauffer M. R. (eds): The early Proterozoic Trans-Hudson Orogen of North America. Geological Association of Canada Special Paper 37: 443-463.

Hald N., Tegner C. (2000): Composition and age of Tertiary sills and dykes, Jameson Land Basin, East Greenland: relation to regional flood volcanism. Lithos 54: 207-233.

Halls H. C. (1976): A least-squares method to find a remanence direction from converging remagnetization circles. Geophysical Journal of the Royal Astronomical Society 45: 297-304. 170

Halls H. C, Kumar A., Srinivasan R., Hamilton M. A. Paleomagnetism and U-Pb geochronology of easterly trending dykes in the Dharwar craton, India: feldspar clouding, radiating dyke swarms and the position of India at 2.37 Ga. Submitted to Precambrian Research.

Halls H. C, Davis D. W., Stott G. M., Ernst R. E., Hamilton M. A. (in press): The Paleoproterozoic Marathon large igneous province: New evidence for a 2.1 Ga long-lived mantle plume along the southern margin of the North American Superior Province. In press in Precambrian Research.

Halls H. C., Davis D. W. (2004): Paleomagnetism and U-Pb geochronology of the 2170 Ma Biscotasing dyke swarm, Ontario, Canada: evidence for vertical-axis rotation across the Kapuskasing Zone. Canadian Journal of Earth Sciences 41: 255-269.

Hamilton M. A., Buchan K. L. (2007): U-Pb geochronology of the Western Channel Diabase, Wopmay orogen: implications for the APWP for Laurentia in the earliest Mesoproterozoic. Geological Association of Canada/Mineralogical Association of Canada Joint Meeting, Yellowknife 2007 Program with Abstracts.

Hamilton M. A., Buchan K. L., Garde A. A., Connelly J. N. (2004): U-Pb age and preliminary paleomagnetism of a Melville Bugt diabase dyke, West Greenland, and implications for Mid- Proterozoic Laurentia-Baltica reconstructions. Program with Abstracts, AGU-CGU-SEG-EEGS Joint Assembly.

Hamilton M.A., Davis D.W., Buchan K.L., Halls, H.C. (2002): Precise U-Pb dating of reversely magnetized Marathon diabase dykes and implications for emplacement of giant swarms along the southern margin of the Superior Province, Ontario. Geological Survey of Canada Current Research 2002-F6, 8 pp.

Harrison J. C. (2006): In search of the Wegener fault: re-evaluation of strike-slip displacements along and bordering Nares Strait. Polarforschung 74: 129-160.

Harrison J. C, Brent T. A., Oakey G. N. (2006): Overview of the bedrock geology of Nares Strait region, Arctic Canada and Greenland. Geological Survey of Canada Open File 5278.

Harrison J. C, Damaske D., Mayr U., Nowlan G. S., Oakey G. (2003): Re-evaluation of Wegener Fault displacements. Program with Abstracts: 4th International Conference on Arctic Margins.

Harrison J. C, Mayr U., McNeil D. H., Sweet A. R., Mclntyre D. J., Eberle J. J., Harington C. R., Chalmers J. A., Dam G., Nohr-Hansen H (1999) Correlation of Cenozoic sequences of the Canadian Arctic region and Greenland; implications for the tectonic history of northern North America. CSPG Bulletin 47, 3: 223-254.

Heaman L. M., LeCheminant A. N. (1993): Paragenesis and U-Pb systematics of baddeleyite (Zr02). Chemical Geology 110: 95-126. 171

Heaman L. M., LeCheminant A. N., Rainbird R. H. (1992): Nature and timing of Franklin igneous events, Canada: Implications for a Late Proterozoic mantle plume and the break-up of Laurentia. Earth and Planetary Science Letters 109: 117-131.

Higgins A. K., Soper N. J. (1989): Short paper: Nares Strait was not a Cenozoic plate boundary. Journal of the Geological Society of London 146: 913-916.

Hodych J. P., Cox R. A., Kosler J. (2004): An equatorial Laurentia at 550 Ma confirmed by Grenvillian inherited zircon dated by LA-ICP-MS in the Skinner Cove volcanics of western Newfoundland: implications for inertial interchange true polar wander. Precambrian Research 129:93-113.

Hoffman P. F., Kaufman A. J., Halverson G. P., Schrag D. P. (1998): A Neoproterozoic snowball Earth. Science 28: 1342-1346.

Hofmann H. J., Jackson G. D. (1996): Notes on the geology and paleontology of the Proterozoic Thule Group, Ellesmere Island, Canada, and North-west Greenland. Geological Survey of Canada Bulletin 495: 1-26.

Hunt C. P., Moskowitz B. M., Banerjee S. K. (1995): Magnetic properties of rocks and minerals. In: Clauser C, Huenges E. (eds.): Rock Physics and Phase Relations: A Handbook of Physical Constants. AGU Reference Shelf 3: 189-204.

Irving E. (1964): Paleomagnetism and its application to geological and geophysical problems. Juhn Wiley & Sons: New York. 399 pp.

Irving E., Baker J., Hamilton M., Wynne P. J. (2004): Early Proterozoic geomagnetic field in western Laurentia: implications for paleolatitudes, local rotations, and stratigraphy. Precambrian Research 129: 251-270.

Irving E., Donaldson J. A., Park J. K. (1972): Paleomagnetism of the Western Channel Diabase and related rocks, Northwest Territories. Canadian Journal of Earth Sciences 9: 960-971.

Jackson G. D., Iannelli T. R. (1981): Rift-related cyclic sedimentation in the Neohelikian Borden basin, northern Baffin Island. In: Campbell, F. H. A. (ed.), Proterozoic Basins in Canada. Geological Survey of Canada Paper 81-10, pp. 269-302.

Jackson G. D., Hunt P. A., Loveridge W. D., Parrish R. R. (1990): Reconnaissance geochronology of Baffin Island, N.W.T. GSC Paper 89-2: 123-148.

Jackson G. D. (2000): Geology of the Clyde-Cockburn Land map area, north-central Baffin Island, Nunavut. GSC Memoir 440: 303 pp.

Jackson H. R., Harmon T., Neben S., Piepjohn K., Brent T. (2006): Seismic reflection profiles from Kane to Hall Basin, Nares Strait: Evidence for faulting. Polarforschung 74: 21-39. Jackson H. R., Dickie K., Marillier F. (1992): A seismic reflection study of northern Baffin Bay: Implications for tectonic evolution. Canadian Journal of Earth Science 29: 2353-2369.

Jackson H. R., Koppen L. (1985): The Nares Strait gravity anomaly and its implications for crustal structure. Canadian Journal of Earth Sciences 22: 1322-1328.

Jackson H. R., Keen C. E., Falconer R. K. H., Appleton K. P. (1979) New geophysical evidence for seafloor spreading in central Baffin Bay. Canadian Journal of Earth Science 16: 2122-2135.

Jackson K. C, Halls H. C. (1988): Tectonic implications of paleomagnetic data from sills and dykes in the Sverdrup basin, Canadian Arctic. Tectonics 7: 463-481.

Jaffey, A. H., Flynn, K. F., Glendenin, L. E., Bentley, W. C, and Essling, A. M. (1971): Precision measurements of half lives and specific activities of U and U. Physics Reviews C4: 1889-1906.

Jakobsson M., Cherkis N. Z., Woodward J., Macnab R., Coakley B. (2000): New grid of Arctic bathymetry aids scientists and mapmakers. Eos Transactions 81: 89.

Jefferson C. W., Hulbert L. J., Rainbird R. H., Hall G. E. M., Gregoire D. C, Grinenko L. I. (1994): Mineral resource assessment of the Neoproterozoic Franklin igneous events of Arctic Canada: Comparison with the Permo-Triassic Noril'sk-Talnakh Ni-Cu-PGE deposits of Russia. Geological Survey of Canada Open File Report 2789.

Jensen L. S. (1976): A new cation plot for classifying subalkalic volcanic rocks. Ontario Geological Survey Miscellaneous Paper 66, 22 p.

Johnson G. L., Srivastava S. P. (1982): The case for major displacement along Nares Strait. In: Dawes P. R., Kerr J. W. (eds.) Nares Strait and the drift of Greenland: A conflict in plate tectonics. Meddelelser om Grenland Geoscience 8: 364-368.

Jourdan F., Feraud G., Bertrand H., Watkeys M. K., Kampunzu A. B., Le Gall B. (2006): Basement control on dyke distribution in Large Igneous Provinces: Case study of the Karoo triple junction. Earth and Planetary Science Letters 241: 307-322.

Kah L. C, Sherman A. G., Narbonne G. M, Knoll A. H., Kaufman A. J. (1999): 513C stratigraphy of the Proterozoic Bylot Supergroup, Baffin Island, Canada: implications for regional lithostratigraphic correlations. Canadian Journal of Earth Sciences 36: 313-332.

Kalsbeek F., Pulvertaft T. C. R., Nutman A. P. (1998): Geochemistry, age and origin of metagreywackes from the Palaeoproterozoic Karrat Group, Rinkian Belt, West Greenland. Precambrian Research 399:383-399.

Kalsbeek F., Taylor P. N. (1986): Chemical and isotopic homogeneity of a 400 km long basic dyke in central West Greenland. Contributions to Mineralogy and Petrology 93: 439-448. 173

Kalsbeek F. (1986): The tectonic framework of the Precambrian shield in Greenland: a review of new isotopic evidence. Geological Survey of Greenland Report 128: 55-64.

Kamo S. L., Gower C. F. (1994): Note: U-Pb baddeleyite dating clarifies age of characteristic paleomagnetic remanence of Long Range dykes, southeastern Labrador. Atlantic Geology 30: 259-262.

Kirschvink J. L. (1980): The least-squares line and plane and the analysis of paleomagnetic data. Journal of the Royal Astronomical Society 62: 699-718.

Krogh, T.E. (1973): A low contamination method for hydrothermal decomposition of zircon and extraction of U and Pb for isotopic age determinations. Geochimica et Cosmochimica Acta 37: 485-494.

Krogh, T.E., Corfu, F., Davis, D.W., Dunning, G.R., Kamo, S.L., Greenough, J., Nakamura, E. (1985): Precise U-Pb isotopic ages of diabase dykes and gabbros using trace baddeleyite. In: Mafic Dyke Swarms, H.C. Halls and W.H. Fahrig (editors), Geological Association of Canada Special Paper 34: 147-152.

Lawver L. A., Grantz A., Gahagan L. M. (2002): Plate kinematic evolution of the present Arctic region since the Ordovician. In: Miller E. L., Grantz A. Klemperer S. L. (eds.) Tectonic Evolution of the Bering Shelf- Chukchi Sea - Arctic Margin and Adjacent Landmasses. GSA Special Paper 360: 333-358.

LeCheminant A. N., Heaman L. M. (1991): U-Pb ages for the 1.27 Ga Mackenzie igneous events, Canada, support for the plume initiation model. Geological Association of Canada/Mineralogical Association of Canada Joint Meeting, Toronto 1989. Program with Abstracts.

LeCheminant A. N., Heaman L. M. (1989): Mackenzie Igneous events, Canada: Middle Proterozoic hotspot magmatism associated with ocean opening. Earth & Planetary Science Letters 96: 38-48.

Lepvrier C. (2000): Drift of Greenland and correlation of Tertiary tectonic events in the West Spitsbergen and Eurekan fold-thrust belts. Polarforschung 68: 93-100.

Ludwig, K.R. (2003): User's manual for Isoplot 3.00, a geochronological toolkit for Excel. Berkeley Geochronological Center Special Publication 4, 71 pp.

Maher H. D. (2001): Manifestations of the Cretaceous High Arctic Large Igneous Province in Svalbard. Journal of Geology 109: 91-104.

Manby G. M., Lyberis N. (1996): State of stress and tectonic evolution of the West Spitsbergen fold belt. Tectonophysics 267: 1-29. McElhinny M. W., McFadden P. L. (2000): Paleomagnetism: Continents and oceans. Academic Press (2nd ed.). 386 pp.

McWhae J. R. H. (1981): Structure and spreading history of the northwestern Atlantic region from the Scotian Shelf to Baffin Bay. in Kerr J. W., Fergusson A. J. (eds.) Geology of the North Atlantic Borderlands. CSPG Memoir 7: 299-332.

Merrill R.T., McElhinny M. W. (1983): The Earth's magnetic field: Its history, origin and planetary perspective. Academic Press, New York. 401 pp

Mertanen S., Pesonen L J. (1995): Palaeomagnetic and rock magnetic investigations of the Sipoo Subjotnian quartz porphyry and diabase dykes, southern Fennoscandia. Physics of the Earth and Planetary Interiors 88: 145-175.

Meschede M. (1986): A method of discriminating between different types of mid-ocean ridge basalts and continental tholeiites with the Nb-Zr-Y diagram. Chemical Geology 56: 207-218.

Miall A. D. (1985) Stratigraphic and structural predictions from a plate-tectonic model of an oblique-slip orogen: The Eureka Sound Formation, Northwest Canadian Arctic islands, in Biddle K. T., Christie-Blick N. (eds.) Strike-slip deformation, basin formation and sedimentation. SEPM Special Publication 37: 361-374.

Miall A. D. (1983): The Nares Strait problem: A re-evaluation of the geological evidence in terms of a diffuse oblique-slip plate boundary between Greenland and the Canadian Arctic Islands. Tectonophysics 100: 227-239.

Mosar J., Eide E. A., Osmundsen P. T., Sommaruga A., Torsvik T. H. (2002): Greenland- Norway separation: A geodynamic model for the North Atlantic. Norwegian Journal of Geology 82: 281-298.

Narbonne G. M., James N. P. (1996): Mesoproterozoic deep-water reefs from Borden Peninsula, Arctic Canada. Sedimentology43: 827-848.

Neben S., Damm V., Brent T., Tessensohn F. (2006): New multi-channel seismic reflection data from Northwater Bay, Nares Strait: Indications for pull-apart tectonics. Polarforschung 74: 77- 96.

Neben S., Brent T. A., Harrison, J. C, Tessensohn F., Jackson H. R. (2003): New marine reflection seismic data in North Baffin Bay, Kennedy Channel and Hall Basin: subsurface constraints on structure and stratigraphy of the Nares Strait region. Program with Abstracts: 4l International Conference on Arctic Margins.

Nielsen T. F. D. (1990): Melville Bugt dyke swarm: A major 1645 Ma alkaline magmatic event in west Greenland. In: Parker A. J., Rickwood P. C, Tucker D. H. (eds.) Mafic dyke swarms and emplacement mechanisms. A. A. Balkema, Rotterdam: 497-505. 175

Nutman A. P, Dawes P. R., Kalsbeek F., Hamilton M. A. (in press): Palaeoproterozoic and Archaean gneiss complexes in northern Greenland: a terrane constellation for the High Arctic. In press in Precambrian Research.

Oakey G. N., Damaske, D. (2006): Continuity of basement structures and dyke swarms in the Kane Basin region of central Nares Strait constrained by aeromagnetic data. Polarforschung 74: 51-62.

Oakey G. N, Damaske D. (2004): Tectonic elements of the Kane Basin Region of Nares Strait constrained by aeromagnetic data. Program with Abstracts: 4th International Conference on Arctic Margins.

Oakey G. N. (2006): Cenozoic evolution and lithosphere dynamics of the Baffin Bay-Nares Strait region of Arctic Canada and Greenland. Ph.D. thesis, Free University of Amsterdam. 233 pp.

Oakey G. N. (1994): A structural fabric defined by topographic lineaments: Correlation with Tertiary deformation of Ellesmere and Axel Heiberg Islands, Canadian Arctic. Journal of Geophysical Research 99, B10: 20311-20321.

Ogg J. G. (1995): Magnetic polarity time scale of the Phanerozoic. In: Ahrens, T. J. (ed.): Global Earth Physics: A Handbook of Physical Constants. AGU Reference Shelf, 1: 240-270.

Okulitch A. V. (1991): Geology of the Canadian Archipelago and North Greenland, Scale 1:2 000 000. In: Trettin H. P. (ed.) Geology of the Innuitian Orogen and Arctic Platform of Canada and Greenland. Geological Survey of Canada, Geology of Canada 3.

Okulitch A. V., Dawes P. R., Higgins A. K., Soper N. J., Christie R. L. (1990): Towards a Nares Strait solution: Structural studies on SE Ellesmere Island and NW Greenland. Marine Geology 93: 369-384.

Osadetz K. (1982): Eurekan structures of the Elkblaw Lake area, Ellesmere Island, Canada. In: Embry A. F., Balkwill H. R. (eds.): Arctic geology and geophysics. Canadian Society of Petroleum Geologists Memoir 8: 219-232.

Park J. K. (1981): Paleomagnetism of basic intrusions from the Brock Inlier, Northwest Territories, Canada. Canadian Journal of Earth Sciences 18: 1637-1641.

Patchett P. J., Embry A. F., Ross G. M., Beauchamp B., Harrison J. C, Mayr U., Isachsen C. E., Rosenberg E. J., Spence G. O. (2004): Sedimentary cover of the Canadian Shield through Mesozoic time reflected by Nd isotopic and geochemical results for the Sverdrup basin, Arctic Canada. Journal of Geology 112: 39-57.

Payne A. V., Baadsgaard H., Burwash R. A., Cumming G. L., Evans C. R., Folinsbee R. E. (1965): A line of evidence supporting continental drift. In: International Union of Geological Sciences' Upper Mantle Symposium, New Delhi: 83-93. 176

Pearce J. A. (1996) A user's guide to basalt discrimination diagrams, in: Wyman D. A. (ed.) (1996) Trace Element Geochemistry of Volcanic Rocks: Applications for Massive Sulphide Exploration. Geological Association of Canada, Short Course Notes 12: 79-113.

Pearce J. A., Cann J. R. (1973): Tectonic setting of basic volcanic rocks determined using trace element analyses. Earth and Planetary Science Letters 19: 290-300.

Pehrsson S. J., Buchan K. L. (1999): Borden dykes of Baffin Island, Northwest Territories: a Franklin U-Pb baddeleyite age and a paleomagnetic reinterpretation. Canadian Journal of Earth Sciences 36: 65-73.

Piepjohn K., Von Gosen W. (2001): Polyphase deformation at the Harder Fjord Fault Zone (north Greenland). Geological Magazine 138: 407-434.

Piepjohn K., Tessensohn F., Harrison C, Mayr U. (1998): Involvement of a Tertiary foreland basin in the Eurekan foldbelt deformation, NW coast of Kane Basin, Ellesmere Island, Canada. Polarforschung 68:101-110.

Piper J. D. A., Steam J. E. F. (1977): Paleomagnetism of the dyke swarms of the Gardar Igneous Province, south Greenland. Physics of the Earth and Planetary Interiors 14: 345-358.

Reid I., Jackson H. R. (1997): A review of three transform margins off Eastern Canada. Geo- Marine Letters 19: 87-93.

Rickwood P. C. (1990): The anatomy of a dyke and the determination of propagation and magma flow directions. In: Parker A. J., Rickwood P. C, Tucker D. H. (eds.) Mafic dyke swarms and emplacement mechanisms. A. A. Balkema, Rotterdam: 81-100.

Roest W. R., Srivastava S. P. (1989): Seafloor spreading in the Labrador Sea: a new reconstruction. Geology 17: 1000-1003.

Rowley D. B., Lottes A. L. (1988): Plate-kinematic reconstructions of the North Atlantic and Arctic: Late Jurassic to Present. Tectonophysics 155: 73-120.

Rubin A. M. (1995): Propagation of magma-filled cracks. Annual Review of Earth and Planetary Sciences 23: 287-336.

Saalmann K., Thiedig F. (2002): Thrust tectonics on Br0ggerhalv0ya and their relationship to the Tertiary West Spitsbergen fold-and-thrust belt. Geological Magazine 139: 47-72.

Shellnutt J. G., Dostal J., Keppie J. D. (2004): Petrogenesis of the 723 Ma Coronation sills, Amundsen basin, Arctic Canada: implications for the break-up of Rodinia. Precambrian Research 129: 309-324. 177

Skaarup N, Jackson H. R., Oakey G. (2006): Margin segmentation of Baffin Bay/Davis Strait, eastern Canada based on seismic reflection and potential field data. Marine and Petroleum Geology 23: 127-144.

Skoog D. A., Holler F. J., Niemann T. A. (1997): Principles of instrumental analysis (5th ed.). Saunders College Publishing: Toronto. 849 pp.

Sobczak L. W., Hearty D. G., Forsberg R., Kristoffersen Y., Eldholm O., May S. D. (1990): Gravity from 64°N to the North Pole. In: Grantz A., Johnson L., Sweeney J. F. (eds) The Arctic Ocean Region. GSA: The Geology of North America, v. L: 101-118.

Soderlund U., L. Johansson (2002): A simple way to extract baddeleyite (Zr02) Geochemistry, Geophysics, Geosystems 3: (2) 10.1029/2001GC000212.

Soper N. J., Higgins A. K. (1990): Models for the Ellesmerian mountain front in North Greenland: a basin margin inverted by basement uplift. Journal of Structural Geology 12: 83-97.

Srivastava S. P., MacLean B., Macnab R. F., Jackson H. R. (1982) Davis Strait: Structure and evoltion as obtained from a systematic geophysical survey, in Embry A. F., Balkwill H. R. (eds.) Arctic Geology and Geophysics. CSPG Memoir 8: 267-278.

Srivastava S. P., Falconer R. K. H., MacLean B. (1981): Labrador Sea, Davis Strait, Baffin Bay: Geology and geophysics - a review. In: Kerr J. W., Fergusson A. J. (eds.) Geology of the North Atlantic Borderlands. CSPG Memoir 7: 333-398.

Srivastava S. P. (1985): Evolution of the Eurasian basin and its implications to the motion of Greenland along Nares Strait. Tectonophysics 114: 29-53.

Stacey, J. D., Kramers, J. D. (1975): Approximation of terrestrial lead isotope evolution by a two-stage model. Earth and Planetary Science Letters 26: 207-221.

Stephenson R. A., Ricketts B. D. (1990) Bouguer gravity anomalies and speculations on the regional crustal structure of the Eurekan orogen, Arctic Canada. Marine Geology 93: 401-420.

St-Onge M. R., Jackson G. D., Henderson, I. (2006): Geology, Baffin Island (south of 70°N and east of 80°W), Nunavut. Geological Survey of Canada Open File 4931, scale 1:500 000.

St-Onge, M. R., van Gool J. A. M., Garde A. A. (2005): Correlation of Archaean to Mesoproterozoic units and structures across Baffin Bay, Davis Strait, and the Labrador Sea: Kangerlussuaq workshop 2005 report and literature review. Danmarks og Granlands Geologiske Undersogelse Rapport 2006/6: 21-57.

Talwani M., Eldholm O. (1977): Evolution of the Norwegian-Greenland Sea. GSA Bulletin 88: 969-999. 178

Tarduno J. (1998): The high Arctic large igneous province. Program with Abstracts: 3r International Conference on Arctic Margins.

Taylor F. B. (1910): Bearing of the Tertiary Mountain Belt on the Origin of the Earth's Plan. GSA Bulletin 21: 179-226.

Tessensohn F., Jackson H. R., Reid I. D. (2006): The tectonic evolution of Nares Strait: implications of new data. Polarforschung 74: 191-198.

Tessensohn F., Piepjohn K. (1998): Eocene compressive deformation in Arctic Canada, North Greenland and Svalbard and its plate tectonic causes. Polarforschung 68: 121-124.

Torsvik T. H., Mosar J., Eide E. A. (2001): Cretaceous-Tertiary geodynamics: a North Atlantic exercise. Geophysical Journal International 146: 850-866.

Torsvik T. H., Smethurst, M. A. (1999): Plate Tectonic modeling: Virtual Reality with GMAP. Computer & Geosciences 25: 395-402.

Torsvik T. H., Smethurst M. A., Meert J. G., Van der Voo R., McKerrow W. S., Brasier M. D., Sturt B. A., Walderhaug H. J. (1996): Continental break-up and collision in the Neoproterozoic and Palaeozoic A tale of Baltica and Laurentia. Earth-Science Reviews 40: 229- 258.

Trettin H. P., Mayr U., Long D. G. F., Packard J. J. (1991): Cambrian to Early Devonian basin development, sedimentation, and volcanism, Arctic Islands. In: Trettin H. P. (ed.) Geology of the Innuitian Orogen and Arctic Platform of Canada and Greenland. Geological Survey of Canada, Geology of Canada 3: 165-238.

Trettin H. P. (1991a): Tectonic Framework. In: Trettin H. P. (ed.) Geology of the Innuitian Orogen and Arctic Platform of Canada and Greenland. Geological Survey of Canada, Geology of Canada 3: 59-66.

Trettin H. P. (1991b): Middle and late Tertiary tectonic and physiographic developments. In: Trettin H. P. (ed.) Geology of the Innuitian Orogen and Arctic Platform of Canada and Greenland. Geological Survey of Canada, Geology of Canada 3: 493-496.

Upton B. G. J., Emelaus C. H., Heaman L. M., Goodenough K. M, Finch A. A. (2003): Magmatism of the mid-Proterozoic Gardar Province, South Greenland: Chronology, petrogenesis and geological setting. Lithos 68: 43-65.

Walker G. P. L. (1999): Volcanic rift zones and their intrusion swarms. Journal of Volcanology and Geothermal Research 94: 21-34.

Weber J. R., Sweeney J. F. (1990): Ridges and basins in the central Arctic Ocean. In: Grantz A., Johnson L., Sweeney J. F. (eds) The Arctic Ocean Region. GSA: The Geology of North America, v. L: 305-336. 179

Wegener A. (1929): The origin of continents and oceans. Transl. by John Biram from the 4 German edition. Dover: New York.

Wynne P. J., Irving E., Osadetz K. (1983): Paleomagnetism of the Esayoo Formation of North Ellesmere Island: Possible clue to the solution of the Nares Strait dilemma. Tectonophysics 100: 241-256.

Zijderveld J. D. A. (1967): A. C. demagnetization of rocks: Analysis of results, in Collinson D. W. (ed.) Methods in Paleomagnetism. Elsevier: Amsterdam, pp. 254-286. 180

Appendix A: Paleomagnetic data by specimen

Specimen numbers and locations as described in Table 3.1. D = Declination; I = Inclination; Nm = Number of remanence measurements contributing to the direction; MAD = Maximum Angle of Deviation from the measured direction (according the program of Kirschvink (1980)). NSEP denotes specimens for which No Stable End Point could be obtained. When multiple specimens from the sample were measured (e.g., BB6-2, BB6-4), the average was taken for the sample direction. Specimen numbers ending with "t" (e.g., BE3-3t) denote specimens that were thermally demagnetized. CM = chilled margin, N, S, E, W, T = north, south, east, west, top.

Sample D I MAD Nm Description

BB: Brae Bay dyke, Devon Island (75°43.102'N, 82°59.232'W), 20 m thick, trend 285° BB1-2 280.8 -22.8 1.0 8 20 cm SCM BB2-2 300.8 -9.8 8.1 5 2 m SCM BB3-2 284.5 -17.8 1.8 6 6 m SCM BB4-1 280.5 -20.7 1.5 7 10 m SCM BB5-3 285.1 -8.3 2.3 4 lmNCM BB6-2 300.1 -20.1 2.0 6 at SCM BB6-4 285.2 -18.9 1.2 7 at SCM BB7-7 280.9 -22.0 1.9 9 atNCM BB7-8 294.5 -21.1 1.9 7 atNCM BB8-6 280.6 -22.0 2.0 9 4cmNCM BB9-3 280.4 -18.2 2.4 8 at SCM

BE: Belcher Glacier dyke, Devon Island (75°34.16'N, 81°25.01'W), 3 m thick, trend 175° BE1-2 NSEP atECM BE2-2 NSEP at ECM BE3-1 101.2 -26.0 3.8 5 2cmWCM BE3-3t 97.7 -29.5 2.9 5 2cmWCM BE4-1 124.3 -31.8 2.4 5 lmWCM BE5-1 104.2 -30.0 4.1 4 atWCM BE6-1 NSEP dykelet, 1 m away from main dyke BE7-2 NSEP 60 cm ECM

BG: Belcher Glacier dyke, Devon Island (75°35.707'N, 80°11.718'W), 75 m thick, trend 270° BG1-2 268.1 43.3 7.8 7 3mNCM BG2-2 272.3 29.4 11.6 5 2mNCM BG3-4 272.9 23.8 3.0 10 60 cm NCM BG4-4 260.4 7.7 5.9 8 lOcmNCM BG5-1 283.3 37.8 2.0 9 18 m NCM BG6-1 270.0 41.1 2.5 5 25 m NCM BG7-3 NSEP 30 m NCM 181

BG8-3 275.6 56.4 2.6 35mNCM

BP: Belcher Point dyke, Devon Island (75°46.088'N, 81°19.397'W), 34 m thick, trend 264c BP1-1 286.5 -25.3 5.4 4 centre of dyke BP2-2 281.5 -3.0 3 5 14mSCM BP3-1 270.0 -1.0 3.8 8 8 m SCM BP4-2 280.0 -2.5 1.7 8 3.5 m SCM BP5-2 274.2 -13.8 2.7 4 1 m SCM BP6-3 286.3 -11.5 1.1 6 3 cm SCM BP7-3 278.8 -6.9 0.6 13 3 cm SCM BP8-2 285.7 -7.6 1.1 8 at SCM BP9-4 270.3 3.9 2.2 9 atNCM

BR: Belcher Glacier dyke, Devon Island (75°28.25N, 81°33.13W), 30 m wide, trend 275c BR1-1 276.7 -10.5 1.6 6 at NCM BR2-1 NSEP 60 cm NCM BR3-1 280.8 -10.5 2.0 6 10 cm NCM BR4-1 287.6 -4.1 4.2 5 60 cm NCM BR4-3t NSEP 60 cm NCM BR5-2 282.5 -5.5 2.3 7 1.5 m NCM BR6-1 NSEP centre of dyke BR7-1 NSEP atNCM BR8-2 280.2 -8.0 1.6 6 at NCM BR8-3t 287.0 -16.9 6.2 7 at NCM BR9-4 286.4 -1.6 3.2 6 20 cm NCM BR10-1 283.6 -12.3 2.4 9 2 cm NCM BR11-1 NSEP 1.5 m NCM BR11-2 NSEP 1.5 m NCM

CA: Carey Islands sill, Greenland (76°44.382'N, 73°13.477'W), >15 m thick, dip 15°N CA1-2 305.0 46.7 5.3 5 2mTCM CA2-1 293.5 13.6 4.9 4 20 cm TCM CA3-2 NSEP 5mTCM CA4-2 295.3 13.6 6.5 9 1.5 m TCM CA5-1 299.4 7.0 2.3 7 5 cm TCM CA6-2 296.1 4.9 2.2 7 10 cm TCM CA-7A NSEP 5 m TCM

CG: Cadogan Glacier dyke, Ellesmere Island (78°17.11N, 77.06.62'W), 25 m thick, trend 280° Margins not accessible, fine-grained float collected for geochem CGl-lt 229.4 -7.8 6.5 4 interior, N side CGI-2 259.3 -1.6 6.3 4 interior, N side CG1-3 247.7 1.9 3.7 5 interior, N side CG2-2 248.0 -0.8 0.7 5 interior, N side CG3-1 257.8 7.2 3.8 5 interior, N side 182

CG4-1 249.4 3.5 3.0 5 interior, N side CG5-2 236.2 7.3 5.1 4 interior, N side (arrow redrawn - cracked) CG6-2 269.3 -6.3 5.3 5 interior, nearest SCM CG6-3 271.1 -5.4 4.0 8 interior, nearest SCM CG7-2 255.8 3.5 3.5 4 centre of dyke

CH: Craig Harbour dyke, Ellesmere Island (76°18.382'N, 80°54.018'W), 60 m thi« 020° CHI-3 240.7 22.2 2.3 5 30mECM CHl-4t NSEP 30mECM CH2-1 NSEP 15mECM CH2-5t NSEP 15mECM CH3-2 NSEP centre of dyke CH4-3 241.1 31.1 6.3 4 1.5mWCM CH5-2 239.4 28.6 4.0 11 4 cm WCM CH6-3 NSEP 3.5mWCM CH7-2 229.5 47.7 4.8 5 8 m WCM

CL: Clarence Head dyke, Ellesmere Island (76°42.95'N, 77°50.97'W), 25 m thick, 355° CL1-1 156.4 11.1 3.0 7 centre of dyke CL2-2 164.7 16.7 4.0 4 centre of dyke CL3-1 169.4 20.3 2.1 4 near WCM CL4-2 163.4 10.2 6.7 6 near ECM CL5-2 164.6 9.2 4.6 8 centre of dyke CL6-1 NSEP 30 cm WCM CL6-2 NSEP 30 cm WCM CL6-3t NSEP 30 cm WCM CL7-2t NSEP 40 cm WCM CL7-3 NSEP 40 cm WCM CL8-2 NSEP at WCM CL8-3t NSEP at WCM CL9-2 169.3 3.7 0.9 5

CM: Cunningham Mountains dyke, Devon Island (74°37.825'N, 81°09.003'W), 25 m thick, trend 220° Cloudy, no sun compass orienting possible CM1-3 284.5 -2.3 3.5 6 40 cm SCM CM2-1 269.6 1.3 2.9 11 5 cm SCM CM3-2 281.2 -7.9 2.5 6 10 cm SCM CM4-2 282.3 -0.9 1.2 7 10 cm SCM CM5-2 270.4 26.4 4.6 4 6 m SCM CM6-2 291.1 16.6 4.0 5 7 m SCM CM7-2 286.5 3.7 3.5 8 2 m SCM 183

CV: Cape Faraday dyke, Ellesmere Island (77°53.863'N, 76°38.837'W), 120 m thick, trend 038° Initially called "Cape Verde", samples labelled CV CV1-1 57.2 -33.1 4.7 11 lmWCM CV2-1 61.7 -36.0 4.1 7 lOcmWCM CV3-1 55.7 -37.6 3.0 5 2 cm WCM CV4-1 NSEP 12mWCM CV5-2 NSEP 47 m WCM CV6-1 NSEP 33 m WCM CV7-2 NSEP 10 cm WCM

CW: Cape Warrender dyke, Devon Island (74°28.583'N, 81°51.863'W), 20 m thick, trend 305° CW1-2 311.9 26.6 2.3 7 atECM CW2-2 332.5 -21.2 7.9 5 atECM CW3-2 NSEP lmECM CW4-4 NSEP 20 cm ECM CW6-2 301.7 -38.8 5.4 6 lmECM CW7-1 NSEP 30 cm ECM CW8-2 280.3 14.7 3.7 6 5 cm ECM CW9-1 286.3 15.8 4.1 6 at WCM CW10-1 NSEP centre of dyke

DH: Dundas Harbour dyke, Devon Island (74°31.106'N, 82°23.497'W), 29 m thick, trend 270° DH1-2 14.9 -1.9 3.1 6 60 cm SCM DH2-2 19.8 -12.5 5.5 4 1.5mSCM DH3-1 NSEP lmSCM DH4-1 22.9 1.8 2.5 at SCM DH5-2 NSEP centre of dyke DH6-2 40.3 -17.4 3.0 5 centre of dyke DH7-1 28.2 18.6 3.5 5 centre of dyke DH8-1 29.8 36.1 9.4 5 centre of dyke DH9-2 NSEP centre of dyke DH10-1 NSEP at SCM DH11-2 NSEP centre of dyke DH12-2 NSEP 6mNCM DH13-2 44.2 -18.6 7.3 5 60 cm NCM DH14-1 14.7 3.5 2.0 6 at SCM

EA: Eastern Glacier dyke, Devon Island (75°49.464'N, 82°04.735'W), 40 m thick, trend 260° EA1-3 261.9 28.1 3.4 11 20 cm SCM EA2-2 270.4 30.1 5.3 6 50 cm SCM EA3-3 270.9 31.4 5.8 7 lmSCM EA4-2 269.0 28.2 2.2 4 2.5 m SCM 184

EA5-3 282.3 41.7 2.4 5 7 m SCM EA6-3 276.4 34.5 4.3 5 10 m SCM EA7-2 268.2 34.8 3.8 7 centre of dyke EA8-3 297.7 28.5 3.7 5 30cmNCM EA9-2 290.8 26.7 4.4 6 interior, N side EA10-2 275.2 36.6 3.1 6 interior, N side EA11-1 270.7 20.9 2.7 9 20 cm SCM

EG: Eastern Glacier dyke, Devon Island (75°44.41'N, 82°04.24'W), 21 m thick, trend 255° EG1-1 281.1 17.1 4.4 14 3 cm NCM EG2-3 280.9 26.3 1.6 10 2 cm NCM EG3-3 NSEP 2 m NCM EG4-4 279.2 24.2 2.0 10 at SCM EG5-3 NSEP 1.2 m SCM EG5-4t NSEP 1.2 m SCM EG6-4 276.5 30.6 3.9 8 10 cm SCM EG7-1 281.1 23.8 0.9 7 at SCM EG8-2 278.6 16.4 1.5 7 at SCM EG9-3 283.5 35.0 1.6 6 centre of dyke

GF: Granville Fjord sill, Greenland (76°51.38'N, 69°55.35'W), 10 m thick, dip 10°S GF1-1 NSEP 7cmTCM GF2-2 NSEP 7 cm TCM GF3-1 294.3 7.4 4.6 6 10 cm TCM GF3-2t NSEP 10 cm TCM GF4-2 288.5 8.3 1.8 8 7 cm TCM GF4-2t NSEP 7 cm TCM (continuation of GF4-2) GF5-2 287.6 14.0 1.0 7 interior GF6-lt 287.9 9.6 3.6 9 interior GF6-2 294.3 18.6 2.2 5 interior GF7-lt 293.7 10.6 5.4 8 interior GF7-2 290.4 11.4 4.3 4 interior GF8-U NSEP interior GF8-2 290.6 9.0 5.2 4 interior GF8-2t NSEP interior (continuation of GF8-2)

GR: Grise Fiord dyke, Ellesmere Island (76°25.923'N, 83°01.283'W), 40 m thick, trend 315° Host rock (anorthosite) used for baked-contact test GR1-2 271.3 18.8 4.5 8 at SCM GR.2-2 270.5 30.5 2.9 14 at SCM GR3-2 275.6 19.6 4.4 7 host at SCM GR4-3 270.0 28.3 2.7 8 40 cm SCM GR5-2 276.4 22.7 1.4 7 at SCM GR6-3 275.3 28.4 2.3 9 3.5 m SCM GR7-2 276.4 21.3 1.3 5 8 m SCM GR8-2 271.6 15.1 2.0 7 centre of dyke 185

GR9-2 271.9 20.9 2.5 13 60cmNCM GR10-2 NSEP host 4 m NCM GR11-2 280.7 20.2 4.7 8 host 40 cm NCM GR12-3 NSEP host5mSCM GR13-5 NSEP host 45 m NCM GR14-1 190.3 18.1 3.2 12 host 45 m NCM GR15-1 NSEP host 35 m NCM GR16-1 175.4 43.5 9.4 5 host 26 m NCM GR17-lt NSEP host 23 m NCM GR17-2 187.8 16.9 4.4 6 host 23 m NCM GR18-2 NSEP host 18 m NCM GR19-2 NSEP host 15 m NCM GR20-2 NSEP host 8 m NCM GR21-2 NSEP host 4.5 m NCM GR22-1 274.7 32.7 4.8 13 host 10 m NCM GR23-2 180.9 31.3 4.6 7 host 45 m NCM GR24-3 NSEP host 45 m NCM GR25-2 188.7 33.4 6.7 7 host35mNCM

HE; Herbert Island dyke, Greenland (77°25.301'N, 70°08.472'W), 21 m thick, trend 100c HE 1-2 287.9 12.5 4.6 7 interior, S side (not included in site average) HE2-1 NSEP interior HE3-1 126.9 -17.5 6.5 7 interior HE4-lt NSEP 3 m NCM HE4-2 96.5 3.8 7.9 6 3 m NCM HE5-1 121.5 29.5 4.9 6 lmNCM HE6-2 NSEP 50 cm NCM HE7-1 NSEP 10 cm NCM HE8-1 NSEP 30 cm NCM HE9-2 NSEP 1 m NCM HE10-2 90.8 -17.2 9.8 5 3 cm NCM (possibly not in place) HE11-2 88.6 65.4 4.8 6 interior, N side

HM: Hope Monument dyke, Devon Island (74°38.253'N, 80°12.600'W), 55 m thick, trend 80° No margins visible, much rubble HM1-2 275.3 15.8 3.7 7 interior, N side (not included in site average) HM2-1 274.7 13.9 3.6 5 interior, N side (not included in site average) HM3-1 262.2 10.6 2.5 5 6 m NCM (not included in site average) HM4-2 NSEP interior HM5-1 NSEP interior, S side HM6-2 240.3 -8.2 3.2 7 <5 m SCM HM7-3 243.5 -3.1 3.7 6 <6 m SCM HM8-1 242.1 -4.0 3.5 8 <4 m SCM HM9-2 250.9 -3.2 2.7 7 <5 m SCM 186

JP2: Josephine Peary Island dyke, Greenland (77°35.639'N, 66°52.980'W), 65 m thick, trend 325° No margins accessible JP2-1-1 208.6 48.2 2.4 6 interior, W side JP2-2-2 NSEP interior, W side JP2-3-2 203.5 48.1 0.5 5 interior, W side JP2-5-2 209.5 44.0 1.9 5 interior JP2-6-2 216.7 43.2 7.2 5 interior, E side JP2-7-2 208.7 41.5 3.8 5 interior, E side JP2-8-2 209.6 43.0 2.0 6 closest to E margin JP2-9-2 206.9 41.2 5.0 5 closest to W margin (possible orientation error)

KA: Kangeq dyke, Greenland (77°12.31'N, 70°46.81'W), 55 m thick, trend 300° Exposed margins atop very high scree-covered cliff KA1-2 306.3 0.7 1.9 8 ECM of dykelet coming off main dyke KA1-1 309.3 -3.9 2.1 14 at ECM of main dyke

KC: Kap Cleveland dyke, Greenland (77°33.990'N, 70°15.359'W), 16 m thick, trend 300° No margins visible KC1-2 NSEP interior, S side KC2-2 NSEP interior, S side KC3-2 290.2 14.0 2.3 6 interior KC4-2 286.9 16.6 2.6 6 interior KC5-1 288.5 17.7 4.0 6 interior KC6-2 NSEP interior KC7-1 NSEP interior, N side KC7-3t NSEP interior, N side KC8-3 296.6 14.2 3.6 4 interior KC9-1 NSEP interior, N side KC10-2 NSEP interior, N side

KL: Kap Leiper dyke, Greenland (78°41.0'N, 70°40.0'W), 54 m thick, trend 270° Difference observed between interior samples and margin samples - margin samples used for site mean KLl-lt 135.5 13.3 3.3 8 44 m SCM (not included) KL1-2 290.2 9.3 0.9 4 44 m SCM (not included in site average) KL2-1 NSEP lOmNCM KL2-2 277.9 9.7 3.1 4 lOmNCM KL3-2 291.1 7.5 2.3 5 43 m SCM KL4-lt 290.7 14.4 1.7 6 43 m SCM KL4-3 NSEP 43 m SCM KL5-2 283.5 12.8 0.5 8 30 m SCM KL6-3 281.5 9.9 1.0 7 29 m SCM KL7-1 279.9 15.2 1.7 9 25 m SCM KL8-1 266.9 4.7 0.8 7 10 m SCM KL9-1 294.0 16.3 1.7 6 6 m SCM 187

KL10-1 262.6 41.2 2.6 10 2 m SCM (not included in site average) KL10-2t NSEP 2mSCM KL10-3 262.0 35.0 4.0 5 2 m SCM (not included in site average) KLll-lt NSEP 50 cm SCM KL11-2 246.3 40.9 6.7 9 50 cm SCM (not included in site average) KL11-3 NSEP 50 cm SCM KL12-1 NSEP lmSCM

KT: Kap Trautwine dyke, Greenland (77°12.267'N, 70°46.471'W), 50 m thick, trend 300c Very low intensity, no margin samples KT1-1 NSEP interior KT2-1 103.3 4.1 1.3 3 interior KT3-1 NSEP interior KT3-2 NSEP interior KT4-1 NSEP interior KT5-2 NSEP interior KT6-lt NSEP interior, reddish portion of dyke KT7-1 NSEP ~1 m SCM (no sun) KT8-1 NSEP ~3 m SCM KT9-1 133.0 -8.0 4.9 ~5 m SCM

LG; Leffert Glacier dyke, Ellesmere Island (78°43.608'N, 75°41.012'W), 50 m thick, trend 080° LG1-1 227.9 17.2 8.3 5 centre of dyke LG2-2 247.7 7.8 8.2 5 centre of dyke LG3-1 240.7 16.3 5.1 5 centre of dyke LG4-1 NSEP 18mNCM LG5-4 251.4 10.3 5.3 5 16mNCM LG6-3 244.9 14.3 1.1 5 14.5 mNCM LG7-1 241.7 9.8 2.9 9 3 cm SCM LG8-1 247.0 19.2 1.7 9 at SCM LG9-1 NSEP host, 2 m SCM LG10-2t NSEP host, 15 cm SCM

NS: Norton Shaw dyke, Ellesmere Island (76°39.70'N, 78°04.88'W), 20 m thick, trend 350° NS1-3 137.5 22.4 6.7 5 20 cm WCM NS2-2t NSEP 15 cm WCM NS2-3 299.9 31.5 4.9 7 15 cm WCM (not included in site average) NS3-1 137.0 35.0 4.2 8 2 cm WCM NS4-1 NSEP 60 cm WCM NS4-2t 186.9 33.2 8.6 4 60 cm WCM NS5-2 172.1 13.3 2.8 4 1 m WCM NS5-4 162.0 26.1 3.2 5 1 m WCM NS6-3 133.0 37.5 2.7 8 at WCM NS6-4t 126.6 34.9 5.8 9 at WCM NS7-1 NSEP host, 4 cm WCM NS7-3 NSEP host, 4 cm WCM NS8-1 NSEP 4mWCM NS8-2 NSEP 4 m WCM NS9-1 165.4 26.1 0.9 4 8 m WCM NS9-6 168.2 32.8 3.7 3 8 m WCM

NU1: Northumberland Island dyke, Greenland (77°21.499'N, 71°33.733'W), 40 m thick, trend 290° No margin samples NU1-1-1 102.2 44.3 1.6 3 interior NU1-2-2 118.5 37.1 3.4 4 interior NU1-3-1 111.1 32.6 4.1 4 interior NU1-4-1 107.1 41.2 4.0 6 interior NU1-5-1 117.0 40.7 2.6 5 interior NU1-6-2 116.9 49.4 2.0 3 interior NU1-7-1 94.1 30.1 0.9 3 interior NU1-8-1 106-1 24.3 4.0 5 interior

NU2: Northumberland Island dyke, Greenland (77°22.78'N, 71°29.038'W), 40 m thick, trend 290° No margin samples NU2-1-2 287.5 18.4 1.6 5 interior NU2-2-2 286.3 19.9 1.2 8 interior NU2-3-2 284.6 18.8 1.2 6 interior NU2-4-2 301.8 15.2 1.2 9 interior NU2-5-2 295.6 15.3 1.3 12 interior NU2-6-1 280.9 18.2 1.6 5 interior NU2-7-1 288.1 10.6 1.4 4 interior NU2-8-2 293.1 19.6 2.1 6 interior

OF: Olrik Fjord dyke, Greenland (77c 10.0'N, 66°20.0'W), 200 m thick, trend 330° OFl-l-2t 15.2 -2.9 4.0 8 5 cm WCM OFl-l-4t 14.3 -0.5 4.2 8 5 cm WCM OFl-3-lt NSEP 5 cm WCM OF2-2-U 34.2 -16.0 4.1 8 50 cm WCM OF2-4-1 26.1 -14.9 3.2 9 50 cm WCM OF2-4-2t NSEP 50 cm WCM OF2-5 27.8 -4.7 4.5 4 50 cm WCM OF3-3 7.8 -11.9 3.9 6 2 m WCM OF3-4t 17.5 -4.3 3.3 10 2 m WCM OF3-7 16.3 -2.3 4.1 5 2 m WCM OF4-lt NSEP 20 m WCM OF4-2 NSEP 20 m WCM OF5-1 NSEP 30 m WCM OF5-2t NSEP 30 m WCM OF6-1 356.2 3.6 2.2 5 centre of dyke, slightly altered OF6-2t NSEP centre of dyke, slightly altered OF7-3t 353.9 -12.8 2.0 3 centre of dyke 189

OF8-2 54.1 3.9 2.3 centre of dyke OF8-3t NSEP centre of dyke OF9-2-7t 11.0 -22.9 4.2 atWCM

OR: Orne Island dyke, Ellesmere Island (77°52.230'N, 76°19.056'W), 60 m thick, trend 250° No sun compass readings, and apparent sunspot activity caused large swings in magnetic compass readings due to magnetic storm. Site therefore not included in Canadian mean pole ORl-lt 64.2 40.7 3.4 12 2 cm SCM ORl-2t 56.3 34.2 3.8 6 2 cm SCM OR2-2t NSEP 30 cm SCM OR3-1 NSEP 6 m SCM OR4-2 324.0 72.9 1.4 3 interior, PEF overprint OR5-1 NSEP interior OR6-1 345.3 60.2 2.3 4 interior, PEF overprint OR6-U NSEP interior OR7-1 88.2 36.3 6.5 9 3 cm NCM OR7-2t 87.4 39.3 4.8 13 1.5 m NCM OR8-2 80.0 76.1 2.5 8 1.5 m NCM OR9-2t 92.1 44.8 3.4 14 5 cm NCM

PH: Philpotts Island dyke, Devon Island (75°00.88'N, 79°37.07'W), 30 m thick, trend 075° No exposed margins PH1-1 NSEP interior PHl-2t 275.5 27.5 4.5 4 interior PH2-lt NSEP interior PH2-2 267.4 44.5 3.6 6 interior PH3-1 273.1 -2.6 8.1 5 interior PH3-2t NSEP interior PH4-1 333.3 24.4 8.5 6 interior PH4-2 335.0 24.5 5.3 5 interior PH5-1 265.9 26.1 6.5 6 interior PH5-2t NSEP interior

PI: Philpotts Island dyke, Devon Island (75°00.486'N, 79°46.192'W), 50 m thick, trend 225° No exposed margins, no sun for samples 4-7 PI1-3 304.9 33.4 8.6 7 interior PI2-1 291.0 44.5 2.0 8 interior PI3-3 297.9 8.6 1.7 4 interior PI4-2 NSEP interior PI5-3 285.6 47.3 0.9 14 interior PI6-2 283.3 43.1 0.9 12 interior PI7-2 NSEP interior 190

PK: Kap Powell dyke, Greenland (77°56.305'N, 72°12.511'W), 65 m thick, trend 280° No margin samples PK1-2 284.2 -24.3 1.9 7 interior PK2-1 287.2 -21.8 2.4 6 interior PK3-1 295.0 -18.4 0.8 7 interior PK4-1 285.1 -21.1 1.6 8 interior PK5-2 289.8 -17.7 2.1 6 interior PK6-1 296.7 -10.1 2.1 8 interior PK7-1 282.3 -15.3 5.9 5 interior PK8-3 NSEP interior PK9-4 NSEP interior

PW: Kap Powlett dyke, Greenland (77c'11.946'N , 70°50.083'W), 10 m thick, trend 290 PW1-3 292.8 12.4 2.5 4 near WCM PW2-2 294.9 8.0 3.3 5 interior, W side PW3-2 NSEP interior PW4-2 NSEP 2mECM PW5-1 288.9 13.8 3.9 4 50 cm ECM PW6-1 NSEP lOcmECM PW6-lt NSEP 10 cm ECM PW6-2 296.0 16.1 4.4 4 10 cm ECM PW7-1 NSEP 10 cm ECM PW7-2 290.6 28.0 3.8 5 10 cm ECM PW8-1 NSEP lmECM PW8-2t NSEP lmECM PW9-2 295.7 15.2 4.0 6 1.5 m ECM PW10-2 296.0 17.9 1.1 3 10 cm ECM

QA: Qaanaaq dyke, Greenland (77°29.064'N, 68°51.860'W), 33 m thick, trend 305° QAM NSEP 10 cm ECM QA1-2 NSEP 10 cm ECM QA2-2 298.5 23.5 2.6 7 12 cm ECM QA3-2 297.1 23.2 2.7 10 35 cm ECM QA4-2 302.8 21.3 3.7 6 70 cm ECM QA5-1 288.6 35.3 5.7 5 1.2 m ECM QA6-3 304.7 20.6 8.5 3 4 m ECM QA7-1 NSEP 9 m ECM (top fell off of core) QA8-2 NSEP interior, W side QA9-3 NSEP interior, W side QA10-1 NSEP 4 m WCM QA11-2 297.8 30.9 3.4 4 7 m ECM QA12.3 NSEP at ECM OT: Qeqertat dyke, Greenland (77°28.688'N, 66025.543'W), 100 m thick, trend 310° QT1-1 196.7 42.3 2.5 6 interior, W side QT2-1 197.9 35.9 4.0 5 interior, W side QT3-2 NSEP interior, W side QT4-2 191.3 35.6 3.4 5 interior, W side QT5-1 193.5 34.2 3.2 6 interior, W side QT6-1 204.8 21.3 7.1 5 6mECM QT7-2 202.6 18.3 3.6 6 4.5 m ECM QT8-3 203.8 20.4 1.8 6 3mECM QT9-2 173.5 44.4 2.2 5 at ECM QT9-7 175.8 20.1 1.4 7 at ECM

RF: Robertson Fjord dyke, Greenland (77°42.890'N, 70°35.472'W), 7 m thick, trend 105° Most samples very low intensity (0-10 mA/m) RF1-1 NSEP 2mNCM RF2-2 NSEP lmNCM RF3-2 NSEP centre of dyke RF4-1 NSEP 5 cm SCM RF4-2t NSEP 5 cm SCM RF5-1 299.0 7.8 4.8 10 cm SCM RF6-1 NSEP at SCM RF7-lt NSEP 1.5mNCM RF7-2 NSEP 1.5mNCM RF8-2 NSEP atNCM RF9-2 NSEP 2 cm NCM RF10-2 289.7 -6.0 6.8 20 cm NCM RF11-2 NSEP 4 cm NCM

SGI: Sverdrup Glacier dyke, Devon Island (75°42.747'N, 83°15.759'W), 40 m thick, trend 100° Too hard to drill, red-weathered rind, no margins, low intensity SG1-1 NSEP interior

SG2: Sverdrup Glacier dyke, Devon Island (75°34.901'N, 83°24.510'W), 20 m thick, trend 305° no sun SG2-1-4 95.5 -30.8 3.3 3 5 m SCM SG2-2-5 97.5 -25.6 5.6 4 centre of dyke SG2-3-5 102.3 -33.4 4.5 4 3.5 m NCM SG2-4-3 105.6 -24.6 2.4 5 2.5 m NCM SG2-5-3 95.4 -17.0 3.8 3 lmNCM SG2-6-1 110.9 -26.5 2.4 6 2 m SCM

SG3; Sverdrup Glacier dyke, Devon Island (75°37.202'N, 83°22.222'W), 20 m thick, trend 240° Many samples very weak, viscous PEF component makes up almost entirety of NRM for some. Therefore not included in Canadian mean pole SG3-1-2 62.2 67.0 4.8 8 at SCM, PEF overprint SG3-2-1 NSEP atSCM SG3-3-2 NSEP limSCM SG3-4-2 74.1 55.1 2.6 5 m SCM, PEF overprint SG3-5-2 NSEP centre of dyke SG3-6-3t NSEP 2 m SCM SG3-7-3 NSEP 2 m SCM SG3-8-3 351.9 61.6 6.7 at SCM, PEF overprint SG3-9-2 NSEP at SCM SG3-10-1 NSEP at SCM

SK: Siorapaluk dyke, Greenland (77°43.938'N, 70°29.820'W), 50 m thick, trend 115° Many samples with low intensity SK1-1 NSEP interior SK2-2 NSEP interior SK3-2 107.8 35.2 4.8 interior SK4-1 NSEP lOcmNCM SK5-3 NSEP lmNCM SK6-2 NSEP interior SK7-2 NSEP interior, N side SK8-2 NSEP interior, N side

TB: Thule Air Base dyke, Greenland (76°27.562'N, 69°14.474'W), 30 m thick, trend 295c No margin samples TB1-1 297.7 23.7 0.5 5 interior TB2-1 292.2 17.9 2.9 3 interior TB3-2 299.8 29.6 4.0 5 interior TB4-1 287.2 25.5 2.0 6 interior TB5-1 293.8 21.1 4.3 6 interior TB6-1 299.6 0.5 7.0 5 interior TB7-6 300.1 37.4 1.1 10 interior TB8-4 300.4 43.7 1.5 8 interior TB9-1 298.7 35.9 3.1 6 interior ST1 10 »- T- v 10 10 in n •

^TjNiodriNri^^riNaiiiiw^^NiDuinNddddinWNNCN^Nridcid IT) T-T- t ID in V CO T- "J f N t- (O O O T- CM CO T- in in co co T-

P «iii (OONOc

NinNCDNOsq^tDNiDOonini-ip^m^nin^oooiqoiNNninsvco^; C-IdH driNNdosnddiDOiD^co^^d^ridcodiiONdiriNd^^^iDsdwd •

0 Dao oN^ridodNodb^^dffl^dNrdsNriddiiiirirciridririoiioNd •tf *- T- T- * N eg SJOlOllDtlOlONv T- CM CO

t taj Nw^^dorNddsriroio'r^eo^riD^dcdoioiDrd^ddiofldriro f T-T- t- m CO O COfflOIBflDttMT- *- CM CO

ujin oso)T-Nr-tcosq*qi(|0(!iS[oni»Nion«iONi-(oq«i^inNN*«)0)N dioViodiooNddrsiiiisri^NiosrdidoicNiociidrirSrNSosjivd * »- T- n in f s T- "tiocoiOT-^Niiit *-

^^owNSNqiDNOininnnwqiDnNN^q^intqin^cq^^aiinqqqin Z-ldQ QiritNiio'd^ioriddMNNsoiciriijioidid^aiNdrid^^ •* T- T- o ID ^ o T- nococoM^^coin

c-frc\a ^ oon^ior-iDr-NqiDNMniMOT-n^q /v^a ddl6i-OBrNbdNlblb*SrdfflN«)IOtS^dfflNIDBJj) O mr-T- r- coini n iim) f* r o>i ao> Oa) > CM CM •* T- T-

Ni-OC0*10T^in^NIDN^C0S(J)CV|v;N(D , 9-^08 oomT-ooOT-cNOOcor -ocor--o*-mcococoi^CO CO CO - ir> *- *- r- co *- in incorooco 0) O) _o CM •* v- CM

i^^c\g NNioinMN»T^iqt-_qiONnN**oiNr;nqn*Niqo)*^ioWT; cMO^oocdocMdooiricNiocMoocdcdo^cd^ocdocnT^o m r-T- T- co *- in T- incocDcoococM^CM o CM CM •* i- CM CO d /-/flq qqt^Nnoasqnipi-ooioo^^Oi-in^eoiDionnnqiONOj^NO : : : ; ; : : ; §0 z. ina sdNin^iridcJdT ddnsT T riwiriN(D^'tN«)rir rir NT i iDioidv *T T- T- O N t (D r rSOIMNtOin* *- co f m *- < 9|dlllBS a a onSSoooSoo ^r yy£iio)£S>tococMCMi2i_ •_ .Q .C-Q >-._nsoi_o-o3.Q.i-c3.c-aroE-Q 'siiirodiiioicidoiiisaiiisd^Ndiosiii

^^ o^cMcooirirocMoocdcd^oicsio^cMcdcDcdoi IOT-T- T- T- m CO T- CM a> CM co -a- ^r i-

Z-993 NOJ^qNNnoicqionciinnneq^ocD^so)

^- T- T- V Tf O f T"

IAI0-90 c!"a?,^cv!l^T^c?ir?a:ic?T*tc!a?l'?cv!tcl<::!,:^,;1!u?'r) oim^ioduioiNddririoiaioidr-tiodss •**-*- CM m n n 0) s CN o) ^- m to T-

Q-7C)'} E^^ON^iDNCMoooii-oiONOONOinion r:*Ni:dKisriv;dti!iti)iisd'»«i<,ir;rd IOT-T- CMOO'1-I^CM O •* O> CM mm co CN

i-qyg ooiococMCMCoco050iCDinoO'«ta>cMco^-cooocOT-

in T— T— rCMy Bin) *f fflCO N(M CM u~IO» sN- r\] ^- IT) CO CN

ONOt-NNOSM-COnCBniOCDOOS^DM-CM e-eya coiocii^dcdciicM'ddrico'ddNdriiriaicd'js •* T- T- CM CM ^" Q> OOt-O

CMC0T-C)NCDlI)S- CO TO CM

c.^s co^wsdddNddr-cM'dcv'sddcMdndddcM'cM^dM^iri^^io'io^^d •* T- T- CM S * CO r- CMSCOCOCMCOV^CM CM T- •«- CM CO "tf fe.i.oc)c; ooo^N^ffloM'iOT-onNco^cMcoT-nooBt-int-Noiticoin^Noioorin c.^^> s^rdddNNddsr-N^NdNT-rrnNrdid^inNriiiT-T-'dsiiird ^3" r- r- N I- * CO r- OKCOlOCOlO^inn *" CO CM ^f

°^a avnridcdaiNddcddd^ai^ria^a^oNuiMiriifl'NrisrT-iricd^^d •*T-T- CM O 't CO COSOCOCMT-COM-CM V- T- C0 CO •* T- i- T- a-C-7C\c> OniOSN^OIO

EEEEEEEEEEEEEEEEEEEEEEEEEEE ^$>d~^^i-$>^^o

ejdujBS SgOgo 00808 ~y !.-2^ £i-:! acu! £^ ^£°>' cvou cwo CI.MN (.CNM -" i_ i- -Q .C -Q 1- — COOi-OCD'OD.QM-ca-CJScoE-0- WH-Wtri5HQ->OZmwOOOZllJl-XHHa:-JC0>- Yb Th Ta Tb Sm Ce Co Cr Sc V Th AI203 Ti02 Lu La Rb Hf Eu Nd Ba Ni Cr Sr Y Zr CaO MnO Si02 Pb U Rb Nb Na20 MgO P205 K20 Fe203

Sample ppm ppm ppm ppm ppm ppm ppm ppm PPm PPm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm % % % ppm % % % % % % % 380.3 460.5 131.0 524.9 78.8 38.6 47.9 26.5 49.6 12.7 15.5 11.7 2.7 0.0 0.8 0.6 3.1 8.5 5.5 0.2 4.2 1.4 LG8-4 434.5 373.1 287.9 88.3 95.0 66.2 33.6 50.3 13.6 14.2 12.9 0.8 3.1 9.3 4.8 0.2 4.1 7.7 0.6 0.4 1.2 1.1 NS4-3 398.5 255.4 386.9 64.7 93.2 97.1 30.1 50.3 14.3 11.3 13.4 13.3 2.9 0.5 0.5 9.6 5.5 0.2 3.6 1.8 1.4 1.0 NS8-1 440.2 470.8 415.3 138.8 83.5 49.6 44.7 29.6 48.4 13.0 15.4 11.5 2.4 0.6 2.7 3.0 0.7 0.6 7.9 7.3 0.2 5.2 PI-CM 431.2 354.4 341.5 146.8 46.6 23.8 49.0 66.0 62.0 13.8 14.9 12.5 2.9 4.3 0.9 0.7 8.5 5.7 0.2 4.6 0.2 1.2 PI3-3 273.5 185.1 77.5 52.7 54.9 50.7 76.4 16.3 10.8 10.8 16.1 7.2 2.2 2.7 0.2 4.0 0.8 0.5 0.2 0.3 7.2 0.9 HM8-1 451.8 443.9 325.6 139.2 94.2 97.8 40.5 24.0 49.2 12.3 14.7 12.5 3.0 8.3 0.2 4.7 3.8 2.1 0.6 0.7 6.0 1.7 EA1-3 424.2 387.4 545.2 152.8 107.6 44.8 48.4 69.5 27.4 15.1 14.8 12.3 2.9 8.3 6.4 0.2 0.9 0.8 0.8 5.0 1.5 1.7 CM4-2 497.9 416.3 422.4 168.1 115.5 40.3 43.8 48.5 32.0 18.4 14.9 11.5 2.3 4.3 0.7 0.6 3.3 7.6 7.6 0.2 5.0 0.7 CW4-4 397.4 190.5 130.8 69.0 32.6 22.3 93.4 50.5 14.7 10.1 14.5 12.9 2.6 4.0 0.1 0.5 7.1 0.2 5.6 0.4 1.6 1.4 HM-CM 196

CM

KL1-11- 2 NU1-9- 2 RF5- 1 t Sampl e CA5- 2 CA6- 1 PK9- 5 GF2- 1 TB9(8)- 3 ni c Frank l 3 Q- Si02 % 1 49.3 49.1 47.8 48.4 47.4 49.4 57.5 44.3 50.7 48.6 49.1 Ti02 % I 3.8 3.8 5.8 4.4 5.1 5.1 0.8 5.8 4.4 5.3 5.1 AI203 % o 12.1 12.7 11.5 12.4 12.2 11.4 17.3 12.4 11.4 12.7 12.2 Fe203 % 15.5 15.5 12.5 15.7 15.9 15.8 9.5 15.0 14.5 14.5 13.4 MnO % 0.2 0.2 0.0 0.2 0.2 0.2 0.1 0.6 0.1 0.2 0.1 MgO % 5.6 5.1 13.3 5.7 5.5 4.8 5.7 8.0 6.2 9.6 6.8 CaO % 9.3 9.4 1.3 9.6 9.1 8.2 2.6 10.0 7.5 5.4 8.2 Na20 % 2.9 3.0 0.3 2.6 2.7 3.1 6.1 2.7 3.1 1.4 3.0 K20 % 0.6 0.6 6.6 0.6 1.3 0.9 0.1 0.5 1.2 1.5 1.2 P205 % 0.5 0.5 0.7 0.5 0.5 0.9 0.1 0.5 0.6 0.5 0.7 Nb ppm 18.8 18.9 44.4 18.0 24.0 40.4 21.3 34.3 27.2 31.0 32.7 Zr ppm 272.2 263.5 451.9 241.6 279.8 485.4 196.4 458.3 376.3 437.6 392.6 Y ppm 34.8 35.1 37.0 46.9 40.0 46.1 45.7 53.8 47.1 44.1 46.9 Sr ppm 348.4 358.4 54.2 292.5 350.8 550.3 305.0 476.9 485.5 327.2 567.0 Rb ppm 12.0 13.3 46.3 8.0 28.0 20.1 3.1 10.0 43.7 19.7 34.5 U ppm 0.6 0.3 2.6 0.0 0.0 0.0 3.9 0.4 1.2 3.7 1.0 Th ppm 1.7 2.6 2.1 0.0 0.0 1.4 17.6 1.9 3.4 2.1 3.8 Pb ppm 5.4 3.8 4.0 0.0 0.0 4.6 8.1 9.4 4.6 12.3 2.4 V ppm 450.8 445.0 531.9 0.0 0.0 411.7 145.5 589.8 396.2 584.7 406.4 Cr ppm 99.7 92.0 71.2 0.0 0.0 25.2 97.4 133.1 31.3 221.7 41.9 Ni ppm 108.6 119.4 120.7 0.0 0.0 90.3 31.3 100.6 82.7 113.1 107.6 Ba ppm 122.0 101.1 361.4 0.0 0.0 158.1 112.2 192.9 323.4 156.8 446.2 Sc ppm 31.28 31.66 28.06 27.82 14.12 Cr ppm 79.29 80.67 55.46 34.41 50.92 Co ppm 42.86 45.09 38.47 45.34 6.75 Ce ppm 48.8 50.14 77.28 99.96 92.29 Nd ppm 33.55 40.82 46.76 66.01 37.35 Eu ppm 2.72 2.59 2.04 4.91 1.7 Tb ppm 1.19 1.45 1.37 1.89 1.07 Hf ppm 6.82 6.62 10.2 12.61 3.61 Ta ppm 1.3 1.32 2.2 2.57 0.79 Th ppm 1.5 1.8 2.34 2.5 12.2 Rb ppm 10.3 13.02 33.35 12.8 13.6 La ppm 18.01 18.81 29.02 26.87 33.34 Sm ppm 8.79 9.5 12.12 8.68 5.14 Yb ppm 3.1 3.22 3.09 2.35 2.04 Lu ppm 0.41 0.45 0.41 0.31 0.31 Yb Th Ta Tb Th Sm Rb Ce Co Cr Sc Y Zr AI203 Lu La Hf Eu Nd Cr V Pb Sr Ti02 Ba Ni U Rb Nb P205 K20 Na20 CaO MgO MnO Si02 Fe203

Sample ppm PPm PPm PPm PPm ppm ppm PPm PPm PPm PPm PPm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm % % % % % % % % % % 250.3 445.4 438.5 464.3 109.2 48.5 96.1 55.9 30.7 49.5 13.0 11.3 2.5 0.2 0.9 3.1 0.2 1.9 7.9 7.3 5.3 1.4 HE7-1 503.1 391.4 394.2 171.6 29.2 76.9 64.4 54.7 46.0 19.7 16.1 10.7 2.2 3.0 0.7 2.3 8.2 9.6 0.3 4.8 1.5 1.1 SK5-3 271.3 490.5 478.1 383.8 88.1 86.9 59.8 45.6 26.5 47.8 13.3 12.0 2.5 2.0 0.9 3.5 4.9 1.2 8.2 7.4 0.2 1.7 KC10-2

Mel ville Bugt 464.0 685.4 222.4 21.6 23.6 86.6 47.9 63.8 81.0 12.4 17.4 4.8 0.4 3.9 3.4 8.2 6.8 0.1 1.5 1.9 1.5 1.6 JP2-9-2 484.5 345.4 300.9 20.0 27.6 39.6 55.3 92.7 46.1 17.1 13.9 2.7 0.1 6.7 0.6 3.3 8.2 8.2 0.2 1.0 1.2 1.4 QT9-2 360.0 895.2 251.8 157.2 140.2 37.8 28.0 69.8 12.7 15.3 14.4 9.1 2.4 0.5 4.7 0.0 3.2 7.6 0.2 2.1 1.9 t OF3-2 357.9 900.5 249.2 156.5 39.5 28.4 50.7 49.7 69.4 10.2 15.4 14.4 6.7 0.5 2.0 3.1 4.8 3.3 0.9 7.6 0.2 2.1 O U- o> 4 CO 1

1 Idas Harboi 299.9 224.7 410.8 163.2 26.5 24.0 70.7 91.0 11.0 49.1 15.7 12.7 4.9 0.2 0.6 2.9 0.4 0.4 9.7 6.5 0.2 2.3 Q T i