The Origin of Massive Ground Ice in Raised Marine Sediments Along the Eureka Sound Lowlands, Nunavut, Canada

The origin of massive ground ice in raised marine sediments along the Eureka Sound Lowlands, Nunavut, Canada

A thesis submitted to McGill University in partial fulfillment of the requirements of the degree of Master of Science

Department of Geography

Acknowledgements This project was largely funded by the Natural Sciences and Engineering Research Council of Canada (NSERC), as part of Professor Wayne H. Pollard’s research program to investigate processes related to the formation and stability of permafrost and periglacial landforms in polar environments. Additional student support was supplied by a NSERC Canada Graduate Scholarship (Master’s) and by the Northern Scientific Training Program (NSTP). Logistical support in the field was provided by the Polar Continental Shelf Program (PCSP). This thesis was supervised by Prof. Pollard (Department of Geography, McGill University) and Prof. Denis Lacelle (Department of Geography, University of Ottawa). Thank you to Prof. Pollard for conceiving the project and organizing the field campaign. I appreciate the unparalleled guidance, knowledge and opportunities that I have gained as a direct result of your supervision. Thank you to Prof. Lacelle for outstanding supervision both in the field and in the laboratory. This work would not have been possible without your support. Thank you to Jean Bjornson at the University of Ottawa for indispensable direction and assistance in the laboratory. Thank you to Nimal de Silva and Smitarani Mohanty of the Geochemistry Laboratory at the Advanced Research Complex (Ottawa) for running the cation and anion samples on the ICP-AES. Thank you to Sarah Murseli and Christabel Jean at the A.E. Lalonde AMS Laboratory for running the radiocarbon samples. Thank you also to Mike Dalva and Paula Kesterman for laboratory and equipment support at McGill. Prof. Tim Moore (Department of Geography, McGill University) was the internal committee member for this thesis. Thank you for introducing me to the world of academic research and providing sage advice along the way. Thank you also to Melissa Ward Jones for ‘showing me the ropes’ – your mentorship in matters academic, bureaucratic and logistic has been much appreciated. Thank you to Fanny Amyot for comradeship and acting as my agent in Burnside. Special thanks to Beth Ann Clarke for providing me with accommodation during my extended visits to Ottawa. Finally, thank you to my parents, Doug and Diane, for unwavering support and encouragement.

Abstract/Resumé

Massive ground ice is a feature of periglacial environments throughout Siberia, Alaska, northern Canada, Antarctica and Mars. The occurrence of massive ground ice raises questions about: i) its origin, ii) its extent across the landscape and iii) its response to climatic change. In the Canadian High Arctic Archipelago, tabular massive ground ice is found extensively throughout the Eureka Sound Lowlands (ESL) on Ellesmere and Axel Heiberg islands. Previous research by W.H. Pollard interpreted the ice as having a segregated origin related to deglaciation and Holocene marine regression. This study used a geochemical approach (stable water isotopes, major ions) to further explain the water source and freezing history of massive ground ice exposures in the ESL. Active layer sediments and permafrost cores from massive ice bodies were collected along an elevation gradient, from modern sea level to Holocene marine limit (~143 m a.s.l.). Results show that ESL ground ice formed during the Holocene in an open-system as raised marine sediments slowly froze from the ground surface downwards. Massive ice formed where the input rate of isotopically-depleted groundwater (from an underlying aquifer fed by the disintegration of the Innuitian Ice Sheet) was much greater than the downwards freezing rate. The extent of geochemical mixing and of the ice segregation process are reflections of the freezing rate – thus we can reconstruct permafrost aggradation along an elevation gradient of 143 m, which represents a span of ~8,000 years since deglaciation.

La glace massive au sol est une caractéristique des environnements périglaciaires à travers Sibérie, Alaska, le nord du Canada, l’Antarctique et Mars. L’occurrence de la glace massive au sol produit des questions par rapport à : i) son origine, ii) son importance à travers le paysage et iii) sa réaction aux changements climatiques. Dans l’archipel de la haute arctique canadienne, on se trouve de la glace au tabulaire massive partout dans les basses terres des îles Ellesmere et Axel Heiberg. Des recherches précédentes par W.H. Pollard ont interprété la glace à avoir une origine intra-sédimentaire reliée à la déglaciation et régression marin pendant l’Holocène. Cette étude a utilisé une approche géochimique (isotopes stables d’eau, ions majeurs) pour expliquer la source d’eau et l’histoire de congélation des exposés de glace massive au sol dans les basses terres du Eureka Sound. Sédiments de la couche active et carottes de pergélisol ont été pris des masses de glace au sol à travers un gradient d’élévation – de le niveau de mer moderne jusqu’à la limite marine de l’Holocene (143 m a.s.l.). Nos résultats démontrent que la glace massive au sol de Eureka Sound est provenue d’un système ouvert durant l’Holocene, pendant que des sédiments marins élevés gelaient lentement de la surface vers le bas. La glace massive a formé quand la contribution de eaux souterraines appauvries en isotopes était beaucoup plus grand que la vitesse de congélation en direction vers le bas. L’ampleur du mixage géochimique et du processus de ségrégation de glace sont des résultats de la vitesse de congélation – donc on peut reconstruire la rapidité d’aggradation du pergélisol à travers un gradient de 143 m, ce qui représente un espace d’environ 8,000 ans après déglaciation.

TABLE OF CONTENTS

1. INTRODUCTION ...... 8

1.1) AN UNUSUAL MINERAL...... 8 1.2) TYPES OF GROUND ICE ...... 10 1.3) MASSIVE GROUND ICE ...... 13 1.3.1) Buried surface ice ...... 14 1.3.2) Ice segregation in soils ...... 15 1.3.3) Wedge ice ...... 19 2. STUDY AREA ...... 20

2.1) PHYSIOGRAPHY AND GENERAL SETTING ...... 20 2.2) CLIMATE AND VEGETATION ...... 23 2.3) QUATERNARY HISTORY ...... 23 2.4) PERMAFROST AND GROUND ICE ...... 26 3. RESEARCH OBJECTIVES ...... 29 4. METHODS ...... 30

4.1) INTRODUCTION ...... 30 4.2) GEOCHEMICAL THEORY – AS APPLIED TO GEOCRYOLOGY ...... 31 4.2.1) Stable water isotopes ...... 31 4.2.2) Major ion geochemistry ...... 34 4.2.3) Occluded gases ...... 35 4.3) FIELD METHODS ...... 36 4.3.1) Site selection ...... 36 4.3.2) Field sampling ...... 39 4.4) LABORATORY METHODS ...... 41 4.4.1) Active layer (AL) samples ...... 41 4.4.2) Permafrost cores ...... 41 4.4.3) Ice wedge (IW) samples ...... 44 4.4.4) Collection of ice blocks ...... 45 5. RESULTS ...... 46

5.1) ELEVATION TRANSECT ...... 47 5.1.1) Dump Slump ...... 47 5.1.2) Station Creek ...... 57 5.1.3) Black Top...... 64 5.1.4) Gemini ...... 74 5.2) EUREKA SOUND LOWLANDS WATER CHEMISTRY ...... 84 5.3) ICE WEDGE 14C-DOC ...... 89 6. DISCUSSION ...... 90

6.1) ORIGIN OF MASSIVE ICE ...... 90 6.1.1) Gemini ice: a special case? ...... 96 6.2) CAN ICE SHEET MELTWATER BE TRANSFERRED INTO A SUBGLACIAL AQUIFER? ...... 102 6.3) PERMAFROST EVOLUTION ...... 104 6.4) DISSOLVED ORGANIC CARBON IN GROUND ICE: 14C DATING ...... 107 7. CONCLUSIONS ...... 109 8. REFERENCES ...... 111 APPENDIX ...... 120

LIST OF FIGURES

1. INTRODUCTION ...... Figure 1.2.1: Ground ice classification from MacKay (1972)...... 11 Figure 1.3.1: An example of an ice wedge viewed in cross-section. Photograph courtesy of W.H. Pollard...... 13 Figure 1.3.2: An example of tabular massive ice. Photograph courtesy of W.H. Pollard...... 13 Figure 1.3.3: A schematic representation taken from Rampton (1988, pp. 68), depicting the development of massive ground ice associated with permafrost aggradation following the retreat of the Laurentide Ice Sheet ...... 15 2. STUDY AREA ...... Figure 2.1.1: Queen Elizabeth Islands in the Canadian High Arctic. Red arrow points to Eureka Sound Lowlands. Modified from Pollard et al. (2015)...... 20 Figure 2.1.2: Aerial view of a typical ESL landscape ...... 21 Figure 2.1.3: Aerial view of Eureka area, Ellesmere Island ...... 22 Figure 2.3.1: Diagram depicting the Innuitian Ice Sheet at LGM (from England et al. 2006)...... 24 Figure 2.3.2: Deglaciation of the Innuitian Ice Sheet (from England et al. 2006)...... 25 4. METHODOLOGY ...... Figure 4.2.1: Meteoric relationship between 18O and D in precipitation across Earth’s latitudes (from Clark & Fritz 1997) ...... 32 Figure 4.2.2: Stable water isotopes in precipitation at Eureka Weather Station collected on a monthly basis from January 1989 to December 1993 (IAEA/WMO 2018) ...... 32 Figure 4.2.3: Local meteoric water line (LMWL) for Eureka Weather Station, Nunavut, Canada. (IAEA/WMO 2018) ...... 33 Figure 4.3.1: Schematic diagram of elevation transect of massive ice exposures on the Ellesmere Island side of the ESL...... 36 Figure 4.3.2: Map of Eureka Sound Lowlands (1:400,000) ...... 37 Figure 4.3.3: Map of Eureka vicinity (1:50,000) ...... 38 Figure 4.4.1: Circular saw with 0.8 mm diamond blade at the CryoLab, University of Ottawa...... 42 5. RESULTS ...... Figure 5.1.1a: Dump Slump (DS) massive ice exposure in a retrogressive thaw slump...... 49 Figure 5.1.1b: Gravimetric water content (GMC) of samples in AL and core sub-samples from massive ice (DS-1)...... 49 Figure 5.1.1c: Depth profiles of ice from DS-1 core: a) Photograph of DS-1 core (partial), b) volumetric 18 ice content (fVi), c) δ O and d) deuterium excess...... 51 18 Figure 5.1.1d: Depth profiles of ice from DS-2 core. a) volumetric ice content (fVi), b) δ O and c) deuterium excess...... 52 Figure 5.1.1e: Co-isotope plot (D-18O) of DS-1 core samples...... 53 Figure 5.1.1f: Co-isotope plot (d-D) of DS-1 core samples...... 53 Figure 5.1.1g: Co-isotope plot (D-18O) of DS-2 core samples...... 54 Figure 5.1.1h: Co-isotope plot (d-D) of DS-2 core samples...... 54 Figure 5.1.1i: Depth profiles of major cations in DS-1 core...... 55 Figure 5.1.1j: Depth profiles of major anions in DS-1 core...... 56 Figure 5.1.1k: Geochemical facies of massive ice from DS-1 core...... 56 Figure 5.1.2a: Station Creek massive ice exposure in a large retrogressive thaw slump...... 58 Figure 5.1.2b: Sampling sites: Station Creek ...... 58 Figure 5.1.2c: Depth profiles of ice from SC-1 core: a) Photograph of SC-1 core, b) volumetric ice content 18 (fVi), c) δ O and d) deuterium excess...... 59 Figure 5.1.2d: Co-isotope plot (D-18O) of SC-1 core samples...... 60 Figure 5.1.2e: Co-isotope plot (d-D) of SC-1 core samples...... 60 Figure 5.1.2f: Depth profiles of major cations in SC-1 core...... 61 Figure 5.1.2g: Depth profiles of major cations in SC-1 core...... 62

Figure 5.1.2h: Geochemical facies of massive ice from SC-1 core...... 62 Figure 5.1.3a: Black Top (BT) massive ice exposure in a retrogressive thaw slump ...... 66 Figure 5.1.3b: Gravimetric water content (GMC) of samples in AL and core sub-samples from massive ice (BT-1 & BT-2)...... 66 Figure 5.1.3c: Depth profiles of ice from BT-2 core: a) Photograph of BT-2 core (partial), b) volumetric 18 ice content (fVi), c) δ O and d) deuterium excess...... 68 18 Figure 5.1.3d: Depth profiles of ice from BT-1 core: a) volumetric ice content (fVi), b) δ O and c) deuterium excess...... 69 Figure 5.1.3e: Co-isotope (D-18O) plot of BT-2 core samples...... 70 Figure 5.1.3f: Co-isotope (d-D) plot of BT-2 core samples...... 70 Figure 5.1.3g: Co-isotope (D-18O) plot of BT-1 core samples...... 71 Figure 5.1.3h: Co-isotope (d-D) plot of BT-1 core samples...... 71 Figure 5.1.3i: Depth profiles of major cations in BT-2 core...... 72 Figure 5.1.3j: Depth profiles of major anions in BT-2 core...... 73 Figure 5.1.3k: Geochemical facies of massive ice from BT-2 core...... 73 Figure 5.1.4a: a) Photograph of Gemini-1 sampling site. b) Gravimetric moisture content (GMC) of core sub-samples from massive ice (G-1)...... 75 Figure 5.1.4b: Photograph of Gemini-3 sampling site at a retrogressive thaw slump...... 76 Figure 5.1.4c: Gravimetric moisture content (GMC) of core sub-samples from G-3 core...... 76 Figure 5.1.4d: Depth profiles of ice from G-1 core: a) Photograph of G-1 core, b) volumetric ice content 18 (fVi), c) δ O and d) deuterium excess...... 78 Figure 5.1.4e: Depth profiles of ice from G-3 core: a) Photograph of G-3 core, b) volumetric ice content 18 (fVi), c) δ O and d) deuterium excess...... 79 Figure 5.1.4f: Co-isotope (D-18O) plot of G-1 core samples...... 80 Figure 5.1.4g: Co-isotope (d-D) plot of G-1 core samples...... 80 Figure 5.1.4h: Co-isotope (D-18O) plot of G-3 core samples...... 81 Figure 5.1.4i: Co-isotope (d-D) plot of G-3 core samples...... 81 Figure 5.1.4j: Depth profiles of major cations in G-3 core...... 82 Figure 5.1.4k: Depth profiles of major anions in G-3 core...... 83 Figure 5.1.4l: Geochemical facies of massive ice from G-3 core...... 83 Figure 5.2.1: Total dissolved solids (TDS) in melted ground ice in permafrost cores from the ESL ...... 85 Figure 5.2.2: Piper plot showing hydrochemical facies present in ground ice and environmental samples in the ESL...... 86 Figure 5.2.3: Stiff diagrams reveal three distinct hydrofacies present in ESL ground ice ...... 87 Figure 5.2.4: All stable water isotope data from ESL ground ice and environmental sources...... 88 6. DISCUSSION ...... Figure 6.1.1: Exposed massive ice in a retrogressive thaw slump. Close to G-3 core sampling site...... 98 Figure 6.1.2: Shaw’s (1977) facies model for alpine glacial lakes...... 100

APPENDICES ...... Figure A1: Gravimetric and volumetric moisture/ice content in AL and permafrost in a short core (MF-1) from a Mokka Fiord massive ice body...... 120 Figure A2: Mokka Fiord (MF) massive ice exposure ...... 120 Figure B1: Collapsed pingo, Mokka Fiord area ...... 121 Figure B2: Photograph and hydrochemical characteristics of collapsed pingo ...... 121 Figure C1: Sampling method micro-study ...... 122

LIST OF TABLES

2. STUDY AREA ...... Table 2.4.1: Areas with many natural exposures of massive ground ice identified by reconnaissance in 1991, 1992 and 1994, as discussed in Pollard (2000a) and via personal communication)...... 28 4. METHODOLOGY ...... Table 4.4.1: USDA texture class boundaries ...... 41 5. RESULTS ...... Table 5.1.1a: Sediment data from AL samples and DS-1 permafrost core samples...... 50 Table 5.1.1b: Texture class data from AL samples and DS-1 permafrost core samples...... 50 Table 5.1.1c: Dissolved organic carbon (DOC) and total nitrogen (TN) present in supernatant water in DS-1 core ...... 56 Table 5.1.2a: Dissolved organic carbon (DOC) and total nitrogen (TN) present in supernatant water in DS-1 core ...... 62 Table 5.1.2b: Water chemistry from SC3 samples ...... 63 Table 5.1.3a: Sediment data from AL samples and BT-2 permafrost core samples ...... 67 Table 5.1.4a: Sediment data from G-3 permafrost core samples...... 77 Table 5.1.4b: Texture class data from G-3 permafrost core samples...... 77 Table 5.1.4c: Dissolved organic carbon (DOC) and total nitrogen (TN) present in supernatant water in G- 3 core...... 83 Table 5.2.1: Stable water isotopes and major ions in local water sources: Slidre Fiord (sampled at 79°58’17.8”N, 85°41’50.7”W), Black Top (BT) Creek (sampled at 79°58’50.4”N, 85°4’36.4”W) and a late-lying snowbank near the Eureka airstrip (79°59’58.3”N, 85°47’47.4”W)...... 85 Table 5.3: 14C-DOC results from ESL ice wedges. The ice wedges are associated with retrogressive thaw slump exposures discussed in section 5.1...... 89 6. DISCUSSION ...... Table 6.1: Summarized values of two major ion ratios for ESL ground ice and other environmental sources. Samples with ratios < 1 are marked in bold...... 96 APPENDICES ...... Table A1: MF-1 core active layer sediment characteristics...... 120 Table A2: Texture class data from MF-1 core...... 120 Table A3: Stable water isotopes in block samples (MF-4, MF-5) taken from a Mokka Fiord retrogressive thaw slump exposure...... 120

1. Introduction

Fieldwork over the last thirty years has discovered massive ground ice underlying large swaths of the Eureka Sound Lowlands in the Canadian High Arctic (e.g. Pollard 2000a; Pollard et al.

2015). The current polar desert climate of the area is unable to supply the water to form such extensive ice deposits. Detailed analysis of the cryostratigraphy, spatial distribution and elevation of ground ice exposures forms the basis of the current explanation of ground ice formation and landscape evolution (Pollard and Bell 1998). The source of the water to supply the ice bodies and the relationship between the massive ground ice and the Last Glacial Maximum remain poorly explained. The aim of my research is to use isotope geochemistry and stratigraphic context to further elucidate the water source and genesis of massive ground ice in the Eureka Sound Lowlands of Ellesmere Island and Axel Heiberg Island, Nunavut.

1.1) Ice as an unusual mineral

Under cryotic conditions, ice should be thought of as a mineral; it is a naturally-occurring, inorganic, homogeneous, crystalline solid (H2O), in the same way quartz (SiO2) or calcite

(CaCO3) are minerals. In the polar regions of the Earth it is an important structural component of the near-surface crust. H2O is a unique mineral in the sense that its melting point is within the range of Earth temperatures and thus highly unstable. Yershov (1990) points out that every planet has a chemical compound in its atmosphere that exists in solid, liquid and gaseous states, for instance CH4 on Uranus. On Earth, it happens to be H2O – which also has an exclusive role in chemical and biochemical processes and is a prerequisite for organic life. Earth and its Moon are exceptional among celestial bodies in our solar system in that they have a discontinuous cryogenic surface, as opposed to a complete cryogenic surface (e.g. Mars, Uranus) or a non-

cryogenic surface (e.g. Mercury, Venus). This means that the dominant substance (H2O) perennially exists in a solid state only in certain parts of the planet – what we call the cryosphere.

Put simply, the cryosphere is the frozen water part of the Earth system, which includes the continental ice sheets of Greenland and Antarctica, regional ice caps, sea ice, glaciers, snow and frozen ground. Within the cryosphere, terrestrial regions with cold, but non-glacial climates are referred to as periglacial environments. Within the scope of periglacial environments (which includes any area largely dominated by frost action) are areas where permafrost is present.

Permafrost is defined as earth material that remains below 0°C for at least two consecutive years

(French and Shur 2010). Permafrost is found across about 20% of the global land area and up to

50% of Canada’s land area (French 2007). The surface horizon that overlies the permafrost and thaws on an annual basis is called the active layer. The depth to which permafrost persists in the

Earth’s crust depends on i) the properties of the soil and rock materials, ii) the amount of heat arriving at the ground surface, iii) the amount of heat actually absorbed at the surface (e.g. albedo) and iv) the amount of heat arriving from the Earth’s interior (Yershov 1990). The insulating presence of a large surface ice mass, such as an ice sheet or a glacier, limits the extent of permafrost development, hence the association of permafrost with periglacial as opposed to glacial landscapes.

Permafrost is defined as either epigenetic (i.e. forming after deposition of the host sediment or rock) or syngenetic (i.e. forming as sedimentation occurs, causing the base of the active layer to aggrade upwards) (French and Shur 2010). In high-latitude areas that were previously covered by thick Late-Pleistocene ice sheets, it is mostly epigenetic permafrost, as permafrost aggradation is relatively recent. Syngenetic permafrost is typically silty and ice-rich and associated with cold- climate sedimentation. In modern times, this occurs primarily in alluvial and deltaic

9 environments of north-flowing rivers in Arctic North America (e.g. Mackenzie, Colville) and northern Siberia (e.g. Lena, Yenisey, Obe, Kolyma). In the Pleistocene, syngenetic permafrost aggraded in loess sediments across large areas of Siberia, forming ice-rich yedoma complexes that have persisted throughout the Holocene (Yershov 1990).

Unique landforms and processes are associated with the freezing of water, frost heaving and ice segregation in the soil. While they are home to less than 0.3% of the global human population

(French 2007), permafrost areas are important to us for their extensive natural resources (gold, diamonds, lead, zinc, copper, etc.), large stocks of soil organic carbon (SOC) and their inherent instability in the face of climatic change. Recent warming in the Arctic has produced noticeable landscape changes essentially due to the phase change of H2O from solid to liquid. It has become important to understand the extent, origin and chemical make-up of ice in order to predict how the system will respond in the future.

1.2) Types of ground ice

All ice found in freezing and frozen ground is referred to as ground ice. For ground ice to persist, it must be within permafrost sediments. When present, ground ice drastically alters the mechanical and thermal properties of the soil or bedrock. It can exist as a cement around other mineral particles, or as a monomineral ‘rock’ (Yershov 1990). French (2007) outlines several reasons why knowledge of the origin, nature and extent of ground ice is useful: i) to understand the evolution of Quaternary landscapes, ii) to retrieve paleoclimatic and paleohydrological information and iii) to assess potential geomorphic responses to disturbances of permafrost thermal regimes, such as the construction of roads or pipelines. Additionally, the discovery of extensive ground ice on Mars (Boynton et al. 2002) has raised questions about the processes

10 involved in its formation (Lacelle et al. 2008) and whether it is a dynamic part of Martian geomorphology (Schorghofer and Aharonson 2005).

There have been many attempts to classify ground ice based on morphology, genesis and relation to the other components of the frozen ground mass (solid matrix, air/voids). Shumskii (1959) describes how early Russian ground ice researchers used poorly-substantiated extrapolations to assign universal significance to specific phenomenon they observed in the field. In response, there was an attempt to emphasize the great range of conditions for the formation of ground ice, which sparked a profusion of classification schemes. The enthusiasm to sub-classify ad nauseam was more rigorous, but this approach made it almost impossible to characterize the nature of ground ice in any specific case. In North America, MacKay (1972) classifies ground ice based

Figure 1.2.1: Ground ice classification from MacKay (1972).

11 upon its three primary sources of water (atmosphere, surface or sub-surface) and the principal transfer process of water to the freezing plane, resulting in 10 genetically distinct types of ice

(Fig. 1.2.1). This classification, while elegant, excludes buried surface ice (e.g. glacial ice) or cave ice. According to French (2007), the most significant types in terms of frequency and volume are pore ice (in-place freezing of subsurface water), wedge ice (formed by infiltration of atmospheric or surface water into thermal contraction cracks), segregated ice (ice lenses formed by sub-surface water under a thermal gradient) and buried ice (e.g. a relict glacier).

The distribution of ground ice across a landscape is highly variable. Coring methods provide considerable detail for one point in space but cannot give an idea of the lateral extent of ice.

Geophysical surveys, such as the use of ground-penetrating radar (e.g. Angelopoulos et al. 2013) are useful for estimating spatial extent but lack detail. Luckily, areas with ground ice tend to have many natural exposures, sometimes up to 20-30 m high. This is due to the unstable nature of ice at Earth temperatures, which makes the terrain inherently susceptible to accelerated erosion and thaw-induced disturbances (Pollard 2000a). Natural exposures provide the most detailed information on ice morphology, structure and stratigraphy.

Cryo-stratigraphy is defined as “the study of frozen layers in the Earth’s crust” (French and Shur

2010). It owes its roots to a long tradition of geocryology in Russia and is based on the premise that perennially-frozen sediment and rock have different structures than unfrozen sediment and rock (Gilbert et al. 2016). The value of cryo-stratigraphy for a geomorphologist is that the thermal history and origin of the substrate can be estimated by examining the structures formed by ground ice in sediments. French and Shur (2010) note how field-observations of cryo- structures can reveal either epigenetic (formed after deposition of host sediment) or syngenetic

(forms at same time) permafrost formation. Reticulate and irregular cryo-structures are

12 associated with the former. The latter is typified by rhythmically organized (layered or lenticular) structures, reflecting a progressively aggrading depositional surface in cold-climate conditions.

1.3) Massive ground ice

When the ground ice content exceeds the saturated-moisture content of its host sediments it is known as excess ice. Wherever excess ice exists, the ground is susceptible to volume loss and subsidence should the ice melt. On a landscape scale, this can lead to rapid, widespread mass wasting to produce thermokarst. The highest excess ice contents are associated with massive ice bodies, which are defined has having a gravimetric ice content > 250%

(MacKay 1989). In basic terms, these are ice bodies in the ground that are close to pure ice (not necessarily pure H2O), or with minimal sediment enclosed. Generally, massive ice can be wedge ice, intrasedimental (i.e. enhanced segregated and segregated-intrusive) ice or buried ice. Wedge ice is easily Figure 1.3.1: An example of an ice wedge viewed in cross- recognized by its V-shape when observed as a cross-section section. Photograph courtesy of W.H. Pollard. (Fig. 1.3.1). Distinguishing between massive intrasedimental

ice and buried surface ice is less

certain. In their most spectacular

forms, both types can exist as tabular

massive ice bodies up to 20 m thick,

extended horizontally across the

Figure 1.3.2: An example of tabular massive ice (white) exposed landscape (e.g. Fig. 1.3.2). in the Eureka Sound Lowlands, Ellesmere Island, Nunavut, Canada. The exposure is approx. 5-10 m high. Photograph Assessments of their origins are courtesy of W.H. Pollard.

13 difficult and have long been one of the most contentious problems in the fields of periglacial geomorphology and geocryology (MacKay 1971; French and Harry 1990; Vaikmae et al. 1993;

Fritz et al. 2011, etc.). While they may appear similar in the field, the processes by which these massive ice bodies form are very different. Indeed, the difference in genesis is analogous to an igneous rock forming in-situ in the crust (segregated ice) vs. a sedimentary rock (buried surface ice).

1.3.1) Buried surface ice

Sediments can be deposited on glaciers, snowbanks, sea ice and lake ice. If enough sediment accumulates, it can protect the underlying ice from seasonal thaw. Everest and Bradwell (2003) provide an example of this occurring in current time in southern Iceland, near the foot of three glaciers: Skeidarájökull, Hrútárjökull and Virkisjökull. Burial proceeds both gradually and in catastrophic events (e.g. large landslides or floods). It is often associated with glacial surges

(Schomacker 2008).

Late-19th century Russian explorers such as von Toll (1897) encountered massive ice exposures in outcrops along Siberian coasts or rivers and speculated that they were relict glacier ice, i.e. buried ice from the last major continental glaciation. Shumskii (1959) argues that most, if not all, of these exposures are intrasedimental in origin. There are, however, a few notable examples of

Late Pleistocene glacial ice preserved as massive ground ice in Siberia. Astakov and Isayeva

(1988) present evidence that thick ice beds exposed in the Yenisei River valley are in fact

“fossil” glacial ice. They dub this an example of “retarded deglaciation”, whereby the landscape is still “recovering” from the last Ice Age. This throws into doubt the whole notion of periglacial environments, and questions whether many landscapes that were assumed to be in geomorphological equilibrium are still responding to the climate amelioration that marked the

14 end of the Pleistocene. This idea is echoed by Everest and Bradwell (2003), who recognize that

“landscape development in areas of stagnant ice topography may post-date initial deglaciation by a considerable degree”.

There is evidence that buried glacial ice persisted in sediments left behind by the Laurentide Ice

Sheet in eastern North America. Eyles (1977) suggests that postglacial permafrost and buried ice were prevalent in Newfoundland for at least 2,000 years after ice retreat. In areas along the margins of the former Laurentide Ice Sheet in the western Canadian Arctic, numerous researchers have found preserved glacial ice in till (e.g. French and Harry 1990; Murton et al.

2005; Lacelle et al. 2007). Buried ice and intrasedimental ice are often found in proximity to each other within these Laurentide tills, making them difficult to tell apart (French and Harry

1990). Moreover, the preserved ice is usually basal glacial ice that would have melted and refrozen multiple times, rendering it more isotopically comparable to segregated ice. Thus, the distinction between buried glacial ice and intrasedimental ice can be surprisingly blurry.

Other places where buried glacier ice is found on Earth include the McMurdo Dry Valleys of

Antarctica (Sugden et al. 1995; Lapalme et al. 2017), the Svalbard archipelago (Brandt et al.

2007) and Bylot Island, Nunavut (Moorman and Michel 2000). It has also been hypothesized to exist in the equatorial regions of Mars (e.g. Scanlon et al. 2015).

1.3.2) Ice segregation in soils

The process of small-scale ice segregation is fairly well known. Mackay (1972) describes how ice lenses will grow in fine-textured soil (<1 mm particles) with a downward-moving freezing front. By a combination of both capillary action and hydrodynamic gradients, soil moisture migrates to the freezing plane. Upon freezing, the radius of the ice crystals is greater than the

15 radius of curvature of the soil grains, thus the ice is ‘segregated’. Smith (1985) specifies that the ice lens grows at some distance behind the freezing plane, as there is unfrozen, migrating water adjacent to soil particles within “frozen” soils. In coarse-textured soil, there is very little moisture migration and ice is able to penetrate between grains (i.e. overcome the Gibbs-Thomson effect of curvature), resulting in pore ice, as opposed to ice segregation. Yershov (1990) emphasises that moisture migration in any soil will only occur if the advance of the freezing boundary is slow enough to maintain a strong thermal gradient. Experiments on ice segregation in soils are notoriously hard to interpret due to high heterogeneity and polydispersity. In a recent breakthrough, Schollick et al. (2016) demonstrate the key factors that determine if ice segregation will occur are the freezing rate and particle size, using a series of laboratory tests on suspended colloidal particles. Under slow freezing conditions, growing ice “rejects” particles, as opposed to engulfment of particles by a fast-moving freezing front.

While the theories described above adequately explain micro-scale ice segregation (e.g. frost heaving), the physics behind the formation of massive, several-metre thick lenses are more complex. Massive ice formed by the migration of water to the freezing plane is still referred to as segregated ice, although this is a bit of a misnomer, because the adjective ‘segregated’ implies separated layers instead of a monolithic body. Yershov (1990) suggests the term ‘migration ice’ is more appropriate, but in this thesis I will use the more common North American term

‘segregated ice’. Massive segregated ice is usually associated with fine-grained sediments underlain by coarser sediments. The underlying, permeable coarse sediment (sand) is necessary to supply the excess water into the clay or silt above. MacKay (1972) presents drill hole data that shows 95% of massive ice bodies in the Tuktoyaktuk Coastlands, N.W.T., Canada, were directly underlain by sand, while the remaining 5% had a thick sand layer not far away.

For massive segregated ice to occur on a

landscape scale, it requires very specific

thermal and hydraulic conditions.

Researchers in North America and Russia

realized that the areas where they found

massive ground ice were near the outer limits

of known glaciation, suggesting a link

between glacial history, permafrost evolution

and ice segregation. Rampton (1988, 1991)

proposes a model where the formation of

massive segregated ice is triggered by

deglaciation (Fig. 1.3.3). Originally

conceived to explain the origin of massive

ice in the Tuktoyaktuk Coastlands, the model

envisages degradation of any previously-

existing permafrost beneath Wisconsinan-era

ice sheets. As deglaciation proceeds,

permafrost begins to aggrade downwards

into newly-exposed Pleistocene sediments.

Subglacial meltwater is driven to the glacier

terminus by a hydraulic gradient created by

the glacier ice overburden pressure. The Figure 1.3.3: A schematic representation taken from Rampton (1988, pp. 68), depicting the development of model explains the source of water necessary massive ground ice associated with permafrost aggradation following the retreat of the Laurentide Ice Sheet.

17 for the growth of such large, tabular ice bodies and the mechanism by which water moves to the freezing plane. For massive ice to grow to thicknesses of 30-50 m, the glacier terminus must remain in more or less the same location for at least 1000 years according to Rampton (1991).

Conceivably, there must also be an equilibrium between the advance of the freezing front and the supply of water. The presence of bands of sediment reflect subtle changes in this dynamic balance.

The hypothesized mechanism of subglacial drainage by groundwater is well-observed under present-day conditions. Boulton et al. (2007) demonstrate that distributed flow is the main process beneath large ice sheets. They frequently observe upwelling of groundwater at or near glacial margins. The subglacial hydrological system is recharged by: i) water derived from surface melt that penetrates to the glacier bed (major source in summer), ii) water derived from basal melting and iii) water produced by melting of ice walls of subglacial drainage tunnels.

French and Harry (1990) argue that a subglacial hydraulic gradient may not always be required, with pore water expulsion from the freezing of saturated coarse-grained sediment capable of supplying enough water to form ground ice up to 25 m thick, but Rampton (1991) doubts whether this could be the case over a large area as is observed in the Tuktoyaktuk Coastlands.

In the cases discussed above, free groundwater is drawn to the freezing front by both thermal and pressure gradients. Sometimes, the thermal gradient is negligible, and the principal transfer process is only due to a pressure potential. A highly-pressurized injection of free groundwater into frozen or freezing ground will produce intrusive ice, analogous to intrusive bodies of magmatic rocks (batholiths, laccoliths, sills, dikes, etc.). The main difference between intrusive ice and segregated ice is that the separation of the ice from the surrounding material is due to the pressure of water alone, and not by the matric potential (suction) associated with the growth of

18 ice crystals (Shumskii 1959). Intrusive ice tends to be highly localized, such as massive laccoliths that form ice-cored hills, known as pingos in North America and bulgunniakh in

Siberia (Yershov 1990). It is generally very pure and transparent ice.

1.3.3) Wedge ice

Wedge, or vein ice is the most common type of massive ice. Ice wedges are formed by the freezing of surface water that penetrates thermal-contraction cracks in the spring (French and

Shur 2010). Active wedges are always found in a near-surface stratigraphic position, although it is possible to find paleo-wedges underlying modern sediments. They are associated with fine- grained sediments due the strong expansion/contraction properties of clays. Viewed from above, the thermal-contraction cracks usually form polygons anywhere from 5 to 30 m across (Fig.

2.1.1). The wedges can be several metres wide at their widest point (i.e. at the top). Unlike tabular massive ground ice, the general process of wedge ice formation is undisputed, although it can be unclear whether it is meltwater, snowfall or hoarfrost that is infiltrating the thermal- contraction crack (St-Jean et al. 2011). Due to the strong seasonality of ice wedge growth, they can preserve useful paleoclimate and paleohydrologic information (e.g. Meyer et al. 2010; Opel et al. 2017). In theory, the youngest ice is found in the centre of the wedge, with progressively older ice further away from the centre. Many studies attempt to reconstruct climate based on

δ18O values dated using 14C from particulate organic matter (POM) in ice wedges, although recently it has been pointed out that 14C ages are often overestimated due to the incorporation of old POM (e.g. Lachiniet et al. 2012). A new method for dating ice wedges is through 14C incorporated in dissolved organic carbon (DOC).

2. Study Area – the Eureka Sound Lowlands

2.1) Physiography and general setting

In the Canadian High Arctic Archipelago, Eureka Sound (~290 km long, 13-48 km wide) separates Axel Heiberg Island from Ellesmere Island. Broadly, it is an intermontane area fringed by massifs (>500 m a.s.l.) with ice caps and alpine glaciers. The region is underlain by poorly

Figure 2.1.1: Queen Elizabeth Islands in the Canadian High Arctic. Red arrow points to Eureka Sound Lowlands. Modified from Pollard et al. (2015). lithified Mesozoic and Cenozoic clastic rocks of the Sverdrup Basin geologic structure, mostly sandstone, siltstone and shale (Bell and Hodgson 2000). Axel Heiberg and Ellesmere islands are

20 within the Queen Elizabeth Islands group (Fig. 2.1.1), which is a sub-plate, broken by rifting in the Late Cretaceous. According to Hodgson (1989), Eureka Sound is a graben within this rift system. The sound is surrounded by rolling lowlands on both sides that cover approximately 750 km2 (see Fig. 4.3.2).

From here on, the Eureka Sound Lowlands (ESL) will be defined specifically as the land below

Holocene marine limit, which is between 140-150 m above modern sea level (see Fig. 4.3.2).

The Fosheim Peninsula, on the Ellesmere side of the

sound, comprises most of the land area of the ESL.

The surficial material of the ESL is fine-grained

sediment (silty clay to fine sand) deposited by a late

Pleistocene/early Holocene marine inundation

(Hodgson 1985), interspersed with some Holocene Figure 2.1.2: Aerial view of a typical ESL landscape. Ice-wedge polygons are fluvial deposits and patches of thin organic muck. ubiquitous in fine-grained sediments below Holocene marine limit. An inactive Massive ground ice is almost exclusively found below retrogressive thaw slump, a thaw pond and a river channel are also visible in this image. within these fine-grained marine sediments (Pollard

1991; Hodgson and Nixon 1998; Pollard 2000a; Pollard et al. 2015), thus the focus on land below marine limit. Viewed from above, the ESL landscape is a smooth plain marked by ice wedge polygons, retrogressive thaw slump scars and short river channels (Fig. 2.1.2).

Archaeological surveys suggest prehistoric occupation of the area from 4000-3000 BP and 1300-

300 BP, at times when sea ice was more open (Sutherland 2000). The first Europeans to pass through the Eureka Sound area were the Norwegians aboard the Fram, led by Captain Otto

Sverdrup in the spring of 1901. The geologist on the Fram, Per Schei, remarked upon the raised marine terraces and extensive, rolling lowlands in the ESL. He did not, however, comment on

21 exposures of ground ice, likely because the ground was covered in snow and they were too busy shooting hares (Sverdrup 1904).

The Canadian military established a weather station (Eureka) on the north shore of Slidre Fiord in 1947 (Fig. 2.1.3). In addition to the weather station (now operated by Environment Canada), current infrastructure at Eureka includes a gravel airstrip, Canadian military base (Fort Eureka) and fuel storage for the Polar Continental Shelf Program. A brief flurry of oil exploration on the

Fosheim Peninsula in the 1970’s was commercially unsuccessful, but it did spark scientific interest in the area and instigated detailed geologic mapping.

Station Creek (SC) thaw slump

Airstrip & Fort Eureka

Eureka Weather Station

Slidre Fiord

Figure 2.1.3: Aerial view of Eureka area, Ellesmere Island.

2.2) Climate and vegetation

The Fosheim Peninsula is perhaps the driest area in Canada with mean annual precipitation less than 68 mm. The lowlands support tundra and sedge-meadow vegetation characterized by 145 vascular plant species (Edlund et al. 2000). This surprisingly high plant diversity is supported by i) high summer temperatures compared to elsewhere in the Queen Elizabeth Islands and ii) the provision of water in dry years by the melting of massive ground ice. In fact, Becker et al. (2016) showed that ground ice melt near Eureka leads to the creation of more microhabitats and promotes greater vascular plant diversity.

The ESL may have higher-than-expected summer temperatures for its latitude (due in part to orographic effects of the mountainous interior of Ellesmere Island), but overall it is still characterized as a harsh, polar climate. Eureka’s mean annual air temperature is -19.7°C with a thaw season (mean air temperatures above 0°C) of only 3-6 weeks (data from Environment

Canada, 1947-2015). The 1947-2015 average July air temperature is 5.6°C, yet Pollard et al.

(2015) show that since 1980 the July average has risen to 6.2°C, and in the period from 2005-

2015 it rose to 7.7°C. The summers of 2011 and 2012 were both exceptionally warm (average

July air temperatures of 9.8°C and 8.7°C, respectively).

2.3) Quaternary history

The Queen Elizabeth Islands region, which comprises the Eureka Sound Lowlands, has a complicated and disputed late-Quaternary history. Most of the known Quaternary record for the region is drawn from raised marine deposits, as glacial indicators are scattered. Blake (1970) first

Figure 2.3.1: Diagram depicting the Innuitian Ice Sheet at LGM (from England et al. 2006). The ESL region is contained within the rectangle. In the ESL, ice flowing eastward from the Axel Heiberg alpine sector would have converged with ice flowing westward from the Ellesmere alpine sector. hypothesized the existence of widespread regional ice (Innuitian Ice Sheet) throughout the Late

Wisconsinan (most recent) glacial period, based on a corridor of maximum Holocene emergence from Eureka Sound to Bathurst Island. This entails a contiguous ice sheet spanning from the

Greenland Ice Sheet to the east and the Laurentide Ice Sheet to the south (Fig 2.3.1) The existence of an Innuitian Ice Sheet was challenged by England (1983; 1987) and Bell (1996), among others, due to a lack of primary glacial geological evidence. Their opposing view advocated for a restricted Late Wisconsinan ice cover, termed the Franklin Ice Complex. In this

Figure 2.3.2: Deglaciation of the Innuitian Ice Sheet (from England et al. 2006). The ESL region is contained within the rectangle. Deglaciation of the Eureka-Slidre Fiord area is estimated >8,500 BP, while the Gemini area is >7,500 yrs BP scenario, existing upland ice caps extended 10-40 km further downslope and other upland areas had ice caps – but much of the Queen Elizabeth Islands region remained unglaciated. England

(1983) proposed the existence of a full glacial sea throughout the Late Wisconsinan period that inundated low-lying areas. This thinking dominated Quaternary science in the Canadian High

Arctic for several decades. Work by Bednarski (1998), Dyke (1999) and Ó Cofaigh et al. (2000) re-instated the widespread ice hypothesis, with new geological evidence and studies that discounted a full-glacial sea. Even England, the leading proponent of a restricted ice cover, has

25 become a vocal supporter of a contiguous ice sheet. In the most recent comprehensive review,

England et al. (2006) describe an Innuitian Ice Sheet that comprised both alpine and lowland sectors. They estimate that retreat began around 11-12,000 years BP, with penetration of the sea into the Queen Elizabeth archipelago around 8,500 years BP (Fig. 2.3.2). Subsequent isostatic rebound has caused marine regression of up to 150 m (e.g. Holocene marine limit at Eureka is

143 m a.s.l.). The ability to reconstruct Quaternary events from non-glacial landforms such as fluvial deposits and ground ice would help in understanding the evolution of the High Arctic landscape. Ground ice, in particular, has potential for inferring Quaternary history due to its preservation of water and sediments – whose chemical and physical signatures reveal environmental conditions at the time of formation.

2.4) Permafrost and ground ice

The ESL are characterized by deep, continuous permafrost. According to Taylor (1991) the permafrost depth is 500 m, based on measurements from the abandoned Panarctic Oil well at the

Gemini site. The active layer is typically ~ 60 cm but varies due to microclimatic conditions.

Mean annual ground temperatures are stable at -16.5°C, and the depth of zero annual amplitude is ~15.4 m (Pollard et al. 2015).

Hodgson (1985) and Pollard (1991) reported that ground ice is an important constituent of surficial deposits on the Fosheim Peninsula of western Ellesmere Island. The most important in terms of volume and frequency are pore ice, ice-wedge polygons (commonly high-centre polygons) and tabular massive ice. Pollard (2000a) notes the strong association of massive ice with the Holocene sea level history and marine sediments. In general, volumetric ice contents in marine sediments exceed 50%. Past study focused on seven distinct areas with numerous

26 massive ice exposures and thermokarst features (Fig. 2.4.1). In the years since, the study of massive ground ice exposures has broadened to include more parts of the ESL (Pollard et al.

2015).

Pollard (2000a; 2000b) interprets the massive ice exposures on the Fosheim Peninsula as intrasedimental ice, as they are all located below Bell (1996)’s Holocene marine limit for the

Eureka (~143 m a.s.l.) and are conformably overlain by marine sediments (the preservation of glacial ice under marine sediments is deemed improbable). Importantly, sediment inclusions in the massive ice are similar to the sediment above the ice, suggesting the sediment predates the ice. Inspired by Rampton’s (1988) model of ground ice aggradation for the Tuktoyaktuk

Coastlands, Pollard and Bell (1998) propose a three-phase model of massive ground ice aggradation for the area: I) a full ice-cap and glacial-sea phase (14,000-9,500 BP)1, II) a period

1 England et al. (2006) have since ruled out a full-glacial sea. Instead, they propose that during the Last

Glacial Maximum (LGM), Eureka Sound was an ice saddle where ice flowing west from the Ellesmere

Ice Divide coalesced with ice flowing east from the Axel Heiberg alpine divide. Glacial isostatic adjustment models predict the Eureka Sound intermontane region would have been inundated by up to

1200 m of trunk ice; the maximum ice loading of the entire Innuitian Ice Sheet (Ó Cofaigh et al. 2000;

Simon et al. 2015). The ice sheet began to break up at its margins ~ 11-10 14C ka. Disintegration of the central part of the ice sheet began in earnest ~9.0 ka and the sea penetrated the Eureka Sound intermontane area between 8.5 and 8.0 ka, inundating the length of Eureka Sound and Nansen Sound

(England et al. 2006). Thus, Phase II and III of Pollard and Bell’s (1998) landscape model – that which concerns massive ground ice formation - remains valid in the context of currently-accepted Late

Quaternary chronology, +/- 1.0 ka.

27 of warmer temperatures and ice cap decay (9,500-7,500 BP) and III) the emergence of submerged lowland basins and aggradation of permafrost into the marine deposits (7,500 BP – present). The formation of ground ice is thought to have occurred only in the third phase.

Pollard et al. (2015) recognize upslope glacial meltwater as the major source for massive ground ice. They note the underlying, unconsolidated Tertiary bedrock would have been saturated by glacial meltwater and been an excellent aquifer. The geochemistry of massive ground ice in the region has yet to be extensively tested.

Massive ground ice area Location (lat., long.)

Eureka 80°00’N, 85°57’W south Slidre Fiord 79°56’N, 86°05’W

Slidre River (Hot Weather Creek; Gemini site) 79°59’N, 84°28’W; 79°58’N, 84°10’W

Blue Man Cape 79°44’N, 85°57’W

Eureka Sound 79°43’N, 84°30’W

south Fosheim 79°28’N, 84°20’W

May Point* 79°25’N, 84°29’W

Mokka Fiord/Buchanan Lake* 79°30’N, 87°30’W

Fig. 2.4.1: Areas with many natural exposures of massive ground ice identified by reconnaissance in 1991, 1992 and 1994, as discussed in Pollard (2000a) and via personal communication (*Axel Heiberg Island).

3. Research Objectives

The aim of my research is to infer the origin of tabular, massive ground ice in the Eureka

Sound Lowlands of Ellesmere and Axel Heiberg Island based on stable isotope geochemistry, ionic geochemistry and cryostratigraphic setting. Geochemical findings will be used to test the hypothesis first proposed by Pollard and Bell (1998), and since updated in

Pollard et al. (2015), that massive ice associated with marine sediments in the ESL is of segregated /intrasedimental origin and formed synchronously with marine regression in the early

Holocene. Based on the geochemical nature of the ice, this research will reconstruct the evolution of Holocene permafrost in raised marine deposits and climate in the Canadian High

Arctic. Additionally, this research will contribute towards understanding the extent and timing of deglaciation.

4. Methods

4.1) Introduction

This thesis is a product of a 2017 geocryological study of exposed massive ground ice in the

ESL. Permafrost and ground ice have been extensively studied on the Fosheim Peninsula since the first days of oil exploration in the 1970’s. Previous research focused on mapping and stratigraphic relationships (Hodgson and Nixon 1998; Pollard and Bell 1998; Pollard 2000a), quantifying ground ice volume (Couture and Pollard 1998), retrogressive thaw slump behaviour

(Ward and Pollard 2016), active layer depths (Bonnaventure et al. 2016), active layer detachments (Lewkowicz 2007) and ground ice microbiology (Steven et al. 2007). Before the current study, high-resolution geochemical profiles (notably of stable water isotopes and major ions) for ground ice bodies were lacking. Despite limited geochemical analysis (Pollard 2000a) there remained uncertainty surrounding ground ice formation processes in relation to the

Quaternary landscape history. The crux of this thesis is an analysis of ESL ground ice chemistry along an elevation gradient from Holocene marine limit (~143 m a.s.l.) to modern sea level. The elevation gradient can also be thought of as a chronological gradient – from oldest and fastest marine emergence (~143 m a.s.l.) to most recent and slowest emergence (modern sea level).

Extensive core samples collected during fieldwork in the summer of 2017 were sent to the

CryoLab at the University of Ottawa with frozen permafrost, unfrozen water and soil samples. In this section I will discuss the research design and methodology employed in this study, including: i) a review of the theory behind the use of stable isotopes and ionic geochemistry in ground ice studies, ii) study site selection and field sampling techniques and iii) laboratory techniques.

4.2) Geochemical theory – as applied to geocryology

4.2.1) Stable water isotopes

Chemical elements are defined by the number of protons in their atomic nuclei, i.e. their atomic number. Isotopes are variants of a particular element that have the same number of protons but different numbers of neutrons, adding atomic mass. Earth’s water – H2O - is largely composed of

1H and 16O, which are the dominant isotopic forms of each respective element, but other stable isotopes of H and O do occur naturally: 2H (also known as deuterium, or D), 17O and 18O. The ratios of D/1H and 18O/16O can be measured at various phases of the water cycle to identify different water masses and track hydrological processes (17O is much less abundant than 18O and acts similarly in the water cycle, so is rarely measured). The use of stable water isotopes in hydrology exploits the reduced mobility of the rarer, heavier isotopes (D and 18O) in isotopic exchange (equilibrium) reactions and kinetic (non-equilibrium) reactions. These isotopic fractionations are most often related to phase changes in the water cycle (Gat 1996).

The oceans contain 96.5% of Earth’s H2O and constitute the largest reservoir in the hydrologic cycle (Hendriks 2010). For this reason, oceanic waters are used as the reference for stable water isotope ratios, specifically the Vienna Standard Mean Ocean Water (VSMOW). In a water

18 3 sample (x), D and O are measured as δ(‰) = (Rx/RVSMOW – 1) × 10 , where R is the atom ratio

D/1H and 18O/16O, respectively2. In global precipitation, δD and δ18O are correlated along what

Dansgaard (1964) called the Global Meteoric Water Line (GMWL), defined as δD = 8(δ18O) +

10 (Fig. 4.2.1). At a local scale, the regression slope may deviate slightly from 8 due to effects of

2 In the ocean, the ratios of the isotopes D/H and 18O/O are 155.95 × 10-6 and 2005.2 × 10-6, respectively, where H and O without a mass assignment refer to the assembly of all isotopic species (Baertschi 1976).

different moisture sources and trajectories of

weather systems over land masses – this is

referred to as the Local Meteoric Water Line

(LMWL). The International Atomic Energy

Agency (IAEA) maintains a database of

monthly δ18O and δD samples from stations

around the globe called the Global Network of

18 Figure 4.2.1: Meteoric relationship between O and D Isotopes in Precipitation (GNIP) and presents in precipitation across Earth’s latitudes (from Clark & Fritz 1997). LMWLs for each station. At Eureka, δ18O and

δD values in modern precipitation range from -42.87‰ to -16.37‰ and -329.86‰ to -162.7‰,

respectively (Fig. 4.2.2). The regression slope of the LMWL at Eureka Weather Station is 7.43

(Fig. 4.2.3).

1988-12-23 1990-05-07 1991-09-19 1993-01-31 1994-06-15 0 0

-5 -50 -10

-15 -100

-20 -150 -25 O (‰) O 18 -200 (‰) δD δ -30

Figure-35 x: Monthly stable water isotopes in precipitation at Eureka Weather Station, Nunavut, Canada-250 from-40 January 1989-December 1993 (IAEA/WMO 2018). 18O/16O ratios in blue, D/H ratios in red. -300 -45

-50 -350

δ18O δD

Figure 4.2.2: Stable water isotopes in precipitation at Eureka Weather Station collected on a monthly basis from January 1989 to December 1993 (IAEA/WMO 2018).

Since the 1970’s, stable water isotopes (δD and δ18O) have been used to investigate the origin of

ground ice (Stuiver et al., 1976). This analysis relies on isotope fractionation that occurs during

0 phase changes. If ice (or snow) formed

-50 by condensation of atmospheric water -100 (e.g. a glacier) is buried and preserved -150

-200 in permafrost, the slope of the

δD (‰) δD -250 regression of δD on δ18O will plot close

-300 y = 7.4336x - 9.1398 to the LMWL (typically ~ 8). However, R² = 0.9696 -350 -45 -35 -25 -15 if the ice formed under slow freezing δ18O (‰) (i.e. equilibrium freezing), as would

Figure 4.2.3: Local meteoric water line (LMWL) for Eureka theoretically occur in Weather Station, Nunavut, Canada. Data points are from precipitation samples measured monthly at Eureka from January 1989-December 1993 (IAEA/WMO 2018). segregated/intrasedimental ice, the

regression slope should be significantly lower (typically 3-7) (Lacelle 2011). Essentially, this is

due to the fact that the equilibrium fractionation factors between water and ice are different for D

and 18O. Similarly, deuterium excess (d) can be calculated as d = δD-8*δ18O (Dansgaard 1964), a

measure that reflects changes in temperature and relative humidity of the air in the source area.

Stable water isotope profiles in an ice body help identify its origin, along with unconformities

that reflect secondary processes such as post-depositional partial melting and refreezing. For

instance, Lacelle et al. (2011) used these techniques to determine sublimation and diffusion rates

in massive ice in the dry valleys of Antarctica.

The relative age of ground ice can be estimated with δ18O values, based on knowledge of

Quaternary climatic conditions. According to Michel (2011), Holocene segregated ice lenses and

ice wedges typically have δ18O values between -18 to -22‰ and -22 to -25‰, respectively. On

33 the other hand, ice formed in the colder Pleistocene can be as low as -36‰. Generally, more negative values are indicative of a colder climate. The δ18O values are also useful in assessing whether seawater is the source of massive ice. Ground ice formed from a marine source would be expected to have much higher δ18O values, approaching VSMOW, whereas glacial meltwater would have strongly depleted 18O.

4.2.2) Major ion geochemistry

In the study of ground ice, traditional stable water isotopes work best when combined with supplementary hydrochemical analyses. In natural waters, more than 90% of dissolved solids can

+ 2+ + 2+ 2- - - 2- be attributed to eight ionic species: Na , Ca , K , Mg , SO4 , Cl , HCO3 and CO3 (Fetter

2001). When groundwater flows through an aquifer it acquires a representative chemical signature due to interactions between the water and the lithologic matrix. Similarly, there are unique chemical signatures associated with meteoric waters and seawater. These signatures are known as hydrochemical facies and help to determine the source of ground ice. Fritz et al. (2011) measured ionic concentrations and electrical conductance in samples from Late Pleistocene massive ground ice on Herschel Island, Yukon Territory. When interpreted alongside stable water isotope data, the authors were able to confidently identify the water source as full-glacial water. Slagoda et al. (2012) used similar techniques on massive ice on the Yamal Peninsula,

Siberia. They identified one ice body as having a marine water source due to elevated δ18O and

Na/Cl values, while most other ice bodies in the area were linked to a fresh groundwater source.

Depth profiles of ion concentrations are also useful for distinguishing cryostratigraphy and unconformities. MacKay (1983) and Lacelle et al. (2014) traced thaw unconformities in permafrost in western Arctic Canada with this technique. In all the above-mentioned studies (and

34 this thesis), the analysis is performed on meltwater after thawing ice-bearing permafrost, with concentration expressed as solute per litre of solution.

4.2.3) Occluded gases in ground ice

Cardyn et al. (2007) used the O2/Ar and N2/Ar ratios of occluded gases trapped in massive ice to determine its origin. Air trapped in firn (i.e. in the case of buried glacier ice or snow bank) is expected to have similar molar ratios to the atmosphere. On the other hand, intrasedimental ice formations can be distinguished based on the differential solubilities of gases in water (Henry’s

Law), thus making this technique a valuable tool for identifying the origin of massive ice bodies.

Extracting the gases from ice samples is tricky; the wet extraction method works best for O2, N2 and Ar gases (Sowers et al. 1997). Examples of successful applications of the occluded gas technique (used concordantly with stable water isotopes) to infer the origin of ground ice in the

Canadian Arctic include Lacelle et al. (2009), Lauriol et al. (2010) and St-Jean et al. (2011).

A recent advance in the field of ground ice geochemistry has been the calculation of noble gas ratios. Inert noble gases (He, Ne, Ar, Kr, Xe) comprise less than 1% of the atmosphere. An advantage of noble gases is that unlike O2, N2 or CO2 they are non-reactive and only affected by physical processes. Utting et al. (2016) exploited the different solubilities of Ar, Kr and Xe in water to assess the source of massive ice samples from the Canadian Arctic. The Kr/Xe ratio was found to be most effective at distinguishing a buried-air versus freezing-exsolution origin. The

Ar/Xe was also useful due to the larger difference in solubilities between the two gases, making it potentially more diagnostic of ice formation processes. The Ar/Xe ratio must be interpreted carefully, however, due to the relative mobility of Ar.

4.3) Field methods

4.3.1) Site selection

While massive tabular ice is believed to be widespread across the ESL, it is most easily sampled from natural exposures. In the summer of 2017, several potential sampling sites (exposed massive ice in retrogressive thaw slumps) were identified during an aerial survey by helicopter, based on previous thermokarst mapping by Pollard (2000) and Pollard et al. (2015). Sites were selected that would cover a large spatial extent of the ESL (to account for landscape heterogeneity) and would create a rough elevation transect – from Holocene marine limit (~143 m a.s.l.) to modern sea level (Slidre Fiord). Primary areas of study were the Eureka vicinity, the

Slidre River and Mokka Fiord (see fig. 4.3.2, 4.3.3). Specifically, this consisted of four exposures on the Ellesmere side and one exposure on the Axel Heiberg side of Eureka Sound

(Fig. 4.3.1). In addition, I had geochemical data from an exposure on the South Slidre River sampled in a 2015 reconnaissance survey by D. Lacelle and W.H. Pollard (personal communication). The stratigraphic position of each exposure was described and recorded with particular attention given to contact relationships and boundaries, as well as the site’s elevation relative to the Holocene marine limit.

Figure 4.3.1: Schematic diagram of elevation transect of massive ice exposures on the Ellesmere Island side of the ESL. Marine emergence dates are based on Simon et al. (2015)’s revised Innuitian Ice Sheet model for relative sea level (RSL) in the Eureka area during the Holocene. The geophysical model, which is currently the most up-to-date and rigorous reconstruction available, is constrained by field RSL data derived from the ages and elevations of radiocarbon-dated material, such as marine shells, driftwood and mammal bones.

Figure 4.3.2: Map of Eureka Sound Lowlands (1:400,000)

Figure 4.3.3: Map of Eureka vicinity (1:50,000)

4.3.2) Field sampling

At each exposure, permafrost and active layer (AL) samples were collected at several locations along the headwall to get a complete picture of ground ice chemistry. Two strategies were used to extract cores from massive ground ice:

1. Where there was < 2 m of sediment overlying massive ice, a pit was dug to the top of the

permafrost table. AL soil was sampled with a trowel at 5 cm intervals. AL samples were

stored in sealed polyethylene bags. A SIPRE corer attached to a handheld motorized

power head was used to extract 3” diameter vertical cores from the permafrost. The cores

ranged from 40-160 cm in length, with the goal to sample as much of the massive ice

body as possible. The frozen cores were preserved in 5-75 cm segments, wrapped in

aluminum foil, labelled and stored in thermally-insulated boxes.

2. If overlying sediments were > 2 m, cores were drilled directly into the ablating massive

ice surface by constructing a safe drilling platform with earth material or wooden palettes

above the mudflow. Frozen samples were preserved in the same manner as described

above.

Frozen cores were transported to the CryoLab at the University of Ottawa. Permafrost cores were supplemented with individual grab samples with an ice axe directly from the (cleaned) ablating ice face and overlying frozen sediments. Care was taken to discard surface ice from these samples to avoid contamination from refrozen ablation meltwater. These individual grab samples were melted in polyethylene bags and stored in 20 ml high-density polyethylene (HDPE) vials.

Several large ice blocks (~40 cm2) were also sampled, kept frozen and transported back to the

CryoLab for occluded gas analysis.

In addition to the elevation transect of exposed ground ice, water was sampled (stored in 20 ml

HDPE vials) from several natural sources in the ESL: Slidre Fiord (79°58’17.8”N,

85°41’50.7”W), Mokka Fiord (79°35’6.9”N, 86°46’33.2”W), two late-lying snowbanks

(79°59’58.3”N, 85°47’47.4”W; 79°59’52.8”N, 85°46’20.0”W) and Black Top Creek

(79°58’50.4”N, 85°4’36.4”W).

Blocks of ice (30 cm2) from a collapsing pingo at a river mouth where an unnamed stream reaches a side arm of Mokka Fiord (79°33’7.6”N, 86°47’58.2”W).

Finally, samples were collected from ice wedges that were exposed in cross-section in the headwalls of slumps. “Edge” and “Centre” samples were taken, representing older and younger ice, respectively. The ice was left to melt in sealed polyethylene bags in the field, and then quickly poured into dark HDPE bottles.

4.4) Laboratory methods

4.4.1) Active layer (AL) samples

The 5 cm interval AL soil samples were analyzed for gravimetric moisture content (%), colour

(dry), loss-on-ignition (%), texture (sand, silt, clay %), δ13C (‰), bulk organic C (%), δ15N (‰) and total N (%). AL samples were weighed, dried at 105°C for 24 hours and then re-weighed.

Gravimetric moisture content (GMC) was calculated as:

GMC (%) = (Pre-dry mass – Post-dry mass)/Post-dry mass * 100

The colour of the dried soil was recorded using the Munsell colour system. It was then ground with a mortar and pestle to prepare for loss-on-ignition Name of texture class Diameter limits (mm) (LOI) analysis with a thermogravimetric analyser (Leco Clay <0.02 TGA701). To prepare for texture analysis, 5 ml of AL soil Silt 0.02-0.05 was dispersed with 40 ml of sodium hexametaphosphate. A Sand 0.05-2 laser diffraction analyser (Microtrac S3500) was used to Table 4.4.1: USDA texture class boundaries assess the proportions of sand, silt and clay-sized particles.

Bulk C and N analyses were performed by Y. Wang at the University of Ottawa on acidified sediment samples with an Elementar VarioEL III instrument.

4.4.2) Permafrost cores

The objective was to create high-resolution depth profiles of the massive ice, mainly for isotope analysis. First, the permafrost cores were sliced longitudinally; one half was archived in a -17°C freezer while the remaining half was cut into 2 cm-thick sub-samples (“half-pucks”) with a specially-modified circular saw in the CryoLab (fig. x). I let the sub-samples melt in sealed 50

41 ml sterile centrifuge tubes. Sub-samples were analyzed for excess ice (%), volumetric ice content (%) and gravimetric ice content (%). Sediment was analyzed for texture (sand, silt, clay

%), loss-on-ignition (%), δ13C (‰), bulk organic C (%), δ15N (‰) and total N (%). Supernatant water was analysed for stable water isotopes (δD, δ18O, d excess), major cations (Ca2+, Fe2+, K+,

2+ + 2+ 2- - - Mg , Na , Sr ), anions (SO4 , Cl , NO3 ), dissolved organic carbon (DOC) and dissolved organic nitrogen (DON).

Figure 4.4.1: Circular saw with 0.8 mm diamond blade at the CryoLab, University of Ottawa. Cores were cut longitudinally, then at 2 cm intervals into “half-pucks”.

First, the thawed samples were weighed and then centrifuged for 5 minutes at 5000 rpm to help separate the mixture (soil and water). The volume of supernatant water was recorded. Whenever possible, the supernatant water (filtered through 0.45µm pore diameter) was put into 20 ml vials for stable water isotope and ion analyses. The remaining moist sediment was dried at 105°C for

24 hours, then re-weighed. This was used to calculate the gravimetric ice content (GMC, or W) with the same formula as outlined in Section 4.4.1. The volumetric ice content (fVi) was calculated according to the procedure given by Pollard and French (1980):

fVi = Vi/(Vi + Vs)

where: Vi ≈ W + (0.09W)

Vi = volume of ice

Vs = volume of soil solids

A Los Gatos Research (LGR) liquid water analyser coupled to a CTC LC-PAL autosampler was

18 16 18 used for simultaneous O/ O and D/H ratios for H2O. O and D are measured relative to

VSMOW (‰). The equation from Dansgaard (1964) was used to calculate d excess. Samples were analyzed for stable water isotopes whenever at least 2 ml of supernatant water could be extracted from the sub-samples.

2+ 2+ + 2+ + 2+ 2- - - Major cations (Ca , Fe , K , Mg , Na , Sr ) and anions (SO4 , Cl , NO3 ) were measured on supernatant water from every second sub-sample for at the geochemistry laboratory at the

Advanced Research Centre at the University of Ottawa. Prior to cation analysis, water samples were acidified to a pH of 2 with ultra-pure HNO3 acid. Anion samples were not acidified. Cation and anion samples were diluted with deionized water by a factor of 10. An inductively coupled

43 plasma atomic emission spectrometer (ICP-AES) measured concentrations of the ions, with duplicate samples run (analytical reproducibility < 5%). Ionic concentrations for the ground ice are presented in mg/l.

A Shimadzu TOC analyser (Vcsn model) was used to measure DOC/TN. First, meltwater sub- samples were acidified to a pH of 3 - to help remove dissolved inorganic carbon (DIC) by sparging with Ultra-Zero compressed air for 5 minutes prior to analysis. The DOC measured in this analysis is specifically Non-Purgeable Organic Carbon (NPOC), as opposed to DOC calculated as Total Carbon (TC) – Inorganic Carbon (IC). The analyser uses a platinum catalyst for high-temperature combustion (720°C) of acidified samples, then passes them through a non- dispersive infrared detector. It takes the best 3 of 5 sample injections and provides the mean, with a coefficient of variation <5%. TN is measured off peaks produced by the detection of nitrogen monoxide gas – which itself is the product of the decomposition of TN in the sample upon combustion. DOC/TN was only analyzed on select massive ice samples, ice wedges and environmental samples (e.g. snowbank).

All sediment samples from the cores were measured for LOI and select sediment samples for texture (sand, silt, clay %) using the same methods outlined in Section 4.4.1.

4.4.3) Ice wedge (IW) samples

The melted IW samples were filtered (0.45µm pore diameter) and stored in dark 1 L bottles.

Twelve IW samples for 14C-DOC analysis at the A.E. Lalonde AMS Laboratory at the University of Ottawa. The radiocarbon analyses were performed by S. Murseli, C. Jean and X. Zhao on a

3MV tandem accelerator mass spectrometer built by High Voltage Engineering (HVE), with

12,13,14C3- ions measured at a 2.5 MV terminal voltage on argon stripping. The standard 14C/12C

44 ratio was Ox-II. The fraction modern carbon (F14C) was measured as the ratio of the sample

14C/12C to the standard 14C/12C (Millard 2004). Following Stuiver and Polach (1977), radiocarbon ages are calculated as -8033ln(F 14C) and reported in 14C yr BP (BP = AD 1950).

4.4.4) Collection of ice blocks for occluded gas analysis

Ice samples were preserved as 30 cm3 frozen blocks for future analysis of occluded gases, following the methods of Cardyn et al. (2007), St-Jean et al. (2010) and Utting et al. (2016). The number of frozen samples that could be transported back to Ottawa were limited by logistical constraints.

5. Results

This section is organized into three sections. First, active layer (AL) and permafrost core data from massive ice exposures are presented on a site-by-site basis from lowest to highest elevation above sea level: Dump Slump, Station Creek, Black Top and Gemini (5.1). These data include both sediment and ice chemistry. Second, ground ice chemistry is compared with local water sources across the ESL (5.2). This allows for the ground ice to be situated within the context of the landscape, from a chemical perspective. Third, 14C-DOC results are presented (5.3), to constrain the age of ESL ground ice. Additional data is appended at the end of the thesis, specifically: a core from a massive ice exposure near Mokka Fiord, Axel Heiberg Island (App.

A), hydrochemical data from a collapsing pingo also on Mokka Fiord (App. B) and results of a micro-study compared ice sampling techniques (App. C).

5.1) Permafrost cores along elevation transect – Fosheim Peninsula – p. 44-77

In this section, massive ice cores and overlying AL sediment are presented on a site-by-site basis, in order of increasing elevation: Dump Slump (40 m a.s.l.), Station Creek (45 m a.s.l.),

Black Top (70 m a.s.l.) and Gemini (125 m a.s.l.). At every site except Station Creek, two cores were sub-sampled and analyzed. Ice chemistry was analyzed in each core. Sediment analyses (C,

N) were run only on select cores (marked with an asterisk in the following sections).

5.1.1) Dump Slump (2 cores: DS-1*, DS-2) – 40 m a.s.l.

The Dump Slump (DS) site is located at 79°58'50"N, 85°47'8.75"W. Core locations and physical characteristics are indicated in figs. 5.1.1a and 5.1.1b.

The mean bulk C content of DS-1 sediment is 2.15%. Sediment within massive ice has significantly higher bulk C content (2.49%) compared to AL soil (1.66%) (p=0.004). The mean

δ13C of DS-1 sediment is -25.03‰. There is no significant difference between the δ13C signature of AL sediment and sediment within massive ice. The mean bulk N content and δ15N of DS-1 sediment are 0.13% and 3.48‰, respectively, and there is no significant difference between AL and permafrost sediment for either variable. The mean C:N ratio is 16:1, with a significantly lower mean C:N ratio in AL soil (~14:1) compared to sediment within massive ice (~17:1)

(p=0.04). Table 5.1.1a presents sediment chemistry data for DS-1. Table 5.1.1b presents sediment texture for DS-1.

The volumetric ice content (fVi) of core DS-1 is high, with a mean of 94.0%, ranging from 75.4 to 99.1% (fig. 5.1.1c). Core DS-2 is also ice-rich, with a mean of 94.7%, ranging from 81.8 to

98.7% (fig. 5.1.1d).

Stable water isotopes are depleted in both DS-1 and DS-2 cores. In DS-1, δD ranges from -258.0 to -237.9‰ and δ18O from -33.3 to -31.1‰ (fig. 5.1.1c). In DS-2, δD ranges from -265.6 to -

245.7‰ and δ18O from -34.0 to -31.3‰ (fig. 5.1.1d). The δD-δ18O slopes of DS-1 and DS-2 are

7.78 and 7.26, respectively (figs. 5.1.1e, 5.1.1g). There is no significant relationship between dD and d in DS-1 (fig. 5.1.1f) and a weak negative relationship in DS-2 (fig. 5.1.1h). Total dissolved solids (TDS) in DS-1 ice range from 33 to 8833 mg/l, with a median of 160 mg/l. Depth profiles of cations and anions are presented in figs. 5.1.1i and 5.1.1j. The dominant ions are (Na++K+)/Cl-

(fig. 5.1.1k). DOC concentrations are given in Table 5.1.1c.

Gravimetric moisture content (%) (above) Figure 5.1.1a: Dump Slump (DS) massive ice exposure in a retrogressive 1 100 10000 thaw slump. Cores sampled vertically at 0 white Xs.

Active layer

50 (left) Figure 5.1.1b: Gravimetric water

Permafrost table content (GMC) of samples in AL and core sub-samples from massive ice (DS-1). Note discontinuity of approximately 1 m 100 Ice-rich sediments, not sampled between AL and permafrost samples. This is due to core sample being taken directly into ablating massive ice surface, i.e. at 150 the ‘floor’ of the slump.

Massive ice

Depth from ground surface (cm) surface ground from Depth 200

Sediment depth Munsell Organic Carbonates δ13C (‰) Bulk C (%) δ15N (‰) Bulk N (%) C:N ratio (cm) colour matter (%) content (%) 0-5 5Y 5/3 3.98 0.89 -24.95 0.85 4.65 0.09 9.4 5-10 5Y 4/2 5.28 1.05 -26.29 1.19 4.42 0.1 11.9

10-15 5Y 4/2 4.76 1.02 -24.73 1.2 4.03 0.09 13.3 15-20 5Y 4/2 4.9 0.94 -24.73 1.42 3.79 0.1 14.2 AL 20-25 5Y 4/2 4.98 0.88 -24.73 1.44 3.43 0.11 13.1 25-30 5Y 4/2 6.23 1.1 -24.63 1.76 3.37 0.12 14.7 30-35 5Y 5/2 7.23 1.35 -24.83 2.27 3.24 0.14 16.2

35-40 5Y 5/2 7.17 1.43 -24.6 2.44 3.07 0.14 17.4 40-45 5Y 5/2 6.78 1.34 -25.38 2.35 3.29 0.13 18.1

147 / / / -24.67 2.41 3.1 0.12 20.1 155 / / / -25.15 1.45 3.23 0.07 20.7

163 / / / -25.49 2.87 3.53 0.16 17.9 171 / / / -24.75 2.8 3.48 0.16 17.5 179 / / / -24.79 2.79 3.3 0.15 18.6 DS-1 185 / / / -24.61 2.75 3.34 0.16 17.2 / / / 189 -24.61 2.8 3.33 0.16 17.5 191 / / / -24.56 2.75 3.38 0.15 18.3 203 / / / -25.28 2.69 3.69 0.13 20.7 211 / / / -25.13 3.04 3.88 0.17 17.9 219 / / / -25.5 2.87 3.24 0.17 16.9 / / / 227 -25.62 2.23 3.39 0.15 14.9 231 / / / -25.78 0.95 2.31 0.15 6.3

Table 5.1.1a: Sediment data from AL samples and DS-1 permafrost core samples.

Depth (cm) Sand (%) Silt (%) Clay (%) Texture 5-10 41.91 35.35 22.72 Loam 195 4.43 61.11 34.46 Silty clay loam

Table 5.1.1b: Texture class data from AL samples and DS-1 permafrost core samples.

18 fVi δ O (‰) d excess (‰) 0.7 0.8 0.9 1 -33.5 -32.5 -31.5 7 9 11 13 140

150

160

170

180

190

200 Depth from ground surface (cm) surface ground from Depth

210

a) 220

230

240 b) c) d)

Figure 5.1.1c: Depth profiles of ice from DS-1 core: a) Photograph of DS-1 core (partial), 18 b) volumetric ice content (fVi), c) δ O and d) deuterium excess.

18 fVi δ O (‰) d excess (‰) 0.7 0.8 0.9 1 -34.5 -33 -31.5 2 4 6 8 140

150

160

170

180 Depth from ground surface (cm) surface ground from Depth

190

200

210 a) b) c)

Figure 5.1.1d: Depth profiles of ice from DS-2 core. a) volumetric ice content (fVi), b) 18 δ O and c) deuterium excess.

Figure 5.1.1e: Co-isotope plot (D- 18O) of DS-1 core samples.

y = 7.78x + 2.81 R2 = 0.964 n=45

Figure 5.1.1f: Co-isotope plot (d-D) of DS-1 core samples.

Figure 5.1.1g: Co-isotope plot (D- 18O) of DS-2 core samples.

y = 7.26x – 18.19 R2 = 0.99 n=28

Figure 5.1.1h: Co-isotope plot (d- D) of DS-2 core samples.

Ca2+ (mg/l) Mg2+ (mg/l) Fe2+ (mg/l) Na+ (mg/l) K+ (mg/l) 1 10 1001000 1 10 100 1000 0.001 0.1 10 1 100 10000 1 10 1001000 140

150

160

170

180

190

200

Depth from ground surface (cm) surface ground from Depth 210

220

230

240

Figure 5.1.1i: Depth profiles of major cations in DS-1 core.

2- - - SO4 (mg/l) Cl (mg/l) NO3 (mg/l) 1 10 1001000 1 100 10000 0 4 8 12 140

150

160

170

180 Figure 5.1.1k: Geochemical facies of massive ice from DS-1 core.

190

200 Depth (cm) DOC (mg/l) TN (mg/l) Depth from ground surface (cm) surface ground from Depth 191 2.36 1.9 210 231 2.41 0.4

220 Table 5.1.1c: Dissolved organic carbon (DOC) and total nitrogen (TN) present in supernatant water in DS-1 core. 230

240

FigureDepth 5.1.1j:(cm) DepthDOC profiles(mg/l) ofTN major (mg/l) anions in DS-1 core. 191 2.36 1.9 231 2.41 0.4

5.1.2) Station Creek (2 cores: SC-1*, SC-2) - 45 m a.s.l.

The Station Creek (SC) site is located at 79°59'53"N, 85°53'53"W. Fig. 5.1.2a shows the SC retrogressive thaw slump in its entirety and fig. 5.1.2b gives the exact locations where the cores and grab samples were taken from.

The fVi of core SC-1 is very high, with a mean of 93.1%, a minimum of 56.7% and a maximum of 100% (fig. 5.1.2c).

Stable water isotopes are highly variable in SC-1 ice, with δD ranging from -268.6 to -196.0‰ and δ18O from -34.9 to -25.5‰ (fig. 5.1.2c). The δD-δ18O slope is 7.52 (fig. 5.1.2d) and there is no significant relationship between δD and d (fig. 5.1.2e). TDS values vary from 28 to 1152 mg/l, with a median of 154 mg/l. Depth profiles of cations and anions are presented in figs.

5.1.2f and 5.1.2g. The dominant ions are (Na++K+)/Cl- (fig. 5.1.2h). DOC concentrations in SC-1 ice are presented in Table 5.1.2a. Stable water isotopes of grab samples SC3 1-10 (ice veins and massive ice) are given in Table 5.1.2b.

Figure 5.1.2a: Station Creek massive ice exposure in a large retrogressive thaw slump. Black rectangle indicates location of photograph in Fig. 5.1.2b below.

Figure 5.1.2b: Sampling sites: 2 vertical cores into massive ice (SC-1 & SC-2) marked by white Xs; ice samples from ice veins in reticulated clay (SC3 1-8); 2 block samples from massive ice (SC3 9 & 10). No active layer samples were taken.

18 fVi δ O (‰) d excess (‰) 0.5 0.75 1 -35 -33 -31 -29 -27 -25 2 4 6 8 101214 2

82 Depth from top of core (cm) core of top from Depth

102

122

142

c) a) b) d) Figure 5.1.2c: Depth profiles of ice from SC-1 core: a) Photograph of SC-1 core, b) volumetric ice content 18 (fVi), c) δ O and d) deuterium excess.

Figure 5.1.2d: Co-isotope plot (D- 18O) of SC-1 core samples.

y = 7.52x – 5.47 R2 = 0.98 n=64

Figure 5.1.2e: Co-isotope plot (d- D) of SC-1 core samples.

Ca2+ (mg/l) Mg2+ (mg/l) Sr2+ (mg/l) Fe2+ (mg/l) Na+ (mg/l) K+ (mg/l) 1 10 100 0 10 20 0.01 0.1 1 0.01 1 100 1 10 100 1000 1 10 100 0

Depth from top of core (cm) core of top from Depth 100

120

140

Figure 5.1.2f: Depth profiles of major cations in SC-1 core.

2- Cl- (mg/l) NO - (mg/l) SO4 (mg/l) 3 1 10 100 10 100 1000 0.1 1 10 0

Figure 5.1.2h: Geochemical facies of massive ice 60 from SC-1 core.

80 Depth (cm) DOC (mg/l) TN (mg/l) 54 6.13 2.38 56 4.25 1.84 58 2.34 1.18 Depth from top of core (cm) core of top from Depth 100 60 2.30 1.03 62 1.73 1.03 72 4.74 2.36 78 17.61 4.31 120 90 4.57 1.35 118 19.47 5.4 146 8.18 4.29 140 Table 5.1.2a: Dissolved organic carbon (DOC) and total nitrogen (TN) present in supernatant water in DS-1 core.

Figure 5.1.2g: Depth profiles of major cations in SC-1 core.

Sample δD (‰) δ18O (‰) d excess (‰) DOC (mg/l) TN (mg/l) SC3-1 -119.94 -16.74 14.00 / / SC3-2 -118.00 -16.50 14.03 / / SC3-3 -117.78 -16.52 14.38 / / SC3-4 -117.23 -16.40 13.95 / / SC3-5 -119.19 -16.26 10.90 / / SC3-6 -117.42 -15.83 9.26 1.79 1.42 SC3-7 -118.01 -16.14 11.09 / / SC3-8 -119.61 -16.31 10.86 / /

SC3-9 -195.82 -25.39 7.28 / / SC3-10 -262.13 -34.00 9.86 / /

Table 5.1.2b: Water chemistry from SC3 samples (see Fig. x for corresponding sampling location). SC3-1 to 8 are ice veins in reticulated clay layer directly above SC-1 and SC-2 massive ice. SC3-9 & 10 are clear, massive ice above the reticulated ice layer.

5.1.3) Black Top (2 cores: BT-1*, BT-2) – 70 m a.s.l.

The Black Top (BT) site is located at 79°59'8.17"N, 85°42'3.28"W. Fig. 5.1.3a displays the retrogressive thaw slump and sampling locations. Fig. 5.1.3b shows that the gravimetric moisture contents of BT-1 and BT-2 cores exhibit similar patterns.

The mean bulk C content of BT-2 sediment is 2.72%. Permafrost sediment has a significantly higher bulk C content (2.88%) compared to AL soil (2.27%) (p<0.001). The mean δ13C of BT-2 sediment is -25.12‰. There is no significant difference between the δ13C signature of AL sediment and permafrost sediment. The mean bulk N content and δ15N of BT-2 sediment are

0.14% and 2.77‰, respectively. The δ15N of AL soil is significantly enriched (3.14‰) compared to permafrost soil (2.64‰) (p<0.001). The mean C:N ratio is 19:1, with a significantly lower mean C:N ratio in AL soil (16.5:1) compared to sediment within massive ice (~20:1) (p<0.001).

Sediment chemistry data is presented in Table 5.1.3a.

The fVi of cores BT-1 and BT-2 are moderately high with means of 71% and 57%, respectively

(figs. 5.1.3c and 5.1.3d). BT-1 fVi values range between 39.0 and 89.7%, while BT-2 ranges between 14.8 and 85.8%.

Stable water isotopes are more enriched in BT-1 and BT-2 than cores from other sites. In BT-1,

δD ranges from -213.2 to -171.6‰ and δ18O from -27.8 to -20.8‰ (fig. 5.1.3d). In BT-2, δD ranges from -187.2 to -175.8‰ and δ18O from -24.5 to -22.7‰ (fig. 5.1.3c). The δD-δ18O slopes of BT-1 and BT-2 are 6.52 and 5.75, respectively (figs. 5.1.3e and 5.1.3g). There are significant negative relationships between δD and d in both cores (figs. 5.1.3f and 5.1.3h). BT-2 ice has very high TDS concentrations, ranging from 6,913 to 34,152 mg/l, with a median of 14,094 mg/l.

Depth profiles of cations and anions are presented in figs. 5.1.3i and 5.1.3j. The dominant cation

64 is Na+. The dominant measured anion is Cl-, although charge balance analysis suggests that

- HCO3 (unmeasured) may be present in high concentrations as well.

Gravimetric moisture content (%) (above) Figure 5.1.3a: Black 1 100 10000 Top (BT) massive ice 0 exposure in a retrogressive thaw slump. Cores sampled vertically at white Xs. Active layer

(left) Figure 5.1.3b: 50 Permafrost table Gravimetric water content (GMC) of samples in AL and core sub-samples from massive ice (BT-1 & BT-2). These are BT-1 continuous cores from the 100 BT-2 ground surface.

Icy sediments/massive ice 150 Depth from ground surface (cm) surface ground from Depth

200

Sediment Munsell colour OMC (%) Carbonates (%) δ13C (‰) Bulk C (%) δ15N (‰) Bulk N (%) C:N ratio depth (cm) 0-5 5Y 4/3 5.93 0.94 -24.72 1.5 4.22 0.12 12.5 5-10 5Y 4/2 6.97 1.11 -24.76 1.79 3.76 0.12 14.9 10-15 5Y 4/2 6.37 1.19 -24.75 1.97 3.6 0.12 16.4 15-20 5Y 4/2 6.77 1.26 -25.47 2.04 3.66 0.13 15.7 20-25 5Y 4/2 6.53 1.47 -24.49 1.99 3.22 0.12 16.6 25-30 5Y 4/2 7.04 1.52 -24.68 2.19 3.29 0.12 18.3 30-35 5Y 5/2 8.47 1.86 -25.21 2.32 3.38 0.14 16.6 35-40 5Y 5/2 10.64 1.77 -25.27 2.52 2.95 0.15 16.8 40-45 5Y 5/2 9.41 1.84 -24.67 2.55 2.77 0.15 17.0 45-50 5Y 5/2 9.01 1.76 -25.15 2.56 2.41 0.16 16.0 50-55 5Y 5/2 7.25 1.68 -24.88 2.52 2.32 0.14 18.0 55-60 5Y 4/2 6.71 1.74 -24.96 2.69 2.24 0.15 17.9 60 / 7.53 1.79 -25.49 2.85 2.94 0.16 17.8 62 / 7.76 1.78 64 / 8.48 1.81 -25.08 3.49 2.94 0.17 20.5 66 / 5.33 1.69 68 / 6.36 1.61 -25.43 2.44 2.79 0.13 18.8 70 / 7.06 1.58 72 / 6.96 1.57 -25.35 2.69 2.88 0.14 19.2 74 / 6.85 1.79 76 / 7.15 1.6 -25.54 2.71 2.94 0.14 19.4 78 / 7.25 1.74 80 / 8.34 1.71 -24.91 3.62 2.84 0.16 22.6 82 / 7.41 1.63 84 / 7.25 1.59 -25.44 2.72 2.9 0.14 19.4 86 / 7.19 1.64 88 / 7.42 1.6 -25.44 2.8 2.88 0.14 20.0 90 / 7.24 1.66 92 / 7.23 1.64 -25.46 2.89 2.88 0.15 19.3 94 / 7.04 1.6 96 / 7.57 1.65 -25.39 3.1 2.86 0.16 19.4 98 / 7.66 1.62 100 / 7.29 1.61 -25.53 2.89 2.89 0.15 19.3 102 / 7.3 1.61 104 / 7.05 1.64 -25.51 2.92 2.86 0.15 19.5 106 / 6.95 1.59 108 / 6.82 1.59 -25.42 2.85 2.94 0.14 20.4 110 / 6.93 1.57 112 / 7.03 1.59 -25.42 2.86 2.85 0.14 20.4 114 / 7.41 1.73 116 / 7.69 1.78 -25.4 2.95 3.03 0.16 18.4 118 / 7.33 1.94 120 / 6.69 1.52 -25.5 2.85 2.93 0.15 19.0 122 / 7.2 1.6 124 / 7.76 1.69 -25.35 2.9 2.94 0.16 18.1 126 / 7.67 1.65 128 / 6.93 1.55 -25.39 2.83 2.94 0.14 20.2 130 / 7.2 1.57 132 / 6.82 1.65 -25.52 2.87 2.94 0.15 19.1 134 / 7.27 1.54 136 / 7.32 1.7 -24.97 2.9 2.3 0.13 22.3 138 / 7.17 1.59 140 / 7.03 1.63 -25.1 2.73 2.17 0.13 21.0 142 / 6.93 1.6 144 / 7.42 1.6 -24.94 3.11 2.4 0.14 22.2 146 / 7.35 1.58 148 / 6.98 1.54 -25.03 2.78 2.3 0.13 21.4 150 / 7.15 1.58 152 / 7.45 1.62 -25.09 2.67 2.19 0.13 20.5 154 / 7.61 1.66 156 / 7.36 1.61 -24.96 2.86 2.66 0.14 20.4 158 / 7.41 1.62 160 / 7.51 1.59 -24.96 2.89 2.35 0.14 20.6 162 / 7.69 1.67 164 / 7.74 1.71 -24.88 2.92 2.4 0.15 19.5 166 / 7.16 1.59 168 / 7.32 1.56 -24.97 2.72 2.4 0.14 19.4 170 / 7.38 1.54 172 / 7.45 1.58 -24.97 2.89 2.36 0.14 20.6 174 / 7.65 1.5 176 / 7.68 1.67 -24.83 2.95 2.29 0.14 21.1 178 / 7.29 1.56 180 / 7.59 1.57 -24.88 2.94 2.35 0.14 21.0 182 / 7.88 1.78 184 / 7.6 1.76 -24.9 2.83 2.55 0.13 21.8 186 / 7.97 1.84 188 / 7.86 1.69 -24.93 3.12 2.46 0.14 22.3 190 / 7.27 1.48 192 / 7.44 1.43 -24.95 2.9 2.4 0.13 22.3 194 / 7.13 1.31 196 / 7.06 1.45 -24.94 3.05 2.47 0.13 23.5 198 / 7.27 1.45 200 / 7.14 1.44 -24.9 2.89 2.32 0.13 22.2 202 / 7.01 1.44 204 / 7.06 1.39 -25.12 2.62 2.4 0.14 18.7 206 / 7.24 1.42 208 / 7.08 1.42 -25.1 2.57 2.53 0.14 18.4

Table 5.1.3a: Sediment data from AL samples and BT-2 permafrost core samples

18 fVi δ O (‰) d excess (‰) 0 0.25 0.5 0.75 1 -26 -25 -24 -23 -22 3 6 9 12 60

100

120

140 Depth from ground surface (cm) surface ground from Depth

160

180

200

a) b) c)

Figure 5.1.3c: Depth profiles of ice from BT-2 core: a) Photograph of BT-2 core (partial), 18 b) volumetric ice content (fVi), c) δ O and d) deuterium excess.

18 fVi δ O (‰) d excess (‰) 0 0.25 0.5 0.75 1 -30 -28 -26 -24 -22 -20 -8 -4 0 4 8 12 70

100

110

120

130 Depth from ground surface (cm) surface ground from Depth

140

150

160

a) b) c)

Figure 5.1.3d: Depth profiles of ice from BT-1 core: a) volumetric ice content (fVi), b) δ18O and c) deuterium excess.

Figure 5.1.3e: Co-isotope (D-18O) plot of BT-2 core samples.

y = 5.75x - 45.19 R2 = 0.81 n=63

Figure 5.1.3f: Co-isotope (d-D) plot of BT-2 core samples.

Figure 5.1.3g: Co-isotope (D-18O) plot of BT-1 core samples.

y = 6.52x – 31.96 R2 = 0.96 n=45

Figure 5.1.3h: Co-isotope (d-D) plot of BT-1 core samples.

Ca2+ (mg/l) Mg2+ (mg/l) Sr2+ (mg/l) Fe2+ (mg/l) Na+ (mg/l) K+ (mg/l) 10 100 10 100 0.1 1 10 0.01 1 100 1000 10000 100000 10 100 60

100

120

140

Depth from ground surface (cm) surface ground from Depth 160

180

200

Figure 5.1.3i: Depth profiles of major cations in BT-2 core.

2- - - SO4 (mg/l) Cl (mg/l) NO3 (mg/l) 1 10 100 100 1000 10000 0 2 4 6 60

100

120 Figure 5.1.3k: Geochemical facies of massive ice from BT-2 core.

140

Depth from ground surface (cm) surface ground from Depth 160

180

200

Figure 5.1.3j: Depth profiles of major anions in BT-2 core.

5.1.4) Gemini (2 cores : G-1, G-3*) – 120 m a.s.l.

The location of the Gemini site is 79°58’N, 84°10’W. Fig. 5.1.4a shows the sampling site of G-1 and its consistently high gravimetric moisture content. Fig. 5.1.4b shows the sampling site of G-

3, which (unlike G-1) is at the upper stratigraphic contact of the Gemini massive ice body. Fig.

5.1.4c presents gravimetric moisture content of G-3.

Mean bulk C content of sediment in the G-3 core is 2.99%. Mean δ13C is -25.5‰. There is no significant difference in C content or isotopic signature between sediment from above or within the massive ice. Mean bulk N content is 0.22% and the mean δ15N is 2.54‰. The mean C:N ratio is 14:1. N analyses were only done on sediment within the massive ice. Table 5.1.4a presents sediment chemistry data. Table 5.1.4b gives some texture analyses on G-3 sediment.

Massive ice in the G-1 and G-3 cores have very high fVi, with means of 98% and 96%, respectively (excluding samples of sediment above massive ice). There is very little variation in fVi within the massive ice of G-1 and G-3. G-1 ranges between 97.0 and 98.9% (fig. 5.1.4d), while G-3 ranges between 87.4 and 97.6% (fig. 5.1.4e).

Stable water isotopes are consistently depleted. In G-1, δD ranges from -263.3 to -249.1‰ and

δ18O from -33.9 to -32.1‰ (fig. 5.1.4d). In G-3, δD ranges from -243.0 to -223.5‰ and δ18O from -31.5 to -29.8‰ (fig. 5.1.4e). The δD-δ18O slopes of G-1 and G-3 are 7.33 and 8.43, respectively (figs. 5.1.4f and 5.1.4h). There is no significant relationship between δD and d in either G-1 or G-3 (figs. 5.1.4g and 5.1.4i). The TDS concentration in G-3 ice is low, between 25 and 606 mg/l, with a median of 36 mg/l. Depth profiles of cations and anions are presented in figs 5.1.4j and 5.1.4k. The dominant cation is Ca2+ and a charge balance analysis implies that

- HCO3 is the dominant anion (fig. 5.1.4l). Table 5.1.4c gives selected DOC concentrations in G-3 ice.

Gravimetric moisture content (%)

0.1 1 10 100 1000 0

50 Depth from permafrost table (cm) 60

80 b) a) 90

Figure 5.1.4a: a) Photograph of Gemini-1 sampling site. Core was sampled directly into massive ice body after removal of active layer sediment above. Overlying sediment is believed to be an inactive mudflow from a previous thaw slump further upslope. b) Gravimetric moisture content (GMC) of core sub-samples from massive ice (G-1).

G-1 Core

100 m

Gravimetric moisture content (%) (above) Figure 5.1.4b: Photograph of Gemini-3 0.1 1 10 100 1000 sampling site at a retrogressive thaw slump. AL 0 sediment removed – thus the core begins at top of permafrost table.

20 (left) Figure 5.1.4c: Gravimetric moisture content (GMC) of core sub-samples from G-3

core. The top ~20 cm of the core are frozen silty clay loam that overlie the massive ice 40 body.

80 Depth from permafrost permafrost from table (cm) Depth

100

120

Sediment Organic matter Carbonates (%) 13C Bulk 15N Bulk N C:N ratio depth content (%) (‰) C (%) (‰) (%) (cm) 2 8.07 1.95 -24.59 4.51 / / / 4 8.08 1.76 -24.78 3.97 / / / 6 9.07 1.64 -25.12 3.75 / / / 8 7.44 1.49 -26.13 3.92 / / / 10 6.29 1.27 -25.13 2.82 / / / 12 8.86 1.61 -25.39 2.89 / / / 14 9.49 1.73 / / / 16 9.07 1.76 -25.1 2.52 / / / 18 8.63 1.67 -27.05 2.42 / / / 20 8.67 1.72 -27.25 2.32 / / / 22 9.69 1.84 -25.57 2.6 / / / 24 / / -25.53 1.79 / / / 26 / / -25.87 2.82 / / / 34 / / -25.77 2.74 / / / 42 / / -25.73 2.33 / / / 50 / / -25.65 2.62 / / / 58 / / -24.96 3.05 2.78 0.23 13.26 66 / / -25.68 3.06 2.88 0.23 13.30 74 / / -25.42 3.22 2.53 0.24 13.41 82 / / -25.14 3.54 2.92 0.24 14.75 90 / / -24.72 2.95 2.5 0.2 14.75 98 / / -25.16 3.3 2.4 0.22 15 106 / / -25.26 2.9 2.27 0.21 13.80 114 / / -25.75 2.81 2.03 0.21 13.38

Table 5.1.4a: Sediment data from G-3 permafrost core samples.

Depth (cm) Sand (%) Silt (%) Clay (%) Texture 14 1.5 61.8 36.7 Silty clay loam Table 5.1.4b: Texture class data from G-3 permafrost core samples.

fVi δ18O (‰) d excess (‰) 0.4 0.6 0.8 1 -35 -34 -33 -32 -31 2 4 6 8 1012 0

50 Depth from permafrost (cm) table permafrost from Depth

a) c) d) 90 b)

Figure 5.1.4d: Depth profiles of ice from G-1 core: a) Photograph of G-1 core, b) 18 volumetric ice content (fVi), c) δ O and d) deuterium excess.

18 fVi δ O (‰) d excess (‰) 0 0.5 1 -32 -31 -30 -29 4 6 8 10 12 0

60 Depth from permafrost (cm) table permafrost from Depth

100

120 a) b) c) d)

Figure 5.1.4e: Depth profiles of ice from G-3 core: a) Photograph of G-3 core, b) 18 volumetric ice content (fVi), c) δ O and d) deuterium excess.

Figure 5.1.4f: Co-isotope (D-18O) plot of G-1 core samples.

y = 7.33x + 22.80 R2 = 0.88 n=43

Figure 5.1.4g: Co-isotope (d-D) plot of G-1 core samples.

Figure 5.1.4h: Co-isotope (D-18O) plot of G-3 core samples.

y = 8.43x + 22.80 R2 = 0.96 n=47

Figure 5.1.4i: Co-isotope (d-D) plot of G-3 core samples.

Ca2+ (mg/l) Mg2+ (mg/l) Fe2+ (mg/l) Sr2+ (mg/l) Na+ (mg/l) K+ (mg/l) 0 5 10 15 20 02468 0.1 1 10 0 0.05 0.1 1 10 100 1000 0 4 8 12 0

Depth from permafrost (cm) table permafrost from Depth 80

100

120

Figure 5.1.4j: Depth profiles of major cations in G-3 core.

2- - - SO4 (mg/l) Cl (mg/l) NO3 (mg/l) 0 15 30 0 15 30 0 1 2 3 4 0

Figure 5.1.4l: Geochemical facies of massive ice 60 from G-3 core.

Depth from permafrost (cm) table permafrost from Depth 80

Depth (cm) DOC (mg/l) TN (mg/l) 36 4.40 2.56 100 96 3.26 1.76

Table 5.1.4c: Dissolved organic carbon (DOC) and total nitrogen (TN) present in supernatant water in G-3 core. 120

Figure 5.1.4k: Depth profiles of major anions in G-3 core.

5.2) Eureka Sound Lowlands environmental water chemistry This section provides an overview of stable water isotopes and major cations/anions of ice and water sources across the ESL (including ground ice). Table 5.2.1 gives stable water isotopes and major ions of local water sources: seawater, stream water and snow. Fig. 5.2.1 compares the ionic concentrations of massive ground ice. Fig. 5.2.2 compares the hydrochemical facies of ESL massive ground ice, along with some of the local sources referred to in Table 5.2.1. Fig. 5.3.3 compares the dominant ions of various ground ice and local water sources. Fig. 5.3.4 compiles the stable water isotope values presented in this thesis with precipitation data from the IAEA.

18 2+ 2+ + 2+ + 2+ 2- - - Source δD δ O d Ca Fe K Mg Na Sr SO4 NO3 Cl ‰ mg/l Slidre Fiord -98.35 -13.21 7.34 117.0 <0.01 79.61 199.3 1621.1 1.33 595.8 nd 4414.6

BT Creek -189.36 -24.83 9.24 24.47 <0.01 1.48 10.21 10.84 0.08 73.00 0.08 13.51

Late-lying -220.23 -28.77 9.93 0.99 0.08 0.69 0.28 3.12 <0.01 0.55 0.12 1.70 snowbank Table 5.2.1: Stable water isotopes and major ions in local water sources: Slidre Fiord (sampled at 79°58’17.8”N, 85°41’50.7”W), Black Top (BT) Creek (sampled at 79°58’50.4”N, 85°4’36.4”W) and a late- lying snowbank near the Eureka airstrip (79°59’58.3”N, 85°47’47.4”W).

Total dissolved solids (mg/l) 1 10 100 1000 10000 100000

Black Top Dump Slump Station Creek Gemini Slidre Fiord water Snowbank

Figure 5.2.1: Total dissolved solids (TDS) in melted ground ice in permafrost cores from the ESL. TDS in Slidre Fiord water and a late-lying snowbank are given as reference.

Figure 5.2.2: Piper plot showing hydrochemical facies present in ground ice and environmental samples in the ESL. Concentrations of each ion were converted into milliequivalents (mEq) per litre in - order to produce this diagram. Carbonate species (HCO3 ) were estimated based on the charge balance - equation. The HCO3 concentration may be overestimated for the Black Top (BT) ground ice, due to TDS > 104. For a detailed explanation of Piper plots and charge balance equations, see Fetter (2001). Slidre Fiord water, Station Creek (SC) ground ice and Dump Slump (DS) ground ice are classed as ‘marine’ waters.

Figure 5.2.3: Stiff diagrams reveal three distinct hydrofacies present in ESL ground ice.

-50

-100

-150 DS-1 DS-2 SC-1 -200 BT-1

δD (‰) δD BT-2 G-1 G-3 -250 MF-4 precipitation Slidre Fiord -300

-350 -45 -40 -35 -30 -25 -20 -15 -10 δ18O (‰)

Figure 5.2.4: All stable water isotope data from ESL ground ice and environmental sources. All data is from samples collected in 2017 and analyzed at the University of Ottawa, except for precipitation data from IAEA/WMO (2018).

5.3) Ice wedge 14C-DOC data This section presents 14C-DOC values found in ice wedges across the ESL. DOC in ESL wedge ice ranges from 2.0-6.7 ppm. Ice wedge 14C-DOC ages range from 2624-28948 yr BP. Results are presented in the table below and discussed in Section 6.4.

Sample ID Location in ice DOC 14C yr BP +/- F14C +/- wedge (ppm) BT-IW1-centre centre 2.8 5251 30 0.5201 0.002 BT-IW1-edge edge 2.0 6480 38 0.4463 0.0021 BT-IW2-edge edge 2.8 22043 117 0.0643 0.0009 BT‐IW2‐Centre centre 2.5 28948 269 0.0272 0.0009 DS-IW2 / 2.0 14641 64 0.1616 0.0013 Gem-IW1-edge edge 6.7 2624 28 0.7213 0.0025 Gem-IW2-centre centre 4.5 2980 29 0.69 0.0025 Gem-IW2-edge edge 3.7 4619 29 0.5627 0.002 MF-IW1-centre centre n/a 7296 47 0.4032 0.0024 MF-IW2-centre centre n/a 3047 29 0.6843 0.0025 MF-IW2-edge edge n/a 5237 34 0.521 0.0022 DS-1 (95-134 cm) intrasedimental ice 2.0 12729 52 0.205 0.0013

Table 5.3: 14C-DOC results from ESL ice wedges. The ice wedges are associated with retrogressive thaw slump exposures discussed in section 5.1. The bottom row (in italics) is from massive tabular ice, extracted from the base of the DS-1 permafrost core (see 5.1.1).

6. Discussion

In this section, the geochemical findings will be discussed in the context of Pollard’s (2000b) landscape model explaining the formation of massive ground ice in the Eureka Sound Lowlands

(ESL). Besides the freezing history and origin (6.1), other topics of discussion include the movement of water prior to freezing (6.2), climatic implications (6.3) and ground ice age derived from radiocarbon (6.4).

6.1) Origin of massive ice: segregated or buried?

The primary aim of this research is to identify the origin of massive ice in the ESL. In the landscape model presented by Pollard (2000b), which forms the framework of this study, groundwater is driven by a hydraulic and thermal gradient from an aquifer underlying Holocene marine sediments to a freezing front induced by marine emergence.

Co-isotope plots (δD-δ18O) are often used to help determine the origin of ground ice, based on the theory that buried glacier ice will have a freezing slope close to the meteoric line (~ 8), while segregated ice (i.e. formed by slow equilibrium-freezing) will have a slope <6. The LMWL measured at Eureka Weather Station follows a slope of 7.4 (fig. 4.2.3, IAEA 2018). At every

ESL site except Black Top, the massive ice has a co-isotope slope ≥ 7.4 (figs. 5.1.1e, 5.1.1g,

5.1.2d, 5.1.4f, 5.1.4h). This would suggest that much of the massive ice in the ESL is buried glacial or snowbank ice, contrary to the hypothesis of Pollard (2000a&b) and Pollard et al.

(2015).

However, a buried ice origin is rejected for most ESL massive ice, with the possible exception of ice close to Holocene marine limit at the Gemini site (see 6.1.1. Gemini ice: a special case?). A segregated-intrusive origin of ESL massive ice is more likely for several reasons:

1) In almost all cases, ESL massive ice is found stratigraphically below marine sediments.

The upper contact between the ice and the marine sediments is conformable and

gradational, as would be expected in intrasedimental ice that formed after the marine

sediments were deposited.

2) Preservation of substantial volumes of glacial ice in a submarine setting is highly

unlikely, although theoretically possible. Recent studies by Overduin et al. (2012) and

Stendel et al. (2016) have demonstrated the widespread existence and persistence of sub-

sea permafrost, but large volumes of relict glacier ice have not been documented under

the sea. In most cases submarine permafrost is associated with the post glacial inundation

of coastal lowlands that already contained extensive permafrost (e.g. the Beaufort Shelf)

3) There is chemical continuity between the massive ice and ice lenses found in the

overlying marine sediments, as well sediment inclusions in the massive ice match the

texture of the overlying marine sediments.

4) Glacial till is conspicuously absent in massive ice exposures.

None of these points explain the “meteoric” freezing (co-isotope) slopes observed in ESL massive ice. However, in an open system where there the input of isotopically-depleted water is much greater than the freezing rate, “meteoric” freezing slopes can be produced in segregated ice, as would be the case if local glacial meltwater was the primary source for ground ice formation.

Souchez and Jouzel (1984) give the following equation for a freezing slope (s) in an open system with an initial reservoir (i) and input (A) at t=0:

훼[(훼 −1)(1+ 훿푖)−(퐴/푆)(훿퐴 − 훿푖)] 푠 = 훽[훽 − 1)(1 + 훥푖)−(퐴/푆)(훥퐴 − 훥푖)]

where (values relevant to this study in italics):

β = fractionation factor of 18O i-w = 1.003 (O’Neil 1968)

α = fractionation factor of D i-w = 1.0175 (O’Neil 1968)

δA = input D (‰)

δi = reservoir D (‰)

ΔA = input 18O (‰)

Δi = reservoir 18O (‰)

A = coefficient for input

S = coefficient for freezing

This equation shows that the freezing slope depends on (δA-δi) and (ΔA-Δi), i.e. the difference between the δ value of the input and that of the reservoir at t=0. Souchez and Jouzel (1984) write that “in a natural reservoir, generally there is no reason for a change of input during the formation of this reservoir and during it subsequent freezing”. They argue that in natural settings

δA=δi and ΔA=Δi, so the slope equation reduces to:

훼[(훼 − 1)(1 + 훿푖)] 푠 = 훽[훽 − 1)(1+ 훥푖)]

Thus, the freezing slope for an open system is essentially the same as it is for a closed system, as long as the input is not significantly different in its isotopic composition from that of the initial

92 reservoir. This model predicts that under equilibrium freezing (as would be expected for the ice segregation), the freezing slope should plot significantly below the meteoric slope. The 1984

Souchez and Jouzel paper is the theoretical foundation of most subsequent studies using co- isotope plots to determine ground ice origin.

Souchez and de Groote (1985) published evidence of a natural exception to the assumption of similar isotopic signatures in an open freezing system. At the Gruben-gletscher in Valais, western Switzerland, basal ice forms where subglacial meltwater enters a permafrost zone near the glacier margin and freezes at the glacier sole. There is also seepage of water from an ice- dammed lake into this zone near the glacier margin. They found the δD-δ18O slope of the basal ice to be ~8.3, much higher than predicted by Souchez and Jouzel’s (1984) model for open- system freezing. To explain this, Souchez and de Groote (1985) ran a computer simulation for the evolution of δD and δ18O in ice samples if the reservoir is allowed to mix over time with an isotopically-lighter input. They found that under the conditions observed at the base of Gruben- gletscher (where the input of isotopically-depleted water is much greater than the freezing rate), basal ice samples will plot along a “meteoric” slope – albeit with slightly lower d excess values.

This effectively negates the use of the slope of co-isotope plots to distinguish between glacial ice and ice formed by equilibrium freezing, under the scenario observed at Gruben-gletscher.

Another natural example of “meteoric” slopes produced by an open freezing system with an isotopically-light input is found at the Mammuthöhle cave in the Alps of central Austria. Kern et al. (2011) found a freezing slope of 8.13 in a 5 m core from perennial cave ice deposits. They realized that most of the ice accumulation occurred in the early spring when karst spring water

(which represents an average isotopic signature of yearly precipitation) receives an influx of snow meltwater (with a more depleted isotopic signal), as it slowly freezes on the cave floor.

The two conditions outlined by Souchez and de Groote (1985) required to produce “meteoric” slopes in ice formed by equilibrium freezing are: 1) an open system where a reservoir is well- mixed with an isotopically-lighter input and 2) a slow freezing rate. The landscape model for massive segregated ice formation in the ESL (Pollard et al. 2015) or, for that matter, any massive segregated ice associated with the deglaciation of an ice sheet and permafrost aggradation along its margins in the style of Rampton (1988), fit these criteria. In the case of the ESL:

1) The two isotopically-disparate water sources would be an initial reservoir of isotopically-

heavy remnant seawater within the raised marine sediments and isotopically-depleted

glacial meltwater being driven to the glacier terminus. Modern Slidre Fiord seawater is

measured at δD ≈ -98‰ and δ18O ≈ -13‰, compared to glacial meltwater which has an

isotopic signal closer to δD ≈ -250‰ and δ18O ≈ -30‰ (e.g. Lecavalier et al. 2017).

2) The theory behind the growth of massive segregated ice lenses, as explained in Section

1.3.2 (pp. 11-13), necessitates a slow freezing rate, i.e. a stable freezing front, to allow

unfrozen groundwater to migrate upwards to the freezing front over long periods of time

(102-103 yrs).

More evidence for a mixed water source is given by the hydrochemical facies observed in ESL massive ice. The facies of massive ice samples from Dump Slump and Station Creek (~40-50 m a.s.l., emergence ~5 kya) are Na-Cl dominant and almost identical to modern Slidre Fiord water

(fig. 5.2.3), but with TDS concentrations lower by more than one order of magnitude (fig. 5.2.1).

The highly-depleted stable water isotope values in the ice confirm that in spite of the clear marine influence, the primary water source for the ice has to be meltwater, not marine water.

The only massive ice that shows a "classic" freezing slope is at Black Top slump (~70 m a.s.l., emergence ~7,000 yrs BP), with slopes of 5.75 and 6.52 from cores BT-1 and BT-2, respectively. Black Top ice is also more sediment-rich (i.e. lowest volumetric and gravimetric ice content among ESL massive ice samples). According to the theory of ice segregation, the inclusion of more sediment into the Black Top ice must be an indicator of a faster freezing rate, as there is no discernible difference in sediment particle size at Black Top compared to Dump

Slump or Station Creek. A faster freezing rate would prevent the long-term migration of underlying groundwater – thus with the Pollard et al. (2015) model in mind, we would expect less of a ‘glacial’ signal and more of ‘marine’ signal. The reduced mixing also explains why the

δD-δ18O slope plots along a lower line – the site is more representative of a closed freezing system. The TDS concentrations measured in Black Top ice are in the range of modern Slidre

Fiord water (>10,000 mg/l) (fig. 5.2.1) and its Na+ concentrations exceed all other ESL massive ice by two orders of magnitude. The δ18O values of Black Top ice range from -27 to -20‰ (fig.

5.1.3c, 5.1.3d), significantly heavier than Dump Slump and Station Creek ice, but also lighter than modern fiord water (-13‰). However, it is expected that during the Holocene the isotopic signal of Slidre Fiord would become heavier over time. At the time of when marine sedimentation was occurring at Black Top, Slidre Fiord would have been receiving substantial amounts of isotopically-depleted meltwater from the deteriorating Innuitian Ice Sheet that fringed it. There is a high probability that the δ18O value of ~ -22‰ VSMOW found in Black

Top ice is closer to early Holocene Slidre Fiord δ18O than our modern measurement.

+ + 2+ - - As further evidence of a marine influence, the calculated (Na +K )/Ca and Cl /HCO3 ratios for

ESL massive ice and other sources are presented in Table 6.1. In their study into the origin of

massive ground ice on Herschel Island, N.W.T., Fritz et al. (2011) write that “(Na++K+)/Ca2+ and

- - Cl /HCO3 ratios greater than unity [>1] indicate an enrichment of ions with a likely marine

+ + 2+ - - Site (Na +K )/Ca ratio Cl /HCO3 ratio N Mean Median Min. Max. N Mean Median Min. Max. Dump Slump massive ice 14 7.17 2.34 0.60 66.23 14 4.40 3.38 1.94 10.96 Black Top massive ice 32 343.33 350.20 166.18 507.75 32 51.18 54.22 24.72 71.26 Station Creek massive ice 28 4.23 2.41 1.07 22.12 28 3.94 2.36 1.11 18.25 Gemini massive ice 23 0.66 0.58 0.29 1.31 24 0.44 0.36 0.13 1.15 Pore ice in sediment above Gemini ice 1 68.28 / / / 1 2.73 / / / Slidre Fiord water 1 12.43 / / / 1 21.32 / / / Black Top Creek water 1 0.33 / / / 1 0.57 / / / Glacial ice (west Greenland)* 3 0.72 / 0.64 0.78 2 0.36 / 0.29 0.43 *from Yde et al. (2014)

Table 6.1: Summarized values of two major ion ratios for ESL ground ice and other environmental sources. Samples with ratios < 1 are marked in bold.

origin derived from dissolution from sediments with a marine influence.” Dump Slump and

Station Creek massive ice samples have ratios > 1, while Black Top ice is >> 1. The only

massive ice that deviates from this trend is at the Gemini site.

6.1.1) Gemini ice: a special case?

Massive ice at Gemini (~120-130 m a.s.l., emergence ~8.5 kya) is very pure (few to no sediment

inclusions) and has a thin sediment overburden (<2 m). Stable water isotope profiles in cores G-1

and G-3 reveal depleted δD and δ18O values that remain constant with depth (fig. 5.1.4d &

5.1.4e). δ18O values are on average ~ 3‰ more depleted in G-1 core (lower) compared to G-3

core (higher). The stratigraphic vertical difference between G-1 and G-3 is about 10-15 m.

The TDS in Gemini ice is low (~10 mg/l). Unlike other ESL ground ice, Gemini ice is composed

+ + 2+ of Ca-HCO3 waters (fig. 5.2.3), which represent a strong terrestrial signal. The (Na +K )/Ca

- - and Cl /HCO3 ratios of G-3 ice samples are less than unity (Table 6.1). In fact, Gemini ice has virtually the same major ion ratios as glacial ice from the Greenland Ice Sheet (Yde et al. 2014).

Interestingly, only one sub-sample from G-3 core showed an Na-Cl marine signal – ice extracted from sediment directly overlying the massive ice. This is consistent with surficial mapping which identifies sediments in the Gemini area as having a marine origin (Aitken and Bell 1998; Bell and Hodgson 2000).

There is little evidence that glacial meltwater mixed with marine waters or sediments at Gemini.

The geochemical signal in the massive ice is purely glacial. Furthermore, the freezing slopes produced by co-isotope plots for G-1 and G-3 ice equal or exceed the LMWL (fig .5.1.4f &

5.1.4g). If the Gemini ice has a segregated origin, freezing would have had to proceed in one of two scenarios: i) very slowly, to allow for constant migration of unfrozen water to the freezing front in order to produce such massive, pure lenses (but after such slow freezing, the effects of equilibrium isotopic fractionation should be apparent in the isotopic signature of the ice, instead of “meteoric” slopes), or ii) it could have frozen quickly if there was water being supplied as quickly as it could freeze, likely under a strong pressure gradient.

Recalling the four major stratigraphic/contextual arguments for a segregated origin of lower elevation massive ice (pp. 91):

1) A conformable upper contact between massive ice and marine sediments. At Gemini,

massive ice is overlain by a thin cover of marine sediments and for the most part the

contact is conformable. There are a few exceptions. For example, fig. 6.1.1 shows a clear

unconformity and some recumbent folds in the ice near the G-3 sampling site. This type

of unconformity, however, can easily be explained by local unloading process related to

sea level change, isostatic rebound, deflation and surface thermokarst. Folding is also

observed in other sites across the ESL.

Figure 6.1.1: Exposed massive ice in a retrogressive thaw slump. Close to G-3 core sampling site.

2) Unlikely preservation of large volumes of relict glacier ice under the sea. Based on its

position close to Holocene marine limit, the Gemini site would have been submerged for

the least amount of time. The preservation of glacial ice is therefore more plausible here

than at the lower elevation sites.

3) Similarity of sediment characteristics above and below ice. Due to the lack of sediment

contained within Gemini ice, it is difficult to compare. However, there are appears to be

no difference in δ13C signature or bulk C % in sediment above and within G-3 massive

ice (Table 5.1.4a).

4) Absence of glacial till. There is no glacial till or debris visible at the Gemini site.

So is Gemini massive ice buried Innuitian ice? Is there even a process by which this is possible?

While the burial of relict glacial ice is well recorded and discussed in Section 1.3.1, the long- term preservation of glacial ice in subaqueous sediments has never been directly observed.

Syvitski et al. (1996) and Dowdeswell et al. (2000) present examples of how ice can be buried in glacimarine sediments during the breakup or re-advance of major ice shelves. Often, the buried ice is from icebergs that run aground or from sikussaks (multiyear landfast sea ice). In these cases, however, the ice is fragmented, aligned haphazardly and eventually thaws, leaving melt- out features in the sediments. This is not the type of ice that is found at Gemini, stratigraphically or geochemically.

In his study of alpine lake sediments in British Columbia, Shaw (1977) postulated that i) the final attitudes of the sediment bodies, ii) the deep basins occupied by modern lakes and iii) terraces underlain by lacustrine deposits are all “the result of melt out of large masses of glacier ice buried beneath the lacustrine sediments”. He proposed a sedimentation model during deglaciation, illustrated in fig. 6.1.2. In the retreating ice margin phase, stages a and b in fig. x reflect a high rate of sedimentation and low rate of glacier retreat. The sedimentary succession is characterized by deltaic sequences with some turbidity current deposits.

If we apply Shaw’s (1977) high-alpine glaciolacustrine model to the High Arctic glacimarine setting of the Gemini site, it actually fits quite well. The Gemini site, at 130 m a.s.l., is very close to Holocene marine limit, where we can envisage an ice margin similar to the one illustrated in fig. 6.1.2. Aitken and Bell (1998) describe the area as a marine limit delta with extensive turbidity flow deposits. They describe an early Holocene environment where deglacial sedimentation rates exceeded 10 m kyr-1. Glacial ice could have easily been buried in the manner illustrated by Shaw (1977).

Several researchers (Hodgson 1985; Aitken and Bell 1998; England et al. 2006; Simon et al.

2015) have noted the extremely rapid rates of isostatic rebound near marine limit on the Fosheim

Figure 6.1.2: Shaw’s (1977) facies model for alpine glacial lakes.

100

Peninsula. This could have raised the buried ice and overlying sediments above sea level relatively quickly and exposed them to colder subaerial temperatures – allowing for the glacial ice to be preserved.

The massive ice exposed near Holocene marine limit at the Gemini site has similar geochemistry of what would be expected in glacial ice, either terrestrial, grounded or floating. For instance,

Gemini ice has the same δ18O profile and low salinity as the top 10-40 m of ice cores collected from the Ward Hunt Ice Shelf and the former Hobson’s Choice Ice Island (Jeffries 1992).

A segregated or segregated-intrusive origin cannot be ruled out, however, and further analysis should be undertaken on the archived samples including one entire core (G-2) and two half-cores

(G-1 & G-3). For instance, the ratios of occluded gases (discussed in Section 4.2.3) can be measured in Gemini ice to compare with atmospheric concentrations. A more detailed cryostratigraphic field analysis must be undertaken as well. The largely conformable upper contact and absence of glacial till in the exposure are still the most convincing arguments against the presence of buried ice.

101

6.2) Can ice sheet meltwater be transferred into a subglacial aquifer?

The model of landscape-scale massive ice formation by the process of ice segregation proposed by Pollard et al. (2015) requires the infiltration of large volumes of subglacial meltwater into a permeable, conductive geologic unit beneath an ice sheet. Is this even possible?

Ice sheets produce meltwater on their surface (supraglacial zone) and at their base (subglacial zone). The rate of supraglacial meltwater production is largely controlled by seasonal and local weather patterns, while rates of subglacial meltwater production are explained by friction due to ice flow and melt from water flow (Ravier and Buoncristiani 2018). Even cold-based, polar ice sheets have large areas of basal melting (Paterson 1994). According to Boulton et al. (2007), 35-

45% of meltwater derived from the surface of Breidamerkurjökull (which they describe as an analogue for sediment-based Pleistocene ice sheets of North America) penetrates to the glacier bed. The modern Greenland Ice Sheet (GIS) is a better analogue for the former Innuitian Ice

Sheet. It would be expected that the snow, firn and ice layers of a colder, polar ice sheet, such as the GIS, would be less permeable than the more temperate Bredamerkurjökull, making basal melting a larger source to subglacial drainage. However, recent work has emphasized the importance of moulins, large shaft-like openings, for transferring surface meltwater to the subglacial zone on the GIS (Gulley et al. 2009; Catania and Neumann 2010), so subglacial

Innuitian meltwater may have been composed of both basal and surface melt.

In any case, when ice sheets overlie permeable materials, such as the unconsolidated Tertiary sandstones and shales of the ESL, infiltration of subglacial meltwater into the bed and into groundwater circulation is expected. In fact, some argue that beneath large ice sheets most, if not all subglacial meltwater enters the bed and flows as groundwater in the direction of the glacier

102 terminus (e.g. Boulton and Dobbie 1993; Lemieux et al. 2008; Ravier and Buoncristiani 2018).

The flow pattern of groundwater beneath the ice sheet is characterized by a net downwards flow vector due to the ice overburden pressure values. Beyond the glacial margin, there is strong upward flow due to decreasing overburden pressure (Ravier and Buoncristiani 2018). Upwelling of groundwater flow at or near the glacial margin has been reported in the field at

Bredamerkurjökull by Boulton et al. (2007). In the Canadian High Arctic, upwelling of large volumes of meltwater through subglacial sediments at the glacier terminus has been observed at

John Evans Glacier on eastern Ellesmere Island (Skidmore and Sharp 1999). Groundwater flow paths are affected by the hydraulic conductivity of overlying sediments and the existence of proglacial permafrost. Where proglacial permafrost exists, groundwater discharge will only occur away from the glacial margin at taliks (Ravier and Buoncristiani 2018). The fine-grained silts and clays of the ESL, a consequence of marine transgression, acted as a hydraulic aquitard as the ice margin retreated, limiting the loss of meltwater by upwelling. Rapid permafrost aggradation effectively turned the overlying sediment into an aquiclude, capping the ever- increasing ice lens growing at the aquifer-aquiclude interface.

103

6.3) Permafrost evolution

A transect of ground ice profiles across an elevation gradient allows us to track the evolution of

Holocene permafrost aggradation in ESL sediments. Pollard and Bell’s (1998) landscape model predicts that downwards freezing would have started as land first emerged at marine limit (~143 m a.s.l.). The sample sites in this study listed in chronological order of permafrost aggradation would be thus: Gemini, Black Top, Station Creek, Dump Slump (see fig. 4.3.1 for estimated dates of emergence). The nature of the ground ice at each elevation gives clues into the rate of permafrost aggradation at the time of emergence.

Climate reconstructions have revealed that starting at about 10,000 yrs BP, during the initial break-up of the Innuitian Ice Sheet, the High Arctic experienced rapid warming. Air temperatures exceeded modern values by up to 5-8°C (Lecavalier et al. 2017). It must be noted that even with an increase of 5-8°C, the mean annual air temperature at Eureka would still be in the range of -14 to -11°C, much colder than current annual air temperatures in the Low Arctic.

Even in a “warmer” Arctic, subaerial freezing of recently emerged sediments would be expected to proceed quickly, as there would exist a large thermal disequilibrium between the air and ground. Early Holocene Arctic temperature reconstructions are based on proxies such as Agassiz

Ice Cap δ18O and pollen records, bowhead whale bones, lake sediments and peat sequences

(Briner et al. 2016; Lecavalier et al. 2017). Besides a few sudden cooling episodes (e.g. the 8.2 kya event), the High Arctic remained warmer than present through the early Holocene (although still cold enough for permafrost to aggrade into recently emerged sediments). Following a period of exceptionally warm temperatures around 6-5,000 yrs BP (coinciding with the Holocene

Thermal Maximum), a general cooling trend began that culminated in the Little Ice Age in 1700

104

AD. The current warming trend is unprecedented in at least the last 5,000 yrs (Briner et al. 2016;

Lecavalier et al. 2017).

Permafrost began aggrading at the Gemini site between 8,500 and 8,000 yrs BP. Overall, this was a period of rapid warming associated with the break-up of highland ice caps. Assuming the

Gemini massive ice is segregated, this would coincide with a remarkably stable freezing front to allow for the constant supply of upwelling meltwater to form pure ice at the base of the marine sediments. On the other hand, it is widely accepted that ~8,200 yrs BP there was an abrupt climate fluctuation where temperatures in central Greenland dropped by 4-8°C over ~150 years

(Barber et al. 1999; Briner et al. 2016). The cause of the sudden cooling is believed to be associated with a massive outflow of fresh water from the Hudson Strait following the collapse of the northern ice dam of glacial lakes Agassiz and Ojibway (Barber et al. 1999). In many

Arctic locales, this climatic event is linked with stable or re-advancing glacial margins, known as the Cockburn substage (Young et al. 2012; Briner et al. 2016). Interestingly, glacial re-advances are linked with burial of glacial ice. In any case, the rate of freezing is hard to infer based on the

Gemini exposure, as it is unclear if the massive, pure ice is a reflection of a slow freezing rate or pressurized groundwater.

The Black Top slump is situated between 60-70 m a.s.l., corresponding with a period of emergence ~7-7,500 yrs BP. This corresponds with a moderate warm period – a minor lull in the warming trend following the onset of deglaciation. The BT cores contain dirty massive ice with volumetric ice contents from 50-80%. Hodgson and Nixon (1998) found similar ice content in cores taken from the same elevation range (excluding those taken from ice wedges). In general,

Hodgson and Nixon (1998) recorded pure (>95% fVi) massive ice at lower (35-55 m a.s.l.) or higher (90-130 m a.s.l.) elevations across the ESL. It is possible that the dirtier ice found at this

105 elevation across the ESL is due to a more rapidly advancing freezing front during the time of emergence. The migration of isotopically-depleted groundwater was limited – thus most of the water supplying the massive ice was brackish water within the marine sediments. There is no discernible difference in sediment texture between the Black Top site and the Station Creek or

Dump Slump sites (silty clays coarsening with depth to silty loams).

Relative sea level curves (e.g. Simon et al. 2015) place Station Creek and Dump Slump exposures (40-50 m a.s.l.) as emerging at approximately 5,000 yrs BP. Both are examples of pure

(>95% fVi) massive ice. Their approximate time of formation coincides with a relatively warm period in the Canadian High Arctic. Air temperatures are estimated to be 3-4°C warmer than pre- industrial modern values. Downwards freezing of the ground would proceed at a slower rate under warmer air temperatures. This created a stable freezing front, allowing unfrozen groundwater (glacial melt) to migrate through the aquifer (fractured, unconsolidated Tertiary bedrock) and form massive ice at the aquifer-marine sediment interface over the span of 102-103 yrs.

106

6.4) Dissolved organic carbon in ground ice: 14C dating

Radiocarbon dating using DOC in ESL ice wedges gives mixed results (Table 5.3). In most cases where ‘edge’ and ‘centre’ samples were analyzed, ice sampled from the edge of the wedges gives older 14C ages than ice from the centre of the wedge. This is consistent with the theory of ice wedge growth where thermal contraction cracks form in the middle of the wedge and local melt water infiltrates a thermal contraction crack each spring. The use of DOC to date ice wedges relies on the assumption that the ice wedge grows by snow meltwater. We assume that in the spring, meltwater flows over and leaches through a shallow active layer where modern plants are growing, thereby incorporating DOC that corresponds with the year of infiltration.

DOC concentrations in ESL ice wedges ranged from 2.0-6.7 mg/l. This is lower than the mean

DOC concentration (9.6 mg/l) of ice wedge samples collected across Siberia, Alaska and the

Yukon by Fritz et al. (2015), although within their reported range (1.6-28.6 mg/l). Our 14C-DOC results suggest that ice wedges began to form in the ESL during the early-to-mid-Holocene.

Specifically, ice wedge formation began no later than 4,500 yrs BP at the Gemini site, 6,500 yrs

BP at the Black Top site and 5,500 yrs BP at the Mokka Fiord site. The initiation of ice wedge formation is expected to have been synchronous with permafrost aggradation into raised fine- grained sediments. The 14C-DOC ages associated with the wedges post-date the estimated time of emergence for their respective sampling sites (fig. 4.3.1) and are expected to coincide with massive ground ice formation.

The only ice wedge where this pattern is not observed is BT-IW2, which was sampled at the

Black Top slump. Ice from the centre of the wedge gives a 14C age older than the edge

(28,948±269 vs. 22,043±117 14C yr BP). These ages are anomalously old and pre-date marine

107 emergence by >15,000 yrs, when the region was glaciated. The other ice wedge sampled at the

Black Top slump (BT-IW1) dates at 6,480 14C yr BP (edge) and 5,251 14C yr BP (centre), consistent with Holocene permafrost aggradation. An unspecified (centre or edge) ice sample from a wedge at the Dump Slump site also has an old 14C age (14,641 yr BP). We believe these abnormally old values are a result of “old” organics being incorporated into the meltwater instead of contemporary plant-derived DOC. This could occur if the meltwater flowed through a fine-grained medium ‘contaminated’ with old marine carbon. Before the 14C-DOC method can be used confidently in paleo-reconstruction studies, more research needs to be done on the factors that effect the transfer of snow to ice wedge.

The use of 14C-DOC as a dating tool in geocryology is designed for ice wedges, as opposed to tabular ice bodies. Nevertheless, we analyzed a sample of non-wedge massive ice from Dump

Slump. It had a 14C age of 12,729 yrs BP (DOC concentration = 2.0 mg/l). Based on our interpretation of Dump Slump ice as having a segregated origin and a strong glacial meltwater source, a late-Pleistocene age for DOC in the ice makes sense.

In their survey of DOC in Arctic ground ice, Fritz et al. (2015) write that “an open question remains as to how much DOC can be found in intrasedimental ice”. DOC concentrations in ESL intrasedimental ice range from 1.01-31.61 mg/l, with a mean of 6.67 mg/l. This is the first time

DOC concentrations have been reported in High Arctic intrasedimental ice.

108

7. Conclusions

The aim of this thesis was to learn about the origin of massive ground ice in the Eureka Sound

Lowlands through field observations, permafrost core analysis (especially, stable water isotopes and major ions) and radiocarbon dating of ice wedges. The results lead to the following conclusions:

I. Massive ground ice in the Eureka Sound Lowlands formed synchronously with marine

regression, as water migrated upwards to a stable freezing front near the interface of fine-

grained marine sediments and an underlying, permeable geologic unit.

II. The massive ground ice formed in an open freezing system with an initial reservoir of

remnant marine water within the fine-grained sediments, mixed with an input of

isotopically-depleted glacial meltwater supplied by the permeable layer below.

III. Massive ground ice exposures with the highest volumetric ice contents are associated

with the most isotopically-depleted ice. In these cases, the input of glacial meltwater is

expected to have been much greater than the downwards freezing rate.

IV. “False” meteoric co-isotope slopes can be produced in ground ice that formed in an open

system with an input that has a significantly different isotopic signature than an initial

reservoir.

V. There is a possibility that massive ground ice close to Holocene marine limit (specifically

at the Gemini site) could be buried glacial ice, due to its unique hydrochemistry.

VI. The rate of permafrost aggradation into newly-emerged marine sediments was faster in

the early-Holocene (~7,500 yrs BP), as indicated by lower volumetric ice contents,

heavier isotopic signatures and stronger marine hydrochemical signals in massive ice at

corresponding elevations. In contrast, permafrost aggradation rates are inferred to be

109

slower during the mid-Holocene (~5,000 yrs BP) due to the presence of thick bands of

“pure”, isotopically-depleted massive ice at these elevations, a reflection of a slow

freezing front.

• Vol. ice contents > 95% 18 • Depleted isotopes (δ O <30‰) • Vol. ice contents > 95% • Co-isotope slope ≈ 8.3 18 • Depleted isotopes (δ O <30‰) • Co-isotope slope = LMWL = 7.4 • Ca-CO 3 • Low TDS (<10 mg/l) • Na-Cl • Moderately high TDS (100-1000 mg/l) • Thick, “pure” segregated massive ice • Thick, “pure” segregated massive ice • Large input of glacial meltwater • Large input of glacial meltwater Or… • Buried glacial ice?

• Vol. ice contents ≈ 50-70% 18 • Heavier isotopes (δ O: -25 to -20‰) • Co-isotope slope ≈ 5

• Na-CO3 • High TDS (>10,000 mg/l)

• “Dirty” massive ice • Less input of glacial meltwater, more remnant seawater

Fig. 7.1: Summary of key geochemical findings along elevation transect (refer to Fig. 4.3.1)

110

8. References Aitken AE, Bell T (1998) Holocene glacimarine sedimentation and macrofossil paleoecology in the Canadian High Arctic: environmental controls. Marine Geology 145: 151-171 Angelopoulos MC, Pollard WH, Couture NJ (2013) The application of CCR and GPR to characterize ground ice conditions at Parsons Lake, Northwest Territories. Cold Regions Science and Technology 85: 22-33 Astakhov VI, Isayeva LL (1988) The ‘Ice Hill’: an example of ‘retarded deglaciation’ in Siberia. Quaternary Science Reviews 7: 29-40 Baertschi P (1976) Absolute 18O content of Standard Mean Ocean Water. Earth Planet. Sci. Letters 31: 341-344 Barber DC, Dyke A, Hillaire-Marcel C, Jennings AE, Andrews JT, Kerwin MW, Bilodeau G, McNeely R, Southon J, Morehead MD, Gagnon JM (1999) Forcing of the cold event of 8,200 years ago by catastrophic drainage of Laurentide lakes. Nature 400: 344-348 Bednarski J (1998) Quaternary history of Axel Heiberg Island bordering Nansen Sound, Northwest Territories, emphasising the Last Glacial Maximum. Canadian Journal of Earth Sciences 35: 520-533 Becker MS, Davies TJ, Pollard WH (2016) Ground ice melt in the high Arctic leads to greater ecological heterogeneity. Journal of Ecology 104: 114-124 Bell T (1996) Late Quaternary glacial and sea level history of Fosheim Peninsula, Ellesmere Island, Canadian High Arctic. Canadian Journal of Earth Sciences 33: 1075-1086 Bell T, Hodgson DA (2000) Quaternary geology and glacial history of Fosheim Peninsula, Ellesmere Island, Nunavut. In Garneau M, Alt B (eds.) Environmental Response to Climate Change in the Canadian High Arctic. Geological Survey of Canada Bulletin 529: 175-187 Blake W (1970) Studies of glacial history in arctic Canada. I. Pumice, radiocarbon dates, and differential postglacial uplift in the eastern Queen Elizabeth Islands. Canadian Journal of Earth Sciences 7: 634-664 Bonnaventure P, Lewkowicz A, Smith S, Ednie M (2016) Influence of coastal proximity on ground thermal regime in a High Arctic environment, Eureka, Nunavut, Canada. Proceedings of the XI International Conference on Permafrost, Potsdam, Germany: 401-402 Boulton GS, Dobbie KE (1993) Consolidation of sediments by glaciers: relations between sediment geotechnics, soft-bed glacier dynamics and subglacial ground-water flow. Journal of Glaciology 39: 26-44 Boulton GS, Lunn R, Vidstrand P, Zatsepin S (2007) Subglacial drainage by groundwater- channel coupling, and the origin of esker systems. Part I – glaciological observations. Quaternary Science Reviews 26: 1067-1090

111

Boynton WV et al. (2002) Distribution of hydrogen in the near surface of Mars: evidence for subsurface ice deposits. Science 297: 81-85 Brandt O, Langley K, Kohler J, Hamran SE (2007) Detection of buried ice and sediment layers in permafrost using multi-frequency ground-penetrating radar: a case examination on Svalbard. Remote Sensing of the Environment 111: 212-227 Briner JP, McKay NP, Axford Y, Bennike O, Bradley R, de Vernal A, Fisher D, Francus P, Frechette B, Gajewski K, Jennings A, Kaufman DS, Miller G, Rouston C, Wagner B (2016) Holocene climate change in Arctic Canada and Greenland. Quaternary Science Reviews 147: 340-364 Cardyn R, Clark ID, Lacelle D, Lauriol B, Zdanowicz C, Calmels F (2007) Molar gas ratios of air entrapped in ice: a new tool to determine the origin of relict massive ground ice bodies in permafrost. Quaternary Research 68: 239-248 Catania GA, Neumann TA (2010) Persistent englacial drainage features in the Greenland Ice Sheet. Geophysical Research Letters 37: L02501 Couture N, Pollard WH (1998) An assessment of ground ice volume near Eureka, NWT. Proceedings of the 7th International Conference on Permafrost 55: 195-200 Dansgaard W (1964) Stable isotopes in precipitation. Tellus 16(4): 436-468 Dowdeswell JA, Whittington RJ, Jennings AE, Andrews JT, Mackensen A, Marienfeld P (2000) An origin for laminated glacimarine sediments through sea-ice build-up and suppressed iceberg rafting. Sedimentology 47: 557-576 Dyke AS (1999) The Last Glacial Maximum and deglaciation of Devon Island, Arctic Canada: support for Innuitian Ice Sheet. Quaternary Science Reviews 18: 393-420 Edlund SA (1989) Vegetation indicates potentially unstable Arctic terrain. GEOS 18(3): 9-13 Edlund SA, Alt BT, Garneau M (2000) Vegetation patterns on Fosheim Peninsula, Ellesmere Island, Nunavut. In Garneau M, Alt B (eds.) Environmental Response to Climate Change in the Canadian High Arctic. Geological Survey of Canada Bulletin 529:129-143 England J (1983) Isostatic adjustments in a full glacial sea. Canadian Journal of Earth Sciences 20: 265-270 England J (1987) Glaciation and the evolution of the Canadian high arctic landscape. Geology 15: 419-424 England J, Atkinson N, Bednarski J, Dyke AS, Hodgson DA, O Cofaigh C (2006) The Innuitian Ice Sheet: configuration, dynamics and chronology. Quaternary Science Reviews 25: 689-703 Everest J, Bradwell T (2003) Buried glacier ice in southern Iceland and its wider significance. Geomorphology 52: 347-358 Eyles N (1977) Late Wisconsinan glacitectonic structures and evidence of postglacial permafrost in north-central Newfoundland. Canadian Journal of Earth Sciences 14: 2797-2806

112

Fetter CW (2001) Applied Hydrogeology, 4th edition. Prentice Hall: Upper Saddle River, NJ French HM (2007) The Periglacial Environment. Third Edition. John Wiley and Sons Ltd.: Chichester, UK French HM (2011) Periglacial environments, in Gregory KJ, Goudie AS (eds.) The SAGE Handbook of Geomorphology. Sage Publications: London, pp. 393-411 French HM, Harry DG (1990) Observations on buried glacier ice and massive segregated ice, western Arctic coast, Canada. Permafrost and Periglacial Processes 1: 31-43 French HM, Schur Y (2010) The principles of cryostratigraphy. Earth-Science Reviews 101: 190-206 Fritz M, Wetterich S, Meyer H, Schirrmeister L, Lantuit H, Pollard WH (2011) Origin and characteristics of massive ground ice on Herschel Island (Western Canadian Arctic) as revealed by stable water isotope and hydrochemical signatures. Permafrost and Periglacial Processes 22: 26-38 Fritz M, Opel T, Tanski G, Herzschuh U, Meyer H, Eulenberg A, Lantuit H (2015) Dissolved organic carbon (DOC) in Arctic ground ice. The Cryosphere 9: 737-752 Garneau M (2000) Introduction, in Environmental Change in Response to Climate Change in the Canadian High Arctic. Geological Survey of Canada – Bulletin 529: pp. 1-16 Gat JR (1996) Oxygen and hydrogen isotopes in the hydrologic cycle. Annu. Rev. Earth Planet. Sci. 24: 225-262 Gilbert GL, Kanevskiy M, Murton JB (2016) Recent advances (2008-2015) in the study of ground ice and cryostratigraphy. Permafrost and Periglacial Processes Gulley JD, Benn DI, Mueller D, Luckman A (2009) A cut-and-closure origin for englacial conduits in uncrevassed regions of polythermal glaciers. Journal of Glaciology 55: 66-80 Hendriks MR (2010) Introduction to Physical Hydrology. Oxford University Press: Oxford. Hodgson DA (1985) The last glaciation of west-central Ellesmere Island, Arctic Archipelago. Canadian Journal of Earth Sciences 22: 347-368 Hodgson DA (1989) Quaternary Geology of the Queen Elizabeth Islands, in Fulton RJ (ed.) Quaternary Geology of Canada and Greenland. Geological Survey of Canada: Ottawa, Canada, pp. 441-478 Hodgson DA, Nixon FM (1998) Ground ice volumes determined from shallow cores from western Fosheim Peninsula, Ellesmere Island, Northwest Territories. Geological Survey of Canada - Bulletin 507. Natural Resources Canada: Ottawa. IAEA/WMO (2018) Global Network of Isotopes in Precipitation. The GNIP Database. Accessible at: http://www.iaea.org/water

113

Jeffries MO (1992) Arctic ice shelves and ice islands: origin, growth and disintegration, physical characteristics, structural-stratigraphic variability, and dynamics. Review of Geophysics 30: 245-267 Kaplyanskaya F, Tarnogradskiy V (1986) Remnants of the Pleistocene ice sheets in the permafrost zone as an object for paleoglaciological research. Polar Geography and Geology 10: 65-72 Kern Z, Fórizs I, Pavuza R, Molnár M, Nagy B (2011) Isotope hydrological studies of the perennial ice deposits of Saarhalle, Mammuthöhle, Dachstein Mts, Austria. The Cryosphere 5: 291-298 Lacelle D (2011) On the 18O, δD and D-excess relations in meteoric precipitation and during equilibrium freezing: theoretical approach and field examples. Permafrost and Periglacial Processes 22: 13-25 Lacelle D, Bjornson J, Lauriol B, Clark ID, Troutet Y (2004) Segregated-intrusive ice of subglacial meltwater origin in retrogressive thaw flow headwalls, Richardson Mountains, N.W.T., Canada. Quaternary Science Reviews 23: 681-696 Lacelle D, Lauriol B, Clark ID, Cardyn R, Zdanowicz CH (2007) Nature and origin of a Pleistocene-age massive ground-ice body exposed in the Chapman Lake moraine complex, central Yukon Territory. Quaternary Research 68: 249-260 Lacelle D, Fisher D, Clark ID, Berinstain A (2008) Distinguishing between vapor- and liquid- formed ground ice in the northern martian regolith and potential for biosignatures preserved in ice bodies. Icarus 197: 458-469 Lacelle D, Davila AF, Pollard WH, Andersen D, Heldmann J, Marinova M, McKay C (2011) Stability of massive ground ice bodies in University Valley, McMurdo Dry Valleys of Antarctica: Using stable O-H isotope as tracers of sublimation in hyper-arid regions. Earth and Planetary Science Letters 301: 403-411 Lacelle D, Vasil’chuk YK (2013) Recent progress (2007-2012) in permafrost isotope geochemistry. Permafrost and Periglacial Processes 24: 138-145 Lacelle D, Fontaine M, Forest AP, Kokelj S (2014) High-resolution stable water isotopes as tracers of thaw unconformities in permafrost: a case study from western Arctic Canada. Chemical Geology 368: 85-96 Lachiniet MS, Lawson DE, Sloat AR (2012) Revised 14C dating of ice wedge growth in interior Alaska (USA) to MIS 2 reveals cold paleoclimate and carbon recycling in ancient permafrost terrain. Quaternary Research 78: 217-225 Lauriol B, Lacelle D, St-Jean M, Clark ID, Zazula GD (2010) Late Quaternary paleoenvironments and growth of intrusive ice in eastern Beringia (Eagle River valley, northern Yukon, Canada). Canadian Journal of Earth Sciences 47: 941-955

114

Lapalme C, Lacelle D, Pollard WH, Fisher D, Davila A, McKay CP (2017) Distribution and origin of ground ice in University Valley, McMurdo Dry Valleys, Antarctica. Antarctic Science 29: 183-198 Lecavalier BS, Fisher DA, Milne GA, Vinther BM, Tarasov L, Huybrechts P, Lacelle D, Main B, Zheng J, Bourgeois J, Dyke AS (2017) High Arctic Holocene temperature record from the Agassiz ice cap and Greenland ice sheet evolution. Proceedings of the National Academy of Sciences 114: 5952-5957 Lemieux JM, Sudicky EA, Peltier WR, Tarasov L (2008) Dynamics of groundwater recharge and seepage over the Canadian landscape during the Wisconsinan glaciation. Journal of Geophysical Research 113: F01011 Lewkowicz A (2007) Dynamics of active-layer detachment failures, Fosheim Peninsula, Ellesmere Island, Nunavut, Canada. Permafrost and Periglacial Processes 18: 89-103 MacKay JR (1971) The origin of massive icy beds in permafrost, western arctic coast, Canada. Canadian Journal of Earth Sciences 8: 397-422 MacKay JR (1972) The world of underground ice. Annals, Association of American Geographers 62: 1-22 MacKay JR (1983) Downward water movement into frozen ground, western Arctic coast. Canadian Journal of Earth Sciences 20: 120-134 MacKay JR (1989) Massive ice: some field criteria for the identification of ice types. Geological Survey of Canada, Current Research Part B (Paper 83-1B): 67-74 Meyer H, Schirrmeister L, Andreev A, Wagner D, Hubberten HW, Yoshikawa K, Bobrov A, Wetterich S, Opel T, Kandiano E, Brown J (2010) Lateglacial and Holocene isotopic and environmental history of northern coastal Alaska – Results from a buried ice-wedge system at Barrow. Quaternary Science Reviews 29: 3720-3735 Michel FA (2011) Isotope characterisation of ground ice in Northern Canada. Permafrost and Periglacial Processes 22: 3-12 Millard A (2014) Conventions for reporting radiocarbon determinations. Radiocarbon 56(2): 555-559 Moorman BJ, Michel FA (2000) The burial of ice in the proglacial environment on Bylot Island, Arctic Canada. Permafrost and Periglacial Processes 11: 161- 175 Murton JB, Whiteman CA, Waller RI, Pollard WH, Clark ID, Dallimore SD (2005) Basal ice facies and supraglacial melt-out till of the Laurentide Ice Sheet, Tuktoyaktuk Coastlands, western Arctic Canada. Quaternary Science Reviews 24: 681-708 Ó Cofaigh C, England J, Zreda M (2000) Late Wisconsinan glaciations of southern Eureka Sound: evidence for extensive Innuitian ice in the Canadian High Arctic during the Last Glacial maximum. Quaternary Science Reviews 19: 1319-1341

115

Opel T, Laepple T, Meyer H, Yu Dereviagin A, Wetterich S (2017) Northeast Siberian ice wedges confirm Arctic winter warming over the past two millennia. The Holocene 27: 1789- 1796

Overduin PP, Westermann S, Yoshikawa K, Haberlau T, Romanovsky V, Wetterich S (2012) Geoelectric observations of the degradation of nearshore submarine permafrost at Barrow (Alaskan Beaufort Sea). Journal of Geophysical Research: Earth Surface 117: F02004

Paterson WSB (1994). The Physics of Glaciers, 3rd Edition. Elsevier Science Ltd.

Pollard WH, French HM (1980) A first approximation of the volume of ground ice, Richards Island, Pleistocene Mackenzie delta, Northwest Territories, Canada. Canadian Geotechnical Journal 17: 509-516

Pollard WH (1990) The nature and origin of ground ice in the Herschel Island area, Yukon Territory. In Proceedings of the Fifth Canadian Permafrost Conference, Collection Nordicanada 54: 23-30

Pollard WH (1991) Observations on massive ice ground ice on Fosheim Peninsula, Ellesmere Island, Northwest Territories, in Current Research, Part E, Geological Survey of Canada

Pollard WH (2000a) Distribution and characterization of ground ice on Fosheim Peninsula, Ellesmere Island, Nunavut. In Garneau M, Alt B (eds.) Environmental Response to Climate Change in the Canadian High Arctic. Geological Survey of Canada - Bulletin 529: 207-233

Pollard WH (2000b) Ground-ice aggradation on Fosheim Peninsula, Ellesmere Island, Nunavut. In Garneau M, Alt B (eds.) Environmental Response to Climate Change in the Canadian High Arctic. Geological Survey of Canada - Bulletin 529: 325-333

Pollard WH, Ward MK, Becker M (2015) The Eureka Sound lowlands: an ice-rich permafrost landscape in transition. Proceedings of GeoQuebec Conference, Quebec QC (September 20- 23, 2015)

Rampton VN (1988) Quaternary geology of the Tuktoyaktuk Coastlands, Northwest Territories. Geological Survey of Canada: Memoir 423

Rampton VN (1991) Short Communication: Observations on buried glacier ice and massive segregated ice, Western Arctic Coast, Canada. Permafrost and Periglacial Processes 2: 163- 165

Ravier E, Buoncristiani JF (2018) Glaciohydrogeology. In Menzies J, Van der Meer JJM (eds.) Past Glacial Environments, 2nd Edition. Elsevier Ltd., pp. 431-466

116

Scanlon KE, Head JW, Marchant DR (2015) Remnant buried ice in the equatorial regions of Mars: Morphological indicators associated with the Arsia Mons tropical mountain glacier deposits. Planetary and Space Science 111: 144-154

Schollick JMH, Style RW, Curran A, Wettlaufer JS, Dufresne ER, Warren PB, Velikov KP, Dullens RPA, Aarts D (2016) Segregated ice growth in a suspension of colloidal particles. Journal of Physical Chemistry B 120: 3941-3949

Schomacker A (2008) What controls dead-ice melting under different climate conditions? A discussion. Earth-Science Reviews 90:103-113

Schorghofer N, Aharonson O (2005) Stability and exchange of subsurface ice on Mars. Journal of Geophysical Research 110 (E05003), doi: 10.1029/2004JE002350

Shaw J (1977) Sedimentation in an alpine lake during deglaciation, Okanagan Valley, British Columbia, Canada. Geografiska Annaler 59: 221-240

Shumskii PA (1959) Principles of Geocryology, Part 1. General Geocryology, Chapter IX: Ground (subsurface) ice. Translated from Russian by C de Leuchtenberg. Academy of Sciences of the U.S.S.R, Institute of Permafrost Studies: Moscow

Simon K, Dyke A, James TS (2015) A new glacial isostatic adjustment model of the Innuitian Ice Sheet, Arctic Canada. Quaternary Science Reviews 119: 11-21

Skidmore ML, Sharp MJ (1999) Drainage system behaviour of a High-Arctic polythermal glacier. Annals of Glaciology 28: 209-215

Slagoda EA, Opokina OL, Kurchatova AN, Rogov VV (2012) Structure and composition of complex massive ice bodies in Late Pleistocene-Holocene sediments of the Marre-Sale Cape, West Yamal. In Proceedings of the Tenth International Conference, Melnikov VP, Drozdov DD, Romanovsky VE (eds.), Salekhard, Russia: 415-420

Smith MW (1985) Models of soil freezing, in Church M, Slaymaker O (eds.) Field and Theory: Lectures in Geocryology. UBC Press: Vancouver, pp. 98-119

Souchez RA, Jouzel J (1984) On the isotopic composition in δ18O of water and ice during freezing. Journal of Glaciology 30: 369-372

Souchez RA, JM de Groote (1985) δD-δ18O relationships in ice formed by subglacial freezing: paleoclimatic implications. Journal of Glaciology 31: 229-232

Sowers T, Brook E, Etheridge D, Blunier T, Fuchs A, Leuenberger M, Chappellaz J, Barnola JM, Wahlen M, Deck B, Weyhenmeyer C (1997) An interlaboratory comparison of techniques for extracting and analyzing trapped gases in ice cores. Journal of Geophysical Research 102: 26527-26538

117

St-Jean M, Lauriol B, Clark ID, Lacelle D, Zdanowicz C (2011) Investigation of ice wedge infilling processes using stable oxygen and hydrogen isotopes, crystallography and occluded gases (O2, N2, Ar). Permafrost and Periglacial Processes 22: 49-64

Stendel M, Rodehacke C, Romanovsky V, Nicolski D, Marchenko S, Christensen JH, Mottram R, Oleson M (2016) How permanent is East Siberian Arctic Shelf permafrost? Proceedings of the XI International Conference on Permafrost, Gunther F, Morgenstern A (eds.), Potsdam, Germany: 814

Steven B, Pollard WH, Greer CW, Whyte LG (2008) Microbial diversity and activity through a permafrost/ground ice core profile from the Canadian High Arctic. Environmental Microbiology 12: 3388-3403

Stuiver M, Yang IC, Denton GH (1976) Permafrost oxygen isotope ratios and chronology of three cores from Antarctica. Nature 261: 547-550

Stuiver M, Polach HA. Discussion: reporting of 14C data. Radiocarbon 19: 355-363

Sutherland PD (2000) Preliminary results of archaeological research on Fosheim Peninsula, Ellesmere Island and in adjacent areas of western Ellesmere and eastern Axel Heiberg islands, Nunavut, in Garneau M, Alt BT (eds.) Environmental Response to Climate Change in the Canadian High Arctic. Geological Survey of Canada – Bulletin 529: pp. 319-324

Sverdrup O (1904) New Land, Vol. II. London: Longmans, Green and Co. (translated from Norwegian by EH Hearn)

Syvitski JPM, Andrews JT, Dowdeswell JA (1996) Sediment deposition in an iceberg-dominated glacimarine environment, East Greenland: basin fill implications. Global and Planetary Change 12: 251-270

Utting N, Lauriol B, Lacelle D, Clark I (2016) Using noble gas ratios to determine the origin of ground ice. Quaternary Research 85: 177-184

Vaikmae R, Michel FA, Solomatin VI (1993) Morphology, stratigraphy and oxygen isotope composition of fossil glacier ice at Ledyanaya Gora, Northwest Siberia, Russia. Boreas 22: 205-213

Van Everdingen RO (1998) Multi-language glossary of permafrost and related ground-ice terms. International Permafrost Association, Terminology Working Group

Vasil’chuk YK (2011) Homogeneous and heterogeneous massive ground ice in permafrost. The Earth’s Cryosphere 15(1): 40-51

118

von Toll EV (1897) Ancient glaciers of New Siberian Islands, their relation to mammoth corpses and the Glacial period. Notes of the Russian Imperial Geographical Society 32: 1-137 (in Russian)

Ward MK, Pollard WH (2016) Using a mix-methods approach to monitor thermokarst activity in the Eureka Sound Lowlands, Ellesmere Island, Nunavut. Proceedings of the XI International Conference on Permafrost, Potsdam, Germany: 936

Yde JC, Knudsen NT, Hasholt B, Mikkelsen AB (2014) Meltwater chemistry and solute export from a Greenland Ice Sheet catchment, Watson River, West Greenland. Journal of Hydrology 519(B): 2165-2179

Yershov ED (1998) General Geocryology. Originally published in Russian. Cambridge University Press: Cambridge, UK

Young NE, Briner JP, Rood DH, Finkel RC (2012) Glacier extent during the Younger Dryas and 8.2-ka event on Baffin Island, Arctic Canada. Science 337: 1330-1333

119

Appendix A: Mokka Fiord – Axel Heiberg Island – 60 m a.s.l.

Gravimetric moisture content (%)

0 100 200 300 0

30 Figure A2: Mokka Fiord (MF) massive ice exposure

Sediment depth Munsell colour Organic matter Carbonates 40 (cm) content (%) (%) 0-5 5Y 4/2 8.79 1.37 Permafrost table 5-10 5Y 4/2 8.24 1.70 10-15 5Y 4/2 8.16 1.52 15-20 5Y 4/2 8.59 1.57 50 20-25 5Y 4/2 8.80 1.48 25-30 5Y 4/2 6.64 1.37 30-35 5Y 4/2 6.28 1.50 35-40 5Y 4/2 5.35 1.83 60 40-45 5Y 4/2 6.44 1.57 Depth from ground surface (cm) surface ground from Depth

Table A1: MF-1 core active layer sediment characteristics. 70

Sediment Sand (%) Silt (%) Clay (%) Texture class 80 depth (cm) 5-10 (AL) 11.48 56.33 32.19 Silty clay loam 45-50 (pmf) 6.5 63.61 29.89 Silty clay loam Table A2: Texture class data from MF-1 core. 90

Block samples δD (‰) δ18O (‰) d excess (‰) 100 MF-4 -235.22 -29.95 4.41 0 0.25 0.5 0.75 1 MF-5 -232.74 -29.61 4.12 fVi Table A3: Stable water isotopes in block samples (MF-4, MF-5) taken from a Mokka Fiord retrogressive thaw slump exposure. GMC Volumetric ice content

Figure A1: Gravimetric and volumetric moisture/ice content in AL and permafrost in a short core (MF-1) from a Mokka Fiord massive ice body. 120

Appendix B: Collapsing pingo, Mokka Fiord, Axel Heiberg Island (79°33’7.6”N, 86°47’58.2”W)

121

120

100

40 Concentration (mg/l) Concentration 20

0 Ca Na+K Mg SO4 Cl

Surface removed Not removed

D 18O d excess 50

-50

-100

-150 per mil (‰) mil per -200

-250

-300

Surface removed Not removed

Appendix C: Comparison of hydrochemistry of massive ice sampled using two methods at Dump Slump exposure. Standard practice when sampling ground ice from a natural exposure is to remove approximately 5 cm of surface ice to avoid samples contaminated by refreezing. In this micro-study, we observe significantly higher concentrations of dissolved ions in samples where surface ice was not removed. This difference was observed in all 2+ + + 2+ 2- major cation species (Ca , Na , K , Mg ). Concentrations of SO4 were more than double in samples where surface ice was not removed – but there was no difference in Cl-. The differences in stable isotope ratios were negligible. These findings serve as a reminder to remove surface ice when sampling ground ice to avoid contamination.

122