Optical Dating of Stabilized Parabolic Dunes, Savary Island, British Columbia

Optical dating of stabilized parabolic dunes, Savary Island, British Columbia

by Libby Biln

Bachelor of Science, University of the Fraser Valley, 2014

Thesis Submitted in Partial Fulfillment of the Requirements for the Degree of Master of Science

in the Department of Earth Sciences Faculty of Science

Copyright in this work rests with the author. Please ensure that any reproduction or re-use is done in accordance with the relevant national copyright legislation. Approval

Name: Libby Biln Degree: Master of Science Title: Optical dating of stabilized parabolic dunes, Savary Island, British Columbia Examining Committee: Chair: Dirk Kirste Associate Professor Brent Ward Senior Supervisor Professor Olav Lian Supervisor Adjunct Professor John Clague Committee Member Professor Emeritus Ian J. Walker Committee member Professor, Arizona State University Stephen A. Wolfe External Examiner Natural Resources Canada Geological Survey of Canada

Date Defended/Approved: June 13, 2017

ii Abstract

Research has shown that the south coast of British Columbia (BC) has experienced changes in relative sea level and climate since deglaciation (~15 ka ago); however, there has been little study of the landscape’s response to these changes. On Savary Island, in the Strait of Georgia, there exist large parabolic dunes that are unique to the region. These dunes are stabilized, supporting mature forest growing in well- developed soil, and they contain eroded palaeosols indicating that their formation was punctuated by periods of episodic stabilization and soil formation. Optical ages from K- feldspar indicate that dune formation began prior to 7.69 ± 0.71 ka and stabilized by about 5.47 ± 0.36 ka ago when relative sea level lowering was slowing and climate was becoming cooler and moister. Periods of landscape stability during dune formation were brief, probably lasting only a few hundred years.

Keywords: Parabolic sand dunes; optical dating; postglacial landscape change; radiocarbon dating; Coastal British Columbia; relative sea level change

iii Dedication

Firstly, I would like to dedicate this work to my loving and supportive parents. You have always pushed me to be the best version of myself, supporting me in everything I do. Thank you for understanding the importance of a good education and encouraging me to pursue my passions.

Secondly, this work would not have been possible without the support of my husband, Alec. You have been by my side throughout this entire journey, listening to my crazy OSL talk, putting up with my field-work absences, dealing with my frequent break downs, and editing my poor grammar. Your intense work ethic has encouraged me to push myself to new levels and get the most out of every day. I could not have accomplished this feat without you, I am beyond grateful.

♥ 

iv Acknowledgements

This work would not have been possible without the help, guidance and support of Dr. Olav Lian. His patience in teaching me the complicated nuances of research and optical dating are invaluable. I am very grateful to Dr. Brent Ward for taking me on as a masters student and allowing me to work with ‘black magic’. Your constructive criticism has helped me grow as researcher. I am also indebted to Dr. John Clague for allowing me to build on his initial findings and providing guidance in the realm of Quaternary sciences. Thanks are also due to Dr. Ian Walker for his insight into coastal processes and dune dynamics. Your advice has helped me to become a more well-rounded scientist. I am very grateful to Liz Webster from the Savary Island Land trust for introducing me to Savary Island, showing me the trails and letting us stay in her beautiful cabin. Many thanks to the external examiner, Dr. Stephen Wolfe, whose comments and advice greatly improved this thesis.

This work would not have been possible without the financial support of the Hakai Institute and Tula Foundation, notably Eric Peterson and Christina Munck. Many thanks for supporting Science on the coastal margin and in the Discovery Islands.

A big thank-you to Dr. Christina Neudorf who taught me the ways of analysing luminescence data, more than once, and answered my never ending questions in record breaking time. Your insight and expertise greatly assisted this research and I would not have optical ages without you. I would also like to acknowledge Dr. Paul Sanborn from the University of Northern British Columbia who provided soil science expertise and advice in the field and throughout this thesis. Additionally, Dr. Brian Menounos and Rob Vogt from the UNBC Lidar research group for facilitating lidar acquisition. Thanks to Derek Heathfield for his assistance with lidar processing.

To my Luminescence Dating Lab mates, Jordan Bryce and Travis Gingerich, thank you for all of your help in the field and laboratory, you made the ‘Bat-cave’ exciting. To my fellow grad students, thanks for keeping my graduate work caffeinated, full of punny moments, dance parties, and good conversation, it would not have been the same without you: Carie-Ann, Snowy and Patt. A big thanks to everyone who helped me in the

v field, and finally, to all the friends and family who have encouraged and supported me throughout this journey, your interest in what I do means the world.

vi Table of Contents

Approval ...... ii Abstract ...... iii Dedication ...... iv Acknowledgements ...... v Table of Contents ...... vii List of Tables ...... x List of Figures...... xi

Chapter 1. Introduction ...... 1 1.1. Study area: location, access and general character ...... 2 1.2. Regional overview ...... 4 1.2.1. Geology ...... 4 1.2.2. Climate and vegetation ...... 4 1.2.3. Coastal environment and conditions ...... 7 1.2.4. Quaternary history ...... 8 Late Quaternary history ...... 8 Olympia Interstade...... 8 Fraser Glaciation ...... 9 Postglacial period ...... 11 Glacial and postglacial climate ...... 12 Sea level history ...... 13 1.2.5. Savary Island lithostratigraphy and previous research ...... 15 1.3. Thesis objectives and rationale ...... 16 1.4. Thesis format ...... 17

Chapter 2. Methods...... 18 2.1. Pedology ...... 18 2.2. Radiocarbon dating ...... 19 2.3. Lidar analysis ...... 20 2.4. Optical dating ...... 20 2.4.1. Sample collection and preparation ...... 22 2.4.2. Estimating the equivalent dose (De) ...... 24 2.4.3. Estimating the environmental dose rate ...... 28 2.4.4. Anomalous fading ...... 30 2.5. Grain size analysis ...... 31 2.5.1. Palaeowind speed calculation ...... 32

Chapter 3. Results ...... 35 3.1. Introduction ...... 35 3.2. Radiocarbon ages ...... 35 3.3. Landforms and their interpretation ...... 37 3.4. Optical dating ...... 44 3.4.1. Savary Island quartz ...... 44 3.4.2. Developing a SAR protocol for Savary Island potassium feldspar ...... 47 Preheat plateau test ...... 48

vii Recycling ratio and recuperation dose test ...... 49 Dose recovery test ...... 49 3.4.3. Correction for anomalous fading ...... 50 3.4.4. Tests of the laboratory lighting ...... 51 3.4.5. Tests for grain size dependence on optical ages ...... 52 3.4.6. Optical ages ...... 53 3.5. Grain size and wind speed ...... 57 3.5.1. Grain size ...... 57 3.5.2. Palaeowind speed for aeolian sand transport ...... 64

Chapter 4. Postglacial landscape evolution of sand dunes on Savary Island, British Columbia ...... 66 4.1. Introduction ...... 66 4.2. Study area ...... 67 4.3. Methods ...... 71 4.3.1. Remote sensing and landform identification ...... 71 4.3.2. Lithostratigraphy ...... 72 4.3.3. Dune chronology ...... 72 Optical dating sample preparation and measurement ...... 74 Determination of the equivalent dose ...... 75 Determination of the environmental dose rate ...... 76 Optical age determination ...... 77 Radiocarbon dating ...... 77 4.4. Results ...... 78 4.4.1. Geomorphology ...... 78 4.4.2. Lithostratigraphy ...... 78 4.4.3. Radiocarbon and optical ages ...... 85 4.5. Summary and discussion ...... 90 4.5.1. Geochronological age estimation ...... 90 4.5.2. Landscape evolution ...... 94 4.6. Conclusions and future research ...... 98

Chapter 5. Conclusions ...... 100 5.1. Geomorphology ...... 100 5.1.1. Summary ...... 100 5.1.2. Future work ...... 101 5.2. Optical dating ...... 101 5.2.1. Summary ...... 101 5.2.2. Future work ...... 102 5.3. Pedology ...... 103 5.3.1. Summary ...... 103 5.3.2. Future work ...... 104 5.4. Summary of key findings and outcomes ...... 105

References ...... 106 Appendix A Site and pedon descriptions ...... 121 Site 1 ...... 121 SIDS5 ...... 129

viii SIDS6 ...... 131 SIDS7 ...... 133 SIDS9 ...... 135 SIDS10 ...... 137 SIDS11 ...... 139 Swale01 ...... 141 Swale02 ...... 143 Swale03 ...... 145 Swale04 ...... 147 Appendix B Lidar ...... 149 Appendix C XRF and XRD ...... 151 Appendix D Thin section analysis ...... 153 Appendix E. Grain size analysis...... 159 Appendix F Optical dating radial plots...... 163 Appendix G Dosimetry results...... 165 Appendix H Charcoal transport hypothesis ...... 166

ix List of Tables

Table 2.1 Depth of optical dating samples below the modern surface and their relation to palaeosols...... 23 Table 2.2. Values of d (Guérin et al. 2011) andH coefficients1 ...... 30

Table 2.3 Sample calculation of palaeowind speed using D50 grain size from SIDS1C...... 34 Table 3.1 Radiocarbon ages from charcoal fragments collected from B- horizons of palaeosols at sites 1 and 2...... 36

Table 3.2 Optical dating samples, total dose rates, equivalent dose (De), overdispersion (OD) values, fading rates (g), and optical ages (uncertainties are ±1) ...... 55 Table 3.3 Grain size statistics in phi (φ) units, after Folk and Ward (1957). Percentiles for verbal classifications are presented in Folk (1968) ...... 61 Table 3.4 Grain-size data calculated using the Malvern Mastersizer 2000...... 62

Table 3.5. Wind speed calculations from grain size data using D50 values. Sample calculation for SIDS1C shown in Table 2.3...... 65 Table 4.1 Radiocarbon AMS ages from charcoal fragments collected from soil B horizons at sites 1 and 2 (Figure 4.5)...... 85 Table 4.2 Optical dating sample depths, 40K, Rb, Th, and U concentrations, and water contents ...... 86

x List of Figures

Figure 1.1. Satellite image of Savary Island showing the region in which the vegetated parabolic dunes occur...... 3 Figure 1.2. Hourly wind speeds recorded from 1992 to 2016 by Environment Canada at Sentry Shoal (Station #46131)...... 6 Figure 1.3. Example of dune vegetation in the Coastal Douglas Fir biogeoclimatic zone on Savary Island...... 7 Figure 1.4. Geoclimatic and associated lithostratigraphic units for the Fraser Lowland...... 11 Figure 1.5. Sea level curves for the Strait of Georgia, near Savary Island. Curves were produced from data in Shugar et al. (2014) accessed from http://dx.doi.org/10.1016/j.quascirev.2014.05.022...... 14 Figure 1.6 Simple section diagram based on previous work (Clague et al. unpublished data) ...... 16 Figure 2.1 Locations of study sites on Savary Island (Figure 1.1)...... 18 Figure 2.2 Example of the radial plot, which is commonly used to visualize the spread in De values from individual aliquots of a sample ...... 28 Figure 2.3 Comparison of grain size distributions determined by different methods ...... 32 Figure 3.1 ‘Bare Earth’ digital elevation model of Savary Island created in GlobalMapper16...... 38 Figure 3.2. Locations of study sites on Savary Island. Detailed site descriptions and coordinates can be found in Appendix A...... 39 Figure 3.3 Identified, ground-truthed, and interpreted geomorphic features on Savary Island...... 40 Figure 3.4 Bare-Earth DEM produced from lidar data viewed in GlobalMapper16...... 43 Figure 3.5 Viewpoint of (A) and (B) show in in Figure 3.4 ...... 44 Figure 3.6 LM-OSL curves of quartz from two different regions in BC ...... 45 Figure 3.7. Graph showing recycling ratios and recuperation values of quartz aliquots measured for sample SIDS2 (180-250 µm) ...... 46 Figure 3.8 (A) Dose response curve for aliquot #39 of sample SIDS2 quartz (180-250 µm), which is typical of the curves for this sample ...... 47 Figure 3.9 Flowchart showing optical dating protocols applied to KF grains in this study include ...... 48 Figure 3.10 A) Results from the preheat plateau test on sample SIDS2 (180- 250 µm) KF...... 50

xi Figure 3.11 (A) Equivalent dose distribution of samples prepared under regular and dimmed room lighting conditions (see text), to test for partial bleaching during sample preparation...... 52 Figure 3.12. (A) Radial plot showing the De distributions resulting from two grain size ranges, obtained using SAR protocol 1 ...... 53 Figure 3.13 Grain-size distribution histograms of optical dating samples SIDS1 to 4...... 58 Figure 3.14 Grain-size distribution histograms of optical dating samples SIDS5 to 8...... 59 Figure 3.15 Grain-size distribution histograms of optical dating samples SIDS9 to 11...... 60 Figure 3.16 Cumulative percent grain size curves for of all samples ...... 63 Figure 3.17. Grain size distributions calculated using the Malvern Mastersizer 2000...... 64 Figure 4.1 (A) Location of Savary Island (inside box), approximately 145 km northwest of Vancouver, BC...... 69 Figure 4.2 Sea level curves for the Strait of Georgia near Savary Island ...... 71 Figure 4.3 Location of dune heads sampled for optical dating...... 74 Figure 4.4 Oblique ‘three-dimensional’ view of the central 4 km of Savary Island with associated topographic profile (A – A’)...... 78 Figure 4.5 Composite section diagram of sites 1 and 2 showing locations of optical dating samples and ages ...... 82 Figure 4.6 (A) Top of palaeosol 1 showing small scattered, charcoal fragments (indicated by arrows) and mottling; locations shown in Figure 4.3. (B) Close-up of palaeosol 2 and its upper contact, as indicated by the abrupt change in colour. The scale is in cm...... 84 Figure 4.7 A) Example of a luminescence decay curve (sample SIDS5) typical of those derived from the Savary Island samples ...... 87 Figure 4.8 Possible scenario for postglacial evolution of Savary Island since deglaciation...... 97 Figure 4.9 Optical ages (round black data points; error bars are ± 2) of dune head stabilization plotted with climate periods determined from regional palaeoecological studies, as compiled by Galloway et al. (2009) and based on radiocarbon dating...... 98

xii Chapter 1. Introduction

During the late Quaternary Period, coastal British Columbia (BC) experienced significant fluctuations in climate (Clague et al. 2004) that influenced sea level (Shugar et al. 2014), ice volume, and modified the landscape, which in turn impacted the presence and distribution of humans, animals, and vegetation in the area. The Pleistocene record of long-term climate change in BC and its impact on the landscape are recorded in sedimentary sections exposed in river valleys and sea cliffs (e.g. Armstrong and Clague 1977; Clague et al. 1980; Hicock and Armstrong 1981, 1983; Lian and Hickin 1993, 1996; Mathewes 1993; Ward and Thomson 2004). Although fluctuations in postglacial climate in coastal BC have been well documented through palaeoecology (e.g. Galloway et al. 2009; Lacourse 2005; Mathewes and Heusser 1981), the sedimentary history of coastal BC during this period has not received much attention.

Savary Island is the most southern of the Discovery Islands in the Strait of Georgia. The central portion of the island is capped with a thick sequence of aeolian sand that rests unconformably on sediments deposited during the last glaciation. The aeolian sediments form large (~600 m long) stabilized parabolic dunes that cover ~1 km2 of the island’s centre. Parabolic dunes are elongate, U- or V-shaped, vegetation- controlled features that consist of a deflation basin between elongate ridges (or arms) that form parallel to the dominant wind direction (Hesp and Walker 2013). Vegetation works to anchor the arms as the dunes migrated down wind. Parabolic dunes are rare in the region, and constitute the primary study area for this research. They are sensitive to environmental changes and require specific environmental conditions to form and become stabilized. As such, aeolian dunes can provide information about long-term environmental change. A reconnaissance investigation of the dunes in 2000 revealed the presence of several buried soils (palaeosols); charcoal fragments from the B horizons of some of them were radiocarbon dated to ~10,000 calendar years, which suggested that the soils are of this approximate age (Clague et al. unpublished data).

1 The palaeosols record periods of relative landscape stability and soil formation while episodes of sand accumulation record periods of relative instability and aeolian activity.

This research applies remote sensing (lidar), radiocarbon and optical dating, and pedology techniques to gain an understanding of the character and timing of postglacial geomorphic change on Savary Island, with specific focus on the parabolic sand dunes. As the island undergoes increased land development pressures and reduced biodiversity resulting from habitat loss caused by increased erosion due to sea level rise, urgency exists to understand the unique dune field. This research will add to our knowledge of postglacial environments in the region and give further insight into the effect of past climate and sea level change on landscape development.

1.1. Study area: location, access and general character

Savary Island, the southernmost of the Discovery Islands, lies at the north end of the Strait of Georgia in the Georgia Depression between Vancouver Island and the Coast Mountains of BC (Figure 1.1). Primary access to the island is via water taxi based out of the coastal town of Lund, situated about 4 km to the northeast. Savary Island is home to between 30 and 60 residents year round, but it receives a large influx of semi- permanent residents and tourists in the summer months who use the island for recreation. Its sandy beaches and mild climate have drawn tourists for over 100 years resulting in the subdivision of 1,111 acres of land into more than 1,700 lots. This very high lot density now poses a risk to groundwater contamination through septic fields and overuse of water (Harrington et al. 2005). There are no lakes on the island but it has one main aquifer, three shallow perched aquifers, and four springs, which together supply residents and visitors with limited fresh water (Tupper 1996).

Figure 1.1. (A) Savary Island (inside box) is approximately 145 km northwest of Vancouver. About 7.5 km long, it varies in width from 0.3 to 1 km. (B) Satellite image showing the region in which the vegetated parabolic dunes occur (outlined in white box); these dunes cover ~1 km2 of the land near the centre of the island.

Land ownership of the dune field is divided into two. One owner, the Nature Trust of BC, aims to preserve the dunes and the unique and fragile ecosystems associated with them. The other portion, is privately owned by a family whose aim is to

3 develop the property and create a gated community on the dunes, a goal that has been considered unsustainable considering the island’s limited resources.

1.2. Regional overview

1.2.1. Geology

British Columbia consists mainly of ancestral North American rocks and accreted terranes that make up the morphogeological belts of BC; these are from east to west: the Foreland (Rocky Mountain), Omineca, Intermontane, Coast Plutonic and Insular belts (Monger et al. 1972). Its coastal margin and islands belong to the most tectonically active Insular Belt, which consists of the exotic Wrangellia and Alexander terranes that are composed of a mixture of oceanic basalts, volcanic arc related rocks, minor limestone, and mid-Paleozoic continental crust (Monger et al. 1972). Savary Island is in the Insular Belt; however, is mainly sediment with the only exposed bedrock cropping out on its most eastern point, which consists of 200-400 Ma granite. The outcrop is colloquially referred to as Mace Point (Kennedy 1992). The majority of the island consists of sediments associated with the growth and decay of the last Cordilleran Ice Sheet (CIS) (Clague 1976) during the last (Fraser) glaciation. According to digital maps developed by Massey et al. (2005), a northwest- to southeast-trending fault ~5.3 km long lies ~5 km west of the centre of Savary Island; whether or not this fault is active is unclear.

1.2.2. Climate and vegetation

Savary Island lies in the rain shadow of Vancouver Island, at the northern end of the Coastal Douglas Fir biogeoclimatic zone. It experiences warm and dry summers and cool and moist winters (Tupper 1996). The mean annual temperature is 9.5 °C, with mean annual precipitation of 1230 mm, of which less than 20 % falls between May and September (Wang et al. 2012). The dominant wind directions are from the northwest in the summer months and from the southeast in the winter months (Thomson 1981) (Figure 1.2). In general, overall air circulation controls wind regimes; in coastal environments wind is influenced by gradients in pressure and temperature between land

4 and water, the morphology of the region, and cyclonic disturbances (Cooper 1958). On the Pacific coast, the Aleutian low pressure system influences present-day coastal climate, particularly in the winter months. Migration of this system in the past would have significantly impacted weather and climate along the coast and has been associated with glacier fluctuations (Heusser et al. 1985; Mann and Hamilton 1995; Anderson et al. 2005;). Soil environments in these climates (warm and dry) are typically dry and nutrient poor (Nuszdorfer et al. 1991).

Douglas fir (Pseudotsuga menziesii var. menziesii) and a dense salal (Gaultheria shallon) understory dominates the study area (Figure 1.3). The area also hosts large arbutus (Arbutus menziesii) and western red cedar (Thuja plicata) with an Oregon-grape (Mahonia nervosa) understory (Henderson 2003). Detailed studies of Savary Island’s dune ecology have been performed (i.e. Dunster 2000; Henderson 2003), and coastal processes have been examined (Bornhold et al. 1996). Groundwater resources have been investigated (Tupper 1996), and Savary Island has been included in a regional study of coastal sand ecosystems (Page et al. 2011), which details specific species at risk on Savary Island and how they compare to those in other sand ecosystems. A community perspective and historical look at Savary Island and its sensitive ecosystem are given by Harrington (2005) and include other islands in the Salish Sea (the body of water including the Juan de Fuca Strait, the Strait of Georgia, Puget Sound, and all of their connecting channels). The unique forest-dune ecosystem on Savary Island is the largest and the most intact and undisturbed ecosystem of its type in the Georgia Basin (Page et al. 2011); it hosts a number of endangered species such as the contorted-pod evening-primrose (Camissonia contorta), the Island Tiger Moth (Grammia complicata) (Page et al., 2011) and some rare moss and bryophyte species (e.g. Homalothecium arenarium) (Sadler 2000).

Figure 1.2. Hourly wind speeds recorded from 1992 to 2016 by Environment Canada at Sentry Shoal (Station #46131) graphed using the openair package for R programming language (Carslaw and Ropkins 2012). Data have been separated into seasons: spring (March to May), summer (June to August), fall (September to November), and winter (December to February). Winter has the strongest winds, predominantly from the southeast.

Figure 1.3. Example of dune vegetation in the Coastal Douglas Fir biogeoclimatic zone on Savary Island. This photo shows a dense salal (Gaultheria shallon) understory, small pines (Pinus contorta var. contorta) and Douglas fir (Pseudotsuga menziesii).

1.2.3. Coastal environment and conditions

Savary Island is known by locals as “Hawaii of the North” or the “Sandbar” because of its long sandy beaches and warm ocean temperatures. Ocean surface temperatures at Sentry Shoal (Stn # 46131), an Environment Canada Weather Buoy 10 km west of Savary Island, average 17°C in the summer months with an annual average just above 11°C. The ocean temperatures close to Savary Island are likely even warmer because of the shallow sedimentary shelf surrounding the island that allows the water to heat up in the sunlight. An additional factor suspected to aid in the warming of the ocean around Savary Island is the lack of local currents. Thomson (1981) modeled the strength of maximum flood and ebb tides in the Strait of Georgia and concluded that there is a significant reduction in strength around Savary Island. Here the southward-

7 propagating tides from the northern channels and the northward-propagating tides from the south meet. Thomson (1981) suggests the northward-propagating tides dissipate and ‘swirl’ around, likely in clockwise direction, near Lund. The mean tidal range on Savary Island is ~3.35 m (Thomson 1981).

Wave heights, and their erosive power, in the Strait of Georgia are proportional to the wind fetch and its strength and duration (Holthuijsen 2007). Sheltered significantly by Texada Island to the south and the other Discover Islands to the north, the fetch is small for Savary Island (11 to 22 km from the southeast), preventing waves from reaching over 2.5 m in height (data for Sentry Shoal, Environment Canada Station # 46131). Thurber Engineering Ltd. (Smith et al. 2003) found erosion rates of up to 0.41 m/a on the south-facing side of the island (Drawing 14-197-0-2 in Smith et al. 2003). The island will continue to slowly erode if sea level rises, projected to rise up to 50 cm by 2100, based on a variety of greenhouse gas emissions (e.g, James et al. 2014). Clague et al. (unpublished data) suggested that Savary Island was once much larger in extent and has since eroded back to its current state.

1.2.4. Quaternary history

The Quaternary stratigraphy and inferred glacial history of southwestern BC has been extensively studied (see reviews by Clague and Ward 2011; Clague and Mathewes 1989; Clague 1981). The Olympia Interstade (formally defined by Hebda et al. 2016), the Fraser Glaciation, and the postglacial period are the most recent geoclimatic periods in the region. The last of these three periods is of primary interest in this study, however understanding the nature and timing of local glaciation and the effect it had on the landscape is important for interpreting and understanding changes in the postglacial landscape.

Late Quaternary history

Olympia Interstade

The Olympia Interstade (commonly referred to as the Olympia nonglacial interval) spanned the period from ~52 ka 14C BP (57 ka cal BP) until 26 ka 14C BP (30 ka cal BP) (e.g. Armstrong and Clague 1977; Hebda et al. 2016), and is a time when the

8 Fraser Lowland and southwestern BC coast were ice free (Armstrong et al. 1965; Clague 1981). The Cowichan Head Formation (Armstrong and Clague 1977) refers to sediments deposited during the Olympia Interstade in southwest BC. This formation has two members, a lower member that consists of fluvial and marine sediments and an upper terrestrial member of plant-bearing silt, sand, and gravel with some bog, lagoonal, and littoral sediments (Clague 1976). Stratigraphic evidence for this interstade is found along the east coast of Vancouver Island (Alley 1979), near the mouths of some of the valleys bordering the Fraser Lowland (Lian and Hickin 1993, 1996; Hebda et al. 2016), and farther south in the Puget Lowland of Washington State (Easterbrook 1992); there is also evidence of it in the interior of BC (Fulton and Smith 1978; Clague et al. 1990; Lesemann et al. 2013). Previous studies (Kennedy 1992; Dunster 2000) suggest that Olympia Interstade sediments occur on Savary Island although they are not exposed at the surface. In general, deposits associated with the Olympia Interstade are rare, as glaciers scoured and eroded the landscape during the subsequent Fraser Glaciation, removing much of the sediments or buried them under thick blankets of drift. At the end of the Olympia Interstade, climate began to deteriorate (Hebda et al. 2016) and the Fraser Glaciation began.

Fraser Glaciation

The Fraser Glaciation refers to the last period of ice sheet development in BC. At its maximum, the ice sheet extended as far west as the edge of the continental shelf and as far south as ~47° N in the Puget Lowland of Washington State. The Cordilleran Ice Sheet reached its maximum southern extent by ~14.5 ka 14C BP (~17 ka cal BP) (Porter and Swanson 1998). At that time southern parts of the Laurentide Ice Sheet, which covered the rest of Canada, had already begun to retreat, although it was still advancing in NE BC (e.g.,Dyke and Prest 1987; Hickin et al. 2016). Changes in climate caused rapid retreat of the ice, leaving the coastal areas of southwestern BC ice free by 13 ka 14C BP (~15 ka cal BP) (Clague 1981). Although deglaciation was characterized by rapid retreat of the ice sheet, glaciers in some southern Coast Mountain valleys briefly re-advanced during Late-Glacial time (Clague et al. 1997; Eamer et al. 2017).

At the onset of the Fraser Glaciation ice travelled from north to south down the Strait of Georgia. During this time, sandy outwash known as Quadra Sand was

9 deposited at the front of the advancing glaciers (Clague 1976). This time-transgressive unit documents the onset of major climate deterioration in coastal BC and the onset of the Fraser Glaciation; it is the oldest geoclimatic unit in the Fraser Glaciation drift package (Figure 1.4). These generally well sorted, horizontally and cross stratified sands are widely present throughout the Georgia Depression. The type section for Quadra Sand is on Quadra Island, 45 km northwest of Savary Island, in 40-m-high coastal bluffs. Clague (1976) shows a 200-m-long section of Quadra Sand on the south side of Savary Island overlain by a thin unit of till and ice-contact and glaciomarine sediments. The section does not show the capping aeolian unit, likely because the study was primarily interested in glacial stratigraphy. On Hernando Island, about 4 km northwest of Savary Island, water well logs document Quadra Sand persisting to a depth of 25 m below relative sea level (RSL) (Clague 1976).

Fraser Glaciation drift includes sediments deposited during advance, retreat, and in some areas re-advance of ice. Figure 1.4 shows the different lithostratigraphic units associated with the Fraser Glaciation in the Fraser lowland (in blue) and bounding units. Commonly overlying the Quadra Sand is Vashon Drift (Hicock and Armstrong 1985), which includes sediments deposited during the Vashon Stade. In the western Fraser Lowland there was a short-lived older advance of Fraser Glaciation ice, called the Coquitlam Stade (Hicock et al. 1999; Ryder and Clague 1989). In that region, Port Moody Interstade sediments (Hicock and Armstrong 1981, 1983; Hicock and Lian 1995; Lian et al. 2001; Ward and Thomson 2004) separate sediments deposited during the Coquitlam and Vashon stades. At least two minor re-advances of Fraser Glaciation ice occurred during general retreat of the glacier in the eastern Fraser Lowland: the Sumas advance occurred between about 12 and 11 ka 14C BP (14 – 11 ka cal BP) (Clague et al. 1997) and is represented by Sumas drift, which consists of till, outwash, and ice contact sediments (Easterbrook 1969). A second re-advance in the Squamish river valley of the Squamish valley glacier occurred shortly after 12.8 ka cal BP (Friele and Clague 2002). Sediments representing the Coquitlam and Sumas stades, and the Port Moody Interstade, have not been found outside of the Fraser Lowland. According to Clague and James (2002, their Fig. 5(b)), Savary Island became ice free by ~12 ka 14C BP (~14 ka cal BP).

Figure 1.4. Geoclimatic and associated lithostratigraphic units for the Fraser Lowland. Ages are in radiocarbon years; time divisions are not to scale. Cells filled with blue indicate formal lithostratigraphic units associated with the Fraser Glaciation, marked by the onset of deposition of Quadra Sand, the lowermost unit directly observed on Savary Island today. The yellow area highlights the postglacial period, which includes the Holocene, when sand dunes were likely forming in places along the shoreline in the Strait of Georgia region. A discussion of older units can be found elsewhere (e.g. Clague 1981). Modified after Table 44.1 of Clague and Ward (2011).

Postglacial period

As climate began to warm after the Fraser Glaciation, BC entered the postglacial period, the time following disappearance of glaciers in the region. In southwestern BC ice in the lowlands had disappeared before the start of the Holocene, the formal period that documents a shift in global climate to warmer conditions 11.7 ka cal BP ago (Walker et al. 2009). In southwest BC, palaeoecological studies (Mathewes and Heusser 1981; Clague and Mathewes 1989; Hebda 1995; Mandryk et al. 2001; Pellatt et al. 2001; Lacourse 2005; Galloway et al. 2007, 2009) have increased our understanding of the climate and vegetation of the postglacial period (discussed later). Salish Sediments (Armstrong 1984) is the formal lithostratigraphic unit that represents postglacial sedimentation in the Fraser Lowland. These sediments blanket much of the surface of modern southwestern BC and include all sediments deposited since the relative stabilization of RSL and after ice left the region (McCammon 1977). In the Fraser Lowland, Fraser River sediments form a lithostratigraphic unit separate from Salish

11 Sediments (Armstrong 1981; 1984; Clague et al. 1983), but were also deposited during postglacial time.

Aeolian deposits can contribute to the postglacial sedimentary record significantly in areas that experience high winds and available sediment supply. They can form thick deposits or thin veneers on the surface, commonly forming the parent material for modern soils (e.g. Saunders 1985). Within a stratigraphic section, these sediments can reflect environmental change and give insight into past climates.

Glacial and postglacial climate

Past climate in southwest BC has undergone notable fluctuations largely influenced by Milankovitch solar insolation cycles (Walker and Pellatt 2003) and fluctuations in the Aleutian low pressure system (Galloway et al. 2009). Based on palaeoecological studies, the postglacial period in southwest BC can be divided into three broad climate periods (Hebda 1995; Heusser 1983; Mathewes and Heusser 1981): After a cold and dry glacial interval beginning ~26.5 ka 14C BP (~30.8 ka cal BP) and ending ~12 ka 14C BP, (1) the region first experienced a cool and moist period from about 12 to 10.5 ka 14C BP (14 to 11.5 ka cal BP) (Clague 1981; Galloway et al. 2009; Heusser 1983; Mathewes and Heusser 1981). (2) This was followed by a warm and dry period until ~6 ka 14C BP (~6.8 ka cal BP) (Clague 1981; Galloway et al. 2009; Heusser 1983; Mathewes and Heusser 1981; Walker and Pellatt 2003). Evidence of a warm and dry climate during this time is confirmed by the presence of Douglas fir (Pseudotsuga menziesii var. menziesii) and alder (Alnus rubra) pollen (Heusser 1983, 1985). The most dramatic rise in temperatures occurred between 10.5 and 10.0 ka 14C BP (12.5 to 11.3 ka cal BP) when there was rapid and pronounced warming after the Younger Dryas cold period. In southern BC, the highest mean July temperatures (~16 °C), and lowest precipitation near 8 ka 14C years BP (~8.9 ka cal BP) (Heusser et al. 1985; Mathewes and Heusser 1981). Mathewes and Heusser (1981) call this warm and dry period the Holocene xerothermic interval. Clague and Mathewes (1989) concluded, upon reviewing palynological studies in the southern Canadian Cordillera, that maximum Holocene warmth likely occurred between 9.1 and 7.6 ka 14C BP (10.3 – 8.4 ka cal BP). Research from northwestern Washington State also indicated that the warmest and driest climate in this interval occurred between about 10 and 7 ka 14C BP (11.6 – 7.8 ka cal BP). This warm and dry period ended ~6 ka 14C BP (~6.8 ka cal BP) and climate

12 transitioned into (3) cooler and wetter conditions, similar to that of today (Mathewes and Heusser 1981). Pollen records showing an abundance of western hemlock (Tsuga heterophylla) and western red cedar (Tsuga heterophylla) which provide evidence of these moister conditions (Galloway et al. 2007).

Sea level history

Eustatic, steric, and isostatic effects, sediment loading, and tectonic uplift or subsidence affect sea level relative to land positions. Eustatic, or global, sea level is affected by a change in the mass or volume of the oceans (Rovere et al. 2016); it is the change in sea level relative to the centre of the Earth (James et al. 2014). Relative sea level (RSL) change refers to changes in the land surface in relation to the sea surface (Rovere et al. 2016). Regionally, in addition to factors affecting global sea level, erosion, sedimentation, compaction of sediments, differential loading of the CIS and local variations in the viscosity of the mantle have resulted in regions of crustal uplift and areas of depression. Together, these have caused major fluctuations in regional RSL since deglaciation of BC’s coast (Clague et al. 2005). A recent review of RSL change over the entire Pacific Northwest region from northern California to southern Alaska (Shugar et al. 2014) illustrates how the patterns of RSL change since deglaciation had varied spatially. Postglacial RSL studies in the Strait of Georgia (e.g., Clague et al. 1982; Hutchinson et al. 2004; James et al. 2005; see also review by Shugar et al. 2014) provide insight into the RSL history near Savary Island.

The sea level curves reflect differences in the time of deglaciation in the region, which are not the same at the north and south end of the Strait of Georgia (Figure 1.5). As glacial ice retreated, the isostatically depressed land began to rebound and RSL began to fall. In the southern areas, sea level dropped below the modern datum ~12 ka cal BP (Hutchinson et al. 2004), while in the northern parts, near Quadra Island and Campbell River, it appears to have remained above modern datum at that time. Savary Island lies on the northern edge of a zone where Shugar et al. (2014) suggest RSL fell below the present datum approximately 12.2 ka cal BP (Figure 1.5) although no evidence has been found to support a sea level below present-day sea level on Savary Island or on Quadra Island, 45 km northwest of Savary Island. In fact, a low stand in RSL has not been identified in the northern Strait of Georgia (Figure 1.5B) (Shugar et al. 2014), but it is suggested to have occurred before 10 ka cal BP (James et al. 2005). All

13 RSL curves for the region (Hutchinson et al. 2004; James et al. 2005; Shugar et al. 2014) show that sea level rapidly dropped from a high stand at ~150 m from 14 to ~12 ka cal BP.

In the central Strait of Georgia, RSL was more or less stable after 11 ka cal BP as isostatic rebound slowed and began to match the rate of eustatic sea level rise (Hutchinson 1992; James et al. 2005). Prior to this, isostatically depressed land allowed for rapid marine transgression of low-lying areas (Mann and Hamilton 1995). The sea level curves (Figure 1.5) reflect isostatic rebound of the land followed by gradual stabilization from ~8 ka cal BP to present.

Figure 1.5. Sea level curves for the Strait of Georgia, near Savary Island. Curves were produced from data in Shugar et al. (2014) accessed from http://dx.doi.org/10.1016/j.quascirev.2014.05.022.

14 1.2.5. Savary Island lithostratigraphy and previous research

Stratigraphic evidence exists primarily in eroding bluffs along the southern side of the island and at some exposures on its north side (Clague 1976), while evidence of older units is recorded in water well logs. The exposed lithostratigraphic succession consists of three units (Clague, unpublished data): Unit 1 is exposed at and above present sea level, the lower contact is not observed. It consists of 13 m of horizontally and cross-bedded sand and gravel. Unit 1 is overlain unconformably by ~0.5 m of highly consolidated, clast- to matrix-supported diamicton (unit 2), which, in turn, is overlain sharply by up to 10 m of weakly horizontally bedded and cross-bedded well-sorted sand (unit 3). In places, the diamicton (unit 2) is entirely or partially replaced by boulders, cobbles, and pebbles. In the central part of the island, unit 3 forms stabilized parabolic dunes at the surface. In exposures on the south side of the island, unit 3 contains two prominent yellowish-brown zones that grade downwards to a light grey colour over ~70 cm. They contain rare, scattered, charcoal fragments, typically a few millimetres in diameter, most of which are found within 20 cm of the top of the zones; several fragments were collected for radiocarbon dating (Clague et al. unpublished data). Higher in the unit 3 there is evidence of at least two other yellowish-brown zones, but these are much thinner (15 to 20 cm thick) and weaker (less developed).

Unit 1 is interpreted to be Quadra Sand, and unit 2 as Vashon till. Unit 3 consists of aeolian sand; the yellowish-brown zones are interpreted to be palaeosols (Clague, unpublished data)(Figure 1.6). These units were analysed again during the course of the present research, and additional radiocarbon and optical dating samples were collected from unit 3 to complement the radiocarbon ages provided previously by J.J. Clague. This information is presented in Chapter 4.

Figure 1.6 Simple section diagram based on previous work (Clague et al. unpublished data). A) Yellowish-brown zone 2 referred to as Paleosol 1 B) Yellowish-brown zone 1 referred to as Paleosol 2. Photographs of palaesols were taken in 2015

1.3. Thesis objectives and rationale

A better understanding of the character and timing of postglacial geomorphic change on Savary Island was achieved based on the following primary objectives, which were to:

1) Interpret Light Detection and Ranging (lidar) imagery collected from the island during the summer of 2014 and use it to map and interpret landforms. Lidar imagery was also used to identify key sites for optical dating, and, where possible, for radiocarbon dating. Since the island’s landscape is covered by

16 thick forest, lidar data allows accurate detailed maps to be made showing the ‘bare earth’ landscape, which makes previously undiscovered geomorphic features visible.

2) Develop and test optical dating laboratory protocols. Optical dating was needed to estimate the time of dune stabilization and to provide limiting ages for palaeosol formation. Optical dating is inherently experimental so laboratory protocols have to be developed and tested each time the method is applied for the first time at a particular locality.

3) Apply pedology techniques to investigate the environmental conditions in which the palaeosols formed, and link the palaeosols to other palaeoclimate records. Soils can provide information about the environment in which they formed and they indicate periods of land-surface stability. They may also be a key component in understanding why this rare forest-dune ecosystem exists on Savary Island.

1.4. Thesis format

The objectives are addressed in the following chapters. Chapter 2 describes the methods used in this study, focussing on optical dating. Chapter 3 details the results. Chapter 4 is a manuscript intended to be submitted to a journal and thus incorporates a shortened version of the methods from Chapter 2 and many of the results from Chapter 3. The last chapter in this thesis, Chapter 5, summarises the main findings and discusses the potential implications they have on the region in addition to potential avenues of future research.

17 Chapter 2. Methods

Fieldwork was done in June of 2014, and in April, August, and September of 2015. The stratigraphy at select sites was described, measured, and photographed, and dune sands were sampled for optical dating and radiocarbon dating to develop a chronology of dune activity and stabilization. Figure 2.1 shows site locations, and detailed site and pedon descriptions can be found in Appendix A.

Figure 2.1 Locations of study sites on Savary Island (Figure 1.1). The ‘SIDS’ (‘Savary Island Dune Sand’) labels also refer to optical dating samples. Imagery from Google Earth.

2.1. Pedology

Modern soils on Savary Island were described and interpreted in order to compare them to the palaeosols (sites 1 and 2) that developed during dune formation. The goal was to gain an understanding of the environment under which each soil developed. In general, the primary factors influencing soil formation include climate,

18 biota, topography of the land, parent material (texture and mineralogy), and time (Jenny 1994; Pye and Tsoar 2009). In the case of soils developed on sandy parent material, rates of pedogenesis are typically higher because of the lower surface area caused by narrow grain size distribution (Schaetzl and Anderson 2005). Leaching depth typically increases with overall percentage of sand in parent material. Soils were investigated and described on dune heads (where optical dating samples were collected) and in the low-lying areas (swales) between dunes.

Soils at swale sites were compared to soils studied at dune heads; although close in distance, the topography, and presumably the microclimate, between dune heads and the swales may differ. At each site a soil pit was dug to a depth below the lowest pedogenic horizons; this transition usually occurred about 1 m below the modern surface. Soil horizons were identified, measured, and described using the Canadian System of Soil Classification (Soil Classification Working Group 1998); the palaeosols and incipient palaeosols at sites 1 and 2 were described in the same manner (Table A1 and A2). Descriptions and interpretations of some of the palaeosols, and some of the modern soils, were made with the assistance of Dr. Paul Sanborn, University of Northern BC (UNBC). Detailed site and pedon descriptions and interpretations can be found in Appendix A.

2.2. Radiocarbon dating

Charcoal fragments collected for radiocarbon dating during the 2014 and 2015 field seasons (see section 4.4.2 Lithostratigraphy) were intended to lend support to the ages obtained from similar samples collected and dated by J.J. Clague in 2000. The samples, which consisted of rare pieces dispersed in the sediment matrix, typically <1 cm in diameter, were inspected and cleaned by Alice Telka (Paleotec Services). Those deemed suitable for dating were forwarded to the Keck Carbon Cycle Facility (University of California, Irvine) where they were given a standard acid-base-acid treatment to remove any contaminating carbonates and soil humics. Samples were then washed with purified water, dried, and converted to graphite for radiocarbon dating by accelerator mass spectrometry (AMS). All ages were subsequently calibrated using OxCal 4.2 and the IntCal 13 data set.

19 2.3. Lidar analysis

Topographic maps, aerial photographs, and satellite images that existed at the beginning of this study were not sufficient to clearly resolve the geomorphology beneath the forest on Savary Island. To overcome this problem, the UNBC lidar Research Group was commissioned to fly lidar over Savary Island in August 2014. A Global Navigation Satellite System station established on Quadra Island, 45 km northwest Savary Island, was used as a base station for spatial referencing. Scanner specifications used during the flight are detailed in Appendix B. Rob Vogt (UNBC) processed the raw lidar data; ground classification of lidar data was done in LAStools software suite using automatic filter ‘lasground’ with the setting ‘extra_fine’. Notably, this ground classification algorithm struggles particularly for sharp ridges. The average point density covered was 12.5 pts/m2 of which 17 % were classified as ground points and extracted for use in creating the ‘bare earth’ digital elevation model (DEM). Derek Heathfield (University of Victoria) processed data further to develop a DEM that was used for further analysis. In Quick Terrain Modeler, adaptive triangulation with 1-meter grid spacing was used for DEM creation.

ArcGIS 10.0 software was used to perform principal component analysis, based on methods described by Devereaux et al. (2008), on eight hillshade layers of the DEM with azimuths ranging from zero to 360° at 45° intervals, each with an ‘altitude’ of 45°. This analysis resolved geomorphic features on the island, in particular the sand dunes. Using ArcGIS and GlobalMapper 16, digital geomorphic maps were created. These maps, together with a GPS-enabled tablet computer, were used in the field to locate specific landforms, for example the dune heads to be sampled for optical dating. Landform interpretation and key site identification were performed in ArcGIS with the assistance of the ‘contour with barriers’ line generation algorithm. Global Mapper was used for visual analysis, cross-section analysis and quantification of landforms.

2.4. Optical dating

Optical dating is used to determine when mineral grains (typically quartz and feldspar) were last exposed to sunlight or heat. It is based on the fact that mineral

20 grains (sediment) contain structural defects and impurities within their crystal lattices, some of which can act as traps for free electrons. When exposed to sunlight, light- sensitive electron traps are emptied (the sample is ‘bleached’). When sediment grains become buried and are shielded from sunlight, the traps begin to accumulate free electrons (see reviews by Lian 2013; Roberts and Lian 2015; Roberts et al. 2015). Free electrons are produced from the valence bands of atoms when they absorb radiation from the immediate environment and from cosmic rays. In practice, sediments are kept in the dark when collected. In the laboratory, prepared and separated (concentrated) grains are stimulated by light of a specific wavelength or wavelength range, typically infrared (IR) light for feldspar and blue light for quartz; during stimulation the electrons in light-sensitive traps are released, and they promptly recombine at other sites where excess energy is given off as luminescence of a wavelength that is shorter than that of the stimulating beam (Aitken 1998). The luminescence signal decays (bleaches) exponentially over time with the most light-sensitive traps being emptied first. The luminescence decay curves (or ‘shine-down’ curves) are therefore the sum of many such curves, each representing a population of traps with various sensitivities to light (Aitken 1998). For quartz it has been found that the signals represented by some of the component decay curves are not suitable for dating, as they represent traps that are not thermally stable over geologic time. Fortunately it is a type of trap, or population of traps, that bleaches quickly that is most suited to dating, and it is responsible for the so- called ‘fast component’ of the quartz decay curve; only a few seconds of direct sunlight exposure is needed to reset its signal. Other principal components are referred to as ‘medium’ and ‘slow’ (and there are also other less well-defined components) (Bailey et al. 1997), and these are usually avoided: If the fast component is missing from the decay curve of a particular quartz sample, then that sample is usually deemed unsuitable for dating. The presence or absence of these decay components can be assessed using linear modulation, where the power of the stimulation light is steadily increased over measurement time (Bulur 1996). For feldspar, however, such well-defined decay components have not been identified in the IR-stimulated signal, and it is therefore likely that its decay curve results from a continuum of light-sensitive traps (Wintle 2008), and they all appear to be appropriate for dating. The intensity of the luminescence signal is proportional to the number of trapped electrons and therefore the time elapsed since the sand grains were last exposed to sunlight (i.e., the burial age). In practice, an optical

21 age is determined by recording how a sample’s luminescence intensity varies with increasing doses of laboratory radiation (γ or β radiation), and from that, the dose of laboratory radiation that gives the same intensity of luminescence as did the environmental dose (which came from of α, β, and γ radiation, and from cosmic rays) is estimated – the so-called equivalent dose, De. The De (measured in grays, Gy) is divided by the measured environmental dose rate, 퐷̇ 푇 (measured in Gy/ka), which gives the optical age in ka (1) (Galbraith and Roberts 2012). Reviews can be found in Lian and Roberts (2006), Roberts and Lian (2015), and Roberts et al. (2015).

퐷 (퐺푦) Optical age (in ka) = 푒 (1) 퐷̇ 푇 (퐺푦/푘푎)

2.4.1. Sample collection and preparation

Optical dating samples were collected in opaque aluminum tubes ~23 cm in length and ~7 cm in diameter. Each sample tube was labeled and sealed with opaque black plastic caps and Duct Tape to keep the sediment from mixing and to preserve water content. A total of 11 optical dating samples were collected (Figure 3.2), five of which were associated with palaeosols formed in dune sediments (SIDS1 to 4 and SIDS8). Samples collected beneath palaeosols came from an excavated face of a coastal bluff. Sample tubes were inserted horizontally into clean aeolian sand below palaeosols near the locations where samples for radiocarbon dating were collected. Samples from six dune heads (SIDS5 to 11, not including SIDS8) were taken horizontally from pits ~1 m below the surface (see Figure 2.1 for locations and Table 2.1 for depths). Dune heads were selected for sample collection because they are expected to give an age that closely approximates the time of most recent dune activity and thus the age of dune stabilization (Wolfe et al. 2002). If samples had instead been collected from the stoss slope of the dunes, optical ages would date an older part of the dune in an area of erosion (Wolfe et al. 2002). If samples had been collected from the arms of the dunes, optical ages would also date an older part of the landform as the arms form and stabilize first as the dune migrates down-wind. Appendix A gives additional details on sample site locations and associated soil descriptions.

22 Table 2.1 Depth of optical dating samples below the modern surface and their relation to palaeosols. Depth below Sample Relation to palaeosols surface (cm) SIDS1 690 40 cm below palaeosol 1 (Site 2) SIDS1C1 690 40 cm below palaeosol 1 (Site 2) SIDS2 920 45 cm below palaeosol 2 (Site 2) SIDS3 296 20 cm below insipient palaeosol 1 (Site 2) SIDS4 250 26 cm above insipient palaeosol 1 (Site 2) SIDS5 140 N/A SIDS6 110 N/A SIDS6B2 110 N/A SIDS7 120 N/A SIDS8 480 70 cm below insipient palaeosol 2 (Site 1) SIDS9 105 N/A – collected below modern surface/soil SIDS10 140 N/A – collected below modern surface/soil SIDS11 135 N/A – collected below modern surface/soil 1 Duplicate samples collected from the same site as SIDS1. 2 Duplicate samples collected from the same site as SIDS6.

In the laboratory, under dim orange light (several layers of Lee 158 ‘deep orange’© optical film wrapped around masked fluorescent tubes), quartz and K feldspar (KF) grains were extracted from the samples using standard procedures (Wintle 1997). KF was chosen over other feldspar species because it has received the most attention, and therefore testing, as a chronometer. After HCl and H2O2 treatments to remove carbonates and organics, respectively, grains were sieved into multiple grain size ranges for testing. The primary grain size range used for dating was 180-250 µm diameter fraction, with some further tests run on the 300-400 µm fraction of SIDS1C. After the grains of the desired size range were isolated, ‘heavy liquid’ (lithium metatungstate) at a density of 2.62 g/cm3 was used to separate quartz from heavy minerals, and then KF was separated from quartz using a density of 2.58 g/cm3. Quartz grains were then treated with 50 % HF acid for 40 minutes while KF grains were treated with 10 % HF acid for 5 minutes; this procedure etches (removes) the outer portion of the grains affected by alpha radiation during burial, and thus simplifies the dose rate calculation. For the quartz fractions, the HF acid treatment also serves to dissolve any contaminating feldspar grains not removed during heavy liquid separation. After HF acid treatment,

23 samples were subsequently subjected to an HCl acid wash to remove fluorides produced during the digestion of KF, which may form luminescence-blocking coatings on the grains (Preusser et al. 2009). Prepared grains were mounted onto ~1 cm diameter aluminum disks coated with silicone oil using a 2 mm mask, which resulted in aliquots each containing ~100 grains.

Since other feldspar species have densities close to that of KF, it was important to confirm the presence of KF in the separated samples. This was done by semi- quantitative X-ray diffraction (XRD) (Appendix C) at the Saskatchewan Research Council (SRC) Advanced Microanalysis Centre. The presence of KF in the samples was also confirmed by thin section analysis: bulk sub-samples from optical dating samples (SIDS1C, SIDS2, SIDS5, and SIDS9) were sent to Vancouver Petrographics LTD for preparation. Thin sections were mounted on 3 cm wide microscope slides using blue resin to aid in identification of KF grains. Under a petrographic microscope, mineral identification was performed using Mackenzie and Adams (1994) as a guide. Digital microscope images were recorded under plain and polarized conditions at various angles (Appendix D). To facilitate quantification of the KF abundance, grains were traced (using Adobe Illustrator) to help identify them and to sort them, to give an estimate of percent composition. These results were then compared to results obtained from XRD analysis (Appendix D, Table D1).

2.4.2. Estimating the equivalent dose (De)

Several methods can be used to estimate De; most of them have been reviewed by Lian and Roberts (2006), Wintle (2008), Lian (2013), and Roberts et al. (2015). For this work a variation of the basic single-aliquot regenerative-dose (SAR) method (Murray and Wintle 2003) was used. The basic SAR method involves measuring the luminescence from a single aliquot held at an elevated temperature (usually 125 °C for quartz and 50 °C for feldspar) several times, each measurement occurring after a laboratory dose of radiation and a heat treatment (‘preheat’), each subsequent dose being different (usually higher) than the previous one. Measurement at an elevated temperature serves to keep the thermal conditions of the grains constant and, for quartz, it also prevents freed electrons from becoming re-trapped in less stable traps during stimulation. The preheat treatment is required to assure that only electron traps that are

24 thermally stable over geological time are measured, it removes electrons from thermally unstable traps that may become filled during laboratory irradiation. The preheat has, however, an unwanted effect in that some electrons can be transferred from light insensitive traps to the light sensitive traps of interest, which can result in the measured luminescence and the calculated age being higher than they should be; the degree at which this thermal transfer affects the luminescence signal therefore has to be assessed. An unwanted effect of repeated irradiation, heating, and measurement, is that the sensitivity of the mineral grains can change. To correct for this, the aliquot is given a small dose of radiation (a ‘test dose’) following the initial luminescence measurement, and then another, usually less severe, heat treatment, and the luminescence is measured again. The luminescence measured after each test dose is used to normalize the signal measured following each regenerative dose. In practice, the luminescence recorded over the initial part of the signal is measured (usually recorded over the first 0.4 s for quartz, while longer intervals can be used for feldspar), to capture the most light-sensitive part of the signal (and the ‘fast component’ for quartz), with the background signal (typically that recorded between 90 and 100 s of stimulation) subtracted from it. For this work KF aliquots were measured for a total 200 s, and the first 5 s of the signal, with the last 40 s subtracted. These data are used to plot a curve (a ‘dose response’ or ‘growth’ curve) of normalized luminescence intensity as a function of laboratory dose. The normalized luminescence signal measured initially from the aliquot (the ‘natural’ signal), before any regenerative doses are given, is interpolated onto the dose-response curve, and the De value for that aliquot is read off the dose axis. A variation of this procedure involves measuring the IRSL signal from feldspar at high temperature, for example at 225 °C, which remains after measurement at standard temperatures (50 °C) for 100 s (Thomsen et al. 2008). This so-called post-IR-IR signal has been observed to suffer far less from anomalous fading (discussed below) for most samples, but bleach much more slowly, than the initial signal commonly used to find a

De.

Optical dating is inherently experimental, and the laboratory protocols needed to provide a reliable age differ between mineral types, and, regionally, even between the same mineral species. The choice of preheat temperature and duration over which it is applied are therefore important: a temperature-duration combination that empties electrons from all thermally-unstable traps, while minimizing both erosion of electrons

25 from thermally-stable traps and changes in sensitivity is desired. To get an idea of the range of suitable preheat temperature-duration values for a particular sample, a preheat plateau test is often completed. This involves either observing how De values or dose- recovery ratios (discussed below) change with increasingly stringent temperature- duration combinations; the latter approach was used for this work. The choice of test dose is also important; if it is too low, the luminescence measured as a result of it may not be adequate to correct for sensitivity change. To assess the adequacy of the chosen preheat temperature-duration and test dose, various quality-control tests are built into the SAR method. They include (i) repeating one of the regenerative doses, measuring the luminescence, and comparing the resulting signal intensity to that measured after it was given that dose the first time, and (ii) observing the signal measured after stimulation and the preheat treatment, but without a prior dose of radiation (often called the recuperation signal). The former, referred to as the recycling ratio test, serves to check the effectiveness of the test dose to correct for sensitivity change, and the latter gauges the degree of thermal transfer. Typically, the recycling ratio should be within 10% of unity, and the signal measured as a result of thermal transfer (recuperation) should be no more than 5% of that measured from the ‘natural’ (Murray and Wintle 2000; Wintle and Murray 2006). For young samples recuperation can be significant, but it may be mitigated by adding a step to the end of the SAR sequence where the aliquot is stimulated for 40 s while held at a temperature higher (typically 280 °C) than that of the preheat. This step, sometimes referred to as a ‘hot wash’ treatment, serves to transfer, and concurrently eject, electrons that are thermally transferred from any light insensitive traps into the traps measured for dating, during the preheat (Wintle and Murray 2006). Once all of the built-in quality-control conditions are satisfied, an additional quality control test is performed, which involves exposing a number of aliquots of the grains to light (usually natural sunlight) so that all the relevant traps are emptied. The aliquots are then given a dose of laboratory radiation similar to, or slightly higher than, the De found using the chosen SAR protocol. The ratio of the applied dose to the recovered dose is called the dose-recovery ratio, and if it is consistent with unity, the SAR protocol is deemed viable (Wintle and Murray 2006). Dose-recovery experiments cannot, however, account for effects related to differences between laboratory and field conditions, for example effects that may result from dose-rate differences at and near the sample site, so dating some samples from the sedimentary unit of interest, for which their ages are

26 well known by independent methods, is advisable whenever possible. An account of the experiments performed in order to develop and test a viable SAR protocol for quartz and feldspar on Savary Island is detailed in Chapter 3.

Once a suitable SAR protocol has been found, many prepared aliquots of a sample are measured. The number measured depends to some degree on depositional environment: for aeolian sediments, where it is expected that all grains experienced sufficient sunlight exposure prior to burial, fewer aliquots usually need to be measured compared to, for example, a fluvial deposit where grains might be expected to have different bleaching histories. The size of the aliquots also matters – the smaller the aliquot the greater the possibility of different De populations being identified. Aliquots that pass the recycling ratio test and recuperation test are accepted, and plotted on a radial plot (Figure 2.2) The distribution of De values is then analyzed using the central age model (CAM) or another suitable model (see below), which calculates the weighted- mean equivalent dose for the set and takes into account the extra spread

(‘overdispersion’, OD) in the De values above and beyond that associated with the measurement uncertainties (Galbraith et al. 2005; Roberts et al. 1998). It has been found that even samples that have been well-bleached before burial can have OD values of 20%, sometimes even as high as 30% (see Table 4 in Arnold and Roberts 2009; Galbraith et al. 2005; Jacobs and Roberts 2007; Olley et al. 1998). The rule of thumb is therefore for sample De distributions with OD values ~20 or less, the CAM is applied, otherwise another model is used in an attempt to resolve De population(s), such as the minimum age model (MAM) or the finite mixture model (FMM) (Galbraith et al. 1999; Galbraith 2005).

For all the experiments, sample aliquots were measured using a Risø TL/OSL DA-20 reader/irradiator equipped with a calibrated 90Sr/90Y source that delivered β particles to the mineral grains at a rate of ~5.4 Gy/min. KF grains were stimulated with 130 mW/cm2 of infrared (IR) light (~800 nm), and luminescence emissions (violet, ~400 nm) were detected using a 9235QA photomultiplier tube (PMT) fitted with Schott BG-39 and Corning 7-59 optical filters. The BG-39 filter absorbs scattered IR light from the stimulating beam, while the 7-59 filter absorbs yellow-green light (~570 nm) emitted from plagioclase feldspars. Optical dating of feldspars using IR-stimulated luminescence is commonly called ‘IRSL dating’. Quartz grains were stimulated by 45 mW/cm2 blue

27 (~470 nm) light, and ultraviolet (~350 nm) emissions were detected using the same PMT, but with a 7.5 mm thick Hoya U-340 optical filter to absorb scattered light from the stimulation beam.

Figure 2.2 Example of the radial plot, which is commonly used to visualize the spread in De values from individual aliquots of a sample. Shown are the De values of 24 aliquots (~100 grains each) of KF sample SIDS2 measured during a dose-recovery experiment where the applied laboratory dose was 24.7 Gy. Points that plot farther from the origin are more precise. To read the De (Gy) of an aliquot, for example the aliquot associated with the red data point, a straight line is drawn from the origin, through the point, to where it intersects the curved axis. The uncertainty in this De value is found by sliding the straight (left-side) x-axis along the red line to where it coincides with the data point, and then projecting lines from the origin through its endpoints to the curved axis (e.g., as in Figure 6 of Galbraith et al. 1999). Points that plot within the shaded region are within 2-sigma of the weighted mean. Unlike histograms, radial plots take into account the precision of the individual points enabling the spread in the data to be properly visualized and analyzed (Galbraith 2010).

2.4.3. Estimating the environmental dose rate

To calculate the environmental dose rate, the concentrations of U, Th, Rb, and 40K are required. To this end, a commercial laboratory (Maxxam Analytics, Mississauga, Ontario) was contracted to perform neutron activation analysis (NAA) on representative

28 subsamples of each bulk sample prepared for dating. These analyses indicated that U, Th, and Rb concentrations were lower than what is typically found in natural sediments, and that U concentrations were near or below the laboratory’s detection limit, which is about 0.5 ppm. To corroborate these values, fractions of the same subsamples were sent to the Australian Nuclear Science and Technology Organization (ANSTO, Lucas Heights, NSW, Australia) laboratory where 40K, Th, and Rb concentrations were determined again by NAA, and U concentrations were found by delayed neutron activation analysis (DNAA), a technique that is more suited to measuring low concentrations of this radioisotope.

Pore water in the sediment matrix attenuates radiation, so knowledge of each sample’s water content is required for the dose rate calculations. Water contents were determined by filling perforated 1 cm3 plastic cubes of known mass with sediment which had retained its ‘as-collected’ water content. The filled cubes were weighed, dried in an oven at 40 °C, and then placed in water to allow the sediment to saturate. This technique was developed at SFU about 30 years ago and was adopted by the UFV lab 푤 (O.B. Lian, personal communication,July 2017). The saturated (Δ푠푎푡) and as-collected 푤 (Δ푎푐) water contents were determined using the following relations:

푤 푀푎푐−푀푑푟푦 ∆푎푐= (2) 푀푑푟푦−푀푐

푤 푀푠푎푡−푀푑푟푦 ∆푠푎푡= (3) 푀푑푟푦−푀푐

Where Mac, Mdry and Msat, are the as-collected, dry and saturated masses of the sediment plus container, and Mc is the mass of the empty container. Since all of the samples were collected from well-drained sand, the as-collected water contents were used in each case, with an uncertainty (± 10%) to account for any reasonable variation.

For each sample, environmental dose rates due to  and  radiation originating from outside the grains were determined using standard formulae (e.g., Berger 1988;

Lian et al. 1995) that convert radioisotope concentrations to dose rates 퐷̇ 𝑖, where 𝑖 = 훽, 훾,and take into account water content (4):

29 ̇ 1 퐷𝑖 = (푑𝑖퐾[퐾] + 푑𝑖푇ℎ[푇ℎ] + 푑𝑖푈[푈]) 푤 푤 (4) 1+퐻푖 ∆

Where [Th], and [U] are concentrations of Th and U in µg/g (ppm), respectively and [K], is % of 40K, the d coefficients are the dose rates per unit concentration of K, Th, and U. The H coefficients are the ratios of specific stopping powers, or attenuation, for  or  radiation (Aitken 1985), of water to dry sediment (Table 2.2), and w is the water content. For the  dose rate, attenuation factors (Brennan 2003) appropriate for an HF- acid-etched 200 µm diameter grain (0.0791 for 40K, 0.2112 for Th, and 0.1567 for U) or a 350 µm grain (0.1197 for 40K, 0.2637 for Th, and 0.2051 for U) was applied. Larger grains have larger surface areas and therefore absorb more  radiation from the surroundings. The effect of the two grain size ranges (180-250 and 300-400 µm) on optical ages will be tested in this study.

Table 2.2. Values of d (Guérin et al. 2011) and H coefficients1

풘 i diK diU diTh 푯풊  0.7982 0.1457 0.0277 1.25  0.2491 0.1116 0.0479 1.14 1 d’s have units of Gy·ka-1 per g·g-1 for U and Th; and Gy·ka-1 per 1% K for K. H is dimensionless.

Components of the total dose rate internal to the grain as a result of K and Rb were estimated using the concentrations provided by Huntley and Baril (1997) and Huntley and Hancock (2001), and an internal  dose rate estimate of 0.08 Gy/ka was used (Ollerhead et al. 1994). The contribution of cosmic rays to the total dose rate, 퐷̇ 푐, was estimated using present sample burial depths, a sediment density estimate of 2.5 g/cm3, and the relation of Prescott and Hutton (1994). The total environmental dose rate, 퐷̇ 푇, is therefore:

퐷̇ 푇 = 퐷̇ 훽 + 퐷̇ 훾 + 퐷̇ 훼 + 퐷̇ 퐶 (5)

2.4.4. Anomalous fading

Anomalous fading is a malign effect that was first observed in the thermoluminescence signals of feldspar by Wintle (1973); it appears to be negligible or absent in quartz. It refers to the loss of electrons from traps that are thermally stable at

30 ambient temperatures over geologic time. The accepted mechanism for anomalous fading is quantum-mechanical tunnelling of the trapped electrons to nearby recombination centres (Aitken 1985; Huntley and Lamothe 2001). Anomalous fading of the IRSL signal typically measured for dating has been found to occur in most feldspars (Huntley and Lamothe 2001; Huntley and Lian 2006). The rate at which anomalous fading occurs in a sample depends on the nature of the trap, and it can be a function of laboratory dose (Huntley and Lian 2006), stimulation temperature, the part of the luminescence decay signal that is sampled (Thomsen et al. 2008), the intensity of the luminescence signal (Lamothe et al. 2012), and the calcium content in plagioclase feldspars (Huntley et al. 2007). For samples for which the intensity of the natural signal falls within the linear part of its dose response curve (typically samples with optical ages up to 20–50 ka), optical ages can be corrected for anomalous fading. This involves determine a sample’s fading rate, g, which is the percent of the original signal that is lost per decade of time (%/decade), a ‘decade’ being a factor of 10 in time since irradiation. Huntley and Lamothe (2001) proposed two methods of doing this, each involving irradiating, preheating, and measuring several aliquots of the sample at various times over days, weeks, and months. Auclair et al. (2003) proposed an alternative method of determining fading rate, which involves using a modified SAR protocol. This method allows fading rates to be determined for several individual aliquots over the course of only a few days, and it is therefore much more time-efficient. The method of Auclair et al. (2003) was used to determine fading rates for this research. Fading corrections were applied using the model of Huntley and Lamothe (2001).

2.5. Grain size analysis

Grain size analysis was performed on each sample collected for optical dating in order to determine the abundance of the preferred grain size range (180-250 µm) used for optical dating, and to use it as a proxy for palaeowind speed. Grain size analysis was also performed in order to better understand the environmental conditions under which the dunes formed. Samples were analysed in distilled water solution using a Malvern Mastersizer 2000 with obscuration values around 20. Two samples (SIDS1C and SIDS8) were also analyzed using dry sieves and a shaker table, as well as with wet sieves after chemical treatments for optical dating preparation to provide a check on the

31 Mastersizer results. Grain size fractions separated with dry sieves and a shaker table were >710, 710-500, 500-425, 425-355, 355-300, 300-250 and <250 µm (Figure 2.3, Table E1). Grain size fractions separated with wet sieving after chemical treatments were >700, 700-600, 600-500, 500-400, 400-300, 300-350, 350-180, and <180 µm (Figure 2.3, Table E2). The results using the three methods were in agreement. Coarse grains (≥ 2000 µm diameter) constituted less than 0.2 % of 4 samples, the remaining samples had no grains over 1500 µm. Data acquired using the Malvern Mastersizer 2000 were used for further analysis (Figure 2.3) (see Appendix D for raw data).

Figure 2.3 Comparison of grain size distributions determined by different methods. The dotted lines show a moving average of the points, which makes the points at the top of the graph appear as if they were not included, which is not the case.

2.5.1. Palaeowind speed calculation

Grain size data were used to calculate the wind speed necessary for the grains to be transported by saltation during dune activity. The wind speed needed for the initiation of grain movement, and thus dunes to develop, can be determined using Bagnold’s formula for shear velocity (equation 6) (Bagnold 1941) and equation (7) to convert between 푢∗푡 and 푢10 (Hsu 1974), based on D50 values. The threshold friction velocity (푢∗푡) can be estimated using:

32 푝푠𝑔푑 푢∗푡 = 퐴√ (6) 푝푎

Where A is a coefficient equal to 0.1 in air for any for grain ≥ .25 mm in diameter

3 (Pye 1983), Ps is the density of quartz (2650 kg/m ), g is the acceleration due to gravity

2 3 (9.81 m/sec ), Pa is the density of air at 15°C at sea level (1.225 kg/m ), and d is the mean grain size of the sand (D50 in m). According to Hsu (1974), the average hourly wind velocity (u) in the constant direction of transport at an anemometer height at 10 m can be determined if D50 ≥ 0.00025 m. This is done by substituting the threshold friction velocity from equation 6 (푢∗푡) into the shear (or friction velocity) variable (U∗) in the following equation:

U∗ = 0.037 U10푚 (7)

Sample wind speed calculation using the above equations are shown in Table 2.3.

33 Table 2.3 Sample calculation of palaeowind speed using D50 grain size from SIDS1C. A threshold friction velocity (풖∗풕) is found using Bagnold’s (1941) equation (6) and then used in Hsu’s (1974) equation (7) to determine a palaeowind velocity at 10 m above the surface.

Sample SIDS1C

−4 D50 (m) 5.6E 푚

푘𝑔 푚 2650 × 9.81 × 5.6E10−4푚 푚3 푠2 푚 풑풔품풅 0.1√ = .3447 풖 풕 = 푨√ 푘𝑔 푠 ∗ 풑 1.225 풂 푚3

U 퐔 = ퟎ. ퟎퟑퟕ 퐔 ∗ = U ∗ ퟏퟎ풎 0.037 10푚

. 3447 m = 9.32 m/s 0.037

푚 60푠 60 푚𝑖푛 1 푘푚 푘푚 Unit conversion to km/hr 9.32 × × × = 33.5 푠 1 푚𝑖푛 1 ℎ푟 1000푚 ℎ푟

Palaeowind speed in constant direction of transport 10 m 34 km/ hr above the surface for sample SIDS1C

34 Chapter 3. Results

3.1. Introduction

This chapter presents new radiocarbon ages from the palaeosols and the results of the experiments performed to obtain reliable optical age estimates for the stabilization of sand dunes on Savary Island. Both KF and quartz grains were analyzed using a modified version of the standard single-aliquot regenerative dose (SAR) protocol and a post-infrared infrared-stimulated luminescence (post-IRIR) SAR protocol for KF.

3.2. Radiocarbon ages

Radiocarbon ages of charcoal fragments collected during this study from eroded B-horizons of palaeosols 1 and 2 at site 2 (see section 4.4.2. for stratigraphy and description of palaeosols) are in stratigraphic order, and are consistent with the ages of the charcoal fragments collected in 2000 from the same stratigraphic positions (Clague, unpublished data), if their analytical uncertainties are taken into account at 2(Table 3.1). The radiocarbon age of charcoal collected from the B-horizon of the incipient palaeosol at site 1 is about ~1000 to 1500 years older than the age of the charcoal associated with palaeosol 1 at site 2. It should be noted that the incipient palaeosol was not traced to the positions where palaeosols 1 and 2 were exposed, but based on degree of soil development, elevation above sea level, and their orientation along the arm of a dune, they are thought to be distinct soil horizons. Charcoal fragments at all sites are interpreted to be detrital in origin, having been reworked by wind from a different location. If this interpretation is correct, the ages from charcoal provide maximum ages for soil development. However, it is possible that the charcoal fragments had been transported to their current positions from soil organic horizons by bioturbation, and in this case their ages would more closely date the palaeosols.

35 Table 3.1 Radiocarbon ages from charcoal fragments collected from B- horizons of palaeosols at sites 1 and 2.

Site Lab No. Date 14C Age, yr Age range (cal Collected BP yr BP) 1

Site 1 (incipient UCIAMS1599762 6/21/2015 8245 ± 20 9290 – 9130 palaeosol)

UCIAMS1495872 01/07/2014 7400 ± 25 8320 – 8180 Site 2 (palaeosol 1) TO-91552 11/26/2000 7180 ± 120 8300 – 7740

UCIAMS1495882 02/07/2014 8550 ± 40 9560 – 9480 Site 2 (palaeosol 2) TO-91542 11/26/2000 8730 ± 120 10160 – 9530

1Calibrated using Oxcal 4.2 and the IntCal 14 data set. Age ranges are 2. 2 Accelerator mass spectrometry (AMS) radiocarbon age. TO – Isotrace (University of Toronto); UCIAMS - W.M. Keck Carbon Cycle Accelerator Mass Spectrometer Facility (University of California, Irvine). The UCIAMS samples were collected during this study, while the TO samples were collected and submitted for dating by J.J. Clague in 2000.

3.3. Landforms and their interpretation

Lidar data were used to create a ‘bare earth’ map of Savary Island (Figure 3.1). This map was used to locate sites for this research within the dunes, particularly dune heads, to be sampled for optical dating (Figure 3.2). Through digital analysis and field verification, previously unidentified geomorphic features were mapped (Figure 3.3).

Figure 3.2

Figure 3.1 ‘Bare Earth’ digital elevation model of Savary Island created in GlobalMapper16. The focus of this study was the large parabolic sand dunes near the centre of the island, which are shown at a larger scale in Figure 3.2.

Figure 3.2. Locations of study sites on Savary Island. Detailed site descriptions and coordinates can be found in Appendix A

Figure 3.3 Identified, ground-truthed, and interpreted geomorphic features on Savary Island.

The centre of the island is characterized by a 1 km2 area of large stabilized parabolic dunes, the focus of this study (Figure 3.3A). They reach lengths of up to 1000 m and vary in width from 90 to 180 m. Dune heads rise above the adjacent topography between 5.5 and 12.5 m in height. Samples were collected for optical dating on the heads of specific dunes (Figure 3.2). The dunes consist of predominantly coarse sand and are covered by mature coastal Douglas fir (Pseudotsuga menziesii) forest. East of the parabolic dunes is a relatively planar undifferentiated sand surface (Figure 3.3B) that slopes from ~22.5 m in elevation westward to ~12 m above sea level (asl). This surface has landforms that appear to be remnants of dunes (Figure 3.3C). They vary in length and width and typically protrude from the surrounding landscape by 4 to 5 m. The west side of the planar surface promptly ends with an old gravel-topped air strip that extends 700 m in a southeast to northwest orientation, the same orientation as the long axis of the dunes. Remnant dune features do not appear to exist west of the field of parabolic dunes.

Four raised terraces (Figure 3.3D) cover 0.7 km2 on the northeast end of the island, are oriented to the northeast, and dip slightly as they extend to the southwest (Figure 3.4 B). The lowermost terrace (T1) is ~750 m long and has an average elevation of 6 m above sea level (asl) in the northeast and 5 m asl on the southwest. Ocean-front homes line the base of this terrace, one of the first areas of the island inhabited by non- First Nations peoples. The average elevation of the second terrace (T2) ranges from 13 m asl in the northeast to 12 m asl in the southwest. This terrace has a weak crescentic shape and is ~550 m in length. The third terrace (T3) gradually decreases in elevation from 21 m asl in the northeast to 20 m asl in the southwest and is ~700 m long. The highest terrace (T4) is the smallest in area (~450 m in length and has an average elevation ~30 m asl). These terrace features are interpreted to be wave-cut terraces representative of palaeobeaches, formed when RSL was higher than today followed by subsequently lowering, although variation in horizontal elevation suggests there could be another origin. The average elevations of the terraces on the northeastern part of the island correspond to those of erosional, crescentic scarp-like features at the western end of the island (Figure 3.3 D and F; Figure 3.5 A). The smaller and weaker crescentic terraces (Figure 3.3F) notch the landscape west of the dunes (Figure 3.3F); their elevations are mapped in Figure 3.5A. Similarly to terraces on the east (Figure 3.3D),

these features are interpreted as wave-cut paleobeaches, although weaker in landscape expression. They likely formed during a period of RSL fall; the alternative, they formed during a period of RSL rise would have destroyed the lower features.

More features interpreted to have formed when RSL was higher than today include the crescentic, raised platform facing north just west of the parabolic dunes (Figure 3.4E). This platform is interpreted to be a raised, embayed, wave-cut palaeobeach. It slopes from 22 m asl at the north side on a gradient of 2° south- southwest to 30 m asl. As this feature is higher in elevation than modern sea level, it formed when RSL was higher than present-day. The sloping difference in elevation is likely a product of continual beach formation during RSL lowering.

The landform interpreted to be the youngest on Savary Island is at the northwest point and is a referred to locally as Indian Point. This landform (Figure 3.3G) tapers from 600 m wide at the south end to 150 m wide to the north over a distance of 580 m. The southern edge of this feature is 5 m asl and lies at the base of a steep 30° slope. It slopes down to the north to 4 m asl at its most northern point. Small ridges on this landform can reach up 9 m asl, which may be a result of construction of roads and buildings or, alternatively, the presence of remnant foredunes or beach ridges. This landform is interpreted to be a prograding series of beach ridges and foredunes forming a spit (Figure 3.3G). Beach ridges and foredunes form by different processes, a more detailed field analysis of the site is required to conform their origin. This spit is believed to form from sediment eroded from the south side of the island and transported westward by longshore drift and then accreted at Indian Point (Bornhold and Conway 1996). Currently slow accretion rates allow this feature to grow; however, growth is punctuated by periods of rapid erosion, commonly associated with storms (Bornhold and Conway 1996).

Figure 3.4 A) Bare-Earth DEM produced from lidar data viewed in GlobalMapper16. Eye symbols indicate viewpoints in Figure 3.5A, in Figure 3.5B and in B of this figure B) Raised wave-cut terraces, profiled in (C) along A – A’. Terraces were defined based on visual evidence and inflection points.

Figure 3.5 Viewpoint of (A) and (B) show in in Figure 3.4 A) Crescentic wave cut terraces interpreted to be palaeobeaches. B) Raised, crescentic depression, likely a former beach. Circular depression indicates location of Indian Springs, one of the shallow perched aquifers on the island (Tupper 1996).

3.4. Optical dating

3.4.1. Savary Island quartz

The results from LM-OSL measurement on Savary Island quartz showed that it has a dim fast component and a relatively strong medium component. In Figure 3.6 LM- OSL curves for Savary Island quartz are compared with those from quartz collected from Quaternary deposits in the Peace River region of northeast BC near Dawson Creek that

have provided optical ages that are thought to be reliable (Hickin et al. 2015, 2016; Huntley et al. 2017); both samples show a distinct and relatively intense fast component, while the two Savary Island quartz samples do not. To further test the suitability of Savary Island quartz for optical dating, a sample (SIDS2, 180-250 µm quartz) was run using a standard SAR protocol (Figure 3.8); of the 24 aliquots analyzed, only three failed the recycling ratio test, but 21 failed the recuperation test (Figure 3.7). An example of the dose response curve created for one of the aliquots that passed both the recycling ratio and dose recovery tests (aliquot #29) is shown in Figure 3.8A. Because of these results, Savary Island quartz was deemed unsuitable for optical dating and all subsequent experiments focussed on KF.

Figure 3.6 LM-OSL curves of quartz from two different regions in BC. Savary Island quartz (sample SIDS2) has a dim fast component and is dominated by slow and/or medium components. In contrast the fast component of quartz from the Peace River region is relatively bright.

Figure 3.7. Graph showing recycling ratios and recuperation values of quartz aliquots measured for sample SIDS2 (180-250 µm). For aliquots to be suitable for optical dating they must have recycling ratios within 10 % of unity, and recuperation values should be ≤ 5% (grey band on graph). Black data points show the results of the recycling ratio test, of which only two aliquots failed (labelled on graph); however, only three out of 24 aliquots passed the recuperation test. Only three of 24 quartz aliquots passed both recycling ratio and recuperation tests (red points). Error bars are ± 1σ.

Figure 3.8 (A) Dose response curve for aliquot #39 of sample SIDS2 quartz (180-250 µm), which is typical of the curves for this sample. Luminescence measured over the first 0.4 s, minus that measured over the last 20 seconds, was used to determine the dose-response points. This aliquot had a recycling ratio of 1.03 ± 0.07 and recuperation value of 0.3 ± 2.1 %. B) Luminescence decay curves for the two accepted aliquots of Savary Island quartz showing a dim fast component (the abrupt decay seen over the first few seconds) that is probably affected significantly by the prominent medium component (the subsequent, more slowly decaying part of the signal).

3.4.2. Developing a SAR protocol for Savary Island potassium feldspar

Initial experiments for developing an optical dating protocol for Savary Island potassium feldspar (KF) were based on recent studies done by Neudorf et al. (2015) for the same mineral on Calvert Island (~275 km to the northwest) and is referred to as protocol 1 in this study (Figure 3.9). Protocol 1 uses a low preheat temperature and duration (160 °C for 10 s). This protocol is different from that used in most other regions/studies, in which SAR sequences employ higher preheat temperatures and longer durations, typically 200 to 250 °C for 60 s (i.e. Murray and Wintle 2003). A preheat plateau test was therefore performed to determine the suitability of 160°C/10 s preheat for Savary Island KF. Two samples (SIDS1C and 2) were also dated using the 250 °C / 60 s preheat combination, and also using the post-IRIR SAR method (Figure 3.9).

Step SAR protocol 1 SAR protocol 2 Post-IRIR protocol 3 Natural/ Regenerative dose, Natural/ Regenerative dose, 1 Natural/ Regenerative dose, Di Di Di 2 Preheat (160°C, 10s) Preheat (250°C, 60s) Preheat (180°C, 10s)

3 IRSL (50°C, 200 s)  Ln, Li IRSL (50°C, 200 s)  Ln, Li IRSL (50°C, 100 s)  IRSL (150°C, 200 s)  L , 4 n Li 5 Test dose (Beta 10 s) Test dose (Beta 10 s) Test dose (Beta 10 s) 6 Preheat (160°C, 10 s) Preheat (250°C, 60 s) Preheat (180°C, 10s)

7 IRSL (50°C, 200s)  Tn, Ti IRSL (50°C, 200s)  Tn, Ti IRSL (50°C, 100 s) IRSL (150°C, 200 s)  T , 8 n Ti 9 Hotwash (180°C, 40 s) Hotwash (200°C, 40 s) 10 Return to step 1. Return to step 1. Return to step 1.

Figure 3.9 Flowchart showing optical dating protocols applied to KF grains in this study include: protocol 1, a SAR protocol developed and tested by Neudorf et al. (2015) with a relatively low preheat temperature and duration and a hotwash treatment; protocol 2, a variation of protocol 1, but with a higher preheat temperature and duration based on that of Auclair et al (2003), and no hotwash; and protocol 3 a post-IRIR protocol based on that used by Colarossi et al. (2015) and Thomsen et al. (2008). Each protocol was applied to 180-250 µm KF grains. Protocol 1 was also applied to 180-150 µm quartz grains and 300-400 µm KF grains. Ln and Li are the luminescence signals measured from the ‘natural’ dose and after each regenerative dose (i = 1 to 5), respectively, and Tn and Ti are those measured after the test dose.

Preheat plateau test

The preheat plateau test shows the capability of a given SAR protocol to accurately recover a given dose at different preheat temperatures. The closer the ratio of given dose to applied dose is to one, the better the SAR protocol performed at the particular preheat temperature. A laboratory dose of 18 Gy was measured using preheat temperatures of 120, 140, 160, 180, 200, 240, and 260 °C for 10 s. The ratios of the measured dose to the given dose are in agreement with those of Neudorf et al.’s

(2015) SAR protocol and show that 160 °C is likely the most suitable preheat temperature for Savary Island KF (Figure 3.10A).

Recycling ratio and recuperation dose test

The recycling ratio assesses the ability of the test dose to correct for any changes in luminescence sensitivity, while the recuperation test assesses the effect of thermal transfer on the luminescence signal (Chapter 2.4.2) (Figure 3.10A inset graph). For samples analyzed using protocol 1, 100 % of 10 samples and between 93 and 96% of 3 samples passed the recycling ratio test. Only one aliquot from one sample did not pass the recuperation test using protocol 1; 93.4 % of aliquots passed both recycling and recuperation tests using the protocol, and recuperation values ranged between 0.4 to 0.7%. For samples analyzed using protocol 2, 100 % of the aliquots passed the recycling and recuperation tests. Values of recuperation for both samples tested using protocol 2 were 1.41 and 1.49 %. Three samples were analyzed using protocol 3, 87 to 96 % of aliquots passed the recycling ratio test. All aliquots passed the recuperation test; recuperation values were between 0.9 and 1%.

Dose recovery test

The dose recovery test helps one determine if the chosen SAR protocol can recover a given (laboratory) dose. Using protocol 1, a dose of 6 Gy (24 aliquots) was successfully recovered from sample SIDS2 (180-250 µm) KF grains. A dose of 24.7 Gy was chosen after a successful recovery of 6 Gy based on an estimate of how old suspected samples might be (~ 10 ka) assuming an environmental dose rate of 2 Gy/ka. The recovered dose was calculated using the CAM and gave dose-recovery ratios of 1.01 ± .03 and .97 ± .04, for the 6 Gy and 24.7 Gy applied doses, respectively. An example is shown in Figure 3.10B.

Figure 3.10 A) Results from the preheat plateau test on sample SIDS2 (180-250 µm) KF. Measurements were recorded at 50°C at 90% optical power, and the cut-heat was 20 °C higher than the preheat temperature being tested. This shows that the protocol accurately recovers a known dose (18 Gy) using preheat temperatures of 160, 180, and 200 °C for 10 s each. Aliquots outlined in red were given the preheat temperature of 160°C for 10 s, the temperature used in the majority of dating experiments in this study (protocol 1). Inset graph shows the results of the recycling ratio and recuperation tests for all aliquots of this sample during the preheat plateau test. All aliquots passed recycling and recuperation tests. B) Dose-recovery test performed on 24 aliquots of SIDS2 depicted as a radial plot of De values. Grains were bleached in direct sunlight for 3 hours and given a laboratory dose of 24.7 Gy; this dose was then recovered successfully using protocol 1. Inset graph shows a typical dose- response curve (normalized luminescence signal, Li/Ti, as a function of laboratory dose) from one of the SIDS2 KF aliquots. These results suggest that SAR protocol 1 is suitable for dating KF grains on Savary Island.

According to preheat plateau and dose recovery tests, protocol 1, developed and tested by Neudorf et al. (2015), is expected to provide reliable optical ages for the Savary Island KF samples and was therefore used in subsequent experiments.

3.4.3. Correction for anomalous fading

G-values were determined using the method of Auclair et al (2003). Each optical age was corrected for anomalous fading using its own fading rate (g-value) where possible and the method described by Huntley and Lamothe (2001). Protocol 1 g-values

range from 4.9 ± 0.1 % to 7.5 ± 0.1 % per decade (Table 3.2) and are typical of those found in other studies along the BC coast (Wolfe et al. 2008; Neudorf et al. 2015, 2017). Only one g value was calculated from the two samples tested using protocol 2, 6.1 ± 0.2 % per decade (uncertainties are at 1σ); it is similar to those determined using protocol 1. The fading rates found using protocol 3 (the post-IRIR protocol) were 2.3 ± 0.2, 4.0 ± 0.4, and 5.6 ± 2 % per decade (Table 3.2), and are generally lower than those found using protocols 1 and 2, as expected (Thomsen et al. 2008). The fading rate calculated for SIDS5 using protocol 3 was 5.59 ± 1.75 %/decade which is uncharacteristically high for a post-IRIR signal. Indeed, it is consistent with that determined using protocol 1. The advantage of using a post-IRIR sequence is that a part of the signal is isolated that has a negligible fading rate, thus removing the need to correct for it. This is not the case for SIDS5, nor were the fading rates negligible for other samples (SIDS1C, SIDS2, and SIDS6B), which all had to be measured and corrected for.

3.4.4. Tests of the laboratory lighting

To test whether the filtered laboratory room lighting affected the samples during their preparation, an additional sample SIDS1C (collected adjacent to SIDS1) was prepared under room lights that were visually estimated to be about half as intense as normal. The results (Figure 3.11) suggest a slightly lower De value is obtained under dimmer room lights, but it is consistent with that determined under normal conditions at only 1. Lower laboratory (i.e. dimmed) lighting conditions were therefore maintained when preparing all subsequent samples.

Figure 3.11 (A) Equivalent dose distribution of samples prepared under regular and dimmed room lighting conditions (see text), to test for partial bleaching during sample preparation. The dashed lines show the weighted mean De values found using the CAM; points that fall within the grey bands are within 2 of the mean. (B) The results indicate no significant differences. Analytical uncertainties are ± 1; the fading-corrected age error is 1. n – number of aliquots accepted, De – equivalent dose in Gy found using CAM.

3.4.5. Tests for grain size dependence on optical ages

Grains of different sizes may have different transport histories and may even differ in mineral composition. To see if grain size has an effect on optical age, KF in two different grain size ranges (180-250 µm and 300-400 µm) was prepared from sample

SIDS1C and dated using protocol 1 (Figure 3.12). Although the resulting De values are significantly different, variation in the total dose rate (see section 2.4.3), and anomalous fading rate, results in calculated optical ages that are consistent (Figure 3.12; Table 3.2).

Figure 3.12. (A) Radial plot showing the De distributions resulting from two grain size ranges, obtained using SAR protocol 1. The dashed lines show the weighted mean De values (in Gy) using the CAM. (B) Calculated ages show no statistical difference even after correction for anomalous fading. SIS1C 180-250 um is also the SID1C dimmed sample from Figure 3.11. Analytical uncertainties are ± 1, n – number of aliquots accepted, De – Equivalent dose in Gy found using CAM.

3.4.6. Optical ages

The fading-corrected optical ages for each sample and protocol are shown in Table 3.2. Ages obtained using protocol 1 have an average OD of 16.5 ± 0.5 %, with a range between 10 and 22 %; these are typical values for well bleached samples (cf. Arnold and Roberts 2009). All of the dune head samples were dated using protocol 1. These samples (SIDS5, SIDS6B, SIDS7, SIDS9, SIDS10, and SIDS11) ranged in age from 4.88 ± 0.33 ka (SIDS10) to 6.29 ± 0.40 ka (SIDS7) (Table 3.2), with a weighed mean age of 5.47 ± 0.15 ka.

Ages obtained using protocol 2 gave results that were the same as those found using protocol 1, after correction for anomalous fading. The fading rates measured using protocols 1 and 2 for sample SIDS1C are 7.49 ± 0.10 and 6.13 ± 0.17 %/decade, respectively. Overdispersion rates found using protocol 2 were lower than all those found using protocol 1, but only two samples were tested using protocol 1 so generalizations should not be made.

Optical ages obtained with protocol 3 are slightly older than those obtained using protocols 1 and 2, but are within error (Table 3.2). Four samples were run using this protocol, with fading-corrected ages of 7.44 ± 0.71 ka (SIDS1C), 8.00 ± 0.69 ka (SIDS2), 8.32 ± 0.79 ka (SIDS5), and 5.89 ± 0.40 ka (SIDS6B). Sample SIDS1C gives the same age using protocol 3, within error, as using protocol 1; however it is slightly older than ages obtained using protocol 2. Sample SIDS 2 gave the same age using protocol 3, within error, as using protocols 1 and 2. Sample SIDS6B yielded the same age, within error, using protocol 3 as it did using protocol 1 (Table 3.2). Overall more aliquots measured using protocol 3 were rejected for recycling or recuperation violations than using the other protocols. A large discrepancy exists between the fading-corrected age calculated for SIDS5 using protocol 1 (5.76 ± 0.36 ka) and that calculated using protocol 3 (8.32 ± 0.79 ka), a difference of about 2.6 ka. This difference may result from the inclusion of grains with inherited signals, as the post-IRIR signal is considerably harder to bleach.

Table 3.2 Optical dating samples, total dose rates, equivalent dose (De), overdispersion (OD) values, fading rates (g), and optical ages (uncertainties are ±1). Grey shading indicates results using different SAR protocols on the same samples. Dashed line separates samples from sites 1 and 2 (above dashed line) from the dune head samples (below dashed line).

Fading- Total dose rate Uncorrected age Sample P1 N2 D (Gy) OD (%) g (%/decade) corrected age (Gy/ka) e (ka) (ka)

SIDS1 1 25/25 2.20 ± 0.15 7.47 ± 0.30 19.4 ± 2.9 6.20 ± 0.36 3.40 ± 0.27 5.75 ± 0.42 SIDS1C3 1 36/36 2.20 ± 0.15 6.62 ± 0.23 20.3 ± 2.5 7.49 ± 0.10 3.02 ± 0.23 6.38 ± 0.45 SIDS1C4 1 30/30 2.75 ± 0.17 10.00 ± 0.36 19.5 ± 2.6 5.51 ± 0.11 3.64 ± 0.26 6.13 ± 0.42 SIDS1C 2 24/24 2.20 ± 0.15 7.11 ± 0.32 15.6 ± 3.4 6.13 ± 0.17 3.30 ± 0.25 5.93 ± 0.44 SIDS1C 3 21/24 2.20 ± 0.15 11.29 ± 0.63 24.5 ± 4.1 3.99 ± 0.40 5.15 ± 0.45 7.44 ± 0.71 SIDS2 1 22/24 1.91 ± 0.13 7.32 ± 0.32 21.2 ± 3.1 6.92 ± 0.10 3.83 ± 0.41 7.69 ± 0.71 SIDS2 2 27/27 1.91 ± 0.13 8.18 ± 0.24 14.3 ± 2.1 6.13 ± 0.17(5) 4.29 ± 0.43 7.81 ± 0.67 SIDS2 3 23/24 1.91 ± 0.13 12.50 ± 0.40 40.8 ± 6.0 2.31 ± 0.16 6.50 ± 0.66 8.00 ± 0.69 SIDS3 1 27/27 2.21 ± 0.15 6.63 ± 0.14 10.4 ± 1.5 4.99 ± 0.09 3.01 ± 0.21 4.76 ± 0.30 SIDS4 1 25/25 2.29 ± 0.15 6.26 ± 0.37 16.6 ± 2.4 4.99 ± 0.09(6) 2.74 ± 0.21 4.32 ± 0.29 SIDS8 1 30/30 2.09 ± 0.14 7.86 ± 0.27 18.8 ± 2.5 5.99 ± 0.10(7) 3.76 ± 0.28 6.70 ± 0.45

SIDS5 1 28/30 2.38 ± 0.16 8.17 ± 0.20 12.9 ± 1.7 5.60 ± 0.09 3.39 ± 0.24 5.76 ± 0.36 SIDS5 3 21/24 2.38 ± 0.16 11.57 ± 0.52 20.0 ± 3.3 5.60 ± 0.40 4.85 ± 0.40 8.32 ± 0.79 SIDS6B 1 30/30 2.14 ± 0.15 7.11 ± 0.16 11.6 ± 1.6 5.83 ± 0.10 3.32 ± 0.24 5.78 ± 0.36 SIDS6B 3 23/24 2.14 ± 0.15 9.59 ± 0.28 13.3 ± 2.2 3.07 ± 0.18 4.46 ± 0.33 5.89 ± 0.40

SIDS7 1 35/35 2.14 ± 0.14 7.86 ± 0.22 16.3±2.0 5.66 ± 0.10 3.67 ± 0.26 6.29 ± 0.40 SIDS9 1 28/30 2.27 ± 0.15 7.15 ± 0.23 17.6±2.3 5.47 ± 0.09 3.09 ± 0.31 5.22 ± 0.34 SIDS10 1 36/36 2.35 ± 0.19 7.06± 0.18 14.6±2.9 5.22 ± 0.08 3.00 ± 0.26 4.88 ± 0.33 SIDS11 1 30/30 2.52 ± 0.18 8.37 ± 0.22 14.1±1.9 4.89 ± 0.10 3.31 ± 0.26 5.20 ± 0.34

1 Protocol details outlined in Figure 3.9. 1 = SAR with a 160°C, 10s preheat, 2= SAR with a 250°C, 60s preheat, 3 = Post-IRIR. 2 Number of aliquots accepted/measured. 3 Duplicate sample collected from same pit as SIDS1, prepared under dimmed laboratory lighting conditions.. 4 300-400 µm grain size used. 5 Using fading rate of SIDS1C protocol 2. 6 Using the fading rate of SIDS3, located directly below SIDS4. 7 Using an average fading rate of all samples collected from section face(SIDS1, SIDS1C, SIDS2, SIDS3, SIDS4).

3.5. Grain size and wind speed

3.5.1. Grain size

Histograms of grain size distribution of each sample can be seen in Figure 3.13, 3.14 and 3.15. The analyzed samples are well sorted to moderately well sorted (Folk 1966), with an average sorting coefficient of 0.55 phi (Table 3.3). Most beach sands have sorting values between 0.25 to 0.30 phi, and associated dune sands are typically slightly better sorted (Folk 1980). Measures of skewness (asymmetry) are close to one, showing approximately symmetrical curves. Those that have negative skewness values display a tail of coarse grains, whereas positively skewed samples have a tail of finer grains (Folk 1966). Samples SIDS-5 and SIDS-11 had the highest skewness values (Table 3.3), reflected in distributions with small increases in grain percentages between 45 to 150 µm (Figure 3.14 and Figure 3.15 respectively). The kurtosis, or peakedness, of the samples is predominantly mesokurtic (eight samples); three samples are platykurtic with an average value of 0.93 ± 0.03, slightly lower than values for coastal dunes studied in Mexico by Kasper-Zubillaga and Carranza-Edwards (2005). Sediments dominantly from one source will show relatively normal distributed curves, whereas sediments from multiple source areas will have a more bimodal distribution and extreme skewness and kurtosis values (Folk 1980). Dune samples typically have positively skewed mesokurtic distributions (Folk 1980), and the best sorting attained naturally in sediment is between 0.20 and 0.25 (Folk 1968). Cumulative grain size distribution of all samples is compared in Figure 3.16.

The average D50 value of all the samples was 467 µm (Table 3.4). The surface weighted mean and volume weighted mean were close in range. This highlights the narrow grain size range. Particle size distribution is represented in phi units (Figure 3.17) as a function of volume (%). The average grain size peaks for all samples is around 1.2 phi.

Figure 3.13 Grain-size distribution histograms of optical dating samples SIDS1 to 4.

Figure 3.14 Grain-size distribution histograms of optical dating samples SIDS5 to 8.

Figure 3.15 Grain-size distribution histograms of optical dating samples SIDS9 to 11.

Table 3.3 Grain size statistics in phi (φ) units, after Folk and Ward (1957). Percentiles for verbal classifications are presented in Folk (1968). Sorting values from 0.35 to 0.5 are well sorted, 0.50 to 0.70 are moderately well sorted, and 0.70 to 1.00 are moderately sorted. Skewness values between 0.1 and -0.1 are approximately symmetrical. Kurtosis values equal to 1 indicate normal distributions; 0.67to 0.90 indicate platykurtic distributions; 0.90 to 1.11 indicate mesokurtic distributions; and > 1.11 indicate leptokurtic distributions.

Verbal Verbal Sample Sorting Verbal Sorting Skewness Skewness Kurtosis Kurtosis Moderately well Approximately SIDS1 0.56 sorted 0.04 symmetrical 0.96 Mesokurtic Moderately well Approximately SIDS2 0.60 sorted -0.09 symmetrical 0.92 Mesokurtic Approximately SIDS3 0.49 Well sorted 0.01 symmetrical 0.87 Platykurtic Moderately well Approximately SIDS4 0.59 sorted 0.03 symmetrical 0.97 Mesokurtic Moderately well Approximately SIDS5 0.63 sorted 0.06 symmetrical 0.96 Mesokurtic Moderately well Approximately SIDS6 0.51 sorted -0.09 symmetrical 0.88 Platykurtic Moderately well Approximately SIDS7 0.53 sorted 0.09 symmetrical 0.92 Mesokurtic Moderately well Approximately SIDS8 0.55 sorted -0.09 symmetrical 0.96 Mesokurtic Moderately well Approximately SIDS9 0.51 sorted -0.09 symmetrical 0.86 Platykurtic Moderately well Approximately SIDS10 0.52 sorted -0.09 symmetrical 0.93 Mesokurtic Moderately well Approximately SIDS11 0.62 sorted 0.07 symmetrical 0.98 Mesokurtic Average 0.56 ± 0.05 -0.01 ± 0.08 0.93 ± 0.04

Table 3.4 Grain-size data calculated using the Malvern Mastersizer 2000. The surface weighted mean or the surface moment mean diameter is the size of the particles that contribute to the majority of the total surface area. The volume weighted mean constitutes the size of the particles that contribute to the majority of the bulk sample volume, also known as the mass moment mean diameter (Seville and Wu 2016).

Surface weighted Volume weighted 1 2 Sample Location mean (µm) mean (µm) D50(µm) D50(φ) SIDS1C Bluff 522 606 557 0.84 SIDS2 Bluff 363 430 391 1.36 SIDS3 Bluff 449 504 473 1.08 SIDS4 Bluff 513 602 546 0.87 SIDS5 Dune Head 464 586 532 0.91 SIDS6 Dune Head 383 430 406 1.30 SIDS- Dune Head 487 554 516 0.96 SIDS8 Bluff 353 403 375 1.41 SIDS9 Dune Head 395 445 419 1.25 SIDS10 Dune Head 364 409 386 1.38 SIDS11 Dune Head 450 588 539 0.89 Average 431 505 467 1.11 1 The mid-point of the distribution. 2 Phi (φ) is a dimensionless unit of grain size measurement. Φ = -log2 D/D0 where D is the diameter of the particle in mm and D0 is the reference diameter equal to 1 mm.

Figure 3.16 Cumulative percent grain size curves for of all samples. Bracket indicates the range in D50 values of all the samples, from ~400-600 µm.

Figure 3.17. Grain size distributions calculated using the Malvern Mastersizer 2000. This graph compares grain size distributions in phi units, showing that most samples are predominantly composed of grain size around 1.0 phi. Two samples, SIDS5 and SIDS11 have a larger portion of small grains indicated by the small bump in the distribution curves around 4.2 phi.

3.5.2. Palaeowind speed for aeolian sand transport

Based on D50 grain size values, the wind speed required for aeolian activity during the later stages of dunes activity was 31 km/hr. These calculations represent a palaeo-wind threshold value that is probably close to a maximum monthly average wind speed (IJ Walker, pers comm.). When modern wind speeds are grouped by seasons, we can see that stronger winds occur during the winter months and are from the southeast, with weaker summer winds coming from the northwest (Figure 1.2). The strong winter winds blow in the same direction as winds responsible for parabolic dune development.

Table 3.5. Wind speed calculations from grain size data using D50 values. Sample calculation for SIDS1C shown in Table 2.3.

Sample D (m) u*t (m/s)1 U (m/s)2 km/hr Name 50 10m SIDS1C 0.0006 0.34 9.29 33.5 SIDS2 0.0004 0.29 7.78 28.0 SIDS3 0.0005 0.32 8.56 30.8 SIDS4 0.0005 0.34 9.20 33.1 SIDS5 0.0005 0.34 9.08 32.7 SIDS6 0.0004 0.29 7.93 28.6 SIDS7 0.0005 0.33 8.94 32.2 SIDS8 0.0004 0.28 7.63 27.5 SIDS9 0.0004 0.30 8.06 29.0 SIDS10 0.0004 0.29 7.73 27.8 SIDS11 0.0005 0.34 9.14 32.9 Average 31 ± 2 1 Found using Bagnold’s (1941) equations for threshold friction velocity (Equation 6) 2 Founding using Hsu (1974) equation 5. (Equation 7)

Chapter 4. Postglacial landscape evolution of sand dunes on Savary Island, British Columbia

4.1. Introduction

Geomorphological and palaeoecological studies have shown that the south coast of British Columbia (BC) has experienced pronounced changes in relative sea level and climate since local deglaciation began about 15 ka cal ago (Shugar et al. 2014). Coastal sand dunes are sensitive to changes in sea level, wave regime, sediment supply, climate, and regional vegetation (Davidson-Arnott and Law 1996; Hesp 2002), and, as such, they provide an opportunity to learn more about the response of coastal systems to environmental change (Billy et al. 2015). This is important because they provide habitats for a variety of plant and animal species, and are areas for recreation and tourism. Also, they can have cultural significance, and they help provide protection against coastal erosion (Page et al. 2011). Widely used as archives for environmental change in North America (e.g. Halfen and Johnson 2013; Wolfe et al. 2016) and elsewhere, sand dunes in coastal BC have received little attention, likely because they are rare in this region; most (~80%) of the coastline is rocky (Clague and Bornhold 1980). In coastal BC, active dunes currently exist on the west coast of Vancouver Island, in isolated areas on the central coast (e.g., Calvert Island; Neudorf et al. 2015), and on the eastern side of Haida Gwaii (Wolfe et al. 2008); stabilized dunes exist along the Strait of Georgia at Cape Lazo near Comox (Mathews et al. 1970) and on Savary Island (Page et al. 2011), the latter are the focus of this research.

In coastal environments, parabolic dune formation almost always begins with the development of a foredune in the presence of vegetation. After a blowout occurs, the dune begins to elongate parallel to the direction of the wind (Pye 1983; Hesp 2002, 2011; Hesp and Walker 2013). The newly formed U- or V-shaped lobe advances, commonly into pre-existing vegetation, and long ridges (or arms) extend shoreward and

are anchored by vegetation on the outer edges (Hesp and Walker 2013). Stabilized, or stabilizing, parabolic dunes can indicate a transition to more humid conditions that allow for the establishment of vegetation on the dune surface. A return to drier conditions may result in dunes becoming active again. In an ancient deposit, periods of episodic stabilization may be represented by soil development (palaeosols), whereas episodes of activity are represented by sand accumulation (e.g. Wolfe et al. 2001; Halfen and Johnson 2013). Previous research on Savary Island includes a detailed study of dune ecology (Dunster 2000; Henderson 2003); however, knowledge of dune morphology is limited to relatively subjective interpretations derived from contour maps and ecological field observations (Tupper 1996; Dunster 2000), with no specific measurements. Additional research in the region including Savary Island focuses on the coastal sand ecosystems of BC (Page et al. 2011), coastal processes (Bornhold et al. 1996), and a study of sensitive ecosystems in the Salish Sea region (Harrington et al. 2005).

The key objectives of this research are to: 1) determine the best materials and protocols for the optical dating of aeolian sediments on Savary Island, 2) compare the optical ages to radiocarbon ages on charcoal and explain any discrepancies, and 3) describe the evolution of Savary Island and link periods of dune formation to climate variations and changes in relative sea level. This paper presents the first detailed study of the evolution of the parabolic dunes on Savary Island and its relation to regional environmental change.

4.2. Study area

Savary Island, the southernmost of the Discovery Islands, lies at the north end of the Strait of Georgia in the Georgia Depression between Vancouver Island and the Coast Mountains (Figure 4.1). Much of the island is capped by aeolian sand that rests unconformably on sediments deposited during the last glaciation. Near the centre of the island, the aeolian sediments form large (~1000 m long), stabilized parabolic dunes that cover ~1 km2; in other places, dunes appear to have been eroded or possibly never formed. A reconnaissance investigation of the dunes in 2000 revealed the presence of several buried soils (palaeosols) that contain scattered charcoal fragments in their B

horizons. Some of the charcoal was radiocarbon dated to ~10 ka cal BP, suggesting that the soils are of this approximate age (Clague et al. unpublished data).

Figure 4.1 (A) Location of Savary Island (inside box), approximately 145 km northwest of Vancouver, BC. (B) Satellite image of Savary Island (imagery from Google Earth); it is about 7.5 km long, ranges in width from 0.3 to 1 km, and is completely vegetated. C) ‘Bare-Earth’ lidar image shows parabolic dunes (within box) which cover ~1 km2 in the centre of the island.

Currently, Savary Island lies at the northern end of the Coastal Douglas Fir biogeoclimatic zone in the rain shadow of Vancouver Island. It is dominated by a mature Douglas fir (Pseudotsuga menziesii) forest growing in a well-developed Dystric Brunisol soil. The island experiences warm dry summers and cool moist winters. The mean annual temperature is 9.5 °C and the mean annual precipitation is 1230 mm, of which less than 20 % falls between May and September (Wang et al. 2012). The dominant wind directions are from the northwest in the summer months and from the southeast in the winter months (Figure 1.2) (Thomson 1981). Average monthly wind speeds near Savary Island (10 km to the west) are ~18 km/h. Maximum hourly wind speeds of 90 km/h and a maximum wind gust of 99 km/h were recorded in January of 2000 (Environment Canada Stn # 46131).

Postglacial relative sea level (RSL) change along the coast of BC is complex (e.g. Clague et al. 1982; Mathews et al. 1970; Shugar et al. 2014), mainly reflecting the effects of removal of ice loads after glaciation. Glaciers had retreated from Savary Island by about 14 ka cal BP (Clague and James 2002), and relative sea level (RSL) in the region fell quickly as the land rebounded isostatically (Figure 4.2); it was about 15 m lower than present in the early Holocene. By about 8 ka cal BP, RSL began to rise, and it had reached a position within a few metres of present sea level by about 5 ka cal BP ago (Mathews et al. 1970; Clague et al. 1982; Hutchinson 1992; Shugar et al. 2014); however, recent sea level data from Quadra Island (~25 km to the NW) shows only a decrease to the present level during the early Holocene (D. Fedje, unpublished data).

Figure 4.2 Sea level curves for the Strait of Georgia near Savary Island. Curves were reproduced from data compiled by Shugar et al. (2014) accessed from http://dx.doi.org/10.1016/j.quascirev.2014.05.022

4.3. Methods

4.3.1. Remote sensing and landform identification

Savary Island is covered with dense conifer forest, which makes identification of landforms from aerial photographs and satellite imagery difficult. To overcome this difficulty, lidar imagery was acquired in August 2014 with a Riegel 580 scanner flown at an elevation of about 1200 m above the ground surface. A Global Navigation Satellite System station established on Quadra Island, 45 km northwest of Savary Island, was

used as the base station for spatial referencing. Point cloud data were interpolated to a 1 m grid for both terrain surface and ‘bare earth’ models. The average point density was 12.5 pts/m2 of which 17 % were classified as ground points and extracted for use in construction of a ‘bare earth’ digital elevation model (DEM).

In order to resolve geomorphic features on the island, principal component analysis was performed using ArcGIS 10.0 software, based on the methods described by Devereaux et al. (2008). Eight hillshade layers of the DEM were used, with azimuths ranging from 0 to 360° at 45° intervals, each with an inclination of 45°. Using ArcGIS 10.0 and GlobalMapper 16, maps of the geomorphology were created. These maps, together with a global positioning system (GPS)-enabled computer tablet, were used to locate features of interest on the ground. Global Mapper 16 was also used for cross- section analysis and 3D modeling of sites.

4.3.2. Lithostratigraphy

Natural exposures of the uppermost ~25 m of the surficial deposits on Savary Island are present in coastal bluffs on the north and south sides of the island. Clague (1976, Fig. 3B) provides a description of some of the bluffs on the south side of the island. For this research, the lithostratigraphy was examined on the south side of the island in order to complement additional observations made in 2000 by J.J. Clague (unpublished data). Lithostratigraphic units were established based on thickness, colour, texture, sedimentary structures, and the nature of the contacts between units. In areas of suspected soil development, soil horizons were identified, measured, and described using the Canadian System of Soil Classification (Soil Classification Working Group 1998).

4.3.3. Dune chronology

Eleven samples were collected for optical dating; five (samples SIDS1 to 4 and SIDS8) were collected from natural wave-cut sections exposed in the flank, or arm, of a parabolic dune; they bracket eroded oxidized units interpreted to be palaeosol B horizons and are associated with charcoal fragments that were collected for radiocarbon

dating. The remaining six samples (samples SIDS5, 6, 7, and 9 to 11) were collected from pits dug in stabilized heads of parabolic dunes (Figure 4.3), and were used to date the last period of dune activity (cf. Wolfe et al. 2011). Samples were collected by inserting aluminum tubes (23 cm long and 7 cm in diameter) horizontally into clean aeolian sand ~1 m below the surface, 20 to 40 cm below B horizons for dune heads (Table 2.1), or in cleaned section faces.

Optical dating measures the time elapsed since mineral grains (typically quartz or feldspar) were last exposed to sufficient sunlight (e.g. Roberts and Lian 2015). The method works on the principle that minerals contain impurities and structural defects, some of which can act as traps for free electrons. Free electrons are produced when minerals absorb radiation from the environment. Some of the traps can hold electrons over geological time (i.e. they are thermally stable at ambient temperatures over geological time), and of these, some can be emptied by a short (few seconds) exposure to direct sunlight. What is required to determine an optical age is a measure of the rate at which radiation in the environment is absorbed by the mineral grains of interest (the environmental dose rate, measured in Gy/ka), and an estimate of the dose of laboratory radiation (usually  or  radiation, measured in Gy) that produces the same amount of luminescence as did the environmental radiation absorbed by the grains since burial (the equivalent dose). An optical age (in ka) is simply the equivalent dose divided by the environmental dose rate. The basic physics behind optical dating are discussed by Aitken (1998), Lian and Roberts (2006), and Wintle (2008).

Figure 4.3 Location of dune heads sampled for optical dating. The number next to each yellow dot corresponds to the SIDS sample number. Samples SIDS1 to 4 and SIDS8 were collected from lithostratigraphic sections at sites 1 and 2 (Figure 4.5).

Optical dating sample preparation and measurement

In the laboratory, under dim orange light, quartz and potassium feldspar (KF) grains were extracted and prepared from the bulk samples using standard procedures

(Wintle 1997): After HCl and H2O2 treatments to remove any carbonates and organic material, respectively, grains were sieved into various grain size fractions. Quartz and KF grains were then concentrated by density separation using lithium metatungstate (2.62 g/cm3 for quartz and 2.58 g/cm3 for KF). To remove the surfaces of the quartz and KF grains that had been affected by  radiation, and to remove any KF contamination from the quartz concentrates, samples were treated with 40% HF acid for 40 minutes for quartz and 10% HF acid for 5 minutes for feldspar. Grains were subsequently given an HCl acid wash to remove fluoride precipitate (Preusser et al. 2009), rinsed several times with deionized water, then with methanol, and finally dried and mounted onto 0.9 cm

diameter aluminum disks using silicone oil as an adhesive. Each disk (aliquot) contained about 100 grains. Aliquots were measured using a Risø TL/OSL DA-20 reader equipped with a calibrated 90Sr/90Y β source that delivered β particles to fine sand mounted on an aluminum substrate at a rate of ~5.4 Gy/min. Luminescence was detected using an Electron Tubes Ltd. 9235QB photomultiplier tube (PMT). Quartz and KF aliquots were stimulated with 72 mW/cm2 of blue (~470 nm) and 130 mW/cm2 of near-infrared light (~870 nm) light, respectively. For the KF aliquots, Schott BG-39 and Corning 7-59 optical filters were placed in front of the PMT; the BG-39 filter absorbs scatter infrared light, the 7-59 filter absorbs the yellow-green (~570 nm) luminescence emitted from plagioclase feldspars, while allowing the violet (~400 nm) emission from KF to be recorded. The ultraviolet (~350 nm) emission from quartz grains was recorded using a single 7.5 mm thick Hoya U-340 optical filter in front of the PMT, which absorbs scattered blue light.

Determination of the equivalent dose

Equivalent dose was determined using a variation of the basic single-aliquot regenerative-dose (SAR) method (Murray and Wintle 2003). The luminescence was recorded over the initial 0.4 s of stimulation for quartz, and over the first 5 s for KF to capture the most light-sensitive part of the signals, with the background signals recorded between 90 and 100 s of stimulation subtracted from it. These data were used to plot a curve (a ‘dose response’ or ‘growth’ curve) of normalized luminescence intensity as a function of laboratory dose. The normalized luminescence signal measured initially from the aliquot (the ‘natural’ signal), before any regenerative doses are given, is interpolated onto the dose-response curve, and the equivalent dose value for that aliquot is read off the dose axis. The reliability of the chosen SAR protocol was assessed by (i) repeating one of the regenerative doses and comparing the resulting signal intensity to that measured after it was given the dose the first time, and (ii) observing the signal measured after stimulation and the preheat treatment, but without a prior dose of radiation (often called recuperation). The former, referred to as the recycling ratio test, serves to check the effectiveness of the test dose to correct for sensitivity change, and the latter gauges the degree of thermal transfer. Typically, the recycling ratio should be within 10% of unity, and the signal measured as a result of thermal transfer should be no

more than 5% of that measured from the ‘natural’ (Murray and Wintle 2000; Wintle and Murray 2006).

In order to test the suitability of the Savary Island quartz for optical dating, linearly modulated optically-stimulated luminescence (LM-OSL) (Bulur 1996; Bulur et al. 2000) was measured, which allows for different components of the luminescence signal to be resolved. LM-OSL data showed that the signal from Savary Island quartz is very dim and that the so-called ‘fast component’ desirable for dating is small or absent (Figure 3.6). Quartz was therefore abandoned, and experiments to find a suitable SAR protocol for KF were conducted, one for which all the built-in quality control tests (cf. Wintle and Murray 2006) are satisfied for the majority of aliquots measured. For all samples, the 180-250 µm grain size range was used, and for one sample a larger grain size (300-400 µm) was also used to look for any differences that might occur as a result of vertical grain movement. Three variations of the standard SAR protocol were used (Figure 3.9): The first (protocol 1) is based on the protocol used by Neudorf et al. (2015) and employs a preheat temperature and duration that are lower than are commonly used and a ‘hot wash’ treatment. A low preheat combination is desirable as it erodes less of the stable signal (electron traps) used for dating. The second protocol uses a higher preheat temperature and duration, typical of those used in most studies, and does not include a hot wash (e.g., Auclair et al. 2003; Feathers and Tunnicliffe 2011), while the third employs the post-IRIR signal, which is intended to circumvent or reduce the effects of anomalous fading (Thomsen et al. 2008).

Determination of the environmental dose rate

The radiation dose absorbed by the mineral grains in the environment comes from , , and  radiation produced during the decay of U, Th, their daughter products, and 40K, within the mineral grains and from their immediate surroundings. To determine the concentrations of these radioisotopes, representative portions of the bulk samples were dried, milled, and sent to a commercial laboratories for neutron activation analysis (NAA) for U, Th, 40K, and Rb; for some of the samples radioisotope concentrations were also determined at a different laboratory using delayed neutron counting (DNC) (Table 4.2). Two different laboratories were used to add confidence to the results, and because DNC provides better precision for U, which occurred in low concentrations in these

samples. From the concentrations of these radioisotopes, dose rates due to  and  radiation were calculated using standard formulae (Aitken 1985; Berger 1988; Lian et al. 1995). The dose rate due to  radiation, which is internal to the grains, was estimated to be 0.08 Gy/ka (cf. Ollerhead et al. 2001). Pore water in the sediment matrix attenuates or absorbs radiation differently than minerals do, and it therefore has to be taken into account when dose rates are calculated. All of the samples consisted of well-drained aeolian sand, so in each case the as-collected water contents were used. There is also a contribution from cosmic-ray radiation, and this was estimated using present burial depths and the formula of Prescott and Hutton (1994).

Optical age determination

For each sample equivalent dose values were determined from 20 to 30 aliquots, which were analyzed together using the central age model (CAM) (Galbraith et al. 1999). The CAM determines the weighted mean of all the aliquots in a sample and takes into account the level of overdispersion (OD), which is a measure of the spread in the values beyond those associated with analytical uncertainties. Optical age values for each sample were calculated by dividing representative equivalent dose values (found using the CAM), by their respective environmental dose rates. Optical ages were corrected for anomalous fading using the method of Huntley and Lamothe (2001). The fading rate was found using the method of Auclair et al. (2003) on 12 aliquots from each sample.

Radiocarbon dating

Charcoal samples collected during this study were dried and sent to Paleotec Services (Ottawa, Ontario), where they were inspected for contamination using a microscope. Some samples contained modern roots and fungi and were rejected, samples that were deemed suitable for dating were forwarded to the W.M. Keck Carbon Cycle AMS Laboratory (University of California, Irvine) where they were given a standard acid-base-acid treatment to remove any contaminating carbonates and soil humics. Samples were then washed with purified water, dried, and converted to graphite for radiocarbon dating by accelerator mass spectrometry (AMS). Radiocarbon ages were converted to ka cal BP using Oxcal 4.2 and the IntCal 13 data set.

4.4. Results

4.4.1. Geomorphology

The bare earth DEM enabled identification of geomorphic features on Savary Island. These features include, from east to west, terraces, a relatively flat plain with landform remnants, relict (stabilized) parabolic dunes, raised embayed beaches, and a spit at the west end of the island. The parabolic dunes occur over an area of just under 1 km2 near the centre of the island. They range in length from 300 to 1000 m, and are 70 to 100 m wide (Figure 4.4). East of the dunes the landscape is relatively flat; in places it is interrupted with irregular landforms that look like remnant parts of dunes (Figure 4.4). These landforms appear randomly dispersed and differ in morphology from each other. At lower elevations, they appear more subdued.

Figure 4.4 Oblique ‘three-dimensional’ view of the central 4 km of Savary Island with associated topographic profile (A – A’). Undifferentiated landforms occur in places east of the dunes and may be the remains of eroded dunes.

4.4.2. Lithostratigraphy

Lithostratigraphy was documented at the south-facing bluffs (sites 1 and 2, Figure 4.5) where three units were identified. Unit 1 is up to 15 m thick and consists of horizontally and cross-bedded, moderately to well-sorted, medium to coarse sand. At

the top of this unit the sand grades into gravelly silty sand; the lower contact is below the beach surface and was therefore not observed. Unit 1 is interpreted to be Quadra Sand, glaciofluvial outwash deposited during the advance phase of the Fraser Glaciation about 30,000 years ago (Clague 1976; Armstrong and Clague 1977).

Unit 2 is between 5 and 50 cm thick. Its lower contact is sharp and lies ~13 m above the high tide line at site 2. It consists of highly consolidated matrix-supported diamicton, which, in places grades into, or is replaced by, poorly sorted matrix-supported gravel. Clasts range in size from pebbles to boulders. This unit is interpreted to be basal till deposited during the Vashon Stade of the Fraser Glaciation. The gravel zone at the top of the unit is interpreted to be reworked by waves during deglaciation when relative sea level was higher than present.

Unit 3 ranges in thickness from 3 to 9 m and is composed of well-sorted massive medium to coarse sand, in some places with weak planar cross-bedding. Grain-size analysis indicates 80% of the sediment in samples taken from this unit is between 290 and 770 µm in diameter. Average grain size of all the samples is ~470 µm. A sharp contact exists between unit 3 and the underlying diamicton or gravel. This unit extends to the surface, where it forms large parabolic dunes. Unit 3 is interpreted to be aeolian sand. Within unit 3, there are two prominent (~ 70 cm thick), and two weaker (~20 cm thick), yellowish-brown zones. The top of the first of the prominent zones (zone 1) is at 14.2 m asl, 80 cm above the contact with unit 2; it is olive yellow (2.5Y 6/6) at the top with many, coarse, yellowish brown (10YR 5/8) mottles, most of which are concentrated between 2 and 5 cm from the top of the zone. The olive yellow zone fades downward, becoming light yellowish brown (2.5Y 6/4) over a vertical distance of 70 cm, with common, medium light gray (2.5Y 7/2) mottles throughout. The top of the second yellowish-brown (10 YR 5/4) zone (zone 2) is 2.30 m higher in the unit, 3.10 m above the contact with unit 2. Zone 2 begins with a yellowish brown, abrupt broken boundary at the top, with reddish brown (5 YR 4/4) moderately cemented nodules up to 3 cm in diameter. Below this is a mixed region of light olive brown (2.5Y 5/4) sand and a locally less common, light gray (2.5Y 7/1) sand for 50 cm. Zone 2 fades to light gray (2.5Y 6/3) over a vertical distance of about 70 cm. Both zones 1 and 2 can be traced laterally for several metres, disappearing beneath slumped sediment or modern vegetation.

Charcoal fragments (<1 cm in diameter, most <0.5 cm) are dispersed sporadically throughout the uppermost parts of both of these zones; a charcoal fragment from each zone was collected for radiocarbon dating (samples UCIAMS149587 and UCIAMS149588, Table 4.1), as were samples for optical dating (samples SIDS1, 2, and 1C, Table 4.3). Optical samples were collected 40 cm below the tops of the zones. Two similar, but weaker, yellowish-brown zones (zones 3 and 4) occur higher in the unit. Zone 3 is 4.80 m above the contact with unit 2 and 6.10 m beneath the top of the section. Zone 3 is reddish gray (2.5 YR 7/1.5) at the top and includes few faint light orange (2.5Y 7/3) mottles that disappear about 12 cm below the top of the zone. At 12 cm the faint light orange mottled zone transitions with an abrupt, wavy boundary into light yellowish brown (2.5Y 6/4) with few, medium, faint light gray (2.5Y 7/2) mottles. The total thickness of zone 3, as indicated by its colour is ~60 cm. A charcoal fragment from this zone was collected for radiocarbon dating (samples UCIAMS159976, Table 4.1) 10 cm below the top of the zone, as was a sample for optical dating (samples SIDS8, Table 4.3) ~70 cm from the top of the zone. Zone 4 is higher in unit 3, 7.90 m above the top of the contact with unit 2 and 3 m beneath the top of the section. Zone 4 fades downwards and becomes indistinguishable from the surrounding sand over a distance of 15 cm. Samples believed to be charcoal were collected for radiocarbon dating from zone 4; however, after close inspection they were found to consist of darkened brown ‘peds’, with clumps of modern ectomycorrhizae and rhizomes fragmented with ‘fuzzy’ fungal stringers surrounding the peds (A. Telka, Paleotec Services, personal communication 2015). This material is interpreted to be contaminated with modern vegetation through bioturbation and therefore is not suitable for radiocarbon dating.

The yellowish-brown colouration of these zones is interpreted to be the result of oxidation, and this, together with the presence of mottling and cementation, indicate that these zones are soil B-horizons. No evidence of O or A horizons was observed; these horizons therefore have been removed, likely together with some of the B-horizons, by subsequent aeolian erosion. The charcoal fragments found scattered within the B- horizons are interpreted to have been transported to the site with the aeolian sand that became the soil parent material. The two prominent B-horizons, zones 1 and 2, are hereafter referred to as palaeosol 1 (upper; zone 2) and 2 (lower; zone 1), respectively

(Figure 4.6). The thinner and weaker zones 3 and 4 are believed to be incipient palaeosols, hereafter referred to as incipient palaeosol 1 (upper; zone 4) and 2 (lower zone 3), respectively. The incipient palaeosols show weak discoloration and mottling, appearing to reflect the initial stages of soil development. Optical dating samples SIDS3 and SIDS4 bracket insipient palaeosol 1 (Figure 4.5; Table 4.3).

Figure 4.5 Composite section diagram of sites 1 and 2 (location shown in Figure 4.2) showing locations of optical dating samples (black dots) and ages. Brown lines indicate tops of eroded palaeosol B horizons. Horizontal dimension gives grain size as indicated in the lithofacies codes. Beach level refers to the log line.

Figure 4.6 (A) Top of palaeosol 1 showing small scattered, charcoal fragments (indicated by arrows) and mottling; locations shown in Figure 4.3. (B) Close-up of palaeosol 2 and its upper contact, as indicated by the abrupt change in colour. The scale is in cm.

4.4.3. Radiocarbon and optical ages

Radiocarbon ages of charcoal fragments collected during this study (UCIAMS samples) are shown in Table 4.1 together with those of charcoal fragments collected and dated J.J. Clague in 2000 (TO samples; unpublished data).

Table 4.1 Radiocarbon AMS ages from charcoal fragments collected from soil B horizons at sites 1 and 2 (Figure 4.5).

Site Lab No.1 Date 14C Age Age (cal yr Collected (years BP) BP) 2 Site 1 (insipient UCIAMS159976 6/21/2015 8245 ± 20 9290 - 9130 palaeosol 2) UCIAMS149587 01/07/2014 7400 ± 25 8320 - 8180 Site 2 (palaeosol 1) TO-9155 11/26/2000 7180 ± 120 8300 - 7740

UCIAMS149588 02/07/2014 8550 ± 40 9560 - 9480 Site 2 (palaeosol 2) TO-9154 11/26/2000 8730 ± 120 10160 - 9530

1UCIAMS – W.M Keck Carbon Cycle AMS Laboratory, University of California, Irvine; TO – IsoTrace Laboratory, University of Toronto. 2 Calibrated using Oxcal 4.2 and the IntCal 13 data set. Age ranges are 2.

Preheat plateau experiments revealed that the 160 °C / 10 s preheat combination used in protocol 1 results in recycling ratios closest to unity and recuperation values

(Lx/Tx)/(Ln/Tn) below 2.9%, which thus have a negligible effect on De determination (Wallinga et al. 2000). Furthermore, a dose recovery test indicated that protocol 1 could successfully recover a laboratory dose of 24.7 Gy (dose recovery ratio 0.98 ± 04%), a dose corresponding to an age of ~10 ka, slightly higher than that expected for these samples. Grains from two grain size ranges (180-250 and 300-400 µm) were analyzed to see if any differences due to transport history or movement within the sedimentary profile could be seen, but no significant differences were observed. As a result, fading corrected ages have been calculated for all dune head samples using SAR protocol 1 and are shown in Table 4.3.

Table 4.2 Optical dating sample depths, 40K, Rb, Th, and U concentrations, and water contents. Dashed line separates samples from sites 1 and 2 (above dashed line) from dune head samples (below dashed line). Sample locations are shown in Figure 4.3

Depth below 40 1 Sample K (%) Rb (µ/µg) Th (µ/µg) U (µ/µg) Δwac ground (cm) SIDS12 690 1.02 ± 0.21 20.8 ± 3.5 1.09 ± 0.10 0.50 ± 0.12 0.03 SIDS22 920 0.67 ± 0.18 16.7 ± 3.3 1.89 ± 0.15 0.62 ± 0.12 0.06 SIDS32 296 1.01 ± 0.19 18.4 ± 3.0 1.07 ± 0.10 0.35 ± 0.11 0.03 SIDS42 250 1.05 ± 0.21 22.0 ± 3.1 1.01 ± 0.09 0.52 ± 0.10 0.03 SIDS82 480 0.85 ± 0.19 18.5 ± 3.0 1.75 ± 0.14 0.58 ± 0.13 0.06 SIDS52 140 1.02 ± 0.22 20.4 ± 3.6 2.32 ± 0.18 0.68 ± 0.13 0.04 SIDS62 110 0.76 ± 0.18 17.7 ± 3.3 2.09 ± 0.16 0.72 ± 0.13 0.06 SIDS72 120 0.76 ± 0.18 20.2 ± 3.2 2.11 ± 016 0.55 ± 0.13 0.04 SIDS92 105 0.89 ± 0.19 18.0 ± 3.3 2.26 ± 0.17 0.72 ± 0.12 0.05 SIDS103 140 1.00 ± 0.11 14.0 ± 1.3 1.80 ± 0.10 0.90 ± 0.30 0.05 SIDS113 135 1.20 ± 0.11 21.0 ± 1.5 2.0 ± 0.11 0.70 ± 0.30 0.04 1 "as collected" water content (mass water/mass dry minerals). An uncertainty of ±10% was included in the dose rate calculations 2 Australian Nuclear Science and Technology Organization (ANSTO, Lucas Heights, NSW, Australia) laboratory where 40K, Th, and Rb concentrations were determined by Neutron activation analysis, and U concentrations were found by delayed neutron activation analysis. 3 Neutron activation analysis performed at Maxxam Analytics, commercial laboratory in Mississauga, Ontario.

Figure 4.7 A) Example of a luminescence decay curve (sample SIDS5) typical of those derived from the Savary Island samples. Inset graph shows the dose-response curve for the same sample aliquot. An exponential curve is fitted to the data onto which the natural signal is interpolated and the equivalent dose (De) is read from the dose axis. B), C), and D) show radial plots for some of the samples (n = number of aliquots accepted, OD is overdispersion). De values that fall within the grey bands are within 2σ of the weighted mean value.)

Table 4.3 Optical dating samples, total dose rates, equivalent dose (De), cosmic dose (Dc), overdispersion (OD) values, fading rates (g), and optical ages (uncertainties are ±1) found using protocol 1. Dashed line separates samples from sites 1 and 2 (above dashed line) from dune head samples (below dashed line). Sample locations are shown in Figure 4.3

Total dose Uncorrected Fading-corrected Sample N1 D (Gy/ka) D (Gy)3 OD (%) g (%/decade) rate (Gy/ka)2 c e age (ka) age (ka)

SIDS1C 36/36 2.20 ± 0.15 0.07 ± 0.01 6.62 ± 0.23 20 ± 3 7.49 ± 0.10 3.02 ± 0.23 6.38 ± 0.45

SIDS2 22/24 1.91 ± 0.13 0.06 ± 0.01 7.32 ± 0.32 21 ± 3 6.92 ± 0.10 3.83 ± 0.41 7.69 ± 0.71

SIDS3 27/27 2.21 ± 0.15 0.13 ± 0.01 6.63 ± 0.14 10 ± 2 4.99 ± 0.09 3.01 ± 0.21 4.76 ± 0.30

SIDS4 25/25 2.29 ± 0.15 0.14 ± 0.01 6.26 ± 0.37 17 ± 2 4.99 ± 0.09(4) 2.74 ± 0.21 4.32 ± 0.29

SIDS8 30/30 2.09 ± 0.14 0.10 ± 0.01 7.86 ± 0.27 19 ± 3 5.99 ± 0.10(5) 3.76 ± 0.28 6.70 ± 0.45

SIDS5 28/30 2.38 ± 0.16 0.16 ± 0.01 8.17 ± 0.20 13 ± 2 5.60 ± 0.09 3.39 ± 0.24 5.76 ± 0.36

SIDS6B 30/30 2.14 ± 0.15 0.17 ± 0.01 7.11 ± 0.16 12 ± 2 5.83 ± 0.10 3.32 ± 0.24 5.78 ± 0.36

SIDS7 35/35 2.14 ± 0.14 0.17 ± 0.01 7.86 ± 0.22 16 ± 2 5.66 ± 0.10 3.67 ± 0.26 6.29 ± 0.40

SIDS9 28/30 2.27 ± 0.15 0.17 ± 0.01 7.15 ± 0.23 18 ± 3 5.47 ± 0.09 3.09 ± 0.31 5.22 ± 0.34

SIDS10 36/36 2.35 ± 0.19 0.16 ± 0.01 7.06 ± 0.18 15 ± 3 5.22 ± 0.08 3.00 ± 0.26 4.88 ± 0.33

SIDS11 30/30 2.52 ± 0.18 0.17 ± 0.01 8.37 ± 0.22 14 ± 2 4.89 ± 0.10 3.31 ± 0.26 5.20 ± 0.34

1 Number of aliquots measured/accepted.

2 Dose rates due to γ and β radiation were calculated using standard formulae (e.g. Aitken 1985; Berger 1988; Lian et al. 1995), and the dose-rate conversion factors of Guerrin et al. (2011), and the grain-size dependent absorption factors of Brennan (2003). The dose rate due to a radiation, that is internal to the grains, was estimated to be 0.08 Gy/ka (cf. Ollerhead et al. 2001) 3 Found using the Central Age Model (CAM). Recuperation values for all samples were between 1.6 and 2.9 %. 100% of aliquots from ten of the samples passed the recycling ratio test and the remaining three samples had 93% and 96% of aliquots pass. 4 Using the fading rate of SIDS-3, located directly below SIDS-4

5 Using an average fading rate of samples collected from sites 1 and 2 (SIDS1C, SIDS2, SIDS3, SIDS4)

4.5. Summary and discussion

Optical dating of multi-grain aliquots of quartz and KF sand concentrates (180– 250 and/or 300–400 µm diameter size ranges) extracted from the heads of forested parabolic dunes on Savary Island was undertaken to establish when the dunes stabilized. Optical dating samples associated with palaeosols collected from what is likely a dissected dune arm, some near its contact with underlying glacial sediments, were collected to estimate dune initiation and soil formation (intermittent periods of dune stability).

4.5.1. Geochronological age estimation

Observation of the linearly-modulated ultraviolet luminescence signal from Savary Island quartz indicated that it is not suitable for optical dating, and this was confirmed by preliminary SAR experiments that showed that the performance of the built-in quality control tests was low. Quartz was abandoned and all further experiments were done using KF, IR stimulation, and the violet luminescence signal.

Laboratory experiments indicated that for KF, a SAR protocol using a low preheat temperature (160 ºC for 10 s) and a ‘hotwash’ treatment (180 ºC for 40 s) at the end of each SAR cycle (protocol 1, Figure 3.9) yielded optical ages consistent with those found using a more stringent, but more signal-erosive, preheat treatment (250 ºC for 60 s) typical of many of those that have been used in other studies. Fading rates (g-values) were found using the SAR method of Auclair et al. (2003) and optical ages were corrected for anomalous fading using the method of Huntley and Lamothe (2001). Fading rates ranged from 4.89 ± 0.10 to 7.49 ± 0.10 %/decade, and are all typical of those found for KF in other studies in BC (e.g., Huntley and Lamothe 2001; Huntley and Lian 2006; Wolfe et al. 2008; Neudorf et al. 2015). Optical dating of two of the samples (SIDS2 and SIDS5) using an established post-IRIR SAR protocol yielded ages consistent with those found using protocol 1, but with fading rates that were not negligible (2.31 ± 0.16 and 5.60 ± 1.75 %/decade). Furthermore, the relatively dim post-

IRIR signal resulted in higher analytical uncertainties in the De (and age) values. These

results suggest that there is no advantage to using post-IRIR protocols to date KF on Savary Island, although further experiments using different SAR parameters (e.g., preheat combinations) should be performed to confirm this.

Optical ages from dune heads (protocol 1) range from 4.88 ± 0.33 to 6.29 ± 0.40 ka, with a weighted mean age of 5.47 ± 0.36 ka; all ages are consistent at 2, most at 1(Table 4.3). Optical dating samples that bracket an eroded palaeosol B-horizon exposed in what is likely the arm of a stabilized dune (palaeosol 2; Figure 4.3), near the contact with underlying glacial sediments (Figure 4.5), yielded ages of 7.69 ± 0.71 ka (SIDS2, below) and 6.38 ± 0.45 ka (SIDS1, above). Sample SIDS1 also gives a maximum age for palaeosol 1 (Figure 4.6), which occurs about 3 m higher in the section. Duplicate charcoal fragment samples collected about 10 cm below the top of the eroded B-horizon of palaeosol 2 gave age of 8550 ± 40 14C yrs BP (9560-9489 cal yrs BP, UCIAMS149588) and 8730 ± 120 14C yrs BP (10160-9530 cal yrs BP, TO-9154), while those collected near the same position relative to palaeosol 1 gave ages of 7400 ± 25 14C yrs BP (8320-8180 cal yrs BP, UCIAMS149587) and 7180 ± 120 14C yrs BP (8300- 7740 cal yrs BP, TO9155) (Table 4.1; Figure 4.7). Incipient palaeosols 1 and 2, exposed a few metres higher in the sections, are bracketed by optical ages of 6.30 ± 0.42 ka (SIDS8, below), 4.76 ± 0.30 ka (SIDS3, between), and 4.32 ± 0.29 ka (SIDS4, above) (Table 4.3; Figure 4.5). A charcoal fragment collected immediately below incipient palaeosol 2 yielded an age of 8245 ± 20 14C yrs BP (9290-9130 cal yrs BP, UCIAMS159976) (Table 4.1; Figure 4.5).

Optical ages could be too old if all, or a significant number, of the grains in the sample were exposed to an insufficient amount of sunlight before final burial. Since all the samples consist of grains that would have been transported by aeolian processes over a migrating dune field, this is unlikely. This is supported by overdispersion values (Table 4.3) that fall within the range of those expected for well-bleached samples (Arnold and Roberts 2009), and the observation that the distributions of aliquot De values do not show discrete populations (Figure 4.7), although subtle populations could be hidden in the multi-grain aliquot data. Optical ages could also be too old if the environmental dose rate had decreased over time, or if there were inhomogeneities at the samples sites that were not accounted for. An environmental dose rate that is too low can occur if

disequilibrium is introduced to the U decay chain or if pore water contents were estimated incorrectly. However, radioactive disequilibrium and significant variation in water content are unlikely in well drained mineral sand at sites away from organic-rich horizons (Prescott and Hutton 1995; Olley et al. 1998), as was the case here.

Optical ages could be too young if younger grains are introduced to the sample sites. This can happen if grains are able to migrate downward by gravity in the sediment profile through a matrix of larger grains (cf. Berger et al. 1994), and/or by bioturbation. However, consistent ages from two grains size ranges (180–250 and 300–400 µm) from sample SIDS1C suggest this was unlikely. Partial reworking of grains can also result in an optical age being too young, and we think this may be the case for samples SIDS3 and 4, near the top of the section. This is supported by their optical ages being younger than the dune head ages (dune arms are expected to stabilize earlier than dune heads), and the presence of modern roots and modern detrital organic matter near the sample sites (see Appendix A, Site 1, pedon description for SIDS3 and SIDS4); the latter was confirmed by Paleotec Services prior to an attempt at radiocarbon dating at this position. It is also possible that the sand at samples sites SIDS3 and SIDS4 were partially reworked by wind, for example during recent erosion of the section face, or by slumping of the sand above the bluff face. Optical ages could be too young if the measured fading rates are two low. Fading rates were measured using the SAR method of Auclair et al. (2003) which employs much shorter delay times between measurements than the original well-tested method developed by Huntley and Lamothe (2001); although the method of Auclair et al. (2003) has been employed in many subsequent studies, there has been little comparison between fading rates determined using the two methods on the same samples. It is therefore possible that our fading rates could be too low. However, as mentioned earlier, fading rates for all the samples measured during thus study are typical of those measured during other studies along the BC coast.

Radiocarbon ages of two individual (duplicate) charcoal fragment samples associated with each of palaeosols 1 and 2 (Table 4.1) are consistent with each other at 2, and both pairs of ages are in correct stratigraphic order (Figure 4.5). Any difference between duplicate ages may be the result of different death ages, or if the charcoal was derived from different parts of a tree (e.g., Gavin 2001; Frueh and Lancaster 2014), or if

the material was reworked from sources of different age. However, the radiocarbon age of charcoal fragments associated with incipient palaeosol 1 is significantly older than those of palaeosols 1 and 2, which appear to be stratigraphically lower (Figure 4.5).

Calibrated radiocarbon ages are not consistent with associated optical ages at any sample location, the latter always being significantly younger. In other studies, charcoal found in the lower mineral horizons of soils have been interpreted to have originated as a result of the burning of in situ root systems, root throw, or animal burrowing (e.g., Gavin et al. 2003). Few isolated root casts, but no evidence of bioturbation, were found at the sample sites; however, the presence of charcoal fragments at approximately the same positions in palaeosols 1 and 2, and in incipient palaeosol 2, is consistent with them being transported by bioturbation. Although the apparent reversal in the radiocarbon ages of charcoal fragments associated with palaeosol 1 and incipient palaeosol 2 supports that the charcoal fragments were reworked, the age reversal might be explained by the former having been established on a surface that dipped downward to the position of the latter; the palaeosols were not traced laterally between sites 1 and 2 (Figure 4.5) to confirm this. However, the strikingly different degrees of B-horizon development between them (one is incipient), and the fact that modern soils that have developed on dune heads and in dune swales showed little difference, suggest that incipient palaeosol 2 and palaeosols 1 and 2 are all remnants of different soils. No evidence of a palaeosol was found between palaeosols 1 and 2 at site 2. An alternate explanation for why the charcoal fragments occur in these positions is that they were deposited when dune arms began to become stabilized; stabilizing vegetation might have reduced surface wind speeds allowing small charcoal fragments to be deposited. In that case the charcoal fragments would have likely been eroded from older horizons (burned surfaces?) exposed in deflation basins between dunes and between dune arms (Appendix H, Figure H1). Others (Blong and Gillespie 1978; Loope and Swinehart 2000; Halfen et al. 2010) have observed the reworking of older charcoal into aeolian sediments. In that case, the charcoal fragments give a maximum age for aeolian deposition and soil formation at these stratigraphic positions.

4.5.2. Landscape evolution

The major environmental variables controlling coastal dune formation include available sediment, sufficient onshore winds, and the presence and type of stabilizing vegetation. When RSL was at its lowest point, Savary Island would have been considerably larger, its southern shore being about 1.5 km away, based on the size of a shallow erosional platform that now borders the island (Clague et al. unpublished data) (Figure 4.8 C). The lower RSL would have exposed glacial outwash deposits within the Strait of Georgia and provided sediment for construction of the dune field (Figure 4.8 C and D). Sometime after ~12 ka, after RSL drop, vegetation in the area would have started to become established; evidence of this is the charcoal associated with the palaeosols that date between 8200 to 9600 cal yr BP (Table 4.1). Wind direction during this time matched modern winter wind patterns (Figure 1.2), as indicated by the orientation of the dunes on the island (Figure 4.8 D); the D50 grain size from dune head samples indicate dune forming winds were likely around 30 km/hr near Savary Island, although it is recognized that sustained high-speed winds over several hours are required for dune formation of this magnitude. Dunes extend across the entire width of the island today, reaching lengths up to 1000 m indicating an episode(s) of significant aeolian activity. The existence of palaeosols indicates periods of stabilization which allowed for soil development, followed by re-activation and aeolian deposition. The degree of development of the palaeosol B horizons exposed at site 2 suggests that these episodes of stability were short-lived, lasting perhaps only a few hundred years (P. Sanborn, personal communication, April 2015), and this is consistent with the optical ages. This is in contrast to the present surface, which is dominated by a mature Douglas fir (Pseudotsuga menziesii) forest growing in a well-developed Dystric Brunisol soil. The rate of RSL lowering began to slow appreciably around ~ 8 ka cal ago, and this would have reduced sand supply (Figure 4.8 E).

Several palaeoecological studies (e.g., Mathewes 1973, 1985; Heusser et al. 1980; Mathewes and Heusser 1981; Hebda 1983; Heusser 1985; Lacourse 2005; Galloway et al. 2009;) have shown that climate following deglaciation was cold and moist until about 10 ka 14C BP (12.5 to 11.3 cal ka BP) when conditions became warm and dry. This warm and dry period is commonly referred to as the Holocene xerothermic

interval (HXI). It was a time when mean summer temperature was about 2°C warmer than present, and mean annual precipitation may have been 20-30% lower than that of today (Heusser et al. 1985). These conditions might have affected vegetation and would have likely contributed to increased aeolian activity and dune development. The HXI persisted until about 8000 years ago when climate transitioned to cooler and moister conditions, and then to cool and wet by about 4000 years ago (Figure 4.9). All of the charcoal associated with the dune palaeosols studied during this work date to HXI, with ages from 8200 to 9600 cal yr BP. Accepting that our optical ages are correct, dune head stabilization occurred about at 5.47 ± 0.36 ka, around the time when the rate of RSL lowering was slowing, and much after the end of the HXI when climate was cool and moist. The oldest dune optical age (sample SIDS2), collected near the contact with underlying glacial sediments (Figure 4.5), suggests that perhaps the combination of steady wind, high sediment supply, and warm and dry conditions (vegetation type) resulted in repeated reworking of the pre-existing aeolian cover until about 7.69 ± 0.71 ka when an episode of periodic stabilization occurred, but more dating is needed to confirm this. However, if the charcoal samples more closely date palaeosol development, then periodic stabilization of the aeolian cover began sometime before ~9.5 ka during the HXI.

Figure 4.8 Possible scenario for postglacial evolution of Savary Island since deglaciation. (A) By ~14 ka cal BP, ice is retreating up the Strait of Georgia (based on Clague and James 2002). Modern shoreline shown in dashed line for reference. (B) As ice vacated the area, the isostatically depressed land was immediately transgressed by the sea, RSL was ~ 150 m above modern (Clague and James 2002). (C) As the land isostatically rebounded and RSL dropped to between -10 to -20 m below present (Barrie and Conway 2002), sediment became exposed and available for aeolian erosion, entrainment, transport, and deposition. (D) Dune formation began and persisted in conjunction with moist climate (Figure 4.9), likely on some of the exposed outwash or Quaternary sediment surfaces prior to this time. (E) As time progressed, the sediment source diminished with associated RSL rise (Figure 4.2B) while climate transitioned to cool and moist conditions. Soil moisture increased, larger pioneer species became established, and dunes stabilized ~5.5 ka cal BP (dune head ages Figure 4.9). (F) The modern relationship between sea level and land in the Savary Island region.

Figure 4.9 Optical ages (round black data points; error bars are ± 2) of dune head stabilization plotted with climate periods determined from regional palaeoecological studies, as compiled by Galloway et al. (2009) and based on radiocarbon dating; original radiocarbon ages have been calibrated for use here using OxCal 4.2 and the IntCal 13 data set. Error bars on climate zones indicate one standard deviation of uncertainty in the transition between climate periods. This uncertainty can be large because of the nature of palaeoecological studies, where the local effects (topography and environment) can influence pollen type, abundance, and preservation, and therefore climate interpretations. Wet climate classifies environments with increased precipitation. Calibrated radiocarbon age range from charcoal in palaeosol B-horizons indicated by light grey line.

4.6. Conclusions and future research

Optical dating of forested parabolic dunes on Savary Island indicates that they stabilized 5.47 ± 0.36 ka ago when RSL was stabilizing, inferred sediment supply was decreasing, and climate was becoming increasingly wet. A single optical age near the base of a dune arm suggests that a pronounced transition toward stabilization began 7.69 ± 0.71 ka ago through a series of episodes of instability (indicated by clean sand) and stability (indicated by palaeosol B-horizon remnants). This is the first study of dune stabilization in the central Strait of Georgia, coastal British Columbia, and it has implications for the establishment of vegetation and our understanding of local landscape response to sand supply and climate.

The discrepancy between the radiocarbon and optical ages associated with the palaeosols gives some doubt to the optical dating protocols used, and the resulting ages, despite the fact that the samples passed all the built-in quality-control tests associated with the optical dating SAR methods used. In light of this, well defined chronostratigraphic beds of known age should be sought in the aeolian sediments on Savary Island that can be used to further test optical dating protocols. Single-grain analysis of the quartz fraction should be undertaken to identify and date grains showing a prominent thermally-stable fast component of the signal; this would not only avoid having to deal with anomalous fading, but it would also allow grain populations of different ages to be more readily identified. Additional experiments on the post IRIR signals from KF would also be useful in order to minimize the effect of anomalous fading.

As the organic material associated with the palaeosols that was radiocarbon dated during this study is detrital, it would be useful to look for more complete buried soils in the aeolian sand that caps Savary Island. This might provide better radiocarbon dating control and further insight into earlier palaeoenvironments. Additional optical dating at the base of the aeolian unit could be performed to better establish the age for the onset of dune formation. Dating of sea-level indicators on Savary Island such as raised terraces and aggradational spit surfaces (Figure 3.3) should be done to establish a better link between local sea-level position and dune formation and stabilization. The establishment of a local palaeoecological record would increase our understanding of the links between shifts in climate and the character of dune formation.

Chapter 5. Conclusions

The main goal of this thesis was to a develop a better understanding of the character and timing of postglacial geomorphic change on Savary Island through remote sensing, landform interpretation, optical and radiocarbon dating, and pedology, with a focus on the timing of the stabilization of the dune field that occupies the centre of the island. This chapter summarizes the principal results from this study and suggests future avenues for research on the island and in the region.

5.1. Geomorphology

5.1.1. Summary

Acquisition of lidar imagery and subsequent generation of a ‘bare-earth’ DEM, enabled the clarification and identification of geomorphic features on Savary Island. These features include, from east to west: raised terraces and cliff top deposits; relatively flat sand plains with local raised areas interpreted to be remnants of landforms; parabolic dunes; raised palaeobeaches; bluffs, and a prograding spit (Figure 3.3). The 1 km2 area of stabilized parabolic dunes that cover the centre of the island was the main focus of this study. The dunes reach lengths up to 1000 m and vary in width from 90 to 180 m. Samples were collected for optical dating from the heads of dunes to determine stabilization ages while soils pits were dug in dune heads and swales to get a better understanding of modern environmental conditions. These dunes are unique to the coast of BC, the only known stabilized and preserved parabolic dunes in the Strait of Georgia that currently are accessible.

100

5.1.2. Future work

Many of the landforms on Savary Island have been identified for the first time and have not yet been the subject of further study, as they were outside the scope of this research. These new landforms, in particular the wave-cut terraces interpreted to be palaeobeaches, may hold valuable insight into our understanding of the land’s response to isostatic depression and the RSL history of the area. By using optical dating (the protocol developed in this research), researchers may be able to refine the known RSL curve of this region (Figure 1.5). Additional landforms discovered on Savary Island that require attention in the future are the raised areas on the planar surface east of the dunes interpreted to be remnants of landforms; some of which appear to be eroded dunes (Figure 3.3C). To understand their genesis their composition and age should be investigated to see if similarities exist between them and the large parabolic dunes. This may have implications for the previous extent of the existing parabolic dune field and for understanding of the RSL history.

The RSL record is becoming an increasingly important topic as it has been proven to largely impact where and when first peoples settled along the coast (e.g., Josenhans et al. 1995, 1997; McLaren et al. 2014; Fedje and Josenhans 2000; Fedje and McLaren 2016). There are known archeological sites on Savary Island including middens, gardens, and culturally modified trees. Stories indicate that the island was an important resource for the coastal people in the past, providing a summer location where they could fish, harvest clams, and bathe in the warm waters (Kennedy 1992). Further investigations into the archaeological history of this island, in light of the geomorphic history, would complement those in the surrounding regions.

5.2. Optical dating

5.2.1. Summary

This study underscored the inherently experimental nature of optical dating by examining the different characteristics of quartz and KF minerals on Savary Island to develop a reliable dating protocol. KF was found to be a suitable mineral for dating,

101

while the luminescence signal from quartz was found to be dim and have only a minor or negligible ‘fast’ component, similar to that observed for this mineral on Calvert Island to the north (Neudorf et al. 2015). The SAR optical dating protocol developed by Neudorf et al. (2015) for KF on Calvert Island was found to be suitable for this mineral on Savary Island as well (i.e., all the built-in SAR quality control tests were satisfied, as were dose- recovery tests).

Optical dating results suggest that the use of post-IRIR SAR protocols to date KF from Savary Island holds little advantage over more traditional SAR protocols. Since the intention of using the post-IRIR signal is to minimize or remove the impact of anomalous fading, its application had no benefit for dating these sample as measured fading rates were not low enough to avoid correction. In fact, as the post-IRIR signal bleaches much more slowly, this protocol increases the risk of measuring a residual signal. Moreover, overdispersion was higher with the post-IRIR signal, possibly due to lower counting statistics or from grains with different bleaching histories, as the post-IRIR signal bleached more slowing than the main signal.

5.2.2. Future work

Additional work on refining the chronology of landforms associated with RSL changes on the island (see section 5.5.1) could be performed using the optical dating protocol developed and applied during this research. This would increase our understanding of the evolution of the island and possibly help refine the RSL curve in the area. When applying any optical dating technique, or protocol, in a region for the first time, it is recommended that some comparison be made between optical ages and independently dated material in the same context (Lian 2013). The discrepancy between our radiocarbon ages of isolated charcoal fragments in the B-horizons of palaeosols developed in dune sand, and optical ages of samples collected directly below these horizons, suggest that future work to this end should be undertaken, despite the fact that the samples passed all the built-in quality-control tests associated with the optical dating SAR methods used.. In light of this, well defined chronostratigraphic beds of known age should be sought in the aeolian sediments on Savary Island that can be used to further test optical dating protocols.

102

Further development and testing of optical dating protocols for Savary Island quartz and KF could improve the efficiency, and perhaps the accuracy, by which landforms are dated using this technique. This could include single-grain analysis of the quartz fraction to identify and date grains showing a prominent thermally-stable fast component of the signal; this would allow grain populations of different ages to be more readily identified and circumvent the effects of anomalous fading. Additional experiments on the post IRIR signals from KF would also be useful in order to minimize the effect of anomalous fading. Additional methods could be applied like the multiple- aliquot additive-dose (MAAD) method that has been successfully applied to coastal deposits in BC in the past (e.g., Lian et al. 1995; Wolfe et al. 2008; Mathewes et al. 2015).

5.3. Pedology

5.3.1. Summary

Field analysis of soils on Savary Island suggests modern soils belong to the Gleyed Eluviated Dystric Brunisols or Orthic Dystric Brunisols order. Although soil samples for laboratory analysis were not taken, field investigations of these sandy soils provide a good comparison to other sandy soils along the coast, particularly in dune environments on Calvert Island and Haida Gwaii. Interestingly, and previously undocumented in the region, mottling structures identified in Savary Island soils by Dr. Paul Sanborn during this research are thought to be “depletion mottles” caused by the selective removal of organic matter by bacteria, fungi, and actinomycetes in the soil (see Appendix A for site and pedon descriptions). The palaeosols display similar mottling structures to those present in the modern soils. Combined with the enriched iron B horizons and the beginning formation of iron cemented nodules, I suggest the palaeosols are a form of palaeo-orthic Brunisol. The palaeosols are suspected to belong to the Orthic Brunisol group due to similar properties and characteristics they share with the modern soils, but because the palaeosols lack A horizons (they have been eroded away) a definitive interpretation is difficult to make.

103

Overall, the palaeosols are weak and suggest development over a short time, likely a few hundred years. These periods of soil development and relative stability likely coincide with wetter conditions. Increasingly moist conditions would allow for increased vegetation cover, contributing to soil development. The presence of incipient palaeosols in the aeolian sequence indicated that even shorter periods of landscape stability existed during dune formation.

5.3.2. Future work

Full diagnostic soil classification still needs to be performed on Savary Island. This would include a chemical analysis of the horizons to understand the abundance of various minerals throughout them, notably carbon and iron, which are diagnostic features of soil classification. A diagnostic classification would also help in understanding the weathering history of the soils, which in turn could lead to information about precipitation and temperature, changes in salinity and pH, and soil fertility. If the same analysis were performed on the palaeosols, a more direct comparison could be drawn between the them and the modern soils on the Savary Island, and soils in other sandy environments, leading to more detailed interpretations.

The soils on Savary Island display rare features that could be the subject of future research, in particular the mottling structures (See Appendix A soil descriptions). To confirm the cause and genesis of the mottling structures, a more in-depth analysis would have to be performed. Similar mottling features have been documented previously in some Brunisols formed in aeolian parent material in Yukon (Paul Sanborn, personal communication, April 2015) and in Podzols in the Netherlands (Buurman et al. 2007). Additionally, the extent of palaeosols within the dunes should be verified. This can be done with Ground Penetrating Radar (GPR) and subsequent auguring at sites where GPR data (reflectors) suggest that palaeosols might exist. The presence of more palaeosols may allow for a more detailed association between episodes of dune stabilization and activity and regional shifts in palaeoclimate as deduced from the existing regional palaeoecological record.

104

5.4. Summary of key findings and outcomes

This study is the first to use lidar imagery to identify and interpret landforms on Savary Island, and it is the first to establish a chronology for the stabilization of parabolic dunes on the island. Key findings/outcomes include: 1) A detailed lidar-based DEM of the forested parabolic dune field on Savary Island; 2) the identification for the first time of landforms indicative of past sea level positions; 3) modern soils on Savary Island appear to by Dystric Brunisols and show what soils developing over ~5500 years in sandy parent material in a Coastal Douglas Fir biogeoclimatic zone look like; 4) that KF, but not quartz, grains are suitable for optical dating on Savary Island; 5) the post-IRIR SAR protocol offers no advantage over more traditional SAR protocols for dating KF from Savary Island; 6) optical dating of aeolian sand on Savary Island indicates that the existing aeolian cover began forming form prior to 7.69 ± 0.36 ka. The existence and character of palaeosols in these deposits indicate that the aeolian landscape experienced short (likely 100s of years long) episodes of stability between periods of activity until 5.47 ± 0.36 ka ago when large parabolic dunes stabilized as the rate of lowering RSL was slowing, and climate was transitioning from being cool and moist to cool and wet.

There is a discrepancy between radiocarbon ages from charcoal in paleosols and associated optical ages. Through extensive protocol testing it is concluded that the optical ages better reflect aeolian deposition ages and that the charcoal is reworked; however it could also be that, due to poorly understood factors, the optical ages are too young and the radiocarbon ages better constrain aeolian deposition.

105

References

Aitken, M.J. 1985. Thermoluminescence Dating. Academic Press, London.

Aitken, M.J. 1998. An Introduction to Optical Dating. Oxford University Press, Oxford.

Alley, N.F. 1979. Middle Wisconsin stratigraphy and climatic reconstruction, southern Vancouver Island, British Columbia. Quaternary Research, 11: 213-237.

Anderson, L. Abbott, M.B., Finney, B.P., and Burns, S.J. 2005. Regional atmospheric circulation change in the North Pacific during the Holocene inferred from lacustrine carbonate oxygen isotopes, Yukon Territory, Canada. Quaternary Research, 64: 21-35.

Armstrong, J.E., 1981. Post-Vashon Wisconsin Glaciation, Fraser Lowland, British Columbia. Geological Survey of Canada. Bull 322: 34 p.

Armstrong, J.E. 1984. Environmental and Engineering Applications of the Surficial Geology of the Fraser Lowland, British Columbia. Geological Survey of Canada, Paper 83-23, Vancouver, BC, Canada.

Armstrong J.E. and Clague, J.J., 1977. Two major Wisconsin lithostratigraphic units in southwest British Columbia. Canadian Journal of Earth Sciences, 14: 1471-1480.

Armstrong, J.E., Crandell, D.R., Easterbrook, D.J., and Noble, J.B. 1965. Late Pleistocene stratigraphy and chronology in southwestern British Columbia and northwestern Washington. Bull. Geol. Soc. Am. 76: 321–330.

Arnold, L.J. and Roberts, R.G. 2009. Stochastic modelling of multi-grain equivalent dose (D e) distributions: Implications for OSL dating of sediment mixtures. Quaternary Geochronology, 4: 204-230.

Auclair, M., Lamothe, M., and Huot, S. 2003. Measurement of anomalous fading for feldspar IRSL using SAR. Radiation Measurements, 37: 487-492.

Bagnold, R.A. 1941. The Physics of Wind-blown Sand and Desert Dunes. Methuen and Co., London. 265 p.

106

Barrie, J.V. and Conway, K.W. 2002. Contrasting glacial sedimentation processes and sea-level changes in two adjacent basins on the Pacific margin of Canada. In Dowdeswell, J. A. & O’Cofaigh, C. (eds.): Glacier-Influenced Sedimentation on High-Latitude Continental Margins, 181–194. Geological Society, London, Special Publication 203.

Berger, G.W. 1988. Dating Quaternary events by luminescence. In Dating Quaternary sediments. Edited by D.J. Easterbrook. Geological Society of America, Special Paper 227, pp. 13-50.

Dating Quaternary events by luminescence. In Dating Quaternary Sediments, Easterbrook DJ (ed.). Geological Society of America: Denver 13–50.

Berger, G.W., Pillans, B.J., and Palmer, A.S. 1994. Test of thermoluminescence dating of loess from New Zealand and Alaska. Quaternary Science Reviews, 13: 309- 333.

Bevington, A., Clague, J.J., Millard, T., Walker, I.J., and Geertsema, M. 2017. The diverse landscapes of Haida Gwaii. In Slaymaker, O. (Ed) Landscapes and Landforms of Western Canada. Springer International Publishing, 291-302.

Billy, J., Robin, N., Hein, C.J., Certain, R., and FitzGerald, D.M. 2015. Insight into the late Holocene sea-level changes in the NW Atlantic from a paraglacial beach- ridge plain south of Newfoundland. Geomorphology, 248: 134-146.

Blong, R. and Gillespie, R. 1978. Fluvially transported charcoal gives erroneous 14C ages for recent deposits. Nature, 271: 739-741.

Bornhold, B.D. 2008. Projected sea level changes for British Columbia in the 21st century. Fisheries and Oceans Canada, Natural Resources Canada, Province of British Columbia. Available from https://www8.nos.noaa.gov/reserves/doc/pdf/ training/projected_sea_level.pdf (Accessed May 2017)

Bornhold, B.D., Conway, K.W., and Sagayama, T. 1996. Coastal sedimentary processes, Savary Island, British Columbia - A preliminary assessment. Geological Survey of Canada, Open File 3400. 62 p.

Brennan, B.J. 2003. Beta doses to spherical grains. Radiation Measurements, 37: 299- 303.

Bulur, E. 1996. An alternative technique for optically stimulated luminescence (OSL) experiment. Radiation Measurements, 26: 701-709.

Bulur, E., Bøtter-Jensen, L., and Murray, A.S. 2000. Optically stimulated luminescence from quartz measured using the linear modulation technique. Radiation Measurements, 32: 407-411.

107

Buurman, P., Schellekens, J., Fritze, H., and Nierop, K. 2007. Selective depletion of organic matter in mottled podzol horizons. Soil Biology and Biochemistry, 39: 607-621.

Carslaw, D.C. and Ropkins, K. 2012. Openair—an R package for air quality data analysis. Environmental Modelling & Software, 27: 52-61.

Clague, J.J. 1976. Quadra Sand and its relation to the late Wisconsin glaciation of southwest British Columbia. Canadian Journal of Earth Sciences, 13: 803-815.

Clague, J.J. 1981. Late Quaternary Geology and Geochronology of British Columbia: Part 2: Summary and Discussion of Radiocarbon Dated Quaternary History. Geological Survey of Canada. Paper 80-35, 41 pp. Ottawa.

Clague, J.J. 1986. The Quaternary stratigraphic record of British Columbia—evidence for episodic sedimentation and erosion controlled by glaciation. Canadian Journal of Earth Sciences, 23: 885-894.

Clague, J.J. and Bornhold, B.D. 1980. Morphology and littoral processes of the Pacific coast of Canada. In: McCann, S.B. (Ed.) The Coastline of Canada, Geological Survey of Canada, Paper 80-10. 339-380. Ottawa.

Clague, J.J. and James, T.S. 2002. History and isostatic effects of the last ice sheet in southern British Columbia. Quaternary Science Reviews, 21: 71-87.

Clague, J.J. and Mathewes, R.W. 1989. Early Holocene thermal maximum in western North America: New evidence from Castle Peak, British Columbia. Geology, 17: 277-280.

Clague, J.J., and Ward, B. 2011. Pleistocene glaciation of British Columbia. In Quaternary Glaciation - Extent and Chronology: A Closer Look. Edited by J. Ehlers, P.L. Gibbard, and P.D. Hughes. Developments in Quaternary Science 15. Elsevier, Amsterdam. pp. 563–573.

Clague, J.J., Armstrong, J.E., and Mathews, W.H. 1980. Advance of the late Wisconsin cordilleran ice sheet in southern British Columbia since 22,000 yr BP. Quaternary Research, 13: 322-326.

Clague, J.J., Froese, D., Hutchinson, I., James, T.S., and Simon, K.M. 2005. Early growth of the last Cordilleran Ice Sheet deduced from glacio-isostatic depression in southwest British Columbia, Canada. Quaternary Research, 63: 53-59.

Clague, J.J., Harper, J.R., Hebda, R.J. and Howes, D.E. 1982. Late Quaternary sea levels and crustal movements, coastal British Columbia. Canadian Journal of Earth Sciences, 19: 597-618.

108

Clague, J.J., Hebda, R.J., and Mathewes, R.W. 1990. Stratigraphy and paleoecology of Pleistocene interstadial sediments, central British Columbia. Quaternary Research, 34: 208-226.

Clague, J.J., Luternauer, J.L., and Hebda, R.J. 1983. Sedimentary environments and postglacial history of the Fraser delta and lower Fraser Valley, British Columbia. Canadian Journal of Earth Sciences, 20, 1314-1326.

Clague, J.J., Mathewes, R., Guilbault, J., Hutchinson, I., and Ricketts, B. 1997. Pre‐ Younger Dryas resurgence of the southwestern margin of the Cordilleran Ice Sheet, British Columbia, Canada. Boreas 26, 261-278.

Clague, J.J., Tupper, D., and Webb, S. 2000. Early Holocene dune fields on the south east coast of British Columbia, unpublished report

Clague, J.J., Wohlfarth, B., Ayotte, J., Eriksson, M., Hutchinson, I., Mathewes, R.W., Walker, I.R., and Walker, L. 2004. Late Holocene environmental change at treeline in the northern Coast Mountains, British Columbia, Canada. Quaternary Science Reviews 23: 2413-2431.

Colarossi, D., Duller, G., Roberts, H., Tooth, S., and Lyons, R. 2015. Comparison of paired quartz OSL and feldspar post-IR IRSL dose distributions in poorly bleached fluvial sediments from South Africa. Quaternary Geochronology 30: 233-238.

Cooper, W.S. 1958. Coastal Sand Dunes of Oregon and Washington. Geological Society of America Memoir 72: 1-162.

Creagh, D. and Bradley, D.A. (Eds.) 2000. Radiation in Art and Archaeometry, Elsevier Science B.V., Amsterdam

Davidson-Arnott, R.G. and Law, M.N. 1996. Measurement and prediction of long-term sediment supply to coastal foredunes. Journal of Coastal Research 12: 654-663.

Devereux, B., Amable, G., and Crow, P. 2008. Visualisation of LiDAR terrain models for archaeological feature detection. Antiquity, 82: 470-479.

Dunster, K. 2000. Sand Dune for the Savary Island Land Trust. Retrieved from http://www.silts.ca/pdfs/DunsterDunesMarch2000.pdf

Dyke, A. & Prest, V. 1987. Late Wisconsinan and Holocene History of the Laurentide Ice Sheet*. Géographie physique et Quaternaire, 41(2), 237–263. doi:10.7202/032681ar

Eamer, J.B. R., Shugar, D.H., Walker, I.J., Lian, O.B., Neudorf, C.M., and Telka, A.M. 2017. A glacial readvance during the retreat of the Cordilleran Ice Sheet, British Columbia central coast. Quaternary Research, 87: 468-481.

109

Easterbrook, D.J. 1969. Pleistocene chronology of the Puget lowland and San Juan Islands, Washington. Geological Society of America Bulletin, 80: 2273-2286.

Easterbrook, D.J. 1992. Advance and retreat of cordilleran ice sheets in Washington, USA. Géographie physique et Quaternaire, 4:, 51-68.

Feathers, J. and Tunnicliffe, J. 2011. Effect of single-grain versus multi-grain aliquots in determining age for K-feldspars from southwestern British Columbia. Ancient TL, 29: 53-58.

Fedje, D.W. and Josenhans, H. 2000. Drowned forests and archaeology on the continental shelf of British Columbia, Canada. Geology, 28: 99-102.

Fedje, D. and McLaren, D. 2016. Hakai, a late-glacial to early Holocene paleoshoreline ‘sweet spot’ on the west coast of Canada. in marine ventures archaeological perspectives on human-sea relations. In: Bjerck, H.B., Breivik, H.M., Fretheim, S.E., Piana, E.L., Skar, B., Tivoli, A.M. and Zangrando, A.F.J. (Eds.), Marine Ventures: Archaeological Perspectives on Human–Sea Relations. Equinox Publishing Ltd., 70 Enterprise Drive, Bristol, CT 06010. 94-104.

Folk, R.L. 1966. A review of grain‐size parameters. Sedimentology, 6: 73-93.

Folk, R.L. 1968. Graphic presentation of size analysis data, manuscript for teaching students. In Pangea, Bremerhaven. Accessed May 2017: http://epic.awi.de/29149/1/Fol1968a.pdf pp. 41-64.

Folk, R.L. 1980. Petrology of Sedimentary Rocks. Hemphill Publishing Company, Austin, Texas, USA.

Folk, R.L. and Ward, W.C. 1957. Brazos River bar: A study in the significance of grain size parameters. Journal of Sedimentary Research, 27: 3-26

Friele, P. A. and Clague, J. J. 2002. Younger Dryas readvance in Squamish river valley, southern Coast mountains, British Columbia. Quaternary Science Reviews, 21(18): 1925-1933.

Frueh, W.T. and Lancaster, S.T. 2014. Correction of deposit ages for inherited ages of charcoal: Implications for sediment dynamics inferred from random sampling of deposits on headwater valley floors. Quaternary Science Reviews, 88: 110-124.

Fulton, R.J. and Smith, G.W. 1978. Late Pleistocene stratigraphy of south-central British Columbia. Canadian Journal of Earth Sciences, 15: 971-980.

Galbraith, R.F. 2005. Statistics for Fission Track Analysis. Chapman and Hall/CRC, London. 224 p.

Galbraith, R. 2010. On plotting OSL equivalent doses. Ancient TL, 28: 1-10.

110

Galbraith, R.F. and Roberts, R.G. 2012. Statistical aspects of equivalent dose and error calculation and display in OSL dating: an overview and some recommendations. Quaternary Geochronology, 11: 1-27.

Galbraith, R.F., Roberts, R.G., Laslett, G., Yoshida, H., and Olley, J.M. 1999. Optical dating of single and multiple grains of quartz from Jinmium rock shelter, northern Australia: Part I, experimental design and statistical models. Archaeometry, 41: 339-364.

Galbraith, R., Roberts, R., and Yoshida, H. 2005. Error variation in OSL palaeodose estimates from single aliquots of quartz: A factorial experiment. Radiation Measurements, 39: 289-307.

Galloway, J.M., Doherty, C.T., Patterson, R.T., and Roe, H.M. 2009. Postglacial vegetation and climate dynamics in the Seymour‐Belize Inlet complex, central coastal British Columbia, Canada: Palynological evidence from tiny lake. Journal of Quaternary Science, 24: 322-335.

Galloway, J.M., Patterson, R.T., Doherty, C.T., and Roe, H.M. 2007. Multi-proxy evidence of postglacial climate and environmental change at Two Frog Lake, central mainland coast of British Columbia, Canada. Journal of Paleolimnology, 38: 569-588.

Gavin, D.G. 2001. Estimation of inbuilt age in radiocarbon ages of soil charcoal for fire history studies. Radiocarbon, 43: 27-44.

Gavin, D.G., Brubaker, L.B., and Lertzman, K.P. 2003. Holocene fire history of a coastal temperate rain forest based on soil charcoal radiocarbon dates. Ecology, 84: 186-201.

Geiss, C. 2015. Mineral identification with X-ray diffraction (XRD). Accessed from:http://www.trincoll.edu/~cgeiss/GEOS_112/XRD. Septemper 2016

Guérin, G., Mercier, N., and Adamiec, G. 2011. Dose-rate conversion factors: Update. Ancient TL, 11: 5-8.

Halfen, A.F. and Johnson, W.C. 2013. A review of Great Plains dune field chronologies. Aeolian Research, 10: 135-160.

Halfen, A.F., Fredlund, G.G., and Mahan, S.A. 2010. Holocene stratigraphy and chronology of the Casper dune field, Casper, Wyoming, USA. The Holocene; 20: 773-783.

Halfen, A.F., Lancaster, N., and Wolfe, S. 2016. Interpretations and common challenges of aeolian records from North American dune fields. Quaternary International, 410: 75-95.

111

Harrington, S., Stevenson, J., and Bateman, R. 2005. Islands in the Salish Sea: A Community Atlas. TouchWood Editions. Victoria, BC

Hebda, R.J. 1983. Late-glacial and postglacial vegetation history at Bear Cove Bog, northeast Vancouver Island, brutish Columbia. Canadian Journal of Botany, 61: 3172-3192.

Hebda, R.J. 1995. British Columbia vegetation and climate history with focus on 6 ka BP. Géographie physique et Quaternaire, 49: 55-79.

Hebda, R.J., Lian, O.B., and Hicock, S.R. 2016. Olympia interstadial: Vegetation, landscape history, and paleoclimatic implications of a mid-Wisconsinan (MIS3) nonglacial sequence from southwest British Columbia, Canada. Canadian Journal of Earth Sciences, 53: 304-320.

Henderson, P. 2003. Savary Island Dune and Shoreline Study: Ecological Component. Strix Environmental Consulting. Fort Langley, BC

Hesp, P. 2002. Foredunes and blowouts: Initiation, geomorphology and dynamics. Geomorphology, 48: 245-268.

Hesp, P.A. 2011. 3.08 - dune coasts. In: Wolanski, E. and McLusky, D. (Eds.), Treatise on Estuarine and Coastal Science. Academic Press, Waltham, 193-221.

Hesp, P.A. and Walker, I.J. 2013. 11.17 coastal dunes. In: Shroder, J.F. (Ed.), Treatise on Geomorphology. Academic Press, San Diego, 328-355.

Heusser, C.J. 1985. Quaternary pollen records from the Pacific Northwest coast: Aleutians to the Oregon-California boundary. In Quaternary North American Sediments. V.M. Bryant Jr. and R.G. Holloway. American Association of Stratigraphic Palynologists Foundation, Dallas, pp. 141-165.

Heusser, L.E. 1983. Palynology and paleoecology of postglacial sediments in an anoxic basin, Saanich inlet, British Columbia. Canadian Journal of Earth Sciences, 20: 873-885.

Heusser, C.J., Heusser, L., and Peteet, D. 1985. Late-Quaternary climatic change on the American North Pacific coast. Nature, 315: 485-487.

Heusser, C., Heusser, L., and Streeter, S. 1980. Quaternary temperatures and precipitation for the north-west coast of North America. Nature, 286: 702-704.

Hickin, A. S., Lian, O. B., and Levson, V. M. 2016. Coalescence of late Wisconsinan Cordilleran and Laurentide ice sheets east of the Rocky Mountain Foothills in the Dawson Creek region, northeast British Columbia, Canada. Quaternary Research, 85(3): 409-429.

112

Hicock, S.R. and Armstrong, J.E. 1981. Coquitlam drift: A pre-Vashon Fraser glacial formation in the Fraser lowland, British Columbia. Canadian Journal of Earth Sciences, 18: 1443-1451.

Hicock, S.R. and Armstrong, J.E. 1983. Four Pleistocene formations in southwest British Columbia: Their implications for patterns of sedimentation of possible Sangamonian to early Wisconsinan age. Canadian Journal of Earth Sciences; 20: 1232-1247.

Hicock, S.R. and Lian, O.B. 1995. The Sisters Creek Formation: Pleistocene sediments representing a nonglacial interval in southwestern British Columbia at about 18 ka. Canadian Journal of Earth Sciences, 32: 758-767.

Hicock, S.R., Lian, O.B., and Mathewes, R.W. 1999. 'Bond cycles' recorded in terrestrial Pleistocene sediments of southwestern British Columbia, Canada. Journal of Quaternary Science, 14: 443-449.

Holthuijsen, L.H. 2007. Waves in Oceanic and Coastal Waters. West Nyack, GB: Cambridge University Press. Retrieved from http://www.ebrary.com.proxy.lib.sfu.ca

Hsu, S. 1974. Computing eolian sand transport from routine weather data. Proceedings 14th Conference on Coastal Engineering, ASCE, New Yorks 1619-1626.

Huntley, D.J., Godfrey-Smith, D.I., and Thewalt, M.L. 1985. Optical dating of sediments. Nature, 313(5998): 105-107

Huntley, D. and Baril, M., 1997. The K content of the K-feldspars being measured in optical dating or in thermoluminescence dating. Ancient TL, 15: 11-13.

Huntley, D. and Hancock, R. 2001. The Rb contents of the K-feldspar grains being measured in optical dating. Ancient TL, 19: 43-46.

Huntley, D., Baril, M., and Haidar, S. 2007. Tunnelling in plagioclase feldspars. Journal of Physics D: Applied Physics, 40: 900-906.

Huntley, D. and Lian, O.B. 2006. Some observations on tunnelling of trapped electrons in feldspars and their implications for optical dating. Quaternary Science Reviews, 25: 2503-2512.

Hutchinson, I. 1992. Holocene Sea Level Change in the Pacific Northwest: A Catalogue of Radiocarbon Dates and an Atlas of Regional Sea-Level Curves. Institute of Quaternary Research, Simon Fraser University, Burnaby, BC.

Hutchinson, I., James, T.S., Clague, J.J., Barrie, J.V., and Conway, K.W. 2004. Reconstruction of late Quaternary sea‐level change in southwestern British Columbia from sediments in isolation basins. Boreas, 33: 183-194.

113

Jacobs, Z. and Roberts, R.G. 2007. Advances in optically stimulated luminescence dating of individual grains of quartz from archeological deposits. Evolutionary Anthropology: Issues, News, and Reviews, 16: 210-223.

James, T.S., Henton, J.A., Leonard, L.J., Darlington, A., Forbes, D.L. and Craymer, M. 2014. Relative Sea Level Rise Projections for Canada and the Adjacent Mainland United States. Geological Survey of Canada, Open File 7737, 67 p.

James, T.S., Hutchinson, I., Vaughn Barrie, J., Conway, K.W., and Mathews, D. 2005. Relative sea-level change in the northern Strait of Georgia, British Columbia. Géographie physique et Quaternaire, 59: 113-127.

Jenny, H. 1994. Factors of Soil Formation: A System of Quantitative Pedology. McGraw- Hill, New York, 281 p.

Jol, H.M. and Bristow, C.S. 2003. GPR in sediments: Advice on data collection, basic processing and interpretation, a good practice guide. In Bristow, C.S. and Jol. H.M. (eds) Ground Penetrating Radar in Sediments. Geological Society, London, Special Publications, 211: 9-27.

Josenhans, H.W., Fedje, D.W., Conway, K.W., and Barrie, J.V. 1995. Post glacial sea levels on the western Canadian continental shelf: Evidence for rapid change, extensive subaerial exposure, and early human habitation. Marine Geology, 125: 73-94.

Josenhans, H., Fedje, D., Pienitz, R., and Southon, J. 1997. Early humans and rapidly changing Holocene sea levels in the Queen Charlotte Islands-Hecate Strait, British Columbia, Canada. Science, 277: 71-74.

Kasper-Zubillaga, J.J. and Carranza-Edwards, A. 2005. Grain size discrimination between sands of desert and coastal dunes from northwestern Mexico. Revista Mexicana De Ciencias Geológicas, 22: 383-390.

Kennedy, I. 1992. Sunny Sandy Savary. Kennell Publishing, Comox, BC.

Lacourse, T. 2005. Late Quaternary dynamics of forest vegetation on northern Vancouver Island, British Columbia, Canada. Quaternary Science Reviews, 24: 105-121.

Lamothe, M., Barré, M., Huot, S., and Ouimet, S. 2012. Natural luminescence and anomalous fading in K-feldspar. Radiation Measurements, 47: 682-687.

Lesemann, J., Brennand, T.A., Lian, O.B., and Sanborn, P. 2013. A refined understanding of the paleoenvironmental history recorded at the Okanagan Centre section, an MIS 4 stratotype, south‐central British Columbia, Canada. Journal of Quaternary Science, 28: 729-747.

114

Lian, O.B. 2013. LUMINESCENCE DATING | optical dating. In: Elias, S. (Ed.), Encyclopedia of Quaternary Sciences (Second Edition). Elsevier, Amsterdam, 653-666.

Lian, O.B. and Hickin, E.J. 1993. Late Pleistocene stratigraphy and chronology of lower Seymour valley, southwestern British Columbia. Canadian Journal of Earth Sciences, 30: 841-850.

Lian, O.B. and Hickin, E.J. 1996. Early postglacial sedimentation of lower Seymour valley, southwestern British Columbia. Géographie physique et Quaternaire, 50: 95-102.

Lian, O.B. and Roberts, R.G. 2006. Dating the Quaternary: Progress in luminescence dating of sediments. Quaternary Science Reviews, 25: 2449-2468.

Lian, O.B. and Shane, P.A. 2000. Optical dating of paleosols bracketing the widespread Rotoehu tephra, North Island, New Zealand. Quaternary Science Reviews, 19: 1649-1662.

Lian, O.B., Hu, J., Huntley, D., and Hicock, S.R. 1995. Optical dating studies of Quaternary organic-rich sediments from southwestern British Columbia and northwestern Washington state. Canadian Journal of Earth Sciences, 32: 1194- 1207.

Lian, O.B., Mathewes, R.W., and Hicock, S.R. 2001. Palaeoenvironmental reconstruction of the Port Moody interstade, a nonglacial interval in southwestern British Columbia at about 18 000 14C years BP. Canadian Journal of Earth Sciences, 38: 943-952.

Long, H., Haberzettl, T., Tsukamoto, S., Shen, J., Kasper, T., Daut, G., Zhu, L., Mäusbacher, R., and Frechen, M. 2015. Luminescence dating of lacustrine sediments from Tangra Yumco (southern Tibetan Plateau) using post‐IR IRSL signals from polymineral grains. Boreas, 44: 139-152.

Loope, D.B. and Swinehart, J.B. 2000. Thinking like a dune field: Geologic history in the Nebraska Sand Hills. Great Plains Research, 10: 5-35.

MacKenzie, W.S. and Adams, A.E. 1994. A Color Atlas of Rocks and Minerals in Thin Section. Wiley Publishers, New York

Mandryk, C.A.S., Josenhans, H., Fedje, D.W., and Mathewes, R.W. 2001. Late Quaternary paleoenvironments of northwestern North America: Implications for inland versus coastal migration routes. Quaternary Science Reviews, 20: 301- 314.

115

Mann, D.H. and Hamilton, T.D. 1995. Late Pleistocene and Holocene paleoenvironments of the North Pacific coast. Quaternary Science Reviews, 14: 449-471.

Massey, N.W.D., MacIntyre, D.G., Desjardins, P.J., and Cooney, R.T. 2005. Digital Geology Map of British Columbia: Tile NM10 Southwest B.C. B.C. Ministry of Energy, Mines and Petroleum Resources, Geological Survey Branch, GeoFile, 2005–1

Mathewes, R.W. 1973. A palynological study of postglacial vegetation changes in the University Research Forest, southwestern British Columbia. Canadian Journal of Botany, 51: 2085-2103.

Mathewes, R.W. 1985. Paleobotanical evidence for climatic change in southern British Columbia during late-glacial and Holocene time. In Climate change in Canada 5: critical periods in the Quaternary climatic history of northern North America. Syllogeus, 55: 397-422.

Mathewes, R.W. 1993. Evidence for Younger Dryas-age cooling on the North Pacific coast of America. Quaternary Science Reviews, 12: 321-331.

Mathewes, R.W. and Heusser, L.E. 1981. A 12 000 year palynological record of temperature and precipitation trends in southwestern British Columbia. Canadian Journal of Botany, 59: 707-710.

Mathews, W.H., Fyles, J., and Nasmith, H. 1970. Postglacial crustal movements in southwestern British Columbia and adjacent Washington state. Canadian Journal of Earth Sciences, 7: 690-702. Mathewes, R.W., Lian, O.B., Clague, J.J., and Huntley, M. 2015. Early Wisconsinan (MIS 4) glaciation on Haida Gwaii, British Columbia, and implication for biological refugia. Canadian Journal of Earth Sciences, 52: 939–951.

McCammon, J.W. 1977. Surficial Geology and Sand and Gravel Deposits of Sunshine Coast, Powell River, and Campbell River Areas. BC Ministry of Mines and Petroleum Resources, Bulletin 65.

Monger, J.W.H., Souther, J.G., and Gabrielse, H. 1972. Evolution of the Canadian Cordillera; a plate-tectonic model. American Journal of Science, 272: 577-602.

Murray, A.S. and Wintle, A.G. 2000. Luminescence dating of quartz using an improved single-aliquot regenerative-dose protocol. Radiation Measurements, 32: 57-73.

Murray, A.S. and Wintle, A.G. 2003. The single aliquot regenerative dose protocol: Potential for improvements in reliability. Radiation Measurements, 37: 377-381.

116

Neudorf, C.M., Lian, O.B., Walker, I.J., Shugar, D.H., Eamer, J.B. and Griffin, L.C., 2015. Toward a luminescence chronology for coastal dune and beach deposits on Calvert Island, British Columbia central coast, Canada. Quaternary Geochronology, 30: 275-281.

Neudorf, C.M., Smith, N., Lepofsky, D., Toniello, G., and Lian, O.B. 2017. Between a rock and a soft place: Using optical ages to date ancient clam gardens on the Pacific Northwest. PloS One, 12: e0171775.

Nuszdorfer, F., Klinka, K., and Demarchi, D. 1991. Coastal Douglas-fir zone. Ecosystems of British Columbia. D.Meidinger and J.Pojar (Editors).Research Branch, BC Ministry of Forests, Victoria, BC Special Report Series 81-93.

Ollerhead, J., Huntley, D.J., and Berger, G.W. 1994. Luminescence dating of sediments from Buctouche Spit, New Brunswick. Canadian Journal of Earth Sciences, 31: 523-531.

Ollerhead, J., Huntley, D., Nelson, A.R., and Kelsey, H.M. 2001. Optical dating of tsunami-laid sand from an Oregon coastal lake. Quaternary Science Reviews, 20: 1915-1926.

Olley, J., Caitcheon, G., and Murray, A. 1998. The distribution of apparent dose as determined by optically stimulated luminescence in small aliquots of fluvial quartz: Implications for dating young sediments. Quaternary Science Reviews, 17: 1033-1040.

Page, N., Lilley, P., Walker, I.J., and Vennesland, R.G. 2011. Status report on coastal sand ecosystems in British Columbia. Report prepared for the Coastal Sand Ecosystems Recovery Team, vii + 83 p.

Pellatt, M.G., Hebda, R.J., and Mathewes, R.W. 2001. High-resolution Holocene vegetation history and climate from hole 1034B, ODP leg 169S, Saanich Inlet, Canada. Marine Geology, 174: 211-222.

Pellatt, M.G., Mathewes, R.W., and Clague, J.J. 2002. Implications of a late-glacial pollen record for the glacial and climatic history of the Fraser lLowland, British Columbia. Palaeogeography, Palaeoclimatology, Palaeoecology, 180: 147-157.

Porter, S.C. and Swanson, T.W. 1998. Radiocarbon age constraints on rates of advance and retreat of the Puget Lobe of the Cordilleran Ice Sheet during the last glaciation. Quaternary Research, 50: 205-213.

Prescott, J.R. and Hutton, J.T. 1994. Cosmic ray contributions to dose rates for luminescence and ESR dating: Large depths and long term time variations. Radiation Measurements, 23: 497-500.

117

Prescott, J.R. and Hutton, J. 1995. Environmental dose rates and radioactive disequilibrium from some Australian luminescence dating sites. Quaternary Science Reviews, 14: 439-448.

Preusser, F., Chithambo, M.L., Götte, T., Martini, M., Ramseyer, K., Sendezera, E.J., Susino, G.J., and Wintle, A.G. 2009. Quartz as a natural luminescence dosimeter. Earth-Science Reviews, 97: 184-214.

Pye, K. 1983. Coastal dunes. Progress in Physical Geography, 7: 531-557.

Pye, K. and Tsoar, H. 2009. Aeolian Sand and Sand Dunes. Springer, Berlin, Heidelberg.

Roberts, R.G. and Lian, O.B. 2015. Illuminating the past. Nature, 520: 438-429.

Roberts, R.G., Jacobs, Z., Li, B., Jankowski, N.R., Cunningham, A.C., and Rosenfeld, A.B. 2015. Optical dating in archaeology: Thirty years in retrospect and grand challenges for the future. Journal of Archaeological Science, 56: 41-60.

Roberts, R., Yoshida, H., Galbraith, R., Laslett, G., Jones, R., and Smith, M. 1998. Single-aliquot and single-grain optical dating confirm thermoluminescence age estimates at Malakunanja II rock shelter in northern Australia. Ancient TL, 16: 19- 24.

Rovere, A., Stocchi, P., & Vacchi, M. 2016. Eustatic and Relative Sea Level Changes. Current Climate Change Reports, 2(4): 221-231.

Ryder, J.M. and Clague, J.J. 1989. Quaternary geology of Canada and Greenland. In: Fulton, R. (Ed.), British Columbia (Quaternary Stratigraphy and History, Cordilleran Ice Sheet). Geological Survey of Canada, Geology of Canada, no. 1 (also Geological Society of America, The Geology of North America, v. K-l), 48- 58.

Sadler, K. 2000. Bryophytes of Savary Island D.L. 1375. Letter to Liz Webster, Savary Island Land Trust. September 14, 2000.

Saunders, I.R. 1985. Late Quaternary Geology and Geomorphology of the Chilliwack River Valley, British Columbia. British Columbia. M.Sc. Thesis, Simon Fraser University, Vancouver, B.C.

Schaetzl, R. and Anderson, S. 2005. Soils: Genesis and Geomorphology. Cambridge University Press, New York.

Seville, J. P. and Wu, C. Y. 2016. Particle Technology and Engineering: An Engineer's Guide to Particles and Powders: Fundamentals and Computational Approaches. Butterworth-Heinemann. Oxford UK, 61 p.

118

Shugar, D.H., Walker, I.J., Lian, O.B., Eamer, J.B.R., Neudorf, C., McLaren, D., and Fedje, D. 2014. Post-glacial sea-level change along the Pacific Coast of North America. Quaternary Science Reviews, 97: 170-192.

Smith, D., Regehr, D., and Gerath, R. 2003. Savary Island dune and shoreline study: Report to Powell River Regional District. Thurber Engineering Ltd. File: 14-197-0 Vancouver, BC.

Soil Classification Working Group. 1998. The Canadian System of Soil Classification. Agriculture and Canada Agri-Foods Canada, Ottawa: NRC Research Press, Ottawa, ON.

Thompson, C.H. 1981. Podzol chronosequences on coastal dunes of eastern Australia. Nature 291, 59-61.

Thomsen, K.J., Murray, A., Jain, M., and Bøtter-Jensen, L. 2008. Laboratory fading rates of various luminescence signals from feldspar-rich sediment extracts. Radiation Measurements 43, 1474-1486.

Thomson, R.E. 1981. Oceanography of the British Columbia Coast. Canadian Special Publications in Fishery and Aquatic Sciences 56: 1-291.

Trenberth, K.E. and Hurrell, J.W. 1994. Decadal atmosphere-ocean variations in the Pacific. Climate Dynamics, 9: 303-319.

Tupper, D. 1996. A preliminary assessment of the groundwater resources of Savary Island, British Columbia for Powell River Regional District and Savary Island Planning Committee British Columbia, Ministry of Environment, Lands and Parks, British Columbia Ministry of Health, Victoria, BC..

Walker, I.R. and Pellatt, M.G. 2003. Climate change in coastal British Columbia—A paleoenvironmental perspective. Canadian Water Resources Journal, 28: 531- 566.

Walker, M., Johnsen, S., Rasmussen, S.O., Popp, T., Steffensen, J., Gibbard, P., Hoek, W., Lowe, J., Andrews, J., and Björck, S. 2009. Formal definition and dating of the GSSP (global stratotype section and point) for the base of the Holocene using the Greenland NGRIP ice core, and selected auxiliary records. Journal of Quaternary Science, 24: 3-17.

Wallinga, J., Murray, A., and Wintle, A. 2000. The single-aliquot regenerative-dose (SAR) protocol applied to coarse-grain feldspar. Radiation Measurements, 32: 529-533.

Wang, X. 2008: Accounting for autocorrelation in detecting mean-shifts in climate data series using the penalized maximal t or F test. Journal of Applied Meteorology and Climatology, 47 (9): 2423-2444.

119

Wang, T., Hamann, A., Spittlehouse, D.L., and Murdock, T.Q. 2012. ClimateWNA-high- resolution spatial climate data for western North America. Journal of Applied Meteorology and Climatology, 51: 16-29.

Ward, B.C. and Thomson, B. 2004. Late Pleistocene stratigraphy and chronology of lower Chehalis River valley, southwestern British Columbia: Evidence for a restricted Coquitlam Stade. Canadian Journal of Earth Sciences, 41: 881-895.

Wintle, A.G. 1973. Anomalous fading of thermo-luminescence in mineral samples. Nature, 245: 143-144.

Wintle, A.G. 1997. Luminescence dating: Laboratory procedures and protocols. Radiation Measurements, 27: 769-817.

Wintle, A.G. 2008. Luminescence dating: Where it has been and where it is going. Boreas, 37: 471-482.

Wintle, A.G. and Murray, A.S. 2006. A review of quartz optically stimulated luminescence characteristics and their relevance in single-aliquot regeneration dating protocols. Radiation Measurements, 41: 369-391.

Wolfe, S.A., Huntley, D.J., David, P.P., Ollerhead, J., Sauchyn, D.J., and MacDonald, G.M. 2001. Late 18th century drought-induced sand dune activity, Great Sand Hills, Saskatchewan. Canadian Journal of Earth Sciences, 38: 105-117.

Wolfe, S.A., Huntley, D.J., and Ollerhead, J. 2002. Optical dating of modern and late Holocene dune sands in the Brandon Sand Hills, southwestern Manitoba. Géographie physique et Quaternaire, 56: 203-214.

Wolfe, S.A., Walker, I., and Huntley, D. 2008. Holocene coastal reconstruction, Naikoon peninsula, Queen Charlotte Islands, British Columbia. Current Research (Online) no. 2008-12, Geological Survey of Canada, Ottawa.

120

Appendix A Site and pedon descriptions

Site locations are shown in Figure 3.2. Soil classifications are based on field observations only. To correctly classify soils to the subgroup one must have chemical data from the horizons, in particular measurements of the pyrophosphate-extractable Al and Fe in the B horizons. Paleosols lacked enough structure to be assigned to a soil order although similarities can be drawn between the modern soils and the palaeosols.

Site 1 Suggested soil classification: Insipient paleosol

Location: Savary Island, side of eroding cliff, west of site 2 Coordinates: 49°56'13.52"N 124°49'6.06"W Elevation: 22 m asl Aspect: SW Slope: 38 ° Position: cliff face Parent Material: Clean aeolian sand Vegetation: Dune vegetation, sage brush. Comments: Optical samples SIDS-8 collected 480 cm below surface, ~ 70 cm below top of insipient palaeosol formation. Charcoal radiocarbon dated from top of B horizon to be 8245 ± 20 14C yr BP (9,294- 9,130 cal yr BP). Section diagram of site 1 depicted in Figure 4.5

Table A1. Insipient palaeosol 2 description.

Horizon Depth (cm) Description C >30-0 Pale red (2.5 YR 7/2); coarse sand; very friable; abrupt wavy boundary Bm 0-12 Reddish gray (2.5 YR 7/1.5); fine sand; few faint light orange (2.5Y 7/3) mottles, very friable; abrupt wavy boundary; 6 – 12 cm thick BC 12-57 Light yellowish brown (2.5Y 6/4) with few, medium, faint light gray (2.5Y 7/2) mottles; sand; very friable; sand; single grains; gradual smooth boundary C 57-120+ Medium light gray (2.5Y 7/2); sand; very friable; horizontally, laminated bedding with slight dip to the west Note: no samples were sent for laboratory analysis.

121

122

Site 2 Suggested soil Classification: Palaeosol

Pedon: SIDS-1 – Paleosol 1

Location: Savary Island, side of eroding cliff, inside of parabolic dune arm Coordinates: 49°56'13.52"N 124°49'6.06"W Elevation: 17.5 m asl Aspect: SW Slope: 38 ° Position: cliff face Parent Material: Clean aeolian sand Vegetation: Dune vegetation, sage brush Comments: Optical samples SIDS-1 and SIDS-1C were collected 680 cm below surface, ~ 40 cm below top of paleosol formation. Charcoal radiocarbon dated from top of B horizon to be 7400 +25 14C yr BP (8,316 – 8,176 cal yr BP). Paleosol dipping into the slope at 14° with an apparent dip of 8.5°.

Table A2. Description of palaeosol 1.

Horizon Depth (cm) Description C I >200-10 Pale red (2.5 YR 7/2); coarse sand; very friable; abrupt wavy boundary C II 10-0 Reddish gray (2.5 YR 7/1.5); fine sand; few faint light gray (2.5Y 7/3) mottles, very friable; abrupt wavy boundary; 6 – 12 cm thick II Ahb 0-5 Yellowish brown (10 YR 5/4); sand; single grains; abrupt broken boundary; presence of charcoal fragments II Bfcb 0-10 Reddish brown (5 YR 4/4); sand; friable; firm; moderately cemented; abrupt broken boundary; 0 to 3 cm thick II Bg 20-50 Mixed matrix of more common weak red (2.5Y 5/4) less common light gray (2.5Y 7/1); sand; common coarse mottles dark yellowish brown (10YR 4/6); friable; gradual smooth boundary; 30 to 40 cm thick BC 50-70+ Light gray (2.5 Y 6/3); sand; very friable Note: no samples were sent for laboratory analysis.

123

124

Site No: 2 Suggested soil classification: Palaeosol

Pedon: SIDS-02 – Palaeosol 2 Location: Savary Island, side of eroding cliff, inside of parabolic dune arm, below SIDS-01 Coordinates: 49°56'13.52"N 124°49'6.06"W Elevation: 16.5 Aspect: SW Slope: 38 ° Position: Crest Parent Material: Clean aeolian sand Vegetation: Dune vegetation, sage brush. Comments: Optical samples SIDS-2 collected 920 cm below surface, ~ 45 cm below top of paleosol formation. Charcoal fragments from top of B horizon radiocarbon dated to 8,550 ± 45 14C yr BP (10,165 – 9,530).

Table A3. Description of palaeosol 2.

Horizon Depth (cm) Description Cgj +30-10 Pinkish grey (2.5 YR 7/2) matrix with faint pale yellow (2.5Y 7/4) mottles; sand; very friable; clear wavy boundary; > 20 cm thick Cg 10-0 Light gray (2.5YR 7/1) matrix with common, medium, yellowish brown (10 YR 5/6) mottles; fine sand; abundant fine, med, coarse, horizontal and oblique roots; some hydrologic discontinuity – water hang up on enough few roots – mottles are moderately cemented; matrix is friable. wavy clear boundary; 0-15 cm thick II Bg 0-10 Olive yellow (2.5Y 6/6) with many, coarse, yellowish brown (10YR 5/8) mottles - almost forming layer; friable; few fine, very fine roots; 2 – 5 cm thick II Bgjb 10-80 Light yellowish brown (2.5Y 6/4) with common, medium light gray (2.5Y 7/2) mottles; fine sand; friable; no roots, concentrated above contact 80 + Gravel lag

125

126

Site No: 2 Suggested soil classification: Insipient palaeosol

Pedon SIDS3 & SIDS4 Location: Slightly west of SIDS1 and SIDS2 along same cliff face Coordinates: 49°56'13.84"N 124°49'6.60"W Elevation: Aspect: SW Slope: 38 ° Position: cliff face Parent Material: Clean aeolian sand Vegetation: Dune vegetation, sage brush Comments: Optical samples SIDS 3 collected 296 cm below surface. SIDS 4 was collected 1.5 m directly above SIDS 3. Insipient paleosol located between SIDS3 and SIDS-4 was too weak too weak be described. Charcoal samples were collected from sand in June of 2014 and sent to Alice Telka for cleaning and preparation prior to radiocarbon dating. Upon analysis sample was determined unsuitable for radiocarbon dating. In fact, there was no charcoal and consisted mostly of modern ectomycorrhizae and rhizome fragmented with ‘fuzzy’ fungal stringers attached. According to Telka’s analysis the sample contained some darkened brown peat/soil ‘peds’, crumbly with clumps of ectomycorrhizae surrounding the small peds (Figure X).

127

128

SIDS5 Suggested soil classification: Dystric Brunisol

Pedon: SIDH-01 Location: Head of vegetated parabolic dune field on Savary Island, BC Coordinates: 49° 56' 35.2299" N 124° 48' 59.3101" W Elevation: 30 m Aspect: SW Slope: ~2° Position: Crest Parent Material: Clean aeolian sand Vegetation: Dense salal understory, forest cover Comments: OSL sample SIDS-05 taken 1.4 m below surface

Table A4. Soil pit description of the dune head labeled SIDS5 in Figure 3.2.

Horizon Depth (cm) Description L 4-3 Needle litter and feather moss. Single particle; loose; mossy; matted structure held together by fungal mycelia Fm 3-0 Dark reddish brown (5YR 3/2 m) Partially decomposed organic matter; Compact matted; abundant very fine, fine roots; abundant mycelia; abrupt smooth boundary; 2-4 cm thick Ae 0-5 Gray (10YR 6/1)very friable; loamy sand; 0% coarse fragments; wavy abrupt contact; abundant fine, medium, coarse roots; 0-5 cm thick Bfj 5-40 Dark yellowish brown (10YR 3/6) with lighter Dark yellowish brown (10YR 4/4) areas; very friable; moderate medium sub-angular blocky; sand to loamy sand; abundant fine, medium roots, mostly oblique; gradual wavy lower contact; 30-40 cm thick Bm 40-95 Yellowish brown (10YR 5/6) with lighter coloured pale brown (10YR 6/3) splotched that cover ~50% of horizon; single grain; sand; no roots, no structure; clear boundary;48 to 50 cm thick BC 95-130+ Pale red (2.5YR 7/3); sand; no roots; no structure; very friable; single grain

129

130

SIDS6 Suggested soil classification: Dystric Brunisol

Pedon: SIDH-02 Location: Head of vegetated parabolic dune field on Savary Island, BC. Coordinates: 49° 56' 18.7080" N 124° 48' 57.4320" W Elevation: 23m Aspect: n/a Slope: flat Position: Crest Parent Material: Clean aeolian sand Vegetation: Mossy clearing in dense forest, surrounded by cedar, hemlock and Douglas fir Comments: Optical samples SIDS-6 and SIDS-6B were collected 100 cm below surface. Pedon at this site was not described in detail due to lack of daylight. Soil was interpreted to be very similar to that of SIDS-5

Table A5. Horizon descriptions of SIDH-02, labeled SIDS-6 in Figure 2.1.

Horizon Depth Description (based on photo and brief field observations) (cm) L 3-2 Feather moss. Single particle; loose; mossy; matted structure held together by fungal mycelia Fm 2-0 Dark grayish brown (10&R 4/2); 1-2 cm thick Ae 0-2 Gray (10YR 6/1) very friable; loamy sand; wavy abrupt contact; abundant fine, medium, coarse roots; 0-2 cm thick Bf 0-10 Dark yellowish brown (10YR 3/6); very friable; sand to loamy sand; abundant fine, medium roots; gradual wavy lower contact; 30-40 cm thick Bfj 10-30 Brownish yellow (10YR 6/6) with lighter; single grain; sand; plentiful fine and medium roots, no structure; clear boundary; 10 – 20 cm thick Bm 30-70 Strong brown (7.5YR 5/6) with abundant coarse mottles; sand; no roots; no structure; very friable; single grain BC 70-100+ Pale red (2.5YR 7/3); sand; no roots; no structure; very friable; single grain

131

132

SIDS7 Suggested soil classification: Dystric Brunisol

Pedon: SIDH-03 Location: Head of vegetated parabolic dune field on Savary Island, BC Coordinates: 49° 56' 23.1497" N 124° 48' 43.8465" Elevation: 27 m Aspect: n/a Slope: flat Position: Crest Parent Material: Clean aeolian sand Vegetation: Dense forest cover but in close proximity to cleared road. Immature pine trees and cedar seedlings, salal predominates low lying vegetation Comments: Optical samples SIDS-7 was collected 120 cm below surface.

Table A6. Soil horizon descriptions of SIDH-03 labeled SIDS-7 in Figure 2.1.

Horizon Depth (cm) Description L 12-11 Needle litter; single particle; loose; 1 – 2 cm thick F 11-10 Dark brown (10 YR 3/3); roots; 1 – 2 cm thick Ha 10-0 Very dark brown (7.5 YR 2.5/2) abundant very fine, fine and medium roots; structure; abrupt wavy boundary; 1 – 2 cm thick Ae 0-5 Light grey (10 YR 7/2 D); Loamy sand; Hydrophobic; friable; plentiful very fine, fine roots; single grain; boundary; 5 – 9 cm thick Bm 5-35 Grayish brown (10 YR 5/4 crushed colour, dry); loamy sand – slightly hydrophobic; single grain; abundant fine, very fine roots; single grain; wavy clear boundary; 1 – 2 cm thick. Lots of roots removing water from horizon causing increased dryness and colour difference Bfj 35-70 Yellowish Brown (10 YR 5/6); loamy sand; weak fine sub angular blocky; many, coarse brown (10 YR 5/3) mottles; few fine, very fine roots; friable; gradual boundary; uniform thickness BC 70-100+ Weak red (2.5YR 6/3); sand; single grain; friable

133

134

SIDS9 Suggested soil classification: Dystric Brunisol

Pedon: SIDH-04 Location: Head of vegetated parabolic dune field on Savary Island, BC Coordinates: 49° 56' 25.1417" N, 124° 49' 07.3604" W Elevation: 23 m Aspect: SE Slope: 1° Position: crest Parent Material: Clean aeolian sand Vegetation: Dense forest cover but sparse understory vegetation. Lots of fallen branches, twigs, moss covered logs. Primarily small hemlock Comments: Optical samples SIDS-9 was collected 105 cm below surface

Horizon Depth (cm) Description L 5-0 Needle litter; single particle; loose; 1 – 2 cm thick Ae 0-3 Light grey (10 YR 7/2 D); loamy sand; friable; plentiful very fine, fine roots; single grain; boundary; 0-3 cm thick Bm 3-38 Grayish brown (10 YR 5/4 crushed colour, dry); loamy sand single grain; abundant fine, very fine roots; single grain; wavy clear boundary; 1 – 2 cm thick BC 38-78 Yellowish Brown (10 YR 5/6); loamy sand; many, coarse brown (10 YR 5/3) mottles; few fine, very fine roots; friable; gradual boundary; uniform thickness C 78-90+ Weak red (2.5YR 6/3); sand; single grain; friable

135

136

SIDS10 Suggested soil classification: Dystric Brunisol

Pedon: SIDS-10 Location: Head of vegetated parabolic dune field on Savary Island, BC Coordinates: 49° 56' 19.6337" N, 124° 49' 10.6784" W Elevation: 33m Aspect: n/a Slope: flat Position: Crest Parent Material: Clean aeolian sand Vegetation: Dense forest cover but sparse understory vegetation. Lots of fallen branches, twigs, moss covered logs. Primarily small hemlock Comments: Optical samples SIDS-10 was collected 140 cm below surface

Table A8. Horizon descriptions of soil pit dug where SIDS-10 was collected.

Horizon Depth (cm) Description L 22-23 Needle litter; single particle; some moss; loose; 1 – 2 cm thick F 22-0 Dark blackish brown; partially decomposed Ae 0-28 Light gray; loamy sand; few fine, very fine roots Bm 28-98 Yellowish brown; loamy sand; few, brown mottles; few fine, very fine roots; many medium roots; friable; gradual boundary; uniform thickness BC 98-120+ Yellowish gray; sand; single grain; friable

137

138

SIDS11 Suggested soil classification: Dystric Brunisol

Pedon: Location: Head of small parabolic dune set inside of larger parabolic dune on Savary Island, BC Coordinates: 49° 56' 26.2396" N, 124° 48' 38.0925" W Elevation: 18 m Aspect: n/a Slope: Flat Position: Crest Parent Material: Clean aeolian sand Vegetation: Dense forest cover but sparse understory vegetation. Lots of fallen branches, twigs, moss covered logs. Primarily small hemlock Comments: Optical samples SIDS-11 was collected 135 cm below surface

Table A9. Horizon descriptions of soil pit dug where SIDS-11 was collected.

Horizon Depth (cm) Description L 12-13 Needle litter; single particle; some moss; loose; 1 – 2 cm thick F 12-0 Dark brown; partially decomposed; plentiful medium and few fine roots Ae 0-5 Light gray; loamy sand; plentiful fine, very fine roots Bm 5-82 Yellowish Brown; loamy sand; few, light grey mottles; plentiful fine, very fine roots; friable; gradual boundary; uniform thickness C 82-137+ Yellowish gray; sand; single grain; friable

139

140

Swale01 Suggested soil classification: Dystric Brunisol

Pedon: SWALE-01 Location: Low lying area between parabolic dunes, southeast of SIDS-9 Coordinates: 49° 56' 23.0957" N, 124° 49' 08.0084" W Elevation: 18.1 m Aspect: n/a Slope: Flat Position: swale Parent Material: Clean aeolian sand Vegetation: Forest composed of thin hemlock and cedars creating a sparse canopy. No significant understory vegetation aside from mosses. Lots of fallen branches and twigs on the forest floor Comments: As depth increases towards the bottom of the pit clasts become more glacially shaped although there are fewer stones at the bottom of the pit

Table A10. Horizon descriptions of SWALE-01.

Horizon Depth (cm) Description L 3-2 Needle litter; moss; loose; 1 – 2 cm thick F 2-0 Dark brown; partially decomposed; and few medium roots Ae 0-8 Light gray; loamy sand; plentiful fine to medium roots Bm 8-50 Yellowish Brown; loamy sand; abundant fine, very fine roots; friable; high concentration of sub-angular gravel and cobbles in one area ~80%, 10 cm thick; 35 to 40 cm thick BC 50-70+ Yellowish brown, sand; single grain; firm

141

142

Swale02 Suggested soil classification: Dystric Brunisol

Pedon: SWALE-02 Location: Low lying area between dunes, southeast of SIDS-5 and northwest of SIDS-11 Coordinates: 49° 56' 27.1576" N, 124° 48' 42.9945" W Elevation: 21.1 m Aspect: n/a Slope: Flat Position: swale Parent Material: Coarse aeolian sand with interbedded gravels Vegetation: Dense salal at least 1 m high, fallen logs nearby. Comments: Location is adjacent to old logging road, remnants of glass and old telephone wire were found in the uppermost part of the soil. The soil does not look to be disturbed any further

Table A11. Horizon descriptions of SWALE-02.

Horizon Depth (cm) Description L 11-9 Needle litter; moss; loose; clear, wavy boundary 2-3 cm thick Ahe 9-0 Pale grayish brown; and few fine roots; sand; friable; clear, smooth boundary; 9 to 2 cm thick. Ae 0-6 Light gray; loamy sand; few medium roots; clear, smooth boundary; 5 to 2 cm thick Bm 6-61 Yellowish brown; loamy sand; abundant fine, very fine roots; friable; 30-40% sub-angular gravel and cobbles, concentrated near bottom of horizon; gradual smooth boundary; 55 to 60 cm thick BC 61-70+ Grayish orange; sand; single grain; firm

143

144

Swale03 Suggested soil classification: Dystric Brunisol

Pedon: SWALE-03 Location: In the dune swale of SIDS 9. SE 600 m. Coordinates: 49° 56' 15.971'' N, 124° 48' 45.72'' W Elevation: 9.9 m Aspect: n/a Slope: Flat Position: Swale Parent Material: Aeolain sand and/or beach sand Vegetation: Forest canopy with salal understory. Exposed coastline lies 280 m southeast Comments: Beach sand was reached at ~2 m below the modern surface. An optical dating sample (SIBS-01) was collected for interest sake.

Table A12. Horizon descriptions of SWALE-03.

Horizon Depth (cm) Description L 9-7 Needle litter; feather moss; pine cones; clear, smooth boundary 2-3 cm thick Fm 7-0 Dark rich brown; few fine roots; matted; mycelia present; abrupt wavy boundary; 5 to 8 cm thick Ae 0-9 Light gray; loamy sand; few medium roots; abrupt, wavy boundary; 5 to 2 cm thick; very dry Bh 9-24 Dark reddish brown; few fine and very fine roots; sand; gradual smooth boundary; 10-15 cm thick Bm 24-76 Light grey with reddish orange mottles; sand; friable; gradual smooth boundary; 55 to 60 cm thick; Large depletion mottle of light grey in colour, suspected root cast BC 76-126+ Gray with interbedded dark grayish black heavy minerals; sand; single grain; firm; noticeable horizontal bedding of heavy minerals indicative of beach sands; water table is reached at this depth

145

146

Swale04 Suggested soil classification: Dystric Brunisol

Pedon: SWALE-04 Location: In the lowlying portion between two dune ridges. West of swale-03; the southernmost site Coordinates: 49° 56' 11.8217" N, 124° 48' 46.2465" W Elevation: 10.5 m Aspect: n/a Slope: Flat Position: Swale Parent Material: Clean aeolian sand Vegetation: Sparse salal understory, surrounded my large Douglas fir that are slowly dying Comments:

Table A13, Horizon descriptions of SWALE-04.

Horizon Depth (cm) Description L 4-0 Needle litter; moss; loose; abrupt, smooth boundary 2-4 cm thick Ae 0-25 Light gray; loamy sand; few fine and very fine roots; clear, smooth boundary; 20 to 25 cm thick Bm 25-70 Reddish brown; loamy sand; abundant fine, very fine roots; friable; many fine and medium light grey mottles; clear, smooth boundary; 45-50 cm thick BC 70-85+ Grayish orange; sand; single grain; friable

147

148

Appendix B Lidar

Acquisition of lidar data from Savary Island was facilitated by Dr. Brian Menounos (UNBC). Data were collected on August 7, 2014, following a pre- determined flight plan with specific scanner specifications (Table B1). There were no data gaps bordering the shore and it was noted that due to low reflectance of the scanner off of water, some instances show no returns. The lidar technician, Rob Vogt (UNBC), provided a report of data analysis performed prior to delivery. This included solving a precise trajectory for the plane using PosPac MMS. A calibration line across the base station both before and after the survey was used to aid in forward and reverse processing of the trajectory with a position accuracy of the trajectory at ±5 cm. Lidar points were processed in CSRS NAD83 UTM Zone 9, but use ellipsoid heights. Conversion to CGVD28 was completed after DEM creation using GPS-H, but attempts at converting the point cloud to geoid heights resulted in a program crash.

Derek Heathfield from the Hakai Institute then converted all point cloud data, classified and raw, from ellipsoidal heights to Orthometric using Merrick MARS 7 GeoCalc tool. The coordinate system remained the same as delivered (NAD83(CSRS) UTMz10), and the vertical datum was converted from ellipsoidal heights to CGVD28 using the HTv2.0 geoid. This converted point cloud was used to produce new DEMs generated using Quick Terrain Modeller, using Adaptive Triangulation (Min-Z) with 1- m grid spacing. The DEMs developed by Derek Heathfield were then then used by myself (Libby Griffin) for geomorphic interpretation and mapping.

149

Table B1. Scanner specification used during flight, note variations in flying speed and height did occur. Processing of flight lines was completed in RiProcess. Location Savary Island

Scanner type Riegl VQ-580 Max. scan angle ± 30 ° Flying speed 225 km/hr Flying height 475 m Pulse rep. rate 380 kHz Flight line overlap 60 %

Table B2. The average point density per flight line is calculated from random samples by dividing the total number of data points by the total non- water acquisition area. Savary Area of coverage 8.75 km2 Density/flight line 9.4pts/m2 Average density 12.5 pts/m2 Ground point % 17%

150

Appendix C XRF and XRD

X-ray diffraction (XRD) relies on the fact that visible light and x-rays have a known wavelength. When X-rays interact with the atoms that make up a crystal, it is split into various waves to create a distinct refractive pattern. This refractive pattern is unique to individual mineral composition and therefor can be used to identify which minerals are making up a specific samples (Geiss 2015). Selected samples were analyzed using X-Ray diffraction at the Saskatchewan Research Council (SRC) Advanced Microanalysis Centre. Dried samples were powdered using a rock mill prior to being sent for analysis to ensure that many randomly oriented crystals were available for interaction with the beam. The beam shines wavelengths of light on the crystal, which leads to a characteristic refraction pattern, the results of which are then compared to known refractive patterns. These patterns rely on the way ion arrangements occur within a crystal structure and chemical structure. Because of this, we can differentiate the presence of different mineral subgroups like plagioclase and alkali feldspars (potassium-rich) (Geiss 2015).

X-Ray fluorescence (XRF) was also performed at SRC Advanced Microanalysis Centre on subsamples of the same samples sent for XRD. Since each element has an electronic orbital of characteristic energy, by bombarding the electrons with radiation, transitions occur between valence electrons. These transitions emit a characteristic fluorescence with wavelength abundance determined using Planks law. Once the abundance of energies are sorted, a percentage relating to the amount of each element can be determined since XRF does not depend on chemical state (Creagh and Bradley 2000). XRF was performed on 40 mm of loose powder giving weight percentages of

Na2O, MgO, Al2O3, SiO2, P2O5, K2O, CaO, TiO2, MnO, Fe2O3, and S (Table A7-1).

151

Table C1. XRF results on 40 mm of loose powder (SRF Advanced Microanalysis Center); group number Amc2015-151 reported September 28, 2015.

Sample Na2O MgO Al2O3 SiO2 P2O5 K2O CaO TiO2 MnO Fe2O3 S Sum Wt% Wt% Wt% Wt% Wt% Wt% Wt% Wt% Wt% Wt% Wt% Wt% GSP2 2.85 1.07 14.5 68.5 0.24 5.34 2.09 0.62 0.04 4.70 0.02 100.0 SIDS1C 5.29 0.67 16.4 71.0 0.08 1.16 4.31 0.14 0.03 1.28 0.00 100.4 SIDS2 4.98 0.66 15.7 71.4 0.07 0.93 4.46 0.18 0.04 1.74 0.01 100.2 SIDS5 5.16 1.33 16.9 67.0 0.09 1.07 5.36 0.34 0.06 2.75 0.01 100.1 SIDS5 Dup 4.95 1.32 16.9 67.1 0.09 1.06 5.41 0.34 0.06 2.76 0.01 100.0 Wt% - Weight percentage of the entire sample.

Table C2. From the SRC Advanced microanalysis centre, XRD results backpacked random mount (From the SRF Advanced Microanalysis Center); group number: Amc2015-151.

Anorthite Quartz Albite Cordierite Microcline Muscovite Sum Sample wt% wt% wt% wt% wt% wt% wt% SIDS1C 40.3 33.9 14.0 1.30 10.5 - 100 SIDS2 43.1 35.0 13.4 1.40 7.10 - 100 SIDS05 46.9 24.3 15.0 1.40 7.80 4.70 100 SIDS-05Dup 45.1 24.2 20.1 0.900 6.70 3.10 100

152

Appendix D Thin section analysis

To ease in the quantification of the abundance of KF, grains were traced (using Adobe Illustrator) in order help identify them, and to sort them, to give an estimate of percent composition (Figure D1). These results were then compared to results obtained from XRD analysis (Table D1).

153

154

Figure D1. Sequence of steps performed to determine an estimate of relative percentage of different minerals in thin section, sample SIDS-5. A) First a clear microscope photo was taken of the thin section. B) Individual grains were traced. C) Once the grains were traced they were identified by comparing multiple microscope images at different angles under plain and polarized light. Since optical dating relies on dating quartz and feldspar grains, those minerals were targeted for identification. D) After grain identification, they were sorted to determine a relative abundance. The left side of this section (D) shows the percentage values associated with different areas of the circle. These values were then compared to data received from XRD analysis. No potassium feldspar was identified in this sample.

There were some discrepancies between the abundance of minerals observed in thin section and the results from XRD. This may be due to the fact that some grains appeared to be composed of more than one mineral. These multi-mineral grains would have been accounted for in the XRD analysis since the bulk sample was ground into a find powder prior to analysis. In addition, identification of minerals in thin section is subjective and therefore is directly related to the knowledge and experience of the analyst. It should be noted that thin section analysis of bulk sample only confirms the presence of KF in the bulk sample, and this may be different from that in the prepared fraction used for dating.

155

Table D1. Comparison of mineral composition determined by XRD and thin section analysis.

Sample XRD (wt %) Thin section analysis (%) – Bulk sample Quartz SIDS-1C 33.9 40 SIDS-2 35.0 53 SIDS-5 24.3 50 SIDS-9 - 55 Plagioclase feldspar SIDS-1C 54.3 23 SIDS-2 56.5 14 SIDS-5 61.9 15 SIDS-9 - 13 Potassium feldspar SIDS-1C 10.5 - SIDS-2 7.10 3 SIDS-5 7.80 - SIDS-9 - 2 Other SIDS-1C 1.30 38 SIDS-2 1.40 30 SIDS-5 6.1 30 SIDS-9 - 35

156

Figure D2. This sections of bulk sediment sub-sampled from optical dating samples.

Samples of prepared, ready to optically date, 180-250 µm KF were sent to SRC for XRF and XRD mineralogical testing. The KF was successfully concentrated through laboratory methods making up ~ 34 % of the sample. Plagioclase comprised ~40% of the samples. The remaining was composed of quartz at ~ 26% (Table D2). Although the presence of quartz in our sample does not affect age determination because IR light does not stimulate it, it does bear on our laboratory processing methods. In the laboratory, mineral density separation is performed once at 2.62 g/cm3 to separate

157

heave minerals from quartz, and a second time at 2.58 g/cm3 to separate the KF from plagioclase and quartz. Since the density of quartz ranges from 2.65-2.66 one would expect two separations to remove all of the quartz. However, amorphous quartz has a density of 2.196. It must also be noted that XRD is a semi-quantitative estimation.

Table D3. XRF Analysis on 40 mm loose powder of Savary Island KF (810-250 µm) done at the SRC Advanced Microanalysis Centre. Report No: AMC2016-062.

Mineral Weight (%)

Na2O 3.12 MgO 0.08

Al2O3 15.8

SiO2 69.55

P2O5 0.05

K2O 7.12 CaO 1.49

TiO2 0.05 MnO 0.01

Fe2O3 0.13 S 0 Sum 97.4

158

Appendix E. Grain size analysis.

Table E1. Grain size analysis results using dry sieves and a shaker table for SIDS1.

Grain size (µm) weight (g) % >710 28.3 13.2 710-500 98.9 46.4 500-425 26.7 12.5 425-355 30.4 14.3 355-300 11.1 5.2 300-250 13.6 6.4 <250 4.3 2.0 Total 213.3 100

Table E2. Grains size distribution samples using wet sieves with distilled water after HCl and H202 treatments.

SIDS8 SIDS1C Grain size (um) weight (g) % weight (g) % >700 9.8 2.00 68.7 16.85 600-700 23.2 4.73 65.9 16.16 500-600 78.2 15.94 92.1 22.58 400-500 102.2 20.83 72.9 17.88 300-400 181.6 37.01 68.2 16.72 250-300 55.5 11.31 21.3 5.22 180-250 36.6 7.46 16.9 4.14 <180 3.6 0.73 1.8 0.44 Total 490.7 100 407.8 100

159

Table E3. Grain size statistics from Malvern Mastersizer (raw data in Table D.4).

Sum of total volume Volume weighted Statistics d (0.1) (µm)2 d (0.5) (µm)3 d (0.9) (µm)4 between 500 and 750 mean (µm)1 µm (%) SIDS-1C 606 345 557 928 51 SIDS-2 430 230 391 684 33 SIDS-3 504 307 473 743 49 SIDS-4 602 332 546 943 48 SIDS-5 586 303 532 943 45 SIDS-6 430 260 406 634 38 SIDS-7 554 326 516 832 51 SIDS-8 403 234 375 609 32 SIDS-9 445 266 419 660 41 SIDS-10 409 247 386 604 34 SIDS-11 588 310 539 937 46 1 Calculated by laser diffraction; comparisons of particle size based on weight; values are the mean size of particles that make up the bulk of the sample. 2 Grain size represented at the 10th percentile (10% of the volume). 3 Grain size represented at the 50th percentile (50% of the volume). 4 Grain size represented at the 90th percentile (50% of the volume).

160

Table E4. Raw Malvern Mastersizer 2000 grain size data. The overall grain size distribution of each sample is better represented visually in Figure 2.4.

Grain sizes SIDS1C SIDS2 SIDS3 SIDS4 SIDS5 SIDS6 SIDS7 SIDS8 SIDS9 SIDS10 SIDS11 (µm) 0.02 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.49 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.98 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 1.95 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 3.91 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 5.86 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 7.81 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 11.72 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 15.63 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 23.44 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 31.25 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 46.88 0.00 0.00 0.00 0.00 0.12 0.00 0.00 0.00 0.00 0.00 0.17 62.50 0.00 0.00 0.00 0.00 0.34 0.00 0.00 0.00 0.00 0.00 0.54 78.13 0.00 0.00 0.00 0.00 0.25 0.00 0.00 0.00 0.00 0.00 0.56 93.75 0.00 0.00 0.00 0.00 0.04 0.00 0.00 0.00 0.00 0.00 0.38 109.38 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.17 125.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 156.25 0.00 0.41 0.00 0.00 0.00 0.00 0.00 0.22 0.00 0.02 0.00 187.50 0.00 2.49 0.04 0.01 0.18 0.62 0.00 2.05 0.56 1.22 0.04 218.75 0.08 4.80 0.43 0.21 1.01 2.53 0.20 4.57 2.24 3.53 0.57 250.00 0.83 6.62 1.84 1.15 2.08 4.82 1.23 6.92 4.31 6.02 1.53 312.50 4.86 15.95 8.73 5.79 7.31 15.15 6.51 17.81 13.83 17.15 6.30 375.00 9.12 16.00 14.18 9.92 10.29 18.19 11.31 18.32 17.14 19.01 9.76 437.50 11.75 13.81 16.08 12.08 11.52 16.98 13.71 15.52 16.56 16.73 11.47 500.00 12.50 11.02 15.04 12.34 11.37 13.75 13.79 11.80 13.89 12.91 11.62 625.00 22.30 14.59 22.15 21.21 19.44 17.18 22.79 14.02 18.28 15.19 20.19 750.00 15.82 7.70 12.01 14.70 13.79 7.48 14.57 5.96 8.63 6.05 14.35

161

875.00 9.89 3.81 5.65 9.22 8.93 2.62 8.14 2.15 3.38 1.89 9.20 1000.00 5.77 1.76 2.42 5.54 5.50 0.63 4.18 0.59 1.03 0.27 5.58 1250.00 5.03 0.99 1.31 5.21 5.24 0.04 2.93 0.07 0.15 0.00 5.16 1500.00 1.50 0.04 0.13 1.82 1.80 0.00 0.55 0.00 0.00 0.00 1.69 1750.00 0.43 0.00 0.00 0.63 0.60 0.00 0.09 0.00 0.00 0.00 0.53 2000.00 0.11 0.00 0.00 0.18 0.18 0.00 0.00 0.00 0.00 0.00 0.16

162

Appendix F Optical dating radial plots.

Figure F1. Radial plots of equivalent doses for samples collected from site 1 and site 2 (see Figure 3.12 for how to interpret). (CAM is central age model of equivalent doses, AGE is a fading corrected age, OD is overdispersion)

163

Figure F2. Radial plots of equivalent doses for dune head samples (see Figure 3.12 for how to interpret). (CAM is central age model of equivalent doses, AGE is a fading corrected age, OD is overdispersion).

164

Appendix G Dosimetry results.

Table G1. Comparison of dosimetry results obtained from two different facilities on the same samples.

1 Sample Facility Depth K (%) Rb (ppm) Th (ppm) U (ppm) wac2 wsat3 (cm) SIDS-1 A 1.02 ± 0.21 20.8 ± 3.5 1.09 ± 0.10 0.50 ± 0.12 690 0.03 0.27 SIDS-1 M 0.9 ± 0.09 22.0 ± 2.2 1.2 ± 0.07 <0.5 SIDS-2 A 0.67 ± 0.18 16.7 ± 3.3 1.89 ± 0.15 0.62 ± 0.12 920 0.06 0.31 SIDS-2 M 0.9 ± 0.09 14.0 ± 2.1 2.0 ± .11 <0.5 SIDS-3 A 1.01 ± 0.19 18.4 ± 3.0 1.07 ± 0.10 0.35 ± 0.11 296 0.03 0.22 SIDS-3 M 0.9 ± 0.09 15.0 ± 1.8 1.0 ± .06 <0.5 SIDS-4 A 1.05 ± 0.21 22.0 ± 3.1 1.01 ± 0.09 0.52 ± .10 250 0.03 0.29 SIDS-4 M 1.1 ± 0.11 25.0 ± 2.3 1.1 ± .07 <0.5 SIDS-5 A 140 1.02 ± 0.22 20.4 ± 3.6 2.32 ± 0.18 0.13 0.04 0.32 SIDS-6 A 110 0.76 ± 0.18 17.7 ± 3.3 2.09 ± 0.16 0.72 ± 0.13 0.06 0.31 SIDS-7 A 120 0.76 ± 0.18 20.2 ± 3.2 2.11 ± 016 0.55 ± 0.13 0.04 0.32 SIDS-8 A 480 0.85 ± 0.19 18.5 ± 3.0 1.75 ± 0.14 0.58 ± 0.13 0.06 0.32 SIDS-9 A 105 0.89 ± 0.19 18.0 ± 3.3 2.26 ± 0.17 0.72 ± 0.12 0.05 0.31 SIDS-10 M 140 1.00 ± 0.11 14.0 ± 1.3 1.80 ± 0.10 0.9 ± 0.3 0.05 0.31 SIDS-11 M 135 1.20 ± 0.11 21.0 ± 1.5 2.0 ± 0.11 0.7 ± 0.3 0.04 0.34 1Facility where dosimetry results were calculated. A = ANSTO, M=Maxxam Analytics. 2 "as collected" water content. Thought to be the most representative of the depositional environment during formation. 3 Saturated water content value for comparison. Note: All radioisotope concentrations were determined using neutron activation analysis (NAA), except for the U concentrations measured at ANSTO, which were determined by delayed neutron counting (DNC).

165

Appendix H Charcoal transport hypothesis

An hypothesis for the origin of the charcoal fragments that were found in the B- horizons of palaeosols at sites 1 and 2 (Figure H1). In each case charcoal fragments produced calibrated radiocarbon ages that are significantly older than the enclosing dune sand.

166

Figure H1 Proposed scenario for the incorporation of charcoal fragments into aeolian sand. The thickness of the burned horizon and the size of the charcoal fragments have been exaggerated for clarity. A) Parabolic dune formation begins. B) Over time, dunes arms elongate and the dune migrates downwind. Sediment is eroded from the deflation basin, between the dune arms, and is transported and deposited downwind. Erosion into old burned horizons, beneath the modern surface occurs in places, and charcoal is entrained an transported downwind. C) As dune migration continues, charcoal fragments are buried beneath more recently transported sediment. Deposition of charcoal fragments (all observed <5 mm in diameter) may have been facilitated by reduced surface wind speeds as a result of the initiation of stabilizing vegetation.

167