UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Long-term human-environment interaction on dynamic coastal landscapes : examples from 15,000 years of… Letham, Bryn 2017

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2017_november_letham_bryn.pdf [ 13.88MB ]
Metadata
JSON: 24-1.0355838.json
JSON-LD: 24-1.0355838-ld.json
RDF/XML (Pretty): 24-1.0355838-rdf.xml
RDF/JSON: 24-1.0355838-rdf.json
Turtle: 24-1.0355838-turtle.txt
N-Triples: 24-1.0355838-rdf-ntriples.txt
Original Record: 24-1.0355838-source.json
Full Text
24-1.0355838-fulltext.txt
Citation
24-1.0355838.ris

Full Text

LONG-TERM HUMAN-ENVIRONMENT INTERACTION ON DYNAMIC COASTAL LANDSCAPES: EXAMPLES FROM 15,000 YEARS OF SHORELINE AND SETTLEMENT CHANGE IN THE PRINCE RUPERT HARBOUR AREA  by  Bryn Letham  B.A. (Honours), Anthropology, The University of British Columbia, 2008 M.Sc., Anthropology, University of Toronto, 2011  A DISSERTATION SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY  in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Anthropology)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) September 2017  © Bryn Letham, 2017 ii  Abstract  This dissertation explores the intersections of past human settlement and the dynamism of coastal landscapes in the Prince Rupert Harbour area, in Tsimshian territory on the northern Northwest Coast, British Columbia. Taking relative sea level (RSL) and shoreline change as a major physical force in coastal people’s lives, both past and present, I explore how coastal fisher-hunter-gatherers occupied this transforming landscape and ultimately consider ways in which people’s engagement with the shores they lived upon may have been generative of new relationships to place and people. A reconstruction of the history of RSL change over the last 15,000 years is developed and presented. This is used to design a predictive model for landforms ideal for human habitation associated with raised paleoshorelines. A field survey of several of these targets identified three archaeological sites dating between 9500 and 6000 BP, pushing the archaeologically-recorded occupation of the area back 3000 years. These early Holocene sites indicate persistent use of places into the later Holocene as shoreline positions shifted with regressing RSL; it is proposed that this is associated with notions of territorial proprietorship acquired through historical precedence of use.   The second half of the dissertation presents a study of the developmental history of several large late Holocene village sites associated with massive anthropogenic shell-bearing components. It is argued that these sites themselves are significant anthropogenic transformations to shorelines, and it is demonstrated that there are instances where shell was very rapidly accumulated to raise, extend, or level landforms, and likely to buffer against foreshore erosion. I contend that many of the landscapes of the Northwest Coast are ‘fisher-hunter-gatherer built environments’, and argue iii  that increased physical investments in modifying coastlines may be associated with a transformation in the way territorial proprietorship is conceived. Specifically, I suggest it is associated with the formalization of institutionalized proprietorship systems similar to the rigid systems observed ethnographically for the Tsimshian, and that this may have resulted from the arrival of large numbers of newcomers to the region over the last 3000 years. The institutionalization of systems governing access to territory and resources enhanced social inequalities. iv  Lay Summary I explore the intersections of ancient human settlement and the dynamism of coastal landscapes in the Prince Rupert Harbour area, on the northern coast of British Columbia. I reconstruct the history of sea level change since the end of the last Ice Age, 15,000 years ago. Knowing the position of shorelines in the past and how shorelines have changed allows archaeologists to predict where ancient coastal settlements were located; I present the results of a survey that identified the three oldest currently recorded sites in the study area on ancient shorelines now stranded inland. I also document how humans transformed shorelines by mounding large amounts of shell, often to flatten or extend terraces on which they built villages. I hypothesize that the long-term occupation of changing shorelines and the physical modification of shorelines by humans may have been related to the development of land ownership systems among the ancient inhabitants.  v  Preface  This dissertation is composed of six chapters, four of which have been written as journal articles for publication on which I am principle author, and these are at various stages of the publication process as of this writing. Chapter 1 is an introduction and overview of the research. Chapters 2-5 are each studies that stand alone as journal articles but which build cumulatively on each other to form the broader narrative of this dissertation. Chapter 6 is a summary and conclusion to the work presented. All research presented herein is the original work of the author, Bryn Letham.  The research program was designed and implemented by me with the direction and facilitation of Andrew Martindale and Kenneth M. Ames, both members of my supervisory committee. Permits to conduct archaeological field work from which the data for this research derived were held by Andrew Martindale (BC Archaeology Branch Permit 2011-0207) and me (BC Archaeology Branch Permit 2015-0171). Research ethics approval was not required.  The content of chapters 2-4 has in the main not been modified from that which was submitted for peer-reviewed publication, or, in the case of Chapter 2, from that which is published. Because of this, the reader will note several repeated general content or points between chapters, mostly background or framing information. Citation formatting and some other journal-specific conventions have been modified for consistency within this entire dissertation, and some inter-chapter references have been added for clarity.  Chapter 2 is adapted from a multi-authored article published in Quaternary Science Reviews Vol. 153, in December 2016, entitled “Postglacial Relative Sea-level History of the Prince Rupert vi  Area, British Columbia, Canada”. Co-authors on this article are Andrew Martindale, Rebecca Macdonald, Eric Guiry, Jacob Jones, and Kenneth M. Ames. I piloted field work, collected field samples, and conducted laboratory analyses. This involved learning and conducting diatom identifications and other geological core sample analyses. I participated in stable isotope sample collection and preparation. I interpreted all results and wrote the paper. Jacob Jones aided in the collection of sediment samples from Livingstone cores and helped identify and quantify diatoms in the core sample from Tsook Lake. Eric Guiry helped collect field samples and facilitated stable isotope sampling and analysis. Rebecca Macdonald also facilitated stable isotope sampling, and performed stable isotope analyses in the Archaeology Isotope Laboratory at the University of British Columbia. Rebecca Macdonald also wrote the stable isotope analysis methods section and aided in the interpretation of stable isotope analysis results. Andrew Martindale and Kenneth M. Ames provided supervision, funding, and editing assistance.   Chapter 3 is adapted from a multi-authored manuscript that was accepted for publication in the Journal of Field Archaeology in August 2017. The paper is entitled “Archaeological Survey of Dynamic Coastal Landscapes and Paleoshorelines: locating early Holocene Sites in the Prince Rupert Harbour area, British Columbia, Canada”; the co-authors are Andrew Martindale, Nick Waber, and Kenneth M. Ames. I designed and implemented the archaeological survey, led the field work, and analyzed the results. I interpreted the data and wrote the paper. Nick Waber helped to design the GIS workflow for quantifying ‘theoretically habitable land’ associated with various paleoshorelines, and co-wrote the description of the workflow. Andrew Martindale and Kenneth M. Ames provided supervision, funding, and editing assistance.  vii  Chapter 4 is adapted from a multi-authored manuscript entitled “Assessing the Scale and Pace of Large Shell-Bearing Site Occupation in the Prince Rupert Harbour Area, British Columbia” that was accepted for publication in The Journal of Island and Coastal Archaeology in September 2017. The co-authors on this paper are Andrew Martindale, Kisha Supernant, Thomas J. Brown, Jerome S. Cybulski, and Kenneth M. Ames. Andrew Martindale, Kenneth M. Ames, and I developed the percussion coring and radiocarbon dating research design. I conducted field core sampling and laboratory analysis and designed the 3D site models. I also designed and conducted the archival analysis of the Boardwalk material, with guidance from Kenneth M. Ames, who excavated the site in the 1960s. I interpreted the data and wrote the paper. Kisha Supernant gathered 3D mapping data for GbTo-34 and GbTo-70. Thomas J. Brown aided with OxCal modelling and helped to write the methods section describing those analyses. Jerome S. Cybulski provided unpublished radiocarbon dates on human remains from Boardwalk. Andrew Martindale and Kenneth M. Ames provided supervision and funding. All co-authors provided editorial suggestions.  Chapter 5 is a manuscript that has been prepared for submission to a peer-reviewed journal. It is entitled “‘Culturally Modified Coastlines’: Long-term human-environment interactions and the generation of systems of territorial proprietorship in the Prince Rupert Harbour”. I am the sole author of this paper. Andrew Martindale and Kenneth M. Ames contributed substantially to the development of ideas in the paper and extremely useful guidance in the structure of the argument.  viii  Information on assistance received from non-co-authors is detailed in the acknowledgments section of this dissertation. ix  Table of Contents  Abstract .......................................................................................................................................... ii Lay Summary ............................................................................................................................... iv Preface .............................................................................................................................................v Table of Contents ......................................................................................................................... ix List of Tables .............................................................................................................................. xvi List of Figures ........................................................................................................................... xviii List of Abbreviations ............................................................................................................... xxiii Acknowledgements .................................................................................................................. xxiv Dedication ................................................................................................................................. xxvi Chapter 1: Coastal Archaeology in the Prince Rupert Harbour Area – An Introduction .....1 1.1 Passing Garden Island ........................................................................................................... 1 1.2 Aims and Scope of Research ................................................................................................ 5 1.3 Previous Archaeological Research in the Prince Rupert Harbour Area ............................... 8 1.4 Interpretive Framework ...................................................................................................... 18 1.5 Structure of the Dissertation ............................................................................................... 25 Chapter 2: Postglacial Relative Sea-level History of the Prince Rupert Area, British Columbia, Canada .......................................................................................................................32 2.1 Introduction ......................................................................................................................... 32 2.1.1 Study Area ................................................................................................................ 33 2.1.2 Regional Setting: Glacial History and RSL Change ................................................. 36 2.1.2.1 General Patterns for Coastal British Columbia ................................................. 37 x  2.1.2.2 Northern British Columbia ............................................................................... 38 2.1.3 Previous Sea-level Work Around Prince Rupert ...................................................... 40 2.2 Data and Methods ............................................................................................................... 41 2.2.1 Limiting Points, Index Points, and Indicative Meanings .......................................... 41 2.2.2 Measuring Elevation and RSL Change ..................................................................... 43 2.2.3 Measuring Age .......................................................................................................... 45 2.2.4 Field and Lab Methods ............................................................................................. 46 2.2.4.1 Livingstone Sediment Cores ............................................................................. 46 2.2.4.1.1 Diatom analyses of core sediment ............................................................... 48 2.2.4.1.2 Stable isotope analyses of core sediment .................................................... 48 2.2.4.2 Relict Paleomarine Sediments in Exposures and Raised Shoreline Landforms ...  ……………………………………………………………………………………………51 2.2.4.3 Basal Dates from Archaeological Sites............................................................. 51 2.3 Results ................................................................................................................................. 52 2.3.1 Livingstone Sediment Cores ..................................................................................... 82 2.3.1.1 Tsook Lake Core (TL#1, 49.7 m asl) ................................................................ 82 2.3.1.2 Rifle Range Lake 1 Core (RR1#2, 35 m asl) .................................................... 83 2.3.1.3 Cores From Bogs on Northern Digby Island (DIB1#1, 17.2 m asl; and NDB#1, 17 m asl) ........................................................................................................................... 84 2.3.1.4 Digby Island Lake 1 Core (DL1#1, 15.2 m asl) ............................................... 85 2.3.1.5 Bencke Lagoon Cores (BL#1 and BL#4, 2.4 m asl) ......................................... 86 2.3.1.6 Optimism Bay Cores (OB#1 and OB#2, -1.36 m asl) ...................................... 91 xi  2.3.1.7 Other Isolation Basin Cores At or Around Current Sea Level (SL#1, 2.2 m asl; PL#1, 0.75 m asl; RA#2, 0 m asl; GLP#1, 0 m asl).......................................................... 95 2.3.2 Paleomarine Deposits in Geological Exposures and Relict Paleoshoreline Landforms  ................................................................................................................................... 96 2.3.3 Archaeological Sites ............................................................................................... 100 2.4 Discussion ......................................................................................................................... 102 2.4.1 Prince Rupert RSL History and the Processes Driving RSL Change ..................... 102 2.4.2 Significance for Regional Studies ........................................................................... 110 2.4.2.1 Regional Glacial and RSL Histories ............................................................... 110 2.4.2.2 Implications for Early Human Occupation and Archaeological Survey ........ 111 2.5 Conclusion ........................................................................................................................ 113 Chapter 3: Archaeological Survey of Dynamic Coastal Landscapes and Paleoshorelines: Locating Early Holocene Sites in the Prince Rupert Harbour Area, British Columbia, Canada ........................................................................................................................................115 3.1 Introduction ....................................................................................................................... 115 3.2 Addressing Challenges to Surveying Geomorphically Dynamic Coasts .......................... 116 3.3 Study Area ........................................................................................................................ 118 3.4 Relative Sea Level History ............................................................................................... 126 3.5 Methods: Prince Rupert Harbour Area Paleoshoreline Survey Design ............................ 128 3.5.1 GIS and LiDAR Modelling ..................................................................................... 129 3.5.2 Field Methods ......................................................................................................... 132 3.6 Paleoshoreline Survey Results .......................................................................................... 134 3.6.1 GbTo-82/P024, Henry Point ................................................................................... 136 xii  3.6.2 GcTo-67, Scott Inlet................................................................................................ 138 3.6.3 GbTo-185, North Digby Island ............................................................................... 142 3.6.4 Other Paleoshoreline Sites Identified: GbTo-186, GbTo-183, GbTo-184, GcTo-66 ……………………………………………………………………………………..145 3.7 Discussion ......................................................................................................................... 146 3.7.1 Survey Assessment and Recommendations for Sampling Dynamic Coastlines .... 146 3.7.2 Northern Coast Early-Mid Holocene Record ......................................................... 148 3.8 Conclusions ....................................................................................................................... 156 Chapter 4: Assessing the Scale and Pace of Large Shell-Bearing Site Occupation using Percussion Coring and 3D Mapping: Examples from the Prince Rupert Harbour Area, British Columbia ........................................................................................................................158 4.1 Introduction ....................................................................................................................... 158 4.2 Large Shell-Bearing Sites around Prince Rupert Harbour ................................................ 160 4.3 Methods............................................................................................................................. 166 4.3.1 Documenting Site Layout with 3D Surface Mapping ............................................. 167 4.3.2 Sampling the Subsurface with Percussion Coring .................................................. 168 4.3.3 Radiocarbon Dating ................................................................................................ 170 4.3.4 Archival Analyses ................................................................................................... 180 4.4 Results ............................................................................................................................... 181 4.4.1 Kitandach/GbTo-34 Surface Topography............................................................... 181 4.4.2 Kitandach/GbTo-34 Paleotopography .................................................................... 183 4.4.3 Kitandach/GbTo-34 Developmental History .......................................................... 184 4.4.4 GbTo-70 Surface Topography ................................................................................ 191 xiii  4.4.5 GbTo-70 Paleotopography ...................................................................................... 192 4.4.6 GbTo-70 Developmental History ........................................................................... 193 4.4.7 Boardwalk/GbTo-31 Surface Topography ............................................................. 195 4.4.8 Boardwalk/GbTo-31 Paleotopography ................................................................... 197 4.4.9 Boardwalk/GbTo-31 Developmental History ......................................................... 200 4.5 Discussion: The Varying Histories of Large Shell-Bearing Sites and the Anthropogenic Engineering of Coastal Landforms in the Prince Rupert Harbour .......................................... 203 4.6 Conclusion ........................................................................................................................ 208 Chapter 5: ‘Culturally Modified Coastlines’: Long-Term Human-Environment Interactions and the Generation of Systems of Territorial Proprietorship in the Prince Rupert Harbour .........................................................................................................................210 5.1 Introduction ....................................................................................................................... 210 5.2 Territorial Proprietorship on the Northwest Coast ............................................................ 214 5.3 Endowment and Investment in Landscapes and Places .................................................... 216 5.4 A Dynamic Coastal Environment ..................................................................................... 221 5.5 A Fisher-Hunter-Gatherer Built Environment: Human-Landscape Modifications on the Northwest Coast ...................................................................................................................... 224 5.6 The Prince Rupert Harbour Case Study ............................................................................ 226 5.6.1 The Prince Rupert Harbour Built Environment ...................................................... 233 5.6.2 Ethnographic and Ethnohistoric Tsimshian Territorial Proprietorship ................... 242 5.6.3 Archaeological Correlates of Territorial Proprietorship ......................................... 245 5.6.3.1 Archaeological Evidence for Proprietorship-though-Endowment in the Prince Rupert Harbour ............................................................................................................... 249 xiv  5.6.3.2 Warfare and Territorial Defence ..................................................................... 258 5.6.3.3 Cemeteries and Legitimizing Genealogical Attachments to Place ................. 260 5.6.3.4 Inter-site Artifact and Raw Material Variation and Unequal Access to Trade ………………………………………………………………………………..261 5.6.3.5 Inter-site Faunal Assemblage Variation and Restrictions over Food Resources or Resource Gathering Areas .......................................................................................... 262 5.6.3.6 Archaeological Evidence for Proprietorship-though-Investment in the Prince Rupert Harbour?.............................................................................................................. 266 5.6.4 A Historical Model for the Development of Territorial Proprietorship in the Prince Rupert Harbour Area........................................................................................................... 270 5.7 Conclusion ........................................................................................................................ 272 Chapter 6: Conclusion ...............................................................................................................276 6.1 Returning to Garden Island ............................................................................................... 276 6.2 Summary and Overview of Contributions ........................................................................ 286 6.3 Future Directions .............................................................................................................. 289 References ...................................................................................................................................291 Appendices ..................................................................................................................................317 Appendix A Supplemental Material for Chapter 2 ................................................................. 317 A.1 Expanded Methods Sections for Chapter 2 ............................................................. 317 A.2 Stable Isotope Measurements from Select Contexts in Livingstone Sediment Cores ..   ................................................................................................................................. 319 A.3 Expanded Core Descriptions................................................................................... 321 A.4 Digby Island Bog Core DIB#1 ............................................................................... 329 xv  A.5 North Digby Bog Core NDB#1 .............................................................................. 330 A.6 Photograph of Didymosphenia geminata Specimen Identified in Core OB#2 ....... 331 A.7 Philip’s Lagoon Core PL#3 .................................................................................... 332 A.8 Auriol Point Lagoon Core GLP#1 .......................................................................... 333 A.9 Salt Lake Core SL#1 ............................................................................................... 334 A.10 Russell Arm Core RA#2 ......................................................................................... 335 Appendix B Supplemental Material for Chapter 3 ................................................................. 336 B.1 Table of Modelled Shoreline Lengths and Modelled Habitable Shoreline Lengths ……………………………………………………………………………………..336 B.2 Steps to Calculate Shoreline Length and Shorelines Adjacent to Theoretically Habitable Land in QGIS ..................................................................................................... 337 B.3 Table of Radiocarbon Dates for Chapter 3 ............................................................. 340 Appendix C Supplemental Material for Chapter 4 ................................................................. 342 C.1 All Radiocarbon Dates Used in Chapter 4 .............................................................. 342 C.2 CQL Code used for OxCal Modelling in Chapter 4 ............................................... 368 C.3 Map of GbTo-31 with Excavation Areas and Individual Excavation Units Labelled ..   ................................................................................................................................. 371 Appendix D Supplemental Material for Chapter 5 ................................................................. 372 D.1 Table of Accumulation Rates for All Sites ............................................................. 372  xvi  List of Tables  Table 1.1 Prince Rupert Harbour (PRH) Culture History as devised by the NCPP ..................... 13 Table 2.1 RSL data point types used in the present study and descriptions of indicative meanings........................................................................................................................................................ 42 Table 2.2 Tidal Parameters and their definitions for Canadian Hydrographic Survey Benchmark Station 9354, predicted over 19 years, start year 2010 ................................................................. 44 Table 2.3 Stable carbon (δ13C) and nitrogen (δ15N) isotope compositions and elemental carbon-to-nitrogen (C/N) ratios of known marine sediments and known freshwater sediments from the study area. ..................................................................................................................................... 50 Table 2.4 Radiocarbon dates for RSL Index and Limiting Points used to constrain the Prince Rupert Harbour area RSL curve ................................................................................................... 53 Table 2.5 Detailed list of most common or key diatoms observed in Livingstone Core Samples........................................................................................................................................................ 75 Table 3.1 Archaeological surveys in the Prince Rupert Harbour through time. ......................... 121 Table 3.2 2014 and 2015 Paleoshoreline survey tests, coverage, and results ............................. 135 Table 3.3 Archaeological sites with radiocarbon dates older than 6000 cal. BP in Tsimshian territory ....................................................................................................................................... 151 Table 4.1 Study site characteristics. ............................................................................................ 165 Table 4.2 All accumulation rates from GbTo-34, GbTo-70, and GbTo-31. ............................... 170 Table 4.3 Accumulation rate summary statistics. ....................................................................... 174 Table 4.4 Incidences of potentially immediate or very rapid deposition .................................... 177 xvii  Table 5.1 Possible archaeological evidence for territorial proprietorship in the Prince Rupert Harbour area................................................................................................................................ 247 Table 5.2 Archaeological sites in Prince Rupert Harbour area with radiocarbon dates earlier than 3000 cal. BP ................................................................................................................................ 255  xviii  List of Figures  Figure 1.1 Location of the Prince Rupert Harbour on the Northwest Coast................................... 2 Figure 1.2 Orthophoto of Venn Pass with recorded archaeological sites with shell-bearing components indicated...................................................................................................................... 3 Figure 1.3 Exposed intertidal zone at the eastern end of Venn Pass .............................................. 3 Figure 1.4 Archaeological sites in the study area known prior to this project (n=476) divided by primary site type and time period of identification ......................................................................... 9 Figure 2.1 Northern coast of British Columbia with study area highlighted and RSL curves for locations across a west-east transect ............................................................................................. 33 Figure 2.2 Study area and location of data points used to reconstruct the Prince Rupert Harbour area RSL history ........................................................................................................................... 35 Figure 2.3 Age-Altitude Plot of all limiting and index points used in this study ......................... 74 Figure 2.4 Tsook Lake Core TL#1 log, photo, and diatom analysis results. ................................ 83 Figure 2.5 Rifle Range Lake core RR1#2 log, photo, and diatom analysis results ...................... 84 Figure 2.6 Digby Island Lake 1 core DL1#1 log, photo, and diatom analysis results .................. 86 Figure 2.7 Orthophoto of a section of northern Venn Pass, showing Bencke Lagoon, Scott Inlet, and Optimism Bay ........................................................................................................................ 87 Figure 2.8 Upper section of Bencke Lagoon core BL#1 log, photo, and diatom analysis results 89 Figure 2.9 A: Plot of δ13C vs CORG/NTOTAL for known marine sediment samples (blue diamonds), known terrestrial samples (green diamonds), a sample of organic-rich sediment from the upper layer in core BL#1 (yellow triangle), and samples from the organic-rich layer at the bottom of cores OB#1 and OB#2 (red triangles). 9B: Plot of δ13C vs δ15N values for the same samples. ... 90 xix  Figure 2.10 Optimism Bay Cores OB#1 and OB#2 logs, photos, and stable isotope analysis sample locations (coloured squares). ............................................................................................ 93 Figure 2.11 Orthophoto of the location of Tea Bay Creek paleomarine exposure and photograph the profile, showing sequence from marine conditions to high intertidal/salt marsh to alluvial/estuarine conditions to the current forest soil buildup. .................................................... 98 Figure 2.12 Left: LiDAR-derived slope-classified map of a portion of northwest Digby Island showing inland linear ridges that likely represent stranded paleoshorelines and archaeological sites in the area. Right: LiDAR-derived hillshaded DEM of the same area. .............................. 100 Figure 2.13 Plot of all data points and the preferred RSL curve for the Prince Rupert Harbour region. ......................................................................................................................................... 102 Figure 2.14 Plot of all data points from the last 5000 years, and two potential RSL interpretations. ............................................................................................................................ 108 Figure 3.1 Northern coastal portion of Tsimshian territory in British Columbia. Archaeological sites with radiocarbon dates older than 6000 cal. BP are indicated. ........................................... 119 Figure 3.2 The Prince Rupert Harbour study area with archaeological sites sorted by principal site type identified through time (n=476) ................................................................................... 120 Figure 3.3 A: Summed probability distribution of all radiocarbon dates on marine and terrestrial material (n=392) for the Prince Rupert Harbour area. B: Prince Rupert area relative sea level curve with early Holocene RSL highstand indicated (modified from Letham et al. 2016: figure 13). .............................................................................................................................................. 125 Figure 3.4 Examples of LiDAR DTM with paleoshoreline and archaeological features ........... 128 Figure 3.5 Paleoshoreline survey targets and results. ................................................................. 133 xx  Figure 3.6 Orthophotos, slope-classified LiDAR DTMs, and hillshaded LiDAR DTMs of Henry Point with GbTo-82/P024 (A-C) and the north shore of Scott Inlet with GcTo-67 (D-F). Survey areas, subsurface tests, and paleoshoreline features indicated. ................................................... 137 Figure 3.7 Stratigraphy in evaluative excavation of a 6500 cal. BP shell component at GcTo-67 (A-B) and a selection of artifacts from each layer ...................................................................... 140 Figure 3.8 Orthophoto, slope-classified LiDAR DTM, and hillshaded LiDAR DTM of GbTo-185 area. Survey area, subsurface tests, and paleoshoreline features indicated. ............................... 143 Figure 3.9 Profile and plan view of 9000-8000 cal. BP occupation layers at GbTo-185. .......... 144 Figure 3.10 Radiocarbon dates for archaeological sites with radiocarbon dates older than 6000 cal. BP in Tsimshian territory. .................................................................................................... 150 Figure 4.1 The Prince Rupert Harbour area with study site locations and orthophotos of study site areas. ........................................................................................................................................... 161 Figure 4.2 Original study site maps with house depressions and previous excavation units indicated. Core test (CT) locations are indicated. ....................................................................... 164 Figure 4.3 GbTo-34 paleotopography and surface map. ............................................................ 182 Figure 4.4 Wave-cut erosion face at the south end of the front of the main site area at GbTo-34 with radiocarbon dated locations indicated and an age-depth model of the same sequence produced in OxCal. ..................................................................................................................... 185 Figure 4.5 Extrapolated site ‘profile’ from cores taken along GbTo-34 Transect A-A’ ............ 186 Figure 4.6 GbTo-34 radiocarbon dates plotted in 3D space and accumulation rate sequences (age vs. depth) for instances with more than 1 date............................................................................ 187 Figure 4.7 Time slices of cumulative accumulation at GbTo-34, with hypothesized RSL position indicated. ..................................................................................................................................... 188 xxi  Figure 4.8 Summed probability distributions (SPDs) of all non-human bone radiocarbon dates from unique contexts at the study sites. ...................................................................................... 190 Figure 4.9 GbTo-70 paleotopography and surface map. ............................................................ 192 Figure 4.10 GbTo-70 radiocarbon dates plotted in 3D space and accumulation rate sequences (age vs. depth) for instances with more than 1 date. ................................................................... 194 Figure 4.11 LiDAR DTM of GbTo-31 with georectified excavation locations. Inset photos indicate recent surface disturbance to the site visible in the DTMs that were used to georectify the site map ................................................................................................................................. 196 Figure 4.12 GbTo-31 extrapolated paleotopography based on patchy data from different excavation areas and current site surface .................................................................................... 199 Figure 4.13 SPD of GbTo-31 radiocarbon dates divided by area, showing shifting occupation of space through the last 5000 years. .............................................................................................. 201 Figure 4.14 Study site accumulation rates through time............................................................. 205 Figure 5.1 Archaeological sites recorded in the Prince Rupert Harbour area ............................ 227 Figure 5.2 Time series maps of archaeological sites with radiocarbon dates with calibrated median ages falling within the indicated timespans. .................................................................. 229 Figure 5.3 Examples of modifications to intertidal zones in Prince Rupert Harbour. ................ 235 Figure 5.4 Examples of shell deposition that significantly transform landforms at Prince Rupert Harbour villages .......................................................................................................................... 237 Figure 5.5 Summaries of accumulation rates at Prince Rupert Harbour area sites. .................... 240 Figure 5.6 Location of GbTo-185, GbTo-183, GbTo-64 and GbTo-65 on Northern Digby Island at different times and different RSL positions through the Holocene. ....................................... 251 xxii  Figure 5.7 Summary timeline of potential archaeological evidence for proprietorship-through-endowment, investment in the built environment, and institutionalized and unequal territorial proprietorship. ............................................................................................................................. 269 Figure 6.1 Headstone on the west beach of Garden Island. ........................................................ 276 Figure 6.2 Garden Island (GbTo-23) looking north from the air at low tide. ............................. 277 Figure 6.3 Cleaned up erosion profile on the southwest side of GbTo-23 showing over 2 m of accumulated cultural deposits and a calibrated age-depth model of 10 radiocarbon dates taken from this location. ....................................................................................................................... 278 Figure 6.4 Garden Island and surrounding area, flooded to show an 8.5 m asl high tide (i.e. high tide during the early Holocene RSL highstand). ......................................................................... 280 Figure 6.5 Stratigraphy at the base of the Garden Island exposure profile................................. 282 Figure 6.6 Headstones with Serbian and Japanese or Chinese text, Garden Island. ................... 285  xxiii  List of Abbreviations  A few abbreviations commonly used in the following pages:  RSL – relative sea level.  m asl – meters above sea level; in this case measured against the Canadian Geodetic Vertical Datum of 1928 (CGVD28).  cal. BP – calibrated years before present. ‘present’ is, with increasing irony as time passes, 1950.  xxiv  Acknowledgements I deeply thank Lax Kw’alaams and Metlakatla First Nations for allowing me to visit and conduct research in their beautiful territories, and for facilitating field logistics. I also extend my gratitude to all the researchers who have worked in and around the Prince Rupert Harbour and who have generated such a rich corpus of knowledge about the area. It is from your shoulders that I humbly try to build upon this work. I am particularly indebted to my supervisor, Dr. Andrew Martindale, who took me to the Dundas Islands for my first archaeological field experience, who sold me on this whole operation, and from whom I continue to learn to this day. Thank you for your guidance and pushing me to be a better scholar. This entire work would not have been possible without you. In addition, I express my sincere gratitude to the rest of my supervisory committee, Dr. Ken Ames and Dr. Zhichun Jing. Thank you all for the inspiration and for knowing when and how to push me to focus and when to give me room to breathe and explore. Special thanks to the field crews that aided data collection at various stages - Relative Sea Level reconstruction (Chapter 2): Steven Dennis, John Maxwell, Eric Guiry, Erika Leighton, Brian Pritchard. Paleoshoreline survey crew (Chapter 3): Steven Dennis, Justin Junge, Jacob Kinze Earnshaw, Eric Guiry, Steve Mozarowski, Tony Leighton, T.J. Brown, Robert Gustas. Coring and mapping GbTo-34 and GbTo-70 (Chapter 4): Steven Dennis, Paul Ewonus, Corey Cookson. There have been many special people in my life over the last five years to whom I owe, to some degree or another, the final outcome of this project: Eric Letham, Rasha Elendari, Haeden Stewart, David Bilton, Jacob Earnshaw, Ian Sellers, Gary Coupland, Jesse Fraser, Eric Guiry, xxv  Mike Blake, Stephanie Huddlestan, Kiara Hart, Dave Archibald, Sascha Gilbert, Gerry Riley, Cait Dub. Cheers to Cowpuccinos, Breakers, Wheelhouse, and the Belmont. Tina and Chris Letham fostered (and put up with!) my fascination with the past from a young age and have encouraged me to follow my passions at every turn in life.  Financial support for this project was provided by a Social Science and Humanities Research Council of Canada Joseph-Armand Bombardier CGS Doctoral Fellowship, a UBC Four Year Fellowship, a UBC Department of Anthropology Dissertation Writing Award, and the UBC Department of Anthropology Charles and Alice Borden Fellowship for Archaeology; as well as research grant funding held by my supervisory committee (SSHRC Grant # 410-2011-0414 PI: Martindale; and NSF Grant # 1216847 PI: Ames). Additional acknowledgments for Chapter 2: Duncan McLaren and Daryl Fedje for allowing the use of the Livingstone Coring equipment and for training in diatom analysis methods. Ian Hutchinson, Thomas James, and John Clague for detailed and productive comments and reviews on drafts of the paper. Audrey Dallimore and Malcom Nichol for photographing the Livingstone cores and allowing us to do sampling at the Pacific Geoscience Centre in Sidney, BC. Eric Letham for assistance with Figures 2.3, 2.4, 2.5, 2.6, 2.8, 2.10, 2.13, and 2.14. Additional acknowledgments for Chapter 3: John Maxwell provided advice for the practicalities of field testing. Steve Mozarowski helped analyse the lithic artifacts from GbTo-185. TJ Brown designed Figure 3A. Three anonymous reviewers greatly improved the final text for publication. Additional acknowledgments for Chapter 4: Gary Coupland for providing additional information on his excavations at Boardwalk. Three anonymous reviewers improved the text for publication. xxvi  Dedication  This work is dedicated to inhabitants of the Prince Rupert Harbour area – past, present, and future. 1  Chapter 1: Coastal Archaeology in the Prince Rupert Harbour Area – An Introduction  1.1 Passing Garden Island If you board a boat in Prince Rupert, British Columbia (Figure 1.1), and travel west, you will cross one of the deepest harbours in the province and then enter Venn Pass, the narrow channel between Digby Island and the Tsimpsean Peninsula (Figure 1.2). If you happen to be travelling through at the low end of one of the harbour’s 7.4 vertical meter tidal swings, you will observe a twice-daily exposed landscape of expansive gravel and mud flats, forested with patches of Fucus and eelgrass, and rife with intertidal shellfish. At the eastern entrance to Venn Pass is a particularly huge intertidal area, emerging from the center of which is a low treeless island covered in grass and shrubs (Figure 1.2 and Figure 1.3). This is Garden Island, and though from a distance it is an inconspicuous islet among a sea of larger islands and a sprawl of rainforested shorelines, there was once an ancient village at this location, standing at the entrance to one of the most densely occupied pre-European landscapes in North America.  2   Figure 1.1 Location of the Prince Rupert Harbour on the Northwest Coast.  3   Figure 1.2 Orthophoto of Venn Pass with recorded archaeological sites with shell-bearing components indicated.   Figure 1.3 Exposed intertidal zone at the eastern end of Venn Pass. Garden Island is the shrubby landform in the center of the photo on the left. The photo on the right is taken from the other side of Garden Island, facing west. The larger treed island visible in both photos is Anian Island. Photos by K.M. Ames.  4  Passing Garden Island into Venn Pass, the boat veers northwest past the ancient Tsimshian villages of Wilgiapshi, Laxmasxwm’ol, Kitandach, K’nu, and many others before routing west again and eventually passing the modern village of Metlakatla and then Pike and Tugwell Islands, beyond which is open ocean. While the only currently occupied community in Venn Pass is Metlakatla, before European contact there were dozens of villages along these shorelines, as well as in the broader area including Kaien Island, Digby Island, and around the Tsimpsean Peninsula. Occupants of these villages used the extensive intertidal resources at their doorsteps, and the spots along the shorelines where they lived are characterized by huge accumulations of shell that was deposited in the forest. Though currently buried by deteriorating forest detritus and forming soils, many of these shell deposits are visible as mounds that rise up several meters from the natural landform. They are also frequently visible in wave-cut banks at the shoreline, where erosion has undercut much of the foreshore and exposed meters of layered shell deposits and other cultural debris. The largest of these villages-upon-shell are frequently over 10,000 m2. A careful eye would also notice that cedar growth is enhanced at these areas (MacDonald 1969:246; Trant et al. 2016), and that they are far better drained than the surrounding forest floor – an important trait in one of the rainiest locations in North America.   These villages, and the cultural deposits accumulated below them, are significant physical transformations to the shape of the shorelines; much of the coast of Venn Pass and parts of many other shorelines in the Prince Rupert Harbour area are the product of human modification operating alongside the constant motion of the ocean. These sites have been venues for archaeological research for over a century (Ames and Martindale 2014; see also discussion below), and these studies have documented nearly 6000 years of human occupation along or near 5  to these shorelines. However, we know from research elsewhere on the Northwest Coast that the very position of these shorelines on which people lived and deposited shell has shifted as a result of relative sea level (RSL) change since the region’s deglaciation towards the end of the Fraser Glaciation (~30-12 kya; the most recent and final advance of the Wisconsin Glaciation) (Clague et al. 1982; Engelhart et al. 2015; Fedje et al. 2005b; Shugar et al. 2014). Thirteen thousand years ago, Venn Pass would have looked entirely different, and it would have appeared entirely different again 9000 years ago. Human occupations from the more distant past are not associated with the current shoreline. The dynamism of the coast must be considered to understand the full breadth of human history in this clearly-important archaeological landscape, as well as to be able to identify potential biases in the well-documented archaeological record for the region. In order to understand how villages such as that at Garden Island came into being, we need to know how both natural and cultural processes have contributed to shoreline positions and the shapes of the landforms along these shorelines.  1.2 Aims and Scope of Research In this dissertation I employ geological and archaeological methods to reconstruct coastline change since the terminal Pleistocene to explore the intersections of past human settlement and the dynamism of coastal landscapes in the Prince Rupert Harbour area, on the northern Northwest Coast, British Columbia. Taking RSL and shoreline change as a major physical force in coastal people’s lives, both past and present, I explore how coastal fisher-hunter-gatherers occupied this transforming landscape and ultimately consider ways in which people’s engagement with the shores they lived upon may have been generative of new relationships to 6  place and people, such as territorial proprietorship1 systems through which groups of people had rights and title over places and could restrict access of these places to others.   To fulfill this agenda, this dissertation research was designed as a series of sequentially linked modules that meet specific aims, each of which is treated in a subsequent chapter: (1) reconstruct the history of RSL change since the region’s deglaciation in order to be able to model where paleoshoreline positions were at different points in time (Chapter 2), (2) use the knowledge of RSL change to design an archaeological survey that targets sites on early Holocene paleoshorelines to assess the time depth of settlement in the area and to better assess how settlement shifted with changing shoreline positions (Chapter 3), (3) document the degrees to which humans themselves physically transformed shorelines through a study of the scale and tempo of anthropogenic deposition at large shell-bearing sites (Chapter 4), and (4) explore the implications of identified patterns of long-term occupations of changing shorelines and investments in modifying coastal landscapes for how fisher-hunter-gatherers conceived relationships of territorial proprietorship (Chapter 5). The temporal scope of this research is therefore broad; I reconstruct a history of RSL change for the last 15,000 years, use it to search for archaeological evidence for human occupation of the area in the early Holocene, and spend a good amount of effort studying how humans transformed coastal landforms with shell over the                                                  1 As detailed in Chapter 5, I follow Trosper (2009) in employing the term ‘contingent proprietorship’ to describe Northwest Coast systems of land tenure, though I drop the word ‘contingent’ for brevity’s sake. My use of this word is synonymous with other scholars’ use of the term ‘ownership’ when specifically referring to Northwest Coast indigenous societies, however, as strictly defined by economists, ownership involves the right to sell what is owned and implies an ontological separation between the owner and that which is owned. Trosper (2009) convincingly argues that this does not accurately describe ethnographically documented Northwest Coast First Nations’ territorial rights and title. 7  latter half of the Holocene. Importantly, this research explores the impacts of physical environmental change on fisher-hunter-gatherer lives while at the same time casting them as historical actors with the potential to modify the landscape in significant ways.   The overall thesis of the research is that an understanding of how coastal landscapes have changed and been shaped by both natural and cultural processes is essential for (1) assessing the representativeness of the archaeological record in such areas and (2) interpreting broader arcs of settlement history, in which human-environment interactions were themselves generative of key aspects of coastal culture. This research demonstrates that in the Prince Rupert Harbour area, the terminal Pleistocene archaeological record is completely unsampled and unknown, the early Holocene record is under-sampled, and the later Holocene record is well sampled but an unknown amount has eroded away. In the Prince Rupert Harbour long-term occupations of particular locations for thousands of years through RSL changes, along with investments in landforms that made them better to inhabit may have contributed to the development of institutionalized systems of land tenure where groups of people had the right to restrict access to territories and resources, which would have accentuated wealth and status inequalities.   This chapter reviews previous archaeological research in the Prince Rupert area and the interpretive and methodological frameworks on which the dissertation is built. It also outlines the chapters that follow. Thematic literature reviews specific to each individual study are embedded within the chapters themselves, and are not repeated here.  8  1.3 Previous Archaeological Research in the Prince Rupert Harbour Area Prince Rupert Harbour is located on the northern British Columbian mainland coast between the Skeena and Nass River Valleys, in the traditional territory of the Tsimshian First Nations (Figure 1.1). Ames and Martindale (2014) argue that the Prince Rupert Harbour is a ‘flagship region’ for archaeology because of its exceptional concentration of archaeological sites (Figure 1.4), its nearly 100-year history of academic research, the enduring cultural legacy and oral records of the Tsimshian descendants, and it’s “central, generative role in theory and explanation building on the Northwest Coast” (Ames and Martindale 2014:166). The area has a century-long history of archaeological study (e.g. Ames 2005; Archer 1992; 2001; Coupland 1988, 2006, 2013; Coupland et al. 1993, 2001, 2003, 2009, 2010; Cybulski 1978, 1992, 2014a, 2014b; Drucker 1943; Eldridge et al. 2014; MacDonald 1969; MacDonald and Cybulski 2001; MacDonald and Inglis 1981; Martindale and Marsden 2003; Patton 2011; Smith 1909). This research has demonstrated that Prince Rupert Harbour was densely occupied by large populations of fisher-hunter-gatherers living in large villages. For many, the ancient occupants of Prince Rupert Harbour came to be viewed as the epitome of ‘complex fisher-hunter-gatherers’ (Ames and Martindale 2014). However, prior to this dissertation research no archaeological site had accepted dates older than 6000 BP. This is largely a result of 100 years of research focused on the modern shoreline. 9   Figure 1.4 Archaeological sites in the study area known prior to this project (n=476) divided by primary site type and time period of identification. Note that nearly all inland sites are culturally modified trees (CMTs) located in the last two decades. Then dense clusters of these sites are found in the footprints of proposed major industrial developments in the Prince Rupert area.  Beginning in 1907, Harlan Smith, one of the first archaeologists to do research on the Northwest Coast, was struck by the exceptionally large ‘kitchen middens’ in the Venn Pass area between Prince Rupert and Metlakatla during what he himself describes as “hasty archaeological reconnaissance” (Smith 1909:595) of the area. During several visits over 20 years, he identified 10  large shell-bearing sites, which were ancient village locations, beneath cleared areas that had been converted to potato gardens during Christian missionizing efforts of the latter 1800s. The naturally-acidic Pacific Northwest rainforest soils that accumulated over these shell-bearing sites were both tempered by the alkalinity of the shell and well-drained by its porous matrix, making the old villages better than other locations for agriculture (Ferguson 1974).  These areas were highly visible from the water, and Smith often observed shell and stone artifacts eroding out of the wave-cut banks at the fronts of these locations. Smith identified several petroglyph sites in the high intertidal areas that are associated with these high-visibility large villages, and he conducted several salvage excavations at the latter, which unfortunately were never published in any detail, nor did he provide any specific detailed reports on the sites he identified and visited (Smith 1909, 1927).  In 1938, Philip Drucker conducted a survey of the north coast of British Columbia, during which he recorded 25 large sites in the Prince Rupert Harbour area (Archer 1989:39) and conducted test excavations at two, GbTo-3 on Anian Island and GbTo-15 on Charles Point (Drucker 1943). Drucker’s recording was much more consistent and rigorous than Smith’s, though his survey methods were similar: he identified sites from a boat by looking for areas that had been cleared and gardened and/or that had steep banks rising from the beach with shell eroding from them (Archer 1989:39-40; Drucker 1943). Similarly, therefore, the sites he identified were mostly large villages, the most visible sites from the water.   During WWII, Prince Rupert Harbour set up defensive infrastructure in order to defend North America’s closest port to Japan (Ferguson 1974), and consequently fell out of the spotlight of 11  archaeological research. However, in the early 1950s, a Prince Rupert local named James Baldwin worked with Charles Borden of the University of British Columbia to conduct salvage excavations at GbTo-10, a large village that was destroyed during the construction of a major commercial port terminal in Prince Rupert (Calvert 1968). Baldwin was also inspired by Drucker’s research, and visited many of the sites that the latter had recorded and in so doing identified several unrecorded sites. Little is known about the methods used to identify the new sites, but we can presume that they were found opportunistically during Baldwin’s explorations of the area around his hometown, likely using similar criteria to those of Drucker.   It was the unparalleled number and density of massive archaeological sites in the passes leading into Prince Rupert Harbour that drew one of the largest archaeological efforts in Canada to the area in the 1960s. In 1966 the National Museum of Canada (now the Canadian Museum of History) initiated the North Coast Prehistory Project (NCPP), which would conduct research around Prince Rupert and other areas of the northern coast of British Columbia over the next two decades and bring the archaeology of Prince Rupert Harbour to international attention (Ames 1981, 2005; MacDonald 1969; MacDonald and Cybulski 2001; MacDonald and Inglis 1981). The NCPP excavated eleven sites in Prince Rupert Harbour and recovered nearly 20,000 artifacts, over 280 human burials, wet site components with spectacular organic preservation, and a large (but poorly sampled and reported) faunal remain assemblage. The excavators also ran 127 radiocarbon dates, which span from nearly 6000 cal. BP to European colonization. Richard Inglis also directed a formal survey and site inventory that identified many new sites and new site types besides the large, highly visible villages that had been identified to date: small campsites, burial caves, stone fish traps, canoe runs, and historic remains (Inglis 1974:13-16; 12  Ferguson 1974). This survey involved more formal site prediction methods than previous surveys, including the use of topographic maps and aerial photographs, as well as considering settlement pattern and land use analogues with ethnographic records of traditional use, though it was fully focused on the modern shoreline (Inglis 1974).  MacDonald and Inglis (1981) used data from the NCPP to devise a three-phase culture history sequence, which has been utilized in later research efforts and reproduced in regional overviews narrating the archaeological history of the area (e.g. Ames and Maschner 1999; Martindale and Marsden 2003; Maschner 1997), though not without some recent criticism and modification (Eldridge et al. 2014). Unfortunately, much of the NCPP’s work remains unpublished, though recently Ames (2005) has published detailed artifact analyses of the excavated material.            13  Table 1.1 Prince Rupert Harbour (PRH) Culture History as devised by the NCPP (MacDonald 1969; MacDonald and Inglis 1981; Archer 1990:47-48). The present project recognizes some of the patterns observed, but does not adhere to any specific culture-history periodization. Culture History Period Dates Characteristics PRH III 5500-3500 BP -cobble tools relatively abundant, bilaterally barbed bone harpoons with line holes, unilaterally barbed bone harpoons with unilateral line guards (characteristic artifacts), also many bone awls, wedges and points that remain continuous throughout the entire sequence   -small shell-bearing sites consisting primarily of mussel and a relative lack of intertidal bivalves  PRH II 3500-1500 BP -chipped stone points and ground slate points and ‘pencils’ frequent, new harpoon forms appear   -most artifact types from PRH III continue  -rapid shell-bearing site accumulation and proliferation   -larger house sizes and larger villages, shell-bearing sites larger and include clams and other intertidal bivalves  -evidence of trade: obsidian, amber, and dentalium shells  -evidence for status differentiation: labrets, burials with grave goods in latter half of the period PRH I 1500 BP-European Colonization -increase in large pecked and ground stone artifacts such as large stone mauls and adzes, zoomorphic art, composite toggling harpoons  -most artifacts from PRH III and PRH II continue  -evidence for trade and status differentiations continue  -large villages with variation in house sizes  Subsequent research has built upon the culture history scheme of the NCPP and directed attention to understanding the origins of the systems of pronounced social inequality evident in the later Holocene archaeological record and in the ethnographic record. Ames (2005) explored the NCPP data within a framework focusing on the origins of social complexity and affluence among fisher-hunter-gatherers. As part of this endeavour, he considered developments and changes in logistical organization, domestic organization, mortuary practices, population demographics, and intensification and storage. With regards to the artifacts, Ames (1976, 2005) 14  found remarkable continuity through time and noted that artifact forms are mostly added through time, rather than replacing others. Considering other data along with the artifact patterns, he argued that ranking, stratified cemeteries, corporate group social organization, food storage, semi- or full sedentism, and intensified use of resource patches were characteristics of societies living in the Prince Rupert Harbour by 3500-3000 BP. He argued that 2000-1000 BP seem to be characterized by further shifts in settlement, including one or several abandonment and reoccupation phases. Jerome Cybulski has conducted bioanthropological studies of human remains excavated by the NCPP to test theories about the emergence of warfare among Prince Rupert Harbour’s inhabitants (Cybulski 1992, 2014a, 2014b). He documents evidence for endemic inter-personal violence between 3000 and 750 BP.  While the NCPP excavations were focused on getting big samples of artifacts by excavating large volumes at  a few sites, in the late 1980s and early 1990s David Archer shifted the research emphasis to settlement patterns and conducted the most thorough and systematic survey and mapping project in Prince Rupert Harbour to date, first as a second phase of NCPP-funded survey in 1982-1983, and then with the North Coast Heritage Inventory Project (NCHIP), funded by the Museum of Northern British Columbia (Archer 1984, 1985, 1989, 1990, 1991). Archer classified the shoreline into zones of different potential and conducted field transects of the intertidal zones and the forest at the shores’ edges, which included subsurface testing. In addition to the types of archaeological material sought by Inglis (above), Archer’s crew also recorded evidence for ancient pathways, cultural debris not necessarily associated with shell-bearing sites (stone and bone artifacts and debitage, fire-cracked rock, and vertebrate faunal remains in the intertidal zone or in tree throws), and culturally modified trees (CMTs). CMTs are an 15  archaeological site type that did not become recognized on the Northwest Coast until the late 1970s, and Archer’s recording of stands of these added significantly more sites to the inventory (Archer 1985:9).   Archer also systematically dated different village types to address the emergence of hierarchical socio-political organization (Archer 1988, 1990, 1992, 2001). Though fraught with sampling issues (Martindale et al. 2017), Archer’s work suggested that a transition from small villages with similar sized houses to large villages with differentially sized houses around 1900 BP signals a transition from egalitarian social organization to hierarchical social organization.   Since Archer’s work, archaeological surveys have been cultural resource management (CRM) projects with more limited areal scopes, related to forestry projects (e.g. Mitchell 1998) and expansions of the commercial port facilities in the area (e.g. Eldridge et al. 2007). The most recent have been impact assessments for proposed large liquefied natural gas terminal facilities (e.g. Morrisey et al. 2015). All of these surveys have been limited to the footprints of the development projects; the most frequently identified sites are CMTs, though the surveys have included limited-to-no subsurface testing away from the current shoreline (Figure 1.4).  In addition to these most recent surveys, in the 1990s and early 2000s Gary Coupland conducted a multi-site study of ‘household archaeology’ and political economy, exploring how subsistence economies were mediated through conflict over and control of resource locations and resource production. Coupland argued that elites’ control over resource production was key to reproducing hierarchical social organization, and through excavating house structures at various 16  villages, he explored the ways in which social organization was spatially reproduced in household organization and architecture (Coupland 2006, 2013; Coupland et al. 1993, 2001, 2003, 2009, 2010; Stewart et al. 2009).   Coupland’s work was the first major excavation project in Prince Rupert Harbour since the NCPP. In 2012 and 2013, Millennia Research conducted a massive-scale mitigative excavation of two entire sites on Kaien Island, GbTo-13 and GbTo-54 (Crockford 2016; Eldridge et al. 2014). In an extensive report, Eldridge et al. (2014) provide a proposed update to the NCPP culture history scheme and Ames’ artifact analysis, and contribute data that substantiates Coupland’s research on differential resource access and control between sites as well as ways in which households embody inequality and reproduce social order.  Another avenue of research spearheaded by Andrew Martindale and Susan Marsden is one that explores the intersections of Indigenous oral records (Tsimshian adawx and Tlingit at.oow) and archaeology as means of narrating history (Marsden 2001, 2002; Martindale 2006, 2009; Martindale and Marsden 2003). This research has focused on narratives of migrations and conflicts, compared them to the archaeological record of settlement patterns as a proxy for such events, and considered how these historical events impacted socio-political organization (Marsden 2001; Martindale and Marsden 2003). In contrast to the behavioural or processual approach of most other research in the area, Martindale and Marsden take a more historical stance and focus on the proximal causes for human action, such as beliefs in the supernatural (Martindale and Nicholas 2014) and the agency of sprits (Marsden 2002).   17  Most recently, Martindale and Ames have undertaken a project that continues exploring the intersection of archaeology and oral histories and expands the archaeological study of the chronology of settlement patterns in Prince Rupert Harbour by radiocarbon dating a large sample of known sites (Ames and Martindale 2014; Edinborough et al. 2016; Martindale et al. forthcoming a, b, c). An assessment of settlement patterns over a broad spatio-temporal scale requires a representative sample of radiocarbon dated sites; Martindale and colleagues have added over 200 dates; doubling the total from previous projects. It is within this larger settlement history project that this dissertation research is situated.  Given this long legacy of research, the Prince Rupert Harbour may be one of the most intensively archaeologically researched locations on the Northwest Coast. However, nearly all of this research has been focused on the modern shoreline and has not taken into consideration the ways in which shorelines have changed through time. This is most apparent when we consider that out of a cumulative total of 448 radiocarbon dates from 49 sites prior to the paleoshoreline survey described in this dissertation (Ames 2005; Archer 1992, 2001; Coupland et al. 2003, 2010; Cybulski 2014; Edinborough et al. 2016; Eldridge et al. 2007, 2014; Letham et al. 2016; MacDonald and Inglis 1981; Southon and Fedje 2003), none have produced accepted ages older than 6000 BP2. Elsewhere on the northern coast, archaeological sites from the terminal Pleistocene and early Holocene have been identified associated with both raised and submerged                                                  2 Two radiocarbon dates run by the NCPP produced ages older than 6000 cal. BP, but both were rejected by the excavators. S-1596 on human bone from GbTo-23 produced a result of 6420 +/-80 years (6872-6446 cal. BP), though was significantly older than all other dates from the site (Ames 2005). The excavators re-dated this individual and got a result of 2629 +/-70 (2261-1879 cal. BP; GSC-2888); which is in line with other dates from the site (Ames 2005). S-1410 on charcoal from GbTo-18 yielded a radiocarbon date of 5555 +/-140 (6658-6003 cal. BP), but was rejected by Sutherland (1978) for being out of stratigraphic sequence with other dates. 18  paleoshorelines, indicating that humans were in the region well before the oldest previously dated Prince Rupert area site (Carlson and Baichtal 2015; Fedje and Christensen 1999; Fedje et al. 2005a; Josenhans et al. 1997; Mackie et al. 2011; McLaren et al. 2011).   This dissertation builds on the foundation set by earlier research and incorporates new methods and a new set of research questions. The relatively recent identification of early sites on paleoshorelines elsewhere indicated that a detailed local RSL history reconstruction would be required in order to design a survey that target sites older than 6000 BP around the Prince Rupert Harbour. I also wanted to explore how processes of human engagement with and participation in the physically transforming coast may have shaped aspects of the culture of these peoples.   1.4 Interpretive Framework While Chapters 2-4 of this dissertation are primarily methods-and-results studies, implicitly the sum of all the parts of this research forms a historical ecology study of shoreline landscapes around Prince Rupert Harbour. Historical ecology is a research approach that focuses on diachronic landscape change as the result of human-environment interactions without limiting itself to only the deterministic impacts of the natural environment on human action, but also incorporates agency and history into studies of the relationships between the two (Ames 2004; Armstrong et al. 2017; Balée 1998, 2006; Balée and Erickson 2006; Crumley 1994; Thompson 2014; Thompson and Waggoner Jr. 2013). Historical ecologists assert that humans are key agents - both intentional and unintentional - in environmental change, and that humans often transform the environment to adapt it to particular social or political systems (Balée and Erickson 2006:4). Thus, this research on human-environment relationships, including this dissertation, 19  takes environmental change as a major force for analysis but moves beyond one-way adaptationism and explores humans’ recursive relationships with it. By first reconstructing the trajectory of RSL change in the Prince Rupert Harbour area and using the basic constraints of sea level position on habitation (i.e. people did not live underwater, but likely lived adjacent to water) to locate evidence of habitation from different times, Chapters 2 and 3 set a base line for Chapters 4 and 5, which explore how the occupants of the Prince Rupert Harbour adapted to changing shorelines as well as modified the shorelines themselves. The culmination of these human-environment interactions may in turn have had a transformative effect on settlement and politics in the region.  The potential for social transformations through physical shoreline transformations is profound. Grier (2014) argues that Coast Salish of the southern Northwest Coast manipulated coastal landforms in such a way as to make them more productive of more diverse subsistence resources (e.g. Deur et a. 2015; Lepofsky et al. 2015), which thereby led to the formalization of territorial proprietorship and rising social inequalities as a result of heightened economic inequalities between villages. Coles (2000) hypothesizes that a delay in the expansion of farming into parts of northern Europe in the mid-Holocene may have been a result of populations of peoples from the now-submerged Doggerland adapted to sedentary coastal lifestyles being pushed inland by rising RSL in the North Sea. He suggests that these groups were pushed into groups of mobile hunter-gatherers, who were also being influenced from the south by agriculturalists, but that within this milieu of various economies coming into contact with each other the coastally-adapted people, who had lived sustainable sedentary lifestyles since the early Holocene, had no immediate reason to adopt agriculture. In Mesopotamia, Kennett and Kennett (2006) argue that 20  the amalgamation of large populations of mobile hunter-gatherers through exceptionally rapid marine transgression (~1 horizontal km/year) between 8500 and 6000 BP was part of the process responsible for generating the region’s first cities of Ur and Eridu, located along newly formed deltas and alluvial floodplains.  Many studies of human-environment interactions focus on agricultural societies and how large-scale economic projects of these societies impacted the environments (e.g. Balée 2013; Balée and Erickson 2006; Hakansson and Widgren 2014; Redman 1999; Walsh 2014), and much of this work explores the ways in which humans modified their landscape to generate a ‘built environment’ (Lawrence and Lowe 1990). However, hunter-gatherer societies are not seen as having had nearly as active a role in landscape modification as urbanized peoples (Erickson 2006). Recent historical ecology research is beginning to contest this notion (Thompson and Waggoner Jr. 2013), and in coastal areas of southeast USA and Brazil, massive shell and earth constructions of hunter-gatherers are being recognized as key landscape features that had impacts on local ecology as well as human history (Fish et al. 2013; Pluckhahn et al. 2015; Schwadron 2013; Thompson 2016; Thompson and Turck 2009; Thompson et al. 2013; Villagran et al. 2011). For example, Thompson et al. (2013; Thompson and Turck 2009) hypothesize that the construction of large shell-bearing sites on the Georgia Bight by fisher-hunter-gatherers was an adaptive strategy to buffer against RSL fluctuation. In Chapter 4 and especially Chapter 5 I present the case that landscapes of the Northwest Coast such as the Prince Rupert Harbour were ‘fisher-hunter-gatherer built environments’ that resulted from generations of enduring modifications and management strategies (see also Grier 2014).    21  This dissertation is also more generally a study in landscape archaeology. Landscape archaeology is itself theoretically and methodologically fragmented (Anschuetz et al. 2001; Knapp and Ashmore 1999; Smith 2003; Wilkinson 2003), though I take landscapes to be the cumulative materialization of human action and natural processes in the environment (Thompson 2013; Wilkinson 2003). As a constellation of places with physical characteristics that limit or afford certain types of action, the landscape structures human occupation, movement, and interactions. However, with the human capacity to modify the environment, people can transform the structuring characteristics of landscapes for themselves and for future generations. In reality, because of the dynamism of both the natural world and human action and the resultant interrelations of each, landscapes are in a constant state of flux.   Despite the focus of this definition of landscape as material, this is not to say that landscapes do not have emotive or affective resonance (Tarlow 2012). Landscapes can indeed evoke culturally-specific sentiments or be mnemonic devices for people familiar with histories surrounding landscapes (e.g. Basso 1996; Harris 2010). Because of this, landscapes can also structure use and occupation with their affective influence, which can be generated through their aesthetic qualities, or through shared real or imagined histories associated with place. Intangible emotions, such as the knowledge that ancestors used a particular area or are buried there, can be powerful in structuring how people use places (Harris 2010). Smith (2003) explores how political authority can be produced and reproduced through landscapes through the ways in which they are imagined, perceived, and experienced, and he argues that politics become materialized in the landscape. For example, palaces or other monuments may influence particular ways of movement or restrictions on access to space as directed by a governing authority, or certain 22  aesthetic qualities of public art or architecture may be used to narrate a particular version of a history that justifies authority (Smith 2003). As such, Smith argues that aspects of politics are materially available for archaeologists to analyze within the landscape. Politics, defined as relations of authority between individuals or groups (Smith 2003), are inevitably influential in structuring settlement patterns and land use, especially in densely occupied landscapes such as the Prince Rupert Harbour area; so theoretically the non-material influences of landscapes on human action may also be materialized in some form within the landscape.   Therefore, human history and environmental history are both structured by and structuring of the landscape; and aspects of these histories can be materialized within the landscape, rendering landscapes long-term historical records of utility for archaeological analysis.  Because of the dynamism of the landscape, and because human history is part and parcel of this dynamism, one of the most important interpretive frameworks for studying this history is an empirically grounded analysis of how the physical landscape has changed through time and how it comes to the state at which an archaeologist encounters it (Stafford 1995). This requires a firm grounding in environmental and geoarchaeological method and theory (Dincauze 2000; Goldberg et al. 2001; Goldberg and Macphail 2006). In slightly different ways in each chapter, I use the material archaeological and environmental record to interpret aspects landscape dynamism through time, and query some of the ways in which the landscape-as-result-of-human-and-natural-processes structured certain aspects of culture, such as settlement patterns, site layouts, and notions of territorial proprietorship.   23  In narrating long-term histories of landscape and settlement change in the Prince Rupert Harbour area – especially in Chapter 5 – this dissertation explores the tensions between intergenerationally-structured notions of place and belonging, and historically specific adaptations to external stimuli such as environmental change and population pressure. In my analysis of how territorial proprietorship existed and was perpetuated in the past I suggest that at some points in time proprietorship emerges or exists non-discursively, through historical precedent of use (Gintis 2007) and persistent use of places (McLaren et al. 2015; Schlanger 1992; Schriever 2012). However, I take history to be a process: human action and practice constructs the social world, and through time, history (Pauketat 2001a, 2001b). Therefore, the persistence of elements of culture requires human action (Latour 2005). Even if cultural rules exist and are respected at a non-discursive level, they are still generated by people doing things; they do not exist a priori. There may be points in time when people need to do different things or change what they do to work to maintain the persistence of certain aspects of culture.  For example, Chapter 5 highlights a time of significant change in the Prince Rupert Harbour area between 3000 and 1300 BP, during which I hypothesize that the ways in which territorial proprietorship was manifest transformed to one in which it needed to be more overtly signaled, legitimized, and defended. I suggest that this transformation may have been catalyzed by large numbers of new people entering the area, who had to situate themselves within a landscape wherein proprietorship was based on heritable historically-endowed rights, but who were unfamiliar with these pre-existing systems. Faced with this disjuncture of understanding, people may have needed to negotiate, assert, and legitimize proprietorship in different, more discursive ways, one of which was through emphasizing physical transformations of the landscape in ways 24  that could be considered investments. These investments were transformations directed towards either increasing the land’s productivity or utility for subsistence or occupation (e.g. Grier 2014; Widgren and Hakansson 2014), asserting historical legitimacy or demonstrating control of places (e.g. Earle 2000), or interacting in culturally-appropriate ways with an environment that is seen as a spiritual agent that can bestow legitimacy (e.g. Marsden 2002; Van der Noort 2011). In many cases, people created or expanded habitation space into areas previously not habitable and deposited sediments in such a way that would buffer against the erosive forces of the ocean. The point is that peoples’ relationships with place – such as notions of proprietorship – while often structured by the landscape, are not static and are often transforming as a result the dynamism of both the landscape itself and the flows of people over and through it. Following Pauketat (2001a), the histories of landscape and settlement that I attempt to unpack in this dissertation are not successions of static structures or phases punctuated by single instances of change, but are instead “processes of cultural construction through practice” (Pauketat 2001a:87). I am only able to narrate partial fragments of aspects of these histories, and others before me reviewed above have narrated other elements. Together, and with future research, a more thorough corpus of historical excerpts may be pieced together.   Overall, this research engages with discussions that are re-assessing how we understand hunter-gatherer relationships with landscapes through an examination of the ways in which human settlement and shoreline change are interrelated and examining the potential relationship to phenomena such as systems of territorial proprietorship. Critically, I consider ways in which human investment in the landscape and generation of a built environment may be germane in the development and institutionalization of political complexity and inequality among fisher-hunter-25  gatherer societies (Grier 2014). In considering how hunter-gatherers actively modified their landscape and how supposedly mobile people invested in and developed long-term relationships with places, I explore the generative power of human-environment interactions for cultural institutions and practices.  1.5 Structure of the Dissertation The cumulative narrative of this research is a history of shoreline and settlement change in the Prince Rupert Harbour area (or rather, necessarily it is a history of several aspects of shoreline and settlement change, with a focus on several select periods or moments), with an emphasis on human-landscape interactions between the dynamic coast and past fisher-hunter-gatherers. I work from a study of landscape change independent of human action (relative sea level change) to one largely controlled by humans (the construction of coastal landforms on which to live), and culminate with the proposal that humans’ engagement with the processes that physically transform coastal landforms played an important role in shaping the social and political organization of people in the past. A longer-term diachronic understanding of settlement and shoreline change in the area is necessary to address the research aim of understanding peoples’ relationships with land and territory. I primarily employ methods from geoarchaeology and settlement/landscape archaeology. The specific methods and techniques used for each component of the research are detailed in their respective chapters; in this section I briefly overview the results and arguments of each chapter and the ways in which data for each were collected. All field data were collected between 2012 and 2015.  26  Chapter 2 reviews the glacial history of northern British Columbia and what is known of post-glacial RSL change in the region, and presents a refined RSL curve for the Prince Rupert Harbour area for the last 15,000 years based on a large set of radiocarbon dated samples that either directly indicate or provide elevational limiting information on the position of RSL. This reconstruction of RSL change is a fundamental building block for all subsequent chapters; it provides guidance to the survey for archaeological sites associated with paleoshorelines and provides essential context for interpreting how people also transformed the landscape. I demonstrate that RSL followed a general trajectory characteristic of mainland coastal locations that were glaciated during the Fraser Glaciation; RSL was well above its current position immediately following ice retreat, rapidly fell to a position below the current sea level in the terminal Pleistocene, rose again to a highstand 5-6 m above its current position during the early Holocene, and slowly fell towards its current position (with perhaps some fluctuations) between 6000 and 1500 BP. Chapter 2 also provides some detailed geological context for the study area.  Data for the RSL reconstruction in Chapter 2 came from geological core samples from lakes, bogs, lagoons, and bays; stratigraphic exposures identified in pedestrian surveys, and core samples from archaeological sites. Transitions from marine to terrestrial environments (or vice versa) identified through macro-sediment characteristics, diatom microfossil analyses, or stable Carbon and Nitrogen isotope analyses were radiocarbon dated from geological core samples and exposure sequences to indicate or provide limiting information on RSL position. Dates from the bases of archaeological sites were used to provide upper constraints on RSL position during their occupation.   27  In Chapter 3 the RSL curve and a precise and accurate LiDAR ground surface Digital Terrain Model (DTM) are used to design a predictive model for archaeological sites associated with early Holocene paleoshorelines – which were located 5-6 m higher in elevation than current shorelines – and test the hypothesis that the lack of archaeological sites pre-dating 6000 BP in the Prince Rupert Harbour is a result of previous surveys’ focus on the modern shoreline, as opposed to any historical reality. The chapter describes a field survey influenced by the predictive model and three pre-6000 BP archaeological sites that were found associated with raised paleoshoreline features, the earliest currently recorded archaeological material in the study area. This includes a site dating older than 9000 BP. Notably, each of these raised paleoshoreline sites is associated with more recent archaeological material at lower elevations that indicates the re-use of these locations as RSL dropped through the later Holocene. This indicates that people maintained attachments to certain places even as the shapes and positions of shorelines changed. Chapter 3 also reviews known early Holocene archaeological data from other coastal areas of Tsimshian territory as well as other parts of the northern Northwest Coast.  The raised paleoshoreline survey was conducted using the RSL reconstruction to guide for target elevations and a precise and accurate 1 m gridded LiDAR bare earth DTM (Airborne Imaging 2013). These were used in a GIS to design a predictive model for locating ideal target landforms that would have been associated with a RSL 5-6 m higher than present, which a small crew and I surveyed during one week in 2014 and three weeks in 2015. We employed auger and shovel testing of these landforms to successfully identify archaeological sites from a 10,000-6000 BP RSL highstand. This is the first survey in the Prince Rupert Harbour area to systematically target paleoshorelines. 28   Chapter 4 explores how people modified coastal landforms around Prince Rupert Harbour with the deposition of marine mollusc shells, and engages with broader discussions of methods for studying the formation of large shell-bearing sites around the world. The large village sites in the area are associated with tens of thousands of cubic meters of shell that has transformed the shape and characteristics of the landform. The chapter focuses on three village sites and describes novel methods used to unpack the scale and tempo of shell deposition at each through time. At two villages, percussion coring, 3D mapping, and intensive radiocarbon dating are used to model the natural land surface beneath sites and the accumulation of cultural deposits at these locations. The third case study in the chapter is a new analysis of data from NCPP excavations at the Boardwalk site (GbTo-31) from 1968-1970, demonstrating how archival data can be used in conjunction with new spatial and chronological modelling techniques to study site developmental histories. At each site, instances are identified where people rapidly built up portions of the site to widen terraces or create ridges, as well as to extend the fronts of sites shoreward to follow regressing RSL and to buffer against wave erosion. At moments within the longer-term occupations of these sites, people were effectively designing landforms on which to build villages, and this site engineering took into account longer term trajectories of RSL change and daily cycles of coastal processes that operated on the landforms.  The data used in Chapter 4 were collected by percussion coring in a grid pattern across two large village sites and making detailed 3D surface models of these sites with a Total Station over a three-week period in 2012. Cores were used to assess the depth of cultural deposits across the sites and to estimate the shape of the natural landform beneath these deposits. A large number of 29  provenienced radiocarbon dates from these cores was used to reconstruct the developmental history of these sites as well as to model accumulation rates at particular locations to assess the pace of deposition. Additional archival analysis and georectification of excavation plans and profiles on the LiDAR bare earth DTM of a third site excavated and dated by the NCPP (GbTo-31) was used for a comparative study of the developmental history of a third site.   In Chapter 5 I build on two main findings from Chapters 3 and 4 – first that people living around the Prince Rupert Harbour occupied particular locations for very long periods, even as the position of the shoreline changed as a result of RSL change; and second, that people came to actively and dramatically change these shorelines themselves with the deposition of large amounts of cultural debris, and that this sometimes took the form of conscious and rapid terraforming – and explore the implications that these might have for understanding how the occupants of the Prince Rupert Harbour related to their landscape. Specifically, I explore notions of territory, and propose, following Grier (2014), that the increased physical investment in landform modification transformed systems of territorial proprietorship. Long-term occupations of certain locations and archaeological evidence for cultural continuity suggest that notions of proprietorship-through-endowment – non-institutionalized intergenerational control of places mediated by historical precedent of use (Gintis 2007; Grier 2014) – may have existed in Tsimshian territory since as early as 9000 years ago. However, I present evidence for an increase in the degree to which people physically modified the landscape that coincides with several independent lines of archaeological evidence for more formalized restrictions on territorial access and stronger overt emphases on proprietorship of places after 3000 BP, and especially between 2000 and 1500 BP. I suggest that increased investments in a built environment seem to 30  have been associated with the formalization of institutionalized territorial proprietorship, wherein politically mediated inequalities in access to resources were more overtly stated and defended, and which seem to have amplified social inequalities. I suggest that the catalyst for the increased degree of investments in a built environment may have been the arrival of large numbers of people from other territories between 3000 and 1500 years ago, referenced in Tsimshian oral histories (Marsden 2001, 2002; Martindale and Marsden 2003) and suggested by a major increase in number of occupied sites during this time (Martindale et al. forthcoming b).  Chapter 5 is an interpretive synthesis of findings from Chapters 2-4 and makes reference to the findings of previous researchers in the area, but it also uses radiocarbon dates and accumulation rates calculated from date pairs from percussion core samples taken from a large sample of sites in the study area to assess settlement history and the timing of investments in shoreline modification through shell deposition (see, for example, Ames and Martindale 2014; Martindale et al. forthcoming a).  The final chapter concludes the dissertation by way of a single case study: that of Garden Island, mentioned in passing at the outset of this chapter. Nearly the entire portion of this island that is currently above the high tide line is composed of archaeological anthropogenic debris; the ‘island’ would be more of an intertidal reef had people not mounded shell upon it for thousands of years. This island was the site of an ancient village and cemetery in the center of Prince Rupert Harbour, and was the location of an early 1900s cemetery as well. A geoarchaeological tour of Garden Island helps to narrate how this landform came into being through processes of RSL change and coastal deposition as well as through human use and occupation; and its current 31  rapid erosion in the face of recent RSL rise and increased motorboat traffic serves as a poignant example of what was once a landform co-constructed by people and the ocean now being reclaimed in the absence of such a relationship.   32  Chapter 2: Postglacial Relative Sea-level History of the Prince Rupert Area, British Columbia, Canada  2.1  Introduction Several decades ago, pioneering regional compilations of radiocarbon dated relative sea level (RSL) data by Mathews et al. (1970) and Clague et al. (1982) demonstrated the variability of RSL histories on the west coast of North America since the end of the Fraser Glaciation, largely related to the location and thickness of ice sheets, the timing of their retreat, and the net result of subsequent isostatic adjustments, eustatic sea level change, neotectonic movements, and sedimentation processes. New compilations have highlighted and re-emphasized this variability (Engelhart et al. 2015; Shugar et al. 2014). RSL histories are key components of paleoenvironmental and landscape reconstructions, and are intimately tied to understanding geomorphological and biological (both human and non-human) change on coastal landscapes through the Holocene. Knowing how RSL changes transform coastal landscapes is a key component for identifying and interpreting the archaeological record along coasts, particularly for the terminal Pleistocene and early Holocene. To date, RSL studies on the northern Northwest Coast mainland have been limited in scope compared to other parts of the region (see summaries in Engelhart et al. 2015; Shugar et al. 2014).  This paper presents new data refining our understanding of the postglacial RSL history of the area around Prince Rupert, on the north coast of British Columbia, Canada (Figure 2.1). We use diverse methods for studying RSL change to generate a robust RSL curve based on a large 33  dataset of limiting and index points. We discuss what this information tells us about postglacial dynamics and coastline change through the Holocene, demonstrate its utility for locating evidence for early human occupation in the study area, and outline the importance of this new data for modelling of glacio-isostatic changes in northern British Columbia.   Figure 2.1 Northern coast of British Columbia with study area highlighted. RSL curves for locations across a west-east transect are shown (modified from Shugar et al. 2014), including the previously hypothesized curve for the Prince Rupert Harbour area. Modern communities are indicated by black dots.  2.1.1  Study Area The study area (Figure 2.2) is on the northern margin of the Hecate Lowlands, a 15–60 km wide area of low relief that extends about 600 km along the northern mainland coast between an 34  offshore coastal trough and the Coast Mountains, and includes many low islands close to the mainland. The surficial geology of the study area is primarily organic (usually peat) veneers or blankets over patches of glaciomarine sediments (clays, silts and dropstones) which in turn overlie metamorphic bedrock (Clague 1984; Massey et al. 2005). In a few areas there are massive deposits of glacial till. Shorelines are crenulated, particularly along the northern shore of Prince Rupert Harbour and through Venn Pass, where there are many sheltered bays, small inlets, and tidal channels. These shorelines often have sand and mud flats extending hundreds of meters at low tides. The Prince Rupert Harbour itself is a deep waterway, one of many glacially carved inlets and valleys in the wider region, the largest of which are Portland Inlet and the Nass River valley to the north and the Skeena River valley to the south.  35   Figure 2.2 Study area and location of data points used to reconstruct the Prince Rupert Harbour area RSL history. Letter and number codes correspond to data points in Table 2.4 and Figure 2.3. For Livingstone Sediment Cores, TL = Tsook Lake, OB=Optimism Bay, BL=Bencke Lagoon, NDB=North Digby Bog 1, DL = Digby Island Lake 1, DIB = Digby Island Bog 1, GLP = Auriol Point Lagoon, PL=Philip's Lagoon, SL=Salt Lake, RA = Russell Arm, RR = Rifle Range Lake 1. For Geological Exposures, A = Swamp Creek, B=Tea Bay Creek, C = estuary north of Optimism Bay, D = shell exposure in creek north of Bencke Lagoon #2, E = shell exposure in creek north of Bencke Lagoon #1, F=Russell Arm/Philip's Lagoon Isthmus, G = Melville Arm, H = McNichol Creek, I=Northwest Digby Island near GbTo-82, J = Pillsbury Cove Lagoon, K=West Kaien Island. For numbered archaeological sites, Borden Numbers or other identifying numbers are in Table 2.4.  Today the two principal communities in the study area are the city of Prince Rupert and the reserve town of Metlakatla, but prior to European contact the area included dozens of contemporaneously occupied villages inhabited by the ancestors of the Tsimshian peoples (MacDonald and Inglis 1981; Ames 2005). Archaeological remains of these villages dot the shorelines along bays and passes. These ancient inhabitants had an intimate relation with the sea, 36  and understanding how shorelines have changed through time is important for locating and interpreting past peoples’ material remains. The rich archaeological record indicates that Prince Rupert Harbour was one of the most densely occupied areas of the Northwest Coast by around 3000 years ago (Ames and Martindale 2014). However, even with a century of archaeological research that includes intensive radiocarbon dating (e.g. Ames 2005; Archer 1992, 2001; Coupland 1988, 2006; Coupland et al. 1993, 2001, 2003, 2009, 2010; Drucker 1943; MacDonald 1969; MacDonald and Cybulski 2001; MacDonald and Inglis 1981; Smith 1909), no archaeological sites dating earlier than 6000 years BP had been identified prior to our research. Elsewhere on the northern coast, terminal Pleistocene and early Holocene archaeological remains are being found with increasing frequency on paleoshorelines in the wake of detailed RSL reconstructions (Carlson and Baichtal 2015; Fedje and Christensen 1999; Fedje et al. 2005a; Fedje et al. 2011; Josenhans et al. 1997; Mackie et al. 2011; McLaren et al. 2011). Our research objectives are similar, and include using RSL data to survey for evidence of earlier occupation in this archaeologically-important place. We also seek to refine the understanding of postglacial landform dynamics in northern British Columbia, which we review next.  2.1.2 Regional Setting: Glacial History and RSL Change The history of glaciation and deglaciation is intrinsically linking to RSL change during our period of interest, particularly during the Terminal Pleistocene. The following sections outline some of the broad patterns for Coastal British Columbia and review more specific data pertaining to the northern coast of the Province.  37  2.1.2.1 General Patterns for Coastal British Columbia Recent compilations of known RSL data for the west coast of North America (Engelhart et al. 2015; Shugar et al. 2014) display a previously recognized (Clague et al. 1982) general pattern for the British Columbia coast in which postglacial RSL histories are largely mirrored between the offshore outer coast and the mainland coast, though these same compilations also demonstrate a high degree of RSL variation through time and space. As with other glaciated areas and their immediate peripheries (see Pirazzoli 1996), terminal Pleistocene RSL change on the Northwest Coast was governed by the location and thickness of ice sheets during the Fraser Glaciation (the most recent glacial period in western North America, ∼30–12 kya, and the latter part of what is more broadly termed the Wisconsin Glaciation in North America, ∼110–12 kya) and subsequent isostatic adjustments during and following deglaciation. The general trend is that mainland and inner coast areas were depressed downward tens to more than 200 meters by an ice sheet hundreds to several thousand meters thick during the Last Glacial Maximum (LGM). At the same time, unglaciated areas of the outer coast were bulged upwards by asthenosphere material displaced outwards by this depression (Clague and James 2002; Fedje et al. 2005b; Hetherington and Barrie 2004). Additionally, during this time global sea level was as much as 125 m lower as a result of ocean water locked up in the ice sheets (Fairbanks 1989).  Deglaciation of the region began around 18,000 or 19,000 cal. BP3 (Blaise et al. 1990) and the ice sheets retreated inland sequentially from the coast (Clague and James 2002; Clague 1984). At this time, RSL was much lower on the outer coast and much higher on the inner coast. Meltwater                                                  3 All dates are discussed in Calendar Years Before Present (i.e. before 1950). 38  caused a rise in global (eustatic) sea level, although this was quickly outpaced by isostatic readjustments caused by the unloading of ice from the land. The uplifted area collapsed, producing an overall rise in RSL on the outer coast, while the once-depressed inner coast rebounded upward, causing rapid RSL fall there. These effects were most pronounced at their outer and inner extremities, and recent work by McLaren and colleagues (2011, 2014) has identified a ‘sea level hinge’ area between the elevated outer coast and the isostatically depressed mainland where RSL position was generally stable through the Holocene.  2.1.2.2 Northern British Columbia Figure 2.1 depicts RSL curves for northern British Columbian locations running west-east. In this region the Cordilleran ice sheet reached its maximum extent sometime after 27,300–25,400 cal. BP (Blaise et al. 1990). Isostatic depression was greatest in the areas with the thickest ice cover, and during this time ice sheets extended out across the northern Hecate Strait into Dixon Entrance (Hetherington et al. 2004). Prince Rupert Harbour was fully glaciated. Offshore, the combined eustatic lowering of the sea level and uplift due to the forebulge resulted in RSL at least 150 m lower at southern Haida Gwaii, and the shallow Dogfish Bank and Laskeek Bank in western Hecate Strait were emerged as a wide coastal plain (Hetherington et al. 2003, 2004; Fedje et al. 2005b; Josenhans et al. 1997).  During deglaciation, glaciers retreated inland and from higher elevations first; the last glaciers to retreat were those that filled the deep inlets and river valleys (Clague and James 2002). This process was rapid, but not constant. There were temperature fluctuations that may have paused glacial retreat periodically, such as the Younger Dryas period between 12,900 and 11,700 cal. BP 39  (Fedje et al. 2011). In the Nass River Valley, McCuaig (2000; McCuaig and Roberts 2006) found several pauses in RSL regression at various highstands that formed now-relict deltas between 230 m asl and 130 m asl during glacial retreat in the area. Melting glaciers caused eustatic sea level to rise until the mid-Holocene (Fairbanks 1989; Smith et al. 2011).  As opposed to the forebulged outer coast, shorelines closer to the depressed mainland and up the valleys were submerged where isostatic depression was greater than the lowered eustatic sea level. Marine mollusc shells dating to 15,000 cal. BP found around Prince Rupert and Port Simpson on the north end of Tsimpsean Peninsula indicate that this part of the outer mainland coast was deglaciated by this time and that RSL was at least 50 m higher (Clague 1984). Radiocarbon dates on shells from Zymagotitz River, near Terrace, 110 km inland from Prince Rupert, indicate that this region was not deglaciated until several thousand years later, around 11,500 cal. BP, but that RSL was 170 m higher at this time in the Kitsumkalum-Kitimat trough (Clague 1984,1985). The highstands in the Nass River Valley remain undated, though their general elevation and distance from the coast are similar to those of the Kitsumkalum-Kitimat trough (McCuaig 2000; McCuaig and Roberts 2006). This illustrates that the timing and pace of deglaciation also caused time-transgressive RSL change. There was considerable discrepancy in deglaciation and RSL position between the outer coast and the heads of the inlets and valleys. Each of these flooded areas experienced rapid RSL drops caused by isostatic uplift, though the rates and timing varied.  The tilting of the crust surface from the uplifted forebulge to the heavily depressed mainland meant that the Dundas Islands, located 40 km northwest of Prince Rupert and 60 km northeast of 40  the northeastern tip of Haida Gwaii, were near to the midway ‘hinge’ point on the deformed continental plate, and maximally submerged by RSL 14.5 m above its current position (Letham et al. 2015; McLaren 2008; McLaren et al. 2011). After 14,000 BP, isostatic uplift and eustatic rise caused RSL to drop gradually from 14.5 m asl to its current position through the Holocene, with a still stand at 7.5 m asl between 8900 cal. BP and 6000 cal. BP. Meanwhile, on Haida Gwaii, isostatic collapse of the forebulge combined with the eustatic sea level rise caused RSL to rise 15 or 16 m above current sea level around 10,000–9500 cal. BP and stabilize there for about 4000 years before slowly dropping towards their current elevation, likely as a result of tectonic uplift (Clague et al. 1982; Fedje et al. 2005b:25).  More recent RSL changes are less known and less well understood in the region, as they were much subtler in comparison to early rapid isostatic and eustatic changes. Late Holocene RSL change is still occurring by way of low-amplitude isostatic, eustatic, steric, and tectonic changes; as well as much more localized processes such as catastrophic tectonic events (earthquakes), sedimentation, compaction, and erosion (Pirazzoli 1996). Late Holocene RSL changes are likely to be more localized, but tracking these smaller scale shifts is relevant for considering their impacts on the shorter timescales of human generations, as well as for understanding the potential impacts of RSL change in the present day.  2.1.3 Previous Sea-level Work Around Prince Rupert Previous RSL research around Prince Rupert was conducted by John Clague for the Geological Survey of Canada (Clague 1984, 1985; Clague et al. 1982), and briefly reassessed by Millennia Research Ltd. (Eldridge and Parker 2007). Clague suggested that RSL dropped from 50 m asl 41  sometime after 15,000 cal. BP and passed below its current elevation sometime after 10,000 cal. BP, before rising again to its current position at 5700 cal. BP (Figure 2.1D). The hypothesis for an early-to-mid Holocene lowstand was based on negative evidence: extensive radiocarbon dating of archaeological sites in the area during the 1970s did not yield any ages older than 5700 cal. BP (Ames 2005; MacDonald and Inglis 1981), leading to the suggestion that RSL had stabilized by this time and that older sites were submerged.  Millennia Research Ltd. tested this hypothesized early Holocene lowstand by examining three core samples from intertidal contexts in the area, and concluded that “sea levels never fell substantially lower than present”; though they allow that even in the absence of evidence, “sea level may still have fallen by a meter or two below modern levels” (Eldridge and Parker 2007:17). Our refined RSL curve for Prince Rupert Harbour includes data from this previous research but demonstrates a fairly different RSL history.  2.2 Data and Methods From 2012 to 2015 we conducted field work to identify RSL index points and limiting points. All RSL data points include location, elevation, age, and indicative meanings.  2.2.1 Limiting Points, Index Points, and Indicative Meanings Sea level index points are data directly indicating the position of RSL at a particular time and space (Hijma et al. 2015; Shennan 2015); they are usually in situ macrofossils or sediments with a known and restricted elevation range relative to the tidal range. For RSL reconstruction, the possible elevation range over which an index point could have formed is calculated and then the 42  difference between that range and current position of the indicator is measured (Table 2.1). Sea level limiting points are fossil or sedimentary indicators of either terrestrial (upper limiting) or marine (lower limiting) environments and constrain but do not directly indicate the position of RSL (Table 2.1). Most of our data are limiting points. Table 2.1 RSL data point types used in the present study and descriptions of indicative meanings. Indicator Sample Type Indicative Meaning Indicative Range for Database Reference Water Level for Database Explanation Index Points     Transitional (mixed fresh and brackish/marine) diatom assemblage in isolation basin sediments Basin sill is either nearly below high tide influence (dropping RSL) or just being inundated by high tides (rising RSL)  HAT to MTL (HAT + MTL)/2 Conservatively assumes that the dated sample represents the time at which the sill of the basin was between mean tide level and the highest astronomical tide level and was on the verge of being isolated/inundated. Growth position butter clam (Saxidomus gigantea) shells Sediment from which the specimen was taken was within the habitat elevation range of S. gigantea when the specimen was alive LLWMT +46 cm to LLWMT -91 cm ((LLWMT+46 cm) + (LLWMT-91 cm))/2 Modification of a method used by Carlson and Baichtal (2015) for estimating RSL based on the known growth range of S. gigantea (Foster 1991). Calculates the elevation range relative to tidal position within which the specimen could have lived.   Limiting Points Indicative Meaning Marine shells in sediment, not in growth position Sediment with shell is from within or below the tidal range. Only brackish/marine diatoms in sediments Sediment was deposited in coastal/marine setting, or, in the case of isolation basin sediments, when the sill was below lowest tide level. Only freshwater diatoms in sediments Sediment was deposited in fully freshwater setting, or, in the case of isolation basin sediments, when the sill was above highest tide level. Terrestrial peat/paleosol Sediment was formed/deposited above high tide. Archaeological site with remains of habitation (shell midden, charcoal concentrations, architectural features) Lowest instance of archaeological material was deposited above high tide. 43   2.2.2 Measuring Elevation and RSL Change The ’zero’ datum against which all elevations are measured relative to is geodetic mean sea level measured by the Canadian Geodetic Vertical Datum of 1928 (CGVD28) benchmark at Prince Rupert, which is 3.85 ± 0.01 m above Chart Datum (http://www.meds-sdmm.dfo-mpo.gc.ca/isdm-gdsi/twl-mne/benchmarksreperes/station-eng.asp?T1=9354&region=PAC). Conveniently, this elevation nearly coincides with Mean Water Level (MWL, 3.849 m above Chart Datum) at Prince Rupert, which is the average of all hourly water levels, and coincides with Mean Tide Level (MTL), the average of High Water Mean Tide (HWMT) and Low Water Mean Tide (LWMT) (Table 2.2; Canadian Hydrographic Survey, personal communication, 2015). Because of this coincidence, geodetic mean sea level, MWL, and MTL are treated as equivalent, and variations around this zero point are expressed as ‘m asl’.   44  Table 2.2 Tidal Parameters and their definitions for Canadian Hydrographic Survey Benchmark Station 9354, predicted over 19 years, start year 2010 (Canadian Hydrographic Survey, personal communication, September 28, 2015). Note that MWL and MTL are essentially the same and are equal to 0 m asl. Note that in Canada, tidal parameters are calculated based on predicted tides, whereas in the USA tidal parameters are calculated based on observed data. Tidal Parameter Abbreviation Definition Measurement above Chart Datum (m CD) Equivalent elevation relative to geodetic mean sea level (m asl; used in this study) Highest Astronomical Tide HAT Highest tide on an 18.6 year cycle. 7.514 3.664 Higher High Water Large Tide HHWLT The average of the highest high waters, one from each of 19 years of predictions. 7.407 3.557 Higher High Water Mean Tide HHWMT The average from all the higher high waters from 19 years of predictions. 6.17 2.32 High Water Mean Tide HWMT The average of the high water levels. 5.897 2.047 Mean Water Level MWL The average of all hourly water levels over the available period of record. 3.849 0 Mean Tide Level MTL The average of HWMT and LWMT. 3.8485 0 Low Water Mean Tide LWMT The average of the low water levels. 1.8 -2.05 Lower Low Water Mean Tide LLWMT The average of all the lower low waters from 19 years of predictions. 1.322 -2.528 Lower Low Water Large Tide LLWLT The average of the lowest low waters, one from each of 19 years of predictions. 0.006 -3.844 Lowest Astronomical Tide LAT Lowest tide on an 18.6 year cycle. -0.125 -3.975   The tidal range at Prince Rupert is 7.40 m, which is very large compared to other areas of the British Columbia coast (Canadian Hydrographic Survey, personal communication, 2015). This introduces uncertainty to measurements on indicators from marine or intertidal contexts that are not in situ, such as re-worked shells or diatoms, which can be pushed to the highest tidal limits 45  by waves or moved below the tidal range by currents or debris flows. In situ indicators, such as molluscs in growth position or salt marsh sediments provide more accurate estimates of RSL position within this wide tidal range.  The elevations of all data were measured using a variety of instruments and methods, including an RTK GPS unit and base station, a Leica Total Station, a clinometer and stadia rod, and hand held GPS units. Elevations were often derived from or double checked against LiDAR digital terrain models (DTMs) of the study area (Airborne Imaging 2013), and all field-derived elevations cross-checked against this dataset showed very good consistency. All measured elevations were converted to m asl on the CGVD28 datum. Vertical measurement errors are applied to all data points in the final dataset and expressed as 95% confidence intervals (see Hijma et al. 2015 for error types and equations).  2.2.3 Measuring Age All RSL limiting and index points in this study have ages measured by radiocarbon dating. All dates have been calibrated using OxCal 4.2 (Bronk Ramsey 2009, 2014), and are presented as 95% (2 sigma) probability ranges in calibrated years before present (BP, i.e. the year 1950). The marine reservoir effect was accounted for by applying a ΔR of 273 ± 38, which is a conservative estimate for at least the last 5000 years in the Prince Rupert area (Edinborough et al. 2016). It has been demonstrated elsewhere that ΔR values can fluctuate through time (e.g. Deo et al. 2004), and that marine organisms from immediate postglacial contexts may have larger offsets than subsequent times as a result of increased deep-water mixing from isostatic depression (Hutchinson et al. 2004b). However, we lack any controlled baseline data from prior to 5000 BP 46  to assess these effects for the study area. We therefore consistently apply a ΔR = 273 ± 38, acknowledging that this value could have been different in the past and some of our early shell dates may be younger than presented. However, most of our calibrated very early shell dates are in accord with early dates on terrestrial material.  Bulk samples of sediment or multiple fragments of macrofossil material were dated when single samples of the appropriate size were not available. Following Törnqvist et al. (2015), in these cases we applied an additional error of ± 100 years before calibration. Bulk organic sediment from immediate postglacial times likely contains carbon taken up from underlying glacial sediments (Hutchinson et al. 2004b). Hutchinson et al. (2004b) find a difference of 625 ± 60 years between postglacial bulk sediments and macrofossils for the southern mainland coast, though this effect varies locally based on the composition of local glacial substrates, and, as with the early postglacial marine shell, no baseline study has been conducted in the study area.  2.2.4 Field and Lab Methods Index and limiting points were derived from sediment cores from bodies of water that contain transitions to or from marine conditions, relict marine sediments identified in geological traverses, and the lowest (earliest) components of archaeological sites identified through excavations or percussion coring.  2.2.4.1 Livingstone Sediment Cores We collected 13 sediment cores from bogs, bays, and isolation basins ranging +49.7 to −1.36 m asl using a hand-driven Livingstone piston corer (Wright 1967). Isolation basins are water-filled 47  basins with a measurable sill over which water drains. In instances of RSL change, these basins are ‘isolated’ from marine conditions when highest high tide levels are below the elevation of the sill, but will be brackish or marine environments during times when tides wash over the sill or the sill is submerged. The bottoms of these basins accumulate sediments containing paleoenvironmental proxies (e.g. diatoms, pollen, foraminifera, ostracods, plant or animal macrofossils) over time. The point at which sediments record a change from a marine to fresh water depositional environment (or vice versa) approximates the time at which water containing those proxies passed over the sill elevation. Dating these transitions is a means of accurately measuring RSL position at certain times (Engelhart et al. 2015; Hafsten 1979, 1983a, 1983b; Hafsten and Tallantire 1978; Hutchinson et al. 2004a; James et al. 2009a; Kjemperud 1981; McLaren et al. 2011; Romundset et al. 2009; Rundgren et al. 1997).  In two instances we cored sphagnum bogs with standing water in which upward-growing peat obscured any definite sill; the surface elevation of standing water is used as a best estimate of elevation. For a tidal bay where a definite sill was not observable due to water depth we selected a well-sheltered location that we anticipated to have good sediment sequence preservation. For estimating the elevation of data points at this location we subtract the depth of dated samples from the elevation of the beach surface at the core location.  We cored basins until we reached an impenetrable obstruction or glacial sediments, which, in the study area, consist of either till or a distinctive blue-gray coloured glacio-marine clay (Clague 1984). Environmental transitions were identified using a combination of lithostratigraphic analyses (physical characteristics of the sediments), diatom microfossil analyses, and sediment 48  stable carbon and nitrogen isotope analyses. Samples were selected for AMS radiocarbon dating from points in the cores that were indicative of transitions.  2.2.4.1.1 Diatom analyses of core sediment Preserved diatom microfossils from core sediment were used as a proxy for changing water salinity and RSL transitions (Battarbee 1986; Zong and Sawai 2015). See Appendix A.1 for a detailed description of sample selection and preparation. A minimum of 300 identifications were made for each sample; species were identified using multiple reference guides (Campeau et al. 1999; Cumming et al. 1995; Fallu et al. 2000; Foged 1981; Hein 1990; Krammer and Lange-Bertalot 1986a, 1986b, 1986c, 1986d; Laws 1988; Pienitz et al. 2003; Rao and Lewin 1976; Tynni 1986). Diatom species were placed in a five-part salinity classification scheme based on the ‘halobian system’ (Hustedt 1953; Kolbe 1927, 1932) outlined by Zong and Sawai (2015:234): 1 = halophobic (salt intolerant freshwater) species, 2 = oligohalobous indifferent (freshwater) species, 3 = oligohalobous halophilic (freshwater but tolerant of salinity levels up to 2‰) species, 4 = mesohalobous (brackish water with salinity levels ranging from 2‰ to 30‰) species, and 5 = polyhalobous (marine water with salinity > 30‰) species.  2.2.4.1.2 Stable isotope analyses of core sediment For key strata where diatom evidence was lacking, we measured stable carbon (δ13C) and nitrogen (δ15N) isotope compositions and elemental carbon-to-nitrogen (C/N) ratios of the organic fraction of sediments as a proxy for paleoenvironmental salinity (for review see Khan et al. 2015b; Lamb et al. 2006). Organic sediments derived from autochthonous inputs of C3-dominated terrestrial materials should have lower δ13C values as well as higher and more 49  variable C/N ratios relative to sediments containing organics derived from marine algae and plants (Khan et al. 2015b). Intertidal and salt marsh areas have δ13C values and C/N ratios that are transitional, reflecting contributions of organic matter from both terrestrial and marine environments (Lamb et al. 2006; Mackie et al. 2005; Khan et al. 2015b). Our results also suggest that δ15N values from organic sediments can be useful for discriminating between marine and terrestrial/freshwater samples as the latter have consistently lower values.  We measured δ13C and δ15N values and elemental compositions of Holocene sediment samples from select cores with known freshwater/terrestrial (n = 8) contexts and marine/intertidal contexts (n = 12) as a comparative baseline for assessing paleosalinity of sediments that lack diatom evidence. Stable isotope compositions were measured using an Elementar vario MICRO cube elemental analyzer (EA) coupled to an Isoprime isotope ratio mass spectrometer in continuous flow mode. Detailed sample contextual details, preparation methods, sample calibration, and analytical uncertainty are discussed in the Appendix A.1. Known freshwater sediment samples yielded lower average δ13C and δ15N values and exhibit a wider range of CORG/NTOTAL ratios than known marine samples (Table 2.3 and Appendix A.2).  50   Table 2.3 Stable carbon (δ13C) and nitrogen (δ15N) isotope compositions and elemental carbon-to-nitrogen (C/N) ratios of known marine sediments and known freshwater sediments from the study area. Bencke Lagoon sample and Optimism Bay samples were of unknown environmental salinity origin and tested against the knowns. Bencke Lagoon is intermediate between fresh and marine values (though closer to freshwater) and suggests a mixture of inputs. Optimism Bay samples fall within the range of known freshwater samples.  δ13C Average δ13C Range δ15N Average δ15N Range CORG/NTOTAL Average CORG/NTOTAL Range Known Marine Samples (n=12) −23.60±2.38‰ −26.51‰ to −19.90‰ +5.3±0.9‰ +4.0‰ to +6.5‰ 17.0±4.8 10.1 to 24.9 Known Freshwater Samples (n=8) −28.78±1.32‰ −31.18‰ to −27.23‰ +0.4±1.1‰ −1.5‰ to +2.0‰ 26.8±16.1 13.3 to 54.4 Bencke Lagoon Sample (n=1) -26.61‰ n/a +2.2‰ n/a 15.3 n/a Optimism Bay Samples (n=6) -28.08±0.76‰ -28.76‰ to -26.86‰ +1.7±0.52‰ +1.1‰ to +2.5‰ 19.15±2.1 17.5 to 23.0    51  2.2.4.2 Relict Paleomarine Sediments in Exposures and Raised Shoreline Landforms Six exposures of marine deposits with abundant marine mollusc shells located above their current habitat range were identified through traverses up creeks or along shorelines and dated. Previous studies in the area have identified an additional four such exposures that we include (Archer 1998; Clague 1984, 1985; Fedje et al. 2005b).  In addition to identifying paleomarine sediments in exposures, LiDAR DTMs were used to identify landforms that could represent relict raised shorelines. These included linear stretches of steeper slope relative to adjacent higher and lower elevations that run parallel to the modern shoreline, which could represent relict wave-cut backbeach berms. These locations were ground-truthed and flat landforms immediately above them were tested for archaeological material.  2.2.4.3 Basal Dates from Archaeological Sites We collected Environmentalist Soil Probe (ESP) percussion core samples from large shell-bearing archaeological sites and dated the lowest instances of cultural material in these cores, operating on the assumption that the dated material represents human occupation on land and therefore above or near the contemporary higher high water mean tide (HHWMT) level (2.32 m asl), which we select as the most meaningful of the highest tide averages on the scale of human lifetimes (see Table 2.2). Several dates on the lowest cultural material from auger samples and test excavations conducted on hypothesized raised paleoshorelines are also included. 62 dates from 28 sites are included as upper limiting constraints on RSL.  52  Earlier compilations of RSL data points for Prince Rupert (Clague 1984, 1985; Shugar et al. 2014) include up to 40 dates from previously excavated archaeological sites, but provenience information for these dates is not available to assign elevations with the level of accuracy that we required (Dan Shugar, personal communication, 2015). We therefore do not include these dates in our analysis; all archaeological data points were collected in this study and carefully controlled for elevation.  2.3 Results One-hundred and twenty-three index and limiting points constrain the inferred RSL curve (Table 2.4, Figure 2.3). Five of these are from previous studies; the rest are new. All index points and key limiting points are reported individually in this section. A summary of diatom analyses is presented on the core log Figures and Table 2.5. A RSL curve interpreted from the entire collection of points, their association to one another, and judgement of their reliability is presented in Section 2.4.1. 53  Table 2.4 Radiocarbon dates for RSL Index and Limiting Points used to constrain the Prince Rupert Harbour area RSL curve. Map ID letters and numbers refer to locations on Figure 2.2 and labelled data points on Figure 2.3. Map ID (see Figure 2.2)     Lab #1 Site        Method and Test 14C Age BP +/- Calendar Range (older, 2 sigma)2 Calendar Range (recent, 2 sigma)2 Calibrated Median Easting Northing Material Proxy Indicator Elevation/(Index Point Paleo RSL Elevation), m asl  Source Index Points TL D-AMS 009955 Tsook Lake Livingstone Core TL#1 12167 47 14782* 13714* 14087* 406914 6023855 Seeds Lake core sediments - mixed diatom assemblage (fresh and brackish) 49.7/(47.9) This study DL D-AMS 008745 Digby Island Lake 1 Livingstone Core DL1#1 12312 41 15013* 13859* 14380* 406751 6017482 Twigs - charred Lake core sediment: freshwater diatom assemblage with slight mixing of brackish/marine diatoms 15.2/(13.4) This study BL D-AMS 009951 Bencke Lagoon Livingstone Core BL#1 11292 46 13255 13065 13142 408247 6021605 Twig Lagoon core sediment: mixed diatom assemblage (fresh and brackish/marine diatoms) 2.4/(0.6) This study B D-AMS 004468 Tea Bay Creek Bulk sample from exposure 9526 34 10196 9901 10073 405524 6022967 Butter/horse clam shell Marine mollusc shells in gray silty sand and clay 2.4/(5.15) This study 54  Map ID (see Figure 2.2)     Lab #1 Site        Method and Test 14C Age BP +/- Calendar Range (older, 2 sigma)2 Calendar Range (recent, 2 sigma)2 Calibrated Median Easting Northing Material Proxy Indicator Elevation/(Index Point Paleo RSL Elevation), m asl  Source C D-AMS 007880 Estuary north of Optimism Bay Bulk sample from exposure 9589 32 10250 9952 10159 407411 6022741 Tresus capax shell Marine molluscs in growth position 0.058/(2.8) This study Lower Limiting Points, intertidal 4 OS-101336   Biogenic shell deposit beneath GcTo-52 CT 2012-515 3010 30 2645 2327 2454 404301 6024878 Saxidomus gigantea or Tresus spp. shell ESP core test: marine mollusc shells in sand below archaeological site 5.02 This study 6 OS-119876 Biogenic shell deposit beneath GcTo-66 CT 2013-035 2800 25 2315 2071 2201 404990 6022028 Clam shell ESP core test: marine mollusc shells in sand below archaeological site 2.92 This study 6 OS-119874 Biogenic shell deposit beneath GcTo-66 CT 2013-035 3170 30 2762 2495 2676 404990 6022028 Clam shell ESP core test: marine mollusc shells in sand below archaeological site 2.38 This study 13 D-AMS 007890  Biogenic shell deposit beneath GbTo-184 CT 2014-526 3426 91 3201 2736 2952 408012 6012786 Saxidomus gigantea or Tresus spp. Shell ESP core test: marine mollusc shells in sand below archaeological site 5.36 This study 55  Map ID (see Figure 2.2)     Lab #1 Site        Method and Test 14C Age BP +/- Calendar Range (older, 2 sigma)2 Calendar Range (recent, 2 sigma)2 Calibrated Median Easting Northing Material Proxy Indicator Elevation/(Index Point Paleo RSL Elevation), m asl  Source OB D-AMS 008753 Optimism Bay OB#1 10350 47 11214 10880 11079 407593 6022454 Shell, unknown marine snail Bay core sediment: brackish and marine diatom assemblage, marine mollusc shells -4.76 This study OB D-AMS 008747 Optimism Bay OB#1 9586 37 11123 10751 10932 407593 6022454 Wood (bark) Bay core sediment: brackish and marine diatom assemblage, marine mollusc shells -4.76 This study OB D-AMS 008748  Optimism Bay OB#1 9866 46 11391 11201 11262 407593 6022454 Plant fibre Bay core sediment: interface between intertidal sediments and peat/gyttja, reedy plant macrofossil with brackish/marine plant δ13C ratio (-19.4)     -4.86 This study 56  Map ID (see Figure 2.2)     Lab #1 Site        Method and Test 14C Age BP +/- Calendar Range (older, 2 sigma)2 Calendar Range (recent, 2 sigma)2 Calibrated Median Easting Northing Material Proxy Indicator Elevation/(Index Point Paleo RSL Elevation), m asl  Source OB D-AMS 008754 Optimism Bay OB#2 10365 38 11219 10923 11098 407593 6022454 Mytilus sp. shell Bay core sediment: brackish and marine diatom assemblage, marine mollusc shells -5.26 This study OB D-AMS 008749 Optimism Bay OB#2 9568 40 11100 10735 10932 407593 6022454 Wood (charcoal?) Bay core sediment: brackish and marine diatom assemblage, marine mollusc shells -5.26 This study SL D-AMS 005838 Salt Lake SL#1 2080 30 2140 1952 2050 411213 6021611 Wood/ Charcoal (twig) Lake core sediment: brackish diatom assemblage 2.2 This study E D-AMS 007877  Shell exposure in creek north of Bencke Lagoon #1 Bulk sample 9908 33 10666 10371 10523 408272 6022534 Mytilus sp. shell Marine molluscs in gravelly sand 1.58 This study E D-AMS 007893  Shell exposure in creek north of Bencke Lagoon #1 Bulk sample 8962 32 10224 9928 10161 408272 6022534 Green wood, twig Marine molluscs in gravelly sand 1.58 This study 57  Map ID (see Figure 2.2)     Lab #1 Site        Method and Test 14C Age BP +/- Calendar Range (older, 2 sigma)2 Calendar Range (recent, 2 sigma)2 Calibrated Median Easting Northing Material Proxy Indicator Elevation/(Index Point Paleo RSL Elevation), m asl  Source D D-AMS 007878  Shell exposure in creek north of Bencke Lagoon #2 Bulk sample 10154 34 11023 10669 10830 408314 6022730 Clinocardium nuttalli shell Marine molluscs in silty clay 1.8 This study D D-AMS 007894  Shell exposure in creek north of Bencke Lagoon #2 Bulk sample 9359 28 10670 10506 10579 408314 6022730 Green wood, twig Marine molluscs in silty clay 1.8 This study Lower Limiting Points, intertidal or subtidal GLP D-AMS 008755 Auriol Point Lagoon GLP#1 7693 38 7990 7757 7881 406581 6020949 Balanus sp. Shell Lagoon core sediment: marine mollusc shells 0 This study BL D-AMS 008751 Bencke Lagoon BL#1 13320 63 15284 14675 15018 408247 6021605 Mytilus sp. Shell Lagoon core sediment: marine mollusc shells 2.4 This study BL D-AMS 009953 Bencke Lagoon BL#1 13131 40 14980 14230 14594 408247 6021605 Balanus sp. Shell Lagoon core sediment: marine mollusc shells 2.4 This study BL D-AMS 008752 Bencke Lagoon BL#4 13116 53 14970 14190 14561 408247 6021605 Balanus sp. Shell Lagoon core sediment: marine mollusc shells 2.4 This study 58  Map ID (see Figure 2.2)     Lab #1 Site        Method and Test 14C Age BP +/- Calendar Range (older, 2 sigma)2 Calendar Range (recent, 2 sigma)2 Calibrated Median Easting Northing Material Proxy Indicator Elevation/(Index Point Paleo RSL Elevation), m asl  Source K GSC-2290 West Kaien Island  12700 120 14211 13569 13908 411751 6015894 Mya Truncata Shell Marine mollusc shell in glaciomarine sediment 11 Clague 1984:46-47; Fedje et al. 2005b G D-AMS 002191 Melville Arm Bulk sample 11197 45 13151 12983 13070 412384 6021389 Charcoal Marine molluscs in silty sand 0 This study G D-AMS 002192 Melville Arm Bulk sample 9704 48 10455 10157 10279 412384 6021389 Marine Mollusc Shell Marine molluscs in silty sand 0 This study G D-AMS 002193 Melville Arm Bulk sample 9444 36 10145 9756 9967 412384 6021389 Marine Mollusc Shell Marine molluscs in silty sand -0.3 This study I D-AMS 011955 Northwest Digby Island, near GbTo-82 AT 2015-006 9936 42 10707 10393 10556 404458 6017413 Marine Mollusc Shell Marine mollusc shell hash 9.045 This study J Beta-221626 Pillsbury Cove Lagoon  7130 70 7525 7225 7370 409774 (approximate) 6021322 (approximate) Marine Mollusc Shell Marine mollusc in sand -0.05 Eldridge and Parker 2007 Off map (see Fig. 2.1) Beta-14465 Port Simpson  13040 70 14863 14080 14419 407339 (approximate)  6045499 (approximate) Marine Mollusc Shell Marine mollusc shell in gray clay  53.55 Archer 1998; Fedje et al. 2005b; Fedje et al. 2004:51    59  Map ID (see Figure 2.2)     Lab #1 Site        Method and Test 14C Age BP +/- Calendar Range (older, 2 sigma)2 Calendar Range (recent, 2 sigma)2 Calibrated Median Easting Northing Material Proxy Indicator Elevation/(Index Point Paleo RSL Elevation), m asl  Source Off map (see Fig. 2.1) Beta-14464 Port Simpson  12970 50 14649 14019 14240 407339 (approximate)  6045499 (approximate) Marine Mollusc Shell Marine mollusc shell in gray clay  53.55 Archer 1998; Fedje et al. 2005b  Off map CAMS-3390 Ridley Island  9480 70 10198 9731 9998 415331 (approximate) 6008382 (approximate) Marine Mollusc Shell Marine mollusc shell in gray clay exposed in a road cut 7.05 Fedje et al. 2005b; Fedje et al. 2004:46-47 RR D-AMS 008741  Rifle Range Lake 1 RR1#2 11908 42 14090* 13458* 13749* 417608 6015559 Mixed plant matter, appears charred and/or decomposing Lake core sediments - brackish and marine diatoms 35 This study RA D-AMS 005842 Russell Arm RA#2 3158 26 3448 3343 3385 411094 6021336 Green wood, twig Lagoon core sediment: marine mollusc shells 0 This study RA D-AMS 005843 Russell Arm RA#2 3681 29 3394 3143 3276 411094 6021336 Clam shell Lagoon core sediment: marine mollusc shells 0 This study RA D-AMS 005841 Russell Arm RA#2 2105 29 2148 1998 2077 411094 6021336 Tree cone (charred?) Lagoon core sediment: marine diatoms and marine δ13C and δ15N values 0 This study 60  Map ID (see Figure 2.2)     Lab #1 Site        Method and Test 14C Age BP +/- Calendar Range (older, 2 sigma)2 Calendar Range (recent, 2 sigma)2 Calibrated Median Easting Northing Material Proxy Indicator Elevation/(Index Point Paleo RSL Elevation), m asl  Source RA D-AMS 005840 Russell Arm RA#2 1751 26 1147 924 1026 411094 6021336 Clam shell Lagoon core sediment: marine mollusc shells 0 This study F D-AMS 005852 Russell Arm/Philip's Lagoon Isthmus CT 2013-503 13263 43 15187 14574 14929 410980 6021304 Marine mollusc shell ESP core test: marine mollusc shells 15.9 This study F D-AMS 004470 Russell Arm/Philip's Lagoon Isthmus CT 2013-503 13141 45 15011 14241 14620 410980 6021304 Balanus sp. Shell ESP core test: marine mollusc shells 15.78 This study SL D-AMS 005839 Salt Lake SL#1 3034 28 2660 2345 2491 411213 6021611 Clam shell Lake core sediment: marine mollusc shells 2.2 This study A D-AMS 007879  Swamp Creek Bulk sample 12954 33 14510 14000 14194 404395 6026537 Marine Mollusc Shell Marine mollusc shell in gray silty clay 3.83 This study B D-AMS 004469 Tea Bay Creek Bulk sample 8472 35 9533 9447 9495 405524 6022967 Small green wood cone Marine mollusc shells in gray silty sand and clay 2.4 This study B D-AMS 005846 Tea Bay Creek CT 2013-513 9559 39 11090 10730 10932 405507 6022978 Green wood ESP core test: marine mollusc shells 1.5 This study B D-AMS 005845 Tea Bay Creek CT 2013-513 9508 43 10197 9862 10045 405507 6022978 Clam Shell ESP core test: marine mollusc shells 1.5 This study 61  Map ID (see Figure 2.2)     Lab #1 Site        Method and Test 14C Age BP +/- Calendar Range (older, 2 sigma)2 Calendar Range (recent, 2 sigma)2 Calibrated Median Easting Northing Material Proxy Indicator Elevation/(Index Point Paleo RSL Elevation), m asl  Source B D-AMS 005847 Tea Bay Creek CT 2013-513 9665 37 10386 10133 10229 405507 6022978 Mytilus sp. shell ESP core test: marine mollusc shells 1.2 This study B D-AMS 005849 Tea Bay Creek CT 2013-513 9748 38 10480 10201 10323 405507 6022978 Barnacle shell ESP core test: marine mollusc shells 0.3 This study B D-AMS 005850 Tea Bay Creek CT 2013-513 9989 41 11695 11268 11451 405507 6022978 Green wood ESP core test: marine mollusc shells -0.6 This study B D-AMS 005851 Tea Bay Creek CT 2013-513 10256 43 11133 10762 10966 405507 6022978 Barnacle shell ESP core test: marine mollusc shells -0.6 This study TL D-AMS 009956 Tsook Lake TL#1 12514 50 15090 14365 14778 406914 6023855 Arctostaphylos sp. seed Lake core sediments - brackish and marine diatoms 49.7 This study Upper Limiting Points, non-archaeological BL UBA-29065 Bencke Lagoon BL#1 12199 49 14833* 13738* 14148* 408247 6021605 Organic-rich sediment Lagoon core sediment: freshwater diatom assemblage 2.4 This study BL D-AMS 009952 Bencke Lagoon BL#1 11589 37 13722* 13160* 13420* 408247 6021605 Seeds Lagoon core sediment: freshwater diatom assemblage    2.4 This study 62  Map ID (see Figure 2.2)     Lab #1 Site        Method and Test 14C Age BP +/- Calendar Range (older, 2 sigma)2 Calendar Range (recent, 2 sigma)2 Calibrated Median Easting Northing Material Proxy Indicator Elevation/(Index Point Paleo RSL Elevation), m asl  Source DIB D-AMS 005844 Digby Island Bog 1 DIB1#1 7345 35 8295 8028 8147 409292 6017796 Green wood Bog core sediment: Terrestrial plant macrofossils 17.2 This study H D-AMS 007892 McNichol Creek Bulk sample 8149 29 9241 9009 9075 413052 6022825 Green wood, stick Terrestrial plant macrofossils and freshwater diatom assemblage 18.5 This study NDB D-AMS 009948 North Digby Bog NDB#1 8718 33 10171* 9521* 9771* 405388 6018934 Multiple chunks of wood Bog core sediment - peat/gyttja 17 This study NDB D-AMS 009950 North Digby Bog NDB#1 7055 41 8169* 7624* 7879* 405388 6018934 Twigs and mixed plant matter Bog core sediment - peat/gyttja 17 This study OB UBA-29067 Optimism Bay OB#2 11922 60 14163* 13436* 13773* 407593 6022454 Organic-rich sediment Bay core sediment: peat/gyttja with terrestrial plant macrofossils -6.31 This study OB D-AMS 008750 Optimism Bay OB#2 11876 42 13772 13572 13677 407593 6022454 Wood Bay core sediment: peat/gyttja with terrestrial plant macrofossils -6.31 This study OB UBA-29066 Optimism Bay OB#2 10450 56 12700* 11823* 12307* 407593 6022454 Organic-rich sediment Bay core sediment: peat/gyttja with terrestrial plant macrofossils -6.31 This study 63  Map ID (see Figure 2.2)     Lab #1 Site        Method and Test 14C Age BP +/- Calendar Range (older, 2 sigma)2 Calendar Range (recent, 2 sigma)2 Calibrated Median Easting Northing Material Proxy Indicator Elevation/(Index Point Paleo RSL Elevation), m asl  Source 18 D-AMS 011947 Paleosol beneath GcTo-67 ST 2015-030 6175 35 7170 6960 7077 407969 6022226 Charcoal Shovel test: terrestrial paleosol beneath shell midden 7.95 This study RR D-AMS 008740  Rifle Range Lake 1 RR1#2 11826 59 14055* 13345* 13666* 417608 6015559 Twig fragments, potentially charred Lake core sediments - freshwater diatom assemblage 35 This study TL D-AMS 009954 Tsook Lake TL#1 11785 42 13971* 13330* 13623* 406914 6023855 Twigs and a cone Lake core sediments - freshwater diatom assemblage 49.7 This study Upper Limiting Points, archaeological  5 D-AMS 009643  GbTo-24 CT 2014-525 3518 27 3204 2919 3063 402621.3 6021540 Clam shell ESP core test: lower boundary of archaeological shell midden 5.26 This study 5 D-AMS 009641  GbTo-24 CT 2014-522 3585 29 3312 3007 3155 402628.7 6021537 Clam shell ESP core test: lower boundary of archaeological shell midden 3.8 This study 20 OS-108969   GbTo-34 CT 2012-030 2350 25 1780 1535 1650 407855.8 6021318 Clam Shell ESP core test: lower boundary of archaeological shell midden 9.53 This study 64  Map ID (see Figure 2.2)     Lab #1 Site        Method and Test 14C Age BP +/- Calendar Range (older, 2 sigma)2 Calendar Range (recent, 2 sigma)2 Calibrated Median Easting Northing Material Proxy Indicator Elevation/(Index Point Paleo RSL Elevation), m asl  Source 20 OS-108967   GbTo-34 CT 2012-017 3460 25 3132 2851 2984 407849.1 6021321 Mytilus sp. shell ESP core test: lower boundary of archaeological shell midden 9.06 This study 20 OS-108970   GbTo-34 CT 2012-019 2910 25 2459 2172 2324 407862.3 6021358 Clam Shell ESP core test: lower boundary of archaeological shell midden 8.86 This study 20 OS-108964   GbTo-34 CT 2012-024 3620 25 3335 3065 3204 407847.9 6021278 Mytilus sp. shell ESP core test: lower boundary of archaeological shell midden 7.57 This study 20 OS-108968   GbTo-34 CT 2012-031 4570 25 4552 4274 4421 407839.4 6021327 Mytilus sp. shell ESP core test: lower boundary of archaeological shell midden 7.36 This study 20 SUERC Average GbTo-34 CT 2012-005 n/a n/a 4767 4626 4696.5 407830.6 6021331 Mytilus sp. shell and charcoal ESP core test: lower boundary of archaeological shell midden 6.04 This study; Edinborough et al. 2016 20 OS-108926   GbTo-34 CT 2012-009 4600 30 4606 4308 4459 407844.9 6021367 Mytilus sp. shell ESP core test: lower boundary of archaeological shell midden 5.6 This study 65  Map ID (see Figure 2.2)     Lab #1 Site        Method and Test 14C Age BP +/- Calendar Range (older, 2 sigma)2 Calendar Range (recent, 2 sigma)2 Calibrated Median Easting Northing Material Proxy Indicator Elevation/(Index Point Paleo RSL Elevation), m asl  Source 20 OS-109689   GbTo-34 CT 2012-002 4300 20 4181 3898 4042 407812 6021293 Mytilus sp. shell ESP core test: lower boundary of archaeological shell midden 5.12 This study 20 OS-109085   GbTo-34 CT 2012-037 1910 30 1287 1077 1200 407864.3 6021233 Mytilus sp. shell ESP core test: lower boundary of archaeological shell midden 3.26 This study 21 D-AMS 009629 GbTo-4 CT 2012-535 3603 25 3326 3042 3180 408570 6020202 Clam shell ESP core test: lower boundary of archaeological shell midden 5.4 This study 21  D-AMS 009635  GbTo-4 CT 2014-504 3549 26 3241 2953 3104 408706 6020197 Clam shell ESP core test: lower boundary of archaeological shell midden 3.2 This study 24 OS-108830   GbTo-57 CT 2013-021 3210 25 2835 2596 2719 409965 6021274 Mytilus sp. shell ESP core test: lower boundary of archaeological shell midden 10.01 This study 23 OS-108829   GbTo-59 CT 2013-014 4470 25 4411 4144 4288 409704 6021142 Mytilus sp. shell ESP core test: lower boundary of archaeological shell midden 5.92 This study 66  Map ID (see Figure 2.2)     Lab #1 Site        Method and Test 14C Age BP +/- Calendar Range (older, 2 sigma)2 Calendar Range (recent, 2 sigma)2 Calibrated Median Easting Northing Material Proxy Indicator Elevation/(Index Point Paleo RSL Elevation), m asl  Source 22 OS-108828   GbTo-6 CT 2013-040 2710 25 2243 1945 2077 408172 6019197 Mytilus sp. shell ESP core test: lower boundary of archaeological shell midden 4.8925 This study 7 D-AMS 011954 GbTo-63 AT 2015-049 3350 29 2969 2740 2845 403172 6020154 Protothaca staminea shell Auger test: lower boundary of archaeological shell midden 9.98 This study 14 D-AMS 013864 GbTo-64 CT 2014-508 3590 24 3315 3020 3162 405966 6020930 Clam Shell ESP core test: lower boundary of archaeological shell midden 3.34 This study 9 OS-101344   GbTo-66 CT 2012-556 4650 40 4766 4380 4525 405015.5 6020414 Mytilus sp. shell ESP core test: lower boundary of archaeological shell midden 5.97 This study 9 OS-101352   GbTo-66 CT 2012-554 4780 25 4821 4562 4707 405001.6 6020395 Saxidomus gigantea or Tresus spp. shell ESP core test: lower boundary of archaeological shell midden 5.84 This study 9 OS-101350   GbTo-66 CT 2012-555 4230 30 4095 3815 3950 405000.1 6020410 Saxidomus gigantea or Tresus spp. shell ESP core test: lower boundary of archaeological shell midden 5.29 This study 67  Map ID (see Figure 2.2)     Lab #1 Site        Method and Test 14C Age BP +/- Calendar Range (older, 2 sigma)2 Calendar Range (recent, 2 sigma)2 Calibrated Median Easting Northing Material Proxy Indicator Elevation/(Index Point Paleo RSL Elevation), m asl  Source 12 OS-101572   GbTo-70 CT 2012-042 2900 25 2439 2158 2312 406785.9 6014176 Saxidomus gigantea or Tresus spp. shell ESP core test: lower boundary of archaeological shell midden 6.8 This study 12 OS-101569   GbTo-70 CT 2012-053 3000 25 2640 2318 2437 406798.8 6014181 Mytilus sp. shell ESP core test: lower boundary of archaeological shell midden 6.55 This study 12 OS-101567   GbTo-70 CT 2012-051 2910 30 2465 2165 2324 406792.2 6014200 Marine mollusc shell ESP core test: lower boundary of archaeological shell midden 6.49 This study 12 OS-101656   GbTo-70 CT 2012-044 2590 25 2061 1814 1932 406793.4 6014158 Mytilus sp. shell ESP core test: lower boundary of archaeological shell midden 6.49 This study 12 OS-101571   GbTo-70 CT 2012-047 3320 25 2930 2728 2814 406778.1 6014195 Mytilus sp. shell ESP core test: lower boundary of archaeological shell midden 6.225 This study 26 D-AMS 009639  GbTo-76 CT 2013-003 2509 26 1955 1710 1839 412957 6021122 Clam shell ESP core test: lower boundary of archaeological shell midden 7.49 This study 68  Map ID (see Figure 2.2)     Lab #1 Site        Method and Test 14C Age BP +/- Calendar Range (older, 2 sigma)2 Calendar Range (recent, 2 sigma)2 Calibrated Median Easting Northing Material Proxy Indicator Elevation/(Index Point Paleo RSL Elevation), m asl  Source 26 D-AMS 009637  GbTo-76 CT 2013-001 2479 24 1922 1686 1804 412948.7 6021073 Mytilus sp. shell ESP core test: lower boundary of archaeological shell midden 3.13 This study 10 OS-101580   GbTo-78 CT 2012-062 2560 20 2021 1786 1898 404667.1 6018835 Saxidomus gigantea or Tresus spp. shell ESP core test: lower boundary of archaeological shell midden 7.87 This study 10 OS-101578   GbTo-78 CT 2012-064 2250 25 1664 1404 1535 404619.3 6018856 Mytilus sp. shell ESP core test: lower boundary of archaeological shell midden 5.28 This study 10 OS-101575   GbTo-78 CT 2012-065 3580 30 3308 2997 3147 404625.2 6018815 Saxidomus gigantea or Tresus spp. Shell ESP core test: lower boundary of archaeological shell midden 5.03 This study 11 D-AMS 011956 GbTo-82 AT 2015-007 6436 34 6728 6463 6596 404466 6017411 Balanus sp. and other marine mollusc fragments Auger test: lower boundary of archaeological shell midden 9.39 This study 8 OS-101346   GbTo-89 CT 2012-549 2440 25 1873 1624 1759 405102.4 6020577 Mytilus sp. shell ESP core test: lower boundary of archaeological shell midden 9.2 This study 69  Map ID (see Figure 2.2)     Lab #1 Site        Method and Test 14C Age BP +/- Calendar Range (older, 2 sigma)2 Calendar Range (recent, 2 sigma)2 Calibrated Median Easting Northing Material Proxy Indicator Elevation/(Index Point Paleo RSL Elevation), m asl  Source 8 OS-101338   GbTo-89 CT 2012-551 3010 35 2650 2325 2456 405091 6020557 Mytilus sp. shell ESP core test: lower boundary of archaeological shell midden 6.19 This study 8 OS-101340   GbTo-89 CT 2012-550 3470 25 3141 2860 2998 405100.2 6020618 Mytilus sp. shell ESP core test: lower boundary of archaeological shell midden 5.44 This study 8 OS-101342   GbTo-89 CT 2012-547 3020 25 2650 2336 2467 405088.9 6020581 Mytilus sp. shell ESP core test: lower boundary of archaeological shell midden 4.8 This study 19 OS-101330   GcTo-1 CT 2012-525 3160 25 2749 2494 2665 407641.4 6021772 Mytilus sp. shell ESP core test: lower boundary of archaeological shell midden 8.81 This study 19 OS-101328   GcTo-1 CT 2012-522 3820 30 3562 3327 3436 407607.5 6021725 Mytilus sp. shell ESP core test: lower boundary of archaeological shell midden 4.21 This study 27 OS-101356   GcTo-27 CT 2012-508 3540 35 3245 2931 3092 403664.6 6025041 Saxidomus gigantea or Tresus spp. shell ESP core test: lower boundary of archaeological shell midden 9.12 This study 70  Map ID (see Figure 2.2)     Lab #1 Site        Method and Test 14C Age BP +/- Calendar Range (older, 2 sigma)2 Calendar Range (recent, 2 sigma)2 Calibrated Median Easting Northing Material Proxy Indicator Elevation/(Index Point Paleo RSL Elevation), m asl  Source 27 OS-101360   GcTo-27 CT 2012-506 4380 25 4314 3995 4160 403530 6025000 Saxidomus gigantea or Tresus spp. shell ESP core test: lower boundary of archaeological shell midden 5.66 This study 27 OS-101358   GcTo-27 CT 2012-507 4040 25 3830 3574 3704 403543 6025015 Saxidomus gigantea or Tresus spp. shell ESP core test: lower boundary of archaeological shell midden 5.29 This study 1 OS-101355   GcTo-28 CT 2012-504 2990 40 2644 2303 2431 403393 6025179 Saxidomus gigantea or Tresus spp. shell ESP core test: lower boundary of archaeological shell midden 5.15 This study 1 OS-101354   GcTo-28 CT 2012-503 3530 30 3223 2929 3079 403427 6025182 Balanus sp. shell ESP core test: lower boundary of archaeological shell midden 4.36 This study 1 OS-101353   GcTo-28 CT 2012-502 3360 30 2984 2744 2857 403523 6025010 Saxidomus gigantea or Tresus spp. shell ESP core test: lower boundary of archaeological shell midden 4.01 This study 17 OS-101551   GcTo-39 CT 2012-542 2630 25 2109 1865 1980 407299.5 6022426 Mytilus sp. shell ESP core test: lower boundary of archaeological shell midden 7.58 This study 71  Map ID (see Figure 2.2)     Lab #1 Site        Method and Test 14C Age BP +/- Calendar Range (older, 2 sigma)2 Calendar Range (recent, 2 sigma)2 Calibrated Median Easting Northing Material Proxy Indicator Elevation/(Index Point Paleo RSL Elevation), m asl  Source 2 OS-101565   GcTo-48 CT 2012-529 3790 25 3541 3295 3403 403972 (approximate) 6025087 (approximate) Saxidomus gigantea or Tresus spp. shell ESP core test: lower boundary of archaeological shell midden 5.4761 This study 3 OS-101559   GcTo-51 CT 2012-519 3350 30 2970 2739 2846 404039 6025083 Saxidomus gigantea or Tresus spp. shell ESP core test: lower boundary of archaeological shell midden 6.41 This study 3 OS-101557   GcTo-51 CT 2012-517 2570 25 2040 1794 1909 403965 6025048 Mytilus sp. shell ESP core test: lower boundary of archaeological shell midden 4.33 This study 3 OS-101561   GcTo-51 CT 2012-518 2760 25 2291 2018 2150 404027 6025072 Saxidomus gigantea or Tresus spp. shell ESP core test: lower boundary of archaeological shell midden 4.12 This study 4 OS-101335   GcTo-52 CT 2012-516 2860 20 2350 2131 2259 404325 6024799 Saxidomus gigantea or Tresus spp. shell ESP core test: lower boundary of archaeological shell midden 5.24 This study 4 OS-101334   GcTo-52 CT 2012-514 3210 20 2827 2605 2719 404310 6024840 Mytilus sp. shell ESP core test: lower boundary of archaeological shell midden 5.2317 This study 72  Map ID (see Figure 2.2)     Lab #1 Site        Method and Test 14C Age BP +/- Calendar Range (older, 2 sigma)2 Calendar Range (recent, 2 sigma)2 Calibrated Median Easting Northing Material Proxy Indicator Elevation/(Index Point Paleo RSL Elevation), m asl  Source 4 OS-101332   GcTo-52 CT 2012-512 2610 30 2096 1836 1956 404295 6024838 Saxidomus gigantea or Tresus spp. shell ESP core test: lower boundary of archaeological shell midden 4.99 This study 25 OS-101646   GcTo-6 CT 2012-560 5780 35 6006 5733 5889 412700.6 6021504 Mytilus sp. shell ESP core test: lower boundary of archaeological shell midden 4.18 This study 25 OS-101555   GcTo-6 CT 2012-557 4490 30 4437 4147 4312 412686 6021480 Mytilus sp. shell ESP core test: lower boundary of archaeological shell midden 3.58 This study 6 OS-119875 GcTo-66 CT 2013-035 2410 35 1858 1584 1724 404990 6022028 Clam shell ESP core test: lower boundary of archaeological shell midden 3.22 This study 18 D-AMS 011948 GcTo-67 ST 2015-030 5732 33 6635 6445 6527 407969 6022226 Charcoal Shovel test: lower boundary of archaeological shell midden 8.14 This study 15 D-AMS 007906 GbTo-183 CT 2014-515 3243 28 3560 3393 3463 405353 6020650 Charcoal ESP core test: lower boundary of archaeological shell midden 6.86 This study 73  Map ID (see Figure 2.2)     Lab #1 Site        Method and Test 14C Age BP +/- Calendar Range (older, 2 sigma)2 Calendar Range (recent, 2 sigma)2 Calibrated Median Easting Northing Material Proxy Indicator Elevation/(Index Point Paleo RSL Elevation), m asl  Source 13 D-AMS 007889  GbTo-184 CT 2014-526 2853 24 2346 2126 2250 408012 6012786 Saxidomus gigantea or Tresus spp. Shell ESP core test: lower boundary of archaeological shell midden 5.51 This study 13 D-AMS 013865 GbTo-184 CT2014-527 3034 26 2660 2345 2491 408020 6012763 Saxidomus gigantea or Tresus spp. shell ESP core test: lower boundary of archaeological shell midden 4.62 This study 28 OS-101563   T627-2 CT 2012-574 4860 25 4925 4645 4805 410002.9 6016085 Mytilus sp. shell ESP core test: lower boundary of archaeological shell midden 5.876 This study 16 D-AMS 011949 GbTo-185 ST 2015-034 7445 35 8348 8186 8266 405975 6020487 Charcoal Shovel test: archaeological material on raised terrace 11.35 This study 16 D-AMS 011950 GbTo-185 ST 2015-034 8220 40 9304 9028 9186 405975 6020487 Charcoal Shovel test: archaeological material on raised terrace 11.25 This study *indicates date on a bulk sample of material that has been calibrated with an additional +/-100 year uncertainty modifier. 1 Labs are indicated by Lab Number prefixes as follows: D-AMS=DirectAMS, Bothell, WA; OS= National Ocean Sciences Accelerator Mass Spectrometry, Woods Hole Oceanographic Institute, Woods Hole, MA; SUERC=Scottish Universities Environmental Research Centre, Glasgow, Scotland; UBA=Queen’s University Belfast 14Chrono Centre for Climate, the Environment, and Chronology, Belfast, UK; Beta=Beta Analytic Inc., Miami, FL; CAMS=Center for Accelerator Mass Spectrometry,  Livermore, CA; GSC=Geological Survey of Canada Radiocarbon Dating Laboratory, Ottawa, ON. 2 All terrestrial samples were calibrated using the INTCAL13 calibration curve, and all marine samples were calibrated using the MARINE13 calibration curve (Reimer et al. 2013). 74     Figure 2.3 Age-Altitude Plot of all limiting and index points used in this study. Letter and number labels correspond with data point site locations in Figure 2.2 and data point details in Table 2.4. Time ranges for data points indicate 2-sigma calibrated ranges, the elevation of these ranges is set at paleo-mean sea level for Index Points, and actual measured elevations for limiting points. Vertical lines indicate 95% confidence ranges for vertical error, and they cross the age range at the median age of each data point.     75  Table 2.5 Detailed list of most common or key diatoms observed in Livingstone Core Samples. Salinity Class (1 = halophobic, 2 = oligohalobous indifferent, 3 = oligohalobous halophilic, 4 = mesohalobous, 5 = polyhalobous) and percent of total sample assemblage given in parentheses after each species name. Location and Core Sample Depth Total diatoms confidently identified ([total IDs], [total n species]) Diatom species contributing >2.5% to assemblage ([Salinity Class], [Percent Abundance]) Paleoenvironment/ Paleosalinity Tsook Lake, TL#1 72 cm 299, 17 Frustulia rhomboides (1, 3.7%), Fragilaria sp. cf. elliptica or pinnata (2, 35.8%), Nitzschia amphibioides (2, 19.7%), Fragilaria construens (2, 10.0%), Navicula sp. cf. radiosa or leptostriata (2, 8.7%), Pseudostaurosira brevistriata (3, 9.0%); also Eunotia spp. and Gomphonema spp. (all salinity class 1-2) Freshwater pond or lake Tsook Lake, TL#1 74 cm 294, 36 Fragilaria sp. cf. elliptica or pinnata (2, 25.3%), F. construens (2, 21.8%), Staurosirella leptostauron (2, 6.1%), Stauroneis phoenicenteron (2, 2.7%), Pseudostaurosira brevistriata (3, 29.3%), Fragilaria construens var. subsalina (3, 3.4%), Coscinodiscus sp.* (5, 1%) Freshwater pond or lake, newly isolated Tsook Lake, TL#1 76 cm 278, 32 Fragilaria construens (2, 15.1%), F. sp. cf. elliptica or pinnata (2, 8.6%), Staurosirella leptostauron (2, 5.0%), %), Pseudostaurosira brevistriata (3, 43.9%), Fragilaria construens var. subsalina (3, 4.0%), Thalassiosira sp. or Coscinodiscus sp.* (5, 1.1%) Freshwater pond or lake, newly isolated Tsook Lake, TL#1 78 cm 287, 36 Fragilaria construens (2, 22.3%), Gyrosigma acuminatum (2, 7.3%), Staurosirella leptostauron (2, 4.5%), Stauroneis phoenicenteron (2, 3.1%), Fragilaria sp. cf. elliptica or pinnata (2, 3.1%), Pseudostaurosira brevistriata (3, 16.7%), Cocconeis placentula (3, 5.9%), Epithemia adnata (3, 4.2%), Amphora libyca (3, 3.1%), Tabularia fasciculata* (4, 2.4%), Thalassiosira sp. or Coscinodiscus sp. (5, 4.5%), Cocconeis costata* (5, 2.4%) Newly isolated freshwater pond, upper estuary, or freshwater-dominant lagoon with marine incursions Tsook Lake, TL#1 82 cm 277, 29 Cymatopleura elliptica (2, 17.7%), Fragilaria neoproducta (2, 5.8%), F. sp. cf. elliptica or pinnata (2, 3.2%), F. construens var. subsalina (3, 17.7%), Epithemia adnata (3, 13.0%), Amphora libyca (3, 5.4%), Cocconeis placentula (3, 4.0%), Craticula cuspidata (3, 2.5%), Navicula digitoradiata (4, 8.7%), Opephora olsenii (4, 4.0%), Thalassiosira sp. or Coscinodiscus sp. (5, 6.1%), Cocconeis costata (5, 2.9%) Mixed fresh, brackish, and marine assemblage, environment transitioning from nearshore/ lagoon/estuary to freshwater    76  Location and Core Sample Depth Total diatoms confidently identified ([total IDs], [total n species]) Diatom species contributing >2.5% to assemblage ([Salinity Class], [Percent Abundance]) Paleoenvironment/ Paleosalinity Tsook Lake, TL#1 84 cm 265, 31 Cymatopleura elliptica (2, 10.9%), Fragilaria neoproducta (2, 4.5%), F. sp. cf. elliptica or pinnata (2, 4.5%), F. construens var. subsalina (3, 9.8%), Epithemia adnata (3, 7.9%), Amphora libyca (3, 4.5%), Navicula digitoradiata (4, 21.1%), Rhopalodia acuminata (4, 3.8%), Cocconeis scutellum (4, 3.0%), Thalassiosira sp. or Coscinodiscus sp. (5, 5.3%), Cocconeis costata (5, 3.0%) Mixed fresh, brackish, and marine assemblage, environment transitioning from nearshore/ lagoon/estuary to freshwater Tsook Lake, TL#1 87 cm 294, 22 Navicula digitoradiata (4, 68.7%), Cocconeis scutellum (4, 7.8%), C. costata (5, 5.4%), Thalassiosira sp. or Coscinodiscus sp. (5, 2.7%) Brackish/Marine, likely nearshore coastal Tsook Lake, TL#1 95 cm 197, 15 Navicula digitoradiata (4, 68.7%), Scoliopleura tumida (4, 7.1%), Tabularia fasciculata (4, 2.5%), Cocconeis costata (5, 22.8%), Rhabdonema arcuatum (5, 20.8%), Thalassiosira decipiens (5, 5.6%), Thalassiosira sp. or Coscinodiscus sp. (5, 5.6%), Trachyneis aspera (5, 2.5%) Brackish/Marine, likely nearshore coastal Rifle Range Lake 1, RR1#1 185 cm 272, 10 Eunotia incisa (1, 18.4%), E. serra var. diadema (1, 4.8%), Aulacoseira spp. (2, 33.1%), Stauroneis phoenicenteron (2, 12.13%), Stauroneis spp. (2, 9.6%), Pinnularia maior (2, 11.4%), P. microstauron (2, 8.5%) Freshwater pond or lake Rifle Range Lake 1, RR1#1 287 cm 292, 15 Fragilaria construens (2, 72.3%), Sellaphora sp. cf. pupula or laevissima (2, 9.9%), Achnanthes exigua (2, 6.5%), also several different Gomphonema spp. (2) Freshwater pond or lake Rifle Range Lake 1, RR1#1 289 cm 299, 28 Fragilaria construens (2, 33.8%), F. sp. cf. elliptica or pinnata (2, 26.1%), Achnanthes exigua (2, 4.3%), Pseudostaurosira brevistriata (3, 11.4%) Freshwater pond or lake, newly isolated Rifle Range Lake 1, RR1#1 291 cm 295, 34 Tabellaria spp. (fenestrata or flocculosa) (1, 5.8%), Fragilaria construens (2, 18.3%), F. sp. cf. elliptica or pinnata (2, 12.5%), Pseudostaurosira robusta (2, 2.7%), P. brevistriata (3, 25.1%), Rhopalodia gibba (3, 3.4%), also many different Gomphonema spp. (2) and Cymbella/Encyonema spp. (2)   Freshwater pond or lake, newly isolated 77  Location and Core Sample Depth Total diatoms confidently identified ([total IDs], [total n species]) Diatom species contributing >2.5% to assemblage ([Salinity Class], [Percent Abundance]) Paleoenvironment/ Paleosalinity Rifle Range Lake 1, RR1#1 293 cm 290, 37 Gyrosigma acuminatum (2, 12.8%), Fragilaria sp. cf. elliptica or pinnata (2, 10.6%), F. construens var. venter (2, 9.3%), F. construens (2, 8.6%), Pseudostaurosira brevistriata (3, 15.9%), Fragilaria construens var. subsalina (3, 4.8%), Fallacia pygmaea (4, 7.6%), Opephora olsenii (4, 3.4%) Newly isolated freshwater pond or lake with some brackish species Rifle Range Lake 1, RR1#1 295 cm  289, 37 Fragilaria construens var. subsalina (3, 6.6%), Cocconeis scutellum var. parva (4, 12.5%), C. scutellum (4, 10.4%), Opephora olsenii (4, 11.4%), Achnanthes delicatula ssp. hauckiana (4, 8.0%), Tabularia fasciculata (4, 6.6%), Mastogloia pumila (4, 3.8%), Campylodiscus clypeus (4, 3.1%), Navicula digitoradiata var. minima (4. 3.1%), Fallacia litoricola (5, 8.7%), Opephora marina (5, 3.8%) Brackish/Marine, likely nearshore coastal Rifle Range Lake 1, RR1#1 297 cm 288, 32 Gyrosigma acuminatum (2, 3.1%), Pseudostaurosira brevistriata (3, 7.3%), Tabularia fasciculata (4, 13.9%), Cocconeis scutellum (4, 12.5%), C. scutellum var. parva (4, 5.9%), Navicula digitoradiata (4, 9.4%), Nitzschia sigma (4, 4.2%), Opephora olsenii (4, 3.1%), Thalassiosira sp. or Coscinodiscus sp. (5, 13.9%), Cocconeis costata (5, 6.9%), Fallacia litoricola (5, 2.8%) Brackish/Marine, likely nearshore coastal Digby Island Lake 1, DL#1 155 cm 295, 21 Tabellaria spp. (fenestrata or flocculosa) (1, 13.6%), Frustulia rhomboides (1, 12.5%), Eunotia incisa (1, 3.1%), Aulacoseira spp. (2, 43.7%), Navicula cf. radiosa (2, 4.7%), Encyonema gracilis (2, 3.7%), Pinnularia interrupta (2, 3.1%) Freshwater pond/ or lake Digby Island Lake 1, DL#1 215 cm 298, 26 Eunotia incisa (1, 10.7%), E. cf. minor (1, 5.4%), other Eunotia spp. (salinity classes 1-2, 6.4%), Aulacoseira spp. (2, 16.4%), Navicula cf. radiosa (2, 3.7%), Encyonema gracilis (2, 3.0%), Gomphonema spp. (2, 2.7%), Cocconeis placentula (3, 33.6%) Freshwater pond/ or lake Digby Island Lake 1, DL#1 220 cm 298, 42 Aulacoseira spp. (2, 31.2%), Fragilaria sp. cf. elliptica or pinnata (2, 18.4%), F. construens (2, 16.4%), Cocconeis placentula (3, 3.0%) Freshwater pond/ or lake Digby Island Lake 1, DL#1   225 cm 296, 22 Fragilaria sp. cf. elliptica or pinnata (2, 36.8%), Achnanthes joursacense (2, 8.8%), A. oestrupii (2, 4.7%), Fragilaria construens (2, 2.7%), Pseudostaurosira brevistriata (3, 17.9%), Martyana martyi (3, 16.2%) Freshwater pond/ or lake, newly isolated  78  Location and Core Sample Depth Total diatoms confidently identified ([total IDs], [total n species]) Diatom species contributing >2.5% to assemblage ([Salinity Class], [Percent Abundance]) Paleoenvironment/ Paleosalinity Digby Island Lake 1, DL#1 230 cm 293, 26 Fragilaria sp. cf. elliptica or pinnata (2, 20.1%), Achnanthes exigua (2, 9.2%), Amphora pediculus (2, 5.1%), Gyrosigma attenuatum (2, 4.8%), Fragilaria construens (2, 4.4%), Staurosirella leptostauron (2, 3.4%), Achnanthes joursacense (2, 3.1%), Pseudostaurosira brevistriata (3, 34.8%), Amphora libyca (3, 2.7%), also one Rhabdonema arcuatum and one Thalassiosira sp. or Coscinodiscus sp. (both salinity class 5) Freshwater pond/ or lake, newly isolated Digby Island Lake 1, DL#1 232 279, 20 Fragilaria sp. cf. elliptica or pinnata (2, 23.0%), F. construens (2, 6.5%), Gyrosigma attenuatum (2, 6.5%), Amphora pediculus (2, 5.7%), Pseudostaurosira brevistriata (3, 41.2%), Amphora libyca (3, 6.1%), also one Rhabdonema arcuatum and several Thalassiosira sp. or Coscinodiscus sp. (both salinity class 5) Newly isolated freshwater pond or lake Digby Island Lake 1, DL#1 236 293, 29 Fragilaria sp. cf. elliptica or pinnata (2, 11.2%), Amphora pediculus (2, 3.1%), %), Pseudostaurosira brevistriata (3, 19.8%), Amphora libyca (3, 9.6%), Fragilaria construens var. subsalina (3, 6.8%), Cocconeis placentula (3, 3.8%), Martyana martyi (3, 3.1%), Cocconeis scutellum (4, 3.4%), Thalassiosira sp. or Coscinodiscus sp. (5, 17.7%), Rhabdonema arcuatum (5, 3.4%), Cocconeis costata (5, 2.7%) Mixed fresh, brackish, and marine assemblage, environment transitioning from nearshore/ lagoon/estuary to freshwater Digby Island Lake 1, DL#1 238 219, 29 Amphora libyca (3, 26.0%), Epithemia adnata (3, 2.7%), %), Fragilaria construens var. subsalina (3, 2.7%), Cocconeis scutellum (4, 5.0%), Thalassiosira sp. or Coscinodiscus sp. (5, 21.0%), Thalassiosira decipiens (5, 10.5%), Rhabdonema arcuatum (5, 6.4%), Cocconeis costata (5, 3.2%) Mixed fresh, brackish, and marine assemblage, environment transitioning from nearshore/ lagoon/estuary to freshwater Digby Island Lake 1, DL#1 245 cm 300, 20 Cocconeis scutellum (4, 13.7%), Navicula digitoradiata (4, 7.7%), Scoliopleura tumida (4, 3.3%), Tabularia fasciculata (4, 3.0%), %), Rhabdonema arcuatum (5, 32.3%), Thalassiosira decipiens (5, 13.3%), Cocconeis costata (5, 10.3%), Thalassiosira sp. or Coscinodiscus sp. (5, 3.7%), Trachyneis aspera (5, 2.7%) Brackish/Marine, likely nearshore coastal Bencke Lagoon, BL#1 5 cm 296, 33 Aulacoseira spp. (2, 31.8%), Fragilaria construens (2, 3.0%), Diploneis finnica (2, 2.7%), Fallacia spp. (4, 6.4%), Paralia sulcata (5, 37.8%) Freshwater pond, upper estuary, or freshwater-dominant lagoon with marine incursions 79  Location and Core Sample Depth Total diatoms confidently identified ([total IDs], [total n species]) Diatom species contributing >2.5% to assemblage ([Salinity Class], [Percent Abundance]) Paleoenvironment/ Paleosalinity Bencke Lagoon, BL#1 10 cm 293, 40 Tabellaria spp. (fenestrata or flocculosa) (1, 2.7%), Aulacoseira spp. (2, 30.4%), Fragilaria construens (2, 3.1%), Diploneis finnica (2, 3.1%), Fallacia spp. (4, 5.8%), Paralia sulcata (5, 26.6%) Freshwater pond, upper estuary, or freshwater-dominant lagoon with marine incursions Bencke Lagoon, BL#1 15 cm 303, 36 Tabellaria spp. (fenestrata or flocculosa) (1, 7.3%), Aulacoseira spp. (2, 34.7%), Fragilaria construens (2, 12.2%), F. sp. cf. elliptica or pinnata (2, 9.6%), Cymbella cistula (2, 3.0%), Gomphonema subtile (2, 2.6%), Navicula radiosa (2, 2.6%) Freshwater pond, close to but above highest high tide Bencke Lagoon, BL#1 20 cm 305, 46 Tabellaria spp. (fenestrata or flocculosa) (1, 3.0%), Aulacoseira spp. (2, 22.3%), Fragilaria construens (2, 11.8%), F. sp. cf. elliptica or pinnata (2, 6.9%), Cymbella cistula (2, 4.6%), Achnanthes joursacense (2, 4.3%), Lindavia radiosa (2, 3.3%), Staurosirella lapponica (2, 2.6%), Rhopalodia gibba (3, 4.9%), Pseudostaurosira brevistriata (3, 4.3%), also a small number of Fallacia spp. (4, 2.0%) and Paralia sulcata (5, 2.0%) Freshwater pond, close to but above highest high tide Bencke Lagoon, BL#1 25 cm 298, 39 Tabellaria spp. (fenestrata or flocculosa) (1, 10.0%), Fragilaria construens (2, 12.1%), F. sp. cf. elliptica or pinnata (2, 10.1%), Aulacoseira spp. (2, 9.7%), Cocconeis pseudothumensis (2, 8.1%), Achnanthes oestrupii (2, 6.0%), Staurosirella lapponica (2, 6.0%), Pseudostaurosira brevistriata (3, 10.1%), Cocconeis placentula (3, 5.0%) Freshwater pond Bencke Lagoon, BL#1 27 cm 286, 31 Fragilaria sp. cf. elliptica or pinnata (2, 14.6%), Amphora pediculus (2, 7.3%), Achnanthes joursacense (2, 5.9%), Gyrosigma acuminatum (2, 5.6%), Cocconeis pseudothumensis (2, 3.8%), Fragilaria construens (2, 3.8%), Navicula aurora (2, 3.1%), Staurosirella pinnata var. intercedens (2, 2.8%), Pseudostaurosira brevistriata (3, 29.7%) Freshwater pond Optimism Bay, OB#1 300 cm 278, 38 Tabularia fasciculata (4, 7.6%), Tryblionella aerophila (4, 5.4%), Gyrosigma balticum (4, 5.0%), Bacillaria socialis (4, 4.7%), Gyrosigma fasciola (4, 4.7%), Seminavis ventricosa (4, 4.0%), Nitzschia sigma (4, 3.6%), Psammodictyon panduriforme var. delicatulum or Tryblionella aerophila (4, 2.9%), Navicula transistans (5, 27.7%), Thalassiosira sp. or Coscinodiscus sp. (5, 8.3%), Cocconeis costata (5, 2.5%), also several freshwater species such as Gyrosigma acuminatum*, Fragilaria sp. cf. elliptica or pinnata*, and Surirella brebissonii*    Brackish/Marine, nearshore or intertidal 80  Location and Core Sample Depth Total diatoms confidently identified ([total IDs], [total n species]) Diatom species contributing >2.5% to assemblage ([Salinity Class], [Percent Abundance]) Paleoenvironment/ Paleosalinity Optimism Bay, OB#1 312 cm 296, 45 Aulacoseira spp. (2, 4.7%), Craticula halophila (3, 4.1%), Achnanthes delicatula ssp. hauckiana (4, 13.5%), Navicula digitoradiata (4, 7.8%), Tabularia fasciculata (4, 6.8%), Psammodictyon panduriforme var. delicatulum or Tryblionella aerophila (4, 4.7%), Melosira sp. cf. nummuloides or moniliformis (4, 4.4%), Amphora coffeaeformis (4, 2.7%), Diploneis interrupta (4, 2.7%), Navicula transistans (5, 11.1%), Thalassiosira sp. or Coscinodiscus sp. (5, 5.7%), Cocconeis costata (5, 3.4%), Odontella sp. cf. rhombus or aurita (5, 3.4%) Brackish/Marine, marine transgressive intertidal zone or estuary Optimism Bay, OB#1 315 cm 5, 2 Nearly barren of diatoms, except for some very poorly preserved specimens that appear to be Nitzschia dissipata (2, 40%) and Craticula halophiliodes (3, 60%). Staple C and N analyses indicate that this is a terrestrial or freshwater peat/paleosol. Samples in the same strata below this and in core OB#2 were barren of diatoms. Terrestrial or freshwater Optimism Bay, OB#2 95 cm 301, 44 Gyrosigma acuminatum (2, 2.7%), Craticula halophila (3, 4.0%), Navicula digitoradiata (4, 12.6%), Amphipleura cf. rutilans (4, 7.3%), Tabularia fasciculata (4, 7.3%), Achnanthes delicatula ssp. hauckiana (4, 3.7%), Gyrosigma fasciola (4, 3.7%), Nitzschia sigma (4, 3.3%), Bacillaria socialis (4, 2.7%), Psammodictyon panduriforme var. delicatulum (4, 2.7%), Navicula transistans (5, 14.3%), Thalassiosira cf. eccentrica (5, 3.7%) Brackish/Marine, marine transgressive intertidal zone or estuary Optimism Bay, OB#2 109 cm 299, 55 Navicula digitoradiata (4, 21.4%), Tabularia fasciculata (4, 8.0%), Achnanthes delicatula ssp. hauckiana (4, 3.3%), A. brevipes (4, 3.0%), A. cf. parvula (4, 3.0%), Gyrosigma fasciola (4, 3.0%), Melosira sp. cf. nummuloides or moniliformis (4, 3.0%), Amphipleura cf. rutilans (4, 2.7%), Navicula transistans (5, 6.4%), Thalassiosira pacifica (5, 3.3%), T. cf. eccentrica (5, 2.7%), Thalassiosira sp. or Coscinodiscus sp. (5, 2.7%), also 12.5% total freshwater species such as Gyrosigma acuminatum*, Craticula halophila*, Fragilaria spp.* Brackish/Marine, marine transgressive intertidal zone or estuary Optimism Bay, OB#2 111.5 cm 287, 42 Navicula digitoradiata (4, 13.9%), Tabularia fasciculata (4, 10.1%), Achnanthes delicatula ssp. hauckiana (4, 6.3%), Melosira sp. cf. nummuloides or moniliformis (4, 4.9%), Achnanthes brevipes (4, 3.1%), Thalassiosira sp. or Coscinodiscus sp. (5, 17.4%), Navicula transistans (5, 4.9%), Thalassiosira cf. eccentrica (5, 4.9%), Achnanthes cf. groenlandica (5, 3.8%), Cocconeis costata (5, 2.8%), Tryblionella acuminata (5, 2.8%),  also a couple halophobic  Tabellaria spp. (fenestrata or flocculosa) and several other freshwater species  Brackish/Marine, marine transgressive intertidal zone or estuary 81  Location and Core Sample Depth Total diatoms confidently identified ([total IDs], [total n species]) Diatom species contributing >2.5% to assemblage ([Salinity Class], [Percent Abundance]) Paleoenvironment/ Paleosalinity Optimism Bay, OB#2 113 cm 290, 29 Navicula digitoradiata (4, 52.4%), Gyrosigma balticum (4, 5.2%), Scoliopleura tumida (4, 4.1%), Nitzschia sigma (4, 3.8%), Thalassiosira sp. or Coscinodiscus sp. (5, 4.8%), Thalassiosira cf. eccentrica (5, 3.1%), Tryblionella acuminata (5, 3.1%), Didymosphenia geminata* (salinity class 2, 1 specimen) Brackish/Marine, marine transgressive intertidal zone or estuary * Indicates species that made up less than 2.5% of the total assemblage but which are either notable for reasons discussed in text or are important to some facet of the interpretation of the sample.82  2.3.1 Livingstone Sediment Cores Expanded descriptions of Livingstone sediment cores not included in the published version of this Chapter (Letham et al. 2016) are provided in Appendix A.3.  2.3.1.1 Tsook Lake Core (TL#1, 49.7 m asl) The highest elevation core is from Tsook Lake, north of Metlakatla on the Tsimpsean Peninsula. The elevation of the basin's sill is 49.7 m asl. Core TL#1 (Figure 2.4) contains a sequence of marine sand and silt transitioning to freshwater gyttja and fragmental herbaceous peat and fragmental granular peat (following the terminology of Schnurrenberger et al. 2003:151, and henceforth ‘peat’). An Arctostaphylos sp. seed (a shrub species known as an initial colonizer of deglaciated landscapes [Mann and Streveler 2008:207]) from a brackish and marine-diatom dominated context dates 15,090–14,365 cal. BP (D-AMS 009956). Seeds from a freshwater diatom-dominated zone with minor brackish/marine influence located below the transition to gyttja and peat date 14,782–13,714 cal. BP (D-AMS 009955), indicating that the Tsook Lake basin was likely only being flooded by exceptionally high tides at this time. A relatively gradual transition from marine/brackish to freshwater diatom assemblages over as much as 1200 years between these two dated samples may be indicative of a gradual RSL decline at this time. Twigs and a small cone from just above the transition to dark brown decomposed peat/gyttja date 13,971–13,330 cal. BP (D-AMS 009954) and provide a latest possible date for the full isolation of this basin from marine influence. 83   Figure 2.4 Tsook Lake Core TL#1 log, photo, and diatom analysis results. Diatom species comprising 7% or greater of the total assemblage of any given sample are shown on the expanded bar graph.  2.3.1.2 Rifle Range Lake 1 Core (RR1#2, 35 m asl) Rifle Range Lake 1 is located on the east side of Kaien Island, the furthest east of any of our core samples. It has an estimated sill elevation of 35 m asl. Core RR1#2 (Figure 2.5) contains a sharp transition from brackish and marine diatom-dominated sand and silt to freshwater gyttja and peat. Mixed but indistinguishable plant matter from a thin dark lens of bedded organics about 15 cm below the transition dates to 14,090–13,458 cal. BP (D-AMS 008741). Several small twig fragments from 11 cm above the transition date 14,055–13,345 cal. BP (D-AMS 008740). The very tight chronological succession of these two dates, along with the abrupt transition to fully freshwater conditions indicates that RSL passed very quickly over this elevation. However, the date in the brackish/marine sediment contradicts other dates in this study that suggest that RSL 84  had passed well below 35 m asl at or before this time. Possible reasons for this are discussed in Section 2.4.2.1.   Figure 2.5 Rifle Range Lake core RR1#2 log, photo, and diatom analysis results. Diatom species comprising 8% or greater of the total assemblage of any given sample are shown on the expanded bar graph.  2.3.1.3 Cores from Bogs on Northern Digby Island (DIB1#1, 17.2 m asl; and NDB#1, 17 m asl) DIB1#1 (17.2 m asl) and NDB#1 (17 m asl) are sphagnum bogs with standing water on northern Digby Island (Figure 2.2). Cores from each contain basal blue-gray clay resembling the glacio-marine sediment observed in the study area overlain by sharp transitions to peat (Appendices A.4 and A.5). However, no diatoms were observed in samples from near to these transitions. A stick of wood lying diagonally in the lowest instance of peat in DIB1#1 yielded a relatively recent age 85  of 8295-8028 cal. BP (D-AMS-005844), suggesting that it is intrusive from above or indicating an erosional unconformity. Two dates on samples from higher up in the peat in NDB#1 yielded ages of 8169-7626 cal. BP (D-AMS 009950) and 10,171–9521 cal. BP (D-AMS 009948). These three dates serve as upper limiting points for RSL during the early Holocene.  2.3.1.4 Digby Island Lake 1 Core (DL1#1, 15.2 m asl) Digby Island Lake 1 is one of several lakes in a larger basin at the centre of Digby Island that would have been isolated from the ocean by a long and narrow channel that runs to the south end of the island with a maximum sill height of 15.2 m asl. Core DL1#1 (Figure 2.6) contains a transition from marine and brackish diatom-dominated clayey sandy silt to brown silty mud with a transitional sequence of mixed diatom assemblages to fully freshwater assemblages, overlain by freshwater peat and gyttja. Organic macrofossils of sufficient size for radiocarbon dating were not found in the marine or brackish sediment. Several small twig fragments from just above the transition to medium brown silty sand produced a date of 15,013–13,859 cal. BP (D-AMS 008745). Sediment from 2 cm above these twigs contains only 3.5 percent brackish and marine diatoms, indicating that this date is a reasonable approximation of the time just before the basin became isolated from marine incursions. 86   Figure 2.6 Digby Island Lake 1 core DL1#1 log, photo, and diatom analysis results. Diatom species comprising 10% or greater of the total assemblage of any given sample are shown on the expanded bar graph.  2.3.1.5 Bencke Lagoon Cores (BL#1 and BL#4, 2.4 m asl) Bencke Lagoon is a shallow ‘L’-shaped body of water located in a low-relief area at the end of Scott Inlet, east of Metlakatla (Figure 2.7). The lagoon currently drains over a 2.4 m asl sill, putting it within the upper tidal range for the Prince Rupert area (i.e. just above higher high water mean tide [HHWMT], Table 2.2), and therefore flooded by several high tides each month. The result is slightly brackish water within the lagoon. 87   Figure 2.7 Orthophoto of a section of northern Venn Pass, showing Bencke Lagoon, Scott Inlet, and Optimism Bay. Note the extensive sand and mudflats exposed at low tide. Livingstone core locations are indicated by yellow circles, paleomarine sediment exposures indicated by yellow squares. Letters in parentheses correspond with test locations in Figure 2.2.  Two cores taken several meters from each other (BL#1 and BL#4) contain a sequence beginning with coarse clastic material that is likely glacial till overlain by laminated gray silty sand transitioning to clay with marine mollusc shells that coarsens upwards to sand with marine mollusc shells. Sand without marine mollusc shells but with reworked fragments of marine diatoms overlies these layers. Subsequent to the deposition of the till, this sequence likely 88  indicates a low energy subtidal environment transitioning to intertidal, and eventually to high intertidal. The upper section of core BL#1 (Figure 2.8) contains a sharp transition to gyttja with a remarkably diverse freshwater diatom assemblage (Table 2.5), indicating a transition to a pond or slow-moving creek. The last few centimeters of sediment above this are light brown/tan silty mud containing a freshwater diatom assemblage similar to the gyttja below with the addition of the brackish-marine species Paralia sulcata and small Fallacia spp. The inclusion of these brackish-marine species into an otherwise diverse freshwater species-dominated context suggests that they are allochtonous, carried in by either very high tides or by storm surges. Stable isotope and elemental composition measurements of this gyttja yielded values that are intermediate between the average values for known marine and freshwater/terrestrial baseline samples, though overall closer to freshwater conditions (Table 2.3, Figure 2.9). This suggests that the organic matter at the top of the core is composed of a mixture of both freshwater and marine-derived materials, and was deposited under conditions similar to those of today. 89   Figure 2.8 Upper section of Bencke Lagoon core BL#1 log, photo, and diatom analysis results. Diatom species comprising 5% or greater of the total assemblage of any given sample are shown on the expanded bar graph.  90   Figure 2.9 A: Plot of δ13C vs CORG/NTOTAL for known marine sediment samples (blue diamonds), known terrestrial samples (green diamonds), a sample of organic-rich sediment from the upper layer in core BL#1 (yellow triangle), and samples from the organic-rich layer at the bottom of cores OB#1 and OB#2 (red triangles). 9B: Plot of δ13C vs δ15N values for the same samples. There are slightly fewer marine samples represented because not all of these samples yielded reliable δ15N values.  A Balanus sp. shell from the lowest instance of shell in BL#4 dates 14,970–14,190 cal. BP (D-AMS 008752), and a Mytilus sp. shell and a Balanus sp. shell from the highest instance of shell in BL#1 date 15,284–14,675 cal. BP (D-AMS 008751) and 14,980–14,230 cal. BP (D-AMS 91  009953), respectively, though marine molluscs from early postglacial times may be slightly younger than measured if they are affected by more deep water mixing from isostatic depression (Section 2.2.3, Hutchinson et al. 2004b). A bulk sample of organic-rich sediment from the lowest instance of freshwater gyttja dates 14,833–13,738 cal. BP (UBA-29065) suggesting that the highest tides passed below 2.4 m asl (and therefore below their current position) by this time, although again, there may be an element of immediate-postglacial old carbon effect affecting this age (see Sections 2.2.3; 2.4.1).  Seeds from just below the transition from freshwater gyttja to the silty mud with apparent marine incursions date 13,722–13,160 cal. BP (D-AMS 009952), and a twig from directly above this transition dates 13,255–13,065 cal. BP (D-AMS 009951). The proximity of these very old sediments to the surface indicates that the Holocene sediment sequence has been truncated in this location.  2.3.1.6 Optimism Bay Cores (OB#1 and OB#2, -1.36 m asl) Optimism Bay is an informal name given to a well-sheltered bay with an extensive intertidal mudflat located north of the entrance to Scott Inlet, about 1 km northwest of Bencke Lagoon (Figure 2.7). Intertidal sediment obscures any sill that may exist at the mouth of the bay, so data point elevations are subtracted from the elevation of the beach surface at the core location (-1.36 m asl).  Cores OB#1 and OB#2 (Figure 2.10) are only a few meters apart. Both contain a dark reddish brown organic-rich layer (-5.0 m asl to -4.86 m asl in OB#1 and -6.36 m asl to -6.0 m asl in 92  OB#2) beneath several meters of intertidal or nearshore marine sand with marine shell hash. The buried organic-rich layer contains only a few poorly preserved oligohalobous indifferent and oligohalobous halophilic diatoms that could be allochtonous in OB#1, and no preserved diatoms in OB#2. Stable isotope analyses of two samples from this layer in OB#1 and four samples in OB#2 yielded δ13C and δ15N values within the range of values for our known freshwater/terrestrial sediments (Table 2.3, Figure 2.9; Figure 2.10). Combined with the notable scarcity of diatoms, these results suggest that this deposit was subaerially exposed near to the shore but without direct tidal influence, and that the deposit is a paleosol or peat.  93   Figure 2.10 Optimism Bay Cores OB#1 and OB#2 logs, photos, and stable isotope analysis sample locations (coloured squares). The sediment directly above this layer contains a diverse assemblage of primarily brackish and marine diatom species, though samples also contain between 4 and 18% freshwater diatom species. Stable isotope values of four samples from this zone all differ from those of the peat/paleosol, though exhibit both δ13C values and C/N ratios closer to freshwater/terrestrial samples than the rest of the marine samples that we tested (Appendix A.2), suggesting some degree of mixing of organic sediments. In both cores, the diatom assemblage and stable isotope results indicate a marine transgression over a terrestrial peat or soil; the 3–4 m of shelly sands 94  above these sequences indicate a full transition to an intertidal or nearshore marine environment. There is no indication of terrestrial conditions in either of the cores again.  Eight radiocarbon dates from both cores date the sequence. A bulk sample of organic-rich sediment from the very lowest instance of terrestrial material in OB#2 dates 14,163–13,436 cal. BP (UBA-29067), though, as with the gyttja in Bencke Lagoon, this sample may also be up to several centuries younger if a postglacial hard water reservoir effect has affected the carbonates in the sediment. The degree of this effect is constrained, however, by the age of the large piece of wood several centimeters above the base of the terrestrial layer: 13,772–13,572 cal. BP (D-AMS 008750). A bulk sample of organic-rich sediment from before the transition from freshwater/terrestrial conditions to the brackish diatom-dominated sediment dates 12,700–11,823 cal. BP (UBA-29066), providing an estimate for the last time this area was above tidal influence. In the brackish/marine sediments above the transition in both cores, four dates on plant macrofossils (D-AMS 008747, D-AMS 008749) and shell (D-AMS 008753, D-AMS 008754) all have calibrated age ranges between about 11,230 cal. BP and 10,700 cal. BP.  Notable amongst the diatom assemblage of the marine transgressive sediment in OB#2 was a single specimen of Didymosphenia geminata (Appendix A.6), a nuisance species once considered invasive to the Northwest Coast, though argued by Bothwell and colleagues (2014; Taylor and Bothwell 2014) to be native to North America. This specimen is in stratigraphically secure context and well constrained by the radiocarbon dates to between 12,000 and 11,000 years old (Figure 2.10), making it the oldest identified specimen in North America and having 95  significant implications for our understanding of the origins of this species’ presence on the continent (Max Bothwell, personal communication, 2016).  2.3.1.7 Other Isolation Basin Cores At or Around Current Sea Level (SL#1, 2.2 m asl; PL#1, 0.75 m asl; RA#2, 0 m asl; GLP#1, 0 m asl)  We cored four other basins at or near current sea level: Salt Lake, Russell Arm, Philip's Lagoon, and an unnamed lagoon east of Auriol Point (Figure 2.2). Cores from the latter two contained only marine and intertidal sediment sequences and provide only limited RSL lower constraining information (Appendices A.7 and A.8). Core SL#1 (Appendix A.9) from Salt Lake, an isolated body of water with a 2.2 m asl sill and a minor tidal influence, contained laminated blue-gray clay, silt, and fine sand directly overlain by coarse sand with marine mollusc shells that date only 2660-2345 cal. BP (D-AMS 005839). Salt Lake is currently too high above sea level to support marine shellfish, so this date indicates that RSL was at least high enough for this area to be fully intertidal in the later Holocene. The lower laminated sediments in the core appear marine and suggest higher RSL earlier than the dated shell, though they resemble glacio-marine sediment observed in other cores. If this is the case then there is a significant erosional unconformity at the contact between these sediments and the shelly sand above, perhaps caused by Holocene RSL fluctuations.  Salt Lake drains into Russell Arm, which has an isolation basin with a bedrock sill that is 0 m asl. The ∼4 m sediment sequence sitting on bedrock in core RA#2 contained only intertidal and marine sediment from the last 3400 years; a shell-rich sandy layer at the bottom dates 3394-3143 cal. BP (D-AMS 005843) and 3448-3343 cal. BP (D-AMS 005842), a massive bed of shell-free 96  well-sorted silt rich with marine diatoms above this dates 2148-1998 cal. BP (D-AMS 005841), and an overlying shell-rich layer at the top of the sequence dates 1147-924 cal. BP (D-AMS 005840) (Appendix A.10). While minimally indicating RSL at or above 0 m asl for the last 3400 years, there is some evidence for a slight upwards fluctuation in this sequence. Low-tide and subtidal sediment in the immediate area is fine gray silt, while the higher intertidal zone (i.e. the adjacent depositional environment) has sand and shell hash pushed up by wave action. These facies provide a modern analogue for the facies in the core, and lateral migration of these facies in response to RSL change is suggested by their vertical succession. Therefore, the transition from sediment rich with intertidal molluscs to well-sorted silt with marine diatoms and then back between 3400 BP and 1150 BP suggests a slight rise and then fall in RSL. This pattern is also suggested by late Holocene archaeological data and discussed in Section 2.4.1.  2.3.2 Paleomarine Deposits in Geological Exposures and Relict Paleoshoreline Landforms  Two previously identified and two newly identified paleomarine sediment beds contain raised terminal Pleistocene-aged deposits. Marine mollusc shells exposed in marine sediment 53.6 m asl near Port Simpson, 30 km north of Prince Rupert date 14,863–14,080 cal. BP (Beta-14465) and 14,649–14,019 cal. BP (Beta-14464) (Archer 1998; Fedje et al. 2004; 2005b). Clague (1984, 1985; Lowdon and Blake 1979) dated a Mya truncata shell exposed at 11 m asl on the west side of Kaien Island that produced a calibrated age of 14,211–13,569 cal. BP (GSC-2290). We identified a terminal Pleistocene paleomarine deposit in a terrestrial ESP core from a 16 m asl terrace on the isthmus between Russell Arm and Philip's Lagoon; two marine shell samples from this core dated 15,187–14,574 cal. BP (D-AMS 005852) and 15,011–14,241 cal. BP (D-AMS 97  004470). Another large bed of reworked marine mollusc shells was found exposed at 3.83 m asl in the bank of Swamp Creek on the west side of Tsimpsean Peninsula. A shell from this exposure dated 14,510–14,000 cal. BP (D-AMS 007879).  Seventeen dated samples from seven exposed paleomarine sediment deposits ranging from -0.6 m to 9 m asl have ages ranging from 11,700 cal. BP to 9000 cal. BP. These show a general trend of increasing elevation with time, tracking a marine transgression above the current sea level position in the early Holocene. Several of these samples were located within the current tidal range but were identified tens to hundreds of meters up creeks and buried under several meters of alluvial sediment and forest soil, indicating that these areas had once been intertidal under higher RSL conditions, and then that a subsequent drop in RSL caused a transition to estuarine and then terrestrial conditions (Figure 2.11). Several other locations contained molluscs dead in growth position (i.e. articulated valves sitting vertically in the sediment) within the current tidal range but above their habitat range, indicating higher sea level. 98   Figure 2.11 Orthophoto of the location of Tea Bay Creek paleomarine exposure and photograph the profile, showing sequence from marine conditions to high intertidal/salt marsh to alluvial/estuarine conditions to the current forest soil buildup. In two cases, we dated in situ butter clam (Saxidomus gigantea) specimens that provide RSL index points because of their known habitat range relative to the tidal range (Table 2.1; Carlson and Baichtal 2015:125; Foster 1991). A specimen from a shell bed containing large senile Protothaca staminea, Clinocardium nuttalli, Tresus capax, and Saxidomus gigantea in growth position exposed by a creek that has incised the intertidal zone in an unnamed estuary north of Optimism Bay dated 10,250–9952 cal. BP (D-AMS 007880, Figure 2.7). The mean elevation of this shell bed is 0.058 m asl, though butter clams are known to prefer living between 0.46 m above and 0.91 m below Lower Low Water Mean Tide (LLMWT, -2.528 m asl at Prince Rupert) (Carlson and Baichtal 2015:1254; Foster 1991), or -2.07 to -3.44 m asl around Prince Rupert. This indicates that RSL was 3.5 to 2.1 m higher when the S. gigantea were alive. An in situ S.                                                  4 Carlson and Baichtal use the range -0.91 and +0.46 m above MLLW (mean lower low water), which is a US measurement based on observed data and generally equivalent to LLMWT, a Canadian measure based on predicted tidal levels (Canadian Hydrographic Survey, personal communication, September 28, 2015). 99  gigantea shell from Tea Bay Creek that dates 10,196–9901 cal. BP (D-AMS 004468) was recovered 2.4 m asl indicates that RSL was 5.8–4.5 m higher at that time (Figure 2.11). Assuming a constant tidal range through time, these estimates place highest astronomical tide (3.66 m asl, Table 2.2) as high as 9.46 m asl by ∼10,000 years ago.  In addition to these paleomarine sediments, frequent 7–10 m asl steep-sloped linear ridges that run parallel to the modern shoreline are visible throughout the study area in LiDAR bare earth DTMs (Figure 2.12). These features resemble relict backbeach berms, and their prominence in the regional topography suggests that RSL was once stable at these positions. Archaeological deposits associated with these paleoshorelines indicate that these features were shorelines during the early Holocene (Section 2.3.3), which is consistent with the 5.8–4.5 m higher RSL indicated by the Tea Bay Creek S. gigantea. 100   Figure 2.12 Left: LiDAR-derived slope-classified map of a portion of northwest Digby Island showing inland linear ridges that likely represent stranded paleoshorelines. GbTo-64 is an archaeological site located on the modern shoreline. GbTo-183 is an archaeological site on a 5–7 m asl terrace with dates from ∼3500 cal. BP to ∼2000 cal. BP, associated with slightly higher RSL in the latter half of the Holocene. GbTo-185 is an archaeological site on a 10–12 m asl terrace from the early Holocene RSL high stand. Solid black line is the modern shoreline; light gray shading indicates ‘flooding’ to 7 m asl for reference. Intensifying colours indicate increasing slope. Right: LiDAR-derived hillshaded DEM of the same area. 2.3.3 Archaeological Sites Sixty-two dates from 28 archaeological habitation sites constrain RSL position during the Holocene. Preliminary archaeological survey of flat landforms immediately above the 7–10 m asl paleoshoreline ridges resulted in the identification of three of the oldest archaeological sites yet recorded in the Prince Rupert area. GbTo-185, on a 10–12 m asl terrace (Figure 2.12), contains evidence of concentrated and repeatedly-used campfires or hearths and stone tool making dating 9304-9028 cal. BP (D-AMS 011950) and 8348-8186 cal. BP (D-AMS 011949). Two more sites on 8–9 m asl terraces, GbTo-82 and GcTo-67, have small cultural shell-bearing 101  components that date 6728-6463 cal. BP (D-AMS 011956) and 6635-6445 cal. BP (D-AMS 011948), respectively. A paleosol directly below the cultural component at GcTo-67 provides a further 7.95 m asl upper limiting RSL point at 7170-6960 cal. BP (D-AMS 011947).  The majority of archaeological data points (n = 57) come from the basal components of large shell-bearing sites and date between 5000 cal. BP and 1000 cal. BP. These data are spread between 3.1 m asl and 10 m asl, and in general suggest RSL close to, but slightly higher than that of today (Figure 2.3).  Several archaeological sites dating 3500-1500 cal. BP were identified on paleoshorelines associated with higher RSL. Three previously unrecorded large shell-bearing sites (GcTo-66, GbTo-183, GbTo-184) were identified 60–130 m back from the modern shoreline in LiDAR DTMs. GcTo-28 is a similar previously recorded village 30 m back from the shoreline. These sites are all located on 5.5–6.5 m asl terraces fronted by low-lying 3.5–4.5 m slopes toward the modern shoreline (e.g. Figure 2.12). Basal dates from these sites vary from about 3500 BP to 1700 BP, but they all appear to have been abandoned between 2000 and 1500 BP. Sandy deposits with marine shell that we interpret to be intertidal or storm surge deposits were identified beneath the cultural layers at three of these sites. Shells from a natural deposit 2.38 m asl and 2.92 m asl beneath GcTo-66, 130 m back from the current shoreline, date 2762-2495 cal. BP (OS-119874) and 2315-2071 cal. BP (OS-119876), respectively, while a shell from 5.36 m asl beneath GbTo-184, 60 m back from the current shoreline, dates 3201-2736 cal. BP (D-AMS 007890). Taken together, the archaeological data from the last 5000 years suggests slightly 102  higher RSL until ∼2000-1500 cal. BP, with some potential fluctuations, discussed in Section 2.4.1. 2.4 Discussion The following sections summarize the RSL history that can be derived from the above data and outline the processes that likely drove the changes observed, as well as overview the significance of this particular study for regional research.  2.4.1 Prince Rupert RSL History and the Processes Driving RSL Change The age-altitude relations of our dated samples and an inferred RSL curve are shown in Figure 2.13. The RSL curve is the most parsimonious interpretation of the data. The calibrated ranges of radiocarbon dates add uncertainty to the timing of inflections and potentially more subtle nuances within the curve, especially for the terminal Pleistocene.  Figure 2.13 Plot of all data points and the preferred RSL curve for the Prince Rupert Harbour region. Time ranges for data points indicate 2-sigma calibrated ranges; the elevation of these ranges is set at paleo-mean sea level for Index Points, and actual measured elevations for limiting points. Vertical lines indicate 95% confidence ranges for vertical error, and they cross the age range at the median age of each data point. Our preferred inferred RSL curve is indicated by the solid (well constrained sections) and dashed (loosely constrained sections) line. 103   RSL was at least 50 m higher than at present when the area was deglaciated, though subsequent isostatic rebound caused a rapid drop in RSL. Marine sediment in Tsook Lake (49.7 m asl) demonstrates that deglaciation occurred at by least 15,090–14,365 cal. BP (D-AMS 009956). A gradual transition from marine to freshwater diatoms between 15,090–14,365 cal. BP (D-AMS 009956) and 14,782–13,714 cal. BP (D-AMS 009955) in Tsook lake indicates a relatively slow RSL regression between these times, though a date of 14,163–13,436 cal. BP (UBA-29067) on the first instance of paleosol/peat -6.3 m asl at Optimism Bay indicates very rapid isostatic uplift of the deglaciated landscape after Tsook Lake was isolated from marine influence. This rapid RSL drop is also indicated or constrained by dates from the paleomarine deposit near Port Simpson (53.55 m asl) and on the isthmus between Russell Arm and Philip's lagoon (15.9 m asl), the transition from marine to freshwater conditions in Digby Island Lake 1 (15.2 m asl), the paleomarine exposures on west Kaien Island (11 m asl) and in Swamp Creek (3.83 m asl), and the transition from marine to freshwater conditions in Bencke Lagoon (2.4 m asl).  There is a large degree of overlap between the date on the freshwater bulk sediment samples from Bencke Lagoon (14,833–13,738 cal. BP, UBA-29065) and Optimism Bay (14,163–13,436 cal. BP, UBA-29067) and the transition from marine to freshwater conditions nearly 50 m higher at Tsook Lake (14,782–13,714 cal. BP, D-AMS 009955), which was dated using plant macrofossils. It is likely that the bulk sample ages have been influenced by an immediate postglacial old carbon effect (Hutchinson et al. 2004b). Even if this effect pushes the dates ahead several centuries, these data demonstrate that around Prince Rupert the immediate postglacial RSL drop caused by isostatic rebound likely took less than 1000 years, and as little as a few 104  centuries. This rapid uplift rate is in line with those observed at other near-field/glaciated areas on the west coast of North America, particularly on the southern British Columbia coast (e.g. Clague et al. 1982; Hutchinson et al. 2004a; James et al., 2005, 2009a; Shugar et al., 2014).  A RSL lowstand below -6.3 m asl following initial isostatic rebound lasted for about a 2000-year interval that encompassed the Younger Dryas period (12,900–11,700 BP). This lowstand is indicated by the transition to fully freshwater conditions in Bencke Lagoon at 14,833–13,738 cal. BP (UBA-29065) and by the buried peat/paleosol in Optimism Bay, 6.3 m below current sea level. The extent of this lowstand below sea level is not constrained by any lower limiting points (Figure 2.13), though stable isotope values for the Optimism Bay peat/paleosol suggest very minor mixing of marine-derived organic material, suggesting that the lowstand did not extend much below −6.3 m asl (Section 2.3.1.6, Table 2.3, Figure 2.9). Evidence for the terminal Pleistocene lowstand is not apparent in the other low elevation cores from Philip's Lagoon (PL#1, 0.75 m asl sill) and the lagoon east of Auriol Point (GLP#1, 0 m asl sill), likely due to the erosion of sediment from this time during the subsequent RSL transgression; erosional unconformities are often produced by slow RSL rise (Green et al. 2014). The preservation of lowstand sediment at Optimism Bay and Bencke Lagoon is likely attributable to fortuitous preservation contexts. The re-introduction of marine diatoms in Bencke Lagoon at 13,255–13,065 cal. BP (D-AMS 009951, Section 2.3.1.5), the middle of the lowstand, may be indicative of fluctuations during this time that are not evident within the Optimism Bay cores, irregular storm events or very high tides, mixing of lower freshwater sediments with younger sediment during the RSL transgression, or a laboratory error. All other radiocarbon dates suggest that RSL 105  did not rise up to and above the lowstand peat/paleosol until after 12,700–11,823 cal. BP (UBA-29066), when intertidal sediments are present in both Optimism Bay cores.  A marine transgression caused RSL to rise to 6–8 m asl between 11,700 and 9000 cal. BP. Four dates on the brackish and marine diatom-rich sediments above the Optimism Bay peat/paleosol and seventeen dates on seven relict paleomarine deposits indicate that Optimism Bay was again intertidal by 11,500 BP, that RSL passed over its current position just before 11,000 BP, and that it continued upward several meters in the early Holocene. Because the elevations of these samples are not controlled by sill elevations, and because marine mollusc shells can be moved anywhere within or below the tidal range by waves, tides, and currents, these data have more elevation scatter (Figure 2.13). This may also partly be attributable to varying marine reservoir effects (Hutchinson et al. 2004b). We lend the most weight to the growth position S. giganteas from the estuary north of Optimism Bay (indicating an RSL 3.5–2.1 m asl) and from Tea Bay Creek (indicating an RSL of 5.8–4.5 m asl) for the position of the inferred RSL curve during this transgression. The similar dates on these samples, 10,250–9952 cal. BP (D-AMS 007880) and 10,196–9901 cal. BP (D-AMS 004468), respectively, and the 2.4 m elevation difference between the two suggest that the transgression was rapid. It occurred earlier and more abruptly than post-lowstand transgressions recorded on the south coast of British Columbia. The RSL rise is likely related to a well-recorded global increase in eustatic sea level between 11,650 and 7000 cal. BP (Smith et al. 2011), which includes a particularly rapid increase at the termination of the Younger Dryas associated with a meltwater pulse (Glacial Meltwater Pulse 1B) caused by dramatic warming at this time (Green et al. 2014; Liu and Milliman 2004; Smith et al. 2011). 106  This eustatic sea level rise outpaced isostatic rebound, even though the now-slower isostatic crustal response continued upward.  The early Holocene is characterized by a RSL highstand, primarily constrained by abundant 7–10 m asl paleoshoreline berms and newly identified archaeological sites on terraces associated with these berms, and loosely constrained by 17 m asl upper limiting dates from the Digby Island bogs and 0 m asl lower limiting dates from Pillsbury Cove and the lagoon east of Auriol Point. The Tea Bay Creek S. giganteas indicate that RSL rose to at least 5.8–4.5 m above its current position by 10,196–9901 cal. BP (D-AMS 004468). Taking into account a high tide of up to 3.66 m above this (Table 2.2), the 9000-8000 BP archaeological remains 10–12 m asl at GbTo-185 suggest that RSL may have continued rising another 1 or 2 m by that time. Factoring in the 6500 cal. BP archaeological remains from 8 to 9 m asl terraces at GbTo-82 and GcTo-67, these data suggest that RSL reached 6–8 m asl by 9000 years ago and remained relatively stable above its current position for the duration of the early Holocene, dropping only a couple of meters by 6500 cal. BP. This contradicts earlier RSL reconstructions for the area that inferred that RSL was below its current position between 10,000 and 5700 cal. BP (see Section 2.1.3; Clague 1984, 1985; Clague et al. 1982; Eldridge and Parker 2007).  RSL dropped to within a few meters of its current position after 6500 cal. BP and continued dropping slowly through the Holocene, albeit with some potential fluctuations. This may have been driven by continuing slower isostatic crustal response overtaking the slowing postglacial eustatic sea level rise, the latter of which completed around 6000 BP. The last 6000 years are primarily constrained by basal dates on archaeological sites that display a wide degree of scatter. 107  There are no lower limiting data between 7525-7225 cal. BP (Beta-221626, Pillsbury Cove) and 3448-3343 cal. BP (D-AMS 005842, Russell Arm). There is only a single upper limiting data point between just after 6500 BP and 5000 BP: a basal date of 6006-5733 cal. BP (OS-101646) from site GcTo-6 at 4.18 m asl suggests that RSL continued to fall from the early Holocene highstand, perhaps at a slightly increased rate. The data from 5000 cal. BP onward can be interpreted in several ways, depending on the weight attributed to specific indicators. Figure 2.14 presents two options, a more conservative general pattern of slow RSL regression that smooths out potential noise in the data, and a second option that attempts to fit all the data so that the lowest basal archaeological dates are close to or above a 2.32 m HHWMT and all lower limiting dates above RSL are at least within the relative tidal range. The latter is an exaggerated curve, but illustrates the maximum inflections from known data. Between 5000 and 3200 cal. BP the majority of archaeological basal dates are 5–6 m asl but show a subtle overall decrease in elevation until 3000 cal. BP, at which time three different sites (GbTo-4, GbTo-24, GbTo-64) have dated basal samples at or below modern HAT. This suggests a continued fall, but that RSL was still 1.5–2.5 m higher than its current position during this period. 108   Figure 2.14 Plot of all data points from the last 5000 years, and two potential RSL interpretations. The dotted line is a conservative general trend of regressing RSL that smooths out potential noise in the data while keeping most of the lowest basal archaeological data points above HHWMT (2.32 m above RSL). The dashed line attempts to fit all the data at 250 year intervals in a way that the lowest basal archaeological dates are close to or above a 2.32 m HHWMT and all lower limiting dates above RSL are at least within the relative tidal range. Time ranges for data points indicate 2-sigma calibrated ranges. Vertical lines indicate 95% confidence ranges for vertical error, and they cross the age range at the median age of each data point. The period between 3200 cal. BP and 1600 cal. BP has the largest vertical spread of data (Figure 2.14). An increase in the overall range of basal elevations during this time indicates that people are initiating settlements on higher ground. There is a slight increase in elevations of the lowest basal archaeological dates in the middle of this age range compared to those immediately preceding and following. The four large shell-bearing sites identified on 5.5–6.5 m terraces 60–130 m back from the modern shoreline (GcTo-28, GcTo-66, GbTo-183, GbTo-184) are all occupied during this time and are all abandoned between 2000 and 1500 years ago. There are five lower limiting data points that suggest higher RSL between 3000 and 2000 years ago from 109  stranded paleomarine deposits beneath archaeological sites GcTo-52, GcTo-66, and GbTo-184, and from the Salt Lake Core. The facies sequence in the Russell Arm core also suggest a slight RSL rise sometime between 3394-3143 cal. BP (D-AMS 005843) and 1147-924 cal. BP (D-AMS 005840).  Minimally, these data indicate that RSL continued to be several meters higher into the late Holocene (Figure 2.14, dotted line), though, depending on how much weight is put on the correlation between archaeological basal dates and RSL changes, they could be suggestive of a modest RSL dip and then rise (∼1–2 m) around 3200 cal. BP before ultimately falling to very close to its current position between 2000 and 1500 years ago (Figure 2.14, dashed line). The overall trend of slow RSL fall from the early Holocene highstand is likely attributed to the final influence of isostatic crustal rebound in the region. More data is required to test possible subtle late Holocene RSL fluctuations and their driving mechanisms, though they may be associated with climate fluctuations or neoglacial periods in the Coast Mountains (i.e. Clague and Mathewes 1996; Desloges and Ryder 1990;  Lamoureux and Cockburn 2005).  Most recently, historical tidal records from 1937 to 2000 indicate that RSL is rising in Prince Rupert Harbour by a rate of 1.72 ± 0.06 mm/yr, and that this is a result primarily of global eustatic sea level rise and a very slight (possibly zero) local subsidence rate of 0.7 ± 1.0 mm/yr (Larsen et al. 2003; see also James et al. 2014 for similar calculations). The measured eustatic sea level rise over the last century is likely partly attributable to anthropogenically accelerated global warming. The effects of this recent RSL rise are visible on actively eroding archaeological sites throughout the area. 110   2.4.2 Significance for Regional Studies The following sections review the significance of this particular study for regional glacial isostatic modelling and understanding regional patterns of post-glacial RSL change, as well as the implications for archaeological research.  2.4.2.1 Regional Glacial and RSL Histories This research highlights spatial variation in the timing of RSL changes not previously anticipated in the study area, particularly immediately after deglaciation. A tightly dated transition from marine conditions to freshwater conditions in the Rifle Range Lake 1 core RR1#2 suggests that RSL passed below 35 m asl between 14,090–13,458 cal. BP (D-AMS 008741) and 14,055–13,345 cal. BP (D-AMS 008740), but at the same time, samples from cores in Bencke Lagoon and Optimism Bay indicate that RSL was below its current position in those locations. One explanation for this discrepancy is a time-transgressive lag in isostatic rebound mediated by the position of eastward-retreating ice sheets. Rifle Range Lake is 12 km east-southeast of Optimism Bay, in a glacially carved channel on the fringe of the transition from the Hecate Lowlands to the Coast Mountain Range (Figure 2.2). Ice sheet cover may have been thicker at Rifle Range Lake 1, and may have melted slightly later than the western edge of the study area, causing a lag of several hundred years before this area experienced full isostatic uplift. This implies at least 3.1 m/km of crustal tilt at this ice margin, a high value that suggests a thin lithosphere in this area (see James et al. 2000 for a discussion of the relationship between crustal tilt and lithosphere thickness at the northern Cascadia subduction zone).  111  The pattern holds for radiocarbon dated barnacle shells found in growth position 30 m asl near the mouth of Khyex River entering the Skeena River, a further 30 km east of Rifle Range Lake 1. Two samples from this location both date about 12,700–12,200 cal. BP (Blackwell et al. 2010), indicating that RSL was still well above its current position here during its lowstand around Prince Rupert. Finally, another 80 km east of Khyex River, the Kitsumkalum-Kitimat Trough south of Terrace was not deglaciated until at least 11,500 BP (Clague, 1984, 1985), and RSL dropped rapidly because of isostatic uplift there at the same time as the RSL transgression was taking place at Prince Rupert.  Clearly, RSL position at single points in time can vary greatly with short distances depending on glacial loading, particularly on axes perpendicular to continental margins. As a result, RSL data may need to be gathered and compiled from relatively spatially limited areas, particularly if it is being used for guiding archaeological surveys for terminal Pleistocene material. Furthermore, compiling multiple RSL histories for more discrete spatial units has the potential to contribute to more robust glacial-isostatic modelling of coastal British Columbia (Hetherington et al. 2003, 2004; Hetherington and Barrie 2004), such as that conducted by James et al. (2009b) for the northern Cascadia subduction zone.  2.4.2.2 Implications for Early Human Occupation and Archaeological Survey Understanding the history of RSL change in the Prince Rupert area is critical for developing surveys for terminal Pleistocene and early Holocene archaeological sites in the area, as well as for understanding the impact of RSL change on the archaeological record. Furthermore, the archaeological potential of paleoshorelines away from the current shoreline has important 112  implications for heritage conservation in and around Prince Rupert Harbour, a major port and hub of industrial development. Detailed archaeological impact assessments that include potential paleoshoreline locations above and below current sea level that may be impacted by future development will help to mitigate the potential destruction of early archaeological sites.  Support for a coastal migration route for the first peopling of the Americas is gaining traction (Dixon 2013; Dixon and Monteleone 2014; Fedje and Mathewes 2005; Fedje et al. 2011; Mackie et al. 2011; Mandryk et al. 2001), and there is now evidence for people having lived on the BC coast as early as 13,500 cal. BP near Calvert Island, 350 km south of Prince Rupert (McLaren et al. 2015). Elsewhere on the Northwest Coast, Late Pleistocene and early Holocene sites are being identified with increasing frequency on paleoshorelines, though very few early sites are recorded on or near the mainland, especially on the northern Northwest Coast (Mackie et al. 2011). The paucity of very early sites on the inner coast may be related to a lag in deglaciation time as well as more extreme isostatic adjustments, but our data indicates that the Prince Rupert area was deglaciated and supporting edible marine molluscs by at least 15,090–14,365 cal. BP (D-AMS 009956), was vegetated shortly after, and had completed its most dramatic period of shoreline change by 14,000–13,500 years ago. We suggest that the study area was amenable to human occupation by at least this time; the presence of humans 350 km south on Calvert Island by 13,500 BP means that it is reasonable to hypothesize contemporaneous human occupation of the Prince Rupert area.  Our data suggest that archaeological evidence of habitation around Prince Rupert immediately after deglaciation is likely to be thinly scattered between at least 50 m asl and current sea level, 113  and from between 13,500 and 11,000 years ago is likely to be below current sea level, potentially buried beneath several meters of intertidal sediment. Furthermore, preservation in well-sheltered areas like Optimism Bay is likely to be excellent, whereas other archaeological material may have eroded away during the marine transgression after the Younger Dryas.  Early Holocene archaeological sites will be stranded on raised terraces above a high tide line that was minimally 8 m asl. GbTo-185 is the earliest currently recorded radiocarbon dated archaeological site on the inner northern coast of British Columbia, though an abundance of terraces associated with the 7–10 m paleoshoreline ridges visible in LiDAR DTMs of the study area suggests a high potential for more early Holocene sites. The refined RSL curve provides an important tool for archaeologists working in the region, and will be necessary for exploring the possibilities for early human dispersals through northern British Columbia, as well as developing an understanding of early- and mid-Holocene occupation, which was until now unknown for the Prince Rupert area.  2.5 Conclusion This paper describes RSL history around Prince Rupert since deglaciation, constrained by 123 RSL index and limiting points gathered from Livingstone sediment cores, geological surveys, and archaeological investigations. The area was deglaciated sometime before 15,090–14,365 cal. BP (D-AMS 009956), after which there was a rapid RSL drop from at least 50 m asl to at least −6.3 m asl between 14,500 BP and 13,500 BP in as little as a few centuries. After a lowstand below current sea level for about 2000 years during the terminal Pleistocene, RSL rose again to at least 6 m asl – and as high as 8 m asl – after the Younger Dryas. RSL slowly dropped towards 114  its current position through the Holocene, though it appears to have remained 1–3 m higher until between 2000 and 1500 years ago. There is equivocal evidence for slight fluctuations on the order of several meters between 3200 and 1500 BP. By collecting a large dataset over a relatively small geographical area we are able to distinguish variable RSL histories across relatively short distances. This detailed dataset contributes to a refined understanding of glacio-isostatic dynamics in the region. We identify what is currently the earliest dated archaeological site on the inner northern BC coast, a small 8000–9000 year old campsite on a 10–12 m asl terrace, though we suggest that the study area could have been inhabited by humans by at least 14,500–13,500 years ago, when we have the first dated evidence for vegetation of the landscape. The new inferred RSL curve for Prince Rupert indicates the probable elevations of early human settlement in the region at different times and gives potential targets for future research.115  Chapter 3: Archaeological Survey of Dynamic Coastal Landscapes and Paleoshorelines: Locating Early Holocene Sites in the Prince Rupert Harbour Area, British Columbia, Canada  3.1 Introduction Coasts are among the most dynamic landscapes on the planet, posing unique challenges to archaeological survey. Coastlines are continually reshaped by daily tidal fluctuations, waves, and currents, which variously cause sedimentation, erosion, and redeposition. In addition to these, longer-term trajectories of relative sea level (RSL) change shift the elevation of shorelines, and therefore the parts of landforms on which the daily processes impact. These changes can produce shifts in human settlement, and coast-adjacent sites can be drowned by rising sea levels or stranded inland by falling ones. Archaeologists studying coastally-focused cultures are faced with the challenge of reconstructing coastal landforms for different times and identifying how these were amenable to human occupation. Consequent to deglaciation, coastal sites from the terminal Pleistocene and early Holocene are often located away from the current shoreline, above paleoshorelines that are now either submerged or stranded inland (Bailey and Flemming 2008; Dixon and Monteleone 2014; Fedje and Christensen 1999; Mackie et al. 2011, 2014; Rick et al. 2013). Paleogeographic reconstructions of coastal landscapes can be used to target sites that may otherwise be missed if only the current shoreline configuration is taken into account, even in areas that have been intensively surveyed.   116  We present a model using Light Detection and Ranging (LiDAR) remote sensing and paleoshoreline reconstruction to locate previously unrecognized early Holocene occupation sites in one of the most surveyed areas of the Northwest Coast, the Prince Rupert Harbour, British Columbia, Canada. Building on a century-long legacy of archaeological survey in the area, new methods and research questions reveal different aspects of the archaeological record and demonstrate how attentiveness to the dynamism of shorelines allows us to locate sites that were not identified in previous surveys. We utilize a refined RSL curve for the area in conjunction with high resolution LiDAR digital terrain models to create a predictive model for identifying early Holocene paleoshorelines. Through a field survey of modelled ideal landforms, we identify previously unrecorded early Holocene archaeological sites associated with raised 7-10 m asl paleoshoreline berms, and document a pattern of changing but persistent use of these locations through the Holocene, even as the shape and position of the shoreline changed. We situate these new data with what we know of the early Holocene on the broader northern Northwest Coast and argue that the mounting evidence demonstrates spatially ubiquitous coastal occupation by this time and that lack of recorded early Holocene sites in some areas is likely a result of survey and preservation bias, rather than historical reality. The Prince Rupert Harbour paleoshoreline survey design can be implemented in other coastal regions, and we offer considerations for successfully locating other such sites in similar environments.  3.2 Addressing Challenges to Surveying Geomorphically Dynamic Coasts Archaeological surveys of geomorphically dynamic landscapes benefit from predictive designs that recognize that the modern landscape often does not reflect that of the past (Banning 2002; Stafford 1995). In areas such as the Northwest Coast, locating sites from times when RSL 117  position differed from that of today requires survey design that deals with both the history of sea level change and the obstruction of visibility produced by centuries or millennia of either rainforest growth, decay, and regrowth, or, conversely, flooding and redeposition of sediment by RSL rise (Mackie et al. 2011). If the position of past shorelines is known and detailed landform topography is available, then the task of predicting where paleoshoreline sites are located becomes much less logistically formidable. Across the Northwest Coast, new evidence for early human occupation is being identified in association with paleoshorelines, and their identification has followed in the wake of surveys designed with local RSL position explicitly in mind (Carlson and Baichtal 2015; Fedje and Christensen 1999; Fedje et al. 2005, 2011; Mackie et al. 2011; McLaren 2008; McLaren et al. 2011).  RSL has fluctuated through time and space since the last glacial maximum as a result of the complex interplay of isostatic adjustments, eustatic sea level change, tectonism, sedimentation and erosion, and changes in water temperature (Pirazzoli 1996; Shennan et al. 2015). Local variability in RSL trajectories is becoming increasingly recognized, as local RSL curves are researched and published for more locations (e.g. Engelhart and Horton 2012; Engelhart et al. 2015; Khan et al. 2015; Shugar et al. 2014). Similarly, generating topographic maps with the detail required to identify paleoshoreline features at specific elevations requires special methods, such as LiDAR for terrestrial zones, and seismic transects, multibeam and sidescan sonar for underwater areas. For terrestrial survey, the most commonly available topographic maps often do not have a resolution high enough to identify paleoshoreline landforms, while forest cover prevents such assessments from air photos. LiDAR provides georectified digital terrain models 118  (DTMs) of ground surface topography with precision within several tens of centimeters (Opitz 2013).  The use of LiDAR to identify paleoshorelines and design paleolandscape surveys on the Northwest Coast is so far limited to preliminary work on Northern Haida Gwaii (Sanders 2009), Calvert Island and Hakai Pass (Duncan McLaren, personal communication 2014), and Quadra Island (Fedje et al. 2016). In this study we detail a survey of paleoshorelines around Prince Rupert Harbour that employs LiDAR-derived landscape modelling and knowledge of the RSL history to successfully identify sites from an early Holocene RSL highstand.  3.3 Study Area Prince Rupert Harbour is located on the northern mainland coast of British Columbia, Canada, between the mouths of the Skeena and Nass Rivers (Figure 3.1). The area this paper focuses on includes Digby Island, Kaien Island and the Tsimpsean Peninsula (Figure 3.2). The region is characterized by coastal cedar and hemlock rainforests abutting crenulated shorelines, and inland bogs and bog forests. The shorelines have abundant wide sandy and muddy tidal flats that fill bays and estuaries between schist bedrock outcrops. The tidal range is 7.4 m, one of the largest for the British Columbia Coast (Canadian Hydrographic Survey, personal communication, 2015), and a major mechanical force acting on the area’s shorelines. The main sediment sources that are being reworked and redeposited by coastal processes include glacial till deposits stranded after deglaciation at least 15,000 years ago (Clague 1984; Letham et al. 2016), and sand, silt, and clay transported from the Skeena River outflow by predominantly north-running currents and winds (Clague and Bornhold 1980:342-344). 119   Figure 3.1 Northern coastal portion of Tsimshian territory in British Columbia. Archaeological sites with radiocarbon dates older than 6000 cal. BP are indicated. 120   Figure 3.2 The Prince Rupert Harbour study area with archaeological sites sorted by principal site type identified through time (n=476). Total coastline length in the pictured area is 375 km.  The area is the territory of the Tsimshian people who currently primarily reside in the city of Prince Rupert and nearby communities, although they once occupied dozens of large plank house villages along the coasts, especially around the two main waterways that lead into Prince Rupert Harbour proper: Venn Passage and Fairview Channel (Ames 2005; MacDonald and Inglis 1981). The area has long been recognized for its remarkable density of archaeological sites; it has been 121  subject to over a century of intense archaeological research that has included both regional surveys and large-scale excavations (Ames 2005; Ames and Martindale 2014; Archer 1989:37-51). Examining the history of the various survey projects over this time demonstrates how changes in knowledge, methods, and technology have impacted the results (Table 3.1).   Table 3.1 Archaeological surveys in the Prince Rupert Harbour through time. Survey Survey Methods Area Surveyed Site Types Identified Limitations/Biases Sources Harlan Smith, 1907-1929 ‘hasty archaeological reconnaissance’ (Smith 1909:595), by boat identifying village sites visible beneath areas gardened during the late 1800s or with shell midden exposures eroding from the front wave-cut banks.  Venn Passage, between Prince Rupert and Metlakatla Large villages, petroglyphs associated with large villages Unsystematic, only identified highly visible very large village sites associated with the modern shoreline. No detailed site records made. Smith 1909, 1927 Philip Drucker, 1938 By boat identifying village sites visible beneath areas gardened during the late 1800s or with shell midden exposures eroding from the front wave-cut banks.  Venn Passage, east Digby Island, Kaien Island north and south of Prince Rupert Villages, petroglyphs associated with villages Only identified highly visible large shell middens (mostly villages) associated with the modern shoreline. Drucker 1943 James Baldwin and Charles Borden, early 1950s No formal survey of any sort, though several unrecorded sites were identified and recorded using the same methods as Drucker. Several new sites were recorded on the north side of Prince Rupert Harbour, and south of GbTo-10 on the west side of Kaien Island, where Baldwin and Borden conducted salvage excavations.    Large villages (north Prince Rupert Harbour) and smaller middens near to GbTo-10 salvage excavations. Not a formal survey. Sites were highly visible and associated with the modern shoreline. Calvert 1968 122  Survey Survey Methods Area Surveyed Site Types Identified Limitations/Biases Sources Richard Inglis for the NCPP, 1974 Aerial photograph and topographic map preliminary consultation. Consideration of previously recorded site location types and analogies with ethnographically recorded settlement pattern and results of surveys from other areas. Field survey conducted rapidly by boat, stopping only at ‘ideal’ locations.  Majority of coastline visible in Figure 3.2. Large and small villages, smaller campsites, petroglyphs, intertidal rock features, burial caves, historical period structures. By only visiting ‘ideal’ locations based on expectations drawn primarily from where previous researchers identified large village sites there is a teleological bias in their approach. Only located sites associated with the modern shoreline.  Inglis 1974 David Archer for NCPP and NCHIP, 1982-1990 Aerial photograph and topographic map preliminary consultation. Classification of coastlines into three-part probability scheme based on slope, drainage, and access. Pedestrian survey in transects along all shoreline types (spot checks at low probability locations). Subsurface testing with small-diameter push probes at intervals along transects. Majority of coastline visible in Figure 3.2. Large and small villages, smaller campsites, culturally modified trees (CMTs), petroglyphs, intertidal rock features, historical period structures. Small diameter of push probes is good for identifying subsurface shell deposits, but not artifacts. Non-shell-bearing subsurface components likely to have been overlooked. Survey transects sometimes conducted as far as 100 m inland, meaning that there was a chance that inland paleoshoreline sites could be identified, though the focus was heavily on the modern shorelines and past sea levels were not explicitly considered.  Archer 1984, 1989, 1990, 1991 Various CRM Projects, 1990s to Present Computer GIS-aided predictive models and high resolution orthophoto analyses. Pedestrian surface survey and (often limited) subsurface testing.    Specific limited areas related to various contract projects (i.e. cut blocks, proposed development footprints) CMTs, small camps, intertidal lithic scatters, intertidal rock features. Varies by project, survey methods are often limited by the requirements of a proponent.  e.g. Eldridge and Parker 2007; Morrissey et al. 2015 123  Survey Survey Methods Area Surveyed Site Types Identified Limitations/Biases Sources Current Project High resolution orthophoto analysis, LiDAR DTM analysis, paleoshoreline modelling using RSL reconstruction and LiDAR DTM using selection criteria similar to those of Archer (slope, drainage, and accessibility). Percussion coring, augering, and shovel testing. See Figure 3.5. Large, medium, and small shell middens. Subsurface lithic and charcoal scatters. Specifically targeted subsurface deposits associated with shorelines between 10,000 and 6000 cal. BP. Limited time and resources and the labour-intensive nature of field methods only allowed for a survey of several small targets.  This paper, Letham et al. 2016  In the first half of the 20th Century Harlan Smith (1909), Philip Drucker (1943), and James Baldwin and Charles Borden (Calvert 1968) each opportunistically identified large shell-bearing sites along the shores of the passages leading into Prince Rupert Harbour. The sites identified by this work were the most archaeologically visible: large occupation sites with dramatic surface topography formed by mounding and terracing of shell and rows of depressions left by the construction of ancient plank houses that often had semi-excavated interiors and/or midden material built up around their exteriors. In 1974 Richard Inglis (1974) directed a formal survey of the area as part of the larger North Coast Prehistory Project (NCPP), a two-decade project conducted by the National Museum of Canada (now the Canadian Museum of History) that also included large-scale excavations at 11 sites in the Prince Rupert Harbour area (Ames 2005; MacDonald 1969; MacDonald and Inglis 1981). The most thorough and systematic survey of the area was directed by David Archer between 1982 and 1990 (Archer 1984, 1989, 1990, 1991). Archer’s attentiveness to more than just the ‘most ideal’ habitation locations and the most archaeologically visible sites allowed him to record dozens of sites that had previously been 124  overlooked, including culturally modified trees (CMTs) and modifications to the intertidal zone. His work yielded a larger and more representative sample of the range of archaeological material in the area and pointed to the ubiquity of human use of the Prince Rupert Harbour coastlines. In the years since these formal research projects, surveys have been cultural resource management (CRM) projects with more limited areal scopes, related to forestry and industrial development projects.   These surveys demonstrate that Prince Rupert Harbour is a densely packed archaeological landscape. Prior to our survey there were 476 archaeological sites recorded in the British Columbia Archaeology Branch’s archaeological site database (BCMFLNRO 2016) within the area selected for this review (Figure 3.2). Of these, 150 have shell-bearing components and likely represent different types of habitation sites. There are 375 km of coastline within the study area (Appendix B.1); therefore, the density of shell-bearing sites is 0.40 sites per km of coastline. 448 radiocarbon dates have been run on samples from 49 sites from this area (Ames 2005; Archer 1992, 2001; Coupland et al. 2003, 2010; Cybulski 2014; Edinborough et al. 2016; Eldridge et al. 2007, 2014; Letham et al. 2016; MacDonald and Inglis 1981; Southon and Fedje 2003). However, of this very large sample of dates, all of the ages accepted by previous researchers fall within the last 6000 years, with the majority clustering between 2700 and 1300 years ago (Figure 3.3A).   125   Figure 3.3 A: Summed probability distribution of all radiocarbon dates on marine and terrestrial material (n=392) for the Prince Rupert Harbour area. Dates on human remains are not included, though all fall within the same age range. B: Prince Rupert area relative sea level curve with early Holocene RSL highstand indicated (modified from Letham et al. 2016: figure 13).  The early Holocene is conspicuously absent from the Prince Rupert Harbour archaeological record, but none of the prior surveys explicitly targeted paleoshorelines. With the exception of several recent CRM surveys that were mostly directed towards locating CMTs, previous surveys have only explored the modern shoreline, which only approximates the shoreline position of the last 1500 years, although it has been relatively close to its current elevation for about 5000 years (see below). Elsewhere on the northern Northwest Coast, archaeological sites from earlier time periods are being discovered with regularity in association with paleoshorelines on the Alexander Archipelago in Southeast Alaska (Carlson and Baichtal 2015), on Haida Gwaii (Fedje and Christensen 1999; Fedje et al. 2005; Josenhans et al. 1997), and on the Dundas Islands (McLaren 126  2008; McLaren et al. 2011). Key to these surveys’ success was the reconstruction of postglacial RSL histories, which allowed the researchers to design surveys that specifically targeted paleoshorelines from the time periods they sought. We hypothesized that the lack of early Holocene sites in the Prince Rupert Harbour was due to a lack of archaeological attention to landscape transformations as a result of RSL change as opposed to a lack of human occupation.   3.4 Relative Sea Level History In order to design a survey of paleoshorelines in the Prince Rupert Harbour area, we reconstructed the history of RSL change since the last glacial maximum (Figure 3.3B; Chapter 2; Letham et al. 2016). Like other heavily glaciated areas of the mainland coast, isostatic depression of the earth’s surface during the Fraser Glaciation caused RSL in the Prince Rupert area to be much higher (50 m asl; datum is geodetic mean sea level measured by the Canadian Geodetic Vertical Datum of 1928 [CGVD28] benchmark at Prince Rupert) immediately after the glaciers melted, at least 15,000 years ago. Subsequent isostatic rebound caused a rapid RSL drop as the once-depressed land uplifted, resulting in a lowstand RSL position of at least -6.3 m asl beginning between 14,500-13,500 cal. BP. Isostatic rebound slowed during this terminal Pleistocene lowstand, and there was a marine transgression after the Younger Dryas period caused by eustatic sea level rise becoming the largest contributor to RSL change (Lambeck et al. 2014). This transgression caused RSL to pass again above its current position around 11,000 cal. BP, and RSL rise continued to an elevation of at least 6 m asl (and as high was 8 m asl) by 10,000 years ago. The Holocene is characterized by a slow RSL regression, with some variations in rates. Contributing forces to Holocene RSL trends in the Prince Rupert Harbour area are poorly understood, but are likely related to an interplay of slower isostatic responses, continued 127  eustatic sea level change, and subtle tectonic changes (Letham et al. 2016:187). From its position of 6-8 m asl at 10,000 cal. BP, RSL dropped slowly to about 5 m asl by 6000 cal. BP. There may have been a slight acceleration in RSL regression between 6000 and 5000 cal. BP, but from then RSL seems to have remained 1-3 m asl until reaching its current position between 2000 and 1500 years ago. There is evidence for a minor fluctuation of several meters between 3200 and 1500 BP (Letham et al. 2016:187).  This RSL curve suggests that archaeological evidence of occupation associated with shorelines from the terminal Pleistocene (ca. 14,000-11,000 BP) is likely to be in the current intertidal or high subtidal zone, while evidence of occupation associated with early Holocene paleoshorelines will be associated with RSL at least 5-6 m above that of today. The highest astronomical tide (HAT) in the area is 3.7 m asl, and the average of all higher high waters (higher high water mean tide, HHWMT) is 2.32 m asl (Letham et al. 2016: table 2). Assuming that people would have lived at least above the most frequent high tide positions, we expect that early Holocene coastal occupation sites at elevations between at least 8-10 m asl (Letham et al. 2016). LiDAR DTMs of the study area show a consistent pattern of linear berms running parallel to the modern shoreline between 6 and 10 m asl (Figure 3.4). We interpret these features as relict wave-cut banks from the interface between high tides and the land’s edge. Their ubiquity and clarity as landform features suggest that they were formed during the relatively stable early Holocene RSL highstand. We designed a survey that targeted these raised shorelines, as opposed to surveying the intertidal area for terminal Pleistocene material, which presents an array of different logistical challenges.  128   Figure 3.4 Examples of LiDAR DTM with paleoshoreline and archaeological features. Left: hillshade visualization of the DTM with modelled theoretically habitable areas associated with an 8.5 m high tide line (i.e. high tide during the early Holocene RSL highstand) indicated. Right: slope-classed model derived from the LiDAR DTM highlighting different types of landforms visible, including paleoshorelines. Note that the large village site GbTo-70 is visible as a topographic anomaly in both maps.  3.5 Methods: Prince Rupert Harbour Area Paleoshoreline Survey Design This section outlines the design of a predictive model for locating paleoshorelines and paleoshoreline archaeological sites in the Prince Rupert Harbour area, as well as the field methods employed to search for sites at predicted survey targets. 129  3.5.1 GIS and LiDAR Modelling LiDAR is an invaluable remote sensing tool for identifying raised paleoshorelines in densely forested landscapes. LiDAR is aerial scanning of the earth’s surface with rapidly emitted laser pulses, usually by airplane; reflections of the pulses are measured with high spatial precision by a receiver connected to an inertial navigation system and GPS (Opitz 2013). Because some beams or portions of beams make it through the forest canopy to the ground surface, LiDAR data can be filtered and interpolated to produce ‘bare earth’ DTMs: detailed models of surface topography stripped of trees. We used a bare earth LiDAR DTM of the entire study area flown and processed by Airborne Imaging (2013) and provided to us by Nexen Ltd. to design landscape models to help us predict areas with a high potential for archaeological material associated with early-mid Holocene shorelines.  The LiDAR data used in this project was collected with two different LiDAR systems – a Leica ALS70 system 179 and a Leica ALS70 system 207 – and several different sets of flight parameters based on terrain over several flight missions between April 21 and May 23, 2013 (Airborne Imaging 2013). All flights were designed to collect 8 measurement points per m2 with overlap; measurements were calibrated against each other and against known benchmarks. The final point cloud was classified and filtered for ground points using Terrascan software, and a bare earth DTM with 1 m grid resolution was interpolated from these points. The 95% vertical accuracy of this dataset is 9 cm on flat hard surfaces (Airborne Imaging 2013:7). We ground-truthed the quality of the DTM by locating relatively small features, such as creeks with widths 1-2 m, and found that the LiDAR models successfully captured these subtle topographic nuances, making it an ideal dataset for locating raised paleoshoreline landforms. We also found 130  that the large village sites and their house depressions stand out very clearly in the DTMs (Figure 3.4).  We designed a survey model targeting archaeological sites directly associated with shorelines from the 10,000-6000 cal. BP 5-6 m asl RSL highstand using QGIS, a free and open-source Geographic Information System software package. Under the basic assumption that people would most likely live on flat landforms close to the shore – an assumption backed by both common sense and empirical observation of more recent sites – the two primary predictive criteria selected for the model were slope and elevation. We first assessed the length of paleoshorelines in the area and modelled how much of this shoreline fronted theoretically ‘habitable’ landforms. We selected flat (slope of 0-7°) areas greater than 50 m2 for which the majority of the landform is within 10 vertical meters of the high tide for a given RSL position, and either the majority of that area is within 30 horizontal meters of the selected shoreline position or some portion of that area is within 10 horizontal meters of the selected shoreline position (Figure 3.4; see Appendix B.2 for a step-by-step QGIS workflow). We calculated the length of all shorelines within 15 meters of these theoretically habitable landforms in order to assess the proportion of various shoreline positions that were theoretically habitable (Appendix B.1). For the modern shoreline, 57% of the total length is theoretically habitable, and 93% of all previously recorded sites with shell-bearing components are associated with these shorelines, indicating that the selection criteria are appropriate. For a paleoshoreline contour of 8.5 m asl, which would estimate highest high tide positions for the early Holocene 5-6 m asl RSL highstand, we found that 58% of that shoreline is adjacent to modelled theoretically habitable areas, essentially the same proportion as that of the modern shoreline. These estimates do not 131  take into account uninhabitable flat areas such as bogs, so they are conservative upper estimates. Furthermore, they do not take into account the quality of the shoreline itself, which is often rocky and steep, and therefore unbeachable by watercraft, an assumed key criterion for human shoreline occupation.  In order to select the most ideal of these modelled theoretically habitable landforms we looked for areas that had analogues along the modern shoreline as being amenable to occupation. We classified the slope of the entire dataset into several grade ranges based on the types of landforms they might represent (Figure 3.4). Surfaces with slopes 0-7° are designated as comfortably habitable by people without significant modification, and bands of this grade below paleoshoreline high tide elevations may also represent low energy fine clast-sized beaches (Mason 1991:60). Linear bands between 7° and 24° may represent higher energy gravel beaches (Mason 1991:60) or the back berms of lower grade beaches, such as the frequent 6-10 m asl features noted above. Slopes 24-40° are steep surfaces, neither ideal for human habitation nor likely to represent beaches. Slopes greater than 40° are very steep surfaces, definitely not ideal for human habitation. The steepest surfaces may be bedrock cliffs. These areas have the potential for overhangs that could have served as rockshelters for humans.  We conducted a visual analysis of the GIS to identify areas of theoretically habitable land immediately adjacent to linear sections of medium (7.1-24°) slope between 6 and 10 m asl that run parallel to the modern shoreline or otherwise look analogous to modern day shoreline shapes and are fronted by low grade areas (Figure 3.4). We hypothesized that these areas would have been habitable landforms immediately associated with beaches that were accessible and 132  beachable for watercraft. After identifying ideal targets, we checked orthophotos of the areas to make sure they were not bogs. Sphagnum bogs are common in the raised inland portions of the study area, and though they may not necessarily have been bogs in the early Holocene – several bog studies in the Prince Rupert region have recorded increased paludification after 8500 cal. BP (Banner et al. 1983; Turunen and Turunen 2003) – they are not logistically feasible targets for traditional shovel testing. Ideal landforms are frequent on the west side of Tsimpsean Peninsula, around Digby Island, and in many of the parts of Venn Pass that would have been bays during the early Holocene RSL highstand.   3.5.2 Field Methods Over a total of four weeks between 2014 and 2015 a small crew of 2-5 individuals surveyed and tested sixteen landforms hypothesized to be associated with raised paleoshorelines on the north half of Digby Island and southern Tsimpsean Peninsula (Figure 3.5). We used a combination of shovel and auger testing with water-screening through 6.3 mm and 2 mm mesh to search for subsurface cultural material such as charcoal, fire-cracked rock, bone, shell, or artifacts. Subsurface tests were placed judgementally so as to avoid trees and logs, though in general we attempted to space them 5-10 m apart. We used an Oakfield soil probe to map the extent of subsurface shell components. 133   Figure 3.5 Paleoshoreline survey targets and results. Identifier names/numbers of survey areas correspond with those in Table 3.2. Large inland shell-bearing sites that were identified on the LiDAR DTMs are also indicated. 134  Shovel and auger testing provide only a small view into the landform being tested (0.70 m by 0.35 m for shovel tests, 0.1 m diameter holes for auger tests); it often took repeated visits to test a location before archaeological material was identified. We calculated coverage (“the total area [tested] in which we could expect to detect targets” [Banning 2002:63]) and density of effort (the total coverage of a survey area divided by its area [Banning 2002:60]) for all survey locations in order to assess the reliability of our results (Banning et al. 2016). Coverage per shovel test was therefore 0.245 m2 and coverage per auger test was 0.0079 m2.   3.6 Paleoshoreline Survey Results  Survey results and radiocarbon dates are presented in Figure 3.5, Table 3.2, and Appendix B.3. We discovered archaeological sites on five of the 16 surveyed landforms, three of which were radiocarbon dated to earlier than 6000 cal. BP. A further nine of the tested landforms yielded areas with abundant charcoal or black greasy cultural-looking strata that was deemed ambiguous evidence. Of the three newly identified pre-6000 cal. BP sites, two (GbTo-82 and GcTo-67) are small shell-bearing components that date to about 6500 cal. BP. The third, GbTo-185, is a scatter of quartzite flakes and other stone artifacts associated with several burning areas that may be hearths or campfires with two calibrated dates of about 9000 and 8000 BP, making it the oldest dated site in the study area by nearly 3000 years.  135  Table 3.2 2014 and 2015 Paleoshoreline survey tests, coverage, and results. Survey area locations are indicated in Figure 3.5. Survey Polygon Survey Polygon Area (m2) # of Shovel Tests (0.7 x 0.35 m) # of Auger Tests (0.05 m diameter) # Test Excavation Units Subsurface Coverage (Total Area Tested, m2) Density of Effort (%)* Archaeological Result P003 3988 0 4 0 0.03 0.0008% Negative P009 5975 1 18 1 (1x1 m) 1.39 0.0232% Positive† P010 393 3 1 0 0.74 0.1889% Positive P011 11,901 6 4 2 (1x0.5 m) 2.50 0.0210% Positive† P012 3582 7 2 0 1.73 0.0483% Negative P021 21,707 2 20 0 0.65 0.0030% Ambiguous P023 5102 0 7 0 0.05 0.0011% Ambiguous P024 6530 4 4 0 1.01 0.0155% Positive† P025 3182 3 3 0 0.76 0.0238% Ambiguous P026 1906 3 2 0 0.75 0.0394% Ambiguous P027 1563 4 0 0 0.98 0.0627% Ambiguous NDig1 2892 1 6 0 0.29 0.0101% Ambiguous TBay 1444 0 3 0 0.02 0.0016% Ambiguous NDig2 1353 2 0 0 0.49 0.0362% Ambiguous WPill 5975 14 1 0 3.44 0.0575% Positive EPill 5153 2 2 0 0.51 0.0098% Ambiguous * Density of Effort is the total proportion of the selected survey polygon’s area that was covered by subsurface testing. Density of Effort = subsurface coverage/survey polygon area x 100 † Indicates positive result with a component dated to earlier than 6000 cal. BP. 136  3.6.1 GbTo-82/P024, Henry Point GbTo-82 is located on a 10 m asl terrace above Henry Point, on the west side of Digby Island. During the early Holocene the landform would have been a flat promontory forming a small point with a semi-protected approach to a low-grade beach on the north and northwest and a steep bedrock shoreline to the west (Figure 3.6A-C). GbTo-82 was previously recorded by Archer and includes a single CMT and a small 7x4 m patch of subsurface shell midden, which he noted was in a strange and less-than-ideal location given its elevation and distance from the shoreline (Archer 1989:158). Furthermore, the rocky headlands at the shore make boat access difficult. While RSL change and paleoshorelines were not likely under consideration by Archer and his co-surveyors at the time, his description suggests that the shell-bearing deposit may have been associated with a higher RSL. This midden fell within a high potential target landform in our model, making it an ideal candidate for a site occupied during the early Holocene highstand. Indeed, a marine shell from the lower instance of midden at this location (80 cm below the surface, 9.4 m asl) dated 6728-6463 cal. BP (D-AMS 011956). A bulk midden sample collected with an auger from this location contained predominantly fragments of acorn barnacle and various species of clam. Two very small fragments of unidentifiable bone were recovered. 137   Figure 3.6 Orthophotos, slope-classified LiDAR DTMs, and hillshaded LiDAR DTMs of Henry Point with GbTo-82/P024 (A-C) and the north shore of Scott Inlet with GcTo-67 (D-F). Survey areas, subsurface tests, and paleoshoreline features indicated.   138  Additionally, there was a charcoal-strewn hard-packed surface with an abraded cobble in a shovel test placed 7 meters southeast of the midden area. While undated, its position on the raised landform, its depth and proximity to the GbTo-82 shell midden suggest a contemporaneous (or potentially even earlier) component. It is also possible that part of the shell-bearing component has dissolved with age as a result of edaphic processes in the highly-acidic Northwest Coast rainforest. These processes can result in a black greasy matrix where the shell once was in which lithic artifacts will still remain (cf. Stein 1992), and may therefore cause early shell-bearing sites to appear smaller than they once were.  We also found an unrecorded rockshelter with a subsurface shell midden just above the current shoreline on the south side of Henry Point that would have been underwater at 6500 cal. BP. Along with the more recent-aged CMT recorded by Archer, this indicates persistence of use of this particular location, likely as a resource gathering site or a sheltered stopover for people traversing the exposed western coast of Digby Island, even as the configuration of the shoreline has changed as a result of falling RSL.  3.6.2 GcTo-67, Scott Inlet GcTo-67 is located on an 8-10 m asl terrace on the northern side of the mouth of Scott Inlet between a steep rocky bluff and an upward slope to the northeast (Figure 3.6D-F). During the early Holocene this terrace would have been a tombolo connecting the bluff to the Tsimpsean Peninsula, and the emerging vegetated landform would have been a point with small sheltered bays on either side (Figure 3.6E). We found two separate shell-bearing components on this landform. One is a medium-sized shell accumulation that extends from the middle of the paleo-139  tombolo down towards the modern shoreline. While not extending all the way to the current wave-cut bank, the lower elevation of the majority of this deposit indicates that it post-dates the 10,000-6000 cal. BP RSL highstand, though it may date prior to 1500 cal. BP, when RSL was slightly higher than today. The second shell component is a deeply buried discrete component to the east of the first on the 8-10 m asl terrace.  In a 1x1 m test unit with a surface elevation of 9 m asl in this latter area there was about 40 cm of humus overlying a black, charcoal-rich greasy layer with vertebrate faunal remains and lithic artifacts from about 40-80 cm (Figure 3.7). There was a thin shell-bearing component that yielded a date of 6634-6445 cal. BP (D-AMS 011948) between 80 and 90 cm DBS. This deposit was composed almost entirely of clam shell, though we also recovered a few pieces of land mammal bone. The most common artifacts were small quartzite and quartz crystal flakes (n=6) and debitage/shatter (n=6), beginning around 70 cm DBS in the black layer and continuing into the shell-bearing component. We also recovered two small bird bone tubes made from a single bone that had been sawed in half. These two pieces are bevelled on the sawed ends and notched at their distal ends (Figure 3.7D). These may have been beads, such as those recorded by Ames (2005) from other late Holocene components in the Prince Rupert Harbour. The shell component sits on a dark brown organic-rich paleosol with charcoal, lithic artifacts, and schist cobbles pressed into the surface. Charcoal from the culturally-sterile middle of this paleosol dates to 7196-6966 cal. BP (D-AMS 011947), indicating that occupation of the surface of this layer began after this. Beneath the paleosol there is a transition to dark-gray water-rolled sandy gravel, within which we found a single crude flaked cobble tool made of porphyritic rhyolite with patterned micro-chipping usewear on one edge (Figure 3.7F-G). This artifact stood out as a 140  larger clast than any of the natural clasts, and the deep red raw material was unique compared to the granitic gravel matrix. Below this stratum, the gravel becomes medium brown and completely culturally-sterile. These gravel layers may have been intertidal sediment deposited when this location was a tombolo during the maximum early Holocene marine transgression around 10,000 years ago. Below this there is a sharp transition to blue-gray clay that is likely glacio-marine at 128 cm below the surface.  Figure 3.7 Stratigraphy in evaluative excavation of a 6500 cal. BP shell component at GcTo-67 (A-B) and a selection of artifacts from each layer (C: quartz and basalt flakes from above the shell layer, D: quartz flakes and bird bone tubes from the shell component, E: andesite cobble core sitting on paleosol surface below the shell component, F: porphyritic rhyolite cobble tool, G: close-up of microchipped usewear on edge).  In a second shovel test 10 m southwest from the test excavation there were abundant quartzite and quartz crystal flakes (n=4) and shatter (n=30) along with several larger flakes of basalt or andesite and mudstone, several cobbles that had been used for hammering and grinding, one basalt or andesite mortar, and at least two manu-ported ovoid cobbles. This shovel test has not 141  been dated, and did not contain a shell component, though the upper portion resembled the upper black layer in the test unit, and the lower portion contained more gravel and cobbles, eventually transitioning to a sterile gravel layer.  The radiocarbon dates indicate that a paleosol formed at 8 m asl on this landform sometime before 7000 BP, likely when RSL fell low enough for the paleo-tombolo to become vegetated. Sometime between 7000 and 6500 cal. BP humans occupied the location, evidenced by charcoal, lithic tools, and seemingly manu-ported cobbles on the surface of the paleosol at the lower interface with the shell component. Around 6500 BP people began depositing shell at the location. There are no radiocarbon dates for the black shell-free deposit above the shell component. Perhaps the most intriguing find at this site is the small flaked cobble tool from below the paleosol. It is not clear whether the artifact is in primary context or if it was transported from elsewhere, but its stratigraphic position indicates human activity on or near the landform earlier than 7000 BP. Unfortunately, we did not identify any dateable material in association with this artifact.  The early shell-bearing component at GcTo-67 is contemporaneous with that at Henry Point, and the preserved portion appears to cover about the same horizontal extent (~9x6 m). As at Henry Point, use of the location continued as the shoreline changed: the larger shell deposit that extends down the slope toward the current shoreline had to have been deposited by occupants later in the Holocene after RSL had fallen. 142  3.6.3 GbTo-185, North Digby Island GbTo-185 is located on a flat 11-12 m asl terrace fronted by a steep berm and then an 80-120 m wide stretch of gradually sloping boggy forest running to the current shoreline. The berm mimics the current shoreline shape, indicating that it is a relict back beach berm (Figure 3.8A-B). The boggy area would have been a wide low grade beach in front of this. We found a dense concentration of charcoal and black greasy sediment adjacent to a creek that bisects and downcuts through the 11-12 m asl terrace. Near the surface of this layer in one shovel test we found a small basin filled with charcoal that dated 3557-3381 cal. BP (D-AMS 011951), more recent than the early-mid Holocene RSL highstand. Beneath this layer, however, were earlier charcoal-rich sands and gravels where we found stone artifacts in four shovel tests, and two hearths or campfire features. In total we recovered 12 flakes and lots of shatter from this site; the majority of these were quartzite or quartz crystal. The largest hearth feature was a mound of charcoal and ash lenses with associated cobbles and quartzite flakes (Figure 3.9). The cobble concentration consists mostly of schist similar to the bedrock outcrops in the area, and two of them appear to have notches ground into their sides. There was also a quartzite cobble, a sandstone hammerstone, and a mudstone spokeshave. 143   Figure 3.8 Orthophoto, slope-classified LiDAR DTM, and hillshaded LiDAR DTM of GbTo-185 area. Survey area, subsurface tests, and paleoshoreline features indicated. GbTo-183, a large shell-bearing site visible as a topographic anomaly in the LiDAR DTMs just above 5 m asl is also indicated. Note the pattern of occupation following the shoreline as RSL drops through the Holocene. 144   Figure 3.9 Profile and plan view of 9000-8000 cal. BP occupation layers at GbTo-185.  Two charcoal samples from this feature yielded stratigraphically consistent dates of 8348-8185 cal. BP (D-AMS 011949, 49 cm below surface), and 9302-9028 cal. BP (D-AMS 011950, 55-60 cm below surface). These are the earliest radiocarbon dates on archaeological material recorded for the Prince Rupert Harbour area. Below this feature the charcoal and artifact density decreases, and there is a transition to sterile basal sands.  While we cannot say much about the early Holocene occupation of this site with our small excavated sample, this may have been a coastal camp taking advantage of the sheltered bay, extensive intertidal zone, and estuary that likely existed at the time. The more recent date 145  indicates later Holocene site use too, perhaps periodically as a camp for inland hunting; the low density of accumulated material does not suggest intensive use after RSL had dropped away from this landform. There are two late Holocene shell-bearing sites nearby though: GbTo-183 is stranded on a lower paleoshoreline from between 3500-1500 BP (see below) about 130 m to the north, and GbTo-65 is located on the modern shoreline 100 m west of GbTo-183 and 200 m northwest of GbTo-185 (Figure 3.8C). A basal date from GbTo-183 of 3560-3363 cal. BP (D-AMS 007906) is essentially identical to that of the small charcoal-filled basin at GbTo-185. This site complex demonstrates exceptional long-term use of a place even as shorelines are changing, and a sequence of habitation following these fluctuating shorelines.  3.6.4 Other Paleoshoreline Sites Identified: GbTo-186, GbTo-183, GbTo-184, GcTo-66 We identified another site, GbTo-186, on an 8.5-11 m asl terrace on the west side of Pillsbury Cove with a concentration of black greasy charcoal-rich sediment like that at GbTo-185, and, in one shovel test, a charcoal-filled basin that overlay an ashy deposit with cobbles and a single large argillite flake. The charcoal concentration dated 2060-1904 cal. BP (D-AMS 007891), significantly younger than anticipated. Nonetheless, GbTo-186 and the late Holocene hearth at GbTo-185 indicate more ephemeral human activity back from the shoreline. This site does date to a period during which there may have been a slight increase in the RSL that was already 1-3 m above its current position between 3200 and 1500 BP (Chapter 2; Letham et al. 2016). The shoreline may have been slightly closer to this position during this time.  An unanticipated discovery made while analyzing landforms on the LiDAR DTMs was several topographic anomalies 45-120 m back in the forest on terraces just above 5 m asl that resembled 146  recorded medium and large shell mounds on the modern shoreline (e.g. Figure 3.8B-C). We visited three of these locations in the field, and confirmed that they were indeed ‘stranded’ shell-bearing sites (GbTo-183, GbTo-184, and GcTo-66; Figure 3.5). Radiocarbon dates on core samples from these sites yielded ages between 3500 and 1500 cal. BP. These sites are associated with a different, slightly lower set of raised paleoshorelines, which appear more ephemeral than those from the early-mid Holocene RSL highstand. Assuming these occupations were located just above 5 m high tides, RSL would have been 1.5-2.5 meters above its current position during this time. Sixty percent of a 5 m asl contour is within 15 m of modelled theoretically habitable landforms associated with this RSL position (Appendix B.1); these inland shell accumulations suggest that one could design a survey for more sites from the same time but which do not have enough topographic relief to show up on LiDAR DTMs using the same methods that we use for 10,000-6000 cal. BP landforms.   3.7 Discussion The following sections critically assess the paleoshoreline survey’s sample coverage, provide recommendations for future work, and compare the new early Holocene archaeological data to the contemporary record from the broader northern Northwest Coast.  3.7.1 Survey Assessment and Recommendations for Sampling Dynamic Coastlines  This research demonstrates the value of predictive modeling and sampling that takes into account the dynamism of coastal landscapes, and provides several guiding principles for implementing similar surveys. By reconstructing the history of RSL change in the Prince Rupert Harbour area and modelling changing shoreline positions we begin to account for one of the main sources of 147  geomorphic dynamism of the coastal landscape. LiDAR has proven invaluable for its ability to ‘see through the trees’ and to model subtle landform features associated with past geological and anthropogenic processes.   On the ground, the rainforests of the Northwest Coast generate thick accumulations of acidic humic soil that can cover and decompose organic remains. In the absence of convenient erosion exposures, subsurface testing is necessary to identify early occupations. In contrast to shell-dense later sites, early sites have low obtrusiveness, and often all that is preserved are low density scatters of lithics and charcoal concentrations. Shovel testing is likely the minimum-sized test required for locating these sites, and the clumpy dark wet soil requires water screening for identification of artifacts. Stafford (1995) has demonstrated that reducing screen mesh size reduces the amount of sediment that needs to be searched to identify artifacts and that micro-fraction analysis is effective for identifying low-visibility, low-obtrusiveness archaeological sites. However, the difficulty of wet-screening Northwest Coast soils in the field requires a balance between dirt searched and size fraction analyzed. Identification of these sites can therefore be time consuming and labour intensive.   An examination of the coverage of the methods being used allows us to assess a survey’s effectiveness (Banning et al. 2016). Of the 58% (208.5 km) of the total 10,000-6000 cal. BP paleoshoreline that we modelled as being ‘theoretically habitable’, we only surveyed landforms adjacent to 1.3% (2.7 km). Furthermore, our actual coverage and density of effort is strikingly low – less than 0.1% of any of our target polygons for the latter (Table 3.2) – which suggests that even in areas where we found no archaeological material, we cannot state with much certainty 148  that there is not an archaeological site there. Early Holocene occupations are often ephemeral and affected by preservation bias, meaning that increased search effort and repeated survey of a single location is often required to locate these sites or to develop an adequate level of certainty that there is not a site (Banning et al. 2016). For example, we did not confirm GbTo-185 as an archaeological site until our 12th subsurface test, on our fifth visit to the location. Additional coverage at survey areas that yielded ambiguous results might increase recovery rates.   Overall, our extremely low survey coverage suggests that our sample is not yet representative of the early Holocene archaeological record in the Prince Rupert Harbour area. However, our success rate in the small proportion of area surveyed suggests that there are likely many more early Holocene sites and that the northern Northwest Coast was more ubiquitously populated during this time than previously assumed.   3.7.2 Northern Coast Early-Mid Holocene Record In addition to the three pre-6000 cal. BP sites now known in the Prince Rupert Harbour area, there are ten sites or site components with early dates recorded elsewhere in coastal Tsimshian territory, all on offshore islands in the Hecate Strait (Table 3.3, Figure 3.1 and Figure 3.10). Although the sample size is small, there are some potential patterns. All these sites are associated with raised paleoshorelines; most of the components are currently located around or just above 10 m asl. The earliest radiocarbon date is 11,204-10,885 cal. BP (UCIAMS 28008), from a shell-free component below a shell-bearing component at GcTr-6, Far West Point, on the Dundas Islands (McLaren 2008; McLaren et al. 2011). However, attempts to reproduce this date have failed (McLaren 2008:243-244; Appendix B.3). The next earliest dates, and the oldest dates on 149  shell-bearing components in the sample are 9695-9335 cal. BP (D-AMS 007883) at GaTp-10 on Stephens Island (Letham et al. 2015: table 2) and 9231-8715 cal. BP (Beta 345573) at GbTp-1 on Lucy Island (Archer and Mueller 2013). GbTo-185 is occupied by essentially the same time in the Prince Rupert Harbour area.   The majority of the sites’ earliest dates are between 8000 and 6500 BP, however. The other two sites in the Prince Rupert Harbour area have dates towards the end of this range. Notably visible in Figure 3.10 is that there are also late Holocene dates at nearly all of these sites, indicating that the pattern of repeated or continued use of places even after major shoreline changes extends beyond the Prince Rupert Harbour area sites. In five cases, large late Holocene village occupations extended overtop of sites on early-mid Holocene paleoshorelines (e.g. Martindale et al. 2009).  150   Figure 3.10 Radiocarbon dates for archaeological sites with radiocarbon dates older than 6000 cal. BP in Tsimshian territory. Locations of each site are on Figure 3.1 and details of each site in Table 3.3. 151  Table 3.3 Archaeological sites with radiocarbon dates older than 6000 cal. BP in Tsimshian territory. See Figure 3.1 for locations. Site Location Elevation Landform/shore shape at time of occupation Exposure Exposed Aspect Area of pre-6000 BP archaeological component Sources GbTo-185 Prince Rupert Harbour Area (north Digby Island)  10-12 m asl terrace along coastal bight, possible with a small estuary semi-exposed west-facing ~200 m2 This study. GcTo-67 Prince Rupert Harbour Area (south Tsimpsean Peninsula)  8-10 m asl mainland end of tombolo/isthmus protected probably south-facing; though the tombolo/isthmus also had a bay to the north ~50 m2 (shell midden extent) This study. GbTo-82 Prince Rupert Harbour Area (west Digby Island) 10-11 m asl terrace on a small point at the south end of a wide embayment exposed northwest-facing ~30 m2 (shell midden extent) This study. GbTp-1 Lucy Island 10.5-16.5 point of small island with tombolo exposed northwest-, west-, southwest- and south-facing ~2500 m2 Archer and Mueller 2013 GaTp-10 upper side terrace Stephen's Passage (Stephens Island) ~10 m asl * terrace in a small inlet or estuary protected south-facing ~200 m2 Letham et al. 2015 GaTp-7a lower component Stephen's Passage (south side of Dolly Island) ~10 m asl * south edge of small bay on sheltered side of small island semi-protected west-facing n/a, beneath a large village with late Holocene dates This study. GaTp-14† Stephen's Passage (north side of Dolly Island)   ~10 m asl * sheltered north side of small island protected north- and northwest-facing ~2000 m2 Letham et al. 2015 152  Site Location Elevation Landform/shore shape at time of occupation Exposure Exposed Aspect Area of pre-6000 BP archaeological component Sources GcTr-6 Dundas Islands (west point of peninsula on Dunira Island) 10-12.5 m asl terrace in a small inlet or estuary semi-protected west-facing ~300 m2 McLaren 2008; McLaren et al. 2011 GcTq-2 Dundas Islands (northern Melville Island)  13.5-17.5 m asl mainland end of tombolo/isthmus protected north-facing and southwest-facing ~300-600 m2 McLaren 2008; McLaren et al. 2011 GcTq-4 upper terrace Dundas Islands (small islet between Baron and Dunira Islands)  12.5-14.5 m asl small islet, surrounded by rocky reefs semi-protected west-, south-, and east-facing (~270° exposure); potentially exposed to the north as well. ~600-700 m2 Martindale et al. 2010; McLaren 2008; McLaren et al. 2011 GdTq-3 upper terrace Dundas Islands (small islet between Dundas and Baron Islands) 9.75-13 m asl point of small island semi-exposed northwest- and west-facing ~700-800 m2 Martindale et al. 2010; McLaren 2008; McLaren et al. 2011 GdTq-1 lower component Dundas Islands (small islet between Dundas and Baron Islands) 10-13 m asl protected bay with offshore rock/islet protected west- and southwest-facing n/a, beneath a large village with late Holocene dates McLaren 2008; McLaren et al. 2011 GcTr-8 lower component Dundas Islands (small islet between Dundas and Baron Islands) 8-13 m asl raised area on edge of small semi-exposed bay semi-exposed northwest-facing n/a, beneath a large village with late Holocene dates Martindale et al. 2009; McLaren et al. 2011 * elevation estimates for Stephens Islands sites are rough but reasonable † GaTp-14 is reported with the temporary site number T416-1 in Letham et al. 2015.153  The majority of pre-6000 cal. BP components are small, usually between 200 m2 and 700 m2, though these are likely rough minimum estimates given that the deposits are usually deeply buried and may be affected by the dissolution of shell or other preservation biases. With the exception of the basal component at GcTr-6, GbTo-185, and GcTq-2 (a lithic and charcoal scatter on the Dundas Islands), these are all shell-bearing components. This is likely related to the higher obtrusiveness of shell in subsurface tests relative to that of lithic artifacts; non-shell-bearing sites are inherently more difficult to identify.   If we reconstruct the likely shape of the paleoshorelines associated with each site, many are located on small islets, often in or on the ends of protected or semi-protected bays (Table 3.3). Several are located on the inside of what were likely once estuaries or the inner shores of long narrow bays, and several are associated with the landward sides of what were once tombolos that would likely have provided shelter and easy boat access from multiple directions. The majority (11 out of 13) had beach aspects facing at least one of the directions northwest, west, or southwest.   Apart from our test excavations, small-scale excavations have only been conducted at GcTr-6, GdTq-3, and GcTq-4, all on the Dundas Islands (Martindale et al. 2010; McLaren 2008; McLaren et al. 2011). These are all shell-bearing sites with components dominated by barnacle and clam. GdTq-3 and GcTq-4 have structural depressions in their surface topography, and excavations revealed features such as roasting hearths, heating hearths, and post holes that suggest relatively significant investment in place and intensive occupation (Martindale et al. 2010). The small excavation at GcTr-6 lacked clear evidence of such features, though contained 154  over 2 m of dense shell that accumulated between 8000 and 7000 cal. BP (McLaren 2008; McLaren et al. 2011). GdTq-3 and GcTq-4 lacked a large sample of vertebrate faunal remains, but GcTr-6 yielded over 4000 bones, vastly dominated (93%) by fish remains (McLaren 2008:249-252). All vertebrate fauna recovered from these offshore island sites would have been available locally, and point to an almost exclusively marine resource base. Very few artifacts were recovered from any of these excavations, though notably bird bone tubes like those recovered from GcTo-67 were found at GcTr-6; these are common at late Holocene components excavated by the NCPP (Ames 2005). No microblades were recovered, even though these artifacts are commonly diagnostic of contemporaneous sites elsewhere on the northern coast (Carlson and Baichtal 2015; Fedje and Mackie 2005).   Quartz crystal and quartzite were the most frequent lithic raw materials in the small assemblages from GcTo-67 and GbTo-185. Bedrock in the Prince Rupert Harbour is primarily amphibolite grade schist, which is weak and flakes poorly, but veins of quartzite with occasional quartz crystals are scattered throughout the area. This is the best immediately available raw material, except for cobbles of higher quality material sourced from a few local till deposits. Quartz is a common raw material at some later sites in the area. It makes up 20% of the total assemblage of chipped stone items that the NCPP excavated from Boardwalk (GbTo-31), on eastern Digby Island (Ames 2005:356), and was the most common lithic raw material recovered in large-scale excavations at GbTo-54 (39% quartzite and 13% quartz crystal), at Casey Point on Kaien Island (Eldridge et al. 2014:129-131). In Tsimshian cosmology quartz crystals are viewed as powerful items and are associated with chiefly powers (Miller 1997:39-41). The use of quartz and quartz crystal over 9000 years is evidence of toolkit continuity, which Ames (2005) emphasized for the 155  last 5500 years based on his analysis of artifacts from the NCPP excavations, and McLaren (2008:252-253) concludes based on the small artifact sample from early Holocene sites on Dundas Islands. Use of quartz crystal could also tentatively be evidence for long-term entrenchment of Tsimshian cosmological principles.   Elsewhere on the northern Northwest Coast, paleoshoreline surveys have identified nearly 30 radiocarbon-dated pre-6000 cal. BP sites on both Haida Gwaii (Fedje et al. 2005; Mackie et al. 2011) and the Alexander Archipelago in Southeast Alaska (Carlson and Baichtal 2015). In both areas the sites are mostly characterized by lithic concentrations. Dense shell components appear rarer than those identified on Dundas, Lucy, and Stephens Islands. The Haida Gwaii sites have a well-recorded early Holocene lithic technological sequence, in which generalized lithic toolkits characterized by leaf-shaped bifaces and an absence of microblades have microblades introduced shortly after 9000 BP and bifaces slowly disappear several centuries later (Fedje and Mackie 2005). In Southeast Alaska, paleoshoreline sites are characterized by microblades often made from obsidian (but also include quartz crystal pieces), as well as pebble/flake tools, scattered shells, and evidence of burning (Carlson and Baichtal 2015). These sites are located on elevated marine terraces and fluvial terraces, often in locations that would have been the sides of estuaries or narrow bays, and often with south-facing aspects. The frequency of sites observed in the paleoshoreline surveys on the Alexander Archipelago have led Carlson and Baichtal (2015) to conclude that populations were quite high in that area by that time, and that people may have been concentrated by the submersion of terminal Pleistocene refugia plains that had previously been exposed by lower global sea levels and isostatic forebulging of the outer coast. The flooding of these plains contracted habitable living areas for the early occupants.  156   The success of this research in locating early sites through targeted paleoshoreline surveys, and the now-identified presence of early Holocene human occupation in the Prince Rupert Harbour area supports the hypothesis of a relatively ubiquitous early human presence on the northern Northwest Coast. In a recent study using summed probability distributions of radiocarbon dates as a proxy for demography, Brown (2016) suggests that population levels were relatively stable on the northern coast between 10,000 cal. BP and 6000 cal. BP, with the exception of an apparent (but unexplained) potential ‘population collapse’ between 8800 cal. BP and 8400 cal. BP. Brown (2016) notes, as we do, that shell-bearing sites become much more frequent after 8000 cal. BP and that there is evidence for increased investments in places (such as the construction of large permanent dwellings) – and therefore decreased residential mobility – after 7000 cal. BP. Our new data from the Prince Rupert Harbour area do not contradict these findings, and suggest that with increased investment in targeted surveys a larger sample of the early Holocene archaeological record will be located. Though our excavated sample size is very small, the apparent differences between the northern coastal Tsimshian territory sites and those in southeast Alaska and Haida Gwaii – namely larger and more frequent shell-bearing components and an absence of microblades in our study area – suggest some interesting historical trajectories worthy of further research.  3.8 Conclusions Understanding the early Holocene human history of the Northwest Coast begins with understanding the history of RSL and shoreline change and developing survey designs that account for the geomorphic processes that buried or modified archaeological remains from these 157  times. High resolution precise digital elevation models such as LiDAR-derived DTMs are invaluable for isolating landforms with a high potential for past occupation.  In our survey we mitigated logistical issues through a predictive model that targeted specific site types associated with paleoshorelines from a specific time. We have identified the earliest archaeologically-recorded sites in the Prince Rupert Harbour area, even after a century of often intensive archaeological survey. Changing methods and approaches have identified entirely new aspects of the archaeological record, from the magnificent large villages that originally attracted archaeologists to the area in the early 1900s, to the ubiquitous culturally modified trees and intertidal features that demonstrate the long-term use and management of the entire landscape that have only begun to be recorded by archaeologists in the last few decades, to the now-recognized but less-visible remains of very ancient human occupation from times when the position of the shore was totally different than that of today. Taken together, this legacy of research demonstrates an ubiquitous ancient human presence in the Prince Rupert Harbour area that now dates to at least 9000 cal. BP, though likely extends even earlier (see Letham et al. 2016:188-189). The evidence for repeated use of nearly all the areas with early and mid-Holocene dates into the late Holocene, even when coastal processes and RSL change caused complete reconfigurations of coastlines, is evidence for deep-rooted local attachments to particular places, a pattern we see on other areas of the coast where long-term continuity of use has been demonstrated (e.g. Cannon 2003; McLaren et al. 2015). Future predictive surveys targeting other paleoshoreline areas and other time periods will likely further flesh out this fascinating story of coastal occupation.  158  Chapter 4: Assessing the Scale and Pace of Large Shell-Bearing Site Occupation using Percussion Coring and 3D Mapping: Examples from the Prince Rupert Harbour Area, British Columbia  4.1 Introduction Archaeological sites composed of massive accumulations of shell are recognized worldwide as much for their significance in the history of past coastal societies (Álvarez et al. 2011; Erlandson 2013) as for the methodological challenges they pose to investigators (Claassen 1991; Lyman 1991; Stein 1992; Waselkov 1987). Shell-bearing sites are long-term records of daily activities and the products of different depositional processes, ranging from unpatterned disposal of subsistence waste to mounding and terracing to create planned landforms. These sites can represent short term or seasonal resource processing areas, inter-generational locations of everyday habitation, venues of feasting and ceremony, cemeteries, or any combination of these simultaneously or in succession. The complexity of their origins matches the heterogeneity of their stratigraphy, and the frequently deep deposits present logistical difficulties in excavating through meters of often loosely consolidated shell (Stein 1992). Over the last few decades archaeologists have worked to develop methods for adequately sampling and dating shell-bearing sites in order to interpret and model their formation histories (Cannon 2000b, 2013; Letham 2014; Martindale et al. 2009; Pluckhahn et al. 2015; Stein 1992; Thompson et al. 2016; Villagran 2014). We build on this work to study large shell sites in the Prince Rupert Harbour area of the Northwest Coast in British Columbia, Canada.   159  This area contains an exceptionally dense concentration of large shell-bearing sites that accumulated in association with large fisher-hunter-gatherer village populations, often over thousands of years of occupation (Ames 2005; Ames and Martindale 2014; Archer 2001; MacDonald and Inglis 1981). Integrating three-dimensional mapping, systematic percussion coring, and intensive radiocarbon dating, we query the temporality of occupation and deposition at two large village sites, GbTo-34 and GbTo-70, asking how space at these sites was occupied through time and the processes by which thousands of cubic meters of shell came to accumulate on what effectively became anthropogenically engineered landforms. We also integrate stratigraphic and radiocarbon data gathered from years of previous excavation at the locally well-known Boardwalk site (GbTo-31) into our spatial modelling framework. We demonstrate both ways in which extant data can be incorporated into our analyses using newer technology, and the efficiency of percussion coring over logistically demanding excavations for site-wide modelling. Through refined data collection and the application of innovative spatial and chronological modelling, we demonstrate the rich and complicated history of occupation and use at these sites. We identify instances of rapid accumulation at each site buried within slower accumulating shell. The rapid accumulations modified the shape of the landform and appear to differ from quotidian discard and site maintenance in that they may have been planned endeavours. This analysis documents the complex histories through which these and other large coastal shell-bearing sites were created via a robust full-site sampling strategy and the integration of finer-scale spatial and chronological modelling.  160  4.2 Large Shell-Bearing Sites around Prince Rupert Harbour Shell-bearing sites are abundant in the Prince Rupert Harbour area (Figure 4.1) and include several dozen strikingly large cases. The biggest extend over 20,000 m2 and have cultural deposits up to 8 m deep. Many of the largest shell-bearing sites are also village sites. Villages had large wooden plank structures erected in rows, the remains of which are visible as rectangular surface depressions resulting from either excavating house foundations into shell deposits and/or the accumulation of shell as midden around exterior house walls (Archer 2001; Coupland 2006; Patton 2011). Large villages frequently had several rows of houses, often on separate terraces increasing in elevation from front to back. The highest elevations in most villages are linear shell ridges at the back (landward) side of the site that drop steeply to the natural land surface behind. These sites often have long occupation histories of several millennia, during which tens of thousands of cubic meters of shell and other cultural debris accumulated. Consequently a site’s surface topography only reflects a small part of that location’s occupation history.  161   Figure 4.1 The Prince Rupert Harbour area with study site locations and orthophotos of study site areas.  Elucidating the developmental histories of these large sites is a critical step for exploring the settlement history of the area and for parsing out the behaviours and formation processes that led to the generation of these massive accumulations. Scholars have highlighted the importance of understanding the behaviours behind the deposition of shell at archaeological sites to refine our understanding of them as potentially being more than simply ‘middens’ – trash heaps of 162  subsistence refuse (Claassen 1991; Marquardt 2010; Stein 2001). We therefore use the more inclusive term ‘shell-bearing site’ (following Claassen 1991; Widmer 2014), and suggest that while there are midden components at the Prince Rupert Harbour villages, there may also be other shell deposits that do not necessarily have their origins as food debris.  A key debate among researchers is whether some shell-bearing sites are the result of intentional engineering (e.g. Blukis Onat 1985; Marquardt 2010; Pluckhahn et al. 2016; Randall 2015; Widmer 2014); the massive scale of some of the Prince Rupert Harbour area villages suggests this possibility. Were these landforms – on which ordered rows of clearly engineered plank houses came to be constructed – themselves engineered as occupation locations? By engineering we refer to acts of planning, contrivance, or management of the landscape. A degree of intentionality and communal coordination is implied by this definition. We evaluate these ideas by exploring both the physical effects of shell deposition on the shape of the natural landform as well as the pace of these transformations.      Most analyses of villages in the Prince Rupert Harbour area have been based on traditional excavation techniques (see Ames and Martindale 2014: table 1), which limit the ability of archaeologists to explore site development and the use of space across entire sites. The North Coast Prehistory Project (NCPP) excavated large portions of 11 major sites in the 1960s and 1970s (Ames 2005; MacDonald and Cybulski 2001; MacDonald an Inglis 1981). In subsequent decades, Coupland and colleagues excavated house depressions at five large village sites (Coupland 1988, 2006; Coupland et al. 1993, 2003, 2009; Patton 2011; Stewart et al. 2009). At the same time, Archer (2001) mapped most of the area’s villages and dated samples collected by shovel testing the upper surfaces of sixteen villages to assess chronological trends in village 163  occupation. Most recently, Millennia Ltd. excavated an entire small village (GbTo-54) as part of a CRM mitigation (Eldridge et al. 2014). None of the earlier studies explicitly addressed the developmental histories of the sites, though Ames (2005) attempted to interpret much of the stratigraphy recorded from the NCPP excavations to place the artifact assemblages into time and space, and Coupland et al. (2003) analyzed the differential use of space at the McNichol Creek village site (GcTo-6). Eldridge et al. (2014) provide a robust analysis of site chronology and use areas at GbTo-54, though that project excavated the entire site with a large crew and a very large budget, resources not normally available.  Our study examines the developmental histories of (1) Kitandach/GbTo-34, a large village excavated by the NCPP in 1971 and 1972 at which several dates suggest nearly 6000 years of occupation; (2) GbTo-70, a smaller village where Archer (2001) obtained terminal dates between 2000 and 1500 cal. BP from the back ridge; and (3) Boardwalk/GbTo-31, a site which was the focus of major excavations by the NCPP from 1968 through 1970 (Ames 2005; MacDonald 1969; MacDonald and Inglis 1981; Stewart and Stewart 1996) and by Coupland in 2000 and 2003 (Coupland et al. 2002, 2006, 2010; Stewart et al. 2009), with occupation dates over the last 5000 years (Figure 4.2, Table 4.1). In contrast to large-scale excavations, our methods are low-impact and cost/time efficient.  164   Figure 4.2 Original study site maps with house depressions and previous excavation units indicated. Core test (CT) locations are indicated. 165  Table 4.1 Study Site Characteristics. Site Shell Area (m2) Volume (m3) Number of Houses Number of House Rows Total House Floor Area (m2) Average House Area (m2) House Size Range (m2) House Size SD (m2) Number of ESP Core Samples Number of C14 Dates GbTo-34 8200 12900 >17 at least 2 >949.1 55.8 18.2-220.0 53.4 40 65 (42 unique contexts) GbTo-70 2800 4170 11 1 448.6 40.8 31.3-54.4 6.1 20 20 GbTo-31 6000 n/a 2* 1* 115.3* 57.7 54.1-61.3 5.1 0 45 *Does not represent total number of houses at the village. Excavators found evidence for houses on terraces where house depressions were not visible. Either the surface of GbTo-31 was disturbed after house occupation, obscuring other depressions, or house construction/post-occupation formation processes differed from at the other villages. 166  4.3 Methods Archaeologists around the world have grappled with the stratigraphic complexity and logistical difficulties of studying shell-bearing sites (Claassen 1991; Lyman 1991; Rick and Waselkov 2015; Stein 1992; Stein et al. 2003; Waselkov 1987). Deep shell accumulations may build up in short bursts, yet excavating to the bottom of these deposits is time consuming, expensive, and often not possible without large soundings and shoring. Furthermore, excavations only provide a window into a small area of a site, limiting our ability to assess intra-site patterning.  Recent research has developed methodological alternatives to excavation for addressing site-scale analyses of large shell-bearing sites or ways to plan targeted excavations, including geophysical remote sensing (Thompson et al. 2014), airborne LiDAR mapping (Randall 2014), and wide-coverage small-scale subsurface sampling such as percussion coring (Martindale et al. 2009; Thompson et al. 2016). Other research approaches include modelling of large sets of radiocarbon dates to refine the understanding of chronology and temporality at sites (Pluckhahn et al. 2015; Thakar 2014).  Recent theoretical debate among coastal archaeologists has centred around whether large shell-bearing sites represent the unconscious accumulation of waste by-products (i.e. middens) or deposition directed toward monumental constructions with shell (Marquardt 2010; Pluckhahn et al. 2016; Randall 2015; papers in Wallis and Randall 2014; Widmer 2014), which has stimulated sophisticated geoarchaeological analyses of sediment characteristics and site formation processes (e.g. Pluckhahn et al. 2015; Villagran 2014; Villagran et al. 2011). We employ systematic percussion coring, 3D surface mapping, and intensive radiocarbon dating to efficiently collect 167  and analyze a more representative sample for evaluating the chronology, tempo, and scale of deposition behaviours at the study sites.  4.3.1 Documenting Site Layout with 3D Surface Mapping We mapped the surface topography of GbTo-34 and GbTo-70 using a Leica Viva GS15 RTK GNSS unit with a base station for open areas and a Leica TC805 Total Station with a prism on a stadia rod for forested areas that received weak satellite signals. Data collected with both devices were combined and all data were post-corrected to be relative to geodetic mean sea level (i.e. m asl) as measured against the Canadian Geographic Vertical Datum of 1928 (CGVD28) benchmark at Prince Rupert. The horizontal and vertical accuracy of measurements made with these devices is ±5 cm for all three dimensions under ideal conditions. In addition to general surface topography, we mapped site boundaries, dimensions of structures and other surface features, and the locations of all core tests. For GbTo-34, 5264 individual mapping points were collected; for GbTo-70, 2303 points were collected. Three-dimensional surfaces were derived from these points using the Kriging interpolation method set at 1 m grid intervals in Golden Software’s Surfer program. The result is a spatially accurate and precise 3D rendering of the ground surface of the sites in which surface structural depressions are readily visible.    We also used bare earth digital terrain models (DTMs) derived from LiDAR remote sensing, flown by Airborne Imaging (2013) and provided to us by Nexen Ltd. The LiDAR DTM is a 1 m resolution grid generated from data collected with a Leica ALS70 system 179 and a Leica ALS70 system 207 and corrected to CGVD28, making it directly comparable with our field mapped measurements. Large shell-bearing sites with topographic relief show up clearly in the 168  DTMs, as do large house depression at these sites, though the field-mapped surfaces are clearer renderings because data were judgmentally gathered. We therefore overlay our RTK- and Total Station-derived surfaces on broader landscape surface models clipped from the LiDAR for areas that we did not map ourselves in the field.  We used the LiDAR to georectify earlier site maps of GbTo-31; many features on the NCPP hand-drawn maps were visible in the DTM, including old excavation units that were not backfilled. This allowed us to place many of the NCPP stratigraphic profiles, unit depths, and radiocarbon date positions in real-world space in order to make them directly comparable to our datasets from GbTo-34 and GbTo-70. The result is a landscape model of the site areas containing accurate and precise surface topography and three-dimensionally located subsurface tests.  4.3.2 Sampling the Subsurface with Percussion Coring We collected core samples in a 10-20 m interval grid pattern across GbTo-34 (n=40) and GbTo-70 (n=20) using a 3 cm diameter JMC Environmentalist’s Sub-soil Probe that collects sequential 95 cm vertical segments. The specifics of this device and its application in Northwest Coast shell-bearing sites are published in Cannon (2000b) and Martindale et al. (2009). At GbTo-34 and GbTo-70 our deepest cores retrieved 5 m of stratified sediment.  We logged stratigraphy in each core based on changes in sediment type, grain size, and colour, as well as mollusc shell abundance, species composition, and fragmentation. These characteristics helped to qualitatively assess macrostratigraphic similarities across the sites. We collected samples of both marine shell and charcoal from areas of stratigraphic interest for 169  radiocarbon dating. We estimated the ‘real’ depth of strata by multiplying the compressed depth by the ratio of the actual depth collected by a given segment to the length of the compressed material in the segment. This value was consistently between 1.2 and 1.8 times.5   Three-point provenienced core locations allowed comparison of the cores in relation to each other to assess for similar sedimentary components or stratigraphic breaks across space. We used the depth of the uppermost and lowermost instances of shell in each core to create interpolated surfaces of both the buried top of the cultural deposit and the natural land surface beneath the site. These latter surfaces allow us to assess the degree to which shell deposition modified the original shape of the landform, and allow us to estimate the volume of shell accumulated at the site.   Having estimates of the elevations of the top and bottom of cultural deposits and the configuration of the landform prior to human occupation allows us to assess the interplay between occupation of the sites and relative sea level (RSL) change. RSL change in the study area was most dramatic during the terminal Pleistocene and early Holocene, although during the last 6000 years (the time during which the study sites were occupied) RSL dropped from <~5 m asl to its present position by about 1500 cal. BP (Letham et al. 2016). For the majority of this time RSL was 1-3 m above its current position. During the latter half of the Holocene humans                                                  5 In a past study on the Dundas Islands (Martindale et al. 2009) we applied a constant expansion factor to humic material based on experimental calculations; however, we found that humic material compressed to very different degrees at different sites in the Prince Rupert area, so we opted not to apply differential expansion factors to different strata in this study. Further experimental work is required to assess if we can determine specific expansion factors for specific sediment types. 170  occupied a gradually changing shoreline, and we are able to estimate shoreline position and its relationship with the locations of human occupation and deposition through time.  4.3.3 Radiocarbon Dating Provenienced core samples allow targeted selection of organic material for radiocarbon dating of various contexts to assess occupation duration and intensity, deposition rates, and changes in use of different spaces across the site. At GbTo-34 we dated 60 samples from 37 unique contexts in 11 cores and a deep profile exposed in a wave cut bank at the shoreline. Additionally, the NCPP dated five samples from their excavations at the site (Ames 2005:65). We dated samples from 18 unique contexts in five cores from GbTo-70 and include two dates from Archer’s research (Archer 2001). We did not date any new samples from GbTo-31, but 45 dates have been published and we include an additional 21 previously unpublished dates on human remains run by Jerome S. Cybulski (Ames 2005; Cybulski 2014; Coupland et al. 2010). All samples collected by us were taken from archaeological strata, and their selection from core samples ensured that they came from secure stratigraphic contexts. The dated samples from the NCPP often lack precise provenience information and have much larger error estimates as a result of being run prior to AMS dating technology, however we have no reason to reject any of these dates as they are for the most part in stratigraphic order and consistent with dating by more recent projects. All radiocarbon dates discussed in the text are presented as 2-sigma calibrated ranges; the marine reservoir effect has been accounted for in the calibration of marine samples through applying a ΔR of 273+/-38, which was derived for the Prince Rupert Harbour area through a multiple-paired sample approach applied to three separate contexts from GbTo-34 that span the last 5000 years 171  (Edinborough et al. 2016). All dates used in this study, along with calibrations, are detailed in Appendix C.1.  Spatial control of radiocarbon dates allows for mapping the depositional history of the sites through both time and space. We calculated accumulation rates for deposition between pairs of dated locations in vertical succession (n=54) by dividing the depth of deposit between the two dated locations by the age difference between the medians of the calibrated ranges of the two dates (Tables 4.2 and 4.3; Stein et al. 2003)6. This allowed us to assess locations and timing of rapid deposition resulting from intensive occupation or potential intentional construction events.                                                     6 We use calibrated medians because the calculations require single ages, though we recognize that medians of calibrated radiocarbon probability ranges do not necessarily represent the age with the highest probability within the range. 172  Table 4.2 All accumulation rates from GbTo-34, GbTo-70, and GbTo-31. Rates are calculated using calibrated median ages for vertically subsequent date pairs within each core/unit.  GbTo-34 Core/Unit Lab Numbers Elevation of Dated Sample (m asl) Median Age (cal. BP)1 Depth Difference (cm)2 Age Difference (years)3 Accum