Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The ecology of Picea glauca (Moench) Voss at its range limits in northwest Canada McLeod, T. Katherine 2002

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

Item Metadata

Download

Media
831-ubc_2002-732088.pdf [ 20.98MB ]
Metadata
JSON: 831-1.0103841.json
JSON-LD: 831-1.0103841-ld.json
RDF/XML (Pretty): 831-1.0103841-rdf.xml
RDF/JSON: 831-1.0103841-rdf.json
Turtle: 831-1.0103841-turtle.txt
N-Triples: 831-1.0103841-rdf-ntriples.txt
Original Record: 831-1.0103841-source.json
Full Text
831-1.0103841-fulltext.txt
Citation
831-1.0103841.ris

Full Text

THE E C O L O G Y OF PICEA GLAUCA (MOENCH) VOSS A T ITS R A N G E LIMITS IN NORTHWEST C A N A D A . by T. KATHERINE M C L E O D B.Sc., McMaster University, 1988 M.Sc., McMaster University, 1991 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE F A C U L T Y OF G R A D U A T E STUDIES DEPARTMENT OF G E O G R A P H Y We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH C O L U M B I A OCTOBER 2001 © T. Katherine McLeod, 2001 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of The University of British Columbia Vancouver, Canada Date flcK 31^ 3.001 DE-6 (2/88) A B S T R A C T The northernmost conifers in North America are Picea glauca (Moench) Voss tree islands, located in the Tuktoyaktuk region, which encompasses Tuktoyaktuk Peninsula and the lower Anderson and Horton River valleys to the east. Using, ecological and dendroecological techniques, the role of these tree islands in the range of Picea glauca was examined. Three alternative hypotheses concerning Picea glauca tree islands were proposed: (i) the tree islands are of recent origin, (ii) the tree islands are a normal component in the range of Picea glauca, and (iii) the tree islands are relicts of past treeline advances. Specific objectives were (a) to locate and describe the northernmost Picea glauca, (b) ascertain the relative importance of vegetative and sexual reproduction, (c) examine current reproduction, and (d) determine the age structures of the tree island populations. Fifteen tree island sites were located in the Tuktoyaktuk region, with one to >300 tree islands at each site. Male cones were produced at all sites and female cones at most sites. There was a significant negative linear relation between seed production and distance from treeline. The germinability of seeds collected was less than 3% at tree island sites and reached a maximum of 33% within the Forest-Tundra. Seedbank soil samples yielded no germinable seed of Picea glauca. The low level of seed germinability at tree island sites was reflected in the lack of seedlings across the region. Survivorship of transplanted seedlings was low over two growth seasons, varying from 20% to 45%, with no significant difference in survival inside versus outside the tree islands. In addition to low levels of seed germinability and limiting microclimatic conditions, other factors, such as competition from tundra vegetation, may contribute to low seedling recruitment levels. Tree islands maintained their populations primarily through vegetative reproduction i i by layering. There was a significant linear relation between the distance north of treeline and the proportion of the tree islands that established by layering or from seed. Reproductive origin and tree ring analysis revealed that large numbers of individuals established by layering in the 20th century. In contrast, only a few individuals established from seed and survived in most decades since 1700. The continuous nature of seedling establishment over the period of record and the lack of coincidence between establishment and published reconstructed climatic conditions, suggest the importance of site specific factors in successful sexual and vegetative regeneration of Picea glauca at its range limit. Light ring chronologies developed for six tree island sites showed a good correlation with other light ring chronologies from published sites in northwest Canada. Cross-dated dead stems and radiocarbon dates of wood remains extended the age of the tree islands back to the latter part of the Little Climatic Optimum (ca. A D 1000 to A D 1350). From these findings, it is suggested that tree islands in the Tuktoyaktuk region are relicts of more favourable climatic conditions in the past, when tree line and/or the species range limit advanced northward. in TABLE OF CONTENTS A B S T R A C T i i LIST OF T A B L E S vii LIST OF FIGURES viii A C K N O W L E D G E M E N T S xii CHAPTER 1 INTRODUCTION 1 1.1 Introduction to the Study 1 1.2 Research Hypotheses and Objectives 3 1.3 Thesis Structure 4 CHAPTER 2 NORTHERN TREES 6 2.1 Northern Vegetation Zones 6 2.2 Circumpolar Forest-Tundra and Treeline 8 2.2.1 Northern North America 8 2.2.2 The Tuktoyaktuk Region 11 2.2.3 Eurasia 12 2.3 Factors Affecting Northern Vegetation 13 2.4 Holocene Vegetation History of Northwestern Canada 17 CHAPTER 3 PHYSICAL ENVIRONMENT OF THE T U K T O Y A K T U K REGION 22 3.1 Introduction 22 3.2 Physiography 22 3.3 Bedrock Geology 25 3.4 Quaternary Record 25 3.4.1 Glacial Chronology, Banks Island and Tuktoyaktuk Coast 27 3.5 Climate 29 3.5.1 Low Arctic Ecoclimatic Region 30 3.5.2 High Subarctic Ecoclimatic Region 33 3.5.3 Homogeneous Time Series of Temperature and Precipitation . .33 3.6 Permafrost Conditions and Periglacial Landforms 35 3.7 Soils 38 CHAPTER 4 DISTRIBUTION A N D DESCRIPTION OF TREE ISLANDS 39 4.1 Introduction 39 4.2 Methods 39 4.3 Tree Islands in the Tuktoyaktuk Region 45 4.4 Identification and Differentiation of Picea glauca from Picea mariana . 51 4.5 Substrate Conditions 53 iv < 4.5.1 Soil Descriptions 53 4.5.2 Active Layer Thickness ; .' 55 4.6 Non-Arboreal Vegetation 56 4.7 Insects 57 CHAPTER 5 REPRODUCTION 59 5.1 Introduction . . . 59 5.2 Methods 62 5.2.1 Current Sexual Reproductive Capacity 62 5.2.2 Seedling Survival 64 5.2.3 Reproductive Origin of Individuals 66 5.2.4 Data Analysis 68 5.3 Results: Current Sexual Reproductive Capacity 68 5.3.1 Cone Production 68 5.3.2 Seed Germination Tests 70 5.3.3 Seedlings 73 5.4 Results: Reproductive Origin of Individuals 77 5.4.1 Propensities for Vegetative and Sexual Reproduction 77 5.4.2 Spatial Pattern of Reproduction within Tree Islands 80 5.5 Discussion 94 CHAPTER 6 A G E STRUCTURE A N D ESTABLISHMENT 102 6.1 Introduction 102 6.2 Methods 105 6.2.1 Data Analyses 119 6.3 Results: Reproductive Origin and Establishment Patterns I l l 6.3.1 Tuktoyaktuk Peninsula I l l 6.3.2 Anderson River Valley Sites 113 6.3.3 Horton River Valley Site 116 6.3.4 Maximum Age of Live Trees 120 6.3.5 Establishment and Climate Change 120 6.4 Discussion 130 CHAPTER 7 TREE RING CHRONOLOGIES : 136 7.1 Introduction 136 7.2 Methods 137 7.3 Tree Ring Chronologies 141 7.3.1 Tree Island Chronologies 141 7.3.2 Comparison with Other Sites in Northwestern North America . 144 7.3.3 Spatiotemporal Patterns and Causes 146 7.4 Mortality * 148 7.5 Establishment Along the Tuktoyaktuk Peninsula 153 7.6 Tree Island Age 161 7.7 Conclusions . 162 v CHAPTER 8 THESIS CONCLUSIONS 164 8.1 Summary 164 8.2 Contributions to the Tree Island Hypotheses 167 8.3 The Role of Tree Islands in Reconstructing Range Limits and Treeline Fluctuations 169 8.3.1 Past Fluctuations: Tuktoyaktuk Peninsula 169 8.3.2 Past Fluctuations: Anderson and Horton River Valleys 171 8.3.3 Future Vegetation Change 171 8.4 Future Research 172 8.4.1 Work in Progress 172 8.4.2 Future Research Endeavours 173 REFERENCES 177 APPENDIX 1 189 A l . Location and Description of Sites 189 A l . l Richards Island Sites 189 A1.2 Tuktoyaktuk Peninsula Sites 189 A1.3 Anderson River Valley Sites 194 A 1.4 Horton River Valley Sites 198 vi LIST O F T A B L E S Table 3.1 Climate data for stations along a west to east transect from the western Beaufort Sea to the eastern end of Amundsen Gulf. 34 Table 4.1 The morphological characteristics used to distinguish Picea glauca from Picea mariana 42 Table 4.2 Characteristics of Picea glauca tree islands in the Tuktoyaktuk region... 46 Table 4.3 Morphological characteristics of Picea glauca at sites within the Tuktoyaktuk Region 52 Table 4.4 Active layer depths measured both inside and outside selected tree islands 56 Table 5.1 The proportion of each tree island population that produced cones in 1993, 1994, 1995 69 Table 5.2 The proportion of live tree islands that produced cones in 1993, 1994, 1995, calculated for sites with multiple tree islands that were not intensively studied 69 Table 5.3 Germination results for Picea glauca seed collected from sites within the Tuktoyaktuk Region 74 Table 5.4 Survivorship of transplanted Picea glauca seedlings over two growth seasons .77 Table 5.5 The proportion of each sampled tree island and Picea glauca population that originated from either vegetative or sexual reproduction 78 Table 5.6 Picea glauca seed germinability from southern sources 95 Table 6.1 Results of Chi-Square Tests between the decadal patterns of reconstructed climate and the decadal patterns of seedling establishment for the Tuktoyaktuk Peninsula 129 Table 6.2 Results of Spearman's rank correlation analysis between the reconstructed climate records and the decadal seedling establishment in the Tuktoyaktuk Peninsula and Anderson River valley areas 129 Table 7.1 Results of the time series analyses of the strongly registered light ring year chronologies developed 145 vii LIST OF FIGURES Figure 1.1 The location of the Tuktoyaktuk region 2 Figure 2.1 The location of vegetation zones and treeline within the Tuktoyaktuk region 7 Figure 2.2 Schematic depiction of terms used in this study to describe the northern vegetation zones and limits 9 Figure 2.3 Summary percentage pollen diagram for a southern Tuktoyaktuk Peninsula site based on corrected pollen totals using R values (from Ritchie, 1984) 19 Figure 2.4 Radiocarbon dates ( 1 4 C BP) of wood reported in Spear (1983), Ritchie (1984), Mackay (1992), and from this study 21 Figure 3.1 Physiographic regions of the study area (after Mackay, 1963; Rampton, 1988a) 23 Figure 3.2 Generalized bedrock geology in the study region. Mapping based on Young et al. (1976), Yorath et al. (1980) and Okulitch (1991) 26 Figure 3.3 Wisconsinan ice limits in northwestern Canada (after Hughes, 1987; Vincent, 1989) 28 Figure 3.4 Locations of climate stations in the study area 31 Figure 3.5 Climate normals from: a) Sachs Harbour A , b) Tuktoyaktuk Peninsula, c) Inuvik A, and d) Norman Wells (data from Environment Canada 2000) 32 Figure 3.6 Mean annual temperature and total precipitation departures for the Mackenzie Valley and the Arctic Tundra climate regions 36 Figure 4.1 The location of the tree island sites in the Tuktoyaktuk region 40 Figure 4.2 Growth forms in P i c e a (from Lavoie and Payette, 1992) 44 Figure 4.3 Tree island T5 located on Tuktoyaktuk Peninsula (see Figure 4.1). This tree island is the only tree island in the area and completely surrounded by shrub tundra 47 Figure 4.4 Site A2 located within the Anderson River valley (see Figure 4.1). There are more than 80 tree islands in the immediate area 47 Figure 4.5 Tree island T4 located along the Tuktoyaktuk Peninsula (see Figure 4.1). The P i c e a g l a u c a at this site typically exhibit a mat growth form, with few infranival erect stems (see Figure 4.2) 49 Figure 4.6 Tree island A3c located within the Anderson River valley (see Figure 4.1). Individuals shown in this photograph exhibit supranival skirted and erect whorled growth forms (see Figure 4.2) 49 Figure 4.7 A map showing the locations of 16 tree islands at site T6 (see Figure 4.1). Tree islands at this site were found on various slopes with east, west, south, or north facing aspects , 50 Figure 4.8 Tree island T2, located along the Tuktoyaktuk Peninsula. The substrate conditions at this site are well drained gravelly sands 54 Figure 5.1 Depiction of vegetative reproduction by layering in P i c e a . A lower branch is buried, roots form adventitiously, and the shoot tip continues to grow, forming a new individual 60 Figure 5.2 The location of tree island and forest stand sites in the Tuktoyaktuk viii region 63 Figure 5.3 This photograph shows three individuals produced by layering and their underground connections 67 Figure 5.4 This photograph shows individuals that had established by seed, as determined by the presence of roots and by the lack of underground connections 67 Figure 5.5 The decline in female cone production with increasing distance north of treeline 71 Figure 5.6 The decline in male cone production with increasing distance north of treeline 72 Figure 5.7 P i c e a g l a u c a seed germination from sites along the Tuktoyaktuk transect (see Figure 5.2) 75 Figure 5.8 P i c e a g l a u c a seed germination from sites along the Anderson River transect (see Figure 5.2) 76 Figure 5.9 The live proportion of each sampled tree island and forest stand that originated from either layering or from seed compared to site location.. 81 Figure 5.10 Map of site T1, Mackay Spruce 82 Figure 5.11 Map of site T3, Drillpad Spruce 83 Figure 5.12 Map of site T5, Hilltop Spruce.. 84 Figure 5.13 Map of site T6i, Brunet Spruce 85 Figure 5.14 Map of site A l a , Armstrong Spruce 86 Figure 5.15 Map of site A l b 87 Figure 5.16 Map of site A2a, Spurned Swan Spruce 88 Figure 5.17 Map of Site A2b 89 Figure 5.18 Map of site A3 a 90 Figure 5.19 Map of site A3b , 90 Figure 5.20 Map of site A3c, Caribou Spruce 91 Figure 5.21 Map of site HI a, Horton Spruce 92 Figure 5.22 This photograph shows a toppled individual at site T6i that has produced many individuals by layering 93 Figure 6.1 The location of tree island sites in the Tuktoyaktuk region that were used in this study 104 Figure 6.2 Internodal growth and age determination (from Fritts, 1976) 107 Figure 6.3 Establishment of P i c e a g l a u c a by vegetative and sexual reproduction at site T l (see Figure 6.1) 112 Figure 6.4 Establishment of P i c e a g l a u c a by vegetative and sexual reproduction at site T3 (see Figure 6.1).. 112 Figure 6.5 Establishment of P i c e a g l a u c a by vegetative and sexual reproduction at site T5 (see Figure 6.1) 114 Figure 6.6 Establishment of P i c e a g l a u c a by vegetative and sexual reproduction at site T6i (see Figure 6.1) 114 Figure 6.7 Establishment of P i c e a g l a u c a by vegetative and sexual reproduction at sites across the Tuktoyaktuk Peninsula (see Figure 6.1) 115 Figure 6.8 Establishment of P i c e a g l a u c a by vegetative and sexual reproduction at site A l a (see Figure 6.1) : 117 ix Figure 6.9 Establishment of Picea glauca by vegetative and sexual reproduction at site A l b (see Figure 6.1) 117 Figure 6.10 Establishment of Picea glauca by vegetative and sexual reproduction at site A2a (see Figure 6.1) 118 Figure 6.11 Establishment of Picea glauca by vegetative and sexual reproduction at site A3c (see Figured. 1) 118 Figure 6.12 Establishment of Picea glauca by vegetative and sexual reproduction at all sites within the Anderson River valley (see Figure 6.1) 119 Figure 6.13 Establishment of Picea glauca by vegetative and sexual reproduction at site H l a (see Figure 6.1) 121 Figure 6.14 Annual temperatures at Inuvik Airport 122 Figure 6.15 June, July, August temperatures at Inuvik Airport 123 Figure 6.16 Annual precipitation at Inuvik Airport 124 Figure 6.17 June, July, August precipitation at Inuvik Airport 125 Figure 6.18 Establishment by seed at Tuktoyaktuk Peninsula and Anderson River sites and climatic reconstructions 127 Figure 6.19 Establishment by layering at Tuktoyaktuk Peninsula and Anderson River sites and climatic reconstructions 128 Figure 7.1 Location of tree island sites in the Tuktoyaktuk region that were used in this study 138 Figure 7.2 Temporal distribution of light rings at six tree island sites (see Figure 7.1). The light ring chronology developed by Szeicz (1996) (labelled CDU) is also depicted 142 Figure 7.3 Summer (June, July, August) temperatures for Inuvik Airport with strongly registered light ring years indicated by open circles 147 Figure 7.4 The distribution of ages at death for trees in tree islands within the Tuktoyaktuk region 149 Figure 7.5 Examples showing the cross-dating of samples from dead individuals located at tree island T3 (see Figure 7.1) 151 Figure 7.6 Mortality (number of dead trees per decade) (a) at tree island sites along the Tuktoyaktuk Peninsula and (b) at H l a within the Horton River valley. 152 Figure 7.7 Establishment of Picea glauca, including currently live and dead individuals, by vegetative and sexual reproduction at site TI (see Figure 7.1) 154 Figure 7.8 Establishment of Picea glauca, including currently live and dead individuals, by vegetative and sexual reproduction at site T3 (see Figure 7.1) 155 Figure 7.9 Establishment of Picea glauca, including currently live and dead individuals, by vegetative and sexual reproduction at site T5 (see Figure 7.1) 156 Figure 7.10 Establishment of Picea glauca, including currently live and dead individuals, by vegetative and sexual reproduction at site T6i (see Figure 7.1) 157 Figure 7.11 Establishment of Picea glauca, including currently live and dead x individuals, by vegetative and sexual reproduction at sites across the Tuktoyaktuk Peninsula (see Figure 7.1) 159 Figure 7.12 Establishment of P i c e a g l a u c a , including currently live and dead individuals, by vegetative and sexual reproduction at site H l a (see Figure 7.1) 160 xi A C K N O W L E D G E M E N T S At each stage of this work, numerous people have contributed their knowledge, time, energy, and support. I have been very fortunate and I would like to express my thanks. To start, I'd like to thank my supervisor, Dr. Greg Henry for his guidance, understanding, and enthusiasm throughout this undertaking. It is an uncommon opportunity to conduct reasearch in the Canadian north, while studying at a Canadian university, and I fully appreciate the opportunity Greg has given me. Invaluable assistance and commentary were provided by my committee: Drs. Brian Klinkenberg, and Valerie LeMay, and by my examination committee: Drs. Michael Church, Roy Turkington, and Dan Smith. Their diverse expertise has strengthened this work dramatically. In the north, Isobel Boothe was a cheery, untiring field assistant. I would like to acknowledge the importance of the Polar Continental Shelf Project (PCSP) base in Tuktoyaktuk in facilitating my field research and in bringing together northern scientists. The interactions I had with colleagues and workers at PCSP were very rewarding and invaluable to my work. In particular, Dr. Terry Armstrong, Mr. Claude Brunet, Dr. Chris Burn, and Dr. Patricia Sutherland provided the locations of several tree island sites that may have otherwise not been located. At UBC, Eryn Smith and Jeanette Young helped in sample preparation, particularly laudable given the cramped and windowless work conditions. Dr. Rob Guy provided valuable information on greenhouse seedling production. I consider myself extremely fortunate that I undertook my PhD at U B C while Dr. Ross Mackay was still actively conducting his research and his graduate course. His enthusiasm for northern work, knowledge of the study area and tree island sites, and amiable xii prodding helped immensely. In addition, his career has been an inspiration. I'd like to thank my grad mates Jill Johnstone, Matthais Jacob, Manon Desforges, and Glenda Jones, who created a friendly, stimulating, inviting, and exciting environment, making it much easier to come in to work on the sunny summer days and the rainy winter ones too. Finally, and most importantly I'd like to acknowledge the huge support I have had from my family. My husband, Jim Hamilton, has been a source of love, inspiration, encouragement, and patience. He is an excellent (and untiring) field worker, with an incredible observational ability. From start to finish, his support of my work and career has been phenomenal. My two girls, having arrived post-field work and pre-dissertation, have been exposed at a very early age to the rather insane life of a grad student in the process of "finishing-up". They have been an amazing source of love and affection, and provided a wonderfully hectic refuge from the academic world. This research was financially supported by Northern Scientific Training Program grants to myself and NSERC Paleoecological Circumpolar Treeline (PACT) grants to Greg Henry. Logistical support was provided by PCSP. xiu 1 C H A P T E R 1 I N T R O D U C T I O N 1.1 Introduction to the Study In North America, the northernmost coniferous trees are Picea glauca (Moench) Voss, located within the Tuktoyaktuk region, an area that encompasses Tuktoyaktuk Peninsula, and the lower Anderson and Horton River valleys to the east (Figure 1.1). Picea glauca trees occur as outlying patches located north of treeline within the Low Arctic Shrub Tundra. These groups of trees, called tree islands, are composed of several short individuals that may have erect or prostrate growth forms. Little is known of the ecology or biogeography of this outlying vegetation, especially in northwest Canada. In eastern Canada, Picea glauca is an important treeline species in coastal areas of Hudson Bay, James Bay, Ungava Bay, and the Atlantic coast. Some isolated tree islands occur north of treeline in these areas (Payette, 1983; Payette and Filion, 1985). Outlying tree islands within the Low Arctic Shrub Tundra are thought to be relict populations, established in the past under more favourable climatic conditions. Within the Forest-Tundra, Picea glauca has been shown to rely primarily on sexual reproduction for regeneration (Payette, 1976; Payette and Gagnon, 1979; Elliott-Fisk, 1983; Payette and Filion, 1985). Correspondence between establishment and tree ring width patterns suggest a close relation between favourable climatic periods and regeneration (Payette, 1976). Over the last 100 years, favourable climatic conditions have resulted in high levels of seedling establishment. This has led to an increase in the density of Picea glauca within the Forest-Tundra and a rise in altitudinal timberline by tens of metres, but there has been no recent 2 Figure 1.1 The location of the Tuktoyaktuk region. Treeline is the boundary between the Low Actic Shrub Tundra and the High Subarctic Forest-Tundra. The location of treeline is approximated from Timoney (1988). 3 change in the position of latitudinal treeline (Payette and Filion, 1985). At the Subarctic timberline in northwestern Canada, there has been a similar minor increase in the upper limit of trees and an increased density of Picea glauca within the Forest-Tundra (Szeicz and MacDonald, 1995a). The focus of this thesis is the ecology of Picea glauca tree islands in the Tuktoyaktuk region of northwest Canada. This research documents the distribution of tree islands north of treeline in the region, and determines how these trees are able to maintain their populations under the physiologically limiting conditions of an arctic environment. 1.2 Research Hypotheses and Objectives There are three alternative hypotheses concerning Picea glauca tree islands located north of treeline: (i) Tree islands are of recent origin, established in response to recent climatic amelioration, perhaps since the Little Ice Age or twentieth century warming. (ii) Tree islands are a normal component in the range of Picea glauca. They are in equilibrium with current climatic conditions, occurring within selected favourable microsites. They are able to maintain their population mainly through sexual reproduction. If the tree islands are destroyed, they could re-establish in the area. (iii) Tree islands are relicts of past treeline advances, perhaps since the Hypsithermal (ca. 10000 to 4500 BP in northwest Canada) or the Little Climatic Optimum (ca. 950 to 600 BP). After treeline retreated to its modern position, small groups of trees survived in stunted form. These remnant trees maintain their population mainly through vegetative reproduction. Establishment of seedlings would be rare. Under 4 this scenario, the tree islands are not in equilibrium with current climatic conditions and i f destroyed they could not re-establish in the area. This thesis will contribute evidence in support of one of these hypotheses. There are four main objectives: (i) to locate and describe the northernmost P i c e a g l a u c a tree islands in the Tuktoyaktuk region, (ii) to ascertain the importance of vegetative and sexual reproduction in the maintenance of tree islands, (iii) to compare the production of cones and germinable seed along transects from the Forest-Tundra northward to the Low Arctic Shrub Tundra tree islands, and (iv) to determine the age structure of tree island populations. This will provide information on establishment patterns and will also provide a minimum date for the origin of tree islands in the Tuktoyaktuk region. 1.3 Thesis Structure In Chapter 2,1 provide context for this research. Definitions of northern vegetation zones and boundaries are presented, circumpolar treeline is described, and a more detailed description of the distribution of trees in northwestern Canada is presented. The Holocene vegetation history for northwestern Canada is summarized. In Chapter 3, I describe the physical environment of the Tuktoyaktuk region, including the physiography, geology, climatic regimes, permafrost conditions, periglacial landforms, and soils. The tree island site locations, and descriptions of tree islands and site characteristics are presented in Chapter 4. In Chapter 5, I focus on reproduction in the northern P i c e a g l a u c a populations. The 5 importance of vegetative and sexual reproduction in the establishment of individuals within the tree islands is assessed. In addition, the results concerning the current sexual reproductive capacity of northern P i c e a g l a u c a are presented. Chapter 6 is an examination of the age structures and establishment patterns of the live P i c e a g l a u c a populations in the Tuktoyaktuk region. Establishment is compared with instrumental and reconstructed climate records from Inuvik and northwestern North America. Chapter 7 concerns the development of light ring chronologies from selected tree islands and the cross-dating of dead P i c e a g l a u c a . Minimum ages of the tree islands are determined using the tree ring information and radiocarbon dates of wood remains. In Chapter 8,1 synthesize the results from Chapters 4 through 7 in order to examine the role of tree islands in the range of P i c e a g l a u c a . Conclusions are presented concerning the importance of tree islands in determining past changes in the distribution of P i c e a g l a u c a , and concerning the potential for tree islands as nuclei for rapid range expansion under ameliorating climatic conditions. 6 CHAPTER 2 NORTHERN TREES 2.1 Northern Vegetation Zones Many definitions have been developed to delineate the northern vegetation of Canada (e.g., Rowe, 1972; Hustich, 1979; Atkinson, 1981; Payette, 1983; Timoney et al, 1992). Terminology and mapped vegetation zones and boundaries used in this thesis follow those of Timoney et al. (1992, Figure 2.1). Variation in arboreal vegetation from southern areas of the Boreal Forest to the northern Forest-Tundra largely consists of changes in species diversity, density, local distribution, crown closure, and size of trees. Within the western Boreal Forest, the dominant tree species are P i c e a m a r i a n a (Mill.) B.S.P., P i c e a g l a u c a , P i n u s b a n k s i a n a Lamb., P i n u s contorta Loud. var. latifolia Engelm., L a r i x l a r i c i n a (DuRoi) Koch, Abies balsamea (L.) Mi l l . , B e t u l a p a p y r i / e r a Marsh, P o p u l u s tremuloides Michx., and P o p u l u s balsamifera L. The forest has a closed canopy and a continuous distribution, with large trees and relatively high densities (Ritchie, 1987; Elliott-Fisk, 1988; Pastor and Mladenoff, 1992). Towards the northern limits of the Boreal Forest and into the Forest-Tundra, tree species diversity decreases, with P i c e a m a r i a n a , P i c e a g l a u c a , and L a r i x l a r i c i n a dominating the landscape. P o p u l u s tremuloides, P o p u l u s balsamifera and B e t u l a papyrifera are also found in small numbers, typically where there have been flood or fire disturbances. The northern boundary of the Boreal Forest is the limit of continuous forest, where forests still occur in upland and lowland areas and the tree to tundra surface cover ratio is 1000:1 (Figure 2.2). North of this boundary is the Forest-Tundra, a transition zone between the Boreal Forest and the Shrub Tundra. Northward across this zone, the canopy 7 Figure 2.1 The location of vegetation zones in the Tuktoyaktuk region. Treeline is the boundary between the Low Actio Shrub Tundra and the High Subarctic Forest-Tundra. The location of treeline is approximated from Timoney 8 opens, the tundra component of the vegetation increases and the arboreal component decreases, with forest stands increasingly limited to protected lowlands and valleys. Trees become shorter and there is an increased occurrence of krummholz. Krummholz is defined as a cluster of coniferous tree individuals that exhibit a prostrate or shrubby growth form and are less than 2 m in height (Payette, 1983; Payette et al, 1989). The Forest-Tundra is divided into the Forest Subzone and Shrub Subzone at the physiognomic forest limit, where the tree to tundra surface cover ratio is 1:1. Treeline forms the northern boundary of the Forest-Tundra and is defined as the northern limit of trees with a tree growth form, i.e., greater than 3 to 4 m in height, and where the tree to tundra surface cover ratio is 1:1000. Trees are typically restricted to valleys and slopes with southern exposures (Ritchie, 1987; Timoney, 1988; Sirois, 1992; Timoney et al, 1992). Beyond treeline, individuals of coniferous species occur northward to their species limit, but have a krummholz growth form (Payette, 1983) or are less than 4 m in height. These individuals will be referred to as trees, though they do not have a tree growth form. They typically occur in tree islands, isolated populations composed of krummholz and occasionally some short, erect stems. For this thesis, the tree species limit is defined as the northernmost extent of a particular coniferous or deciduous tree species. 2.2 Circumpolar Forest-Tundra and Treeline 2.2.1 Northern North America The width of the Forest-Tundra ecotone varies greatly across the northern circumpolar area. In the far northwest of North America, i.e., Alaska and the Yukon, the Forest-Tundra is spatially discontinuous. In this region, the complex mountainous topography is the main 10 determinant of the spatial nature of the transition zone. East of the Yukon, the Forest-Tundra is a more continuous, spatially well-defined zone. In the area of the Mackenzie Delta, the Forest-Tundra is quite narrow, spanning approximately 50 km from south to north. Between the Mackenzie Delta and the eastern arm of Great Slave Lake, the Forest-Tundra averages 112 km in width, with minima around Great Bear Lake and the lower Anderson River valley. The ecotone dips strongly southward east of the Anderson River and again east of the Coppermine River. Continuing southeastward, the Forest-Tundra widens substantially between Great Slave Lake and Hudson Bay, averaging 179 km. The greatest widths in this region vary between 228 and 338 km, and are located between the Dubawnt River and southeast Nunavut/northeast Manitoba (Timoney, 1988; Timoney et al, 1992). In northern Quebec and Labrador, the Forest-Tundra is also quite wide, spanning 45 to 355 km (Payette, 1983). In Alaska, P i c e a sitchensis (Bong.) Carr. and A l n u s sinuata (Regel) Rydb. form treeline along the southwest coast. In western and northern areas of Alaska and in the Yukon, P i c e a g l a u c a is the northernmost conifer forming treeline and timberline, though P o p u l u s balsamifera can occur further north, but only in depressional areas, such as river basins (Tuhkanen, 1993). P i c e a m a r i a n a and, to a lesser extent, L a r i x l a r i c i n a are prominent in the northern forest zones of Alaska and the Yukon. Across Canada, east of the Yukon, there are three main conifer species that occur at latitudinal treeline: P i c e a m a r i a n a , P i c e a g l a u c a , and L a r i x l a r i c i n a (Hustich, 1979). In northwest and north-central areas of Canada, P i c e a g l a u c a and P i c e a m a r i a n a form treeline, with P i c e a g l a u c a prominent west of Great Slave Lake and P i c e a m a r i a n a prominent east of this area. L a r i x l a r i c i n a occurs at treeline along the west coast of Hudson Bay (Hustich, 1979; Elliott-Fisk, 1983; Ritchie, 1984; 11 Timoney et al, 1992). In northeastern Canada, Picea glauca dominates treeline in maritime conditions along the east coast of Hudson Bay, along Ungava Bay, and in Labrador. Picea mariana is present in these areas, but dominates treeline in the continental areas. Larix laricina joins these species at treeline in the Lake Minto/Leaf River area of northern Quebec and in Labrador (Elliott, 1979a; Legere and Payette, 1981; Payette, 1983). Beyond treeline, only Picea glauca and Picea mariana occur and these are the northernmost species of conifers in North America (Black and Bliss, 1980; Legere and Payette, 1981; Ritchie, 1984; Larsen, 1989). 2.2.2 The Tuktoyaktuk Region In the Tuktoyaktuk region, treeline occurs to the southern part of Richards Island, dropping southeast to the west end of Husky Lakes (formerly Eskimo Lakes; Figure 2.1). It follows the eastern shoreline of Husky Lakes, approximately 10 km or more inland, to approximately 69°25'N. Treeline is located at approximately 69°20'N in the Anderson River valley and 69°30'N in the Horton River valley (Timoney, 1988). Within the lower Mackenzie River valley, Picea glauca occurs on recent alluvium to approximately 68°50' N (Mackay, 1963). Populus balsamifera may occur with Picea glauca or form pure stands. However, Populus balsamifera extends farther north than Picea and forms the limit of arborescent trees within the Mackenzie Delta (Mackay, 1974; Ritchie, 1984). In upland areas, Picea glauca forests occur, with varying amounts of Larix laricina, Betula papyrifera, Populus tremuloides, and Populus balsamifera, depending on substrate conditions and fire history (Ritchie, 1984). Betula papyrifera - Populus tremuloides stands occur as an early postfire successional communities. Picea mariana is abundant within the 12 lower Mackenzie River valley. It occurs in pure stands with shrubs and mosses in better-drained sites. In poorly drained areas, Picea mariana muskegs prevail with some Larix laricina (Ritchie, 1984). Out of the Mackenzie River valley, Picea mariana becomes less common. Between the Kugaluk and Coppermine Rivers, Picea mariana is rare or absent in the northern Forest-Tundra. Picea glauca forms treeline and the tree species limit, with rare occurrences of Populus balsamifera and Populus tremuloides (Zoltai et al., 1979; Timoney, 1988). Picea glauca and Picea mariana krummholz have been identified on the Tuktoyaktuk Peninsula by Ritchie (1984) and Picea glauca krummholz within the Anderson and Horton River valleys north of treeline by Timoney (1988). 2.2.3 Eurasia The Forest-Tundra in Eurasia broadly follows a similar pattern to that of North America, where the width in the northwest is quite narrow, ranging between 50 and 200 km, widening substantially east of the Ob River, where it spans 750 to 1150 km (Krebs and Barry, 1970). In northern Eurasia, where oceanic conditions prevail, treeline is formed by Betula pubescens Ehrh. ssp. tortuosa (Ledeb.) Nyman, with Populus tremula L., Picea abies (L.) Karst., and Pinus sylvestris L. (Tikhomirov, 1962; Kryuchkov, 1974; Walter, 1984; Kullman, 1990; Tuhkanen, 1993). In mountainous areas of Europe, Larix decidua Miller, Picea abies, and Pinus cembra L. form alpine timberline (Tuhkanen, 1993). Further to the east in northeastern Europe, two closely related species are found at treeline: Picea abies to the west and Picea obovata Ledeb. to the east (Walter, 1984; Tuhkanen, 1993). Within continental areas of northeastern Asia (north-central and northeastern Siberia), Larix dahurica Turcz. (Larix gmelinii (Rupr.) Litv.) forms treeline; Larix sibirica Ledeb. dominates treeline in 13 western Siberia, east of the Ob River (Tikhomirov, 1962; Kryuchkov, 1974; Walters, 1984). In mountainous and northern areas east of the Lena River in Russia, particularly in the extreme northeastern part of Asia, P i n u s p u m i l a (Pall.) Regel. forms timberline and treeline (Tikhomirov, 1962; Kryuchkov, 1974). However, C h o s e n i a macrolepis (Turcz.) Kom., a hardwood, occurs further north (Tuhkanen, 1993). 2.3 Factors Affecting Northern Vegetation There are several factors that have contributed to the regional distribution and characteristics of the Forest-Tundra, the position of treeline, and the arboreal species range limits. These factors fall within three categories: (a) climatic conditions and air masses, (b) topography and soils/surface geology, and (c) historical conditions. The general location of the Forest-Tundra across northern Canada and Eurasia coincides well with the mean July position of the Arctic Front and the mean July 10° isotherm (Bryson, 1966; Larsen, 1974; Sirois, 1992). The Arctic Front represents the boundary between areas dominated by Arctic air and those dominated by southern Maritime air masses (Bryson, 1966; Krebs and Barry, 1970). The Arctic Front is not static but varies quite widely within a season and between years. This variation is quite high in central Canada. In the far northwest of Canada, the Arctic Front is topographically limited at the northern end of the Cordillera and, therefore, varies less in this area. In northern Quebec and Labrador, the general location of the Forest-Tundra does not coincide as well with the mean July position of the Arctic Front. Atlantic, southern Pacific/Gulf of Mexico and, to a lesser extent, Hudson Bay air masses have a rather high frequency of occurrence in this area, hence, there is a great deal of variability in air mass dominance (Bryson, 1966). In Eurasia, the 14 maximum variation in air mass frequency does not coincide with the area of widest Forest-Tundra. This may relate to the mountain topography (the Urals, Central Siberian Uplands, and mountains of Eastern Siberia) complicating the synoptic climate (Krebs and Barry, 1970; Timoney, 1988). The position of the Arctic Front and the frequency of Arctic and Pacific air masses act as a broad, simplified index of the actual complex climatic conditions that affect plants. These climatic conditions, particularly summer temperatures and growing season length, provide the minimum requirements for tree growth. The degree of continentality or oceanity may affect the width of the Forest-Tundra (Larsen, 1974; Sirois, 1992). The proximity of cold oceanic waters may contribute to the narrowing of the Forest-Tundra by producing cooler summer temperatures and a steep climatic gradient as the water body is approached, especially in areas of significant onshore winds. In the northwest, near the Arctic coast and along the coast of Hudson Bay, much steeper vegetation gradients (and a narrow Forest-Tundra) are found (Payette, 1983; Timoney et al, 1992). Burn (1997) found that a steep climatic gradient occurred over the Mackenzie Delta region during the summer, when growing season air temperatures and growing degree-days >5 °C were reduced in coastal areas. The impact of onshore winds on summer conditions in 1994 and 1995 extended inland 50 to 70 km from the Arctic coast. However, proximity to the coast did not appear to affect winter conditions as greatly. In western Eurasia, oceanic effects have been suggested as contributing to narrow ecotones, though in these areas there have been some human-induced modification of the vegetation (Sirois, 1992). The effect of topography on climatic conditions may account for some of the 15 variation in width of the Forest-Tundra in Canada. The north to northeastward rise in elevation in areas between Great Bear Lake and Great Slave Lake produces an important topoclimatic gradient that results in a steep vegetation gradient, with a spatially narrow transition from forest to tundra. A general northward decrease in elevation between Great Slave Lake and eastern Nunavut produces a broad, gentle topoclimatic gradient and a wider area favourable for tree growth. The Forest-Tundra is quite wide in this area (Timoney, 1988; Timoney et al, 1992). A northward decline in elevation is similarly related to the broad Forest-Tundra zone in northern Quebec and Labrador (Hare, 1950; Elliott-Fisk, 1983; Payette, 1983). The complex topography in eastern Eurasia may contribute to the great width of the Forest-Tundra, providing favourable and unfavourable environmental conditions across extended distances (Krebs and Barry, 1970; Larsen, 1974; Timoney, 1988). Other factors may also affect the distribution of Forest-Tundra. In northwestern Canada, the presence of sedimentary rocks and nutrient-rich, fine textured soils are more favourable for tree growth and may support forest farther north. To the east, within the Canadian Shield, felsic crystalline rocks and acidic, nutrient-poor, coarse soils may restrict the sites conducive to tree establishment and growth, especially in continental areas where the position of the Arctic Front is highly variable. Nutrient-poor soils may contribute to the broadening of the Forest-Tundra in northern Quebec, eastern N.W.T., and Nunavut (Timoney, 1988). Permafrost is an important environmental factor. A shallow active layer and low soil temperatures limit root growth. Soil moisture is more difficult to access at low soil temperatures, which can result in stomatal closure in Picea glauca during the growth season regardless of soil moisture levels and evaporative demand (Goldstein et al., 1985). This 16 reduces photosynthetic gains. However, stomatal sensitivity to soil temperature may help prevent winter desiccation by limiting transpiration when the soil is frozen (Goldstein et al, 1985). Permafrost and low soil temperatures result in low rates of nutrient cycling (Van Cleve et al, 1990). Disturbances common in permafrost areas, such as soil churning, frost heave, formation of ice lenses and hummocks, and active layer detachments, can up-root, damage, or kill trees and other vegetation (e.g., Viereck, 1965). Nonetheless, disturbances may also open new areas for colonization (Viereck, 1965; Burn and Friele, 1989) Past treeline fluctuations north and south of present limits may contribute to the breadth of the Forest-Tundra. Favourable climatic conditions in the past have allowed the establishment of trees, and treeline advance, north of current limits. The northern limit of trees and treeline respond slowly to subsequent climatic deterioration because of vegetative inertia. Once established, tree longevity, and the physical (erect versus krummholz growth forms) and reproductive (vegetative versus sexual reproductive modes) plasticity of northern trees promote persistence of adult trees beyond where they can currently establish. Vegetative inertia may also be important in the slow or negligible response of krummholz to recent climatic warming. It has been suggested that low krummholz produce few or no viable seeds during both unfavourable and favourable climatic periods (Payette et al, 1985, 1989; Lavoie and Payette, 1992,1994). During periods of climatic amelioration, an external seed source would be necessary for seedling establishment to occur in areas of low krummholz, resulting in a delayed response in establishment (Payette et al, 1985, 1989; Lavoie and Payette, 1992, 1994; Szeicz and MacDonald, 1995a). These factors make it difficult to determine whether there is an equilibrium between treeline and current climatic conditions. Many researchers suggest that equilibrium with current climate necessitates 17 recent successful sexual reproduction by local trees as indicated by the presence of seedlings at a site (e.g., Elliott-Fisk, 1983). At a local scale, establishment and survival of seedlings and trees are influenced by microclimate (Wardle, 1974). Once established, the presence of an individual tree moderates the microclimatic conditions over a wide area, affecting snow depths, wind velocities, and soil temperatures. This promotes survival and further establishment either from seed or by layering (Wardle, 1993). Groups of trees and krummholz generate their own internal microclimate, more favourable for survival of trees and seedlings (Tranquillini, 1979; Hadley and Smith, 1987). Historical events, particularly fire, that have occurred within differing climatic regimes, can alter the position of treeline and the structure of the Forest-Tundra. Evidence is available from eastern Canada that the timing of a fire event in conjunction with pre-fire tree stature (which reflects seed availability), and pre- and post-fire climatic conditions have resulted in differential post-fire regeneration within the Forest-Tundra through the late Holocene (Payette and Gagnon, 1979,1985; Payette and Filion, 1985). Historical events and climate change add significant complexity to the determination of the physical factors influencing treeline and tree species limits. History becomes a complicating factor because conditions that limit survival of adult trees and vegetative reproduction differ from the conditions that limit viable seed production, germination, and survival of seedlings and saplings. 2.4 Holocene Vegetation History of Northwestern Canada Reconstructions of postglacial vegetation in the Tuktoyaktuk and Inuvik regions have been 18 presented by Spear (1983) and Ritchie (1984). The reconstructions were based on pollen influx or R r e l corrected percentage records from lake sediments, peat, and pingo deposits, stratigraphically controlled by radiocarbon dating. Pollen data were augmented using radiocarbon dated macrofossils. Based on these reconstructions, at approximately 13000 BP, following deglaciation, the Tuktoyaktuk Peninsula was colonized by predominantly shrub tundra assemblages. B e t u l a g l a n d u l o s a Michx., Shepherdia canadensis (L.) Nutt. and Salix L . pollen levels were high, along with Cyperaceae, and Gramineae (Figure 2.3; Ritchie, 1984).. Birch shrub tundra would have occurred in upland areas, open sedge-herb-grass communities in well-drained sites, and grass-sedge marshes in the wettest sites. After 11000 BP, minor levels of P o p u l u s pollen were recorded, suggesting that P o p u l u s balsamifera and P o p u l u s tremuloides may have been present. At approximately 10500 BP, P i c e a pollen levels began to increase, indicating the arrival of spruce. P i c e a needle and seed macrofossils occur frequently in the sediments of Sleet Lake from 9800 to 5900 BP (Ritchie, 1984). Differentiation of P i c e a g l a u c a from P i c e a m a r i a n a pollen indicates that P i c e a g l a u c a arrived in the area first and dominated the forests; P i c e a m a r i a n a became abundant by approximately 8000 BP (Spear, 1983). From studies in the Inuvik area, P i c e a g l a u c a arrived and attained their maximum population levels prior to P i c e a m a r i a n a (McLeod and MacDonald, 1997). As P i c e a pollen levels increased, so did levels of the ericales and Juniperus L . , and there were decreases in pollen levels of B e t u l a , Shepherdia canadensis, P o p u l u s , and Salix. Shrub tundra was replaced by a P i c e a dominated forest-tundra. Between approximately 9000 (possibly as early as 10000 BP) and 5000 BP, Pzcea-dominated boreal forest conditions prevailed, with the spread of 19 Figure 2.3 Summary percentage pollen diagram for a southern Tuktoyaktuk Peninsula lake site. The pollen levels are based on corrected pollen totals using R values (from Ritchie, 1984). 20 L a r i x l a r i c i n a , B e t u l a papyrifera, and A l n u s B. Ehrh. into the region. The fossil pollen data indicative of forest conditions were augmented by the recovery and dating of P i c e a macrofossils. Radiocarbon dates from stumps, twigs, and cones indicated that P i c e a trees occurred in the region between approximately 6600 and 5000 BP (Figure 2.4). The diameters and ring widths of the P i c e a stumps were similar to the tree diameters and ring widths of P i c e a g l a u c a from mesic sites near Inuvik, again. suggesting boreal forest conditions (Ritchie, 1984). From the data available, P i c e a spread as far northwest as the current coastline southwest of the town of Tuktoyaktuk (Figure 2.1; Mackay, 1992), and at least as far north along Tuktoyaktuk Peninsula as approximately 69°50' N (Ritchie, 1984). P i c e a pollen levels decreased from their mid-Holocene high (between 9000 and 6000 BP, depending on the site) and reached modern levels by approximately 3500 BP. The northern limit of boreal forest and treeline retreated as mortality of coniferous trees increased and forest stands thinned in response to climatic cooling. Between 5000 and 4000 BP, Alnus - B e t u l a - P i c e a forest-tundra prevailed in the region. The current positions of the northern Boreal Forest and Forest-Tundra were attained by approximately 4000 to 3500 BP. It is suggested that since this time, modern dwarf shrub tundra has dominated the region with only scattered P i c e a krummholz. There have been no substantive changes in the pollen record that may reflect treeline or species limit responses to either the Little Climatic Optimum (ca.. A D 1000 to 1350; 950 to 600 BP) or the Little Ice Age (ca.. A D 1400 to 1850; 550 to 100 BP). 21 Figure 2.4 Radiocarbon dates ( 1 4C BP) of wood reported in Spear (1983), Ritchie (1984), Mackay (1992), and from this study (bold dates). CHAPTER 3 PHYSICAL ENVIRONMENT OF THE TUKTOYAKTUK REGION 22 3.1 Introduction Species range limits are strongly influenced by physical environment conditions that, in part, control the potential distribution of a species. This potential range is then modified by the influence of organisms (including site modification) and historical conditions, creating the actual species distribution. To better understand the distribution of tree islands in the region, a review of the physical environment is required. In this chapter, the physical setting of the study region is described within the major areas of geology, climate, permafrost, and soils. 3.2 Physiography The physiography of this area has been described by Mackay (1963), Bostock (1970), and Rampton (1988a). The general physiographic regions are the Mackenzie Delta, Tuktoyaktuk Coastlands, and the Anderson Plain (Figure 3.1). Bostock (1970) mapped the Mackenzie Delta and Tuktoyaktuk Coastlands as a single region (Arctic Coastal Plain). Rampton (1988a) defined a Mackenzie Delta division to represent the active Holocene delta and to differentiate it from the Quaternary sediments of the Coastlands. The Mackenzie Delta is an estuarine delta 13000 km 2 in area. The delta is characterized by highly sinuous distributary channels (anastomosing) and numerous shallow lakes. Flooding is frequent and often related to ice jamming during breakup. The local relief is minimal with only 10 m of elevation loss between the delta apex and northern edge. The Tuktoyaktuk Coastlands (Pleistocene Coastlands of Mackay, 1963) are 23 130° Figure 3.1 Physiographic regions of the study area (after Mackay, 1963; Rampton, 1988a). Boundaries are indicated by thick black lines. 24 characterized by low elevations and little local relief. Rampton (1988a) and Mackay (1963) subdivided the Coastlands into several landscape units. In general, the terrain is variable with coastal plains, glacial outwash plains, tunnel valleys, lake plains, rolling hills, eolian dunes and veneers. Extensive glacial outwash deposits occur through the central and northern portions of the peninsula (Rampton, 1988a). These deposits are related to glacial meltwaters that were routed through a series of tunnel valleys in the present day Husky Lakes. The tunnel valleys are linear, deeply scoured features formed by subglacial meltwater. A belt of ice contact features (kames, eskers, and kettles) runs parallel to the north shore of Husky Lakes in the southern and central portions of the peninsula. Areas of glacial moraine are typically rolling hills with low relief. There are also numerous thermokarst lakes, evidence of fluctuating lake levels and past thermokarst activity, consequently lacustrine deposits are extensive. Closed system pingos are widely distributed. Ground ice features include polygonal ground and involuted hills. Thaw slumping in ground ice rich areas is common along shorelines. Eolian deposits are present as sandy dunes and veneers. Across the area, thick sequences of Quaternary sediments overlie sedimentary bedrock of Tertiary age, and there are few outcrops. The Anderson Plain is an undulating area of low relief that runs parallel to the Coastlands. The cover of Quaternary materials is locally thick but there are extensive areas of outcrop and bedrock control of topography. The area is dissected by a series of broad river valleys. Slopes are locally steep but are mainly gentle. Relief is most pronounced in the eastern portion of the area in the Anderson River Uplands, and in the Mallach Hil l Uplands and Smoking Hills above the Horton River. Along the Anderson River, the floodplain is extensive and the valley sides are 25 moderate to gently sloping. The river has cut into poorly consolidated shales and other fine elastics. The area has been ice covered, although the timing is not certain. In contrast, the valley of the lower Horton River shows more relief. Steep slopes in shale occur in the Smoking Hills region. The surrounding upland is generally flat and well drained. Mapped ice limits suggest the lower Horton River was ice free in the Wisconsinan and possibly for the whole of the Pleistocene (Mathews et al, 1989). 3.3 Bedrock Geology There has been extensive work on the bedrock geology of the region, due in part to the economic interest in petroleum. Bedrock geology has been mapped by Young et al (1976) and Yorath et al. (1980). In addition, a series of Geological Survey of Canada papers, memoirs, and bulletins deal with specific areas {e.g., Yorath and Cook, 1981). The bedrock geology is illustrated in Figure 3.2. Essentially, the bedrock is clastic sedimentary. Sandstones and conglomerates of fluvio-deltaic origin underlie Quaternary deposits along the Tuktoyaktuk Peninsula. Fine-grained siltstones and shales are found through the remainder of the region. 3.4 Quaternary Record The Tuktoyaktuk Coastlands and Anderson Plain have been influenced by the Laurentide Ice Sheet in the Pleistocene Epoch. The stratigraphic record is complex and controversy surrounds the timing of some glacial events. The first extensive surveys of Quaternary sediments and landforms in the region were conducted by Mackay (1958, 1963). Regional glacial limits and chronologies have been described by Hughes (1972, 1987), Rampton 26 LEGEND 130' [M3 Tertiary/Quaternary Clastic sediments IlTertiary (Miocene) Conglomerate, Sandstone • Upper Cretaceous Shale, Oil Shale, Siltstone Figure 3.2 Generalized bedrock geology in the study region. Mapping based on Young etal. (1976), Yorath etal. (1980) and Okulitch (1991). 27 (1988a), and Vincent (1989). In the Wisconsinan Glaciation, Laurentide ice advanced from Great Bear and Great Slave Lakes. The Mackenzie Lobe moved north to the Beaufort Sea. In the Low Arctic, ice lobes flowed through Amundsen Gulf and M'Clure Strait to Banks Island. The approximate Wisconsinan Glaciation limits of the Laurentide Ice Sheet in northwestern Canada are shown on Figure 3.3. These limits suggest much of the Tuktoyaktuk Peninsula was ice-free throughout the Wisconsinan Glaciation. 3.4.1 Glacial Chronology, Banks Island and Tuktoyaktuk Coast Evidence of several glaciations is preserved in sediments in the western Arctic. The Toker Point/Franklin Bay Stadial was initially assigned to the Early Wisconsinan Glaciation (Rampton, 1988a). At its height, glacial ice was absent from much of Tuktoyaktuk Peninsula (Figure 3.3). The Toker Point limit runs east to west across the central part of the Tuktoyaktuk Peninsula. Rampton (1988a) mapped Toker Point glacial sediments across the lower half of the Tuktoyaktuk Peninsula. In places, these sediments overlie the sandy Kittigazuit Formation. Recent work suggests the Kittigazuit Formation is a Mid-Wisconsinan eolian deposit laid down between 33000 and 37000 BP during a period of lower relative sea level (Dallimore et al, 1997). Thus, the overlying glacial sediments must post date this time interval. This interpretation suggests glacial ice was more extensive in the late Middle Wisconsinan or early Late Wisconsinan than generally accepted and that much, perhaps all, of the Tuktoyaktuk Peninsula was ice covered. The Sitidgi Stadial is the Late Wisconsinan Glaciation advance. During this stadial, Laurentide ice advanced from Great Bear Lake to the Mackenzie Delta but did not move onto the Tuktoyaktuk Peninsula (Rampton, 1988a). Two distinct limits are associated with phases 28 Figure 3.3 Wisconsinan ice limits in northwestern Canada (after Hughes, 1987; Vincent, 1989). 29 of the Sitidgi Stadial: the Tutsieta and Kelly Lake Phases (Hughes, 1987). Moraines from the Tutsieta Phase can be traced from Sitidgi Lake eastward onto the Anderson Plain and are assigned an age of 13000 BP (Hughes, 1987). 3.5 Climate The climate of the area has been described by Burns (1973) and reviewed by Ritchie (1984). Climate regions have been defined by the Ecoregions Working Group (1989) and Gullett and Skinner (1992). In general, the Tuktoyaktuk Peninsula and the northern portions of the Anderson Plain are cooler and drier than the Subarctic areas to the south, particularly in the summer months. The former areas are influenced by cold and dry Arctic air masses for much of the year. In the Subarctic, summer temperature and precipitation values are higher due in part to incursions of Maritime Polar and Pacific air masses associated with the passage of cyclones along the Arctic Front. • Gullett and Skinner (1992) differentiate between the climate of the western Northwest Territories and that of the Arctic Tundra. Ritchie (1984) separates Arctic Coastal climates from Continental Subarctic climates, largely using treeline as the border. The Ecoregions Working Group (1989) also places a division approximately along treeline, with the Low Arctic Ecoclimatic region to the north and the High Subarctic Ecoclimatic region to the south. Although there are relatively few long term climate records from stations operating in the study area, there are sufficient data available to highlight the differences between these regions. Figure 3.4 shows the distribution of climate stations from which monthly normals are available and the boundary between the Low Arctic and High Subarctic Ecoclimatic 30 Regions. Monthly temperature and precipitation normals are plotted for Norman Wells, Inuvik, Tuktoyaktuk, and Sachs Harbour on Figure 3.5. 3.5.1 Low Arctic Ecoclimatic Region (Arctic Coastal Climate) The study area is in the Low Arctic Ecoclimatic Region (Ecoregions Working Group, 1989). In general, summers are short, cool, and moist (high relative humidity). The frost free period is typically 25 to 50 days. The winter season is long and intensely cold with mean daily temperatures below 0°C for approximately 38 to 40 weeks. Ice break up on small lakes may not occur until early July. Total annual precipitation in the western portion of the region is less than 200 mm. A transect of climate stations within this region from the western Beaufort Sea eastward into Amundsen Gulf includes: Komakuk Beach, Shingle Point, Tuktoyaktuk, Nicholson Peninsula, Sachs Harbour, Cape Parry, Clinton Point, and Cape Young (Figure 3.4). A l l stations are at low elevation and, with exception of Sachs Harbour, are close to the 69th parallel. Monthly and annual climate data are presented for these stations in Table 3.1. The data show the mean annual temperature ranges between -10.5 (Tuktoyaktuk) to -13.7 °C (Sachs Harbour), with summer (JJA) mean temperatures for most stations in the range 4.1 to 5.9 °C. Tuktoyaktuk and Shingle Point show markedly higher summer temperatures (8.5 and 8.3°C) and more than double the number of growing degree days above 5°C than most of the other stations, largely due to their proximity to the Mackenzie Delta and warm southerly winds from the Mackenzie River valley. Total annual precipitation averages 160 mm, with Shingle Point the only station to exceed 200 mm. The percentage of the annual precipitation that falls as snowfall averages 48% and summer precipitation comprises 40 to Figure 3.4 Locations of climate stations in northwest Canada. Boundaries are shown separating the major ecoclimatic regions. 32 • • M e a n M a x i m u m Month ly Tempera tu re (°C) M e a n Month ly Tempera tu re (°C) Prec ip i ta t ion Month ly Ra in fa l l (mm) * * M e a n M in imum Month ly Tempera tu re (°C) J Month ly Snowfa l l (cm) 20 - i a) S a c h s Harbour A O o 2 =3 -*—' CL E CD 0 -20 H -40 20 - i r 60 20 - i b) Tuktoyaktuk 40 0 20 -20 H J F M A M J J A S O N D c) Inuvik A O o 13 -*—' 03 t_ CD O . E CD 0 -20 H -40 0 -40 r- 60 20 - i J F M A M J J A S O N D d) Norman Wel ls A , > 40 0 h 20 -20 H 0 -40 J F M A M J J A S O N D Month J F M A M J J A S O N D Month Figure 3.5: Cl imate normals from: a) Sachs Harbour A , b) Tuktoyaktuk, c) Inuvik A , and d) Norman Wel ls A (data from Environment C a n a d a , 2000). 33 50% of the annual total. At most stations, snow cover is maximum at the end of March, with values ranging from 17 to 32 cm. 3.5.2 High Subarctic Ecoclimatic Region (Continental Subarctic Climate) Much of the Mackenzie Delta, including Inuvik and the valleys of the Anderson and Horton Rivers to the east, are in the High Subarctic Ecoclimatic Region (Ecoregions Working Group, 1989). Summers are cool and moist with the frost free period extending 50 to 90 days (Ritchie, 1984). Winters are long and very cold with mean daily temperatures below 0°C for approximately 36 to 38 weeks. Total annual precipitation is in the range 250 to 350 mm, with much of the precipitation occurring in summer and early fall. Monthly normals for Inuvik and Norman Wells are shown on Figure 3.5. Data from Inuvik and Norman Wells are also presented in Table 3.1. The data show Inuvik and Tuktoyaktuk vary markedly in mean summer temperatures (11.6 vs 8.5°C) and degree days (682 vs 410). In addition, Inuvik receives more precipitation (257 vs 142 mm) and has a much deeper March snow cover (60 vs 32 cm). The transition from Shrub Tundra at Tuktoyaktuk to Forest-Tundra at Inuvik reflects the climate gradient. Norman Wells is in the Low Subarctic Ecoclimatic Region and the data provide comparison with a Boreal Forest site. 3.5.3 Homogeneous Time Series of Temperature and Precipitation Homogeneous climate data have been adjusted to remove non-climatic variations caused by changes in station locations, site characteristics, observers, observation programs, and instrumentation. Homogeneous time series are designed for climate change trend analyses. cd CD cS <D cn O <D •< 13 2 , 2 5 2 § •§ iS Q o o «s o o 2d U J) u 3 ^ a cud CN CU H I ? S3 •a 0 0 ° Q m S > *H o «! Q ^ < & 2 -e o > • c o o o c o CO —I S O - H — S O CN i n SO o CN OC i n OS - H O SO - H cN ro S O CN ro OO so m CN 1/1 O oo © — i so CN i n 0 0 CO C CO O i n oo i n "i oo ON OS O S o - H so oo <N - H o - H "d- so SO CN so m o so o CN r o OS SO o oo oo CN •si-r s -H. OS (N CN "I-H - H ~H (N i n f3 r o CN B "-H —< OS i n oo o i n i n oo r- ^ H O ^ ^ ^ ^ ^ ^ ^ ^ r o O - H O o r o r o r o oo i n o oo CN SO —< — -H-i n CN CN 0 0 o O CN m i n o SO z z z z z z Z z m f - so © © i n SO ro i n CN i n o ro i n o o o o o o o o OS oo OS OS CN O OS oo so SO SO so f - SO so o co ID CQ CO 6 o g 'o CM 6 0 g z 3 ca o 2 3 H z a u PH B o w "3 o z 3 o • D U l CO a co JS o CO GO i ^ z z ^' .1 O OH CO PH U B o U t/5 CN © o o S O m i n m CN i n OS OS i n CN oo CN SO SO CN OS OO so so o S O O S so OS CN o ro r o oo S O m so 34 'cO > co O d cd a 35 Three homogeneous databases are used in this thesis. The Historical Canadian Climate Database (HCCD) includes temperature and precipitation data from a network of 131 stations across Canada. In the HCCD, temperature records are homogeneity adjusted, stations merged to increase record lengths, and missing data estimated. Precipitation data have not been adjusted nor missing values estimated. The HCCD is used to track annual and seasonal temperature and precipitation trends in climate regions across Canada. Results are reported in the Climate Trends and Variations Bulletin as departures from the 1951-80 mean in % for precipitation and °C for temperature. (Whitewood, 2000). Mean annual temperature and total annual precipitation departures are shown for the Arctic Tundra and Mackenzie Valley climate regions of Gullett and Skinner (1992) for the period 1948 to 1999 on Figure 3.6. The mean annual temperature records are similar. Both regions show slightly warmer than average conditions in the early 1950s, a cooling to the late 1970s and subsequent warming. In the Arctic Tundra, the years 1999, 1998, and 1981 are 2.0°C or more above the mean, while in the Mackenzie Valley the warmest years are 1999, 1998,1993,1987, and 1981. Total annual precipitation shows relatively wet periods in the Mackenzie Valley during the late 1950s to early 1960s, and during the late 1980s. Data from the Arctic Tundra show an overall positive trend with a gradual increase in precipitation throughout the record. 3.6 Permafrost Conditions and Periglacial Landforms The Tuktoyaktuk region is located within the zone of continuous permafrost (Brown, 1970). Permafrost reaches a maximum thickness of 750 m under the eastern portion of Richards 36 37 Island (Taylor et al, 1996) and generally exceeds 300 to 400 m for much of the Tuktoyaktuk Coastlands (Taylor et al, 1996). Taliks are present under some lakes (Mackay, 1990a). To the west, beneath the active Mackenzie Delta, the permafrost thickness averages less than 100 m. In the Coastlands, permafrost tends to be ice-rich, with massive ground ice and icy sediments underlying many hills ("involuted hills") and ridges of the Tuktoyaktuk Coastlands (Rampton, 1988b; Mackay and Dallimore, 1992). The massive ground ice may have an epigenetic segregation origin with glacial meltwater as the water source (Rampton, 1988b). Within the region, there are numerous surface features that result from periglacial processes operating in the presence of ice rich permafrost. There are numerous closed system pingos. In the area of the Tuktoyaktuk Peninsula and the south shores of Husky Lakes and Liverpool Bay, there are an estimated 1450 pingos (Mackay, 1990a, 1994). Large ice wedge polygons are present in the region within low lying, poorly drained areas with fine-grained sediments, but ice wedges also occur on slopes (Mackay, 1990b). Typically, ice wedge polygons are outlined by A l n u s or Salix and B e t u l a g l a n d u l o s a shrubs that grow on the ridges (Mackay, 1958; Ritchie, 1984). Both high-centred and low-centred ice wedge polygons can be found in the region. In the former, the central ground is raised with relatively dry peat and a heath-lichen vegetation cover. In the latter, the central area is lower than the ridge and is considerably wetter with Carex and E r i o p h o r u m vegetation cover and, frequently, standing water (Ritchie, 1984; Mackay, 1992). In the study region, ice wedge polygons dominate the flat, low-lying areas of the Tuktoyaktuk Peninsula, and the active floodplain and terraces of the Anderson and Horton River valleys. The thawing of ice-rich permafrost and the resulting subsidence and collapse have 38 generated thermokarst topography across much of the Tuktoyaktuk Peninsula. Thermokarst lakes develop as depressions fill with water. These lakes are shallow, often less than 5 m deep (Mackay, 1985). In the northern Tuktoyaktuk Peninsula, thermokarst lakes tend to be oriented south to north, primarily due to wind effects promoting further melting on the downwind shoreline (Williams and Smith, 1989; Mackay, 1992). Active layer detachments (slumping) are common on the slopes of the study region, particularly lakeshore, river bank, and coastal slopes. Both slump scar and slump deposits become new substrate for the invasion of plants, including tree seedlings (Burn and Friele, 1989). Earth hummocks are prominent in the study.region resulting from cryoturbation and from the presence of certain plants, such as Eriophorum (Williams and Smith, 1989). Earth hummocks dominate flat-lying areas with frost-prone, fine-grained sediments. Mudboils and palsas are also found in the region. 3.7 Soils Three orders of soils are found within the Tuktoyaktuk region: Organic, Cryosols, and Regosols (Ritchie, 1984; Agriculture Canada Expert Committee on Soil Survey, 1987). Organic soils, characterized by their deep accumulation of organic material (peat), are found in poorly drained low-lying areas. These are prevalent in coastal lowlands. Regosols are found in coarser parent materials, in well-drained topographically raised and upland areas, with sparser plant cover and little moss development. Typically they are weakly developed soils. Cryosols, occurring in fine-grained and organic materials, are prominent in permafrost areas, such as the study region, and are commonly cryoturbated. CHAPTER 4 39 DISTRIBUTION AND DESCRIPTION OF TREE ISLANDS 4.1 Introduction Very little has been reported concerning tree islands in northwest Canada. Published reports of tree islands within the Tuktoyaktuk region include: Mackay (1992) at a coastal site near the abandoned town of Kittigazuit, southwest of the town of Tuktoyaktuk (Figure 4.1); Mackay (1958) just inland from the coast east of the mouth of the Anderson River, near the abandoned town of Stanton; Spear (1983) on the east-central part of the peninsula; Ritchie (1984) on the southern part of Tuktoyaktuk Peninsula; and Timoney (1988) along the lower Anderson and Horton River valleys. In these publications, tree islands were not the primary subject and little information is presented. Therefore, the major objective of this chapter is to document and describe tree islands present in the Tuktoyaktuk region. The tree island and site descriptors presented in this chapter are: site location, number of tree islands, areal size, position within the local topography, species composition, number of live and dead individuals, tree heights, growth forms, tree distribution and densities, soils, active layer depths, and non-arboreal vegetation. 4.2 Methods The initial search for the northernmost tree islands on Tuktoyaktuk Peninsula, and in the Anderson and Horton River valleys was conducted by Drs. Greg Henry and Bi l l Freedman in the summers of 1990 and 1991. I conducted additional searches in the summers of 1993, 1994, and 1995. These searches were made by helicopter survey of areas with potentially Figure 4.1 The location of tree island sites in the Tuktoyaktuk region. 41 favourable microsites, generally south-facing slopes. Each site that was located was given an identifying label consisting of an upper case letter and a number. The upper case letter identified the area in which the site was found (R = Richards Island, T = Tuktoyaktuk Peninsula, A = Anderson River valley, H = Horton River valley) and the number identified the site number in that area (e.g., T l and T2 refer to sites 1 and 2 along the Tuktoyaktuk Peninsula). At sites with multiple tree islands, individual tree islands were identified by lower case letters (e.g.,, A l a and A l b were two tree islands at site 1 within the Anderson River valley). Reports of the occurrence of P i c e a from people familiar with the region established the location of seven sites: sites T l and T2 located by Dr. J. Ross Mackay, site T4 by Dr. Ray Spear, site T6 by Mr. Claude Brunet, site A l by Dr. Terry Armstrong, site R l by Dr. Chris Burn, and site R2 by Dr. Patricia Sutherland (Figure 4.1). Once located, a reconnaissance of the site was made to establish the number of tree islands. The site boundaries were delimited by natural obstructions (hills, large creeks or rivers) that prevented further search. Sketch maps of each site area were made, noting the topography and the relative positions of all tree islands. At sites A2 and A3, where numerous tree islands occurred, the number of tree islands was estimated during several site searches and by taking counts using binoculars. The estimated distance beyond treeline ("north" of treeline) was determined using r reported treeline position (Timoney, 1988). For the Anderson and Horton River valley sites, distance from the tree islands to treeline was measured along the valley. For sites on Tuktoyaktuk Peninsula and Richards Island, the measurement was the shortest distance to treeline in any direction. Identification of the trees as P i c e a g l a u c a (Moench) Voss was based on several 42 morphological characteristics (Table 4.1; Gordon, 1952; Little and Pauley, 1958; Parker and McLauchlan, 1978). In addition, pollen grains of Picea glauca can be differentiated from Picea mariana based on grain size (average 100 urn versus 84 um, respectively) and four morphological characteristics described by Hansen and Engstrom (1985): (1) attachment of the two bladders (sacci) to the main body (corpus) of the grain, (2) shape of the bladders, (3) density and arrangement of the internal reticulate structure in the bladders, and (4) size of the bladders compared to the main body. Table 4.1 The morphological characteristics used to distinguish Picea glauca from Picea mariana. Characteristic Picea mariana Picea glauca twig surface1'2 pubescent glabrous twig colour 1 , 2 dark red-brown light yellow-brown twig ridge shape1'2 externally flattened externally rounded twig pulvini 1 ' 2 flattened raised and rounded needle odour2'3 resinous skunk-like cone location2'3 throughout crown top of crown immature cone colour 2 ' 3 dark purple yellow-green cone scales2'3 not resinous, shiny surface resinous, dull surface cone margin 2 ' 3 dendriculate almost entire 1 Gordon (1952) 2 Little and Pauley (1958) 3 Parker and McLauchlan (1978) Tree island position in the landscape was characterised by topographic location and aspect. Areal size of the tree islands was estimated by measuring the greatest length and 43 multiplying by the width perpendicular to that direction. The tallest and shortest erect stems were measured to determine the range of heights in a tree island and the growth forms present were described based on the terminology of Lavoie and Payette (1992, Figure 4.2). A sixth growth form type was added to this list in order to describe the full range of forms present in the Tuktoyaktuk region: infranival erect, which broadly resembles miniature trees. At sites with only one tree island present, more detailed measurements were made of that tree island. At sites with multiple tree islands, detailed measurements were made of the first tree islands located: A l l individuals in each selected tree island were enumerated. Using a tape measure and compass, a base line was established approximately along the long axis of the tree island. The position of each live or dead individual was mapped by measuring, to the nearest 0.05 m, from this baseline. For each individual, the height of the tallest leader was measured to the nearest 0.005m and the growth form was described. Ground vegetation and active layer depths were quantified along transects. At sites T l , T3, and T5, two intersecting transects were established that spanned the tree island length and width (approximately north/south and east/west) and that extended 10 m outside the tree island, in each of the four major directions. For sites T6i, A l a , A2a, A3c, and Hla , one transect was sampled following the long axis of the tree island and extending 10 m outside the tree island at each end. Vegetation was characterized by percent cover of vascular species in 1 m x 1 m quadrats along the transects. Species identification was based on Porsild and Cody (1980). Active layer thickness was measured in the centre of each quadrat using a steel rod. Soil pits were dug down to the parent material or to the permafrost table, both inside and outside the tree islands. Soils were described by texture, colour, and tentatively identified by soil order (Agriculture Canada Expert Committee on Soil Survey, 44 1) MAT 2) INFRA NIVAL 3) SUPRA NIVAL 4) WHORLED S) TREE CUSHION SKIRTED Figure 4.2 Growth forms in Picea. Dashed line indicates the snow-air interface (from Lavoie and Payette, 1992). 45 1987). 4.3 Tree Islands in the Tuktoyaktuk Region Fifteen tree island sites were located within the Low Arctic Shrub Tundra (Figure 4.1). Two sites were located on the northern part of Richards Island ( R l , R2); eight sites along the southern portion of Tuktoyaktuk Peninsula (Tl to T6, T l 1, T12); three sites within the lower Anderson River valley ( A l , A2, A3); and two sites within the lower Horton River valley (HI, H3). The northernmost tree island was at site A l at 69°42*N 128°59'W. This site was located approximately 43 km north of treeline. However, site R l was located the greatest distance north of treeline (53 km north). Along Tuktoyaktuk Peninsula, the northernmost site, T l , was 49 km beyond treeline. Along the Horton River valley, the northernmost known tree island occurred at site H3 at 69°40'N 126°59'W, 19 km north of treeline. However, further search northward in this area was curtailed by limited helicopter time. Nine of the fifteen tree island sites had only one tree island present (Table 4.2; Figure 4.3). At the remaining six sites, multiple tree islands were present (Figure 4.4). Tree island density at these sites varied. At site T6, there were 16 tree islands in an area of approximately 0.8 km , resulting in a density of 20 tree islands/km . Along the Anderson River Valley, tree island density at site A l was 10 tree islands/km , at site A2 it was 320 tree islands/km2, and at site A3 it was 100 tree islands/km2. The total number of individuals per tree island was found to vary from four to 181. The tree islands ranged in size from 4 m 2 to greater than 1200 m 2 . Growth forms identified were mat, infranival cushion, infranival erect, erect whorled, and supranival skirted. However, not all of these growth forms were present at all sites and there was a great range s o •3 3 o U cj o 3 on PH 8 ca p "35 O i - J o s * 3 * * & 3 Ui U O u , Q tS S a) ^ a 5P bO • 3 ca PH cd CA JL H" ca C J . C J Z j H * IS S o o CN V u i <D O 3 C J > ca CU 43 "3 o OH o C S JO CN © i n CN co so so u ca 6 in o ia* o a 40 3 co 3 CB c o T3 u g '3 -3 •a a) g '3 -3 cj 43 <- 3 8 co O o f X. -a c o 3 ft o • — I co u a a a u 3 3 i n CN o S O o i n © V CN a o o ca co JN o o i n ca o a o a 3 o <u OH o C J 3 o o CN • co © © -c j -© S O _ ^ O S © CO CO + +^  +^  +^  1 1 CN co >Hr m H H H H H 43 cj HH M -3 3 V >< 60 42 C 60 <u 3 3 ca => la •O ti <U CO .a -s rri r? C O tU O ^ S2 — o S « a l * a v 3 I s.a •° 42 -g M O " co c o 3 3 co co 3=" £ ej ej .3 3 O o" 3 3 a -3 C O in co co >n 2 2 © © CO w C O o <r> co 43 o ca S O O S (N + ^ H ^ H ^ H ^ S O H ca ca 3 Ui *H -3 H3 C J ca ca ~s "a 3 o <D O u _g o" 3 ca 43 -•—» 3 O o o 3 CO CO is 1 .a o-^ a ca 43 3 CO -a 3 C 3 CO T 3 U g '3 -3 •s 3 o o o CU 3 o 3 C8 CN <^ © © C N ca ca "a ~a 42 3 co >, 3 ca co T3 o g _o "3 43 C M O U C M o o 3 a a CO m in m >n ^ H H H d d m CN CN ca © © © © /—\ /-i ^H i - H a ca ro + - H ^ H CN (+28) (+10) (+53) (+45) CN Til T12 Rl R2 C ^ o a ca u 43 3 O u ( D c. a. _o "co CO o '3 & _o o o in C O ' c ? 43, & u c j a o _3 O S © © —'< I I co -sr © © ^ / N ca 43 © © m in ca 43 a © ca o a o a •3 3 o 0> a. C J o o o. &, • 3 0 0 in o _ in 'c? 43 CO c o 3 3 CO c o Z St C J C J CO r, 3 C J C J CO 3 3 ^ "^43, cj" o" w 3 3 a e CN co • © C^  w _: © ^ • ro © '—^  m CN 00 CN ©• CN S O 00 w 00 S O in 00 t (b)9 43 —•• 'ca' (b)9 S O C O + © S O a CN < © CN V U! C J >, ca o a o a u 8 O 8 C J C J C J O C J o ca ca ca - . fc! fc TJ5 « «j B & o +J P i cd c c c C J <LT _e o 8 t-- m in - rsi H-' O O O • * - » • « - » • » - » m m m © © © . „ CN CN « 43 5 £ , O CN +^  © © CO a CO < C J in cj O a C J cj M J H 43 S 0 O co c o lis-ca o © C J cj" a o 8 S O in CN m © d 1? CN 00 'a ro w ca "a 0 0 CN - 4 00 CN CO in + cj 3 O 3 ca tl C J ta C J O S + ro 47 Figure 4.3 Tree island T5, located along the Tuktoyaktuk Peninsula (see Figure 4.1). This tree island is the only tree island in the area and is completely surrounded by shrub tundra. Figure 4.4 Site A2, located within the Anderson River valley (see Figure 4.1). There are more than 80 tree islands in the immediate area. 48 in tree island physiognomy. At some sites, it was difficult to identify the presence of a tree island from the surrounding vegetation because of the presence of tall shrubs or the low stature of the tree island. For example, at site T4, only mat and infranival erect forms occurred (Figure 4.5). In contrast, at other sites, such as A3c (Figure 4.6), many individuals were supranival skirted and erect whorled, and rose prominently above the shrub tundra vegetation. Most of the tree islands were located on slopes, though they also occurred oh hilltops, flat river terraces, and valley bottoms. Typically the tree islands on slopes had a southerly aspect, but they were also found on east, west, and north facing slopes within protected valleys (Figure 4.7). The distribution of individuals within a tree island resulted in either elongated or broadly circular tree island shapes. Of twelve tree islands in which all individuals were mapped, six were broadly circular, six were elongated. At site A2, 79 tree islands were located along a southwest-facing slope, of which, 63% were broadly circular in shape, 14% were elongated along the slope (downslope/upslope), and 23% were elongated across the contour of the slope. P i c e a m a r i a n a tree islands were located at two of the southern sites along the Tuktoyaktuk Peninsula (sites T6 and site T12, Figure 4.1). P i c e a m a r i a n a and L a r i x l a r i c i n a were found near treeline along the Anderson and Horton River valleys, though four outlying L a r i x l a r i c i n a were located approximately 10 km south of site A3 (Figure 4.1). Detailed descriptions of each site are presented in Appendix 1, along with descriptions of sites located south of the tree island sites. 49 Figure 4.5 Tree island T4, located along the Tuktoyaktuk Peninsula (see Figure 4.1). The Picea glauca at this site typically exhibit a mat growth form, with few infranival erect stems (see Figure 4.2). Figure 4.6 Tree island A3c, located within the Anderson River valley (see Figure 4.1). Individuals at this site exhibit supranival skirted and erect whorled growth forms (see Figure 4.2). 50 F i g u r e 4.7 Location of 16 tree islands at site T6, Brunet Spruce (see Figure 4.1). Lakes are given informal names. Tree islands at this site were found on various slopes with east, west, south, or north facing aspects. 51 4.4 Identification and Differentiation of Picea glauca from Picea mariana The identification of Picea glauca was made using the morphological characteristics listed in Table 4.1. In the field, female cone size and shape, cone bract morphology, twig colour, external shape, and lack of pubescence were the main criteria used to identify an individual as Picea glauca. A t all sites, Picea glauca exhibited the typical morphological attributes of the species: its twigs were glabrous and pale yellow/green in colour, with raised and rounded exteriors and pulvini; female cones were elongated with almost entire bracts that were resinous and dull; female cones opened at the end of the growth season. Generally, male cones were located towards the base of the tree and female cones were found towards the growing tips. However, there was little distance between the positions on the branches since many individuals were prostrate, dwarfed, and had many dead leaders. Six tree islands, that had been located by 1993, were sampled for detailed study of morphological characteristics: T l , T3, T5, A l A3a/b, and A3c (Figure 4.1). O f all the characteristics, twig colour and the presence of fine hairs on twigs were most variable (Table 4.3). The majority of twig samples were glabrous, with a few samples that had scattered fine hairs. The exception was at site T l , where most of the twig samples were slightly hairy and one sample showed pubescence typical of Picea mariana. A t all sites, twig colour was yellow/brown, but many samples (all samples at T l ) had reddish pulvini. In contrast to published reports, all immature female cones were purple or purple/brown in colour, rather than green/brown as expected. A l l pollen samples were of the size and shape typical of Picea glauca. Two varieties of Picea glauca have been described for the N . W . T (in the former District of Mackenzie; Porsild and Cody, 1980). Picea glauca var. albertiana has a narrow CU Cj w s n o m s a - y lojvj/Auiqs Ji u ' 2 13 8 o o> tyi snouis3"a/{|nQ g o g o o o u o 0 o o o S 2 8 2 2 2 j u 13 o _ CO g ^ a j B j n o i j p u a Q o 5 U 5b *-< u ca £ sjuug jsouqv g o g o o o g § g § § § ui 3 O 2 snouisa'a CU cu o o o o o o o o o o o o o o o o o o cs~ 6 0 ^ J B T . T ^ puno -a /pas iB -g o o o o o o o o o o o o o o o o o o OJ 6 0 eob H p u n o - g o o o o o o o o o o o o o o o o o o qsippa-g u> 3 _o § ! m A I M P3"H 'a/A 6 0 H U M O j g / M O T p A o o o o o o 2 —< O N r-» «o <o ~ vo oo in vi ro ,-. ON os m w-i io ° m I-H rj- vo Co" juaosaqnj o U XjiBH^pqSiis .SP * snaiqtJTfj >o o o o o o S - 2 ?! ^ « o\ - H in in OO OO O N ON • | a i d i n g £ g uMOja/aidjnj S U s 1 - 1 uMcug/usaiQ o O g g g a a 2 2 2 £ 2 g o o o g o g o o o - C 2 " i " H H H ^ <; tu C o u j u 13 6 o 53 and spire-like crown, and forms polar and alpine treeline. P i c e a g l a u c a var. P o r s i l d i i tends to dominate rich alluvial soils in the area and has relatively smoother bark with resin blisters and a full, bulb-shaped or pyramidal crown. It is this variety that had been misidentified as Abies balsamea near the western end of Great Slave Lake at approximately 62°N (Porsild and Cody, 1980). In the Tuktoyaktuk region, clear examples of P i c e a g l a u c a var. P o r s i l d i i were observed at sites T6 and HI (Figure 4.1). These individuals were among the tallest in the tree islands. Their prominence across the region, their presence at the other sites, and differences between varieties with respect to reproduction, growth form, and age warrant further investigation. 4.5 Substrate Conditions Substrates were predominantly organic-rich material with a thick moss layer. However at five sites (T2, T4, T6, R l , R2), tree islands occurred on well-drained sand or gravelly sand material (Figure 4.8). 4.5.1 Soil Descriptions At the tree island sites, there were three broad soil types present, all belonging to the Cryosolic order. Along the Tuktoyaktuk Peninsula, sites TI , T3, T5, T6 (except T6h) had organic-rich Cryosols. Typically these soils had a thick litter and moss layer at the surface that varied between 3 and 20 cm thick. Below this were one or two layers of decomposing organic matter. The upper layer was medium brown, slightly decomposed organic matter, between 0 and 15 cm thick. The second layer was dark brown to black, moderately to well decomposed organic matter, between 0 and 36 cm thick. In some soil pits at site TI , and Figure 4.8 Tree island T2, located along the Tuktoyaktuk Peninsula (see Figure 4.1). The substrate conditions at this site are well drained gravelly sands. 55 likely in the valley bottom sites at T6, the permafrost table was encountered before the parent material was reached. The parent material was a light brown-grey clayey silt material with well-rounded pebble and small cobble inclusions. Cryoturbated soils were observed at sites T l and T6. The second type of soil was a poorly developed soil, tentatively classified as a Regosolic Static Cryosol. It was found at sites T2 and T6h along the Tuktoyaktuk Peninsula and R l and R2 on Richards Island. A shallow, layer of litter and moss overlay a sand to gravelly sand material. The third soil type was found at sites A l , A2, A3, and HI . An upper layer of litter and moss varied between 3 and 20 cm thick. This was underlain by a brown to dark brown moderately to well decomposed organic layer, upwards of 5 cm thick. Below this was a 20 to 30 cm thick layer of dark brown to black silt- and clay-rich material with a high organic content. At site A l , the permafrost table was encountered in some pits prior to reaching the parent material. The parent material was a grey, silty clay or clayey silt material, with pebble and small cobble inclusions. Evidence of cryoturbated soils was present at site A l . At all sites across the region, no visual differences were detected between soils inside or outside the tree islands. 4.5.2 Active Layer Thickness At the Richards Island sites, no active layer thicknesses were measured. However, Dr. Chris Burn, who has worked extensively at site R l , found that snow melt has tended to occur early in the growth season on the slope where the tree island was located. By mid-August the active layer was often greater than 100 cm thick (Burn, personal communication, 1995). At sites T l , T3, T5, T6i, A l a , A2a, A3c, and Hla , active layer thicknesses were measured both inside and outside the tree islands.. On average, active layer thicknesses were 56 similar inside and outside tree islands at sites T3, T5, A l a , A2a, and H l a (Table 4.4). At sites T l , T6i, and A3c, active layers were greater than 5 cm thicker outside the tree islands. However, mean active layer thicknesses were significantly different only at site T l . In general, active layer thicknesses were highly variable, reflecting microscale differences in such factors as topography, drainage, insolation, and vegetation cover (particularly moss). Active layer thicknesses between sites also reflected the sampling date. Table 4.4 Active layer depths measured both inside and outside selected tree islands (see Figure 4.1 for site locations). Site Inside Tree Island(cm) Outside Tree Island (cm) Date Samp] (ymd) Mean Range Mean Range T l * 30 10-45 35 16-51 93 07 19 T3 57 25-111 54 25-69 94 07 17 T5 29 16-53 31 6-53 93 07 14 T6i 54 37-58 70 48-92 94 07 25 A l a 43 38-44 45 25-53 94 07 09 A2a 37 25-51 36 28-48 94 07 04 A3c 34 16-50 47 30-61 93 07 29 H l a 48 38-60 52 44-60 94 08 08 •Significant difference (a=0.05) in mean active layer depths inside versus outside tree island 4.6 Non-Arboreal Vegetation The ground vegetation in the Tuktoyaktuk region was dominated by Betula glandulosa Michx., Salix spp. L. , Arctostaphylos rubra (Rehd. & Wils.) Fern., Vaccinium vitis-idaea L. , V. uliginosum L., Empetrum nigrum L., Dryas integrifolia M . Vahl, Lupinus arcticus Wats., Equisetum variegatum Schleich., Cassiope tetragona (L.) D.Don, Ledum decumbens (Ait.) Lodd., Pyrola grandiflora Radius, and various species of Gramineae, mosses, and lichen. Betula and Salix had clumped distributions. They were absent within some tree islands (Betula was absent from inside site Hla , Salix from inside site T6i), but more generally their 57 cover was less than 30-50%, and occurred in more open (less low krummholz) parts of tree islands. Outside the tree islands, they had a greater areal cover. Vaccinium vitus-idaea tended to have a greater coverage inside the tree islands, though it was absent from the Anderson River valley sites. Vaccinium uliginosum and Dryas integrifolia tended to have a greater coverage outside the tree islands. At most sites, mosses tended to have a greater coverage inside the tree islands, except at sites T l and A2a where higher moss levels occurred outside the tree islands. The coverage of all other common ground vegetation was similar inside and outside the tree islands. Rhododendron lapponicum (L.) Wahlenb. was prominent outside the tree islands at sites T6i, A2a, and A3c, the intensively studied sites nearest treeline along the Tuktoyaktuk Peninsula and Anderson River valley. Alnus crispa (Ait.) Pursh was found lake-ward in tree island T3, and 5 m from site T6i on a steep south-facing slope. Otherwise it occurred along drainage courses in the area of the remaining sites. 4.7 Insects Though the tree islands were isolated, particularly at sites T l to T5, and A l , evidence of insect infestation was present in the Tuktoyaktuk region. Live spruce bark beetles and their larvae (Dendroctonus punctatus) were found at T3 (identification by Tim Boulton and Richard Ring, 1994). The beetles occurred within the bark and vascular cambium and the larvae burrowed into the wood, both leaving tracks. Tracks in wood samples were found at sites T3, T6, A2, A3, and HI . Live larvae were also found at site HI . These larvae were of three types. One larval type was very small and white, and was similar to that associated with the spruce bark beetle at T3. Two other larval types were found at HI : (a) large (approximately 1.5 cm long), ivory coloured, ringed like a worm, one end tapered, the other 58 end wide (~3 mm) and flat with an oral opening and "mandibles", found in the wood region, and (b) small (approximately 0.5 cm long and 1 mm wide), mustard coloured, ringed like a worm, tapered at both ends, found in the bark and vascular cambium. 59 C H A P T E R 5 R E P R O D U C T I O N 5.1 Introduction It has been suggested that tree islands occurring within the Low Arctic Shrub Tundra may be relict populations originating under more favourable climatic conditions in the past and persisting in small patches after climatic deterioration and, possibly, treeline retreat (Payette andGagnon, 1979; Legere and Payette, 1981; Elliott-Fisk, 1983; Spear, 1983; Ritchie, 1984). Central to this hypothesis is the concept of vegetative inertia and the slow change in treeline or species limits in response to climatic change. This inertia is due to the longevity of trees and their ability to maintain a population by vegetative reproduction under stressful climatic conditions that may be unfavourable for sexual reproduction. The prominence of vegetative reproduction and lack of sexually produced individuals and seedlings have been used as indicators that conifers at of beyond treeline are not in equilibrium with current climatic conditions and are, therefore, relicts of past, more favourable climates (Elliott-Fisk, 1983; Payette, 1983). The ability to reproduce sexually has been used to suggest that segments of northern treeline in Labrador and northern Quebec are currently in equilibrium with climatic conditions (Elliott-Fisk, 1983; Payette and Morneau, 1993). Sexual reproduction is the primary mode of reproduction in P i c e a g l a u c a throughout most of its range (Nienstaedt and Zasada, 1990). The occurrence of vegetative reproduction by the process of layering (Figure 5.1) has been infrequently observed. Early reports identified layering in P i c e a g l a u c a growing on rocky outcrops with thin soil mantles in Ontario and Quebec (Cooper, 1911; Fuller, 1913; Bannan, 1942). More recent studies indicate that layering in P i c e a g l a u c a occurs at its range limits, including northern treeline (Elliott-Fisk, 1983; Ritchie, 1984; Filion et al, 1985; Timoney, 1988; Larsen, 1989; 60 * A / / _ ^^—r / / ^ ^ ^ ^ / ^ ^ T ^ T 7 ^ ! ^ r / ' / ' / ' ^ / / / / / / / / M V / / / / / / / / / / / / / / / / / / / /A. / / fry / / / / / / / / / / / / / / / / / / / „ \ / / / / / / ' / / - / . / / / / / s / / s s / / / / A Figure 5.1 Depiction of vegetative reproduction by layering in Picea. A lower branchis buried, roots form adventitiously, and the shoot tip continues to grow, forming a new individual. 61 Nienstaedt and Zasada, 1990; Zasada et al, 1992) and at altitudinal treeline and range limits within northern mountain ranges (Szeicz and MacDonald, 1995a). Information concerning vegetative reproduction in tree islands of northwest Canada is limited to citations of the presence of P i c e a g l a u c a krummholz and clones (Spear, 1983; Ritchie, 1984; Timoney, 1988). In other treeline and timberline species of Canada, vegetative reproduction is more common. P i c e a m a r i a n a commonly reproduces by layering at treeline and in boggy areas throughout its range, though it is less frequent or rare in dense forests (Viereck and Johnston, 1990). P i c e a e n g e l m a n n i i (Parry) Engelm. reproduces by layering most often at altitudinal treeline, but post-fire vegetative reproduction is known to occur in survivors (Alexander and Shepperd, 1990). L a r i x l a r i c i n a reproduces primarily through sexual reproduction, though there are some reports of layering (Payette and Gagnon, 1979; Legere and Payette, 1981; Johnston, 1990). P o p u l u s balsamifera and P o p u l u s tremuloides can reproduce vegetatively quite vigorously. P o p u l u s balsamifera can produce new ramets from intact or broken roots, from buds on stumps and at the tree base, and from buried branches (Zasada and Phipps, 1990). P o p u l u s tremuloides produces ramets by root suckering (Perala, 1990). The focus of this chapter is reproduction in P i c e a g l a u c a at its range limits in northwest Canada. The objectives were: 1) to determine the importance of vegetative and sexual reproduction for establishment within P i c e a g l a u c a tree islands, and 2) to assess the current capacity for sexual reproduction in tree islands, comparing it with P i c e a g l a u c a within the High Subarctic Forest-Tundra. This will contribute, in part, to determining whether these tree islands are relicts of past treeline advances which occurred under more favourable climatic conditions (Hypothesis (iii), Chapter 1), or are normal components in the range of P i c e a g l a u c a which are in equilibrium, at least reproductively, with current climatic conditions (Hypothesis (ii), Chapter 1). 62 5.2 Methods In the greater Tuktoyaktuk region, 23 sites were sampled at various times during four growth seasons, 1991, 1993, 1994, and 1995 (Figure 5.2). 5.2.1 Current Sexual Reproductive Capacity At tree island sites, the presence or absence of female, male, and bisexual cones were noted. Female cones were collected from tree island sites and from forest stand sites near treeline and within the northern Forest-Tundra. Two transects were established that span from the northern Forest-Tundra northward to the tree islands. These are referred to as the Tuktoyaktuk transect (sites T l through T10) and the Anderson River valley transect (sites A l through A8, Figure 5.2). The particular collection years (1991, 1993, 1994, 1995) and the time of year when the cones were collected (early July to mid-August) varied for each site due to logistical constraints of working in these remote areas. P i c e a g l a u c a seed is scattered by wind in autumn and early winter, but some is held in the cone until the following spring and summer (Rowe, 1970). At tree island sites where few cones were present, all cones were collected. At tree islands where cones were plentiful, from five to >30 cones per tree were collected. Within stands of P i c e a g l a u c a located in the Low Arctic Shrub Tundra or the High Subarctic Forest-Tundra, at least 25 individuals were sampled with a minimum of five cones per tree. Cones from the previous year were collected as were cones from the current crop year that were thought to be sufficiently ripened. Cone ripeness was subjectively determined by such factors as cone colour (purple to green/brown as ripening proceeds), amount of sap on the cones, and the degree of cone opening. Cones were stored for several weeks at room temperature to allow further ripening of seed (Caron et al, 1993). In the laboratory, seeds were removed from cones by shaking. Seeds were imbibed Figure 5.2 The location of tree island and forest stand sites in the Tuktoyaktuk region. \ \ 64 for 24 hours, de-winged, and stratified for three weeks at 2 to 5 °C. In each sample year, germination tests were carried out by placing up to 100 seeds on 2 mm thick moist filter paper in petri dishes (12 cm2). From 40 to 850 seeds were tested for each site, depending on seed availability in a particular year. The tests were conducted under constant light conditions, with temperatures varying from 22 to 28 °C. Successful germination was determined as the emergence of the radical and hypocotyl to seed-length. Germination tests were ended after three weeks (Schopmeyer, 1974; Ellis et al, 1985). Soil samples were collected in the summer of 1993 to test for germinable seedbanks, providing an indication of recent sexual reproduction and the potential for future seed germination and seedling establishment. Soil samples (5 cm x 10 cm x 5 cm depth) were taken at one metre intervals along either: (i) two intersecting transects (north-south, east-west) that spanned the width and length of the intensively studied tree islands (sites T l , T3, T5), or (ii) along one transect that spanned the long axis of the tree island (T6a, A l a , A l b , A2a, A3c, Hla). The transects continued for ten metres outside the tree islands into the tundra. In the laboratory, the soil samples were loosened, spread in plastic dishes (10 cm x 15 cm), kept moist, and tested for germinable P i c e a seeds under constant light levels, at 20° to 25°C. The tests were run for three weeks. To assess recent establishment success, seedlings and vegetatively produced juveniles at each tree island were identified and a search for seedlings was conducted in the vicinity. 5.2.2 Seedling Survival Establishment of sexually reproduced individuals may also be limited by the post-germination survival of seedlings. To assess seedling survival, transplant studies were 65 conducted. Cones were collected from P i c e a g l a u c a in the Inuvik area (site T10) at the end of the 1993 growth season. Seeds were extracted and stratified using the techniques outlined in Section 5.2.1. Three stratified seeds were then placed on top of a peat-perlite-dolomite mixture (1 bale: 1/2 bale:660g) within tree seedling cones (SC-10 Super Cells) and topped with 0.5 cm of fine gravel. The seeds were placed under 24 hour light in a greenhouse (temperature variation approximately 16-24 °C) and were misted twice daily until growth was observed above the gravel. Seedlings were watered daily and fertilized with 20-20-20 (NPK) once per month. Seedlings were grown from October 1993 to January 1994. From February to March, 1994, the light period was reduced to 16 hours to promote bud set. Then the seedlings were placed in a refridgerator at approximately 4°C for four weeks to fulfil their chilling requirement. The seedlings were transported to Tuktoyaktuk in coolers. The number of seedlings initially produced in the greenhouse was 680. Seedling loss during the growth period, bud set, chilling, and transportation resulted in 203 healthy seedlings available for transplanting. The seedling number restricted the number of sites to three: TI , T3, and T5. Ten transplant spots were randomly chosen inside the tree islands and ten spots in a circular arrangement outside the tree islands, 1 m from the edge. The 1 m distance from the edge was selected (i) because this allowed easier re-location of the small seedlings within the dense shrub tundra communities, and (ii) most seed would normally fall near the parent trees, therefore, it would be likely that seedlings might establish closer to the tree island. At site T5, four seedlings were transplanted per spot, resulting in 40 seedlings inside the tree island and 40 outside. At site T3, there were 31 seedlings planted inside and outside the tree island. At site TI , 30 seedlings were planted inside, 31 outside. Seedlings were transplanted June 20, 1994 at all three sites, two days after arrival at Tuktoyaktuk. Survival was checked at the end of the 1994 and 1995 field seasons. 66 5.2.3 Reproductive Origin of Individuals In order to assess the importance of sexual versus vegetative reproduction in the establishment of P i c e a g l a u c a , each individual within selected tree islands was mapped and the reproductive origin determined. Tree islands were selected at sites south to north along the Tuktoyaktuk Peninsula and lower Anderson River, creating two transects. Along the Horton River valley, extensive sampling was limited to one site, HI . At sites with multiple tree islands, the tree islands selected for sampling were the first tree islands located. At site A9, located within the Forest-Tundra, six 3 m x 60 m transects were randomly located within the forest stand and the reproductive origin of each tree was determined. Origin of an individual by vegetative reproduction (layering) was determined by the excavation of underground connections between individuals (Figure 5.1, Figure 5.3). If only roots were found, the individual was considered to have originated by sexual reproduction (Figure 5.4). Arseneault and Payette (1992) identified vegetatively reproduced individuals of P i c e a m a r i a n a and P i c e a g l a u c a by the presence of a curved trunk base, rather than by excavation of underground connections. However, in permafrost areas, ice-cored hummocks initiated under P i c e a may occur, resulting in a partial or complete toppling of the tree (Viereck, 1965). If the tree continues to grow, often it will have a curved base. The resultant form is a mass of roots encapsulating a hummock from which a tree curves outward and upward. Curved bases may also result from slumps, soil creep on slopes, ground movement in ice wedge polygons, or snow loading on small, thin-stemmed trees such as those occurring in northern areas. In the Tuktoyaktuk region, some individuals had curved bases, but no subsurface connections, and some individuals appeared to have straight stems, but the curved base was 67 Figure 5.3 The photograph depicts three individuals produced by layering and their underground connections. The finger is pointing to the connection between two individuals in the lower portion of the photograph. The trowel indicates the connection between individuals in the lower right and the upper left parts of the photograph. Figure 5.4 The photograph shows individuals established from seed, as determined by the presence of roots and the lack of underground connections. The individuals located to the bottom, to the right, and in the centre of the photograph originated from seed. 68 ' well covered by the development of a very thick peat layer. 5.2.4 Data Analysis In order to evaluate changes in reproduction from south to north in the study region, linear regressions were conducted. Male and female cone production, seed germination, and reproductive origin were regressed against distance from treeline for each sample year. The fit of the models was assessed using A N O V A , with an a level of 0.05. Kolmogorov- ' Smirnov tests for normality indicated that the errors from the regressions corresponded to normal distributions. Differences between sample years in cone production, seed germination, and transplanted seedling survivorship were tested using analysis of variance (ANOVA), with an a level of 0.05. A l l analyses were conducted using SPSS 10.0. 5.3 Results: Current Sexual Reproductive Capacity 5.3.1 Cone Production In all years visited, all sampled tree islands, except site T l , had individuals that produced male cones (Table 5.1). Typically more individuals produced male cones than female cones. No female cones were produced at site T l in any of the sample years. None of the individuals at the tree island at site T5 had female cones in 1993. Including all sites from which quantitative measures of cone production were made, there were no significant linear relations between tree island cone production (male or female) and distance north of treeline (F = 2.004, p = 0.177 and F = 0.184, p = 0.673). However, results of the analyses changed with the removal of site H l a , located at the easternmost extent of the study region. Including only the data from sites along Richards Island, the Tuktoyaktuk Peninsula, and the Anderson River valley, some significant relations tu tu u. *^  21 '5 3 5 |5 — 13 o ia E to 2 B a .9 2 "3 CU | 5 H O O O I O O O O O O O O i © © \ 0 © i / - > © © 0 © oo CN r o O T t O o o T t o o o i o t— o ro o © o s s S S S £ fe fe fe SO _ l r o oo tN PH o S g fe CN T t n oo H N N N m N © 2 S ' S S ^ S f e - f e f e o S ^ f e ^ ^ S S ^ ^ - n f e O P H o S - fe o r-CN « N M \ f IN * - (N N « o S O N O r» T t T t r o r o + + + + - H CN r o T t H H H H r o t N > D T t T t T t r o t N i — i t N + + + + + + + + + 3 - -« < < a o co ( N r o ^H <! •< ffi T 3 O a tu tu 3 o o tu c o o Xi •3 * -a .a is tu PH . fe CO ts T 3 «4H O M o JH C3 CU 4-» ta o •3 tu JS w CO Q CU tu 4-» 6 M '% CA tu •a tj 3 o >/-i os OS T t os OS ro Os Os tU O CJ — i O w J3 <N •/-> tu T3 c -cs 3 « * tu tu H tu > . <*H tU ° I & s )H tU fe & u . £ S ts — 93 H tu so so N ? oo i s s ^ i n N ? oo OS I D o OS wo 2 s° » - I O tN ^H tN .-H ~H —4 O oo wo ^ Co* \n Cp" ^ Co" r o To —' _ ^ o ^ *0 cN — J ? T t - - . O ^ 0 0 ^ 1 O ' ' s—, ^ / T T ' 1 • SO Co vo i o OS ^ os 2: o s " i T t ^- O ~^  Os oo in 2 2? oo vo r o . c uoOS T t r o 00 m oo T t l v o C o ' t N C o ' J j C o ' r O N O S s ~ V D ~ O 0 5 ; C N ~ r O 'os>n r - m S t o t N ro >^ Co 9s C ^" 2=1 ^ ^ ^ c? 2 H ^H CN - H <C <C ffi 70 became apparent (Figures 5.5 and 5.6). Combining 1993 and 1994 data, there were significant linear relations between distance north of treeline and the production of male cones ( r 2 = 0.359, F = 7.297, p = 0.018) and female cones (r2 = 0.268, F = 5.482, p = 0.033). However, the relation between cone production (female or male) and distance north of treeline was not significant for the individual years (p>0.05 in all cases). In addition, there were no significant differences between sample years (1993 versus 1994) in either male or female cone production (F = 1.266, p = 0.279 and F = 1.525, p = 0.237, respectively). At sites with multiple tree islands, not all tree islands produced cones each year (Table 5.2). At T6, eleven of sixteen tree islands produced female cones in 1993, nine of the tree islands in 1994. At site HI , eight of the thirteen tree islands produced female cones in 1993 and three tree islands produced female cones in 1994. At site A l , five of the twelve tree islands had female cones in 1993, seven tree islands had female cones in 1994, and eight tree islands in 1995. At site A2, 34 of 50 live tree islands (68%) had female cones in 1993, 25 of 50 (50%) in 1994. A2a had female cones in 1993, 1994, and 1995. At forest stand sites (T7 through T10, A4 through A8), all adults were producing male and female cones. Bisexual cones were identified at many sites: in 1993 bisexual cones occurred in tree islands at T3, A l b , A3c, and two tree islands at HI ; in 1994 at tree islands T3, T5, two tree islands at A l , and two tree islands at HI ; in 1995 at one tree island at A l (Table 5.1). A second type of cone abnormality was identified at site T3 in which female cones had ceased to grow, and twigs and/or needles then grew from the tip of the aborted cone. 5.3.2 Seed Germination Tests The results of germination tests using seed collected in the various years are listed in Table 5.3. The northernmost tree islands, particularly those along Tuktoyaktuk Peninsula, 71 • Female Cones 1993 • Female Cones 1994 " Regression Female Cones 1994 Regression Female Cones 1993 "Reg ress i on Female Cones All Years 1994 y = -1.3574x + 71.394 R 2= 0.253 p=0.204 1993 y = -0.8623X + 46.668 R 2= 0.268 p=0.293 All Years y = -1.1213x +58.898 R 2= 0.2677 * p = 0.033 25 30 35 40 45 Distance North of Treeline (km) 50 Figure 5.5 The decline in female cone production with increasing distance north of treeline. The data exclude those from site H 1a. Results of the regressions are shown for sample years 1993,1994, and all years combined. 72 0 H , •—, , , •—I 25 30 35 40 45 50 Distance From Treeline (km) Figure 5 . 6 The decline in male cone production with increasing distance north of treeline. Site H1 a is excluded from the analysis. Results from the regressions are shown for 1993,1994, and all years combined. 73 produced very few germinable seeds. In general, germination levels were low for all sites, varying from 0 to 8% for the northern tree islands, to 0 to 30% at sites closer to treeline and within the Forest-Tundra. Combining the 1993, 1994, and 1995 data, there were significant declines in germinable seeds from south to north, for the Tuktoyaktuk and Anderson River transects (regressions, r 2 = 0.521, F = 22.852, p < 0.001 for Tuktoyaktuk, and r 2 = 0.274, F = 8.296, p = 0.009 for Anderson River; Figures 5.7 and 5.8). There were no significant differences in seed germination between years. However, examination of each sample year and transect, individually, indicated that the relation between germination levels and distance north of treeline was significant only for the 1993 data for the Tuktoyaktuk transect (r2 = 0.972, F = 174.68, p = 0.0001) and Anderson River transect ( r 2 = 0.840, F - 31.55, p = 0.001). There was no germination of Picea seeds in any of the seedbank soil samples collected from inside and outside the tree islands. 5.3.3 Seedlings Across the greater Tuktoyaktuk region, searches in the vicinity of each site resulted in the identification of only two seedlings. Both seedlings were found within a metre of each other at site T6f (Figure 5.2, Appendix 1, Section A l . l ) . The seedlings were in an area of old stumps and 3 m east of a more recently dead stem and a small, chlorotic krummholz. The closest living, healthy individual that was producing female cones (three cone crop years were present in 1994) was 8 m north of the seedlings . No seed collected from this site germinated. In 1994, the seedling ages were estimated by nodal counts as 13 and 30 years, indicating that they established in approximately 1982 and 1965, respectively. The seedling heights were 15 cm and 39 cm, with basal diameters of 0.7 and 0.8 cm, respectively. Their height increments were 1.2 and 1.3 cm per year. They had erect stems, were healthy in 1994, c o -a <D 74 ro s H V ) O N O N I £ "O O N -B - | "3 U o O S ^H T3 co ul O N ON ON co O N ON "S M O N O N | O N fl» O N - J -— U V "o U eu ej 9 S CC S u £ o 2 H a a CA C J o o o a H 00 co • C rs — co — a o C J O s ~ ^ ^ 1 / 1 O N W . p O CN CN <N C J C S o o a C J a o C J O a e j a o o o c O vi m m r-. r— © © so a a a ° o «; * ^ m in so C J g «-O o O <—i CJ o CN CJ c o o o a CJ c o o o a m © C J o — ° © • C J O o a o c CJ C o u o a C J a o o o a o C O d Cd c d a CN sq r-—; CN ^ I O I <— O N O C—• " C l - >—i O N •<d- ^ C O C O C O C N + + + + + + in « , ^ O + ° ? CN NO - H C N c o - c t - i n N o r - o o o N - i co m + + 2 2 co so Os O co - c j - r O C N C N ^ P ^ c N C O + + + + + • • I - H t N c o ^ - m N o r - ~ o o _ c d £ > • M M • f f i W X a CJ >> -1—* cd C J c j a. .a _ c d ' e d > c d •*-« o a c d 75 35 - i < 1 > Northwest Treeline Southeast of Treeline of Treeline Figure 5.7 Picea glauca seed germination from sites along the Tuktoyaktuk transect (see Figure 5.2). Note that at site T5, one of twelve seeds germinated, giving an artificially high germination percentage. Relative location of sites is not to scale. 76 Figure 5.8 Picea glauca seed germination from sites along the Anderson River transect (see Figure 5.2). Note that distances between sites are not to scale. 77 and occurred at the sides of hummocks. For the seedling transplant study, survivorship over the first growth season (1994) was severely diminished as a result of ground squirrels digging and trampling in the area of the seedlings (Table 5.4). It appeared that these animals were attracted to the pink flagging tape that was used to mark seedling locations. In each case, the flagging tape was well chewed and widely distributed. Survivorship was low over the two seasons, varying from 20% to 45% inside the tree islands and 23% to 35% outside the tree islands. No significant differences in survivorship were found between seedlings planted inside or outside the tree islands for 1994, 1995, or for the two years combined (F = 0.07, p = 0.804; F = 2.05, p = 0.225; F = 0.597, p = 0.458, respectively). The limiting influence of seedling survival on establishment of individuals beyond treeline warrants further and more extensive investigation (see Section 8.4). Table 5.4 Survivorship of transplanted P i c e a g l a u c a seedlings over two growth seasons. Site locations are shown in Figure 5.2. Site Inside Tree Island Outside Tree Island 1994 1995 Two Seasons 1994 1995 Two Seasons TI 8/30*10 (27%) 6/8 (75%) 6/30(20%) 13/31*5 (42%) 7/13 (54%) 7/31 (23%) T3 19/31 (61%) 14/19 (74%) 14/31 (45%) 14/31*7 (45%) 8/14 (57%) 8/31 (26%) T5 20/40*6 (50%) 17/20 (85%) 17/40 (43%) 17/40*7(43%) 14/17 (82%) 14/40 (35%) = number of seedlings that died due to ground squirrel digging 5.4 Results: Reproductive Origin of Individuals 5.4.1 Propensities for Vegetative and Sexual Reproduction At the sampled tree island sites, vegetative reproduction produced from 33% to 90% of the individuals (Table 5.5). Including all tree island samples from the Tuktoyaktuk and eu eu > eu o a a i'-c £ ° eu eu S* >> eu /-v • - o -l-J w <« 2^ o n i_ a eu "O = I z £ ss •= - s - 0 O 5 ? a -o w z eu eu 1 t/3 a H - M » - n S3 O _ 1-2 •» 5 S l u -l l £ j , 2 s I 3 2 Z fl — eu 8 -a 43 * * s .2 o Q M eu H-* N ? c x r o oo ro s °° a c x O r o oo O in — TT —• so r- 2 ^ "> os t-~ - H os T J- r o r o C N + + + + -i m ui « H H H H c x r o c x OS ro SO ~ o x c x C ? c x o x CN O oo ro CO CN CN -^^  OS O r v , v , — . , N ? o x c x o N c x o x r o 0 0 m SO m o oo OS 0 0 so in •n ^—s • ' ^—^ SO oo in © SO SO oo 0 0 so SO CN t CN CN o x c x - H r o SO o © m CN c§ c x c? Co" c x Co* o x Co* o x c? o x r- 0 0 >n 0 0 m o -* CN i t ^—^ —^' 1 ro in O m ro ro CN CN CN CN c x C ? c x Co" r - i n 0 0 SO CN CN ^—s SO CN N? v O x = 0 s - c> o~-CN m CN m CN 0 0 m >n m 0 0 •n •n r o >n r o CN NP N f r o r o in CN m so ^ SO r-~ >n 0 0 t^ OS 0 0 f ^ c x N ? o x VO c x 0 0 0\ r o so OO -—- 1 — ' r o CN CN m 1—1 f ^ / ^ , ^ /—^\ o x oo oo s"^ c x v O o x N ? ox -<o c;x NO c x vo •<= o x vo c x -ct r~- CN r - 0 0 O CN CN OS oo s o oo s o SO O r-~ r o ^—^ ^ — ^ »—i ^—y ^—' CN OS ro CN s o o OS oo SO m 1 - 1 r—l so r o N ; s x s o C7\ r o O CN in — i CN in r o r o s o s o - H t - i O s •*t '"d" r o r o r o ro> CN + + + + + + + O •<t »u 4 J TO — i — C N C N r O r o r o ^ 7 8 s o a o 3 a. o u 4 -C3 O 3 bo C3 CU y a, O CJ 4 0 6 3 3 "c3 •«-» o o 1-1 a CJ 1-CJ .53 cu CJ C J 3 a cu C J CJ > CJ o CJ OH T3 CJ £ '•+-» 3 CJ O 3 CJ £ cn cj CJ £ C 3 1) Q 79 Anderson River valleys, the Horton River valley site, and one tree stand site located south of treeline near the Anderson River, the proportion of the trees establishing by either layering or by seed showed significant linear relations (positive and negative, respectively) with distance from treeline (regression, r 2 = 0.490, F = 10.55, p = 0.008; Figure 5.9). The northernmost tree island found along the Tuktoyaktuk Peninsula at site T l , had 90% of its stems established by layering. On the southern part of Tuktoyaktuk Peninsula, site T6i had a greater proportion of individuals established by sexual reproduction, reaching 55%, compared to 10% for site T l . For the Tuktoyaktuk Peninsula, the relations between distance north of treeline and the proportion of the tree island produced by layering or by sexual reproduction were not significant (regression, r 2 = 0.530, F = 2.254, p = 0.272). Along the Anderson River valley, the sample of tree islands at the northernmost site had a high proportion of individuals produced by layering (Ala 72% and A l b 55%, Table 5.5). At site A3c, 29 km north of treeline, 39% of the tree island individuals established by layering. Sites A3 a and b, however, have higher proportions of their stems originating from vegetative reproduction (75% and 72%, respectively, Table 5.5). Southward from site A4, located approximately 20 km north of treeline, P i c e a g l a u c a take on an erect tree form with straight trunks, trees are well-spaced, and live branches do not reach the ground. No tree islands were found at sites A4 through A8. These characteristics suggest a lack of vegetative reproduction at these sites. In addition, seedlings were found at A4, and appeared to be more prominent in populations southward towards treeline and within the Forest-Tundra (sites A5 through A9). At site A9, an upland tree stand located within the northern Forest-Tundra, samples of individuals along transects revealed that 11% of the P i c e a g l a u c a trees established by layering. For the Anderson River valley sites, there was a significant relation between reproductive mode and distance north of treeline. The proportion of the tree island 80 population established by layering showed a significant positive linear relation with increasing distance north of treeline (regression, r 2 = 0.517, F = 6.421, p = 0.044). Establishment by sexual reproduction showed a significant negative linear relation with distance north of treeline (regression, r 2 = 0.517, F = 6.421, p = 0.044). Within the Horton River valley, one tree island was intensively studied, site Hla , and had a high proportion of individuals established by vegetative reproduction (58%, Table 5.5). 5.4.2 Spatial Pattern of Reproduction within Tree Islands Excavation of underground connections between individuals revealed a great deal of subsurface biomass. Figures 5.10 to 5.21 illustrate the connections between individuals for the intensively studied sites. The subsurface connections between mother and offspring individuals within tree islands were, at times, simple linkages covering a short and straight distance underground (e.g., the southwest corner of T5, Figure 5.12). In other instances, the connections took a longer and more convoluted path, even passing over other underground stems (e.g., the western corner of T l , Figure 5.10). Layering did not preferentially occur in upslope, downslope, or across slope directions. Rather, it tended to occur where there was open space for live branches to reach the ground. Layering was also promoted by the toppling of an individual, which remained alive (Figure 5.22; Lutz, 1939). Toppling can occur by such factors as soil creep (downslope direction) or snow loading (generally in the downwind direction). The spatial patterns of establishment at these sites largely reflected the physical constraints of vegetative reproduction, which was the dominant reproductive mode. In vegetatively reproducing populations, new individuals are generally produced outwards from a central individual or individuals (e.g., T5, Figure 5.12). This tendency creates a less dense inner space within the 81 100 is 80 T J C _oo O 60 O C o a 40 o 20 y = 0.8848X + 43.972 r =0.490 p = 0.008 / ' -40 -30 -20 • Established by Layering a Established from Seed - - Regression • • * m • • 0 • • -10 0 10 20 Distance From Treeline (km) 30 40 50 Figure 5.9 The live proportion of each sampled tree island and forest stand that originated from either layering or from seed compared to the site location north (+ km) or south (- km) of treeline. Results of regression are shown. 82 b LU g Figure 5.11 Map of site T3, Drillpad Spruce, located at 69°19'N 132°49'W, on a 12° south-facing slope. Legend as in Figure 5.10. 84 Figure 5.12 Map of site T5, Hilltop Spruce, located along Tuktoyaktuk Peninsula at 69°11'N 133°01'W. The tree island is at the top of a 5° south-facing slope. Legend as in Figure 5.10. 85 Figure 5.13 Map of site T6, Brunet Spruce, Located along Tuktoyuktuk Peninsula at 69°18'N, 132°33'W. The tree island is on a bench within a valley and faces west. Legend as in Figure 5.10. E o 6 m Figure 5.14 Map of site A1a, Armstrong Spruce, located at the mouth of the Anderson River at 69°42'N 128°59'W. The tree island is on a 4.5° slope with a southeastern aspect. Legend as in Figure 5.10. 87 Figure 5.12 Map of site A1b, located at the mouth of the Anderson River at 69°42'N 128°59'W. The tree island is on a 3.5° slope with a southeastern aspect. Legend as in Figure 5.10. 89 Figure 5.17 Map of site A2b is located at 69°39.5'N 128°43'W. The tree island is on a 4°, soutwest-facing slope within the Anderson River Valley. Legend as in Figure 5.10. 90 o 2 m Figure 5.18 Map of site A3a is located at 69°36'N 128°35'W. The tree island is on the flat floodplain within the Anderson River Valley. Legend as in Figure 5.10 4 m Figure 5.19 Map of site A3b is located at 69°36'N 128°35'W. The tree island is on the Anderson River floodplain. Legend as in Figure 5.10. E o 92 6m 1 Figure 5.21 Map of site H1a, Horton Spruce, located at 69°36'N 126°57'W, within the lower Horton River valley. The tree island is on a 2° east-facing slope, on the shoulder of a river terrace. Legend as in Figure 5.10. 93 Figure 5.22 The photograph shows a toppled individual at site T6i (see Figure 5.2) that has produced many individuals by layering, including several dead and four live individuals seen in the background. 94 tree island and dense clumps of trees along the perimeter. Back filling can also occur. As the central, older individuals die and decay or are buried, live branches are able to reach the ground in this open space, promoting layering inwards (e.g., T I , Figure 5.10; A l a , Figure 5.14). 5.5 Discussion In climatically stressed sites, establishment of sexually produced individuals can be curtailed at various stages, from the production of viable seed, to germination at a site, and survival of the seedling and sapling. In northern locations, seed and pollen production may be delayed and infrequent (Nienstaedt and Zasada, 1990). Within the Forest-Tundra of Nunavut, Elliott (1979) and Elliott-Fisk (1983) found Picea glauca trees were producing cones, but these cones yielded no viable seed or pollen. South of the tree growth form limit in the Tuktoyaktuk region (located in the vicinity of sites A4 and T7 for the Anderson River and Tuktoyaktuk Peninsula areas, respectively), all adult trees were actively producing male and female cones. North of the tree growth form limit, many individuals and tree islands were producing some female and male cones, though there were fewer at sites further north and none at site TI . Further evidence for the stressed state of these individuals is the presence of bisexual and aborted cones (Elliott, 1979b). At sites south of treeline, seed germinability varied from as low as 0.5% to upwards of 30%. There was inter-site variability apparent even in this temporally limited data set. Germination of Picea glauca seed from other sites in Canada and Alaska ranged from 9% to 98% when tested under greenhouse or growth chamber conditions (Table 5.6), with most results greater than 50% using standard techniques. In comparison, germination levels reported in this study were very low and were shown to decline with increasing latitude, 95 reflecting deteriorating climatic conditions. However, it was not clear where this decline begins and how sharp the germinability gradient is with latitude. In addition, there was a substantial change in tree growth form, cone production, and tree density with latitude. Even if germination levels are high at one site, i f few seeds are actually produced, there is still a low chance of seedling establishment. Germinable seed production (number of germinable seeds produced in a given area) may be a more appropriate measure of the reproductive potential of sites from the Forest-Tundra to northern species limit. The need for research is explored further in Section 8.4, Future Research. Table 5.6 Picea glauca seed germinability from southern sources. Seed Source Reference Conditions and Seedbed Germinability (%) Southern Manitoba Cram, 1951 Greenhouse, 12-38°C and outside 1-27°C, sand, 30 days 49% to 66% Northern Ontario Burgar, 1964 greenhouse, 10-15°C, 21, 28, 35 days 9%, 68%, 73% Central Interior B . C . Eis , 1967 control in greenhouse, mineral soi l , and litter in forest 70%, 37% to 49%, 3% to 5% Interior Alaska Zasada, 1971 . 6% to 70% (ave. 45%) N e w Brunswick Downe and Bergsten, 1991 Standard test methods, room temperature 45% to 98% Central Ontario Caron etal, 1990 Standard test methods, 27 days 50% to 97% Central Ontario C a r o n e / a / , 1993 Standard seed test methods 89% to 95% Quebec Mercier and Langlois, 1992 Standard seed test methods 35% to 80% Southern Ontario Fowler and Park, 1983 Standard test conditions, sand 37% to 9 1 % (ave. 64%) Schopmeyer, 1974 Standard seed test methods 49% to 70% The production of germinable seed takes two years in Picea glauca. In the first season, bud scales begin to develop early in the growth season. By the middle to latter part of the growth season, buds will differentiate into vegetative, pollen cone or seed cone buds. This differentiation depends on a number of factors, including position on tree, nutritional • 96 and hormonal condition of the tree, and environmental conditions. Pollen cone buds tend to quickly differentiate and become dormant, prior to complete differentiation and dormancy of seed cone and vegetative buds (Owens and Molder, 1977 a,b). In the second season, buds break dormancy, pollen grains form and mature followed by formation and maturation of the female gametophyte. Subsequently, pollination, fertilization, development and maturation of the embryos occur and mature seeds begin to be shed in the fall of the second season (Owens and Molder, 1979). Disruption at any time over this two year process limits the production of viable seed. Production of reproductive buds and high levels of viable seed are generally related to warm, dry summer conditions and a lack of frosts (Zasada et al, 1992). Because seeds mature in late summer and early fall, cold periods in the summer can delay and prevent seed maturation prior to the onset of winter. This is most problematic towards treeline and timberline, where growth seasons are short and summer frosts common (Zasada et al, 1992). In addition to physical environmental factors, biological factors such as insects, diseases, and animal predation (which includes grazing of branches and buds) also affect seed production (Zasada et al, 1992). Internal factors should not be discounted in determining production of viable seed. In southern Canada, self pollination and inbreeding result in great increases in empty seed production and reduce germination (Owen and Molder, 1979; Fowler and Park, 1983). Since a large portion of trees in the study area originate by vegetative reproduction, many of the individuals in the tree island populations are genetically identical. Seed produced by these individuals should be inbred to some degree and contribute to production of empty seed. Establishment of individuals by seed originating from an external seed source would reduce the level of inbreeding. No P i c e a g l a u c a seed germinated from soil samples taken at tree island sites. P i c e a g l a u c a seed, which is shed annually, has a limited period of viability (Wagg, 1964). In 97 conjunction with the limited production of germinable seeds, this makes it unlikely that a seedbank would normally be present at the tree island sites. Payette et al. (1982) found high levels oi P i c e a m a r i a n a and L a r i x l a r i c i n a seed in soils from the forest limit to treeline in northern Quebec. However, only iow? L a r i x seeds and no P i c e a seeds germinated from just under 5000 seeds. Payette et al. (1982) concluded that no seedbank exists for these two species. Both of these species have numerous seedlings in the forest sites and someseedlings at the krummholz sites near treeline, suggesting that, in this region, a seedling bank may be more important in regeneration. Lack of seedbank was also reported by Johnson (1975), Elliott (1979), and Elliott-Fisk (1983) in the Forest-Tundra of the N.W.T. and Nunavut. Rapid seed germination after dispersal may explain the lack of seedbank in these studies. This strategy may reduce seed mortality in the soil and enhance seedling survival by allowing a longer first growing season. Low germinable seed production and a lack of seedbanks at tree islands in the Tuktoyaktuk region are reflected in the lack of seedlings. Further limitations to establishment of sexually produced individuals can occur at the seedling stage. Moisture conditions, soil temperatures, and competition are all important factors in germination and seedling survival (Zasada, 1971). The importance of competition from ground vegetation in preventing the establishment of P i c e a g l a u c a has been shown by Scott et al. (1987b) in the Hudson Bay area. In this study, there was no increase in seedling establishment in older, undisturbed sites in response to 20th century warming. These sites were dominated by lichen ground cover which presented an unsuitable substrate for P i c e a seed to successfully germinate and grow. However, in recently exposed sites, there was an increase in establishment and a northward extension of the forest limit. These sites had a moss-dominated ground cover which acted as a more favourable substrate for P i c e a seed 98 germination and growth. In general, establishment of P i c e a g l a u c a seedlings is higher on mineral soil than on litter (Nienstaedt and Zasada, 1990). In Alaska, organic layers tend to reduce the water supply to seedlings and do not make a good seedbed (Zasada, 1971). Eis (1967), working with P i c e a g l a u c a in interior B.C., found that seed germination was less than 4% on litter, and survival of seedlings on litter over the first four years was greatly reduced compared to seedlings growing on mineral soil. In his study, survival in the first year on mineral soil varied between 54 and 70%; survival on litter was 1 to 8%. Eis also found that, in the first growth season, drought was a major factor in seedling mortality, whereas in the subsequent three years, higher competition from ground vegetation in moist environments resulted in high mortality. In general, survival increased with each year. In year two, 62% to 72% survived, in year three, 73% to 93% survived, and in year four, 88% to 99% survived in mineral soil. Zasada et al. (1992), working in interior Alaska, found that P i c e a g l a u c a seedling mortality on mineral soils was greatest in the first growth season and winter after germination. In the subsequent years, mortality sharply declined, and the survival curve leveled off near 20%. The levels of seedling survivorship reported in these studies are broadly similar to the initial findings presented here for the seedling transplant study (Table 5.4). It is suggested that competition from tundra vegetation is contributing to the lack of seedlings in the Tuktoyaktuk region. Low growth rates of northern P i c e a g l a u c a , such as for the two seedlings located at site T6f, indicate that there would be an extended period when seedlings are small and most vulnerable to competition from surrounding tundra vegetation. This competition would not only be for space, light, water, and nutrients, but ground cover can limit growing space. Thick moss layers can act to reduce soil temperatures, prevent seedling roots from reaching mineral soil and a reliable moisture source, and rapid moss 99 growth can smother seedlings that cannot rapidly gain height (Kenety, 1917). The lack of sexually produced individuals is discussed further in Section 6.5, Establishment and Climate Change, and suggestions for future work concerning seed germination and seedling establishment are presented in Section 8.4, Future Research. Vegetative reproduction is key to establishment within P i c e a g l a u c a tree islands of the Tuktoyaktuk region. Individuals originating by sexual reproduction have been very limited in number. Some degree of error is associated with the determination of reproductive origin. It is not known how long the subsurface connection between individuals may last. Studying P i c e a g l a u c a in Alaska, Densmore (1980) found that the connection between individuals degraded within 30 to 50 years. Szeicz and MacDonald (1995a), working with P i c e a g l a u c a at several sites in the mountains of northwest Canada, found that the connections decomposed after approximately 100 years. This would result in an overestimation of the number of sexually reproduced individuals and an underestimation of the number of individuals from vegetative reproduction. However, underground connections were found to be intact between individuals that were 150 to 350 or more years of age. Severance of subsurface connections is most likely to have occurred for long dead stumps. Therefore, it is assumed that the error in determination of reproductive origin of currently live individuals is small. However, the error is likely quite variable between sites, depending on the site-specific micro-environment and the proportion of old stumps within a tree island. Maintenance of P i c e a g l a u c a tree islands in the Tuktoyaktuk region has been primarily through layering. Similarly, Szeicz and MacDonald (1995a) found layering to be a particularly important reproductive mode in population maintenance at high elevation sites within the Mackenzie Mountains, N.W.T. In the northern Forest-Tundra of the Tuktoyaktuk region, little evidence of vegetative reproduction is found, particularly in the valley areas 100 where P i c e a g l a u c a are open grown with straight trunks and without live branches reaching the ground. In this area, germinable seed and seedlings are actively produced. The trend in changing prominence of reproductive modes from Forest-Tundra stands of trees to the northern tree islands is linked to the decline in climatic conditions favourable to trees. With a reduction in temperatures and length of the growing season from Forest-Tundra to Tundra, there is a reduction in energy available for production of male and female cones and for successful seed ripening, germination, and seedling establishment (Legere and Payette, 1981). Cold, dry, windy winter conditions and summer frosts also contribute to the lack of successful sexual reproduction at tree island sites. Vegetative reproduction requires a substantially greater maternal investment of resources and a longer period of offspring dependency, as compared to sexual reproduction. However, under stressful conditions, it is thought to be energetically cheaper in successfully producing an individual, since establishment of an individual from layering is far more certain than establishment from seed (Cook, 1985). Though it is unclear how long an active connection exists between individuals, the clonal offspring has access to parental nutrients and energy stores at least for some initial period of time (Barnes, 1966; Abrahamson, 1980; Legere and Payette, 1981). Under stressful conditions, clonal offspring are more likely to survive than an independent offspring produced from seed (Abrahamson, 1980; Cook, 1985). The production of sexually reproduced individuals is tightly linked with an extended period of favourable climatic conditions, starting from the time when reproductive buds are initiated to many years after when the individual is established within the population. However, the link between layering and climate is less direct. Decreased tree stature and tree density, from south to north across the Forest-Tundra to the northernmost tree islands, are related to the decline in favourable climatic conditions for tree growth. This results in the 101 increased availability of live branches near the ground, a condition necessary for the occurrence of vegetative reproduction. Within forest stands of the Forest-Tundra, live branches near the ground appear to be infrequent or absent, limiting the potential for vegetative reproduction. However, the triggering factor for the occurrence of layering has not been identified. Layering may be indirectly related or unrelated to climate and reflect such factors as snow loading, (peat) moss growth (Kenety, 1917), sediment, debris, or sand deposition which promote burial of branches (Lutz, 1939; Bannan, 1942), and loss of apical dominance (Schier, 1975) by pruning of the main stem above snowpack. Underground connections, burial of stems and branches by sand and active layering were observed at sites R2, T2, and T6h (Figure 5.2). At sites A3a, A3b, and in the vicinity of A3c, sediment deposits from flood events were observed in the tree islands at the base of trees and on lower branches. This sediment deposition may partially explain the higher incidence of layering at A3 a and A3b. Where tree island populations are maintained primarily by vegetative reproduction, the triggering of a layering event must occur frequently enough to offset mortality. Tree islands should occur where microclimatic conditions allow survival of small individuals and krummholz, where sexual reproduction occurs occasionally, and layering events are triggered frequently. These results on the current lack of sexual reproduction, in conjunction with the prominence of vegetative reproduction in tree island maintenance, are strong evidence for the hypothesis that these P i c e a g l a u c a tree islands are relict in nature. Similar conclusions were made for P i c e a g l a u c a and L a r i x l a r i c i n a krummholz north of treeline and P i c e a m a r i a n a krummholz north of the forest limit in northern Quebec (Payette and Gagnon, 1979; Legere and Payette, 1981; Payette and Filion, 1985). CHAPTER 6 A G E STRUCTURE AND ESTABLISHMENT 102 6.1 Introduction There are several possible explanations for Picea glauca tree islands in the Tuktoyaktuk region: (i) they have a recent origin, established from seed sources to the south in response to climatic amelioration, (ii) they are a normal component in the range of Picea glauca unrelated to past treeline advance, or (iii) they are relicts of a past treeline advance. In Hypothesis (i), all individuals in the tree islands would have recently established, with most from seed. Under this scenario, they would most likely have established in response to climatic warming that ended the Little Ice Age or to 20th century warming. Therefore, they would be less than approximately 150 years old (post-1850). Concerning Hypothesis (iii), treeline had advanced northward during the Hypsithermal and subsequently retreated to its current position by approximately 3500 BP (Ritchie, 1984, 1989). If the tree islands are relicts from this advance, or another minor advance during the Little Climatic Optimum, all tree islands would be of great age, with many old individuals and dead stems, few young individuals, current establishment primarily by vegetative reproduction, and establishment by sexual reproduction infrequent or absent. Given an abundance of dead stems and successful cross-dating, tree islands would date prior to the onset df the Little Ice Age (pre-1400). Hypothesis (ii) is difficult to differentiate from Hypothesis (iii). Some aspects that would be particular to the normal component hypothesis include: (a) a variety of tree island ages across the region, since tree islands would be able to establish (or re-establish after disturbance) under current or recent climatic conditions, (b) recent (post-Little Ice Age) abundant sexual reproduction in all tree islands that would reflect its ability to respond to 103 short-term climatic fluctuations and actively maintain the population by sexual reproduction, and (c) establishment by seed in areas beyond the micro-environmental influence of tree islands. Determination of tree island age structure using dendrochronological techniques, in conjunction with reproductive origin, will provide information on population demographics, health of the populations, and future prospects for the tree islands. Determination of establishment patterns through time will indicate the frequency of establishment necessary to maintain a population to the present. However, several limitations exist in examining a static age structure. The age structures of the live populations are those of the survivors: those that were recruited, established in the population, and survived to the time of sampling (Harper, 1977; Szeicz and MacDonald, 1995a). It does not capture changes in mortality and recruitment through time, and does not show changes in population levels through time. Dead stems that are available reflect a selected component of mortality in a tree island. They are the ones that have been preserved to the time of sampling. Loss of mortality information is most problematic for: (a) seedlings, saplings, and small stems which folly deteriorate more rapidly than large stems, (b) for long dead stems which have had time to deteriorate, and (c) for dead stems in wetter environments which deteriorate more rapidly and may be rapidly buried by moss, as compared to those in drier sites. Because of the loss of dead stems, population age structure in the past is unknown, and the time of origin for old tree islands represents a minimum date. In this chapter, establishment patterns of survivors are presented, along with their reproductive origins, for the intensively studied tree islands (Figure 6.1). These patterns are compared with the available instrumental climatic records from Inuvik, and dendroclimatological records from Inuvik and northwestern North America. Cross-dating 104 Figure 6.1 The location of tree island sites in the Tuktoyaktuk region that were used in this study. 105 of dead stems, mortality, and tree island age are dealt with separately in Chapter 7. 6.2 Methods Tree islands were selected at sites along the Tuktoyaktuk Peninsula and the lower Anderson River valley (Figure 6.1). Along the Horton River valley, sampling was limited to one site, HI . At sites with multiple tree islands, the ones selected for sampling were the first tree islands located. Within the selected tree islands, the reproductive origin (vegetative or sexual reproduction) of all individuals was determined by excavation. The presence of roots and no subsurface connections indicated origin by sexual reproduction. Subsurface connections between individuals indicated origin by layering. From all live individuals within the sampled tree islands, cores of wood were taken as close to the base of the tree (above the root bole) as possible. In most cases, excavation of moss, ground vegetation, and soil was necessary to expose the base. Because tree islands in the Tuktoyaktuk region are few in number, isolated, and in a stressed state, an attempt was made to limit the number of cores taken from each individual to that which was needed to date establishment. For those individuals from which a core was successfully taken from the base of the tree, which had a complete record from bark to pith, only one core was taken. However, multiple cores were taken for most individuals for a variety of reasons. Identification of the base was difficult for some individuals in which a slim stem graded into roots, with little change in diameter. In these cases, multiple cores were taken along the stem, attempting to core above and below what was thought to be the stem base. For those individuals in which the base was rotted, cores were taken above the base where wood was solid. In this instance, multiple cores were taken to obtain a complete core from bark to pith and to allow standardization to the base of the tree. Multiple cores were also taken to check for missing rings. The problem of missing 106 rings is described and discussed in Chapter 7. Multiple cores were generally not taken at right angles due to the prostrate and twisted nature of the specimens. It was often impossible to bore into the wood in more than one direction. For individuals with small stems, nodal counts were made (Fritts, 1976; Figure 6.2). The accuracy of nodal counts was tested using tree ring records of small stems that were cut down (sites T3 and A2a; Figure 6.1) or cored. At site A2, where numerous tree islands occurred, all individuals in tree island A2a were cut down at the base. Hence, this site provided the most complete record of age structure. Ring widths from collected samples were very narrow. It was necessary to sand and polish all cores and wood discs using 600 grit sandpaper. Ring counts and observable ring properties were made under 20 to 80 magnification. Reflection microscopy often aided identification of small, light rings. If no complete core from bark to pith was obtained, the number of rings to the pith was estimated geometrically (Baker, 1992). For individuals with cores from above the stem base, age of the tree at the base was estimated using a variety of means: (1) multiple cores from the same individual, (2) height-age relationship of small individuals that had been cut down, or (3) multiple cores from other individuals in the tree island. The number of rings for each sample was counted at least twice; in many cases, a second person (supervisor or lab assistant) also checked the counts. Establishment histograms were constructed using ten year age classes, because of the error associated with estimation of basal age (e.g., Payette and Filion, 1985; Szeicz and MacDonald, 1995a). The limited number of individuals in each tree island, and within each of the three areas (Tuktoyaktuk Peninsula, Anderson and Horton River valleys) precluded modeling of population age structures and examination of departures. In addition, several researchers have indicated that there are problems in some models concerning the assumptions of constant recruitment (power function) and constant mortality (negative exponential function) YEAH er Figure 6.2 Internodal growth and age determination (from Fritts, 1976). 108 for populations in which vegetative reproduction is prominent (Harper, 1977; Johnson et al, 1994; Szeicz and MacDonald 1995 a). Homogeneous data for Inuvik Airport were drawn from the Canadian Historical Temperature Database (CHTD) (Vincent and Gullett, 1999) and the Canadian Monthly Precipitation Database (CMPD) (Mekis and Hogg, 1999). The CHTD contains homogeneous monthly temperature records that have been lengthened by merging with nearby Aklavik station, missing values estimated, and gaps infilled. The techniques used to identify non-climatic variations and produce the corrected data are described in Vincent (1998) and Vincent and Gullett (1999). Temperature measures (mean daily maximum, mean daily, and mean daily minimum) were available for the period from 1927 to 1995. The CMPD contains records of monthly rainfall, snowfall, and total precipitation that have been developed from daily series that were corrected for gauge type, wind under catch, evaporation, wetting loss, and trace amounts. Records were lengthened by merging stations, and missing values and some data gaps were estimated (Mekis and Hogg, 1999). Precipitation measures (rainfall, snowfall, and total precipitation) were available for the period 1959 to 1994. Reconstructed climatic data from tree rings were available from several sources. Jacoby et al. (1985) reconstructed June and July degree days > 10°C from 1524 to 1975, using Picea glauca ring width measures from several sites in Alaska and Yukon Territory. D'Arrigo and Jacoby (1993) reconstructed annual temperatures for the northern hemisphere from a variety of tree species in northern North America, Scandinavia, and Russia. This record spanned 1682 to 1990. D'Arrigo and Jacoby (1992) presented annual temperature departure reconstruction from tree ring widths for northern North America. Sites were located in Alaska, Yukon, and along the Coppermine and Hornby Rivers. Briffa et al. (1994) used maximum latewood densities from a variety of tree species in Alaska/Yukon, 109 Mackenzie Valley, and northern Quebec/Labrador to reconstruct April to September temperatures. A n age-dependent reconstruction of June-July temperatures using P i c e a g l a u c a tree ring widths from alpine timberline in northwest Canada was produced by Szeicz and MacDonald (1995b). Szeicz and MacDonald (1996) presented a 930 year tree ring width chronology for P i c e a g l a u c a in the Campbell Dolomite Uplands, N.W.T. (near site T10, Figure 6.1). Low radial growth was associated with high moisture stress, as indicated by its inverse relations with temperatures of the previous growth year and during the spring of the growth year, and the positive relations with annual precipitation and early spring precipitation. Increased growth was associated with periods of increased precipitation and/or decreased moisture stress. 6.2.1 Data Analyses Using the instrumental climate records from Inuvik and the reconstructed climate record from sites in northwestern Canada (Szeicz and MacDonald, 1995b, 1996), each decade was classified by temperature (warm or cool), moisture level (wet or dry), and the combination of temperature and moisture (warm and wet; cool and wet; cool and dry; warm and dry). For the instrumental records, summer, temperature and precipitation levels for each decade were classified as above or below the climate normals for Inuvik. In Szeicz and MacDonald (1995b), June-July temperatures were reconstructed using P i c e a g l a u c a ring width measures from sites in northwestern Canada. Each decade was classified as above or below the 1638-1988 normal. In Szeicz and MacDonald (1996), P i c e a g l a u c a ring widths were related to changes in moisture conditions (high temperatures in May of the growth season and the previous June and July; low precipitation in late winter and spring). Decades of above average ring width indices, indicative of low moisture stress, were classified as wet. Decades 110 of below average ring width indices were classified as dry. The presence or absence of seedlings that had established within each decade were recorded from 1770 to present for the Tuktoyaktuk Peninsula and the Anderson River valley sites. For the period 1770 to 1990, the decadal pattern of seedling establishment for the Tuktoyaktuk Peninsula and the Anderson River valley were compared to the decadal temperature and moisture patterns using Chi Square Tests of Independence (Likelihood Ratio, due to low counts in some cases, a = 0.05). A Chi Square Test (Likelihood Ratio, a = 0.05) was applied to determine i f the decadal seedling establishment pattern for the Tuktoyaktuk Peninsula area was significantly different from that for the Anderson River valley area. Statistical relations between seedling establishment and both reconstructed temperature and moisture conditions were tested for by calculating Spearman's rho coefficients. Data used were the number of individuals established from seed in each decade for the Tuktoyaktuk Peninsula and Anderson River valley areas, and the decadal averages of the reconstructed temperature and moisture conditions from 1770 to 1990. Since seedling establishment is affected by postestablishment mortality, the decadal seedling establishment levels may be related to climate for some time following establishment. Szeicz and MacDonald (1995a) associated establishment and survival of Picea glauca at Subarctic timberline to summer temperatures at the time of germination and up to 50 years after germination. To examine the influence of post-germination climate conditions, reconstructed temperature and moisture conditions were averaged forward in time over 20-, 30-, 40-, and 50-year periods and correlated to the decadal seedling establishment levels. For example, the 20 year forward averaged temperature for the 1700-1709 decade was the average temperature for the period 1700-1729, the 20 year.forward averaged temperature for the 1710-1719 decade was the average temperature for the period 1710-1739, and so on. Spearman's rho I l l coefficients were calculated between the decadal seedling establishment and the 20-, 30-, 40-, and 50-year forward averaged temperature and moisture records. These analyses were conducted for the 1770 to 1949 period, as 50-year forward averages were not possible after the 1940-1949 decade. Tests whether the correlations could be zero were conducted (a = 0.05). A l l analyses were conducted using SPSS 10.0. 6.3 Results: Reproductive Origin and Establishment Patterns 6.3.1 Tuktoyaktuk Peninsula The age structure of a population reflects both establishment patterns and survivorship patterns through to the time of sampling. Establishment patterns presented here are those of the survivors. At the northernmost site, TI, one tree island was present and it was dominated by a large group of individuals established by vegetative reproduction since 1920 (Figure 6.3). Periodic establishment of vegetatively produced individuals occurred from 1770 to 1920. There were no individuals produced in the 1990s. There were a few sexually produced individuals which had infrequently established between 1710 and 1850. The last sexually produced individual established in the 1840s. The lone tree island at site T3 had a very large group of individuals that established between 1840 and 1980 (Figure 6.4). The majority of these individuals established by layering. Establishment of sexually produced individuals occurred infrequently in this tree island from 1730 to 1950, with the most recent establishment occurring in the 1940s. The oldest living individuals were established by vegetative and sexual reproduction in the 1560s and 1600s, respectively. At site T5, a few individuals had established in most decades from 1730 to 1980, but there was no large cohort of young individuals at this site (Figure 6.5). Layering has 112 CD .Q E 3 30 -r 28 -26 -24 -22 -20 -18 -16 -14 -12 -10 -8 -6 -4 -2 -0 - 1 Site T1 ^ Layered Origin • Seed Origin i • i • r i • i • i • i • i • i • i • i • i 1980 1900 1800 1700 Year (10 year establishment intervals) 1600 Figure 6.3 Establishment of Picea glauca by vegetative and sexual reproduction at site T1 (see Figure 6.1). 30 -r 28 -26 -24 -22 -20 -18 -16 -14 -12 -10 -8 -6 -4 -2 -0 - 1 CD . O E _3 Site T 3 [\] Layered Origin • Seed Origin 1980 1900 1800 1700 1600 Year (10 year establishment intervals) F igure 6.4 Establishment of Picea glauca by vegetative and sexual reproduction at site T3 (see Figure 6.1). 113 occurred at low levels since the 1830s. There were several sexually produced individuals established between 1730 and 1790, but the most recent two established in the 1820s and 1920s. Site T6i showed a sporadic establishment of one to a few individuals in most decades throughout its record, with very little variation in numbers and no large cohort of young individuals (Figure 6.6). The oldest two live individuals established in the 1620s and 1610s, being produced by layering and by seed, respectively. No sexually produced individuals have established since 1880. Some vegetatively produced individuals have established through the twentieth century, though none since 1980. Over all sampled sites of the Tuktoyaktuk Peninsula, and including the two seedlings identified at site T6j (Figure 6.7), recruitment of sexually produced individuals occurred at a low level between 1710 and 1990, with one to four individuals established and surviving from most decades. Decades in which no sexually produced individuals had established and survived were: 1630 to 1710, 1720s, 1830s, 1910s, 1930s, 1970s, and 1990s. The record of vegetatively produced individuals showed a skewed distribution of ages, with numerous young individuals and a decrease in numbers in the older age classes. However, there was a sharp decline in numbers within the youngest age classes. The young cohort of layered individuals largely reflects establishment at sites T l and T3. 6.3.2 Anderson River Valley Sites At tree island A l a , one older live individual had established by seed in the 1790s (Figure 6.8). Establishment has largely occurred since 1930, mainly through vegetative reproduction, but four individuals were established by sexual reproduction, one in the 1930s and three in the 1950s. 114 30 - i 28 -26 -24 -22 -20 -18 -JO 16 -E 14 -z 12 -10 -8 -6 -4 -2 -o -I Si te T5 [\] Layered Origin • Seed Origin ^ 1 1980 II I ' I ' I ' I ' I 1 I 1 I ' I ' I ' I 1900 1800 1700 1600 Year (10 year establishment intervals) Figure 6.5 Establishment of Picea glauca by vegetative and sexual reproduction at site T5 (see Figure 6.1). CD. JD E 30 -r 28 -26 -24 -22 -20 -18 -16 -14 -12 -10 -8 -6 -4 -2 -0 - 1 Site T6i £3 Layered Origin • Seed Origin 1980 i11 • r i n n i • nr r i • i 11 • i 11 T i • i • i • i 1900 1800 1700 1600 Year (10 year establishment intervals) Figure 6.6 Establishment of Picea glauca by vegetative and sexual reproduction at site T6i (see Figure 6.1). 115 1965 s e e d l i n g 1982 s e e d l i n g \ \ \ \ \ \ \ N In> |x|N 1980 Ml Tuktoyaktuk Establishment • Layered Origin • Seed Origin N l l I 1900 1800 NTT 1700 i • i T I 1 1 1 11 1600 Year (10 year establishment intervals) Figure 6.7 Establishment of Picea glauca by vegetative and sexual reproduction at sites across the Tuktoyaktuk Peninsula (see Figure 6.1). The two seedlings located at site T6j are included (established in 1965 and 1982). 116 Tree island A l b has had a low level of establishment since 1860 (Figure 6.9). The last individual produced by sexual reproduction established in the 1940s. The oldest live individuals established in the 1860s and 1870s by sexual and vegetative reproduction, respectively. The youngest individual established by layering in the early 1990s. At this site there was no large cohort of young individuals. Tree island A2a showed very infrequent establishment by sexual reproduction, with the most recent having occurred in the 1940s (Figure 6.10). The oldest live individual established by seed in the 1620s. Establishment by layering has occurred at low levels since 1910. There was no large cohort of young individuals at this site. Tree island A3c showed an intermittent establishment pattern of sexually produced individuals from 1600 to 1880 (Figure 6.11). In the 20th century, only one individual has established by seed; this occurred in the 1950s. Establishment by layering has occurred at low levels since 1910, with one individual established in the 1790s. There was no large cohort of young individuals at this site. For the Anderson River valley sites as a whole, establishment and survival of sexually produced individuals has occurred at a low, but fairly continuous rate over the period of record (Figure 6.12). There were three major periods in the temporal record when no sexually produced individuals had established and survived: 1630 to 1670,1730 to 1760, and 1960 to present. The number of individuals established by layering has increased since 1910, peaking in the 1960s and decreasing thereafter, being particularly low in the 1990s. However, the cohort of young layered individuals largely reflects establishment at site A l a . 6.3.3 Horton River Valley Site One intensively studied tree island was available from the Horton River valley, tree island 117 30 28 -26 -24 -22 -20 <i> 18—1 ^ 16 J E _3 14 -12 -10 -8 -6 -4 -2 -Site A1 (a) [\] Layered Origin • Seed Origin N \ \ 0? NN N r i 11 11 11 • i 1 1 1 r i 11 11 11 11 11 11 • i • i • i • i • i 11 1980 1900 1800 1700 1600 Year (10 year establishment intervals) F igure 6.8 Establishment of Picea glauca by vegetative and sexual reproduction at site A1a (see Figure 6.1). 30 28 26 24 22 20 >- 18 CD •Q 16 | 14 Z 12 10 8 H 6 4 2 0 Site A1 (b) [\] Layered Origin • Seed Origin Vff l 'Vl^rr I T I T I T I T I T I T I ' I 1980 1900 1800 1700 1600 Year (10 year establishment intervals) F igure 6.9 Establishment of Picea glauca by vegetative and sexual reproduction at site A1b (see Figure 6.1). 118 30 28 -26 -24 -22 -20 -.- 18 -JO 16 -| 1 4 -Z 1 2 -10 -8 -6 -4 -2 -0 Site A2a F\] Layered Origin • S e e d Origin 1900 1980 ' I ' I ' I ' I ' I ' I ' I ' I ' I ' I 1800 1700 " i i i " i i F T T 1600 Year (10 year establishment intervals) Figure 6.10 Establishment of Picea glauca by vegetative and sexual reproduction at site A2a (see Figure 6.1). 30 i 28 -26 -24 -22 -20 -1 18 -CD .Q 16 -um 14 --z. 12 -10 -8 -6 -4 -2 -o -I Site A3c [\] Layered Origin • S e e d Origin fHr r* i ' i n ¥ r*Fr»i • i n i • i n n • i • < • i 1900 1800 1700 1600 1980 Year (10 year establishment intervals) Figure 6.11 Establishment of Picea glauca by vegetative and sexual reproduction at site A3c (see Figure 6.1). 119 30 - r 2 8 -2 6 -2 4 -2 2 -^ 2 ° -a> 1 8 -JQ E 1 6 -^ 1 4 -1 2 -1 0 -8 -6 -4 -2 -o-L Anderson River Establishment • Layered Origin • S e e d Origin • i1111 1600 1980 1900 1800 1700 Year (10 year establishment intervals) F igure 6.12 Establishment of Picea glauca by vegetative and sexual reproduction at all sites within the Anderson River valley (see Figure 6.1). 120 H l a . Establishment by sexual or vegetative reproduction has occurred very sporadically over the period of record (Figure 6.13). The most recent recruitment of a sexually produced individual occurred in the 1910s; all others established between 1700 and 1880. Two periods of establishment by layering were 1820 to 1860 and 1940 to 1960. There was no large cohort of young individuals. 6.3.4 M a x i m u m Age of Live Trees A t all sites, except at sites A l and A 2 , a large number of individuals (at sites T l , T3) or the majority of individuals (at sites T5, T6i , A3c , H l a ) established prior to 1900. A t all sites except A l b , live individuals greater than 200 years of age were present, with the oldest occurring at T3 (433 years old). In addition, there were weathered dead stems at each site. 6.3.5 Establishment and Climate Change The instrumental record of temperature and precipitation for Inuvik was available for a short time period. Several trends for the mid- to latter part of the 20th century are evident (Figures 6.14 to 6.17). On an annual basis and relative to the climate normals, temperatures were high in the late 1920s to early 1930s, low in the 1930s, high in the 1940s, and have risen since the early 1970s, with a peak around 1980 (Figure 6.14). Summer temperatures (June, July, August) have risen since the late 1930s (Figure 6.15). This rise is particularly apparent in the summer mean daily maximum temperatures since the late 1950s. Precipitation data, though available only since 1959, indicate a decline in total annual precipitation after the early 1970s (Figure 6.16). This decline is prominent in total summer precipitation (Figure 6.17). For information on climatic conditions prior to the instrumental record, it is necessary to look to available proxy records. Several dendrochronological records and reconstructed 121 30 - r 28 -26 -24 -22 -20 -18 -16 -14 -12 -10 -8 -6 -4 -2 -0 -I Site H1a \ \ \ \ I i r r 1900 £\] Layered Origin • S e e d Origin ' I' I' I' I' 1 1 1 ' 1 11 1700 1600 1980 " 1 1800 Year (10 year establishment intervals) ure 6.13 Establishment of Picea glauca by vegetative and sexual reproduction at site H1a (see Figure 6.1). 122 (Sm0|90 S99J69P) 9JniBJ9dUJ91 123 124 125 o o o o o o o o o T— 00 l O CNl CO CD CO (LULU) uoflBijdpeJd 126 climatic conditions are available for northwestern North America. These are summarized in Figures 6.18 and 6.19. Comparison of establishment at tree islands in the Tuktoyaktuk Peninsula and Anderson River valley areas with the instrumental records and climatic reconstructions described above can be made. It should be noted that the reconstructions were based on growth response of mature trees (mainly P i c e a g l a u c a , ring widths and latewood density) to simplified climatic parameters. Therefore, the comparison is between a regional climatic regime, as determined by radial growth responses of mature conifers, and establishment by sexual reproduction, a process which encompasses successful seed production, germination, seedling growth and survival to the time of sampling. Also, the disintegration of subsurface connections between individuals may have occurred over time and this is most probable for older, dead individuals. Since these data concern live individuals and subsurface connections were found to be intact between individuals that date to the 16th century, data concerning reproductive origin will be considered reliable. The decadal patterns of the reconstructed and instrumental climate records (Figure 6.18) were compared to the decadal patterns of seedling establishment for both the Tuktoyaktuk Peninsula and for the Anderson River valley. The Chi Square tests indicated that the decadal patterns of seedling establishment and decadal patterns of climate conditions (temperature, moisture, and the combined temperature/moisture records) were independent (Table 6.1). Correlations between decadal reconstructed climate records (temperature and moisture) and seedling establishment in the two areas were not significantly different from zero (Table 6.2). In addition, correlations between decadal seedling establishment in the two areas and the 20-, 30-, 40-, and 50-year forward averaged climate records were not significantly different from zero (Table 6.2). Comparing the two areas, there were no significant differences in the decadal patterns of seedling establishment between the 127 I ' I ' I ' I 1 I 1 I 1 I 1 I 1 I 1700 1600 8 -6 -4 -2 -0 Seed Origin Anderson River Sites I I I r . l . ^ 1 .11^.1 ,j ^ 1900 1800 1700 161 1980 . T T 1600 Jacoby era / . (1985) D"Arrigo and Jacoby (1992) D"Arrigo and Jacoby (1993) Northern Latitudes Coppermine River Briffa ef al. (1994) Sze i cz and MacDonald (1995b) Sze icz and MacDonald (1996) Instrumental Record Inuvik Airport very •ave. warm cool • warm very • very warm • cool very cool warm cool 4, 4,. 4^  . 4 ^ " ^ very ! warm warm warm. 1 1 viery warm OQQI cool cold warm warm cool . 1 cool , warminq 4, 4, 4,4,; * :4-! strong ! warming warm warm i i cool warm • warm cool ! 4. 4,4,. I I strong gve. warming ) 1 cool \ warm 1 cpol 1 4, 4, 4 ^ : j I I warni cool warm -cool cool warm 4^  4 ^ " ^ f ^ 4' 4' ! • cool : J T ^ j i I • very • warm 1 very cool cool warm 1 very coel cool • cool I I 1 4, .4tJ, 4 4^4^  moist : dry \ moist moist dry dry moist i , dryinq l i I I : 1 > • • 1 1980 I 1900 I I 1800 1700 I 1600 Figure 6.18 Comparison of establishment by seed and climatic reconstructions and instrumental records. See text for details of the climatic information. CD E 3 40 30-20-10-0 30 128 n Layered Origin, Tuktoyaktuk Sites 1980 1900 1800 s * 10: o: — i — i • i n—i • i n • i 1700 1600 Layered Origin, Anderson River Sites 1980 r i i • i • r i • i 1900" '1800' ' '1700' ' ' ' 1600' ' Year (10 year establishment intervals) Jacoby et al. (1985) very •ave. warm cool very • very warm warm • cool veFy cool warm cool V V ^ - 4- V- 4^  • ^>F~^ D"Arrigo and Jacoby (1992) very ! warm \ warm warm. I I very warm c o o \ cool cold warm warm cool . 1 cool , warminq ^ ^ W . ' 4, ^ 4 , ; D"Arrigo and Jacoby (1993) Northern ! strong ! warming warm warm "1 cool warm • i warm cool I Latitudes 4, 4, 4 , 4 , . 4" 4 ^ ^ Coppermine River I I strong gve. warming " 1 cool 1 warm cpol 1 H : Briffa etal. (1994) Sze i cz and MacDonald (1995b) Sze icz and MacDonald (1996) Instrumental Record Inuvik Airport watm eool warm • cool cool warm ^ 4? \^ 4 ^ ^ ^ ^ T • cool : JTI, dryirtg 1980 1900 1800 1700 • very • warm very very . cool cool cool warm cool • cool • 4- 4* . 4 ^ ^ ^^ ^^  P^l' I I moist : 1 1 1 dry ; moist moist dry i dry moist 1600 Figure 6.19 Comparison of establishment by layering and climatic reconstructions and instrumental records. See text for details of climatic information. 129 Tuktoyaktuk Peninsula and the Anderson River valley areas (X z=1.167, p=0.280). Table 6.1 The results of Chi-Square Tests of Independence between the decadal patterns of reconstructed climate and the decadal patterns of seedling establishment for the Tuktoyaktuk Peninsula and Anderson River valley areas (1770-1989). Establishment Climate Tuktoyaktuk Peninsula Anderson River Variables Likelihood Ratio Value p-value Likelihood Ratio Value p-value Temperature 0.229 0.632 0.832 0.362 Moisture 3.322 0.068 0.204 0.652 Temperature/Moisture 3.928 0.269 3.771 0.287 Table 6.2 Spearman's rho correlation coefficients between the patterns of reconstructed climate and the decadal patterns of seedling establishment in the Tuktoyaktuk Peninsula and Anderson River valley areas (1770-1949). Period of Tuktoyaktuk Peninsula Establishment Anderson River Valley Establishment Climate Averaging Spearman's rho p-value Spearman's rho p-value Temperature 10-year -0.222 0.247 0.019 0.920 20-year -0.045 0.832 0.014 0.946 30-year -0.073 0.730 -0.070 0.741 40-year -0.100 0.633 -0.098 0.640 50-year -0.166 0.582 -0.062 0.767 Moisture 10-year 0.098 0.607 -0.031 0.871 20-year 0.000 1.000 0.006 0.977 30-year -0.149 0.477 -0.078 0.711 40-year -0.185 0.376 -0.109 0.603 50-year -0.203 0.330 -0.244 0.241 There has been a lack of establishment by seed since 1960 at the Anderson River valley tree islands and no increase in the decadal seedling establishment at the Tuktoyaktuk Peninsula tree islands, although temperatures continued to increase through the latter part of 130 the 20th century. From the reconstructed temperature and moisture records (Szeicz and MacDonald, 1995b; Szeicz and MacDonald, 1996), approximately 1800 to 1880 spans a cold period within the Little Ice Age, and was relatively dry. After approximately 1880, reconstructed climatic conditions indicate a general warming trend and moist conditions prevailed from 1925 to 1940. From 1940 to 1960, moisture levels were still slightly above average. Combining the instrumental and the proxy data, 1920 to 1960 was a warm and moist period. Since 1960, temperatures have risen and precipitation levels have declined, indicating dry conditions particularly during the summer months. Three broad climatic periods are evident: the latter part of the Little Ice Age (1800 to 1879), early 20th century warming (1880-1959), and recent warm and dry conditions (1960-1990). Comparison of these three time periods for the Tuktoyaktuk Peninsula area indicated that the patterns of decadal seedling establishment were similar (x , Likelihood Ratio=0.712, p=0.700). However, for the Anderson River valley area, the decadal patterns of establishment were significantly different between the three time periods (%2, Likelihood Ratio=8.673, p=0.013). The period from 1880 to 1959 had high levels of establishment from seed in most decades, particularly between 1930 and 1959, when reconstructed and instrumental climate records indicated warm and moist conditions. Since 1960, when reconstructed and instrumental records indicate warm and dry conditions, there has been no individuals established from seed. 6.4 Discussion Including all sexually and vegetatively produced individuals for the Tuktoyaktuk Peninsula and Anderson River valley areas, the metapopulation of Picea glauca tree islands has been actively regenerating. Taking the tree islands individually, only two, tree islands TI and T3, 131 appeared to have healthy populations. In the metapopulation, high numbers of young individuals, less than 80 to 100 years of age, have resulted from active vegetative reproduction, particularly at the northernmost tree islands (TI, T3, A l ) . At all tree islands studied, the youngest individuals established by layering. The apparent decline in juveniles over the last two decades is likely an artifact inherent to layering. In the layering process, a live branch that is any number of years old is partially buried. The tip of the branch continues to grow, taking a more erect form and eventually adventitious roots develop at the base. At this point, it is considered to be an established individual. At the time of establishment, the individual will already be a number of years old. The age of the individual at establishment will reflect the age of the branch and the point along the branch at which it was buried (i.e., the amount of the tip left exposed to continue growing). The age of the individual does not coincide with the time of establishment. Therefore, establishment of individuals that will be dated to the 1990s and 1980s has not yet occurred, and this limits the number of individuals currently recorded from these decades. The tree islands in the Tuktoyaktuk region are being maintained over the longterm by either low or high levels of layering, depending on the site, and infrequent establishment by seed, which becomes most infrequent at the northernmost sites. This pattern is similar to that found for P i c e a m a r i a n a in northern Quebec at sites north and south of treeline where layering is the primary reproductive mode (Payette and Gagnon, 1979; Payette et al., 1985). In contrast, at sites near treeline in northern Quebec, and at sites near the Subarctic timberline in the N.W.T., where P i c e a g l a u c a relies mainly on sexual reproduction, establishment has been highly episodic, with high and low establishment periods closely following tree ring width chronologies and reconstructed climatic conditions (Payette, 1976; Payette and Gagnon, 1979; Payette and Filion, 1985; Szeicz and MacDonald, 1995a). 132 It appears from the age structures that the individual tree islands that are in the most uncertain state (as indicated by infrequent establishment and few younger individuals) are not those in which establishment by seed is most infrequent, but are those sites that do not exhibit active layering. It is in these tree islands that mortality may overtake establishment, making tree island death a more likely event. In other parts of northern North America, many P i c e a g l a u c a trees have been recorded older than 300 to 350 years, with individuals reaching 400-600 years in Alaska and the Mackenzie Delta (Critchfield, 1985). Trees nearly 1000 years old have been located above the Arctic Circle within the Forest-Tundra (Giddings, 1962). In this study, the oldest living individual was 433 years. This pattern of the oldest individuals of P i c e a g l a u c a found within the Forest-Tundra, rather than at their range limits, is consistent with that found by Szeicz and MacDonald (1995a) working along a Forest-Tundra/altitudinal range limit transect in the Mackenzie Mountains, N.W.T. In both the Tuktoyaktuk Peninsula and Anderson River valley areas, low levels of seedling establishment occurred in most decades. The few decades without establishment may reflect either lack of recruitment, establishment, or survival to sampling time (resulting in an underestimation of establishment levels). For the Tuktoyaktuk Peninsula, no significant differences in the decadal pattern of seedling establishment could be discerned between three key climatic periods: the end of the Little Ice Age (1800 to 1880); the beginning of the 20th century, and the latter part of the 20th century. However, along the Anderson River valley, there were significant differences in the decadal pattern of seedling establishment between the three time periods. This suggests that there was some positive response to the warm moist conditions in the first half of the 20th century, and that the lack of establishment in the latter part of the 20th century may be related to increasingly dry summer conditions. The 133 moisture condition of the seedbed has been indicated as the most important factor affecting germination and early P i c e a g l a u c a seedling survival (Rowe, 1970; Nienstaedt and Zasada, 1990). Thick humus and moss, such as found at most sites in the Tuktoyaktuk region, can easily dry out to below seedling rooting depths, resulting in high seedling mortality (Rowe, 1970; Nienstaedt and Zasada, 1990). However, the comparison of the three key time periods was weakened by the fact that only three decades were available in the most recent time period. In addition, correlations were not significantly different from zero for the decadal levels of seedling establishment for the two areas with the reconstructed temperature and moisture records for northwest Canada. This suggests that seedling establishment in the northern tree islands is not a direct response to variations in the regional climate conditions and/or there are other, more influential, variables. Over the period of record, the lack of a strong response in seedling establishment to a regional climatic regime may relate to several factors. Firstly, the favourable radial growth response of mature trees may not relate to conditions associated with establishment by seed. It may be reasonable to assume that an increase in growth of adult trees would result in an increase in production of viable seeds. However, the conditions promoting increased radial growth may not be favourable to the germination, successful establishment, and the survival of seedlings and saplings. At a P i c e a m a r i a n a site near the coast of Hudson Bay in northern Quebec, periods of establishment by seed, prior to the 20th century, broadly corresponded to periods of somewhat larger tree rings. However, 20th century warming, along with an increase in radial growth, has not translated into an increase in establishment by seed (Payette et al, 1985). Secondly, climatic parameters other than simple temperature or moisture may be important in recruitment and establishment of individuals from seed. These climatic parameters may include temperature and precipitation seasonality, and changes in the length 134 of the growing season (lengthened either in spring or fall). Thirdly, indirect effects of climate change, acting to increase competition from other plants (Stevens and Fox, 1991) and altering the frequency and/or severity of insect outbreaks (Briffa et al, 1998), may be limiting seedling establishment and survival. The factors determining the patterns of establishment by layering (Figure 6.19) are even less clear. The causes of layering events are complex, incorporating many factors promoting burial (e.g., moss growth, flooding and sediment deposition, sand mobilization and deposition, and heavy snow loads) and the presence of live branches near the surface. This makes it extremely difficult to determine any relation with the environment. It may be that 20th century warming has promoted moss growth and layering at sites where low summer precipitation levels are offset by other factors, such as ground ice melting or topographic position (base of slope). Rapid moss growth is a likely factor promoting numerous layering events at sites T l and T3. At site T l , a stump was excavated from organic soil in the centre of the tree island (Figure 5.10) and was dated to 380 BP +/- 70 BP (WAT# 2797; cal A D 1490, Stuiver and Reimer, 1993). This indicates that moss growth is rapid at T l and could promote the active layering observed. It also suggests that rapid burial of dead stems by moss has lead to the loss of dead stems for cross-dating, thereby limiting the extent of the tree population record and the determination of tree island age. Given that tree island persistence has been due primarily to layering, and tree islands, therefore, occur mainly in microsites where burial and layering are promoted, then tree islands are not areas where seedling recruitment would be favoured. Seedlings would not easily be recruited and survive in areas where burial processes are prominent. However, seedling and young tree survival is thought to be favoured within tree islands where microclimatic conditions are moderated compared to the open tundra (Wardle, 1993). Given 135 the contrary conditions necessary for seedling recruitment and for early survival and establishment, in conjunction with a low availability of propagules=(Chapter 5), it is not unexpected that in many northern treeline and Low Arctic areas, including this one, there has been little response detected by way of seedling establishment to 20th century warming (Elliott-Fisk, 1983; Payette et al, 1985; Scott et al, 1987b; Lavoie and Payette, 1994; Lescop-Sinclair and Payette, 1995; Szeicz and MacDonald, 1995a). C H A P T E R 7 136 T R E E R I N G C H R O N O L O G I E S 7.1 Introduction In the study of population dynamics in temperate and high latitude tree species, determination of age is necessary using annual radial rings of wood. This research requires cross-dating of ring patterns to determine: (i) exact years of annual ring formation in live trees to check for missing and false rings, and (ii) dates of establishment and death of dead trees. Typically, cross-dating is accomplished by using unusually narrow or wide rings as marker years. However, in high latitude and high altitude sites, ring width variation is greatly reduced. In these areas, other ring characteristics that mark particular years must also be used. In northern North America, light rings have been used to successfully cross-date coniferous trees (Oswalt, 1957; Jacoby, 1982; Filion et al, 1986; Delwaide et al, 1991; Yamaguchi et al, 1993; Szeicz and MacDonald, 1995a,b; Szeicz, 1996). Light rings are characterized by poorly developed, light coloured latewood cells, with low maximum density (Filion et al, 1986; Delwaide et al, 1991). In northeastern Canada, formation of light rings has been related to below normal late spring and entire growing season temperatures, particularly to years with a cool May, June, August, and September (Yamaguchi et al, 1993). In northwestern Canada, light ring formation has been related to low June, July, and August temperatures, particularly to years with a cool August (Szeicz, 1996). This chapter presents: (i) the light ring chronologies developed for specific tree islands, (ii) comparison of light ring chronologies with those developed in other areas of northern North America, and (iii) the longterm mortality and establishment patterns, and minimum age of tree islands in the Tuktoyaktuk region. 137 7.2 Methods At eight tree islands, located within six sites (Figure 7.1), increment cores or wood discs were taken from all individuals (live and dead) greater than 2 cm in diameter. Details of sampling methods and sample preparation were presented in Chapter 6. Skeletal plots were constructed and narrow, wide, light, frost, and pit (resin duct) rings were identified by qualitative evaluation. Initial attempts to cross-date dead samples using unusually narrow or wide rings as marker years were unsuccessful. Cross-dating using light rings was accomplished at a very late stage, limiting the number of sites to six of the eight available sites. The two sites not included in this study were site T5 (see Figure 6.1), which had eleven old weathered stumps, and site A2 (see Figure 6.1), which had unchecked light ring chronologies for the 27 dead individuals. Missing rings were identified using patterns of light rings from trees within a site and the light ring chronology developed by Szeicz (1996) for Picea glauca in the Campbell Dolomite Uplands, 15 km southeast of Inuvik. Light ring frequencies were calculated as the number of individuals exhibiting a light ring in a particular year, divided by the total number of individuals available in that year. Delwaide et al. (1991), Yamaguchi et al. (1993), and Szeicz (1996) used a threshold frequency of 5% for separating light ring years from non-light ring years. Therefore, in this study, years with light ring frequencies of 0 to 4% were considered non-light ring years and years with light ring frequencies of 5% or more were classified as light ring years. Years for which only one individual exhibited a light ring were not recorded as light ring years, regardless of the frequency level calculated. Light ring chronologies were constructed for six tree island sites (TI, T3, T6, A l , A3, HI), with chronologies for tree islands A l a and A l b and tree islands T6i and T6j combined. * An arbitrary threshold was used to determine whether a light ring year was strongly 138 Figure 7.1 The location of tree island sites in the Tuktoyaktuk region that were used in the study. 139 registered and represents a regional signal, or whether it was a poorly registered light ring year and reflects local conditions. If fewer than half of the sites registered a light ring in a particular year, or when all sites had light ring frequencies of less than 20%, the light ring year was considered to be a poorly registered light ring year. To detect any temporal pattern in the light ring chronologies, time series analyses were conducted using the strongly registered light rings for the six tree island sites alone and in combination with other available sites in northwest Canada (Szeicz, 1996). First order autocorrelations were conducted using 1 to 16 year lags (a = 0.05). The relations between regional climate and the widespread occurrence of light rings at the tree islands were examined. Differences between the average climatic conditions during strongly registered light ring years and non-light ring years were assessed using Student's 2-tailed t-tests for differences between means (a = 0.05). Light ring year frequencies for the region (the number of individuals exhibiting a light ring in a particular year, divided by the total number of individuals available in that year) were calculated and compared to climatic conditions using linear regression (a = 0.05) and correlation (Pearson correlation coefficient, a = 0.05). These tests were limited by the length of the homogenized instrumental record from Inuvik. Monthly, seasonal (winter, spring, summer, autumn), and annual average temperatures were available for 1929 to 1994 and total precipitation levels for 1959 to 1994. Using the light ring chronologies, cross-dating of dead stumps was considered successful for most samples from tree islands T l , T3, T5, HI and many samples from T6i (in total, 55 of 83 samples were successfully cross-dated). Other criteria aided, to a limited extent, the cross-dating of dead individuals. Initially, an indication of time since death was garnered by the presence of intact bark (death likely within last 100-150 years), small 140 branches (<0.5 cm, death likely within 20th century), fine branches (twigs, death likely within last 50 years), and dead needles (death within last few years). More importantly, for those dead individuals that established by layering or had produced another individual by layering, cross-dating was temporally constrained to the lifespan of the individual to which it was connected. Patterns of light ring years were used to cross-date the dead individual to this time period. In the dendrochronological literature, there were no readily accessible statistical tests that assess the strength of cross-dating that is based on light ring years with temporal constraints of connected individuals. Therefore, a subjective level of matching light ring years was developed to accept a sample as successfully cross-dated. This subjective level was five light ring years matching the chronology (or the individual to which it was connected) and greater than 50% of the light ring years exhibited by the individual had to be matched. For five of the 55 cross-dated individuals, less than five matching light ring years were used. A l l these individuals were under 115 years of age, connected to other dated individuals, and had at least 67% of their light rings match with those of the dated individuals. The remaining 50 individuals had between 6 and 20 matching light ring years. Six of these 50 cross-dated individuals were unconnected to another individual. For these six individuals, the number of matching light ring years ranged from 6 to 13, and all had bark intact and small branches, suggesting death within the last 100 years. Samples that did not meet these criteria (28 of 83 individuals) were not considered to be successfully cross-dated and were not included in the population age structure diagrams and analysis. If a sample from a dead individual was taken above the base due to rotting, the date of establishment was adjusted to the base using the methods outlined in Chapter 6 for live individuals. For samples with bark intact, the mortality date is the year after the last 141 complete ring. For weathered samples with bark and possibly outer rings removed, a correction in the date of mortality was made. Payette et al. (1989) estimated that the error in dating stem mortality may be 10 to 20 years and, therefore, added 20 years to the apparent ages for stems that died through the 16th, 17th, and 18th centuries. Szeicz and MacDonald (1995a) added a weathering loss correction of 20 years to samples from individuals that died in the 19th century and 10 years to those in the 20th century. In this study, the adjustment was made based on the severity of weathering (the amount of bark remaining or the absence of bark) and the century of death. Ten years were added to samples without bark that died in the 19th century; 20 years to those that died in the 18th century. A l l samples dated to the 20th century had bark intact. Mortality at these tree islands is reported in 10-year classes. The distribution of ages at death are presented, with 20 years added to the apparent age for samples not cross-dated and without bark. 7.3 Tree Ring Chronologies 7.3.1 Tree Island Chronologies Tree island T l had 18 live individuals available to develop a light ring chronology. Missing rings (1 to 5) occurred in 12 of the trees (67%). The first light ring in an individual occurred in 1802; the first light ring year in the chronology was 1871. Twenty-eight light rings in the chronology were identified (Figure 7.2); 24 of these occurred in the 20th century. Tree island T3 had 43 individuals available to develop a light ring chronology. Missing rings (1 to 8) occurred in 14 of the trees (32%). The first light ring in an individual occurred in 1601; the first light ring year in the chronology was 1770. Twenty-eight light rings in the chronology were identified (Figure 7.2); 20 of these occurred in the 20th century. At site T6, the light ring records from 15 live individuals from T6i and 5 live . , - i - CO CO CO i - Q Year i- i- t- < < i o T- co co 2 co Q h h H < < I O 1939 . . i - co co co ^ Q Year h- i- i- < < i o 1802 1800 1791 1790 1784 1783 1780 1778 1770 1752 1749 1747 1744 1735 1733 1729 1718 1716 1714 1712 1711 1710 1701 1699 1693 1690 1689 1685 1677 1670 1665 1663 1655 1643 1641 1631 1626 1623 1620 1619 1609 1605 1601 1592 1585 142 Light Ring Year Frequencies • 0-4% 5-19% I 20-49% • 50-100% Figure 7.2 Temporal distribution of light rings at six tree island sites (see Figure 7.1). The chronology for T1 ends at 1871, for A1 it ends at 1946, for A3 it ends at 1778, and for T3, T6, and H1 it ends at 1770. The light ring chronology developed by Szeicz (1996-labelled CDU) is also depicted. 143 individuals from T6j were combined into one chronology. Eight individuals (53%) had 1 to 4 missing rings. The first light ring identified in an individual occurred in 1709: the first light ring year in the chronology occurred in 1770. Forty-five light rings were identified (Figure 7.2), of which 34 occurred in the 20th century. At site A l , 6 live trees from tree island A l a and 12 from A l b were combined to form one chronology. Missing rings (1 to 2) occurred in 8 individuals (44%). The first light ring identified in a living tree occurred in 1922. The first light ring year in the chronology occurred in 1946 (Figure 7.2). Fourteen light ring years were identified. Tree island A3c had 19 live individuals available to form a chronology. Eight individuals (42%) had 1 to 3 missing rings. The first light ring in a live individual occurred in 1718. The first light ring year in the chronology occurred in 1744 (Figure 7.2). Sixty-one light ring years were identified, of which 36 occurred in the 20th century. Tree island H l a had 18 individuals available to form a chronology. Twelve of these trees (67%) had 1 to 12 missing rings. The first light ring in a live individual occurred in 1770, which is also the first light ring year in the chronology (Figure 7.2). Sixty-nine light ring years were identified between 1770 and 1993, with 32 light rings occurring in the 20th century. In total, there were 113 years between 1744 and 1995 in which a light ring is registered for at least one tree island site. There were 39 strongly registered light ring years, for which light rings were present in at least half of the tree island sites and with at least one site having had a light ring frequency >20%. At the 13 year interval, the first order autocorrelation was significantly different from zero (autocorrelation coefficient = 0.239, p = 0.004). 144 7.3.2 Comparison with Other Sites in Northwestern North America A recent compilation of light ring chronologies from a collection of Picea glauca sites in northwestern N.W.T. and the Yukon was available for comparison with the tree island sites in this study (Szeicz, 1996). One of the sites is located 15 km southeast of Inuvik, in the Campbell Dolomite Uplands, and is labelled C D U in Figure 7.2. C D U had 39 light ring years (frequency of 5% or greater) between 1770 and 1985. In comparison, T3, T6, A3, and HI had 26, 43, 53, and 67 light rings, respectively, over this same time period. During the 20th century, tree islands TI , T3, T6, A3, and HI had 24, 20, 34, 36, and 32 light ring years, respectively, compared to 16 for CDU. Good agreement in light ring years between the tree island sites and C D U occurred in 20 years from 1770. This sequence of strongly registered light rings is as follows: 1978, 1975,1974, 1969, 1959,1956, 1955, 1947,1935,1926, 1924, 1917, 1884, 1871, 1870, 1857, 1854, 1836, 1778, 1770. In eight other years, agreement occurred between C D U and only 2 tree island sites (1985, 1888, 1881, 1831, 1803, 1791, 1790, and 1783). As the tree island chronologies are expanded, these eight years may become more prominent. The sequence of 20 strongly registered light ring years was tested for autocorrelation. For 1 to 16 year time lags, the first order autocorrelation was not significantly different from zero (Table 7.1). Szeicz (1996) also presented light ring chronologies from five other sites in the Franklin, Mackenzie, Ogilvie, and Richardson Mountains, located to the southwest and southeast of C D U . Including the six tree island sites with the six sites from Szeicz (1996), the sequence of strongly registered light ring years is: 1969, 1959, 1947, 1924, 1884, 1871, 1870, 1836, 1778, and 1770. These ten strongly registered light ring years showed low autocorrelation coefficients (one year lag) that were not significantly different from zero (Table 7.1). 145 Table 7.1 Results of the times series analysis of the strongly registered light ring chronologies developed, (a) The 20-year chronology developed using the six tree island sites in this study and a published Forest-Tundra site near Inuvik (Szeicz, 1996) and (b) The 10-year chronology developed using the six tree islands and other published Forest-Tundra sites in northwestern Canada (Szeicz, 1996). (a) 20 year Chronology (b) 10 year Chronology Lag (years) Autocorrelation Coefficient p-value Autocorrelation Coefficient p-value 1 0.072 0.280 0.063 0.345 2 -0.039 0.470 -0.042 0.521 3 0.071 0.448 -0.042 0.633 4 0.015 0.607 -0.043 0.710 5 -0.040 0.687 -0.043 0.767 6 -0.040 0.749 -0.043 0.809 7 -0.040 0.796 -0.043 0.842 8 0.014 0.866 0.061 0.827 9 0.128 0.559 -0.039 0.861 10 -0.037 0.620 0.065 0.840 11 -0.038 0.674 -0.039 0.869 12 0.017 0.745 0.065 0.852 13 0.071 0.717 0.065 0.838 14 0.071 0.692 0.065 0.825 15 -0.039 0.731 -0.040 0.850 16 0.015 0.787 -0.041 0.872 Picea glauca and other coniferous tree sites in Alaska (Oswalt, 1957; Briffa et al., 1994; Jacoby and D'Arrigo, 1995) showed high frequencies of light ring formation in 1783, which is seen at four sites reported by Szeicz (1996) and occurs at tree islands A3 and HI . Comparison with a Picea mariana light ring chronology from Lac Bush Quebec (Yamaguchi et al, 1993), indicates little correspondence between northeastern and northwestern Canada. Matching years are 1978 (light ring frequency of 6.8% in northeastern Canada), 1969 (33.9%), 1956 (14.3%), 1904 (9%), 1884 (4.5%), 1836 (15.1%). None of the 146 light ring years prominent in northwestern Canada match prominent years in northeastern Canada. 7.3.3 Spatiotemporal Patterns and Causes Agreement in light ring years between tree island sites and other forest sites in northwestern North America indicates that regional climatic conditions influence the formation of light rings in Picea glauca. Lack of correspondence between light ring chronologies from northwestern and northeastern Canada reflects the differing synoptic conditions and climatic regimes (Bryson, 1966). A more complex pattern of dominance and mixing of Atlantic, Hudson Bay, Gulf of Mexico, Pacific, and Arctic air masses occurs in northeastern Canada, in contrast to the prevalence of Arctic and Pacific air masses in northwestern Canada (Bryson, 1966). Comparison of climatic conditions between strongly registered light ring and non-light ring years was conducted using data available from Inuvik Airport. Summer (June, July, August) and August average temperatures were significantly different in light ring years than in non-light ring years (Figure 7.3; p=0.009 and p=0.0003, respectively). August temperatures in strongly registered light ring years averaged 8.1°C compared with 10.8°C in non-light ring years. Summer temperatures averaged 10.5°C in strongly registered light ring years and 11.8°C in non-light ring years. A l l other months and seasons showed no significant difference in average temperatures between light ring and non-light ring years. Light ring year frequencies, calculated for the region as a whole, showed significant linear relations with August and summer temperatures (p<0.001). Over the period of record, as summer and August temperatures decline, a greater number of individuals in the region exhibit a light ring, resulting in higher light ring year frequencies. The correlations between light ring 147 o o o CM E co t CD O CO CD o Q CO < ^ CO >-o CD o CN CO CO c — Q E c § « ' S o * co > CO o ~ & Q . CD o CN IO 3 »1 D) 5 ^ o £ .22 (D O) :3 — Q) CO c *-*- c i» CD 3 — £.1 «? § . E 2 •S2 "5 * -C W CO . 3 OS W D) Q CO < c = . CO > ID O 3 ^ £ o> 5 -° c -o . 3 - ig CO t r « g o 0 Q •— • -c "O O £ C C . -( 0 2 ID O CO co 3 O) iZ (Snp|90 S99J69P) 9JniBJ9dUJ91 148 frequencies and both August and summer temperatures were also moderate (Pearson correlation coefficients of-0.493 and -0.461, respectively). May, annual, winter (December, January, February), and spring (March, April, May) average precipitation levels were significantly different in strongly registered light ring years compared to non-light ring years (p=0.001, p=0.022, p=0.032, and p=0.008, respectively). For strongly registered light ring years and non-light ring years, respectively, precipitation levels averaged 11.2 mm versus 26.9 mm in May, 294.7 mm versus 349.6 mm annually, 51.0 mm versus 69.1 mm in winter, and 46.0 mm versus 64.0 mm in spring. No other significant differences in precipitation were found between strongly registered light ring years and non-light ring years. Linear regressions between the light ring year frequencies and both annual and winter precipitation data showed significant linear relations (p=0.049, and p=0.011, respectively) and moderate correlations (Pearson correlation coefficients of -0.330 and -0.420, respectively). Szeicz (1996) found that light ring years for Picea glauca in the C D U site near Inuvik were strongly related to low summer growing season temperatures (June, July, August) recorded at the Inuvik Airport climate station and unrelated to any precipitation data. 7.4 M o r t a l i t y In each tree island, there were dead standing snags or stumps present, except at tree island A3b, at which all individuals were live. The number of dead individuals per island varied from 3 at site A3a, to 31 at site T6i, with an average of 17.4 dead individuals. The distribution of ages at death for the Tuktoyaktuk Peninsula sites and the Anderson River valley sites are shown in Figure 7.4. While dead individuals of great age (> 300 years) were present in each area, most of the stems were less than 300 years of age at death. For the 149 12 10 4 8 CD £ 6 -3-I Q Tuktoyaktuk Peninsula Sites • Anderson River Valley Sites • Horton River Valley Site H1 1 j I • | k l | 1 • 1| I • l| I B | f | f j f j I M Ij km lj I I | ! • j • l | [ • | • 111 |I1 | 40 80 120 160 200 240 280 320 360 400 440 480>500 Age at Death (20 year c lasses) Figure 7.4 The distribution of ages at death for trees in tree islands within the Tuktoyaktuk region. 150 Tuktoyaktuk Peninsula, many dead stems were between 40 and 120 years and between 180 and 300 years at death, with small peaks at 200 to 220 years and 240 to 260 years. For the Anderson River valley, large numbers of dead stems were between 60 and 260 years at death, with small peaks at 80 to 100,160 to 180, and 220 to 240 years. These distributions suggest that death was not simply a result of advanced age, but occurred at a variety of ages and may be related to periods of suboptimal environmental conditions. At the Tuktoyaktuk Peninsula sites most dead trees were cross-dated (Figure 7.5). Three of five dead trees were successfully cross-dated at tree island T l , twenty-one of twenty-seven dead trees at T3, eight often dead trees at tree island T5, and thirteen of thirty-one dead trees at T6i. In total, 45 dead trees have been cross-dated for the Tuktoyaktuk Peninsula, with at least 22 of these individuals (49%) originating by layering. Of the dead individuals cross-dated, one died in the 1730s, one in the 1750s, and the remaining individuals died since 1790 (Figure 7.6a). Small peaks in mortality occurred at 1800 to 1810, 1920 to 1940, 1960 to 1970, and 1980 to 1995. It is unclear i f the more recent peaks in mortality are simply related to the lack of time for decomposition or burial of dead remains. At tree island Hla , ten dead trees were cross-dated. One individual died in the 1970s, two in the 1920s and 1940s, three in the late 1870s and 1880s, and one in the 1830s, similar to Tuktoyaktuk Peninsula The oldest mortality event occurred in the 1690s (Figure 7.6b). Mortality patterns were determined from cross-dated remains at two sites located near the Subarctic timberline in the Franklin and Ogilvie Mountains of northwest Canada (Szeicz and MacDonald, 1995a). Above timberline, all sampled trees had died in the 19th century, mainly between 1800 and 1860. In contrast, within the upper Forest-Tundra, mortality had been fairly consistent since 1800. These results provide evidence that mortality has been episodic in areas near the physiological limits of a species. This episodic mortality may o o : to CO o CO CO CD o : o C O k_ J C O : cn c . JC CO : oo t 6S S" 55 00 CO CM I CO CD O o •§ co Y 151 CO CD co 5 o c— . — C ffl 3 > CO CO pf f l a j - c ^ ^ co co ^ W 1— "6 T3 >• 3 CO ^ O H - T , r - 'C CD 00 ^  JO JO <g w c n "D C W> CD ffl "c ffl ffl 52 8 5 >,£ 5 c h "* ffl CD >> 00 _ L o>2 c? CO CD .!= C D ^ - O < » • £ . T J £ - „ , *- O w 3 iz g v - $ CO j= v« T3 co = T3 « ^ 2 5 CD ffl £ •K CO 3 Q-o ffl CO 3 . 2 § co § C X J > , £ S . y rn (—•*-'•— O CO c To - E-c 3 O .P C Sc CL CO ffl — *- S? 1 n a) o o i - 2 CO ffl ^ ffl H ffl O i 8y CO CO S ^ g ^ c o - ^ i E -c -5 5 o o - o ^ c ? c ? = 6 > o | 5 ^ — c cr o co Q 4 5 £ ffl- 3-g.c CO ffl ffl -_ CO CO ? O c ffl 5 r ° CO •j^ crk. p, c co co o> CO £ 5 T5.S O C CD O 3 — o 5 a>-o « o >;X3 5 o ? t i n co = ^  o - "5 o O ffl c ±= o ^ o o CD S J Poo - S wro ° = > < 0 co cnO c LO CD 3 cn CO $ 00 _ . ? ffl CO ffl c O f f l r O « 1 ffl CO « 152 (b) CD 4-- Q E • Z 2-1980 1940 1900 1860 1820 1780 Year AD (10 year classes) 1740 1700 ~[—•—i-•—r i n—•—i—•—i—•—r 1980 1940 1900 1860 1820 1780 1740 Year AD (10 year classes) 1700 Figure 7.6 Mortality (number of dead trees per decade) (a) at tree island sites along the Tuktoyaktuk Peninsula, and (b) at H1a within the Horton River Valley. 153 reflect environmental variation, such as the climatic deterioration during the latter stages of the Little Ice Age (early 1800s to approximately 1880). 7.5 Establishment Along the Tuktoyaktuk Peninsula Successful cross-dating of the 45 dead individuals of the Tuktoyaktuk Peninsula extended the establishment record for each tree island. At TI , two of the dead trees were dated as establishing in the 1670s and 1550s (Figure 7.7) and these individuals originated from seed. Twenty-one dead individuals from tree island T3 established between 1200 and 1950 (Figure 7.8). Several individuals established in the 1660s, 1750s, 1920s and 1930s. Older individuals had established in the 1550s, 1420s, 1350s, and 1200s. Between 1800 and 1900, two individuals originated by layering, two from seed. The oldest individual from layering established in the 1690s. A l l other individuals, established prior to 1800, originated from seed. Within the 20th century, all dead individuals originated by layering. At site T5, eight dead individuals augmented the establishment record (Figure 7.9). Two individuals originating from layering, established in the 1840s and 1890s, one individual originating from seed established in the 1730s, and one in the 1670s. In the 16th century, three individuals had established, one from seed origin and two from layering. The oldest individual dated to the 1350s and originated from seed. At tree island T6i, 12 dead individuals have extended the establishment record (Figure 7.10). Two dead individuals originating from layering established in the 19th century. Two individuals originating from seed established around 1700. Five individuals established between 1630 and 1660, four of which originated from layering. The three oldest individuals established in the 1540s, 1270s and 1240s, all originating from seed. For the Tuktoyaktuk Peninsula as a whole, establishment in tree islands has occurred 28 -26 -24 -22 -20 -18 -CD -Q 16 -E 14 -z 12 -10 -8 -6 -4 -2 -o -L Site T1 I3 Live Individual, Layered Origin [2 Live Individual, Seed Origin [2 D e a d Individual, Layered Origin | Dead Individual, Seed Origin 4 n n i ^ ffi i • i • p i • r i 1 i • i 1 i • i 1 i ,BI 1980 1600 Figure 7.7 1900 1800 1700 Year (10 year establishment intervals) Establishment of Picea glauca, including currently live and dead individuals, by vegetative and sexual reproduction at site T1 (see Figure 7.1). 155 C c "cn s i ° T5 CD ca CO •g '> T3 C CD > O s T3 0) T3 CD CD CO re 3 -q > C <D > re •g '> T3 c T3 re CD Q c cn CD CD CO ro 3 •g > c •o ro CD Q IZ3 • • CO H 0 —^• CO jy / / / ///'//'//// / 7-7—7 mz: v ; / ; / 7-7-r 1—1—r 1—1—1—1—r o e o ( 0 - j c \ i o c o ( o < f o j o c t i ( o n W t \ I W N W T - ^ T - T - T -CM O , o CM -I CC X CD W "D C CO CD > CC -*-* CD cn CD > .Q _c/> CC =3 •g '> C CO LU 00 0) J_ 3 Ol CD jeqiunisi "D S> CD ro ro 3 > "D C CD > C cn £ O c •a CD CD W ro 3 > T3 0 > •o CD >* CO CO 3 •g > C "O CO CD D c 'cn CD CD CO ro 3 TJ > C T J CO CD Q -I i_8 LO rz] • • LO H CD CO i_8 CO m / / \ w / / / n — . i i r T — r OCOIO'<t(MOCOCDrf COCMCVJCMCMCVJi-i-i-T—i—i—r O 00 CO tf o . o C\J o . o co CO CD CO co ro o ro CD Q I- 2 o *— 9 rzj - oo o CO CD CC 3 X CD CO T3 C CO > CO -*—' CD CD CD > _Q tf) CO 3 •g > C " D CO CD T3 "O C CO CD > c CD i_ 3 O CD C ' O CD = 3 8 i f Is O ) CO LO co CD, *5 - c o CD C W 3 CO -4—• co UJ CL CD CD J-3 CJ) o . o C\J c CO c CO - •£ ° CO "D >» ro CO 3 •g > T J c 0 > 0 0 CO ro 3 •g > T J c 0 > "D £ 0 ro ro 3 g '> T J _c "D ro 0 Q c CO T J 0 0 CO ro 3 •g > TJ _tz TJ ro 0 Q IZI • • CO h-CD •+—• CO o . o CO I I I I "i— i—r O C O C O t f C M O O O t O ' t f n ( M C M ( M C \ W i - i - r -I I I I o oo co tf ~~r~ jequjn|\| CO 3 X CD CO X5 c CO CD > 3 in O i SO 0-CD 3 i i 158 continuously since 1530, with numbers increasing after 1730 and peaking in the 1930s (Figure 7.11). Further cross-dating may increase numbers prior to this time. A l l dead individuals that established since 1890 originated from layering, except for the two seedlings located near tree island T6j. Subsurface connections remained intact for many layered individuals that had established within the 16th and 17th centuries at tree islands T5 and T6i. This indicates that, at some sites, longterm preservation of subsurface connections is possible. However, for the dead individuals at tree islands TI and T3, subsurface connections indicative of layering were found only for individuals established in the 19th and 20th centuries, although live layered individuals at T3 occur as far back as the 16th century. For tree island Hla , four dead individuals that originated from layering established in the 18th century (Figure 7.11). Four individuals originating from seed and one originating from layering established in the 16th and 17th centuries. Currently, the establishment record extends back to the 1420s. Including live and dead individuals from Tuktoyaktuk Peninsula, periods of favourable establishment by layering were the 1630s, 1750 to 1780, and since 1800. In the 19th and 20th centuries, higher numbers of layered individuals occurred after 1840, and between 1920 and 1960. These periods may relate to the onset of warmer, moister conditions at the end of the Little Ice Age and warm and moist conditions between 1920 and 1960 (Szeicz and MacDonald, 1995b; Szeicz and MacDonald, 1996). A warm and moist climate may have created favourable conditions for layering events, such as growth of healthy branches near the surface and increased moss growth. After 1960, annual and summer temperatures recorded at Inuvik airport continued to rise, while total annual and summer precipitation levels declined (Figures 6.14 to 6.17) creating dry conditions. Suboptimal tree growth, few healthy branches, and a reduction in moss growth may have ensued, limiting S S O f l O o> _ c -•n "8° a> u 2 T CD <D "D Q) ><& CO <0 _| CO —I CO CO CO CO CO 3 3 3 3 T I n S 2 > > > > =6 =6 T3TJ C C c c "D "D CD CD CO CO > > <D CD • • Q Q • • • • to 159 c CD E 03 CO UJ =3 O ZJ H jequjn[\| 160 O C O C O ^ - C M O C O C O ^ r C V J O C O C O t C O C N J C M C M C M C M i - T - i - T - i -1-jeqiunN 161 potential layering events. For currently live and dead individuals along the Tuktoyaktuk Peninsula, periods of favourable establishment from seed were 1730 to 1780,1810 to 1830, and 1840 to 1890. The first period may reflect the warm moist conditions of the 18th century (Szeicz and MacDonald, 1995b, 1996; Jacoby and D'Arrigo, 1995), the latter two periods generally preceded and followed the cold, dry period of the early to mid 19th century. The conditions of the 20th century, particularly warm and dry since the 1960s, appear to have precluded establishment from seed at these tree islands. However, given the lack of statistically significant correlation to reconstructed climatic conditions, these periods of favourable establishment from seed may be better related to site specific conditions acting to limit production and germination of viable seed, recruitment of seedlings, and/or survival of young individuals. 7.6 Tree Island Age The determination of establishment dates for dead stems increased the age of each tree island. At site T l , tree ring records indicate that the tree island originated more than 450 years ago. The radiocarbon date of 380 +/- 70 BP (WAT # 2797; cal A D 1490, Stuiver and Reimer, 1993) was obtained from the outer rings of an excavated stump in the island (Figure 5.10). Given that the stump was at least 20 cm in diameter, this would place establishment at least 100 to 200 years prior to that date, most likely within the 14th or 13th centuries. At site T2, two in situ stumps excavated from the slope and radiocarbon dated, produced dates of 1570 +/- 70 BP and 1540 +/- 70 BP (WAT # 2996 and 2997; cal A D (430, 450, 540) and cal A D 540, Stuiver and Reimer, 1993). Tree ring records at site T3 indicate that the tree island is more than 790 years old, dating to the 1200s. Tree island T5 dates to the 1350s and tree 162 island T6i dates to the 1240s. Thus, all these tree islands originated prior to the Little Ice Age (ca. A D 1400 to 1850) and at least to the latter part of the Little Climatic Optimum (ca. A D 1000 to 1350). For the Anderson River valley area, the tree ring records from live trees indicate that the tree islands date back to the 1600s. One radiocarbon date was determined for the outer rings of in situ dead remains located at A2 (Figure 7.1). This stump dated to 700 +/- 70 BP (WAT # 2998; cal A D 1290, Stuiver and Reimer, 1993). Given the size of the stump was approximately 25 cm diameter, establishment would be 200 to 300 years earlier, most likely within the 10th or 11th century, near the onset of the Little Climatic Optimum. At site H la , cross-dating of dead remains extends the age of the tree island back to the 1420s, near the onset of the Little Ice Age. 7.7 Conclusions Light ring chronologies provide an alternative method for cross-dating samples from slow growing individuals and krummholz, which exhibit narrow, low variation ring widths. Agreement between strongly registered light ring years at tree islands in the Low Arctic Shrub Tundra and those from sites within the High Subarctic Forest-Tundra indicates the influence of regional climate in light ring formation. For the tree islands, the production of dark wood appears to be limited in dry years, particularly by dry winter and spring conditions, and by low summer temperatures, particularly by cool Augusts. In addition, the good agreement between Forest-Tundra trees and the smaller individuals and krummholz of the tree islands, indicates that light ring chronologies from tree islands can be used to expand the network of tree ring sites northward into the Low Arctic. Tree islands may be a valuable component in summer temperature and winter/spring precipitation reconstructions, 163 particularly since few sources of high resolution proxy climatic data occur within the Low Arctic (Johnstone and Henry 1997). Numerous dead P i c e a g l a u c a in the Tuktoyaktuk region resulted from mortality events that included trees of various ages. At the Tuktoyaktuk Peninsula sites, higher mortality has occurred since 1920. However, this may relate to the lack of time for decomposition of recently dead stems. Establishment occurred in most decades somewhere within the region, and peaks in establishment were relatively small. This indicates the importance of site-specific factors in seed production, the successful recmitment of seedlings, and establishment. The combination of tree ring records and radiocarbon dates places these tree islands back to the latter part of the Little Climatic Optimum. 164 C H A P T E R 8 THESIS C O N C L U S I O N S 8.1 Summary The northernmost conifers located within the Tuktoyaktuk region are P i c e a g l a u c a , distributed in small groups or islands and surrounded by shrub tundra. The tree islands along the Tuktoyaktuk Peninsula are very isolated, with only one site composed of several tree islands. Along the Anderson and Horton River valleys, many tree islands occur at each site. In all areas, the tree islands occur in favourable microclimatic locations: slopes with southerly aspects and in protected valleys. Tree islands located on slopes with northerly, easterly, or westerly aspects occur within protected river or dry valleys. Typically, the tree islands occur on moist organic soils, with a hummocky micro-relief. However, they are also found on well-drained gravelly sand substrates, possibly of fluvio-glacial origin. Active layer depths varied widely both inside and outside the tree islands, from less than 0.5 m to greater than 1.0 m. This likely reflects the patchy distribution of trees within the tree islands, and the fact that the sampling spatial resolution appeared to be insufficient to capture the micro-environmental influence of individual tree canopies on active layer depths. At all but one of the tree island sites, male cones were present on most trees, and many had female cones. Combining 1993 and 1994 data, there were fewer individuals producing cones with increasing distance north of treeline. The presence of bisexual cones and other cone abnormalities at many tree islands were indicative of the physiological limitations experienced by these trees. The germinability of seeds produced at these tree islands was very low in all years. Germinability increased at sites southward to treeline along the Tuktoyaktuk Peninsula and Anderson River valley, although levels were still much lower 165 than those reported for the northern Boreal Forest. No significant differences were detected in seed germinability between years. Across the Tuktoyaktuk region, only two seedlings were found at tree island sites, and both were many metres from live trees. Seedling ages were estimated by nodal counts as 13 and 30 years in 1995. Transplanted seedlings showed low survivorship levels, varying from 20 to 45% at the end of the second growth season, with no significant differences inside versus outside the tree islands. At each tree island, a large percentage of the population originated by layering, from 33 to 90% of live and dead individuals. These numbers are regarded as minimum estimates since subsurface connections may decompose in older and dead individuals, which were numerous at most sampled tree islands. Tree islands in the Tuktoyaktuk region rely on vegetative reproduction for population maintenance. With increasing distance north of treeline, tree island maintenance is increasingly dominated by vegetative reproduction. Only a few individual tree islands have a healthy population structure, with numerous individuals in the younger age classes. These tree islands are the northernmost ones, and all young individuals had established by layering. This reflects conditions conducive to layering events, such as rapid moss growth and healthy branches reaching the surface, and may be related to climate only indirectly. Most tree islands were dominated by older trees. A l l tree islands had one to a few individuals within most decadal age classes over the period of record. Grouping the tree islands by area, the Tuktoyaktuk Peninsula and Anderson River valley areas have healthy metapopulation age structures. Numerous individuals established by layering, particularly in the 20th century. For the Tuktoyaktuk Peninsula and Anderson River valley areas, establishment from seed occurred in most decades. No decade stands out as particularly favourable for establishment by seed and there was no correlation between 166 establishment from seed and reconstructed climate records. The continuous, low level establishment for the region reflects the infrequent and individualistic nature of establishment at each tree island site. The decadal establishment pattern and the lack of correlation with climate suggest the importance of site-specific factors, rather than regional climate conditions, in determining establishment from seed and survival of these individuals to present. This conclusion appears to be strengthened by the maintenance of this establishment partem into the 20th century, particularly for the Tuktoyaktuk Peninsula, despite the general warming trend. The patterns of establishment by the two reproductive modes suggests that longterm survival of individual tree islands depends on site-specific conditions promoting vegetative reproduction, rather than sexual reproduction. Light ring chronologies were developed from live trees at six tree island sites. Comparison between these sites and a site near Inuvik indicates good agreement in 20 light ring years since 1770. Ten of these light ring years were also strongly registered at sites further west into the Yukon Territory and Alaska. Little correspondence occurred between prominent light ring years in northwestern Canada and northeastern Canada. Light ring years at tree island sites agree well with years of low summer temperatures and low annual, winter, and spring precipitation levels. These light ring chronologies were used to cross-date dead trees at four tree island sites. In conjunction with radiocarbon dates of wood remains at several sites, tree island records extend back to pre-Little Ice Age times. The oldest tree island, site T3, dates by tree rings to 1200, while radiocarbon dates of buried stems at site T2 date the presence of this tree island to the 5th century. However, complete destruction of tree island T2 and significant alteration of the sandy deposits would be necessary to construct a record that spans the period from the 5th century through to the present live trees. In addition, it is not known if the 167 record would be continuous through to the present. Along the Anderson River valley, a radiocarbon date from surface wood remains extends the record of tree islands at least to the 13th century. 8.2 Contributions to the Tree Island Hypotheses In Chapter 1, three alternative hypotheses were advanced concerning the nature of P i c e a g l a u c a tree islands within the Tuktoyaktuk region. It is clear from this research that the tree islands are far older than the onset of recent warming, indicating that they are not of recent origin. Therefore, Hypothesis (i) can be discarded. Indeed, there has been no striking response in seedling establishment to post-Little Ice Age climatic warming. Hypothesis (ii) (tree islands are a normal component in the range of P i c e a glauca) and Hypothesis (iii) (tree islands are relicts) are difficult to differentiate. Under both hypotheses, tree islands would occur in climatically favourable microsites allowing tree survival in the Low Arctic, and they would be older than 150 to 200 years. For Hypothesis (ii), some authors believe that equilibrium with climatic conditions is shown by current or recent establishment of seedlings (e.g., Elliott-Fisk, 1983). The idea of sexual reproductive e q u i l i b r i u m with climatic conditions is difficult to define and measure. However, given that trees are long-lived, it is'difficult to state how often seedlings need to establish and survive to consider that a population is in equilibrium with climate. According to this hypothesis, tree islands should be maintained primarily by sexual reproduction, with establishment of seedlings varying closely with climatic conditions. However, the scale of climatic fluctuation, which seedling establishment should reflect, is also unclear as climate varies at many time scales. The current lack of establishment from seed is not necessarily evidence against Hypothesis (ii), rather the longterm record of establishment should be considered. 168 In the Tuktoyaktuk region, the tree islands are maintained primarily by vegetative reproduction, with layering most prominent at the northernmost tree islands. Sexually produced individuals established infrequently at each individual tree island. For the Tuktoyaktuk region, establishment by seed of a few individuals occurred in most decades, with very little variation in numbers between decades. There was no synchronous establishment from seed between sites during climatically favourable decades, and there were no correlations between reconstructed climate and establishment from seed. It does not appear that successful sexual reproduction is sensitively tracking regional climate change. Additional evidence against Hypothesis (ii) is that there appears to be no younger tree islands. If tree islands were a normal component in the range of P i c e a g l a u c a , then establishment of tree islands should be ongoing, with tree islands clearly of a variety of ages. Concerning Hypothesis (iii), i f the tree islands were relicts of more favourable climatic conditions in the past, when treeline had been further north, then perhaps a greater density of dead trees and stumps would be expected. The Tuktoyaktuk Peninsula and the Anderson River valley differ in the number of live and dead individuals and tree islands present at sites within each area. The prominence of dead individuals may relate to: (a) the form of advance, as it may have been either a treeline or a species limit advance in response to favourable climate conditions in the past, and (b) the time since the last advance and retreat of treeline or the tree species limit. The two forms of advance would result in very different areal cover of trees at the time of the advance, and therefore different availability of material to be preserved to the time of sampling. Therefore, the availability of dead trees and stumps affects the interpretation of whether the tree islands are relicts of a treeline or a species range limit advance, and the time since the advance (if it differs between areas). These ideas are considered further in Section 8.3, below. 169 The results from this research support Hypothesis (iii): these tree islands are relicts of more favourable climatic conditions in the past, when either treeline and/or the species range extended further northward. A l l tree islands examined in detail are of great age and rely on vegetative reproduction for population maintenance. The pattern of seedling establishment within the region appears to be stochastic, with infrequent seedling establishment at each tree island that was not synchronous among sites. Given the current lack of seedling establishment, low germinable seed production, and lack of a seedbank, it is unlikely that the tree islands could re-establish if destroyed, or that new tree islands could establish elsewhere in the Tuktoyaktuk region under current environmental conditions. It is not known to what extent there is input of seed from southern (treeline) sources, which in this area would be 15 to 50 km away. Given this distance, external seed inputs would be low, thereby limiting the potential for re-establishment. 8.3 The Role of Tree Islands in Reconstructing Range Limits and Treeline Fluctuations The definition of latitudinal treeline incorporates both a tree morphology component (at least 3 m in height) and an areal tree cover component (tree to tundra ratio of 1:1000) (Timoney et al, 1992). In order to document treeline fluctuation, evidence for changes in both of these parameters must be found {e.g., Lavoie and Payette, 1994; Payette and Gagnon, 1985). The northern range limit is defined by the northernmost extent of the species, regardless of areal cover or tree size. Changes in species range limits require determination of the northern extent of the species at different times in the past. 170 8.3.1 Past Fluctuations: Tuktoyaktuk Peninsula The tree islands of the Tuktoyaktuk Peninsula are isolated and of great age. These tree islands may be (i) remnants from a time when tree cover was greater, or (ii) they may have established in isolated, favourable microsites and persisted without changes in tree island density. In the first case, they may be remnants of treeline advance either during the Hypsithermal (prior to 3500 BP), or they may reflect an undocumented treeline advance between approximately 3500 BP and the Little Climatic Optimum. Treeline advance during the Hypsithermal is well documented by pollen and macrofossil records (Ritchie, 1984). However, there is a lack of evidence in the pollen record for a more recent treeline advance. Examination of live tree islands along the Tuktoyaktuk Peninsula is insufficient to determine treeline fluctuations given the areal cover component in the definition of treeline. In order to document any post-Hypsithermal treeline advance, radiocarbon dates of tree species wood remains from numerous, widely distributed sites that are currently tundra, would be needed. In the second case, the tree islands would reflect a species limit advance between approximately 3500 BP and the Little Ice Age. During this time, the number of tree islands may have been similar to present or greater, but lacked the growth form and areal cover necessary to meet the definition of treeline. Clearly, the dates of origin need to be determined for the tree islands. However, my research showed that the tree islands are of great age, introducing the problem of decompositional loss of tree remains. It appears that tree island age may only be determined by radiocarbon dating of wood fragments located within tree island soil (e.g., Payette and Gagnon, 1985). In conclusion, for the Tuktoyaktuk Peninsula, indications of treeline or species limit advances between the Hypsithermal and the Little Ice Age awaits future work on the location and radiocarbon dating of buried wood remains. 171 8.3.2 Past Fluctuations: Anderson and Horton River Valleys At the Anderson and Horton River valley sites, analyses of samples from several tree islands indicated that they are of great age and mortality has occurred in individuals of all age classes. The prominence of dead tree islands and stumps indicates either: (i) greater tree island density and areal cover in the recent past, or (ii) an equilibrium tree island density over time as tree islands establish and die. These locations provide a better spatial record of tree island establishment and demise over the last 1000 years, and there is a greater possibility of determining any post-Hypsithermal treeline or species limit advance. Cross-dating the numerous live and dead P i c e a g l a u c a will provide greater insight into longterm population decline leading to retreat. Analyses of decadal seedling establishment suggested some correspondence with reconstructed and instrumental climate. These areas may provide the sample sizes needed to model population age structures and relate the departures to the climate reconstructions. 8.3.3 Future Vegetation Change Expansion of a species' range may result from a slow, continuous extension of a species' borders (e.g., Delcourt and Delcourt, 1987), or it may occur at a much faster rate as a consequence of outlying populations acting as expansion nuclei (e.g., Bennet, 1985; Cwynar and MacDonald, 1985; Prentice, 1986; MacDonald et al, 1993) If P i c e a g l a u c a tree islands can act as nuclei for expansion, then there is the possibility that treeline could advance quickly under rapidly ameliorating climatic conditions. For a tree island to act as a nucleus for population expansion, it would be necessary for these trees to produce large amounts of viable seed that would develop into competitive seedlings and survive to become reproductive adults. Seed germinability tests indicate that although viable seed is present at 172 tree island sites, it is being produced at very low levels. Work in northern Quebec suggested that krummholz produce low or negligible levels of germinable seed, and that climate amelioration must first stimulate a vertical growth response transforming krummholz into upright trees, before abundant germinable seed is produced (Payette et al, 1982; Payette and Gagnon, 1985; Arseneault and Payette, 1992). Even if there is seed input by long distance transport from southern sources, there is strong circumstantial evidence that current conditions do not favour establishment and survival of seedlings within tundra communities. In the Tuktoyaktuk region, only two seedlings were located at or near tree island sites, and survival of transplanted seedlings was very low. 8.4 Future Research 8.4.1 Work in Progress Other studies, that were beyond this research endeavour, may be conducted using samples already collected. In addition, some new questions that arose from my study may be addressed, in part, by samples already collected. Given the opportunity and the isolated location of these sites, extra samples were taken in anticipation of further study. For samples that do not show reaction wood, ring widths will be measured in order to gain more information about yearly variations in growth conditions, to aid cross-dating of difficult samples that could not be confidently cross-dated using light ring chronologies alone, and to determine if missing ring years correspond between trees and sites, possibly reflecting regional climatic conditions. The potential use of select samples from tree islands in producing ring width chronologies will be evaluated. The numerous tree islands at T6 will allow comparison of tree islands situated in very different microsites, with different substrates, aspects, and topographic positions, and there are two species present at this site. \ . • 173 These differences may be reflected in varying tree ring and establishment patterns. Along the Anderson and Horton River valleys, additional wood samples will allow augmentation of light ring chronologies for these areas and development of tree ring width chronologies. At site A2, there appears to have been a long period of decline in tree islands, with a resulting change from, what appears to have been, a low stature, clumped forest stand to the current open network of live and dead tree islands. This site will allow an examination of the longterm dynamics of tree island establishment and demise. Farther south along the Anderson and Horton River valleys, wood samples from trees at sites near treeline and within the Forest-Tundra have been made available from Greg Henry and Stephan Kesting. Tree ring chronologies will be produced from these sites, allowing a comparison of responses to climatic conditions between erect P i c e a g l a u c a trees within the High Subarctic and small P i c e a g l a u c a within the Low Arctic. 8.4.2 Future Research Endeavours My research has produced many questions that need to be addressed. The following projects are suggested to further the understanding of conifers at their northern limits. i. Microsite and Tree Health Measurement It is believed that tree islands are located in favourable sites that allow the survival of conifers in a stunted form and that differ from surrounding areas. Microclimatic conditions produced by the presence of trees (e.g., reduced wind speed, increased snow depth) would further promote tree survival in the Low Arctic. In order to determine whether the microsite conditions are more favourable at the northernmost tree islands, and the extent to which tree islands may modify the microenvironment, measurement of year-round environmental 174 conditions should be made over several years. Variables should include: air temperature, wind velocities and directions, snow depths and distribution, active layer depths, soil temperatures, soil nutrients, and soil moisture. Moss thickness should be measured annually. These measurements should be made within and outside tree islands and in areas distant from the tree islands in the open tundra but that have similar topographic positions and substrate conditions. Quantitative measures of tree health at the tree islands should include: shoot elongation, shoot production, loss, and position, needle loss, chlorosis, and desiccation. ii. Sexual Reproduction, Establishment and Growth I established two transects along the Tuktoyaktuk Peninsula and Anderson River valley in order to examine changes in seed germinability from the northern Forest-Tundra to the northernmost tree islands. In future work, I would like to expand these transects southward to the northern Boreal Forest in order to capture the changes, gradual or abrupt, in reproduction, establishment patterns, and tree growth. a) Sexual Reproduction As indicated in Chapter 5, a more appropriate measure of sexual reproductive potential is to estimate germinable seed production. This would entail measuring tree density, estimating the number of cones per tree and the number of seeds per cone, and determining the germinability of seeds. Longterm measures would clarify temporal and spatial variations in germinable seed production that have been suggested in my results and that could be related to changing climatic conditions. b) Seedling Recruitment and Establishment South to north changes in seedling densities may reflect changes in (i) germinable seed production, limiting potential seedlings, (ii) competition from ground vegetation, limiting 175 recruitment and initial survival of seedlings, and (iii) climatic conditions, limiting seedling survival. Across the Forest-Tundra, the composition of ground vegetation changes gradually from boreal-dominated in the south to tundra-dominated in the north (Morisset et al, 1983; Kesting, 1996; Henry et al, in preparation). Changes in seedling densities may reflect competition stress as ground vegetation composition changes. Changes along the transects in seedling densities, ground vegetation composition and abundance, and moss thickness should be determined. In order to examine the importance of the above three factors on seedling establishment and densities, two types of field experiments need to be initiated along the transects: seeding and seedling transplants. The seeding experiment would involve direct seeding at each site using seeds collected from Inuvik, with a known germinable level. Other possible treatments include seeding onto competition reduced sites (vegetation thinning) and seeding onto competition eliminated sites (vegetation removed). Transplant studies, initiated in the present work, need to be expanded along the length of the transect. Similarly, other treatments may be included, such as transplanting seedlings into competition reduced (ground vegetation thinning) or competition eliminated (ground vegetation removal) sites. c) Longterm Establishment Age determination of live and dead trees, sampled along the expanded transects, will allow examination of longterm establishment patterns that may vary from the northern Boreal Forest to the tree islands. d) Tree Growth Changes in tree densities (measured for determining germinable seed production), tree heights, and radial growth rates (measured from wood samples taken for determining tree ages) along the transects will help determine where and how rapidly vegetative growth 176 declines in response to climatic deterioration from south to north. Comparison of the northward decline in tree growth to the decline in sexual reproduction may indicate similar or differing northern limits. iii. Tree Island Densities A strong contrast exists in the numbers of tree islands between major areas of the Tuktoyaktuk region. Most sites along Tuktoyaktuk Peninsula have very isolated single tree islands. The Anderson and Horton River valleys have more tree islands at each site. This suggests that either: (a) Tuktoyaktuk Peninsula has been in a longer period of tree island demise, during which remains have decomposed or been buried, or (b) the area was never densely treed. Extensive search for evidence of dead tree islands would be helpful to distinguish between these two possibilities. At select sites (south-facing slopes) that are without tree islands, but that have similar characteristics to current tree island sites, subsurface excavations should be made to look for charcoal layers and wood fragments. This may produce some indication of the extent of past treed areas. iv. Northern Limits of Picea mariana and Larix laricina The northernmost Picea mariana were located at sites along the Tuktoyaktuk Peninsula and the northernmost Larix laricina were located at a site along the Anderson River valley. Continued search for, and analysis of, more Picea mariana tree islands and Larix laricina would allow determination of establishment and radial growth patterns for these species that could be compared to that of Picea glauca tree islands. REFERENCES 177 Abrahamson, W.G. (1980) Demography of vegetative reproduction In O.T. Solbrig, ed., Demography and Evolution in Plant Populations. Botanical Monographs, Volume 15, Blackwell Scientific Publications, London, pp. 89-106. Agriculture Canada Expert Committee on Soil Survey (1987) The Canadian System of Soil Classification. Publication 1646, Research Branch, Agriculture Canada. Alexander, R.R. and W.D. Shepperd (1990) Picea engelmannii Parry ex Engelm. In Silvics of North America, Vol. 1. Conifers. Agriculture Handbook 654. U.S.D.A. Forest Service, Washington, D.C., pp. 187-203. Arseneault, D. and S. Payette (1992) A postfire shift from lichen-spruce to lichen-tundra vegetation at tree line. Ecology 73(3): 1067-1081. Atkinson, K. (1981) Vegetation zonation in the Canadian subarctic. Area 13:13-17. Baker, W.L. (1992) Structure, disturbance, and change in the Bristlecone pine forests of Colorado, U.S.A. Arctic and Alpine Research 24:17-26. Barman, M.W. (1942) Notes on the origin of adventitious roots in the native Ontario conifers. American Journal of Botany 29(8):593-598. Barnes, B .V. (1966) The clonal growth form of American aspens. Ecology 47(3):439-447. Bennett, K.D. (1985) The spread of Fagus grandifolia across eastern North America during the last 18,000 years. Journal of Biogeography 12:147-164. Black, R.A. and L.C. Bliss (1980) Reproductive ecology of Picea mariana (Mill.) B.S.P., at tree line near Inuvik, Northwest Territories, Canada. Ecological Monographs 50(3):331-354. Bostock, H.S. (1970) Physiographic subdivisions of Canada. In Geology and Economic Minerals of Canada, R.J.W. Douglas (ed.), 10-30. Ottawa: Geological Survey of Canada. Briffa, K.R., P.D. Jones, and F.H. Schweingruber (1994) Summer temperatures across northern North America: regional reconstructions from 1760 using tree-ring indices. Journal of Geophysical Research 99D:25835-25844. Briffa, K.R., F.H. Scheingruber, P.D. Jones, T.J. Osborn, S.G. Shiyatov, and E.A. Vaganov (1998) Reduced sensitivity of recent tree-growth to temperature at high northern latitudes. Nature 391:678-682. Brown, R.J.E. (1970) Permafrost in Canada: Its Influence on Northern Development. University of Toronto Press, Toronto. 178 Bryson, R.A. (1966) Air masses, streamlines, and the boreal forest. G e o g r a p h i c a l B u l l e t i n 8(3):228-268. Burgar, R.J. (1964) The effect of seed size on germination, survival, and initial growth in white spruce. Forestry C h r o n i c l e 43(2): 174-177. Burn, C.R. (1997) Cryostratigraphy, paleogeography, and climate change during the early Holocene warm interval, western Arctic coast, Canada. C a n a d i a n J o u r n a l of E a r t h Sciences 34:912-925. Burn, C.R. and P.A. Friele (1989) Geomorphology, vegetation succession, soil characteristics and permafrost in retrogressive thaw slumps near Mayo, Yukon Territory. A r c t i c 42(1):31-40. Burns, B . M . (1973) The C l i m a t e of the M a c k e n z i e Valley - Beaufort Sea Volumes I and II. Climatological Studies Number 24, Atmospheric Environment Service, Environment Canada, Toronto, Ontario. Caron, G.E., B.S.P. Wang, and H.O. Schooley (1990) Effect of tree spacing, cone storage, and prechilling on germination of P i c e a g l a u c a seed. Forestry C h r o n i c l e 66(4):388-392. Caron, G.E., B.S.P. Wang, and H.O. Schooley (1993) Variation in P i c e a g l a u c a seed germination associated with year of cone collection. C a n a d i a n J o u r n a l of Forest Research 23:1306-1313. Cook, R.E. (1985) Growth and development in clonal plant populations. In J.B.C. Jackson, L.W. Buss, and R.E. Cook, eds., P o p u l a t i o n Biology and E v o l u t i o n of Clonal O r g a n i s m s , Yale University Press, New Haven, pp. 259-296. Cooper, W.S. (1911) Reproduction by layering among conifers. B o t a n i c a l Gazette 52:369-379. Cram, W.H. (1951) Spruce seed viability: dormancy of seed from four species of spruce. Forestry C h r o n i c l e 27(4):349-357. Critchfield, W.B. (1985) The late Quaternary history of lodgepole and jack pines. C a n a d i a n J o u r n a l of Forest Research 15:749-772. Cwynar, L.C. and G.M. MacDonald (1985) A fossil pollen based reconstruction of the late Quaternary history of lodgepole pine (Pinus contorta ssp. latifolid) in the western interior of Canada. C a n a d i a n J o u r n a l of Forest Research 15:1039-1044. D'Arrigo, R.D. and G.C. Jacoby (1992) Dendroclimatic evidence from northern North America. In R.S. Bradley and P.D. Jones eds., C l i m a t e Since A . D . 1500, Routledge, London, pp. 296-311. D'Arrigo, R.D. and G.C. Jacoby (1993) Secular trends in high northern latitude temperature 179 reconstructions based on tree rings. Climate Change 25:163-177. Dallimore, S.R., S.A. Wolfe, J.V. Matthews, Jr., and J-S Vincent (1997) Mid-Wisconsinan eolian deposits of the Kittigazuit Formation, Tuktoyaktuk Coastlands, Northwest Territories, Canada. Canadian Journal of Earth Sciences 34:1421 -1441. Delcourt, P. A . and H.R. Delcourt (1987) Long-Term Forest Dynamics of the Temperate Zone: a Case Study of Late-Quaternary Forests in Eastern North America. Springer Verlag, New York. Delwaide, A. , L . Filion, and S. Payette (1991) Spatiotemporal distribution of light rings in subarctic black spruce, Quebec. Canadian Journal of Forest Research 21:1828-1832. Densmore, D. (1980) Vegetation and Forest Dynamics of the Upper Dietrich River Valley, Alaska. MSc Thesis, North Carolina State University, Department of Botany, Raleigh. Downie, B. and U . Bergsten (1991) Separating germinable and non-germinable seeds of eastern white pine (Pinus strobus L.) and white spruce (Picea glauca [Moench] Voss) by the IDS technique. The Forestry Chronicle 67(4):393-396. Ecoregions Working Group (1989) Ecoclimatic Regions of Canada. Map and Report. Canada Committee on Ecological Land Classification Series 21, Ottawa, Canada, 118p. Eis, S. (1967) Establishment and early development of white spruce in the interior of British Columbia. Forestry Chronicle 43(2): 174-177. Elliott, D.L. (1979a) The current regenerative capacity of northern Canadian trees, Keewatin, N.W.T., Canada: preliminary observations. Arctic and Alpine Research 11:243-251. Elliott-Fisk, D.L. (1988) The boreal forest. In M.G. Barbour and W.D. Billings, eds. North American Terrestrial Vegetation. Cambridge University Press, New York, U.S.A. Elliott-Fisk, D.L. (1983) The stability of the northern Canadian tree limit. Annals of the Association of American Geographers 73(4):560-576. Ellis, R.H., T.D. Hong, and E.H. Roberts (1985) Handbook of Seed Technology for Genebanks. Vol. land Vol. II. Handbook for Genebanks No. 3, International Board for Plant Genetic Resources, Rome. Environment Canada (2000) Canadian Climate Normals 1961-1990. w.w.w.cmc.ec.gc.ca/ climate/normals/. Filion, L. , S. Payette, and L. Guthier (1985) Analyse dendroclimatique d'un krummholz a la limite des arbres, Lac Bush, Quebec nordique. Geographie physique et 180 Quarternaire 39:221-226. Filion, L., S. Payette, L. Gauthier, and Y . Boutin (1986) Light rings in subarctic conifers as a dendrochronological tool. Quaternary Research 26:272-279. Folland, C.K., T.R. Karl, and K . Y . A . Vinnikov (1990) Observed climate variations and change. In J.T. Houghton, G.J. Jenkins, and J.J. Ephraums, eds., Climate Change. Cambridge University Press, Cambridge, England, pp. 199-238. Fowler, D.P. and Y.S. Park (1983) Population studies of white spruce. I. Effects of self-pollination. Canadian Journal of Forest Research 13:1133-1138. Fritts, H.C. (1976) Tree Rings and Climate. Academic Press, New York. Fuller, G.D. (1913) Reproduction by layering in black spruce. Botanical Gazette 55:452-457. Giddings, J.L. (1962) Development of tree ring dating as an archeological aid. In T.T Kozlowski, ed., Tree Growth, Ronald Press, New York. Goldstein, G.H, L .B . Brubaker, and T.M. Hinckley (1985) Water relations of white spruce (Picea glauca (Moench) Voss) at tree line in north central Alaska. Canadian Journal of Forest Research 15:1080-1087. Gordon, A . G . (1952) Spruce identification by twig characteristics. Forestry Chronicle 28:43-45. Gullett, D.W. and W.R. Skinner (1992) The State of Canada's Climate: Temperature Change in Canada 1895-1991. State of the Environment Report No. 92-2. Minister of Supply and Services Canada, 36 p. Hadley, J.L. and W.K. Smith (1987) Influence of krummholz mat microclimate on needle physiology and survival. Oecologia 73:82-90. Hansen, B.S. and D.R. Engstrom (1985) A comparison of numerical and qualitative methods of separating pollen of black and white spruce. Canadian Journal of Botany 62:2159-2163. Hare, F.K. and J.E. Hay (1974) The climate of Canada and Alaska. In R.A. Bryson and F.K. Hare, eds., World Survey of Climatology, Volume 11: Climates of North America, Elesvier, New York, pp. 49-192. Harper, (1977) Population Biology of Plants. Academic Press, Toronto. Henry, G.H.H., W. Freedman, and T.K. McLeod (in preparation) Vegetation change through treeline in the western Arctic, Canada. Submitted to Ecoscience Hughes, O.L. (1972) Surficial Geology of Northern Yukon Territory and Northwestern 181 District of M a c k e n z i e , Northwest T e r r i t o r i e s . Geological Survey of Canada, Paper 69-36. Hughes, O.L. (1987) Late Wisconsinan L a u r e n t i d e G l a c i a l L i m i t s of Northwestern Canada: the Tutsieta L a k e and Kelly L a k e Phases. Geological Survey of Canada, Paper 85-25. Hustich, I. (1979) Ecological concepts and biographical zonation of the north: the need for a generally accepted terminology. H o l a r c t i c E c o l o g y 2:208-217. Jacoby, G.C. Jr. (1982) The Arctic. In M . K . Hughes, P .M. Kelly, J.R. Pilcher, and V.C. LaMarche, eds., C l i m a t e from Tree R i n g s , Cambridge University Press, Cambridge, pp. 107-116. Jacoby, G.C. and R.D. D'Arrigo (1995) Tree ring width and density evidence of climatic and potential forest change in Alaska. G l o b a l B i o g e o c h e m i c a l Cycles 9(2):227-234. Jacoby, G . C , Jr., Cook, E.R., and L.D. Ulan (1985) Reconstructed summer degree days in central Alaska and northwestern Canada since 1524. Q u a t e r n a r y Research 23:18-26. Johnson, E.A. (1975) Buried seed populations in the subarctic forest east of Great Slave Lake, Northwest Territories. C a n a d i a n J o u r n a l of Botany 53:2933-2941. Johnson, E.A., K. Miyanishi, and H. Kleb (1994) The hazards of interpretation of static age structures as shown by stand reconstructions in a P i n u s contorta - P i c e a engelmannii forest. J o u r n a l of E c o l o g y 82:923-931. Johnston, W.F. (1990) Larix laricina (Du Roi) K. Koch. In Silvics of North A m e r i c a , Vol. 1. Conifers. Agriculture Handbook 654. U.S.D.A. Forest Service, Washington, D.C., pp.141-151. Johnstone, J.F. and G.H.H. Henry (1997) Retrospective analysis of growth and reproduction in Cassiope tetragona and relations to climate in the Canadian High Arctic. A r c t i c and Alpine Research 29:459-469. Kenety, (1917) Reproduction of black spruce. J o u r n a l of Forestry 15:446-448. Resting, S.F. (1996) Potential insolation and plant communities through treeline in the Lower A n d e r s o n River valley, N. W.T., C a n a d a . MSc Thesis, University of British Columbia. Krebs, J.S. and Barry, R.G. (1970) The Arctic front and the tundra-taiga boundary in Eurasia. G e o g r a p h i c a l Review 60:548-554. Kryuchkov, V . V . (1973) The effect of permafrost on the northern tree line. In F.J. Sanger and P.J. Hyde eds., Permafrost, 2nd International Conference, USSR Contribution, National Academy of Sciences, Washington, D.C., pp. 136-138. Kullman, L . (1990) Dynamics of altitudinal tree-limits in Sweden: a review. Norsk G e o g r a . 182 Tidsskr. 44:103-116. Larsen, J.A. (1974) Ecology of the northern continental forest border. In J.D. Ives and R.G. Barry, eds., A r c t i c and Alpine E n v i r o n m e n t s , Methuen, London, pp. 341-369. Larsen, J.A. (1989) The N o r t h e r n Forest B o r d e r in C a n a d a and A l a s k a . Springer-Verlag, New York. Lavoie, C. and S. Payette (1992) Black spruce growth forms as a record of a changing winter environment at treeline, Quebec, Canada. A r c t i c and Alpine Research 24(l):40-49. Lavoie, C. and S. Payette (1994) Recent fluctuations of the lichen-spruce forest limit in subarctic Quebec. J o u r n a l of E c o l o g y 82:725-734. Legere, A . and S. Payette (1981) Ecology of a black spruce (Picea mariana) clonal population in the hemiarctic zone, northern Quebec:Population dynamics and spatial development. A r c t i c and Alpine Research 13(3):261-276. Lescop-Sinclair, K. and S. Payette (1995) Recent advance of the arctic treeline along the eastern coast of Hudson Bay. J o u r n a l of E c o l o g y 83:929-936. Little, Jr., E.L. and S.S. Pauley (1958) A natural hybrid between black and white spruce in Minnesota. The A m e r i c a n M i d l a n d N a t u r a l i s t 60(1):202-211. Lutz, H.J. (1939) Layering in eastern white pine. B o t a n i c a l Gazette 101:505-507. MacDonald, G.M. , T.W.D. Edwards, and K . A . Moser (1993) Rapid response of treeline vegetation and lakes to past climate change. Nature 361:243-246. Mackay, J.R. (1958) The Valley of the Lower A n d e r s o n River, N. W.T. Geographical Bulletin No. 11:16-56. Mackay, J.R. (1963) The M a c k e n z i e Delta A r e a , N. W.T. Memoir 8, Geographical Branch, Ottawa. Mackay, J.R. (1974) The M a c k e n z i e Delta A r e a , N.W.T. Geological Survey of Canada, Miscellaneous Report 23, Energy, Mines and Resources, Ottawa. Mackay, J.R. (1985) Mackay, J.R. (1990a) Pingos of the western Arctic coast. Scientific Report No. 1, The Science Institute of the Northwest Territories, pp. 14-16. Mackay, J.R. (1990b) Some observations on the growth and deformation of epigenetic, syngenetic, and anti-syngenetic ice wedges. Permafrost and P e r i g l a c i a l Processes 1:15-29. 183 Mackay, J.C. (1992) Lake Stability in an ice-rich permafrost environment: examples from the western Arctic coast. In R.D. Robarts and M . L . Bothwell, eds., Aquatic Ecosystems in Semi-arid Regions: Implications for Resource Management, N.H.R.I. Symposium Series 7, Environment Canada, Saskatoon. Mackay, J.R. (1994) Pingos and pingo ice of the western Arctic coast, Canada. Terra 106:1-11. Mackay, J.R. and S.R. Dallimore (1992) Massive ice of the Tuktoyaktuk area, western Arctic coast, Canada. Canadian Journal of Earth Sciences 29(6): 1235-1249. Mathews, W.H., J.R. Mackay, and G.E. Rouse (1989) Pleistocene geology and geomorphology of the Smoking Hills Upland and lower Horton River Arctic coast of mainland Canada. Canadian Journal of Earth Sciences 26:1677-1687. McLeod, T.K. and C M . MacDonald (1997) Postglacial range expansion and population growth of Picea mariana, Picea glauca, and Pinus banksiana in the western interior of Canada. Journal of Biogeography 24:865-881. Mekis, E. and W.D. Hogg (1999) Rehabilitation and analyses of Canadian daily precipitation time series. Atmosphere and Ocean, 37. Mercier, S. and C.-G. Langlois (1992) Indice de maturite et conservation en fonction de la periode de recolte des semences d'epinette blanche du Quebec. Canadian Journal of Forest Research 22:1516-1523. Morin, A . and S. Payette (1986) Expansion recente de meleze a la limite des forets. Canadian Journal of Botany 62(7): 1404-1408. Morisset, P., S. Payette, and J. Deshaye (1983) The vascular flora of the northern Quebec-Labrador Peninsula: phytogeographical structure with respect to the tree-line. In P. Morisset and S. Payette, eds., Tree-Line Ecology, Proceedings of the Northern Quebec Tree-Line Conference. Collection Nordicana No. 47, 141-152. Nienstaedt, H. and J.C. Zasada (1990) Picea glauca (Moench) Voss. In Silvics of North America, Vol. I. Conifers. Agriculture Handbook 654. U.S.D.A. Forest Service, Washington, D . C , pp.204-226. Oswalt, W.H. (1957) Volcanic activity and Alaskan spruce growth in A D 1783. Science 126:928-929. Owens, J ; N. and M . Molder (1977a) Bud development in Picea glauca. I. Annual growth cycle of vegetative buds and shoot elongation as they relate to date and temperature sums. Canadian Journal of Botany 55:2728-2746. Owens, J.N. and M . Molder (1977b) Bud development in Picea glauca. II. Cone differentiation and early development. Canadian Journal of Botany 55:2746-2760. 184 Owens, J.N. and M . Molder (1979) Sexual reproduction of white spruce (Picea glauca). C a n a d i a n J o u r n a l of Botany 57:152:169. Parker, W.H. and D.G. McLauchlan (1978) Morphological variation in white and black spruce: investigation of natural hybridization between P i c e a g l a u c a and P. m a r i a n a . C a n a d i a n J o u r n a l of Botany 56:2512-2520. Pastor, J. and D.J. Mladenoff (1992) The southern boreal-northern hardwood forest border. In H.H. Shugart, R. Leemans, and G.B. Bonan, eds., A Systems A n a l y s i s of the G l o b a l B o r e a l Forest. Cambridge University Press, New York, U.S.A., pp. 216-240. Payette, S. (1974) Classification ecologique des formes de croissance de P i c e a g l a u c a (Moench.) Voss et de P i c e a m a r i a n a (Mill.) BSP en milieux subarctiques et subalpins. Naturaliste C a n a d i e n 101:893-903. Payette, S. (1976) Succession ecologique des forets d'epinette blanche et fluctuations climatiques, Poste-de-la-Baleine, Nouveau-Quebec. C a n a d i a n J o u r n a l of Botany 54:1394-1402. Payette, S. (1983) The forest tundra and present tree-lines of the northern Quebec-Labrador Peninsula. In P. Morisset and S. Payette, eds., Tree-Line E c o l o g y , Proceedings of the Northern Quebec Tree-Line Conference. Collection Nordicana No. 47, 3-23. Payette, S., J. Deshaye, and H. Gilbert (1982) Tree seed populations at the treeline in Riviere aux Feulles area, northern Quebec, Canada. A r c t i c and A l p i n e Research 14(3):215-221. Payette, S. and L. Filion (1985) White spruce expansion at the tree line and recent climatic change. C a n a d i a n J o u r n a l of Forest Research 15:241-251. Payette, S., L . Filion, A . Delwaide, and C. Begin (1989) Reconstruction of tree-line vegetation response to long-term climate change. Nature 341:429-432. Payette, S., L . Filion, L . Gauthier, and Y . Boutin (1985) Secular climate change in old-growth tree-line vegetation of Northern Quebec. Nature 315:135-138. Payette, S. and R. Gagnon (1979) Tree-line dynamics in Ungava peninsula, northern Quebec. H o l a r c t i c E c o l o g y 2:239-248. Payette, S. and R. Gagnon (1985) Late Holocene deforestation and tree regeneration in the forest-tundra of Quebec. Nature 313:570-572. Payette, S. and C. Morneau (1993) Holocene relict woodlands at the eastern Canadian treeline. Q u a t e r n a r y Research 39:84-89. Payette, S., Morneau, C , Sirois, L . , and Desponts, M . (1989) Recent fire history of the northern Quebec biomes. E c o l o g y 70(3):656-673. 185 Perala, D.A. (1990) Populus tremuloides Michx. In Silvics of North America, Vol. 2. Hardwoods. Agriculture Handbook 654. U.S.D.A. Forest Service, Washington, D.C., pp.555-569. Porsild, A .E . and W.J. Cody (1980) Vascular Plants of Continental Northwest Territories, Canada. National Museum of Natural Sciences, Canada. Prentice, I.C. (1986) Vegetation responses to past climate variation. Vegetatio 67:131-141. Rampton, V . N . (1988a) Quaternary Geology of the Tuktoyaktuk Coastlands, Northwest Territories. Geological Survey of Canada, Memoir 423, Ottawa. Rampton, V . N . (1988b) Origin of massive ground ice on Tuktoyaktuk Peninsula, Northwest Territories, Canada: A review of stratigraphic and geomorphic evidence. In. International Permafrost Conference. Ritchie, J.C. (1984) Past and Present Vegetation of the Far Northwest of Canada. University of Toronto Press, Toronto. Ritchie, J.C. (1987) Postglacial Vegetation of Canada. Cambridge University Press, New York, U.S.A. Rowe, J.S. (1970) Spruce and fire in northwest Canada and Alaska. Tall Timbers Fire Ecology Conference Proceedings No. 10, pp.245-254. Rowe, J.S. (1972) Forest Regions of Canada. Publication 1300. Department of the Environment, Ottawa, Canada. Sakai, A . (1983) Comparative study on freezing resistance of conifers with special reference to cold adaptation and its evolutive aspects. Canadian Journal of Botany 61:2323-2332 Schier, G.A. (1975) Deterioration of aspen clones in the middle of the Rocky Mountains. Research Paper INT-170, U.S.D.A., Forest Service. Schopmeyer, C.S. (1974) Seeds of Woody Plants in the United States. Agriculture Handbook No. 450. Forest Service, U.S.D.A., Washington, D.C. Scott, P.A., C.V. Bentley, D.C.F. Fayle, and R.I.C. Hansell (1987a) Crown forms and shoot elongation of white spruce at the treeline, Churchill, Manitoba, Canada. Arctic and Alpine Research 19(2): 175-186. Scott, P.A., R.I.C. Hansell, and D.C.F. Fayle (1987b) Establishment of white spruce populations and responses to climate change at the treeline, Churchill, Manitoba, Canada. Artie and Alpine Research 19(1):45-51. Sirois, L . (1992) The transition between boreal forest and tundra. J n H.H. Shugart, R. Leemans, and G.B. Bonan, eds., A Systems Analysis of the Global Boreal Forest. 186 Cambridge University Press, New York, U.S.A., pp. 196-215. Spear, R.W. (1983) Paleoecological approaches to a study of tree-line fluctuations in the Mackenzie Delta Region, Northwest Territories: Preliminary results. In P. Morisset and S. Payette, eds., T r e e - L i n e E c o l o g y , Proceedings of the Northern Quebec Tree-Line Conference. Collection Nordicana No. 47, 61-72. Stevens, G.C. and J.F. Fox (1991) The causes of treeline. A n n u a l Review of E c o l o g y and Systematics 22All ' -191 . Stuiver, M . and P J . Reimer (1993) CALIB3 P r o g r a m . Quaternary Isotope Laboratory, University of Washington, U.S.A. Szeicz, J .M. (1996) White spruce light rings in northwestern Canada. A r c t i c and Alpine Research 28(2): 184-189. Szeicz, J .M. and G.M. MacDonald (1995a) Recent white spruce dynamics at the subarctic alpine treeline of north-western Canada. J o u r n a l of E c o l o g y 83:873-885. Szeicz, J .M. and G.M. MacDonald (1995b) Dendroclimatic reconstruction of summer temperatures in northwestern Canada since A.D. 1638 based on age-dependent modelling. Q u a t e r n a r y Research 44:257-266. Szeicz, J .M. and G.M. MacDonald (1996) A 930-year ring-width chronology from moisture-sensitive white spruce (Picea g l a u c a Moench) in northwestern Canada. The H o l o c e n e 6: 1-7. Taylor, A.E . , S.R. Dallimore, and A.S. Judge (1996) Late Quaternary history of the Mackenzie-Beaufort region, Arctic Canada, from modelling of permafrost temperatures. 2. The Mackenzie Delta-Tuktoyaktuk Coastlands. C a n a d i a n J o u r n a l of E a r t h Sciences 33:62-71. Tikhomirov, B.A. (1962) The treelessness of the tundra. Polar Record 11:24-30. Timoney, K.P. (1988) A G e o b o t a n i c a l Investigation of the Subarctic F o r e s t - T u n d r a of the Northwest T e r r i t o r i e s . Ph.D. Dissertation, Department of Botany, University of Alberta, Edmonton, Alberta, 292 p. Timoney, K.P., G.H. LaRoi, S.C. Zoltai, and A . L . Robinson (1992) The high subarctic forest-tundra of northwestern Canada: position, width, and vegetation gradients in relation to climate. A r c t i c 45(1): 1-9. Tranquillini, W. (1979) P h y s i o l o g i c a l E c o l o g y of A l p i n e . Timberline. Springer-Verlag, Berlin, Germany, 137pp. Tuhkanen, S. (1993) Treeline in relation to climate, with special reference to oceanic areas. In J. Alden, J.L. Mastrantonio, and S. Odum, eds., Forest Development in C o l d Climates, Plenum Press, New York, pp. 115-133. 187 van Cleve, K., W.C. Oechel, and J.L. Horn (1990) Response of black spruce (Picea mariana) ecosystems to soil temperature modification in interior Alaska. C a n a d i a n J o u r n a l of Forest Research 20:1530-1535. Viereck, L . A . (1965) Relationship of white spruce to lenses of perennially frozen ground, Mount McKinley National Park, Alaska. A r c t i c 18(4):262-267. Viereck, L .A. and W.F. Johnston (1990) P i c e a m a r i a n a (Mill.) B.S.P. In Silvics of North A m e r i c a , Vol. 1. Conifers. Agriculture Handbook 654. U.S.D.A. Forest Service, Washington, D.C., pp.227-237. Vincent, J.S. (1989) Quaternary geology of the northern Canadian Interior Plains. In Q u a t e r n a r y Geology of C a n a d a and G r e e n l a n d , Geology of C a n a d a , N o . 1, R.J. Fulton (ed.), 100-137. Ottawa: Geological Survey of Canada. Vincent, J.S. and V . K . Prest (1987) The Early Wisconsinan history of the Laurentide Ice Sheet. G e o g r a p h i e physique et Q u a t e r n a i r e 41:199-213. Vincent, L . A . (1998) A technique for the identification of inhomogeneities in Canadian temperature series. J o u r n a l of C l i m a t e , 11, 1094-1104. Vincent, L . A . and D.W. Gullett (1999) Canadian historical and homogeneous temperature datasets for climate change analyses. International J o u r n a l of C l i m a t o l o g y , 19,1375-1388. Wagg, J.W.B. (1964) White spruce regeneration on the Peace and Slave River lowlands. Publication 1069, Canada Department of Forestry, Forestry Research Branch, Ottawa. Walter, H . (1984) Vegetation of the E a r t h and Ecosystems of the Geo-biosphere. 3rd ed. Springer-Verlag, New York. Wardle, P. (1974) Alpine timberlines. In J.D. Ives and R.G. Barry, eds. A r c t i c and A l p i n e E n v i r o n m e n t s , Methuen and Co., London, pp. 371-402. Wardle, P. (1993) Causes of alpine timberline: a review of the hypotheses. In J. Alden, J.L. Mastrantonio, and S. Odum eds., Forest Development in Cold Climates, Plenum Press, New York, pp. 89-103. Whitewood, R.D. (2000) Climate trends and variations bulletin for Canada, annual 1999 temperature and precipitation in historical perspective, http://www.tor.ec.gc.ca/ ccrm/bulletin/index.html Williams, P.J. and M.W. Smith (1989) The F r o z e n E a r t h . F u n d a m e n t a l s of Geocryology. Cambridge University Press, Cambridge, U.K. Yamaguchi, D.K., L. Filion, and M . Savage (1993) Relationship of temperature and light ring formation at subarctic treeline and implications for climate reconstruction. Q u a t e r n a r y Research 39:256-262. 188 Yorath, C.J. and D.G. Cook (1981) Cretaceous and Teriary stratigraphy and paleogeography, Northern Interior Plains, District of Mackenzie. Geological Survey of Canada, Memoir 398, Ottawa. Yorath, C.J., D.K. Norris, and F.G. Young (1980) Geology of the Beaufort-Mackenzie Basin, District of Mackenzie and Yukon Territory. Geological survey of Canada, Map 1509A, Ottawa. Young, F.G., D.W. Myhr, and C.J. Yorath (1976) Geology of the Beaufort-Mackenzie Basin. Geological Survey of Canada, Paper 76-11, Ottawa. Zasada, J.C. (1971) Natural regeneration of interior Alaska forests - seed, seedbed, and vegetative reproduction considerations. In C W . Slaughter, R J . Barney, and G.M. Hansen, eds., Proceedings - F i r e in The N o r t h e r n E n v i r o n m e n t - A Symposium. Pacific Northwest Forest and Range Experiment Station Forest Service, U.S. Department of Agriculture, pp. 231-246. Zasada, J.C. and H . M Phipps (1990) P o p u l u s balsamifera (L.) In Silvics of North A m e r i c a , Vol. 2. Hardwoods. Agriculture Handbook 654. U.S.D.A. Forest Service, Washington, D.C., pp.518-529. Zasada, J .C, T.L. Sharik, and M . Nygren (1992) The reproductive processes in boreal forest trees. In H.H. Shugart, R. Leemans, and G.B. Bonan, eds., A Systems Analysis of the G l o b a l B o r e a l Forest. Cambridge University Press, New York, U.S.A., pp. 85-125. Zoltai, S.C., D J . Karasiuk, and G.W. Scotter (1979) A Natural Resource Survey of the Horton-Anderson Rivers Area, Northwest Territories. Parks Canada, Ottawa. APPENDIX 1 189 A l . Location and Description of Sites In the Tuktoyaktuk region, 24 tree island and forest stand sites were visited over the three field seasons (1993, 1994, 1995) (Figure A l . l ) . This appendix presents a description of each site. A l . l Richards Island Sites Site R l , Todd Spruce, was located at 69°28'N 134°19'W on the northwest portion of Richards Island. It was a large tree island, greater than 1200 m 2 , on a southwest facing slope above a large lake, composed exclusively of P i c e a g l a u c a . Most of the stems had a mat, infranival cushion, or infranival erect growth form and were between 0.25 and 0.5 m in height. There were a few erect whorled stems, between 1.0 and 1.5 m in height which were found either at the base of the slope, or at a small bench near the top of the hill. Site R2, Sutherland Spruce, was located at 69°23'N 134°02'W on the southeast . portion of Richards Island. The tree island was composed of P i c e a g l a u c a and was approximately 100 m in area, on the south-facing slope of a hollow in a sandy deposit. There were several small blow-outs present. Most of the stems had infranival cushion or erect growth forms, though mat growth form was also present. Heights were generally less than 0.5 m, but there were two erect whorled stems, 1.0 to 1.5 m in height. A1.2 Tuktoyaktuk Peninsula Sites Site T l , Mackay Spruce, was located at 69°23'N 133°30'W, 30 m a.s.L, on the north side of the peninsula, 400 m inland from the coast. The tree island was composed of 71 P i c e a 190 g l a u c a individuals (66 live, 5 dead) and was 32.5 m 2 in area. It was at the base of a 4° slope of a small hill that rose approximately 10 m above a nearby pond. The tree island faced southeast. Stem height ranged from 0.2 to 1.2 m and growth Torms present were mat, infranival cushion, infranival erect, and erect whorled. At the end of the 1995 season, branches at the top of the erect whorled stems were experiencing some or complete needle loss. Many parts of the tree island were chlorotic. Site T2, Kittigazuit Spruce, was located at 69°19'N 133°44'W, 10 m a.s.l., at the top of a west facing slope within a bay on the arctic coast. The tree island was composed of four groups of stems: three groups were aligned south to north at the top of the slope , and one group was to the east, behind these three groups. The total area was estimated to be 62 m . Growth forms present were mat, infranival cushion, infranival erect, and erect whorled. Tree heights ranged from 0.25 to 1.7 m. The largest seven erect, whorled individuals were 1.0 to 1.7 m in height. The P i c e a g l a u c a were being affected by mobilization of sand by wind. Stems (including branches and cones) were being buried at the eastern and southern parts of the tree island. Roots were exposed, and sand covered the needles of many individuals located in the western part of the tree island towards the coast and between the groups of stems where wind was funnelled. Chlorotic needles were prominent on the western side of the tree island and between the groups of stems. The slope base was being actively eroded by wave action and there are small groups of P i c e a g l a u c a on ledges of the cliff face. It may be that these have slumped down from the tree island above. Partially buried stumps were located along the cliff face below the live P i c e a g l a u c a . These stumps were rooted into an organic-rich, sandy substrate with many woody roots present. Four stumps were excavated and two were submitted for radiocarbon dating. Site T3, Drillpad Spruce, was located at 69°19'N 132° 49'W, 28 m a.s.l., on a 12° 191 south-facing slope. The tree island, composed of P i c e a g l a u c a , extended from the hill top to the base of the slope at the shore of a small lake, with some dead stumps submerged within 0.5 m of shore. The tree island was composed of 181 P i c e a g l a u c a individuals (158 live, 23 dead) over an area of 110 m 2 . Many individuals had supranival skirted and whorled growth forms, with some infranival cushion and infranival erect growth forms. Some of the larger individuals were flagged. Stems ranged in height from 0.6 to 2.8 m. The site was the location of a Drillpad and signs of the disturbance were still abundant, including a thermokarst pond at the location of the pad, broad areas of E p i l o b i u m angustifolium L . and E q u i s e t u m spp. L. surrounding the pond, several areas that were metal dumps, and scrap metal within the tree island. At the end of the 1995 season, there were many individuals with chlorotic needles, particularly to the northeast where two individuals had few green needles left. The individuals closer to the lake appeared healthier. Site T4, Spear Spruce, was located at 69°28'N 132° 10'W, 37 m a.s.l., on the top and shoulder of a small hill. To the north, the hill slope was very steep and leads down to a lake. There were seven groups of P i c e a g l a u c a , all had a low mat growth form, except two • infranival erect stems. Stem heights were <0.75 m. There were numerous thin dead stems along the ground. The tree island covered an area of 100 m 2 . Site T5, Hilltop Spruce, was located at 69°1 TN 133°01'W, 40 m a.s.l, on the upper part of a 5° slope with a southern aspect. The tree island was composed of 71 P i c e a g l a u c a individuals (60 live, 11 dead) in a 40 m area, approximately 100 m from a small lake. Growth forms present were infranival cushion, infranival erect, supranival skirted, and erect whorled. Many individuals were flagged. Tree heights ranged from 0.3 to 2.8 m. At the end of the 1995 season, all erect individuals showed some chlorosis, many had lost numerous needles. 192 A P i c e a stump was located to the west of T5 in a valley. It was being exposed on a small slump in an ice wedge melt channel. The stump was rooted into peaty material that was moderately decomposed. The stump was excavated and a sample submitted for radiocarbon dating. Sixteen tree islands were located at site T6, 69°18'N 132°33'W, within an area of 0.8 km (see Figure 4.7). The density of tree islands was approximately 20 tree islands/km . Fifteen of these tree islands were composed solely of P i c e a g l a u c a . At site T6h, Brunet Sand Spruce, there were seven tree islands (hi to h7). P i c e a g l a u c a occurred at each and P i c e a m a r i a n a was found at tree island T6h6. Tree island T6h6 occurred on a southeast-facing slope and had two distinct parts. Upslope was a 120 m 2 area of P i c e a g l a u c a krummholz (infranival cushion) with one erect whorled stem and numerous dead stems. Downslope from this was a 310 m area of P i c e a m a r i a n a with approximately 15 erect individuals and much infranival cushion krummholz. There was a large area of dead stems adjacent to the live P i c e a m a r i a n a and were presumed to be dead P i c e a m a r i a n a . Further downslope and on the flat valley floor, another area of P i c e a m a r i a n a occurred of approximately 20 m 2 . At the base of the slope, many of stems had chlorotic needles. Erect individuals of P i c e a m a r i a n a ranged from 1.0 to 1.7 m in height. At site T6 as a whole, P i c e a g l a u c a ranged in height from 0.5 to 3.5 m. Tree island area ranged from 3 to 450 m and their locations varied from hill top to hillslope to flat valley bottom. Tree islands occurred on slopes that are < 10° and mainly faced west, south, east, and southeast, but two tree islands faced north and northwest. Signs of disturbance by humans were present at T6h with some branches cut and piled. Evidence of disturbance by animals included a bear hair on a snapped branch (presumed that tree was used as a rub), and grazed ends of branches (presumed by caribou). In tree islands T6a through T6h, the number 193 of erect individuals varied from 1 to 33 and growth forms present were mat, infranival cushion, infranival erect, supranival skirted, and erect whorled. Site T6i, Brunet Spruce, had 56 individuals (26 live, 30 dead) in an area of 70 m 2 . It was located on top and shoulder of a small bench on a west-facing slope. Individual heights varied from 0.55 to 3.3 m and growth forms present were infranival cushion, infranival erect, supranival skirted, and erect whorled. On one individual, there were sharp incisions that appear to have been made by an axe. In one area of the tree island, there was a group of 14 dead individuals. In this area, the ground surface was covered by a thick layer of needles, indicative of a once healthy grove of trees. In the 1995 growth season, the trees showed some chlorotic needles. At the end of the 1995 field season, another P i c e a g l a u c a tree island was located at 69°17'N 132°52'W, to the southwest of site T5. It was mapped as site T i l , Hamilton Spruce, and was found on a gentle, south-facing slope area. There were infranival cushion, infranival erect, supranival skirted, and erect whorled growth forms present. The trees showed some chlorotic needles. Site T12, Two Spruce, was located on a large island in a lake at 69°05'N 133°16'W. There were two P i c e a m a r i a n a tree islands present on the southwest-facing slope, separated by a small low-lying area. T12a was a large tree island, approximately 200 m 2 with infranival erect, supranival skirted, and erect whorled individuals from approximately 0.5 to 3 m in height. At T12b, there were three supranival skirted and erect whorled individuals between 4 and 5 m in height, with small patches of infranival cushion and erect krummholz (0.5 to 1 m in height), and some small (<2 m in height) supranival and erect whorled individuals. Site T7, within the Low Arctic Shrub-Tundra, was located at 69°15'N 131°50'W on 194 a south-facing slope and upland area around a small lake. The scattered P i c e a g l a u c a tended to be between 3 and 4 m in height. The upland tree islands had many, densely packed P i c e a g l a u c a . P i c e a m a r i a n a were not found at this site. Site T8, within the High Subarctic Forest-Tundra, was located at 69° 1 l ' N 131°27'W. P i c e a g l a u c a occurred on slope and upland areas as scattered trees, many greater than 5 m tall. Seedlings were present. P i c e a m a r i a n a occurred in the wetter valley bottoms. Site T9 was located at 69°06'N 130°59'W, within the High Subarctic Forest-Tundra. The slopes and upland areas were covered with scattered P i c e a g l a u c a trees greater than 5 m in height, with many seedlings present. P i c e a m a r i a n a occurred in the river valley. Site T10 was located 10 km southeast of Inuvik at 68°17'N 133°15'W. The upland areas were dominated by P i c e a g l a u c a greater than 5 m in height. P i c e a m a r i a n a occurred in the lowland areas. Some non-treed areas were found in the lowland bogs. A1.3 Anderson River Valley Sites Site A l , Armstrong Spruce, was located at 69°42'N 128°59'W, along a south-facing slope. Twelve P i c e a g l a u c a tree islands occurred from the base to mid-slope positions, with slopes varying from 0 to 6°. The density of tree islands was approximately 10 tree islands/km2. Tree island area ranged from 1 m 2 to 50 m 2 , with 3 to 71 live individuals. Growth forms present were infranival cushion and infranival erect. However, in only three tree islands, including A l a and A l b , were infranival erect individuals present. Heights ranged from 0.3 to 0.9 m for A l a and 0.4 to 1.0 m for A l b . At the third tree island, erect stems were less than 1 m in height. In seven of the tree islands, there were dead stumps. At the end of the 1995 season, all of these tree islands showed healthy growth. Site A2, Spumed Swan Spruce, was located at 69°39.5'N 128°43'W on a southwest-195 facing slope above the active floodplain of the Anderson River. Tree islands were numerous (more than 80), with a tree island density of approximately 320 tree islands/km . The tree islands spanned the full length of the slope, which varied from 1.5 to 6.5°. Other tree islands were also found east of the river valley in a nivation hollow. Most of these tree islands occurred along a drainage course, but a few were more isolated and occurred along gentle slopes that faced east and southeast. A l l tree islands were composed of P i c e a g l a u c a . The main sampling at A2 covered a 60 x 150 m area, within which 55 dead stumps, 23 dead tree islands (all individuals dead), and 59 live tree islands (at least one live individual) were identified. The size of the tree islands ranged from 0.3 to 41.3 m , with an average of 7.6 m . Not including krummholz, the number of individuals in each tree island ranged from 1 to 23, with an average of 6.3 individuals per tree island. Most of the tree islands were broadly circular in shape (50 tree islands or 63%); the remaining tree islands were broadly elliptical with the long axis either oriented downslope (11 tree islands or 14%) or across the slope (following a contour line; 18 tree islands or 23%). Growth forms present were mat, infranival cushion, infranival erect, supranival skirted, and erect whorled. Two tree islands were intensively studied: A2a and A2b. Both of these tree islands were located mid slope, where the slope is 4°. A2a was approximately 40 m 2 with the long axis oriented across the slope (along a contour). There were 48 individuals (28 live, 20 dead) with heights that ranged from 0.7 to 3.2 m. Growth forms present were infranival cushion, infranival erect, supranival skirted, and erect whorled. A2b had 9 individuals (1 live, 8 dead) in a 9.4 m area with the long axis oriented downslope. These two tree islands were not randomly selected. A2a was chosen as it is a large tree island, and A2b because one of its dead individuals had an unusually large base (24 cm diameter at 45 cm above root bole). Site A3, Caribou Spruce, was located at 69°36'N 128°35'W, on a flat river floodplain: 196 The number of tree islands in an area of 3.2 km was greater than 300, giving a density of approximately 100 tree islands/km2. A l l tree islands encountered were composed of Picea glauca. Three tree islands were intensively studied. A3 a was a 4 m tree island composed of 8 individuals (5 alive, 3 dead), that were 0.5 to 1.7 m in height. A3b was 15 m , composed of 25 individuals (all live) with heights that ranged from 0.5 to 2.5 m. These two tree islands were chosen as they appeared to be of typical areal size and growth forms. A3c was a large tree island of 122 m 2 with 51 individuals (34 live, 17 dead) and heights ranged from 0.5 to 4.5 m. Growth forms present were infranival cushion, infranival erect, supranival skirted, and erect whorled. At the end of the 1994 growth season, the trees appeared healthy. There was much driftwood present on the floodplain and amongst the tree islands, indicative of a spring flood. At the end of the 1995 growth season, the tree islands had many patches of chlorotic needles and generally looked stressed. Site A4 and A5 were located north of treeline. A4 was located on the floodplain, west of the river at 69°31'N 128°27'W. Scattered individuals of Picea glauca formed an open stand. Various sized individuals were present, with many greater than 5 m in height. No krummholz or seedlings were found; the two smallest individuals were infranival erect stems, 0.4 and 0.5 m in height. No evidence of layering was found: the individuals were well-spaced, straight stems, and no live lower branches reached the ground, and no krummholz were present. Growth forms present were infranival erect, supranival skirted, erect whorled, and trees. At this site there were four Larix laricina trees present, approximately 6 to 7 m in height. Larix 1 and 2 were next to each other, and approximately 80 m west were Larix 3 and 4. Larix 4 had been topped and appeared stressed. The other three Larix were healthy. Picea mariana were not found at this site. Site A5 was located at 69°25'N 128°23'W on the upper part of the floodplain. Picea 197 g l a u c a trees were up to 10 to 15 m in height. Trees were straight and well-spaced with both open and closed canopies at the site. No evidence of layering was found. Seedlings were present (some seedling ages, estimated by nodes, were between 1 and 6 years) and were erect, symmetrical, with branch-free bases. P i c e a m a r i a n a were located at this site on an upper terrace in dense forest groves. At the end of the 1995 season, P i c e a g l a u c a had some patches of chlorotic needles. Site A6 to A9 were within the high subarctic forest-tundra. A6 was located at 69°17'N 128°26'W, on a west-facing slope and flat terrace above the river. Many seedlings and saplings of all sizes were present. P i c e a g l a u c a trees were up to 15 to 20 m in height, they had straight stems and were well-spaced. No evidence of layering was found. P i c e a m a r i a n a were found in the vicinity of this site. At the end of the 1995 season, a few trees had some chlorotic needles. Site A7 was located at 69°07'N 128°18'W on the upper part of the floodplain. Seedlings and saplings of all sizes were present. P i c e a g l a u c a trees ranged up to 20 m in height, trunks were straight and trees were well spaced. Different areas in the site had either open or closed canopies. No evidence of layering was found. P i c e a m a r i a n a occurred in the vicinity of the site. At the end of the 1995 season, a few trees had chlorotic needles. Site A8 was located at 68°60'N 128°24'W on the upper floodplain. Some trees were greater than 15 m in height and seedlings and saplings of all sizes were present. No evidence of layering was found. Trees were straight-stemmed and well spaced. Areas of closed and open canopies were present. P i c e a m a r i a n a occurred in the. vicinity of the site. Some trees had chlorotic needles at the end of the 1995 growth season. Site A9 was located at 69°08'N 129°58'W at 90 m a.s.l. in an flat upland area. Both P i c e a g l a u c a and P i c e a m a r i a n a were present at this site. Some live branches reached the 198 ground and curved bases were present, indicative of layering. A1.4 Horton River Valley Sites Site H3 was located at 69°40' N 126°59' W, on the upper part of the Horton River floodplain. One P i c e a g l a u c a tree island was located at this site. A l l individuals were less than 1 m in height. There were few erect whorled stems and many stems that had an infranival cushion growth form. Site H I , was located at 69°36'N 126°57'W, on the second terrace and the slope between the first and second terraces of the Horton River. Seventeen tree islands occurred at this site, with a density of approximately 68 tree islands/km2. The tree islands ranged in size from 3.8 m 2 to 288 m 2 , with an average of 81.1 m 2 . They occurred on flat to 13° slopes, which faced east, southeast, and south. The tree islands had 4 to 41 erect individuals plus krummholz, with an average of 16.9 erect individuals per tree island. Erect individuals ranged in height from 0.5 to 3.5 m and growth forms present were infranival cushion, infranival erect, supranival skirted, and erect whorled. Tree island H l a was intensively studied. It was located on a 2° east-facing slope. There were 43 individuals (28 live, 15 dead) in a 45 m 2 area. Heights ranged from 0.5 to 2.6 m. At the end of the 1994 growth season, the tree islands appeared healthy with little evidence of chlorotic needles. Site H2 was located at 69°29'N 126°55'W on two terraces of the Horton River. P i c e a m a r i a n a occurred in the vicinity of the site. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0103841/manifest

Comment

Related Items