@prefix vivo: . @prefix edm: . @prefix ns0: . @prefix dcterms: . @prefix dc: . @prefix skos: . vivo:departmentOrSchool "Arts, Faculty of"@en, "Geography, Department of"@en ; edm:dataProvider "DSpace"@en ; ns0:degreeCampus "UBCV"@en ; dcterms:creator "Jones, Glenda A."@en ; dcterms:issued "2009-03-10T05:23:58Z"@en, "1997"@en ; vivo:relatedDegree "Master of Science - MSc"@en ; ns0:degreeGrantor "University of British Columbia"@en ; dcterms:description """Primary plant succession was examined on the foreland of the retreating Twin Glacier at Alexandra Fjord, Ellesmere Island. The position of the glacial front was monitored directly from 1980 to 1995 (except 1986-1991), and airphotos show the position of the glacier in 1959. Hence, there was an excellent opportunity to study primary succession where there was adequate chronological control. Presently, a well preserved pre-Little Ice Age plant community and organically rich (paleo) soil are being released from the Twin Glacier, in addition to glacio-fluvial sediments, rendering this foreland study unique. A terrain age scheme was developed using direct retreat measurements, airphotos and Salix aging. The relationship between the clump diameters of Luzula confusa and terrain age was determined to provide an alternative to lichenometry. Vegetation cover was assessed in 1994 and 1995 using a stratified random design. TWINSPAN and canonical correspondence analysis (detrended and non-detrended) were used together to examine vegetation patterns in relation to environmental variables. Soil seed bank and seed rain patterns were examined in relation to the above-ground vegetation. The seed bank was sampled in 1994, including samples from paleo-soil and glacio-fluvial sediment. To assess the fall-winter seed rain, seeds were collected between mid-August 1994 and early June 1995, using seed traps (artificial turf). Winter seed rain was sampled by collecting snow-core samples in early June 1995. Terrain age accounted for most of the variation in species composition over the study area. By directional-replacement, the succession followed four main stages of dominance in 44+ years: mosses —> graminoid-forb —> deciduous shrub-moss —> evergreen dwarf-shrubmoss. There was little difference in the successional sequences exhibited by the vegetation growing on the paleo-soil compared to that growing on the glacio-fluvial sediment. The relationship between the Luzula confusa clump diameters and terrain age was logarithmic. Luzula confusa dominated the above-ground vegetation, as well as, the germinable seed bank and seed rain. The average germinable seed bank, fall-winter seed rain and winter seed rain densities were 367 ± 32, 384 ± 47 and 180 ± 53 seeds/m2, respectively. The seed bank was significantly positively correlated with the above-ground vegetation cover for all species combined, Luzula confusa (monocotyledons), dicotyledons and Papaver radicatum. There was a significant positive correlation between the fall-winter seed rain and the above-ground vegetation cover for Luzula confusa. No difference was detected in seed bank density between the paleo-soil and the glacio-fluvial sediment. However, the total vegetation cover was significantly higher on the paleo-soil. The relationship determined between Luzula confusa clump diameters and terrain age appears to be a valuable alternative to lichenometry on the Twin Glacier foreland. Although the results showed that directional-replacement is possible in high arctic environments, this mode of succession is likely atypical of such environments; the Twin Glacier foreland is located in one of the very few polar oases in the Queen Elizabeth Islands. The positive correlation of the above-ground vegetation with the seed bank and fall-winter seed rain suggests that colonization is largely constrained by seed availability. The winter seed rain appears to be relatively important on the Twin Glacier foreland. The higher vegetation cover on the paleosoil versus the glacio-fluvial sediment suggests, at least for some species, that the former provides conditions more favourable for establishment and growth than the latter."""@en ; edm:aggregatedCHO "https://circle.library.ubc.ca/rest/handle/2429/5817?expand=metadata"@en ; dcterms:extent "6722144 bytes"@en ; dc:format "application/pdf"@en ; skos:note "P R I M A R Y P L A N T S U C C E S S I O N O N T H E T W I N G L A C I E R F O R E L A N D , A L E X A N D R A F J O R D , E L L E S M E R E I S L A N D , C A N A D I A N H I G H A R C T I C by G L E N D A A . JONES B . S c , The University of Toronto, 1994 A THESIS S U B M I T T E D IN P A R T I A L F U L F I L L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F M A S T E R O F S C I E N C E in T H E F A C U L T Y O F G R A D U A T E STUDIES Department of Geography We accept this thesis as conforming to the required standard T H E U N I V E R S I T Y O F BRITISH C O L U M B I A February 1997 © Glenda A . Jones, 1997 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 Geography The University of British Columbia Vancouver, Canada D a t e February 20, 1997 DE-6 (2/88) Abstract Primary plant succession was examined on the foreland of the retreating Twin Glacier at Alexandra Fjord, Ellesmere Island. The position of the glacial front was monitored directly from 1980 to 1995 (except 1986-1991), and airphotos show the position of the glacier in 1959. Hence, there was an excellent opportunity to study primary succession where there was adequate chronological control. Presently, a well preserved pre-Little Ice Age plant community and organically rich (paleo) soil are being released from the Twin Glacier, in addition to glacio-fluvial sediments, rendering this foreland study unique. A terrain age scheme was developed using direct retreat measurements, airphotos and Salix aging. The relationship between the clump diameters of Luzula confusa and terrain age was determined to provide an alternative to lichenometry. Vegetation cover was assessed in 1994 and 1995 using a stratified random design. TWINSPAN and canonical correspondence analysis (detrended and non-detrended) were used together to examine vegetation patterns in relation to environmental variables. Soil seed bank and seed rain patterns were examined in relation to the above-ground vegetation. The seed bank was sampled in 1994, including samples from paleo-soil and glacio-fluvial sediment. To assess the fall-winter seed rain, seeds were collected between mid-August 1994 and early June 1995, using seed traps (artificial turf). Winter seed rain was sampled by collecting snow-core samples in early June 1995. Terrain age accounted for most of the variation in species composition over the study area. By directional-replacement, the succession followed four main stages of dominance in 44+ years: mosses —> graminoid-forb —> deciduous shrub-moss —> evergreen dwarf-shrub-moss. There was little difference in the successional sequences exhibited by the vegetation growing on the paleo-soil compared to that growing on the glacio-fluvial sediment. The relationship between the Luzula confusa clump diameters and terrain age was logarithmic. ii Luzula confusa dominated the above-ground vegetation, as well as, the germinable seed bank and seed rain. The average germinable seed bank, fall-winter seed rain and winter seed rain densities were 367 ± 32, 384 ± 47 and 180 ± 53 seeds/m2, respectively. The seed bank was significantly positively correlated with the above-ground vegetation cover for all species combined, Luzula confusa (monocotyledons), dicotyledons and Papaver radicatum. There was a significant positive correlation between the fall-winter seed rain and the above-ground vegetation cover for Luzula confusa. No difference was detected in seed bank density between the paleo-soil and the glacio-fluvial sediment. However, the total vegetation cover was significantly higher on the paleo-soil. The relationship determined between Luzula confusa clump diameters and terrain age appears to be a valuable alternative to lichenometry on the Twin Glacier foreland. Although the results showed that directional-replacement is possible in high arctic environments, this mode of succession is likely atypical of such environments; the Twin Glacier foreland is located in one of the very few polar oases in the Queen Elizabeth Islands. The positive correlation of the above-ground vegetation with the seed bank and fall-winter seed rain suggests that colonization is largely constrained by seed availability. The winter seed rain appears to be relatively important on the Twin Glacier foreland. The higher vegetation cover on the paleo-soil versus the glacio-fluvial sediment suggests, at least for some species, that the former provides conditions more favourable for establishment and growth than the latter. iii Table of Contents Abstract ii Table of Contents iv List of Tables vii List of Figures viii List of Plates x Acknowledgements xi Chapter 1 Literature Review and Rationale 1 1.1 Introduction 1 1.2 Models incorporating levels of environmental severity 2 1.3 Research objectives 5 Chapter 2 Terrain Age 7 2.1 Introduction 7 2.2 Study area 9 2.3 Materials and methods 12 2.3.1 Sampling design 12 2.3.2 Terrain dating 12 2.3.2.1 Direct positional methods 12 2.3.2.2 Phytometric method 13 2.3.3 Age calibration of Luzula confusa clumps 13 2.4 Results 14 2.4.1 Terrain dating 14 2.4.2 Age calibration of Luzula confusa clumps 18 2.5 Discussion 20 2.5.1 Terrain dating 20 2.5.2 Age calibration of Luzula confusa clumps 21 Chapter 3 Successional Patterns 24 3.1 Introduction 24 3.2 Study area 27 iv 3.3 Materials and methods 28 3.3.1 Sampling design 28 3.3.2 Vegetation sampling 28 3.3.3 Environmental factors 30 3.3.4 Statistical analysis 32 3.3.4.1 Succession patterns 34 3.3.4.2 Succession on paleo-soil versus glacio-fluvial sediment 35 3.4 Results 36 3.4.1 Succession patterns 36 3.4.1.1 Classification and Ordination 36 3.4.1.2 Species diversity 48 3.4.1.3 Pre-LIA vegetation community 52 3.4.2 Succession on paleo-soil versus glacio-fluvial sediment 52 3.4.3 Soil attributes 60 3.5 Discussion 66 3.5.1 Succession patterns 66 3.5.2 Application to succession models 68 3.5.3 Soil nutrients 74 3.5.4 Succession on paleo-soil versus glacio-fluvial sediment 75 3.6 Summary 76 Chapter 4 Seed Dynamics 77 4.1 Introduction 77 4.2 Study area 79 4.3 Materials and methods 80 4.3.1 Seed bank 80 4.3.2 Fall-winter seed rain 82 4.3.3 Snow-core germination 84 4.4 Results • 84 4.4.1 Seed bank 84 4.4.2 Fall-winter seed rain 88 4.4.3 Snow-core germination 91 v 4.5 Discussion 95 4.5.1 Seed bank 95 4.5.2 Fall-winter seed rain and snow-core germination 97 4.6 Summary 103 Chapter 5 Conclusions 105 References cited 108 Appendices 119 vi List of Tables Table 2.1: Retreat rates of the snout of the western lobe of the Twin Glacier from 1959 to 1995 15 Table 2.2: a) Ages of individual Salix arctica specimens collected over 270 m are shown in relation to belt number and 10 m distance intervals leading away from the glacier front 17 Table 3.1: T W I N S P A N species and sample groups for the overall vegetation data set. A dendrogram with eigenvalues is shown for the species classification 38 Table 3.2: Summary of ordination statistics for the overall, paleo-soil and glacio-fluvial data sets 41 Table 3.3: Weighted correlations between the environmental variables and D C C A axes 1 and 2 for the overall vegetation data set 42 Table 3.4: Spearman's correlation coefficients for the D C C A environmental variables 42 Table 3.5: T W I N S P A N classification for vegetation on paleo-soil showing species and sample groups. A dendrogram is shown with eigenvalues for the species classification 54 Table 3.6: T W I N S P A N classification for vegetation on glacio-fluvial sediment showing species and sample groups. A dendrogram is shown with eigenvalues for the species classification 55 Table 3.7: Weighted correlations between the environmental variables and the first two C C A ordination axes for the paleo-soil and glacio-fluvial data sets 59 Table 3.8: Spearman's rank correlation coefficients for the a) paleo-soil and b) glacio-fluvial data sets 59 Table 4.1: Mean (+SE) seed bank densities (seeds/m2) for increasing distance and terrain age from the Twin Glacier snout 86 Table 4.2: Tukey's multiple comparison results showing the terrain ages (years) which had significantly different seed bank densities for all species and for Luzula confusa; significance level, a = 0.05 86 Table 4.3: Mean ( ± S E ) fall-winter seed rain densities (seeds/m2) for increasing distance and terrain age from the Twin Glacier snout 90 vii List of Figures Figure 1.1: Map showing the location of the Twin Glacier foreland (©) , Alexandra Fjord ( 7 8 ° 5 3 ' N ; 7 5 ° 5 5 ' W ) and extensive survey sites (•) on east-central Ellesmere Island. Modified from Henry et al. (1986) 6 Figure 2.1: Number of Salix arctica bud scale scars in relation to distance from the glacier snout 16 Figure 2.2: Clump diameter of Luzula confusa in relation to terrain age 19 Figure 3.1: Schematic representation of the sampling design on the Twin Glacier foreland 29 Figure 3.2: D C C A species ordination for the overall vegetation data, with environmental variables represented by arrows 43 Figure 3.3: D C C A ordination showing T W T N S P A N sample group centroids labelled with their respective terrain age (see Table 3.1) 45 Figure 3.4: D C C A ordination of T W I N S P A N species group (see Table 3.1 and Figure 3.2) centroids with bars representing 95% confidence intervals. The characterizing terrain ages (according to Table 3.1) for each species group are shown 47 Figure 3.5: Percent cover of the dominant species in each of the four successional stages in relation to terrain age 49 Figure 3.6: Dominance-diversity curves for the four successional stages on the Twin Glacier foreland 50 Figure 3.7: Species richness in relation to terrain age for vascular plants, mosses and the combination of the two growth forms 51 Figure 3.8: Percent cover for crustose, fruticose and foliose lichens in relation to terrain age 53 Figure 3.9: C C A ordination for the paleo-soil vegetation data set with environmental variables represented by arrows 56 Figure 3.10: C C A ordination for the glacio-fluvial vegetation data set with environmental variables represented by arrows 57 viii Figure 3.11: Species richness in relation to terrain age for vegetation growing on paleo-soil versus glacio-fluvial sediment 61 Figure 3.12: Mean (±SE) (n = 4) nitrate-nitrogen, available phosphorous and exchangeable potassium levels in relation to terrain age 62 Figure 3.13: Mean (±SE) soil p H (n = 3), organic matter content (n = 3) and soil moisture content (n = 5) in relation to terrain age for paleo-soil versus glacio-fluvial sediment 64 Figure 3.14: Mean (±SE) percent cover of a) the total vegetation and b) Psilopilum cavifolium in relation to terrain age for the paleo-soil versus the glacio-fluvial sediment 65 Figure 4.1: Mean (+SE) germinable seed bank in relation to terrain age for a) all species combined and b) Luzula confusa 87 Figure 4.2: Germinable seed bank in relation to plant cover for a) all species combined, b) Luzula confusa (monocotyledons), c) dicotyledons, and d)Papaver radicatum. Pearson's (r) or Spearman's (rs) correlation coefficients are shown with their respective p-values 89 Figure 4.3: Mean (+SE) fall-winter germinable seed rain fox Luzula confusa in relation to distance from the glacier snout 92 Figure 4.4: Mean (±SE) germinable seed density from snow-cores for Luzula confusa in relation to terrain age 94 Figure 4.5: Mean (+SE) snow depth in relation to terrain age, on June 2, 1995 94 ix List of Plates Plate 2.1: Airphoto (1959) of the southern (A) and western (B) lobes of the Twin Glacier, Alexandra Fjord, Ellesmere Island ( 7 8 ° 5 3 ' N ; 7 5 ° 5 5 ' W ) 10 Plate 2.2: A n aerial view showing the lichen trimline in front of the western lobe of the Twin Glacier, Alexandra Fjord 11 Plate 2.3: Study area located in front of a portion of the western lobe of the Twin Glacier, Alexandra Fjord 11 x Acknowledgements I am indebted to Josef Svoboda and Esther Levesque for introducing me to the astonishing world of the High Arctic. A very special thanks goes to Esther Levesque for her continual support, assistance and encouragement. The tremendous support and patience of my family and my dear Chris Fratton is valued greatly. The guidance and advice provided by Greg Henry throughout the duration of my study at the University of British Columbia was more than appreciated. I am grateful for the insightful comments made by Greg Henry, Roy Turkington and Esther Levesque on my thesis drafts. Dr. W . B . Schofield cannot be thanked enough for his expert identification of all bryophytic taxa presented in this thesis. Many thanks go to Petra Lange, Jonathan Henkelman, Jackie Bastick, Thor Smedstad, and Kristie Trainor for their help in the field and/or in the lab. Thanks also go to Gary Sprules, Va l LeMay and Anne Tolvanen for providing statistical advice. Nutrient analysis by Janet Randall from Peace Growers' Lab, radiocarbon analyses by the Environmental Isotope Laboratory, University of Waterloo and the help from the staff of the Botanical Gardens Nursery, University of British Columbia was appreciated. Research was funded by a Natural Sciences and Engineering Research Council of Canada grant to G . Henry, a Royal Canadian Geographic Society Studentship and Northern Scientific Training Program funds. Logistic support was provided by the Polar Continental Shelf Project of the Department of Energy, Mines and Resources. Thanks to the Royal Canadian Mounted Police for the use of the Alexandra Fjord buildings. xi Chapter 1 Literature Review and Rationale 1.1 Introduction The concept of vegetation succession has long been of great importance to ecological studies. In the last century successional patterns and the mechanisms driving succession have been relatively well documented through the numerous primary and secondary succession studies completed in temperate and low arctic regions. However, there is a paucity of information regarding successional patterns and their mechanisms in extreme environments such as the High Arctic, due to the lack of studies carried out in such environments. The High Arctic is described as an extreme environment because of its low temperatures and short growing seasons which result in low productivity and low ground cover (Chapin, 1987; Edlund and Alt, 1989; Shaver and Kummerow, 1992). The succession studies which have been completed in the High Arctic are almost exclusively secondary succession or vegetation recovery work (e.g. Babb and Bliss, 1974; Barrett and Schulten, 1975; Bliss and Grulke, 1988; Forbes, 1993). Primary succession has seldom been observed in high arctic environments due its rarity and the slow rate of vegetation change (e.g. Tishkov, 1986; Svoboda and Henry, 1987). For this reason, there exists very few studies and therefore little information on the patterns and processes of primary succession in such regions. However, the rapid melting of most glaciers in the Canadian High Arctic over the past 30-40 years, in response to general climatic warming since the 1 end of the Little Ice Age (ca. 1850), has exposed new surfaces which provide an excellent opportunity to observe primary plant succession in a high arctic environment. Many succession theories and models have been developed in order to understand the underlying patterns and processes which control successional sequences in temperate environments (Clements, 1916; Gleason, 1926; Egler, 1954; Connell and Slatyer, 1977; Drury and Nisbet, 1973; Grime, 1977; Noble and Slatyer, 1980; Tilman, 1985; Huston and Smith, 1987). Because the general mechanisms (e.g. competition, facilitation, life history traits and abiotic factors) used to explain succession vary in importance with differing degrees of environmental severity (Svoboda and Henry, 1987; Walker and Chapin, 1987; Matthews, 1992; Chapin et al., 1994), these temperate-based models are not entirely applicable to extreme environments, such as the High Arctic. For example, in marginal environments the physical environment is thought to be much more important than the biological processes, such as competition, in controlling plant establishment and growth (Muller, 1952; Savile, 1960; Billings and Mooney, 1968; Billings, 1987; Svoboda and Henry, 1987; Matthews, 1992). The following section describes the prominent succession models which have incorporated varying levels of environmental severity. 1.2 Models incorporating levels of environmental severity Svoboda and Henry (1987) introduced a set of models proposing that as environmental stress increases, succession shifts from species replacement to species establishment and survival. They proposed three different models representing increasing degrees of environmental severity: directional-replacement, directional-nonreplacement 2 and nondirectional-nonreplacement, respectively. These models are based on the balance between biological driving forces (BDF) (e.g. germination capacity, establishment ability) and environmental resistance (ER) (e.g. temperatures, growing season length, soil characteristics, wind). Directional-replacement is classical directional succession with species replacement in serai stages, each stage distinguished by a dominant set of species, in which B D F > E R . This model is similar to that of Egler's (1954) relay floristics model and Connell and Slatyer's (1977) facilitation model, in which species replacement takes place due to the modification of the environment by the preceding set of species. When B D F s are just slightly greater than E R , directional-nonreplacement proceeds by a slow directional expansion of the invading populations without species replacement. Nondirectional-nonreplacement denotes the case where B D F < E R in which only a minute fraction of the potential species succeed in survival. Those that do survive occur with variable abundance, and other invading species may occasionally establish, but not permanently. The ideas proposed by Svoboda and Henry (1987) are similar to those proposed by Muller (1952) who considered non-selective and selective autosuccession for increasingly severe environments along altitudinal gradients in Scandinavia. Although Muller's (1952) ideas were developed in the context of secondary succession, it is clear that directional-nonreplacement and nondirectional-nonreplacement succession are comparable to non-selective and selective autosuccession, respectively. Matthews (1992) also recognized the importance of differing levels of environmental severity, when considering the 'geoecological' model of succession. This model is based on the general agreement that succession proceeds with a combination of allogenic and autogenic processes (e.g. Muller, 1952; Billings and Mooney, 1968; Bliss, 3 1987; Svoboda and Henry, 1987; Whittaker, 1991; Whittaker, 1993; Helm and Allen, 1995). The model essentially proposes that the relative importance of allogenesis decreases as the relative importance of autogenesis increases through successional time. Matthews (1992) defines autogenesis as biological processes involving life history traits, facilitation and competition, and allogenesis as external changes in the physical environment such as, acidification, nutrient depletion, freeze-thaw processes, hydrological changes and the formation of stone pavement. Matthews (1992) maintained that with increasing environmental resistance, allogenic processes become more important, while autogenic processes become less important. Allogenic processes are perceived to dominate in extremely severe environments, such as polar deserts. The allogenic and autogenic processes are respectively analogous to the environmental resistance and biological driving forces put forth by Svoboda and Henry (1987). For example, where autogenesis > allogenesis a directional-replacement succession would occur and where autogenesis < allogenesis succession would proceed by nondirectional-nonreplacement. The main limitation of the above models, is that they do not deal directly with the mechanisms driving succession. Walker and Chapin (1987), however, proposed a theoretical framework of hypotheses which deals directly with successional processes. Type of succession and environmental severity (severe = low resource availability and favourable = high resource availability) were used to examine the importance of some of the major successional processes (e.g. facilitation, competition, longevity seed arrival) determining succession change during the colonization, maturation and senescence stages of succession. The authors emphasized that succession is governed by a combination of 4 individual process and that certain processes predominate at different successional stages and at different levels of environmental severity. Knowledge regarding primary plant succession in high arctic environments is limited. As a result, the models which incorporate environmental severity, at present, have little meaning. Primary succession work completed in the High Arctic would increase our limited knowledge of successional patterns and processes in extreme environments, which would eventually allow us to examine and improve these models. 1.3 Research objectives Research was conducted on the Twin Glacier Foreland, Alexandra Fjord, Ellesmere Island, Canadian High Arctic ( 7 8 ° 5 3 ' N ; 7 5 ° 5 5 ' W ) (Figure 1.1). Three main objectives were addressed: (1) To obtain reliable terrain age estimates along the successional gradient, so as to increase the pre-existing chronological control and therefore provide a basis on which to examine successional patterns and seed dynamics (Chapter 2). (2) To describe the successional patterns and to examine the vegetation-environment relationship (Chapter 3). (3) To examine the seed bank and seed rain patterns to gain some insight into the processes driving the succession (Chapter 4). 5 6 Chapter 2 Terrain Age 2.1 Introduction The term recently deglaciated terrain is most often used to define terrain exposed since the end of the Little Ice Age, rather than older landscapes deglaciated earlier in the Quaternary (Matthews, 1992). The greatest benefit in using recently deglaciated terrain (viz. glacier forelands) as a study site for primary succession is the positive relationship between increasing distance from the glacier and the time period for soil and vegetation development. This permits succession patterns to be interpreted through a chronosequence approach in which temporal change is represented by changes along a spatial (distance) gradient. Although the chronosequence approach does not provide direct evidence of succession patterns, the time factor can be quantified rather accurately by using reliable estimates of terrain age (years since released from glacial ice) as a surrogate for successional time. Airphotos and direct measurements of glacial retreat are often used as a basis for determining estimates of terrain age. Age estimates can also be determined from plants that display yearly growth increments. For example, the age of woody species, such as Salix spp, can be determined on site by counting the number of bud scale scars along a branch (S0renson, 1941; Wijk, 1986) or in the laboratory by counting the number of annual rings (e.g. Beschel and Webb, 1963). The age of a plant determines only a minimum estimate of the age of the terrain upon which the plant grows, due to the time lag that exists between the year of glacial retreat and the year of species 7 establishment. Palmer and Miller (1961) have suggested that the addition of one year to the age determined by counting the number of Salix bud scale scars yields a reasonably accurate date of glacial retreat at the Rotmoos Gletscher in Obergurgl, Austria. However, where environmental conditions are limiting, the time lag between terrain exposure (by glacial retreat) and species establishment may be longer than one year. In spite of the fact that lichenometry is one of the most popular methods for dating glacier forelands, this technique is often unsuitable for young terrain (< 200 years), as it contains little crustose lichen due to the extensive amount of time required for lichen thalli (e.g. Rhizocarpon geographicum) to become established. Benedict (1989) used the diameters of Silene acaulis clumps to determine the shape of its growth rate curve and to calibrate this species for dating Little Ice Age moraines on the Arapaho Glacier foreland, Colorado. The rapid colonization, slow radial growth and large maximum diameters made Silene acaulis an excellent alternative to lichenometry. The validity of this clump diameter dating method is backed by Whittaker (1993), who used clump diameters of several pioneer species, as terrain age surrogates for part of a study on population-age structure, on the Storbreen glacier foreland, Norway. During this research, the radially growing forb, Luzula confusa, was examined as an alternative to lichenometry. The objectives of this study were: 1) to produce a chronological terrain age outline for the Twin Glacier study area by using a combination of methods including direct retreat measurements, Salix aging, and the use of data obtained from Bergsma et al. (1984); and 2) to determine the relationship between Luzula confusa clump diameters and terrain age. 8 2.2 Study area The Twin Glacier, located at the southern end of the Alexandra Fjord lowland, is an outlet of the large Agassiz ice cap and is characterized by its two tongues extending from the south and west (Plate 2.1). Research was conducted on the Twin Glacier foreland, Alexandra Fjord, Ellesmere Island in the Canadian High Arctic ( 7 8 ° 5 3 ' N ; 7 5 ° 5 5 ' W ) (Figure 1.1; Plate 2.1). The 8 k m 2 gently sloping lowland adjacent to Alexandra Fjord is bounded by steep cliffs to the east and west, by glacial tongues to the south and by the Arctic Ocean to the north. The maximum glacial extent of the Little Ice Age (ca. 1850 A.D.) marks the limit of the recently deglaciated terrain of the Twin Glacier foreland. This limit is delineated by a distinct lichen fringe- or trimline: the recently deglaciated area is characterized by lichen-free rocks, while the area beyond the glacial limit is characterized by lichen covered rocks (Plate 2.2). Research was carried out on gneiss-granite based terrain in front of a portion of the western lobe of the Twin Glacier (Plate 2.3) which has retreated approximately 210 m in the last 36 years. 2.3 Materials and methods 2.3.1 Sampling design The study area extended 300 m away from the glacier snout to just beyond the lichen trimline, and covered a breadth of 120-180 m. The study area was stratified into 9 Plate 2.1: Airphoto (1959) of the southern (A) and western (B) lobes of the Twin Glacier, Alexandra Fjord, Ellesmere Island ( 7 8 ° 5 3 ' N ; 7 5 ° 5 5 ' W ) . As of 1995, the western lobe had retreated approximately 210 m from its 1959 position to the position shown by the black dotted line. 10 Plate 2.2: A n aerial view showing the lichen trimline in front of the western lobe of the Twin Glacier, Alexandra Fjord. The recently deglaciated area is characterized by lichen-free rocks, while the area beyond the glacial limit is characterized by lichen covered rocks. Plate 2.3: Study area located in front of a portion of the western lobe of the Twin Glacier, Alexandra Fjord. 11 thirty 10 m wide belts, running parallel to the glacier terminus. Aside from the convenience of a 10 m wide unit, the belt width was chosen based on recent (1992-1995) retreat rates of the glacial snout (i.e. ~ 10 m/yr); it was understood that the belts would not all represent a terrain age based on these recent retreat rates. The 10 m belts were used as fixed points in a systematic estimation of terrain age from the dating analyses discussed below. 2.3.2 Terrain dating 2.3.2.1 Direct positional methods Estimates of terrain age were obtained using a combination of direct methods, including in situ measurements of the glacial terminus position and interpolation of airphoto data. A series of reference stakes marking the position of the glacier terminus were used to obtain retreat fates for the years 1981-1983 (Bergsmae? al., 1984) and 1993-1995. The rate of ablation between 1980 and 1983 was determined using nine reference stakes set up in 1980 across a 260 m wide portion of the foot of the glacier, while five stakes marked a 100 m wide portion of the terminus for each of the years between 1992 and 1995. A n average retreat rate of the glacial snout was determined for the years 1986-1992, a period when the position of the glacier was not measured directly, by measuring the distance between the 1985 and 1992 reference stakes and dividing this distance by 7 years. For the years 1959-1981, an average retreat rate of the glacial front 12 was determined by adjusting airphotos taken in those years to the same scale, and aligning them by using distinctive topographic features such as the lichen trimline (Bergsmaer al., 1984). 2.3.2.2 Phytometric method In each of the 10 m belts between the 80 and 270 m marks leading from the glacier snout, the number of terminal bud scale scars were counted on the largest individual specimen of Salix arctica; each scar represented a yearly growth increment and therefore one year of glacial retreat. The relationship between the number of bud scale scars (i.e. years of glacial retreat) and the distance from the glacier front was determined in order to produce an ablation rate of the glacial front that could be compared to the rate determined from airphoto interpretation by Bergsmae? al. (1984). 2.3.3 Age calibration of Luzula confusa clumps In 1994, for each 10 m belt beginning at 20 m from the glacier front and terminating at 250 m, the 10 largest Luzula confusa clumps were marked with labelled metal tags. Each clump was measured across its outermost leaves in two orthogonal directions, to the nearest mm. After calculating the mean diameter for each clump, a growth rate curve was produced by plotting clump diameter against terrain age. The terrain ages used for the plot were those determined from the 'Terrain dating' section. 13 2.4 Results 2.4.1 Terrain dating Table 2.1 shows the retreat rates of the glacial front determined from airphoto interpretation and the use of reference stakes. The rate of retreat appeared to have increased in the last 10 years, compared to the previous 26 years. It should be noted, however, that it is unknown what the exact retreat rates were within the 1986-1992 period, because direct measurements were not taken within these years. It should also be pointed out that the unusually high retreat rate in 1993 was attributed to the fact that the 1992 reference stake was set out during the middle of the growing season (July 12, 1992), as opposed to the end of the growing season (mid-late August) when melting has, for the most part, subsided. For the years between 1959 and 1981, the average retreat rate (4.0 m/yr) determined from airphotos (Bergsma et al., 1984) was lower than the average rate (6.9 m/yr) determined from the exponential relationship between Salix bud scale scars and distance from the glacier snout (Figure 2.1). In comparing the actual Salix ages (Table 2.2a) for belts in which the actual age was known, it was noted that establishment time was variable; for instance, it is known that the actual ages of belts 12 and 21 were 14-15 (1980/81) and 36 (1959) years old, respectively, according to reference stakes and airphotos (Table 2.2b). However, according to the Salix ages, belts 12 and 21 were respectively 2-3 and 8 years younger than the actual ages (14-15 and 36 years), indicating that the retreat rate determined from Salix aging (Figure 2.1) is not entirely reliable. For this reason the 4.0 m/yr rate determined by Bergsma et al. (1984) was used for interpolating ages between 14 Table 2.1: Retreat rates of the snout of the Twin Glacier (western lobe) from 1959 to 1995. Mean (±SE) rates were calculated, where sample size was greater than one. Sample size denotes the number of reference stakes used for direct retreat measurements. Year Sample size Retreat rate (m/yr) 1959-1981 n/a 4 . 0 V 6 . 9 * 1981 9 5 . 6 ± 0 . 8 f 1982 9 5.9 ± 1.9f 1983 9 4 . 7 ± 0 . 9 f 1984 1 4.0 1985 1 2.9 1986-1992 n/a 8.6 1993 5 11.2 + 0.8 1994 5 9.0 ± 0 . 6 1995 5 7.6 ± 1.0 Rates obtained from Bergsma et al. (1984). Rate determined from Salix arctica bud scars (see Figure 2.1) 15 Figure 2.1: Number of Salix arctica bud scale scars in relation to distance from the glacier snout. From this relationship, the ablation rate of the glacier front between 1959 and 1980/81 was determined to be 6.9 m/yr 16 Table 2.2: a) Ages of individual Salix arctica specimens collected over 270 m are shown in relation to belt number and 10 m distance intervals leading away from the glacier front. Dashes (-) indicate where Salix arctica was not found and asterisks (*) indicate where samples were not collected due to time constraints, b) Terrain age and year in relation to belt and 10 m distance intervals. The highlighted values represent years which were marked directly with reference stakes or with the use of airphotos. The years which are not highlighted represent those which were interpolated between the known (highlighted) years and those which were determined using Salix age measurements. A correction of 8 years was added to the original Salix age (see text for further explanation). The lichen trimline is indicated by the dotted line. a) b) Belt Distance (m) Salix age (y) Belt Distance (m) Terrain age (y) Year 1 0-10 - 1 0-10 0.25 1995a 2 10-20 - 2 10-20 1 1994a 3 20-30 - 3 20-30 2 1993a 4 30-40 - 4 30-40 3 1992a 5 40-50 - 5 40-50 4 1991 6 50-60 - 6 50-60 5 1990 7 60-70 - 7 60-70 7 1988 8 70-80 5 8 70-80 8 1987 9 80-90 7 9 80-90 9 1986 10 90-100 - 10 90-100 11/10 1984/85a 11 100-110 9 11 100-110 13/12 1982/833 12 110-120 12 12 110-120 15/14 1980/81a 13 120-130 11 13 120-130 17 1978 14 130-140 13 14 130-140 20 1975 15 140-150 17 15 140-150 22 1973 16 150-160 14 16 150-160 25 1970 17 160-170 14 17 160-170 27 1968 18 170-180 16 18 170-180 30 1965 19 180-190 17 19 180-190 32 1963 20 190-200 25 20 190-200 35 1960 21 200-210 28 21 200-210 36 1959b 22 210-220 * 22 210-220 38 1957 23 220-230 32 23 220-230 32+8=40 1955c 24 230-240 * 24 230-240 42 1953 25 240-250 35 25 240-250 35+8=43 1952c 26 250-260 * 26 250-260 44 1951 27 260-270 36 27 260-270 36+8=44 1951c 28 270-280 44+ -1951 29 280-290 44+ -1951 30 290-300 44+ -1951 dates determined using reference stakes dates determined using airphotos cdates determined using Salix bud scar measurements 17 belts 12 and 21 (Table 2.2b) rather than the 6.9 m/yr rate determined from the Salix ages (Figure 2.1). Salix age estimates, however, were used after belt 21, as this was the only dating method used beyond this point. Because the difference between the calculated Salix age and the actual age of belt 21 was 8 years (a result of the time lag between terrain exposure and plant establishment), this difference was used as a correction factor for the remaining Salix ages beyond belt 21. Table 2.2b also shows the interpolated ages and years between 1985 and 1992, which were determined using the average retreat rate for 1986-1992 shown in Table 2.1. 2.4.2 Age calibration of Luzula confusa clumps The observed Luzula confusa clump diameters increased logarithmically over time (r2 = 0.78; Figure 2.2). The curve shows that young Luzula confusa grew rapidly up to about 10 years, after which time the growth rate began to level off. The growth rates ranged from approximately 14 mm/yr in the first 3 years to 1 mm/yr over the last 3 years. 18 Terrain age (years) Figure 2.2: Clump diameter of Luzula confusa in relation to terrain age. 19 2.5 Discussion 2.5.1 Terrain dating The difference in the average glacial front retreat rate determined from Salix aging (6.9 m/yr) and airphoto interpretation (4.0 rn/yr), for the 1959-1981 range, may be due to the apparent variability in the lag time between terrain exposure and Salix establishment. Table 2.2 indicates that in 1980 the lag time appeared to be 2-3 years, whereas in 1959, the lag time was 8 years. The lack of previously established terrain dates does not allow for further speculation. The discrepancy in retreat rates may also be due to the difficulty in determining the actual number of bud scale scars near the base of older stems. This method of determining age was used because it can be done in situ in relatively short period of time with negligible destruction, while maintaining an adequate sample size. Counting annual rings of wood, another dating method, also has disadvantages, especially for dwarf shrubs (Callaghan and Emanuelsson, 1985), such as Salix arctica. Aside from the destruction to obtain the sample, rings are generally difficult to discriminate because of their narrowness. In addition, short growing seasons may hinder growth, such that annual rings may not form and the inner rings on old stems are often unidentifiable on old stems. The relatively high retreat rates during 1986-1995 compared to the years 1959-1985 is likely an indication of the gradual thinning of the terminus, rather than a response to a successive set of warmer years. As glacial ice thins and decreases in area more rock/sediment surface is exposed, taking the place of the ice surface. This ultimately decreases the albedo, and increases the ambient air temperature of the basin, causing glacial ice to retreat at an increased non-linear rate over time (N. Eyles pers. comm., 20 1996). This exposes the main shortcoming of the terrain dating scheme developed for the Twin Glacier foreland: the assumption that the retreat rate of the glacial front remained constant for the years between 1959 and 1981, and 1986 and 1992. Despite this drawback, the terrain dating scheme is more than adequate for the overall purpose of providing a relative basis on which to examine the change in vegetation (Chapter 3) and seed bank and dispersal patterns over time (Chapter 4). 2.5.2 Age calibration of Luzula confusa clumps Luzula confusa dating was found to be a valuable alternative to lichenometry on the Twin Glacier foreland. Luzula confusa is widespread over the foreland, colonizes rapidly and grows radially up to diameters of 20 cm, before the center portion of the clump (genet) begins to die off. By determining its growth rate, Luzula confusa, like Rhizocarpon geographicum (e.g. Beschel, 1963) and Silene acaulis (Benedict, 1989), can also be calibrated for future use in terrain dating. The age of Luzula confusa can be estimated more accurately by multiplying the number of tillers along the longest rhizome branch by the replacement time of ~ 3.5 years (Addison and Bliss, 1984). However, the amount of time required to complete such meticulous work, while maintaining an adequate sample size, made this analysis impossible. The benefit of using Luzula confusa as a species for terrain dating is discussed using the criteria set out by Matthews (1978), where it is stated that the optimal species for dating are highly time dependent, abundant and have a wide environmental tolerance. Although Figure 2.2 shows that Luzula confusa colonizes within two years of glacial 21 retreat, it actually colonizes within one year {personal observation), indicating that Luzula confusa is extremely time dependent. Sampling began on two-year-old terrain as an adequate number of Luzula specimens was not found on younger terrain. The abundance of Luzula confusa overwhelmed that of any other vascular species on terrain younger than 25 years, and was still relatively abundant on the oldest terrain sampled (see Chapter 3). However, the environmental tolerance of Luzula confusa on the Twin Glacier foreland appeared to become limited on older terrain, indicated in Figure 2.2 by the flattening of its growth rate curve. Matthews (1978) reported that many species are only useful for phytometric dating over a 'limited range of environment'. In a study encompassing terrain older than 50-60 years, the use of successionally mature species would be required to obtain reliable surface ages on the Twin Glacier foreland. Salix arctica, Cassiope tetragona and Saxifraga oppositifolia are the prime possibilities for the Twin Glacier foreland and similar high arctic forelands. Beschel and Webb (1963) described growth ring studies of Salix arctica on Axel Heiberg Island and Palmer and Miller (1961) used bud scale scars to age three Salix species in Austria (1950 m asl). Annual growth increments for Cassiope tetragona are identified by examining differences in leaf growth: leaves produced early and late in the season are shorter than those produced in mid-season (Callaghan et al., 1989); or by patterns in internode lengths between leaves (Johnstone and Henry, 1997). Desrosiers (1991) demonstrated, in a site similar to Alexandra Fjord that Saxifraga oppositifolia produces two new sets of leaves each year and used this fact to show that ages could be obtained by counting leaf scars on stems. Despite its apparent limited environmental tolerance on the Twin Glacier foreland, Luzula confusa showed to 22 be an excellent species for dating recently deglaciated terrain in terms of its rapid colonization and overwhelming abundance. The usefulness of the Luzula confusa growth rate curve for dating on other glacier forelands is unknown, due to the wide variation in the type and degree of environmental influence from site to site (Matthews, 1978; Benedict, 1989). For example, edaphic factors, snow depth, aspect, slope, climate and competition vary widely among sites. Although it is known that Luzula confusa can be found in a wide range of habitats (Polunin, 1959; Porsild and Cody, 1980), and that it tends to predominate in sites of intermediate moisture with abundant cryptogams (Addison and Bliss, 1984), there is still much to learn about the ecology of this species. Until more is known about the environmental tolerance of Luzula confusa, the determined growth rate curve should be used with some caution. 23 Chapter 3 Successional Patterns 3.1 Introduction Little is known about primary plant succession in severe environments, such as the High Arctic which is characterized by cool, short growing seasons and nutrient-poor soils. Most of the research, and therefore most of the knowledge on succession stems from temperate regions where environmental conditions are favourable. Few high arctic primary succession studies are known to date. Bliss and Gold (1994) investigated primary succession along a high arctic coastline on Truelove Lowland, Devon Island. Relatively recently emerged surfaces from the seacoast provided a series of sites that were assumed to represent a chronosequence of successively older sites that increased with elevation. In Svalbard, Tishkov (1986) examined primary succession on raised beach ridges, emergent salt marshes and recently deglaciated moraines on the Werenskiold Glacier foreland. In the last century, a considerable number of primary succession studies have been completed on glacier forelands in relatively low resistance environments (sensu Svoboda and Henry, 1987). The positive relationship with increasing distance from a glacier and the time period for soil and vegetation development is a clear advantage in using forelands to study primary succession. The most well known glacier foreland succession studies have come from Glacier Bay, Alaska, where the earliest work was established by W.S. Cooper in 1916 (Cooper, 1923). Succession studies on the Glacier Bay foreland have continued to present day where the focus is now on the mechanisms driving succession 24 (e.g. Chapin et al, 1994, Fastie, 1995). The Mount Robson moraines in British Columbia have been the subject of many successional studies (e.g. Cooper, 1916; Tisdale et al, 1966; Blundon, 1989) and numerous papers have been published from the relatively recent studies of the Storbreen glacier foreland, Norway (e.g. Matthews, 1978; Matthews, 1979; Matthews and Whittaker, 1987; Whittaker, 1989; Whittaker, 1991; Whittaker, 1993). Examples of other glacier foreland studies include work completed by Viereck (1966) at the Muldrow Glacier in Alaska, Persson (1964) at the Skaftafellsjokull Glacier in Iceland, Smith (1982) at a deglaciated headland on A n vers Island in Antarctica, Vetaas (1994) at the B0ldalen glacier in Norway and Helm and Allen (1995) at Exit Glacier in Alaska. The rapid retreat of Canadian high arctic glaciers, due to general climatic warming since the end of the Little Ice Age (ca. 1850 A.D.) , has exposed new surfaces, and therefore created excellent opportunities to study primary plant succession in high arctic environments. Preserved plant communities and patches of organically rich soil are often associated with recently deglaciated terrain in the High Arctic (Bergsma et al., 1984), due to the nature of advance of polar glaciers (cold-based). Polar outlet glaciers advance predominantly by internal deformation rather than basal sliding (Young, 1989) as they are usually frozen to the underlying bedrock because of the permafrost. Temperate (warm-based) glaciers, on the other hand, advance predominantly by basal sliding causing tremendous erosion to the underlying substrate, leaving only infertile mineral sediments behind (Lawson, 1995). The sliding of the ice mass is due to the melting that occurs at the base of the ice as a result of the pressure of the overlying ice (Young, 1989). Well preserved surficial features have been found in both the Arctic and the Antarctic, and radiocarbon dates have indicated the existence of plant communities which 25 preceded glacial advance during the Little Ice Age (1550-1850 A.D.) . Preserved tundra polygons emerging from beneath glacial ice have been reported in Greenland (Goldthwait, 1960; Swinzow, 1962), and on Northern Baffin Island (Falconer, 1966). Goldthwaith (1960) and Falconer (1966) also noted preserved mosses and lichens. On Ellesmere Island, Beschel (1961) observed intact lichens and plants near melting glaciers and Blake (1981) used preserved debris from Salix spp, Dryas integrifolia and various mosses to examine neoglacial fluctuations of glaciers. Fossil polygons, lichens (Schyatt, 1961) and re-exposed moss banks (Collins, 1976; Fenton, 1982; Smith, 1982) have also been observed in the Antarctic. Presently, a well preserved pre-Little Ice Age (LIA) plant community (Bergsma et al., 1984) and organically enhanced soil (paleo-soil), are being released from the Twin Glacier, in addition to glacio-fluvial sediments. It is thought that this pre-LIA community was entombed in ice between 1410 and 1690 A . D . (Bergsma et al., 1984) and possibly as early as 1120 A . D . (Appendix 1). It is proposed that this relict community was initially killed by the covering of a permanent snow bank (which eventually developed into glacial ice) and not by the glacial movements themselves or temperature limitations prior to the L I A advance (Havstrom et al., 1995). The paleo-soil is characterized by its dark colour and fine texture versus the larger grained, lighter coloured mineral substrate of the glacio-fluvial sediment. Most of the species in this relict community are so well preserved that they are identifiable to the species level; some were observed with green foliage and a few species could still be found with flowers fully intact (Bergsma et al., 1984; pers. obs.). To evaluate the successional patterns of vascular plants and mosses on the Twin Glacier foreland, the following objectives were addressed: 1) to determine if the 26 succession is directional and if the succession shows evidence of species replacement, 2) to determine which environmental factors are most important in explaining the successional patterns, and 3) to determine if there are differences between the paleo-soil and the glacio-fluvial sediment in terms of vegetation cover and successional patterns. 3.2 Study area Research was conducted on the Twin Glacier foreland, Alexandra Fjord, Ellesmere Island in the Canadian High Arctic ( 7 8 ° 5 3 ' N ; 7 5 ° 5 5 ' W ) (Figure 1.1; Plate 2.1). The 8 km 2 gently sloping lowland adjacent to Alexandra Fjord is bounded by steep cliffs to the east and west, by glacial tongues to the south and by the Arctic Ocean to the north. These features, in part, create warm conditions and a large water supply giving rise to a polar oasis, where vegetation diversity, cover and plant production are much greater than the surrounding depauperate polar deserts and semi-deserts (Freedman et al., 1994). Polar oases are rare, as they cover only 1-2% of the Queen Elizabeth Islands (Babb and Bliss, 1974). The Twin Glacier, located at the southern end of the Alexandra Fjord lowland, is an outlet of the large Agassiz ice cap and is characterized by two tongues extending from the south and west (Plate 2.1). The maximum glacial advance of the Little Ice Age (ca. 1850 A.D.) marks the limit of the recently deglaciated terrain. This limit is delineated by a distinct lichen fringe or trimline, where the recently deglaciated area is characterized by lichen-free rocks and the area beyond the glacial limit is characterized by lichen covered rocks (Plate 2.2). The study was carried out on gneiss-granite based terrain in front of a 27 portion of the western lobe (Plate 2.3) which has retreated approximately 210 m in the last 36 years 1 3.3 Materials and methods 3.3.1 Sampling design The study area extended 300 m away from the glacier snout to just beyond the lichen trimline and covered a breadth of 120-180 m. The site was stratified into thirty 10 m belts running parallel to the glacier (Figure 3.1). Intervals of 10 m were chosen based on recent (1992-1995) retreat rates of the glacier snout (i.e. -10 m/yr), with the understanding that the belts would not all represent a terrain age based on this retreat rate. To ensure that the belts were sampled across the entire breadth, each belt was further divided into approximate 20 m x 10 m zones, resulting in a sampling grid for the study area. 3.3.2 Vegetation sampling Vegetation was sampled randomly in each zone using 50 cm x 50 cm quadrats, subdivided into 5 cm x 5 cm sections, to estimate percent cover of vascular plants, mosses and lichens. In total, 1411 quadrats were sampled over the 30 belts. Only those species observed in the quadrats were used for analysis. Percent cover was visually 28 Figure 3.1: Schematic representation of the sampling design on the Twin Glacier foreland. The study area was divided into 10 m belts, parallel to the glacier; to ensure that each belt was covered across the entire breadth (120-180 m), each belt was further divided into 10 m x 20 m zones, resulting in an approximate sampling grid. 29 estimated to the nearest 0.5% for cover values greater than 1%. Cover values less than 1% were divided into a scale of 1-3 representing cover ranges: 1 = 0-0.2%, 2 = 0.2-0.5% and 3 = 0.5-1%, respectively. Scale values were converted to percent cover using the midpoint values. The scale used was the uppermost part of a modified Domin-Krajina cover abundance scale (Scale-B) originally applied to polar-desert vascular vegetation by Levesque (1996). It ensures that minute plants, most notably the mosses in this case, are not overlooked. Due to time constraints and the fact that vegetation cover increases with increasing distance from the glacier, the number of quadrats sampled decreased near the bottom of the study site. The number of quadrats sampled in belts 1-17, 18-23 and 24-30 leading away from the glacier were 48-58 (belt 3:68 quadrats), 38-39 and 24-30, respectively. Wein and Rencz (1976) suggested that the required number of samples decreases with increasing vegetation cover and Levesque (1996) similarly suggested that there is an inverse relationship between total cover and the size of the minimal area required for accurate sampling. In addition, a river canyon severed the end of the study area leaving a smaller area to cover (Plate 2.1). Nomenclature follows that of Porsild and Cody (1980) for vascular plants and Ireland et al. (1980) for bryophytes. Voucher specimens were deposited in the herbarium at the University of British Columbia. 3.3.3 Environmental factors Within each quadrat, percent cover estimates were recorded for relict vegetation, 30 detached litter and standing dead litter; substrate cover was estimated using a 5 level classification scale (1 = 1-8%, 2 = 8-20%, 3 = 20-50%, 4 = 50-75%, 5 = >75%) for each of the following parameters: fine substrate (< 0.5 cm), pebbles (0.5-3 cm), small rocks (3-10 cm), large rocks (10-30 cm) and boulders (> 30 cm). Based on the cover of fine substrate, each quadrat was assigned a ratio of paleo soil to glacio-fluvial sediment. The paleo-soil was identified by its deep brown colour and fine texture (< 0.5 mm), while the glacio-fluvial sediment was characterized by its light colour and coarse texture (-0.5-5 mm). In each of the first 26 belts leading away from the glacier snout, soil samples (5 cm diameter x 5 cm depth; n = 10) were collected on three occasions (June 25, July 19 and August 3, 1994) throughout the growing season. Half of the samples were collected from paleo-soil and half from glacio-fluvial sediment, for the first 18 belts, at which point it became somewhat difficult to distinguish the two substrate types. The samples were collected with a trowel, as a soil auger was not practical due to the rocky nature of the study area. A l l three groups of soil samples were weighed (0.01 g) fresh and reweighed dry ( 6 0 ° C to constant weight) to determine percent soil moisture content. The second group (July 19, 1994) of soil samples was also used for pH, organic matter and nutrient analyses. For the p H and organic matter analyses, soil samples (n = 6) from each of the 26 belts were passed through a 2 mm sieve and a sample splitter. A well mixed 1:2.5 soil:water suspension (Blakemore et al, 1987) was prepared for each sample and allowed to equilibrate for 90 minutes before a p H reading (± 0.1) was recorded. Estimates of total organic matter content were determined by weight loss on ignition, in which each sample was oven dried at 1 0 5 ° C over-night and then burned at 5 5 0 ° C for 2 hours. The samples 31 were weighed (0.0001 g) before and after burning. Soil nutrient analyses were performed on the fine fraction (< 2 mm) of five sets of soil samples (n = 4), which represented increasing intervals from the glacier snout. Nitrate-nitrogen was extracted using KC1 (Page et al., 1982). Nitrate reduction was carried out on a cadmium column and colour development was accomplished using sulfanilamide and N E D ; absorbency was read on a spectrophotometer. Available phosphorous was obtained using a medium-strength Bray extract ( N H 4 F in HC1) and ascorbic acid colour development; absorbency was read on a spectrophotometer (Bray and Kurtz, 1945; McKeague, 1978). Exchangeable potassium was extracted using the ammonium acetate method (pH 7.0) and measured using atomic emission spectroscopy (McKeague, 1978). Snow depth was measured, in late spring (June 2, 1995), every 5 m along five transects (230 m) leading away from the glacier front. Rough terrain age estimates were determined from a combination of direct and indirect methods (see Chapter 2; Table 2.2b). However, distance was used as a surrogate for terrain age in the ordination analyses (described below) as sampling was performed in 10 m belt-intervals with increasing distance from the glacier snout. Distance and terrain age are highly correlated (r = 0.990). 3.3.4 Statistical analysis The following methods pertain to the overall vegetation data set, as well as to the paleo-soil and glacio-fluvial data sets which were extracted from the overall data set on 32 the basis of whether quadrats were on paleo-soil or glacio-fluvial sediment. The subsections below outline the statistical methods applied to each data set in greater detail. Two-way indicator species analysis (TWINSPAN) was used for classifying each of the three data sets (Hill, 1979). T W I N S P A N produces ordered two-way tables by constructing a classification of the samples or sites and uses this classification to sort the species into groups. The method is based on the identification of differential or indicator species. Cut levels of 0%, 0.5%, 1%, 2% and 5% were used. It was necessary to use low cut levels as the dominant species had relatively low mean cover values (rarely greater than 5%). The program C A N O C O (ter Braak, 1987), was used for all ordination analyses. A detrended correspondence analysis (DCA) was initially applied to each vegetation data set to decide whether unimodal (CA, D C A , C C A , D C C A ) or linear models (PCA, R D A ) should be used, by examining the lengths of the ordination axes. Data sets with linear distributions generally have short ordination axes (< 2 standard deviation (s.d.) units) and strongly non-linear data sets have ordination lengths greater than about 4 s.d. units (Jongman et al., 1995). In this study the first and second D C A axes for the overall vegetation data set had gradient lengths of 4.2 and 2.1 s.d., respectively (Table 3.2), confirming the choice of a unimodal model. The first two D C A gradient lengths for the paleo-soil and glacio-fluvial data sets were 2.2 and 0.6 s.d. and 2.0 and 1.0 s.d, respectively (Table 3.2). Despite the short second axis gradient lengths in both substrate data sets, unimodal rather than linear models were used because of the greater importance of the first axis and the fact that these data sets were extracted from the highly non-linear overall vegetation data set. The D C A s also verified that the chosen environmental 33 variables explained most of the species variation for each data set, as the D C A eigenvalues were just slightly higher than that for the direct gradient analyses (Table 3.2). A Monte Carlo permutation test was performed on each of the three data sets to test the significance of the first axis (ter Braak, 1987; ter Braak, 1990). Inflation factors for the environmental variables did not exceed 6.5; values above 20 are considered unacceptable. Spearman's correlation coefficients were used to test the strength of association between the environmental variables for each data set, using S Y S T A T (1992). 3.3.4.1 Succession patterns Before analysis began, the original 1411 quadrat samples were averaged by zone, reducing the number of active samples to 139. T W I N S P A N was used to classify the overall vegetation data set with the minimum group size for division and the maximum levels of division set at 14 and 4, respectively. To analyze the relationship between the overall vegetation data set and the environmental factors a detrended canonical correspondence analysis ( D C C A ) was used. In this direct gradient approach, site scores are restricted to be linear combinations of measured environmental variables (Jongman et al, 1995). The analysis was detrended because of a suspected \"arch effect\". Two samples (130 and 139) were eliminated because of extreme environmental data, leaving a total of 137 active samples for the ordination analysis. Of the 45 species observed, 9 were made passive (Appendix 2) because of their rarity (single occurrence) on the study area. Passive species are allocated species scores, but have no influence on the extraction of ordination axes. Except for two species (Silene acaulis and Luzula arctica), all passive 34 species were eliminated from the final D C C A ordination diagram (and TWINSPAN) as they were situated on the periphery of the diagram, indicating that they contributed little to the actual vegetation patterns (ter Braak, 1987). Lichens were not included in the ordination as it was difficult to separate relict lichen from presently developing lichen. A l l environmental variables were included in the ordination except for the nutrient data, as samples for this variable were not analyzed from every belt-interval as the other variables were. However, statistical tests were used to detect whether there was a difference in N , P and K between belt-intervals (i.e. terrain age). Using S Y S T A T (1992), analysis of variance ( A N O V A ) was applied to the phosphate and potassium data, and the nonparametric Kruskall-Wallis A N O V A was applied to the nitrogen data. A multivariate analysis of variance was not used as Spearman's correlation coefficients showed that there was no significant correlation between N , P and K. 3.3.4.2 Succession on paleo-soil versus glacio-fluvial sediment Vegetation patterns were compared between the two substrate types, paleo-soil and glacio-fluvial sediment. These data were analyzed only up to the 18th belt (170-180 m; approx. 30 years), as it became difficult to decipher the two types of soil beyond this point. Quadrats were averaged by the 10 m belt-intervals rather than the 20 m x 10 m zones. Both the paleo-soil and glacio-fluvial data sets were classified using a maximum of 3 levels of division and the default minimum group size of 5. A canonical correspondence analysis (CCA) was applied to both the paleo-soil (18 samples, 37 species) and glacio-35 fluvial (17 samples, 36 species) data sets, in which all rare species were down-weighted, according to C A N O C O procedures (ter Braak, 1987). Terrain age and a set of soil parameters (pH, organic matter content, moisture and fine substrate) were the environmental variables used in the C C A . Species which were made passive (Appendix 2) in the overall vegetation data set were also made passive in the substrate data sets and eliminated from the C C A diagrams (and T W I N S P A N ) for reasons explained above for the overall vegetation data set. Although not a passive species, Saxifraga tricuspidata was also removed from the ordination diagram due to its outlying position. Through S Y S T A T (1992), a rank transformation-multivariate analysis of covariance ( M A N C O V A ) was used to test whether the paleo-soil and the glacio-fluvial sediment differed with respect to a combined measure of soil parameters (pH, organic matter content, moisture and fine substrate), where terrain age was the covariable. Subsequently, a rank transformation-analysis of covariance ( A N C O V A ) was used to test for differences in the overall vegetation cover between the two substrate types. Conover and Iman (1981) described rank transformation-procedures as those which apply the usual parametric procedure to the ranked data instead of to the original data. It is considered to be a bridge between parametric and nonparametric statistics. 36 3.4 Results 3.4.1 Succession patterns 3.4.1.1 Classification and Ordination T W I N S P A N divided the 139 samples into 11 groups (Table 3.1), that were later labelled with mean terrain ages. Because the samples were recorded in 10 m belt-intervals for which terrain ages (1-44+ years) were determined (Table 2.2b), each sample was matched with an age estimate; subsequently, an average age was calculated for each T W I N S P A N group which ranged from 6-44+ years. O f 38 species, five groups were produced from T W I N S P A N (Table 3.1) at the fourth level of division. A summary of all ordinations are shown in Table 3.2. Terrain age was the variable most closely related to D C C A axis 1 (X\\ = 0.660) and explained most of the variation in the species distribution over the study area (Table 3.3; Figure 3.2). The variance explained by axis 1 was significant (p < 0.01) according to a Monte Carlo permutation test. The strong correlation between terrain age and standing dead litter, fine substrate and relict vegetation (Table 3.4) explains why the latter three environmental variables were also closely related to axis 1. Standing dead litter was positively correlated with terrain age, while fine substrate and relict vegetation were negatively related. To a lesser extent, organic matter content, soil moisture content, pH, snow depth and large rocks were also significantly correlated with terrain age. These variables were less closely related to axis 1. Due to the correlation between terrain age and most of the other environmental variables and the observation that terrain age largely explains D C C A axis 1, the first axis was considered to be a terrain age factor complex (sensu Whittaker, 1991). 37 T f t-T f T f T f , H m 00 m in V C r4 T f rl rl vo in w r-m m r« cs,^ - a do Do § fe 3 S .2 I ! rN « ^ a s s - . 3 . 3 •S-.B a B O ft, ^ S 5 a a a 3 .3 & 9- «! 3 -5 3 wi S fe a <£ £ % Q § is t ! s « a, °- -3 B \"3 2 o a s I, § fe \"a a. ~ V3 ^ 53 Ife -*u c a U C j c is e w a o O Q o; a. C g ws CM *-> s ! u o \" L 38 o so o cn os os os 00 OS r> o\\ so os os CN OS o 00 OS oo cn 00 C N 00 o t - O s I\" - rt i T f CN cn cn CN ~ CN t * cn — CN CS Cn CN CN cn CN in — T f — T f — CN — T f — cn — T f — CN ~ cn ~H cn ~< O S T f os cn O S 00 00 00 f-oo se oo m 00 T f 00 i-l r- 00 r> so t - w T f r- o SO OS SO 00 SO T f S O C N S O vt © T T O S IcN —< CN - - N I- CN -cn CN •—' in —< ~H cn —< CN — — — T f —. — • Tf — — — T f - H — • T f — —. - H m —< T f — — T f • • T f • — IS> 00 i r , T t T f O cn t~-cn so cn T f cn -> n CN n 42±1 >-i ON i - i t-fH to r l © TH 00 iH NO ?H iri iH Tf r l i—t O O ON •H NO to O 00 O r-O m O Tf o r l 1) 3 G a o o O 3 — — — CO — — • — — — CN — — • — — CO CN CO — — — • T f — CN • — — — CN CO — • — — — — — • — — CN CN • — — — CO — — • — T f H — • — — CO T f CN — • co CN CO — • CN — — — CO — — • *n T f CO • — — — • CN — • — *- — CO CN T f — • r - l — T f — — • — — CO CN T f — — — — • — • — — CN — — CN CN — • T f CO CO CO CN - CO — CO — CO • — — — CN T f — — • — — — • T f — T f — — — • CN T f — — — • CN — — — — T f — • — — — — • — — — — • — CN — — T f • — — • — — — T f — — — — • — CN T f _ • co • — • — _ n — — — - CO — CO T f — • — — • — — — — • CN — • — • — CN T f — — — CN • T f — T f CO — co — • CO — — — • — CN CO • — — — — — • — — • T f — CO CO — • — — • CN — . _ . — T f CN • — — — — • CN — • — CN T f — — • CN — • — CO CO — • — —1 — • T f T - l T f CO — —1 — \" 5 « ft, ft, -J \"3 5I> a •2 S \"3 B OH 5 § ii s .3 •11 c o s 5 -c \"a § fe a fe -S - | o, ^ oq co Q Q 0, B B fe £ cl « =3 S 5 l | •S -§ 5 « B B 6 -Q C B B ^ CQ Q 8 I N 5 \"ff . a 3 OH § B 5 ^ S » o S J B 13 ^ B I 5 -g Q & O I .1 i ' i 1 a B o 3 e B a 3 ^ B a 61 U to U Q « , ? B i .1 £ Nj c3 40 < CJ u Q ^ \"H QlO O ON \\ D 0 \\ O » > ^ O r- 3 ~ ^ M C\\ vo co vi ri « » o w-i >—i \\o O T f 0 \\ — 1 \\o C\\ o S 2 oo c\\ »n „ — — m ^ ^ cs ~ »n o o o cs d \" « M ; 4 \"1 ^ r- cs o\\ co r-ri « o \" \" 3 •a O O T3 > o o E tfl G > — 3 a u- a 3 w ^ S e „ > u « § T3 '5 .SP 2 g. W O c3 d ^ ^ o E a e 3 3 3 U M W C3 73 00 41 Table 3.3: Weighted correlations between the environmental variables and D C C A axes 1 and 2 for the overall vegetation data set. See Table 3.2 for ordination summary. Environmental variable ax 1 ax 2 Terrain age -0.942 -0.024 Boulder 0.275 -0.017 Large rocks -0.408 0.066 Small rocks -0.205 -0.128 Pebbles 0.375 0.059 Fine 0.703 -0.190 Snow depth 0.400 -0.061 Relict vegetation 0.566 -0.186 Detached litter -0.284 0.212 Standing dead -0.631 -0.021 Soil moisture content 0.373 -0.003 Organic matter content -0.456 0.213 Soil p H 0.361 -0.030 Table 3.4: Spearman's correlation coefficients for the D C C A environmental variables. The values in bold represent significant correlations. Abbreviations are as follows: A G E , terrain age; B , boulders; L G , large rocks; S M small rocks; P E B , pebbles; FINE, fine substrate; S N O W , snow depth; R E L , relict vegetation; LIT, detached litter; SD, standing dead; M O I S T , soil moisture content; O R G , organic matter content; pH, soil p H . Using the Bonferroni correction for multiple comparisons, 0.05/n, where n = number of comparisons (78), a = 0.001. A G E B A G E 1.000 LG B -0.230 1.000 SM L G 0.279 -0.041 1.000 PEB SM 0.248 -0.248 0.180 1.000 PEB -0.235 -0.257 -0.404 0.230 1.000 FINE -0.601 0.097 -0.536 -0.388 0.173 SNOW -0.392 0.277 0.105 -0.180 -0.130 REL -0.614 0.030 -0.196 -0.219 0.090 LIT 0.121 0.125 0.003 -0.240 -0.330 SD 0.847 -0.263 0.146 0.033 -0.208 MOIST -0.466 0.237 -0.024 -0.129 -0.071 ORG 0.533 -0.113 0.084 0.080 -0.102 pH -0.440 0.130 -0.005 -0.235 -0.029 FINE SNOW 1.000 REL 0.157 1.000 LIT 0.439 0.298 1.000 SD -0.002 0.079 0.022 1.000 MOIST -0.474 -0.421 -0.489 0.236 1.000 ORG 0.159 0.278 0.325 0.007 -0.380 1.000 -0.198 -0.195 -0.306 0.167 0.493 -0.129 1.000 0.165 0.270 0.251 0.117 -0.374 0.091 -0.321 42 o -9 00 CO o a co K CO i~ o s: oo • > • P £ 0 2 if ft* £ M cd-^ 3 • -00 CO .2 w o i> 2 '« C H 0) M O H a co tt O H co O cd cu C CD « -3 o .3 1 fe\"'£ So 3 o _r ~ t>oS ^ a y =^ -0 fi 3 tj P H « . § SS 5 -3 S 3 ' \" ... a, C_) CO '•C O p ed ^ c/T 3 MHZ H o o •a s 6 \"3 -\"3 las CD > 00 T-» • - i ? H o OO O OO^3 e cd i a o a > cn o I S ? \" M « oo co cd o a Pu, Co CO .H .y .a a « ,3, co El -^8 •a ^ -c o ;~ g a\"t?« ^ , ^ S « -S •2 Q ^ 5 : *.§ w . ? £/§. 3 a a G cT^ 1~\\ ^ co U U I . 1) t l s,\" .-cd bo-rt co g g « »•§ 3 — co 1 A •S Q ^ a .2 b §J c5 u y . t s u A ^ *3 s O S 2 c 5 5' 3 d 2 oo O CO Q ^ ti be - a t i l - O co . s ag ^ 1 V? co *e •u3 > co cn ? 00 O O co Q,=3 > S^ cS eyew 0 cr co 3^ 2H fag to .2 . „ c£\"cc! \"5 W cd co -^ ) -H W VJ . S 'o m = H bpA CO >-7 -O t J H | g l . l w cu a, i 3 cZ 3 rr1 -4-» C? 3 s .y •S i g c^ cd \"Q ~\" •-5 r v to CO « Q U ~ s | &, •£ Ps^ ; C H a ~? - co •~ 5^ co oo & S & ° C3 ^ cd Q . w co _ ^ « O - O ^ P H 43 Although with a much smaller proportion of variance explained (X2 = 0.107), D C C A axis 2 was best represented by organic matter content and detached litter (Table 3.3; Figure 3.2). These two variables were not significantly correlated with one another (Table 3.4). When terrain age was fitted as a covariable, to remove the effect of age on the other environmental variables, the first and second D C C A axes had eigenvalues of 0.357 and 0.113, respectively. The covariable, terrain age, explained approximately 21% of the total inertia, while the other 15 environmental variables combined explained another 21%, indicating the importance of terrain age on its own in explaining the variation in species composition. Figure 3.3 shows the distribution of the T W I N S P A N sample group centroids, (calculated from sample scores) in ordination space. Six sample groups depicted terrain ages 6-10 years and were roughly positioned in sequential order, stretching along the positive end of the terrain age factor complex axis, showing little response to the second axis (organic matter content and litter). The 20 and 24 year old sample groups were situated close to the origin on middle-aged terrain. Groups representing 36 and 42 year old terrain, located on the negative side of the terrain age axis, appeared to be associated with relatively low levels of organic matter content and litter, while the sample group on the oldest terrain (44+ years) appeared to be associated with areas higher in organic matter content and litter. Restricted to the oldest terrain were Carex nardina, Festuca brachyphylla and Grimmia spp (Table 3.1). Other species characterizing this group were Racomitrium spp, Dryas integrifolia, Cassiope tetragona and Saxifraga oppositifolia. The 36 and 42 year old groups were dominated primarily by Polytrichum-Pogonatum spp and secondarily by Salix arctica and Luzula confusa. 44 Figure 3.3: D C C A ordination showing T W I N S P A N sample group centroids labelled with their respective mean terrain age (see Table 3.1). The bars are 95% confidence intervals. Abbreviations as in Figure 3.2. 45 Figure 3.2 shows the distribution of the individual species according to the D C C A and their T W I N S P A N group membership. Emphasized in Figure 3.4 is the distribution of the T W I N S P A N species group centroids, calculated from the original species scores. The first species group was chiefly composed of moss species, making this group distinct from the other groups. Psilopilum cavifolium was by far the most dominant species in this group. Funaria-Dicranella spp and Pohlia spp were observed frequently with low cover, while the vasculare Poa arctica and Luzula arctica occurred with both low frequency and low cover in group 1. The centroid for this group was at the positive end of axis 1, emphasizing its location on young terrain (6-9 years; Table 3.1). Comprised of 11 species, group 2 was dominated by Luzula confusa and Papaver radicatum. Aside from Draba subcapitata, all Draba species were classified into this group. Also included in group 2 were Saxifraga cernua, Saxifraga nivalis and Bryum spp which were found relatively infrequently across the study area. The mean species score for this group was found near the origin of the ordination diagram, indicating that this group of species generally characterized middle-aged terrain (10-36 years; Table 3.1). Groups 3 and 4 overlapping to a large degree, were found further along the age gradient (axis 1). Group 3 was found on terrain of 36 years and to a lesser extent on 44+ year old terrain (Table 3.1). This small group was dominated by Salix arctica and Polytrichum-Pogonatum spp with Saxifraga tricuspidata and Tortula ruralis occurring in small percentages. Draba subcapitata, Didymodon spp and Conostomum tetragonum dominated the fourth group. Minuartia rubella, Melandrium apetalum, Desmatodon spp and Bartramia ithyphylla were also present in group 4. Although species in group 4 rarely achieved cover values greater than 0.4%, this group was found predominantly on terrain between 20 and 36 46 T 1 (44+y) 5 I—x-\\ -1.5 axis 1 (36-44+y) 3 11 Luzula-Papaver-Draba (group 2) —> Polytrichum-Pogonatum-Salix (groups 3 and 4) —> Racomitrium-Dryas-Cassiope- (group 5) (Figure 3.5). 3.4.1.2 Species diversity The dominating species in each of the four successional stages, mentioned above, are emphasized in the dominance-diversity curves, based on mean prominence values (% cover x (% frequency)14) for each species over the study site (Figure 3.6). The most dominant species in successional stages 1-4 were Psilopilum cavifolium, Luzula confusa, Polytrichum-Pogonatum spp, and Racomitrium spp, respectively. Figure 3.7 shows that the number of species increased rapidly in the first 20 years to 24 species and remained relatively stable for the next 20 years. A drop in the number of vascular species to 15 was observed after 40 years; however, beyond the trimline the number of vasculare peaked at 25 before decreasing to 19 species. The mosses also 48 Trimline Stage 1 - - • i ^ P— Stage 2 Stage 3 u Stage 4 —H 1 1 ^1 ^ 1 l \\ 1 1 * x • 10 20 30 40 Terrain age (years) Psilopilum cavifolium • Pohlia spp • Funaria-Dicranella spp Luzula confusa Papaver radicatum Draba spp • Polytrichum-Pogonatum spp • Sa/o: arctica Racomitrium spp Dryas integrifolia Cassiope tetragona Figure 3.5: Percent cover of the dominant species in each of the four successional stages in relation to terrain age. Terrain ages are unknown beyond the lichen trimline (> 44 years). 49 10 1 0.1 0.01 0.001 0.0001 Psi cav Poh spp Fun-Die Poa arc Stage 1 Luz arc 1 > o fl 44 years). 51 exhibited this pattern, but to a lesser degree. 3.4.1.3 Pre-LIA vegetation community The cover of crustose, fruticose and foliose lichens were plotted to show the abrupt difference in cover between the lichen-free and lichen-populated zones (Figure 3.8). The convergence of these two zones delineates the lichen trimline, thought to represent the maximum glacial extent of the L I A (ca. 1850 A.D.) . Also illustrated in Figure 3.8 is the decrease in relict pre-LIA lichen cover from the glacier snout to the trimline. The decrease is a reflection of wind and water erosion with increased time of exposure from beneath the glacial ice. The same declining patterns were also evident for relict mosses and vascular species as a whole. Cassiope tetragona followed by Dryas integrifolia were the most dominant relict vasculare within the first few years of glacial retreat. Mosses as a whole were also noteworthy in the relict community. The relict community appeared to be similar to that of the presently developing community. A list of the relict vasculars and mosses found at the ice edge are included in Appendix 2. 3.4.2 Succession on paleo-soil versus glacio-fluvial sediment T W I N S P A N divided both the paleo-soil and glacio-fluvial data sets into 6 species groups (Tables 3.5 and 3.6). The groups are represented by symbols and group centroids (calculated from species scores) in the paleo-soil and glacio-fluvial C C A ordination diagrams (Figures 3.9 and 3.10, respectively). Terrain age and fine substrate 52 Trimline 30 25 20 > o U 10 5 0 foliose It 1 1 1 1 0 10 20 30 40 Terrain age (years) Figure 3.8: Percent cover for crustose, fruticose and foliose lichens in relation to terrain age. Terrain ages are unknown beyond the lichen trimline (> 44 years). 53 Table 3.5: T W I N S P A N classification for vegetation on paleo-soil showing species and sample groups. A dendrogram is shown with eigenvalues for the species classification. Sample groups are shown with their corresponding mean (±SE) terrain age. Cut levels are represented by a scale of 1-5: 1 = 0-0.4%, 2 = 0.5-0.9%, 3 = 1-1.9%, 4 = 2-4.9%, 5 = > 5%. P A L E O - S O I L Sample groups 1 1 1 1 1 1 1 1 1 1 2 3 4 5 9 1 6 7 8 0 2 4 3 5 6 7 8 Species group Species 2±ly 5±ly 10±ly 7± ly 15±2y 20±3y 26±ly 1 Psilopilum cavifolium . 1 4 4 4 4 4 5 5 4 3 2 1 1 1 1 1 1 1 Funaria-Dicranella spp . 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 Saxifraga tricuspidata 1 1 0.445 —2 Draba lactea 1 1 1 1 1 1 1 1 1 . 1 1 Draba spp 1 1 1 1 . 1 1 1 . 1 . 0.157 Saxifraga nivalis T 1 1 . —3 Luzula confusa 1 . 1 1 1 3 4 2 3 3 4 4 4 4 4 4 4 4 Pohlia spp . . 1 1 1 1 1 1 1 1 1 1 1 1 1 1 3 2 Papaver radicatum . . 1 1 1 1 1 1 1 1 1 1 1 2 2 0.629 Racomitrium sppT 1 . 1 Poa arctica 1 1 1 1 1 1 . 1 1 . 1 Bryum spp 1 1 1 1 1 1 1 1 1 1 1 1 1 Draba nivalis 1 1 1 1 1 1 1 1 1 1 1 1 1 1 | 4 Salix arctica . . 1 1 1 1 1 1 1 2 1 1 1 1 Polytrichum-Pogonatum spp . . 1 1 1 1 1 1 1 1 1 1 1 1 2 4 2 0.320 r-5 Cerastium alpinum T 1 1 1 . . 1 Conostomum tetragonum 1 1 1 1 1 1 1 1 1 0.248 Saxifraga cernua 1 1 1 1 1 . 1 . 1 —6 Melandrium apetalum 1 1 1 Dryas integrifoliaT . 1 Cassiope tetragonaT . . 1 Cardamine bellidifolia 1 1 . 1 . Encalyptra sppT . 1 1 . . Bartramia ithyphyUa . 1 1 1 1 Minuartia rubella . 1 1 1 1 Draba subcapitata . 1 . 1 1 1 1 Didymodon spp . 1 1 1 1 1 1 Saxifraga oppositifolia 1 1 1 1 1 fRare species down-weighted for CCA, according to CANOCO procedures (ter Braak, 1987). 54 Table 3.6: T W I N S P A N classification for vegetation on glacio-fluvial sediment showing species and sample groups. A dendrogram is shown with eigenvalues for the species class-ification. Sample groups are shown with their corresponding mean (±SE) terrain age. Cut levels are represented by a scale of 1-5: 1 = 0-0.4%, 2 = 0.5-0.9%, 3 = 1-1.9%, 4 = 2-4.9%, 5 = >5%. G L A C I O - F L U V I A L Species group Species 3 5 6 7 1 2 4 0 Sample groups 6 ± l y 5±3y _8_ 8y 1 1 1 9 1 2 6 17±3y 1 1 1 1 3 4 5 7 23±2y 0.586 0.253 1—3 0.738 0.322 0.179 ^ 6 Bryum spp Papaver radicatum Draba lactea Draba nivalis Salix arctica Luzula confusa Saxifraga cernua Ceratodon purpureus 1 Conostomum tetragonum Cardamine bellidifolia 1 Saxifraga oppositifolia Minuartia rubella Bartramia ithyphylla Draba spp Didymodon spp Cassiope tetragonaT Racomitrium sppT Melandrium apetalum T Cerastium alpinum 1 Draba subcapitata Encalyptra spp' 4 3 1 . . 1 1 1 Psilopilum cavifolium 4 4 Epilobium latifloium 1 Draba oblongata 1 Poa arctica . . Funaria-Dicranella spp 1 1 1 1 Pohlia spp 1 1 1 1 Saxifraga tricuspidata T . . . . Polytrichum-Pogonatum spp 1 1 1 1 . 1 1 1 . . 1 1 . 1 . 1 1 1 . 1 . . 1 1 1 1 2 1 3 3 1 . 1 1 1 1 1 1 1 1 . 1 . . 1 1 . 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 . 1 1 1 1 . . 1 1 2 3 3 2 1 1 1 1 1 . . 1 1 . 1 1 1 1 . 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 . 1 1 1 1 1 2 1 1 1 3 2 3 3 1 1 1 1 1 . 1 1 1 1 . . 1 1 1 . 1 1 1 1 1 1 1 . . . 1 . . 1 . I . 1 . 1 . . 1 I I . 1 I I I . fRare species down-weighted for CCA, according to CANOCO procedures (ter Braak, 1987). 55 PALEO-SOIL A Enc spp A Dry int CM CO • 1—< X rr 2 • Rac spp _l_ Dra sub A • Poa arc O R G Dra lac .El A D id spp Bar ith A . Dra niv axis l h Cas tet — A — -1.5 Car bel A G E * * -^ Luz conL? Sax nivN M i n rub Pap ra ft Con tet A P o h S PP Sal arc Bry spp [41 * n m P o l - P o g s P P B A L J x o p p S a x c e r Fud Cer alp A T W I N S P A N species groups O 1 • 2 3 4 5 6 - L - l • • A A Psi cav F I N E Figure 3.9: C C A ordination for the paleo-soil vegetation data set with environmental variables represented by arrows. Symbols represent the T W I N S P A N species groups (see Table 3.5) and boxed numbers represent species group centroids. Eigenvalues for the first and second C C A axes are 0.581 and 0.037, respectively. Abbreviations as in Figure 3.2. 56 GLACIO-FLUVIAL c T 2 • Enc spp • sal arc A Cer alp • Mel ape HP Poa arc O T Sax cer Did spp A axis lh-AT „ ^ Min rub Castet ^ Drani^ _1.5 Carbel A G E < Rac sppA Bry sppn Luz con TWINSPAN species group O 1 • 2 3 4 5 6 • • A A O R G T T M O I S T F I N E , „ , „ rASax opp J* 0 * 1 SPP Pol-Pog spp » •hjj Fun-Die sprl A Cer pur Con tet A Dra spp A ^ D r a s u b • Pap rad • Dra lac A Bar ith Dra oblO t> Epi lat O Psi cav -L-l Figure 3.10: C C A ordination for the glacio-fluvial vegetation data set with environmental variables represented by arrows. Symbols denote the T W I N S P A N species groups (see Table 3.6) and boxed numbers represent species group centroids. Eigenvalues for the first and second C C A axes are 0.587 and 0.098, respectively. Abbreviations as in Figure 3.2. 57 respectively represent the first C C A axis in both the paleo-soil (hi = 0.581) and glacio-fluvial (A,i = 0.587) ordinations (Table 3.7; Figures 3.9 and 3.10). Table 3.8 shows that age is highly negatively correlated to fine substrate for both the paleo-soil and glacio-fluvial data sets. The strong correlation between the two variables reflects the fact that with increasing terrain age, vegetation cover increases while bare ground cover decreases. The cover of fine substrate depended on visible bare ground cover during sampling. Organic matter content was most closely associated with C C A axis 2 for both the paleo-soil (h2 = 0.037) and glacio-fluvial (h2 = 0.098) data sets, but because the second axes explained so little of the variation, this relationship bears little meaning for either data set. When terrain age was fitted as a covariable, the first axis eigenvalues were lowered dramatically to 0.038 from 0.581 (paleo-soil) and to 0.101 from 0.587 (glacio-fluvial). This emphasizes that age explained a large portion of the total inertia for both the paleo-soil and glacio-fluvial data sets (62% and 56%, respectively), while the other four environmental variables combined explained only a small portion (9% and 16%, respectively). The paleo-soil and glacio-fluvial data sets were similar in that the T W I N S P A N species group centroids for both data sets were positioned roughly in sequential order along the terrain age axis, leading from group 1 on very young terrain to group 6 on old terrain. In addition, there was a large gap along the age gradient between group 1 and groups 2-6 for both data sets. The species composition of the two data sets increased in similarity with increasing terrain age or T W I N S P A N group number. Group 1 of the paleo-soil data set had only one (Psilopilum cavifolium) of three species in common with the four species in group 1 of the glacio-fluvial data set, and group 6 from both data sets 58 Table 3.7: Weighted correlations between the environmental variables and the first two C C A ordination axes for the paleo-soil and glacio-fluvial data sets. See Table 3.2 for ordination summary. Paleo-soil Glacio-fluvial Environmental variable ax 1 ax 2 ax 1 ax 2 Terrain age -0.969 -0.019 -0.965 -0.055 Fine substrate 0.907 0.124 0.763 0.036 Soil moisture content 0.478 -0.003 0.503 0.183 Organic matter content -0.424 0.476 0.101 0.671 Soil p H 0.319 -0.193 0.424 0.153 Table 3.8: Spearman's rank correlation coefficients for the a) paleo-soil and b) glacio-fluvial data sets. Bold values indicate significant correlations. Abbreviations are as follows: A G E , terrain age; FINE, fine substrate; MOIST, soil moisture content; O R G , organic matter content; pH, soil pH. Using the Bonferroni correction for multiple comparisons, 0.05/n, where n = number of comparisons (10), a = 0.005. a) Paleo-soil AGE FINE AGE 1.000 MOIST FINE -0.674 1.000 ORG MOIST -0.508 0.420 1.000 pH ORG 0.534 -0.051 -0.202 1.000 pH -0.572 0.165 -0.056 -0.401 1.000 Glacio-fluvial b) AGE FINE MOIST ORG pH AGE FINE MOIST 1.000 0.261 -0.016 ORG pH 1.000 1.000 -0.781 1.000 -0.587 0.171 -0.276 0.335 -0.534 0.545 1.000 -0.166 59 had nine species in common, which translates into 90% (paleo-soil) and 75% (glacio-fluvial) of the species in group 6. Note that where there were differences in species composition, between the corresponding groups of the two data sets, that the differing species were usually classified in a neighbouring T W I N S P A N group. There were few differences between the two data sets with respect to the first division in the T W I N S P A N species dendrogram (Tables 3.5 and 3.6). Species richness patterns were also used to compare the successional sequences displayed by the two data sets (Figure 3.11). Both substrate types showed a steady increase to 20 species over the 28 year period, with the last 3 data points (24, 26 and 28 years) being identical to one another. However, the paleo-soil generally had a slightly higher number of species than the glacio-fluvial data set, over much of the sampling area. 3.4.3 Soil attributes Nitrogen increased rapidly from 1 ppm on 3 year old terrain to approximately 11 ppm on terrain aged 22 years, but declined to about 2 ppm on terrain aged 35 years and remained at this low position on 44+ year old terrain (Figure 3.12). Phosphorous showed a slight decline from about 4 ppm (3-22 years) to 1 ppm (35-44 years) and potassium showed no trend with terrain age (Figure 3.12). A N O V A (P, K) and a Kruskall-Wallis test (N) revealed that there were no significant differences (a = 0.05) in N , P or K between the different terrain ages (Appendix 3). The M A N C O V A results revealed that the combined measure of soil parameters (pH, organic matter content, soil moisture and fines) for the paleo-soil was significantly different (p = 0.000; Appendix 3) from that of the glacio-fluvial sediment. Soil p H was 60 evergreen dwarf shrub-moss tundra (group 5). The pioneer group was distinct from the remaining three stages, in terms of its position on young terrain and its almost pure composition of mosses. However, in truly marginal environments the difference between the pioneer and remaining stages is not obvious (Svoboda and Henry, 1987; Matthews, 1992). A survey completed on a high arctic glacier foreland, in which the pioneer stage was not immediately distinguishable from the next stage, reinforces this trend (Appendix 4). The observation that the first colonizers were mosses on the Twin Glacier foreland complements that of other foreland investigations in the vicinity of Alexandra Fjord (Appendices 4-6). In addition, many others have recognized this trend on recently deglaciated terrain in mid- and high-alpine zones of northern Sweden (Stork, 1963), in Iceland (Persson, 1964), in high alpine New Zealand (Archer, 1973), on Anvers Island, Antarctica (Smith, 1982) and at Svalbard (Tishkov, 1986). A n exception to the pattern is the Tear Drop glacier foreland, Sverdrup Pass, approximately 80 km west of Alexandra Fjord, where mosses and vascular plants appeared to colonize simultaneously (Appendix 7). This may reflect the favourable moisture conditions, provided by the many melt-water streams leading away from the glacial terminus on this foreland. 66 The relatively strong floristic difference between samples within and beyond the lichen trimline supports the contention that the trimline marked the maximum glacial advance during the LIA. The extended time period for vegetation and soil development beyond the trimline may likely explain the abrupt floristic difference observed between the 42 year old T W I N S P A N group sampled within the lichen-free zone and the 44+ year old T W I N S P A N group, sampled beyond the trimline. Although, the age of one Salix arctica specimen rooted 15 m beyond the trimline inside the lichen populated zone indicated a minimum terrain age estimate of 44 years, it is nevertheless believed that the terrain beyond the trimline is much older than 44 years due to its extensive lichen population compared to the terrain directly inside the trimline which was estimated to be 43 years. The observed species richness patterns on the Twin Glacier foreland appear to be similar to that of many other glacier forelands, in that a peak occurred within the first 20-30 years of deglaciation (e.g. Matthews, 1978; Elven and Ryvarden, 1975). This peak likely reflects the amount of open space for colonization and the overlap between the pioneer and intermediate stages of succession (Matthews, 1992). The apparent increase in species richness beyond the trimline is attributed to the occurrence of species (i.e. Festuca brachyphylla, Carex nardina and Grimmia spp) found only on this older terrain rather than a resurgence of the species that appeared to die off after 40 years. Although a slight decline in the number of species was noted in the last few collection points on the Twin Glacier foreland, further sampling along the age gradient is needed to confirm if a decline actually occurs. A decline in species richness is often evident on old terrain (> 100 years) and is usually attributed to increased competition, due to increased vegetation cover over 67 time (e.g. Elven and Ryvarden, 1975). Whether or not competition causes species richness to decline in high arctic successions is unknown. The observation that the pre-LIA community is similar to that of the species in the mature living community located beyond the trimline suggests that the climate prior to the L I A was comparable to the present day climate (see also Bergsma et al, 1984). Partial evidence for this conclusion was provided by Havstrom et al. (1995) who estimated that the average July temperature at Alexandra Fjord just prior (25 years) to the L I A glaciation was only 0 . 7 ° C cooler than present day July temperatures. Although this evidence only applies to a short period before the plants were entombed by the glacier, it does give some indication that the climates were similar. Sninzow (1962) in Greenland and Smith (1982) in the Antarctic, also found that relict communities released from glacial ice were analogous to existing plant communities located beyond the maximum extent of the L I A . 3.5.2 Application to succession models The successional patterns observed on the Twin Glacier foreland provided evidence that species replacement can occur in a high arctic environment in that each of the four successional stages were relatively distinct from one another. The first stage, which was dominated by the moss Psilopilum cavifolium, was clearly separated from the second stage by the shift in dominance to the graminoid Luzula confusa on the 20 year terrain. Although the dominance of Luzula confusa in stage 2 overlapped somewhat into stage 3, the abrupt increase in the deciduous shrub Salix arctica, and the mosses Polytrichum-Pogonatum spp on the 36 year old terrain defined the third stage from the 68 second. Salix arctica and Polytrichum-Pogonatum spp remained with relatively high abundance into stage 4; however, Dryas integrifolia, Cassiope tetragona and Racomitrium spp showed clear signs of dominance in this stage on the 44+ year terrain. In addition, Carex nardina, Festuca brachyphylla and Grimmia spp were restricted to stage 4 indicating that this stage was in fact separate from stage 3. In addition to the Twin Glacier foreland, a study of plant communities along a coastal toposequence at Truelove lowland, Devon Island (Bliss and Gold, 1994) and a short survey on the Tear Drop glacier foreland, Sverdrup Pass (Appendix 7) showed that directional change with species replacement occurs in the High Arctic. Like the Alexandra Fjord lowland, Truelove lowland and Sverdrup Pass are considered to be arctic oases. In addition, on the west coast Spitsbergen (Svalbard), where the mature community resembles that of the Alexandra Fjord lowland, findings by Tishkov (1986) indicated that succession fit the directional-replacement model on marine beach ridges, recently deglaciated moraines and emergent salt marshes. Because of the relatively favourable conditions of polar oases, it is perhaps not surprizing that directional-replacement successions occur as environmental resistance is low enough to allow the recruitment and expansion of the invading species. Aside from the species themselves, the main difference between high arctic oasis successions and temperate successions appears to be in the rate of succession (Svoboda and Henry, 1987). Succession in high arctic landscapes takes up to 1500 years to reach a dynamic equilibrium (Tishkov, 1986), whereas in north-temperate regions, such as at Glacier Bay, Alaska, all phases of succession can be reached in as little as a century (Lawrence, 1958). 69 Despite the evidence for directional-replacement successions in polar oases, there is some evidence for directional-nonreplacement (Appendices 4 and 6) and nondirectional-nonreplacement (Appendix 5) successions. Whether the severity of the environmental conditions and/or the lack of seed sources are responsible for these successional patterns is unknown. Detailed investigations of the seed dynamics, climate, geology and lithology of these marginal high arctic sites are required in order to understand the successional patterns. Although a classical directional-replacement succession was observed on the Twin Glacier foreland, this does not necessarily suggest that competition was a major successional driving force. There is little evidence to suggest whether competition exists in high arctic environments, except perhaps in wetland communities such as sedge meadows (G. Henry, pers. comm., 1997). In truly marginal environments where plants are extremely sparse, it is most probable that plant competition does not exist, but in tundra environments such as that found on the Alexandra Fjord lowland where vegetation cover reaches 100%, competition is conceivable. However, as a process determining successional change in the High Arctic, competition is considered to be of relatively low importance (Svoboda and Henry, 1987; Walker and Chapin, 1987; Matthews, 1992; Chapin et al., 1994). The ability to tolerate stress on the other hand is thought to be important (Grime, 1977), whereby the sequence of species in a succession is largely determined by the life history characteristics of the plants (a.k.a. the 'tolerance' model of Connell and Slatyer (1977)). Egler (1954), Noble and Slatyer (1980), Walker et al (1986), and Svoboda and Henry (1987) have all suggested that life history characteristics alone are capable of 70 explaining patterns of succession without reference to competitive interactions. In addition, Walker and Chapin (1987), Bliss and Peterson (1992) and Chapin et al. (1994) suggested that in tundra environments, where conditions are cold and soils are nutrient-poor, succession may be determined moreso by facilitative interactions rather than competitive interactions. To determine the relative importance of facilitation, competition and life history traits in succession change, stringent measurements and experimental tests of specific hypotheses are necessary. Much of the experimental work on the processes involved in primary succession to date have been conducted in Alaska, at sites where primary succession has already been described thoroughly, in the past. Walker et al. (1986) examined the importance of life history processes on an Alaskan floodplain through investigations of seed rain, seed sowing experiments and density and aging measurements. At the same site, the relative importance of facilitation and inhibition effects on seedling growth for different successional stages was examined by Walker and Chapin (1986). Chapin et al. (1994) examined the facilitative, competitive and life history processes of succession following deglaciation at Glacier Bay, Alaska. Although the seed dynamics on the Twin Glacier foreland were examined through seed bank, seed rain and snow core germination trials (see Chapter 4), other possible processes governing succession were not. The theoretical model presented by Walker and Chapin (1987) suggests that facilitation, longevity, stochastic events, mycorrhizae and seed arrival are the most important factors determining primary succession change in a low resource environment. On an observational/conceptual basis, these fiVe potential processes are discussed below in the context of the Twin Glacier foreland. 71 Bryophytic facilitation was most likely one of the processes initiating the succession on the Twin Glacier foreland. Sohlberg and Bliss (1984) demonstrated that vascular plant species are generally facilitated by bryophyte and bryophyte-fruticose lichen mats in high arctic plant communities. Bryophytes act to increase soil fertility by increasing the water holding capacity of the soil, decreasing p H and increasing organic matter content, therefore making establishment more successful for the succeeding stage. Because soil moisture is important in maintaining high N 2 fixation rates of cyanobacteria (Billington and Alexander, 1978; Wojciechowski and Heimbrook, 1984), these micro-organisms are often associated with hydrophyllic bryophytes; hence the associated cyanobacteria greatly enhances the bryophytic facilitation process. Facilitation in the later stages of succession is also conceivable as root nodules on the late successional species, Dryas integrifolia, at Alexandra Fjord have been observed to fix N 2 (Henry and Svoboda, unpubl.). In terms of vascular plants, the final successional stage on the Twin Glacier foreland was dominated by slow growing, long-lived plants: Dryas integrifolia, an evergreen cushion plant and Cassiope tetragona, an ericaceous dwarf-shrub. According to Walker and Chapin's (1987) model, the dominance of these long-lived plants in the later stages of succession suggests that longevity is an important process in this succession. Mycorrhizae are common in tundra plants (Miller, 1982). O f the species observed on the Twin Glacier foreland, Salix arctica, Dryas integrifolia, Saxifraga oppositifolia, Cassiope tetragona and possibly Carex nardina and Silene acaulis were noted to have mycorrhizal associations at Alexandra Fjord (Kohn and Stasovski, 1990), while these 72 associations were not observed for Papaver radicatum, Draba lactea, Cerastium alpinum, Poa arctica, Luzula confusa, Saxifraga tricuspidata and Festuca brachyphylla. The mycorrhizae only appear to be associated with species typical of the last stage of succession, indicating that mycorrhizae may be more important in the later stages of succession, contrary to the proposition made by Walker and Chapin (1987) that mycorrhizae are highly important over all successional stages in low resource environments. The plant samples used by Kohn and Stasovski (1990) for mycorrhizal analysis were collected on the Alexandra Fjord lowland, but not directly on the Twin Glacier foreland. Evidence of stochastic variation as a process determining successional change is more difficult to identify, and requires long-term monitoring. Disturbance due to frost churning affected successional patterns somewhat on the Storbreen glacier foreland, Norway (Whittaker, 1989). Moreover, Bliss (1971) argues that needle ice activity is one of the largest environmental controls of seedling establishment and survival in tundra environments. However, neither frost action nor needle-ice, although not measured specifically, did not seem to influence vegetation distribution on the Twin Glacier foreland. Like frost action, soil drought is considered to be a major hazard and therefore an important limitation for seedling establishment in arctic regions (Bliss, 1971; Bell and Bliss, 1980). To avoid drought most germination in tundra regions occurs shortly after spring snow-melt or summer rains when the soil is near saturation (Billings and Mooney, 1968; Bell and Bliss, 1980; Oberbauer and Miller, 1982). Variation in moisture regime due to differences in snow drift patterns may influence germination and seedling establishment patterns on the Twin Glacier foreland. 73 The correlation between the above-ground vegetation cover and seed bank densities for the pioneering forbs, Luzula confusa and Papaver radicatum, indicated that colonization was largely constrained by the availability of seed reaching a particular area (see Chapter 4) and that seed arrival is important during the initial stages of succession. 3.5.3 Soil nutrients Given that the number of soil samples analyzed for N P K was small (n = 20) considering the spatial extent of the study site and that there were no significant differences in N , P or K over the study area, the patterns observed should be considered with caution. Nonetheless, explanations are offered below for the apparent trends. The possible increase in N to its peak on 22 year old terrain may be explained in terms of facilitation by mosses and associated N 2 fixing cyanobacteria in the earlier stages of succession. The apparent decrease in N on older terrain (35 and 44+ years) was somewhat unusual as N often levels off or decreases only slightly at some maximum value on older terrain (Matthews, 1992). Rather than a decline in the release of N from decomposition processes, the observed decrease in N may be attributed to the decreasing trend in soil moisture content over terrain age, as N 2 fixation rates and soil moisture content are positively related (Billington and Alexander, 1978; Wojciechowski and Heimbrook, 1984; Henry and Svoboda, 1986). It is thought that the release of N from decomposition processes may be similar over all successional stages because organic matter content was observed to be relatively constant over the study site, presumably due to the unusual presence of organic paleo-soil in the early stages of succession. 74 The general weathering of the mineral soil over time may explain the apparent decrease in phosphorous over time on the Twin Glacier foreland. Syers and Walker (1969) suggested that larger amounts of weatherable P may exist in primary mineral soils versus that of more weathered soils. Phosphorous uptake by the increasingly frequent late successional species may also explain why P levels decreased. Bormann and Sidle (1990) observed a decrease in P at Glacier Bay and provided evidence that the decrease may be due to uptake by spruce. The stable potassium levels on the Twin Glacier foreland study area were not unexpected as potassium is not a limiting factor to plant growth in natural systems and is easily leached from dead tissues and living leaves (Jeffrey, 1987). 3.5.4 Succession on paleo-soil versus glacio-fluvial sediment In general, the paleo-soil and glacio-fluvial successional sequences themselves were similar; both followed the trends in the overall successional sequence discussed above (section 3.5.1). However, the greater vegetation cover growing on the paleo-soil and the seemingly higher number of species observed on the paleo-soil suggests that the paleo-soil positively influences establishment and possibly the rate of growth, and therefore, the overall rate of succession especially in its early stages. In addition, Psilopilum cavifolium, with its much higher cover values on the paleo-soil versus the glacio-fluvial sediment, likely enhances the productiveness of bryophytic facilitation on the paleo-soil. In a study of several successional seres in Czechoslovakia, Prach (1993) observed that the rate of succession was greater in successional sequences containing higher soil 75 fertility. Although the degree of soil fertility was based on water and total nitrogen, the analogy to the paleo-soil patches can be appreciated. In addition, Chambers et al. (1990) concluded in an alpine study that higher germination, growth and survival were found on dark-coloured soils, due to their more favourable temperature and nutrient regimes. In a succession study on young moraines at Austre Okstinbreen, northern Norway, Worsley and Ward (1974) reported \"very small patches of humic-looking material\" and put forth that this material likely affords favourable sites for initial plant colonization. In terms of vascular plants, the notion that the increased cover could have been due to a higher germinable seed bank in the paleo-soil over the glacio-fluvial sediment was ruled out, as there was no significant difference in the number of seeds/m2 between the two substrate types (see Chapter 4). 3.6 Summary A directional-replacement succession was evident on the Twin Glacier foreland which proceeded through the following stages: moss —> graminoid-forb —> deciduous shrub-moss —> evergreen dwarf shrub-moss. Terrain age was by far the most important factor in explaining the successional sequence. There is little difference in successional sequence between the paleo-soil and glacio-fluvial data sets. Based on the results that the total vegetation cover was significantly higher and that species richness was slightly higher on the paleo-soil over that of the glacio-fluvial sediment, suggests that the paleo-soil may positively influence the overall rate of succession. 76 Chapter 4 Seed Dynamics 4.1 Introduction Cold, short growing seasons limit all phases of sexual reproduction in arctic and alpine environments (Bliss, 1962; Bliss, 1971; Billings and Mooney, 1968; Savile, 1972; Bell, 1975; Grime, 1977; Bell and Bliss, 1980), despite comparable growth rates of arctic and alpine species to that of temperate species (Chapin, 1987). It was previously generalized that sexual reproduction is relatively unimportant in tundra environments compared to vegetative reproduction (Billings and Mooney, 1968; Bliss 1971; Savile, 1972; Grime, 1979; Bell and Bliss, 1980; Archibold, 1984). Although this opinion holds true relative to temperate environments, sexual reproduction within the tundra biome itself is important (McGraw, 1980; Chambers, 1995; Murray, 1987; Diemer and Prock, 1993; Murray, 1995), particularly in disturbed habitats (McGraw and Vavreck, 1989; Freedman etal, 1982; Chambers etal, 1990; Chambers, 1993; Chambers, 1995). Disturbed habitats tend to provide conditions that are relatively favourable for germination and seedling establishment of early successional species, presumably due to increased light levels, higher soil temperatures (McGraw and Vavreck, 1989; Chambers, 1995) and reduced competition (McGraw and Vavreck, 1989). Seed dispersal by wind is thought to be more important than dispersal by animals or birds in tundra environments (Gartner, 1983). Like more favourable regions, the efficacy of seed dispersal by wind in tundra regions depends on a number of mechanisms 77 to enhance dispersability. The most common mechanisms are small light seeds (e.g. Pyrola, some monocotyledons) and plumed seeds (e.g. Salix, Dryas). The final position of wind dispersed seeds determines their fate in the seed rain and seed bank. Life history traits (e.g. seed dynamics, growth rates and longevity) alone are often sufficient enough to explain successional change (Egler, 1954; Noble and Slatyer, 1980; Walker et al, 1986; Svoboda and Henry, 1987). Hence, a comprehension of the seed dynamics (i.e. seed production, seed rain, seed bank, dormancy, germination and seedling establishment) during succession is an integral part of understanding the processes causing successional change. Information regarding the seed bank and seed rain of a site provides an approximate record of the past and present above-ground vegetation community. Early successional or ruderal species are usually abundant in the soil seed bank, whereas late successional dominants are usually under-represented (Harper, 1977; Thompson and Grime, 1979; Freedman et al, 1982; Chambers, 1993; Chambers, 1995). Relatively few seed rain studies have been conducted in the context of succession. Summer seed rain has been found to increase with terrain age (Ryvarden, 1971; Stocklin and Baumler, 1996) and winter seed rain (seeds from snow-cores or -scrapes) appears to be relatively unimportant compared to summer seed rain (Ryvarden, 1971). The presence of pre-Little Ice Age paleo-soil (see section 3.1) on the Twin Glacier foreland, Alexandra Fjord, presents a substrate that is not usually found on temperate and alpine glacier forelands. The paleo-soil is organically rich and fine textured (< 0.5 mm), while the glacio-fluvial sediment (-0.5-5 mm), characteristic of all temperate and alpine glacier forelands, is relatively sterile and coarse textured. Soil particle size affects the ability of soil to trap diaspores (Harper, 1977; Chambers etal, 1991). Larger soil particle 78 sizes, tend to increase the number of diaspores trapped in the soil (Chambers etal., 1991), which suggests that the glacio-fluvial sediment would have a higher seed bank than that of the paleo-soil. Although larger particle size also enhances the downward movement of diaspores through the soil column, the difference in particle size between the glacio-fluvial sediment and the paleo-soil is not great enough to produce a difference in burial depth, according to results reported by Chambers et al. (1991). In addition, the potentially higher nutrient pool in the paleo-soil was ruled out as a factor in increasing the number of seeds that would germinate, as the nutrients provided in the soil are generally not needed for germination (Fenner, 1985). The objectives of this component of the research on the Twin Glacier foreland were: 1) to determine the relationship between the above-ground vegetation cover and the germinable seed bank and seed rain densities, and 2) to determine if there is a difference in the germinable seed bank between the paleo-soil and the glacio-fluvial sediment. 4.2 Study area Research was conducted on the Twin Glacier foreland, Alexandra Fjord, Ellesmere Island in the Canadian High Arctic ( 7 8 ° 5 3 ' N ; 7 5 ° 5 5 ' W ) (Figure 1.1; Plate 2.1). The 8 km 2 gently sloping lowland adjacent to Alexandra Fjord is bounded by steep cliffs to the east and west, by glacial tongues to the south and by the Arctic Ocean to the north. These features, in part, create warm conditions and a large water supply giving rise to a polar oasis, where vegetation diversity, cover, and plant production are much greater than the surrounding depauperate polar deserts and semi-deserts (Freedman et al., 1994). Polar 79 oases are rare, as they cover only 1-2% of the Queen Elizabeth Islands (Babb and Bliss, 1974). The Twin Glacier, located at the southern end of the Alexandra Fjord lowland, is an outlet of the large Agassiz ice cap and is characterized by two lobes extending from the south and west (Plate 2.1). The maximum glacial advance of the Little Ice Age (ca. 1850 A.D.) marks the limit of the recently deglaciated terrain. This limit is delineated by a distinct lichen fringe or trimline, where the recently deglaciated area is characterized by lichen-free rocks and the area beyond the glacial limit is characterized by lichen covered rocks (Plate 2.2). The study was carried out on gneiss-granite based terrain in front of a portion of the western lobe (Plate 2.3) which has retreated approximately 210 m in the last 36 years. 4 . 3 Materials and methods 4.3 .1 Seed bank The sampling area extended 260 m away from the glacier snout to just beyond the trimline and covered a breadth of 100-160 m. The area was stratified into 10 m belts running parallel to the glacier, yielding 26 sampling belts (cf. Figure 3.1). To ensure that the belts were sampled across the entire breadth, each belt was further divided into 20 m x 10 m zones, resulting in a sampling grid. Intervals of 10 m were chosen based on recent (1992-1995) retreat rates of the glacier snout (i.e. -10 m/yr), with the understanding that not all the belts would represent a terrain age based on this retreat rate. Table 2.2b shows the Twin Glacier foreland terrain age scheme up to 1995; however, because the seed data 80 were based on samples collected in 1994, the terrain ages used throughout this chapter have been adjusted by subtracting one year from the terrain ages outlined in Table 2.2b. Seed bank samples (10 cm x 10 cm x 2 cm deep) were collected with a trowel from each belt (2 samples/zone) between July 5 and 10, 1994. Ten samples were collected from each belt for the first 220 m, 8 samples from the following two belts, and 6 samples from the final two belts, totalling 248 samples and a total coverage of 2.48 m 2 . For each of the first 18 belts, half of the samples were collected from paleo-soil and half from glacio-fluvial sediment. The samples were packed in paper bags, kept frozen in the field until they were shipped back to the lab at the University of British Columbia (August 17, 1995) where they were air-dried and stored at room temperature ( ~ 2 0 ° C ) until germination trials began (February 16, 1994). Each soil sample was passed through a 2 mm sieve and from the 2 mm portion, a 60 m L (if available) subsample was weighed (0.01 g). Subsamples were placed into 100 mm plastic petri dishes (2 dishes/subsample) and thereafter watered and monitored for seedling emergence every 2 days for 6 weeks in a greenhouse at the U B C Botanical Gardens Nursery. Seedlings that were unidentifiable were transferred to a standard potting soil and grown until they could be identified. Identification was predominantly based on vegetative characters, as flowering was rare. When seedlings could not be identified, they were classified as unidentifiable monocotyledons or dicotyledons. The number of seeds per m 2 was calculated using the formula (from Levesque and Svoboda, 1995), Wta G = g — *100 Wtb 81 where, G = seeds/m2, g = number of germinated seedlings in a subsample, Wt a = weight of < 2 mm soil portion, Wt b = weight of subsample. Each 10 cm x 10 cm sample was converted to 1 m 2 by multiplying the formula by 100. S Y S T A T (1992) was used for all statistical analyses. A l l data were tested for normality using a Kolmorgorov-Smirnov goodness of fit (Lilliefors) test and for homogeneity of variances using the Fm a x-test. Where the data did not meet the appropriate assumptions, nonparametric tests were used. A paired-sample t test was used to test for a substrate effect (paleo-soil versus glacio-fluvial sediment) on the germinable seed bank for the first 18 belts. A Kruskall-Wallis one-way analysis of variance ( A N O V A ) was used to test for an age (belt) effect on the germinable seed bank for all 26 belts. Following the Kruskall-Wallis A N O V A , Tukey's multiple comparison tests were applied to the ranked data (Conover and Iman, 1981). Using belt averages, the relationship between the germinable seed bank and the overlying vegetation cover was examined using Pearson's and Spearman's correlation coefficients. 4.3.2 Fall-winter seed rain Seed traps consisted of 15 cm x 15 cm squares of artificial turf (n = 210) which were held in place by two nails pushed into the ground. Initially the traps covered a total area of 4.73 m 2 . On August 13, 1994, seed traps were placed randomly at 20 m intervals (n = 15/interval) with increasing distance (280 m) from the glacier snout over a 60 m breadth. The traps were collected the following spring (June 8, 1995) and emptied from the traps into envelopes and stored cool ( - 7 - 1 2 ° C ) until germination trials began. 82 Each seed rain sample was spread evenly over moistened Whatman No. 1 filter paper which lined 100 mm plastic petri dishes. The dishes were monitored for seedling emergence every 3 days for 23 days and moistened daily. Due to the lack of time and space in the field station laboratory, the samples were divided randomly into two subsets and germinated over two different time periods. The first germination trial took place in mid-July in the field station laboratory at Alexandra Fjord where the petri dishes were placed on tables and exposed to room temperatures ( 1 8 - 2 0 ° C ) and 24 hour natural light, which radiated through the laboratory windows. The second germination trial took place in early September at the Department of Geography, University of British Columbia. These petri dishes were exposed to room temperatures ( 2 3 - 2 5 ° C ) and 24 hour artificial (20 watt full spectrum fluorescent) light. The petri dishes were rotated daily, for both germination trials. The data from the two trials were combined into one overall data set; there was no significant difference between the two trials (Wilcoxon paired t-test, p = 0.084): A Kruskall-Wallis A N O V A was applied to the seed rain data to examine the significance of an age effect and thereafter Tukey's multiple comparison tests were applied to the ranked data (Conover and Iman, 1981). Pearson's correlation coefficients were used to test the relationship between the overlying vegetation cover and the seed rain densities, and to examine the significance of the relationship between the seed bank and seed rain densities. Belt averages were used for the correlation analyses. 83 4.3.3 Snow-core germination Snow-cores were collected with a Mont-Rose sampler (7 cm diameter) along five 200 m transects leading away from the glacier front, on June 2, 1995. Samples were collected and snow depths were recorded every 5 m (where possible) along the transects. Care was taken in collecting the samples, so that soil was not drawn up with the snow while coring. A total of 130 cores were collected, covering a total area of 0.50 m 2 . The cores were placed in plastic freezer bags and kept frozen until germination trials began (June 20, 1995). Just prior to germination, the cores were melted at room temperature ( 1 8 - 2 0 ° C ) and filtered through Whatman N o . l filter paper, which were then placed into 100 mm plastic petri dishes. The dishes were monitored for seedling emergence approximately every 3 days for 18 days and were moistened daily under room temperature and 24 hour natural light conditions in the field station lab. Statistical tests were not applied to the snow-core data due to its large variability and numerous zero values. 4.4 Results 4.4.1 Seed bank At least nine of 31 possible vascular species in the above-ground vegetation were observed in the seed bank germination trials. Among the species present in the seed bank were Luzula confusa, Papaver radicatum, Saxifraga cernua, Saxifraga oppositifolia, Saxifraga tricuspidata, Poa arctica, Draba spp, Cardamine bellidifolia, Dryas integrifolia, and several unidentifiable dicotyledons. A l l seed bank species were represented in the above-ground vegetation cover; however, the latter four species were 84 not found in the fall-winter seed rain (see below). Aside from Luzula confusa which was found on all terrain, except for that aged 0-2 years old, most of the seed bank taxa were found on terrain older than approximately 15 years (Table 4.1). Luzula confusa dominated the germinable seed bank, making up 78.7% of the emergent seedlings, followed by Papaver radicatum with 11.8%. O f the emergent dicotyledons, 52% were Papaver radicatum. The total number of germinable seeds observed in all the seed bank samples was 808, which translates into an average seed bank density over the entire site of 367 ± 32 seeds/m2 with values ranging from 0 to 2879 seeds/m2. The average densities (and ranges) for Luzula confusa, Papaver radicatum and all dicotyledons over the study area were 298 ± 2 9 (0-2500), 33 + 9 (0-1335) and 63 ± 11 (0-1335) seeds/m2, respectively. There was no significant difference in the germinable seed bank between substrate types; however, an age effect did exist for all species combined and Luzula confusa (p = 0.000 for both categories; Appendix 8). Dicotyledons and other individual species were not tested due to their small percentage of emergent seedlings in the germination trials. Tukey's multiple comparison tests for all species combined showed that, in general, young terrain had significantly lower seed bank densities than that of mid-aged to old terrain (Table 4.2). There was an overall increasing trend in seed bank density over time for the first 30-35 years (Figure 4.1a). The Luzula confusa seed bank density was significantly lower on young terrain compared to mid-aged terrain (Table 4.2; Figure 4.1b). 85 Table 4.1: Mean (+SE) seed bank densities (seeds/m2) for increasing distance and terrain age from the Twin Glacier snout. Species abbreviations are as follows: Car bel, Cardamine bellidifolia; Dra spp, Draba species; Dry int, Dryas integrifolia; Luz con, Luzula confusa; Pap rad, Papaver radicatum; Poa arc, Poa arctica; Sax cer, Sax cernua; Sax opp, Saxifraga oppositifolia; Sax tri, Saxifraga tricuspidata; unID dicot, unidentifiable dicotyledons. Distance Terrain (m) age(y) n Luz con Pap rad Sax cer Sax opp Sax tri Poa arc Dra spp Car bel Dry int unID dicot 0-10 0.25 10 0 0 0 0 0 0 0 0 0 0 10-20 1 10 0 0 0 0 10110 0 0 0 0 0 20-30 2 10 14±14 0 0 0 0 0 0 0 0 0 30-40 3 10 42±17 0 0 0 0 0 0 0 0 0 40-50 4 10 60+43 0 0 0 0 0 0 0 0 0 50-60 6 10 18+18 0 0 0 0 0 0 0 0 0 60-70 7 10 143±55 0 0 0 0 0 0 0 0 0 70-80 8 10 352+141 0 0 0 0 0 0 0 0 0 80-90 9/10 10 3521110 0 0 0 0 0 0 0 0 0 190-100 11/12 10 254±77 10110 0 0 0 0 0 0 0 0 100-110 13/14 10 449±120 4021402 0 0 0 0 0 0 0 0 110-120 16 10 412±182 0 0 0 0 0 0 0 0 5115 120-130 19 10 496±208 28120 0 0 0 0 0 0 0 26118 130-140 21 10 358+139 0 0 0 0 0 0 0 0 0 140-150 24 10 272±82 120+109 10+10 20113 0 0 0 0 0 0 150-160 26 10 5951253 192+140 0 0 0 0 0 0 0 10110 160-170 29 10 5811164 10110 0 0 0 13113 0 0 0 0 170-180 31 10 4971237 115173 0 40140 13113 0 0 0 0 20113 180-190 34 10 5981131 56133 0 0 0 10110 10110 0 0 86153 190-200 35 10 2811114 19119 0 0 0 0 29129 0 0 29129 200-210 37 10 293+111 62122 0 0 0 0 0 0 0 10110 210-220 39 10 3421127 37137 0 0 12112 0 0 0 0 80180 220-230 41 8 155163 93143 0 0 0 818 0 0 0 156196 230-240 42 8 197190 101138 0 0 0 0 0 0 0 0 240-250 43 6 7121373 0 0 0 64164 20120 0 0 0 17117 250-260 43 6 5491395 17117 0 0 43143 0 0 33133 17117 153183 Table 4.2: Tukey's multiple comparison results showing the terrain ages (years) which had significantly different seed bank densities for all species and for Luzula confusa; significance level, a = 0.05. all species Luzula confusa 0.25 < 9, 13-39, 43 0.25, 1, 2, 6 < 13, 19, 29, 34 1, 2 < 13, 19-34,37,39,43 3, 4 < 29, 34 3 < 13, 19,26-34, 43 4 < 13, 19,26,29, 34,43 6 < 13, 19, 24-34, 39, 43 7 < 3 4 86 Trimline 1200 900 - h 600 - f 300 + 10 15 20 25 30 Terrain age (years) 35 40 Trimline 1200 900 4-600 + 300 10 15 20 25 30 35 Terrain age (years) 40 Figure 4.1: Mean (+SE) germinable seed bank in relation to terrain age for a) all species combined and b) Luzula confusa. Terrain ages are not known beyond the trimline (> 43 years). 87 Correlation coefficients showed that the germinable seed bank was significantly positively correlated (p < 0.001) with the overlying vegetation cover for all species, dicotyledons, Luzula confusa (or monocotyledons less the small seed bank of Poa arctica) and Papaver radicatum (Figure 4.2). Although there were too few data to run correlation analyses on the other taxa represented in the seed bank, a general relationship was noted between the position of the species in the seed bank and those in the above-ground vegetation. The position of the later successional species in the seed bank, such as Saxifraga oppositifolia, Saxifraga tricuspidata, Cardamine bellidifolia and Dryas integrifolia appeared to correspond with their overall position in the existing vegetation. 4.4.2 Fall-winter seed rain O f the 210 seed traps set out in August 1994, 199 were recovered the following spring. The total area covered by the retrieved traps was 4.48 m 2 . A total of 1724 (94.9%) germinable and 92 (5.1%) nongerminable seeds were collected in the traps. Six of 31 possible vascular species in the above-ground vegetation were observed in the seed rain samples and, with the exception of Salix arctica, all were observed in the seed bank. A l l species observed in the seed rain were also observed as established adults in the overlying vegetation cover. The average seed rain density over the sampling area was 384 ± 47 seeds/m2. Mean seed rain densities for each terrain age interval are shown in Table 4.3, for each of the six species. Luzula confusa made up 98.4% (379 ± 46 seeds/m2) of the germinable seeds. The remaining species consisting of Salix arctica, Papaver radicatum, Saxifraga oppositifolia, Saxifraga tricuspidata and Poa arctica 88 1200 800 + 400 + 0 y= 192.24x + 93.196 r = 0.67(p<0.001) all species a) 800 600 y= 177.71x +79.649 r = 0.69(p<0.001) 1 Luzula confusa (monocotyledons) b) 0 1 500 400 300 200 100 0 500 400 300 200 100 y = 236.77x+ 10.883 _|_ rs=0.74(p<0.001) dicotyledons c) 0.0 0.2 0.4 0.6 0.8 1.0 y= 145.04+ 10.650 rs=0.72(p<0.001) Papaver radicatum d) 0.4 Cover (%) 0.8 Figure 4.2: Germinable seed bank in relation to plant cover for a) all species combined, b) Luzula confusa (monocotyledons), c) dicotyledons, and d) Papaver radicatum. Pearson's (r) or Spearman's (rs) correlation coefficients a shown with their respective p-values. 89 Table 4.3: Mean (±SE) fall-winter seed rain densities (seeds/m2) for increasing distance and terrain age from the Twin Glacier snout. Species abbreviations are as follows: Luz con, Luzula confusa; Pap rad, Papaver radicatum; Poa arc, Poa arctica; Sal arc, Salix arctica; Sax opp, Saxifraga oppositifolia; Sax tri, Saxifraga tricuspidata; unID dicot, unidentifiable dicotyledons. Distance Terrain n Luz con Sax opp Poa arc Pap rad Sax tri Sal arc unID (m) age (y) dicot 0-20 0.25-1 15 1061±227 3±3 12±7 0 0 0 0 20-40 2-3 15 293+142 0 0 0 0 0 0 40-60 4-6 11 20+9 0 0 0 0 0 0 60-80 7-8 12 230+53 0 0 0 0 0 0 80-100 9-12 13 205±62 0 0 0 0 0 0 100-120 13-16 14 343+85 0 0 0 0 0 0 120-140 19-21 13 345191 0 0 0 0 7+5 0 140-160 24-26 14 816±270 0 0 0 0 6±4 6+6 160-180 29-31 16 456±96 0 0 0 0 6+3 0 180-200 34-35 12 456+186 0 0 4+4 0 0 0 200-220 37-39 15 687±367 0 0 0 0 6±6 6±4 220-240 41-42 14 105±27 0 0 0 3+3 0 0 240-260 43 15 127±54 0 3 + 3 0 0 6+6 6+4 260-280 43+ 18 110±23 0 0 0 0 3±3 5±3 90 made up 1.6% of the emergent seedlings. Analysis concentrated on the trends observed for Luzula confusa due to its clear domination in the seed rain, and the low representation of all other species. A significant difference was detected in the number of germinable seeds/m2 for Luzula confusa between the 20 m sampling units (p = 0.000; Appendix 9). After its dramatic decrease from the first interval (0-20 m from glacier snout), the Luzula seed rain appeared to increase from the 40-60 m interval to the mid to late intervals (140-220 m) before the numbers drop at 220-240 m (Figure 4.3). Tukey tests showed that the first 20 m interval (0.25-1 years) contained significantly greater seeds/m2 than that of the following intervals: 20-40 m (2-3 years), 40-60 m (4-6 years), 100-120 m (13-16 years), 220-240 m (40-42 years), 240-260 m (43 years) and 260-280 m (43+ years). In addition, interval 40-60 m (4-6 years) contained a significantly lower density than intervals ranging from 100-120 (13-16 years) to 200-220 m (36-39 years). For Luzula confusa, there was no significant correlation between its overlying vegetation cover and seed rain density. However, when the large seed rain density occurring in the first 20 m was removed, a significant correlation was noted at the a = 0.1 significance level (0.1 < p < 0.05). Correlation analyses also showed that there was no significant correlation between the seed bank and seed rain densities along the terrain age gradient. 4.4.3 Snow-core germination A total of 90 (91.8%) germinable and 8 (8.2%) nongerminable seeds were observed in the snow-core germination trial. With the exception of one seed, determined 91 1500 CM o o o o o o o o o o o o O o CM 00 o CN oo o 40 years) the relationship between the seed rain and seed bank appears to be reversed, in that the mean seed rain is relatively low, while the mean seed 101 bank is relatively high. The low annual seed rain likely reflects the decreased cover of Luzula confusa in the existing vegetation, while the high seed bank reflects the slow accumulation of Luzula seeds in the soil over the past 43+ years. In addition, the number of flower heads per Luzula genet appeared to be smaller on older terrain (pers. obs.), indicating that the number of seeds dispersed from individual genets was also relatively low. It should be noted that the relatively high seed output of Luzula confusa on the Twin Glacier foreland contrasts a polar semi-desert study on King Christian Island where no viable seeds were produced from Luzula confusa during the three year study (Addison and Bliss, 1984). The harsher environmental conditions (cooler, shorter growing seasons) of the polar semi-desert site in relation to the polar oasis at Alexandra Fjord likely constrained seed production. Even after omitting the large seed rain density of the 0-20 m interval, no significant correlation was observed between the germinable seed bank and germinable seed rain densities for Luzula confusa. This emphasizes the importance of additional factors affecting the destiny of seeds, such as lateral transport, longevity in the soil and depth of burial. The seed rain captured in this study, like all other seed rain studies, does not take into consideration the lateral transport of a seed after it has dispersed from its seed source. Seed traps contain seeds that have been unnaturally halted at a particular location and do not represent the natural position at which the seeds come to rest (Ryvarden, 1971; Harper, 1977; Spence, 1990). Seed longevity in the field is unknown for Luzula confusa, making it difficult to determine whether all of the seeds (germinable + nongerminable) were accounted for in the seed bank germination trials. Direct counting of the remaining seeds in the soil seed bank samples was not done due to time constraints. Variability in 102 burial depth over the study site could partially account for the lack of correlation between the seed bank and seed rain densities. Seed bank samples should be collected at different depths to alleviate this problem. However, since the majority of seeds are found in the upper 2-3 cm of soil (Harper, 1977), variability in burial depth is likely not crucial in explaining the lack of correlation; seed bank samples were collected to a 2 cm depth on the Twin Glacier foreland. 4.6 Summary The ruderals, Luzula confusa and Papaver radicatum dominated the germinable seed bank. The fall-winter and winter (snow-core) seed rain were almost solely represented by L. confusa. The average germinable seed bank, fall-winter seed rain and winter seed rain densities were 367 ± 32, 384 ± 47 and 180 ± 53 seeds/m2, respectively. The significant positive correlation between the above-ground vegetation cover and the germinable seed bank for all species combined, dicotyledons, Luzula confusa (monocotyledons) and Papaver radicatum, suggests that species colonization is largely constrained by the availability of seed reaching a particular area. Complementing this proposal is the significant positive correlation between the above-ground vegetation cover and the germinable fall-winter seed rain for L. confusa. There was no significant correlation between the seed bank and fall-winter seed rain and there did not appear to be a relationship between the winter seed rain and the above-ground vegetation for L. confusa. The summer seed rain was generally lower and the winter seed rain was generally higher than values reported in other studies. The winter seed rain appears to be 103 as important as the summer seed rain on the Twin Glacier foreland. It is proposed that the large fall-winter seed rain observed at the foot of the glacier was due to the glacier acting as a barrier to wind swept seeds carried over.the smooth snow surface. The similar seed densities for the seed bank and fall-winter seed rain suggest high turnover rates which may explain the low seed bank density compared to the fall-winter seed rain at the foot of the glacier. There was no difference in the germinable seed bank between the finer grained paleo-soil and the coarser grained glacio-fluvial sediment. 104 Chapters Conclusions In the primary succession study on the Twin Glacier foreland, terrain age explained most of the variation in species composition. Over 44+ years, the succession proceeded through four main stages of dominance, following the classical directional-replacement mode of succession noted in most temperate and low arctic regions: mosses —> graminoid-forb —» deciduous shrub-moss —»evergreen dwarf shrub-moss. The directional-replacement mode of succession is likely atypical for most other high arctic habitats as the environmental conditions in the depauperate polar deserts and semi-deserts surrounding the Alexandra Fjord polar oasis are generally much more severe in terms of temperature and soil moisture (Freedman et al. 1994). Truly marginal environments are more likely to exhibit directional-nonreplacement or nondirectional-nonreplacement successions (sensu Svoboda and Henry, 1987). Some insight into the life history processes driving succession were gained by examining seed bank and seed rain patterns. The lack of late successional species (Cassiope tetragona and Dryas integrifolia), and the strong presence of early successional species (Luzula confusa and Papaver radicatum) in the seed bank and seed rain reflected the difference in life history traits between the two groups of species. The significant positive correlations between the above-ground vegetation cover and both the germinable seed bank and fall-winter seed rain suggests that colonization is largely constrained by seed availability. To confirm this proposal, experimental work, such as planting seedlings and sowing seeds into different successional stages, is needed to detect whether seeds and 105 seedlings of different successional stages could germinate and/or establish in successional stages that they are not naturally found (e.g. Walker et al., 1986; Chapin et al., 1994). There appeared to be no correlation between the above-ground vegetation cover and the number of germinable seeds/m2 in the snow-cores for Luzula confusa. Nonetheless, this component of the seed rain appeared to be relatively important on the Twin Glacier foreland compared to other winter seed rain studies (Ryvarden, 1971; Grulke and Bliss, 1983). In addition to the experimental work mentioned above, further work involving reproductive effort and seedling establishment and recruitment are needed in order to gain a more complete understanding of the seed dynamics and their role in the succession on the Twin Glacier foreland. Although competition is often considered to be one of the main mechanisms driving succession in temperate and alpine regions, it is unclear whether competition is important on the Twin Glacier foreland. It has been suggested that competition is negligible in high arctic habitats and that other factors such as facilitation and life history traits may be more important (Svoboda and Henry, 1987; Walker and Chapin, 1987; Matthews, 1992; Chapin et al., 1994). It is probable that competition is relatively unimportant in typical polar and semi-polar deserts, due to the large distances separating plants and the harsh environmental conditions. However, competition is likely more important in lush polar oases, such as the Alexandra Fjord Lowland, Ellesmere Island and Truelove Lowland, Devon Island, where vegetation cover reaches and often exceeds 100% in mature communities. Hence, competition may well be important in the late stages of succession on the Twin Glacier foreland. Aside from competition, it was also hypothesized, based on the theoretical model proposed by Walker and Chapin (1987), that 106 bryophytic facilitation and seed arrival were important in the early stages and that longevity and mycorrhizal associations were important in the later stages of succession on the Twin Glacier foreland. To grasp an understanding of the relative importance of both biotic and abiotic processes in high arctic successions intensive work surrounding successional processes, including experimental work, is required. For example, removal experiments would indicate whether specific processes are acting: removal of a species may cause increased or decreased growth in a neighbouring species which would suggest that competition or facilitation is occurring, respectively. The presence of the pre-LIA paleo-soil, which has significantly higher levels of a combined set of soil parameters (pH, organic matter, moisture and fine substrate) compared to the glacio-fluvial sediment, adds an element to the Twin Glacier foreland study that is not encountered in more temperate glacier foreland studies. Despite the similarity in the germinable seed bank between the two substrate types, the total vegetation cover was significantly higher and the species richness was slightly greater on the paleo-soil. The dominant pioneering moss, Psilopilum cavifolium, influenced the difference in the total vegetation cover greatly, as this species was very abundant on the paleo-soil, which agrees with the fact that Psilopilum cavifolium generally prospers on organic mats (W.B. Schofieldpers. comm., 1996). These observations, in general, suggest that the paleo-soil provides conditions which are more favourable for establishment and growth than that provided by the glacio-fluvial sediment; the paleo-soil may therefore positively influence the overall rate of succession. Despite the differences between the two substrate types, the successional sequences of the paleo-soil and glacio-fluvial data sets were similar. 107 References cited Addison, P.A. and L . C . Bliss. 1984. Adaptations of Luzula confusa to the polar semi-desert environment. Arctic, 37(2): 121-132. Archer, A . C . 1973. Plant succession in relation to a sequence of hydromorphic soils formed on glacio-fluvial sediments in the Alpine zone of the Ben Ohau Range, New Zealand. New Zealand Journal of Botany, 11: 331-348. Babb, T . A . and L . C . Bliss. 1974. Effects of physical disturbance on high arctic vegetation in the Queen Elizabeth Islands. Journal of Applied Ecology, 11: 549-562. Barrett, P. and R. Schulten. 1975. Disturbance and the successional response of arctic plants on polar desert habitats. Arctic, 28: 70-73. Bell, K . L . 1975. Aspects of seed production and germination in some high arctic plants. In Plant and Surface Responses to Environmental Conditions in the Western High Arctic. Edited by L . C . Bliss. A L U R Programme, Department of Indian and Northern Affairs, Ottawa, pp 62-72. Bell, K . L . and L . C . Bliss. 1980. Plant reproduction in a high arctic environment. Arctic and Alpine Research, 12(1): 1-10. Benedict, J.B. 1989. Use of Silene acaulis for dating: the relationship of cushion diameter to age. Arctic and Alpine Research, 21(1): 91-96. Bergsma, B . M . , J. Svoboda, and B. Freedman. 1984. Entombed plant communities released by a retreating glacier at Alexandra Fjord. Arctic, 37(1): 49-52. Beschel, R . E . 1961. Botany; and some remarks on the history of vegetation and glacierization. In Jacobsen-McGill Arctic Research Expedition to Axel Heiberg Island, Queen Elizabeth Islands, Preliminary Report 1959-1960. Edited by F . Muller. M c G i l l University, Montreal, pp 179-199. Beschel, R . E . 1963. Observations on the time factor in interactions of permafrost and vegetation. In Proceedings of the First Canadian Conference on Permafrost, April 17 and 18 1962. Edited by R .J .E . Brown. Ottawa, pp 43-55. Beschel, R . E . and D . Webb. 1963. Growth ring studies on Arctic Willows. In Axel Heiberg Island Research Report, Preliminary report 1961-1962. Edited by F . Muller. M c G i l l University, Montreal, pp 189-198. 108 Billings, W . D . 1987. Constraints to plant growth, reproduction, and establishment in arctic environments. Arctic and Alpine Research, 19(4): 357-365. Billings, W . D . and H . A . Mooney. 1968. The ecology of arctic and alpine plants. Biological Review, 43: 481-529. Billington, M . and V . Alexander. 1978. Nitrogen fixation in a black spruce (Picea mariana [Mill] B.S.P.) forest in Alaska. In Environmental Role of Nitrogen-Fixing Blue-green Algae and Asymbiotic Bacteria. Edited by U . Granhall. Ecological Bulletins, Stockholm Volume 26. pp 209-215. Blake, W . 1981. Neoglacial fluctuations of glaciers, southeastern Ellesmere Island, Canadian Arctic, Archipelago. Geografiska Annaler, 63(A): 201-218. Blakemore, L . C . , P .L. Searle, B . K . Daly. 1987. Methods for Chemical Analysis of Soils. New Zealand Bureau, Lower Hutt, New Zealand. Bliss, L . C . 1962. Adaptations of arctic and alpine plants to environmental conditions. Arctic, 15: 117-144. Bliss, L . C . 1971. Arctic and alpine plant life cycles. Annual Review of Ecology and Systematics, 2: 405-438. Bliss, L . C . and W . G . Gold. 1994. The patterning of plant communities and edaphic factors along a high arctic coastline: implications for succession. Canadian Journal of Botany, 72: 1095-1107. Bliss, L . C . and N . E . Grulke. 1988. Revegetation in the High Arctic: its role in reclamation of surface disturbances. In Northern environmental disturbances. Edited by G.P. Kershaw. Boreal Institute for Northern Studies, pp 43-55. Bliss, L . C . and K . M . Peterson. 1992. Plant succession, competition and the physiological constraints of species in the Arctic. In Arctic Ecosystems in a Changing Climate, an Ecological Perspective. Edited by F.S. Chapin III, R . L . Jefferies, J.F. Reynolds, G.R. Shaver and J. Svoboda. Academic Press, San Diego, California, pp 111-136. Blundon, D.J. 1989. Primary Succession Following Retreat of the Robson Glacier, British Columbia. Ph.D. thesis, Department of Botany, University of Alberta, Edmonton, Alberta. Bormann, B . T . and R . C . Sidle. 1990. Changes in productivity and distribution of nutrients in a chronosequence at Glacier Bay National Park, Alaska. Journal of Ecology, 78: 561-578. Bray, R . H . and L . T . Kurtz. 1945. Determination of total, organic, and available forms of phosphorous in soils. Soil Science, 59(1): 39-45. 109 Callaghan, T . V . and B . A . Carlsson and N.J.C. Tyler. 1989. Historical records of climate-related growth in Cassiope tetragona from the Arctic. Journal of Ecology, 77(3): 823-837. Callaghan, T . V . and U . Emanuelsson. 1985. Population structure and processes of tundra plants and vegetation. In The Population Structure of Vegetation. Edited by J. White. Dr. W . Junk Publishers, Dordrecht, Netherlands, pp 399-439. Chambers, J.C. 1993. Seed and vegetation dynamics in an alpine herb field: effects of disturbance type. Canadian Journal of Botany, 71: 471-485. Chambers, J.C. 1995. Disturbance, life history strategies and seed fates in alpine herbfield communities. American Journal of Botany, 82(3): 421-433. Chambers, J .C. and J.A. MacMahon. 1994. A day in the life of a seed: Movements and fates of seeds and their implications for natural and managed systems. Annual Review of Ecology and Systematics, 25: 263-292. Chambers, J . C , J.A. MacMahon and R.W. Brown. 1990. Alpine seedling establishment: the influence of disturbance type. Ecology, 71(4): 1323-1341. Chambers, J . C , J.A. MacMahon and J.H. Haefner. 1991. Seed entrapment in alpine ecosystems: effects of soil particle size and diaspore morphology. Ecology, 72(5): 1668-1677. Chapin, F.S. III. 1987. Environmental controls over growth of tundra plants. Ecological Bulletins, 38: 69-76. Chapin, F.S. Ill, L .R . Walker, C L . Fastie and L . C . Sharman. 1994. Mechanisms of primary succession following deglaciation at Glacier Bay, Alaska. Ecological Monographs, 64(2): 149-175. Clements, F . E . 1916. Plant succession. Carnagie Institute of Washington Publication, No. 242. Collins, N.J. 1976. The development of moss-peat banks in relation to changing climate and ice cover on Signy Island in the maritime Antarctic. British Antarctic Survey Bulletin, No. 43: 85-102. Connell, J .H. and R.O. Slatyer. 1977. Mechanisms of succession in natural communities and their role in community stability and organization. The American Naturalist, 111:1119-1144. Conover, W.J. and R . L . Iman. 1981. Rank transformations as a bridge between parametric and nonparametric statistics. The American Statistician, 35(3): 124-129. 110 Cooper, W.S. 1916. Plant succession in the Mount Robson region, British Columbia. Plant World, 19: 211-238. Cooper, W.S. 1923. The recent ecological history of Glacier Bay, Alaska. Ecology, 4: 93-128; 223-246; 355-365. Densmore, R. and J. Zasada. 1983. Seed dispersal and dormancy patterns in northern willows: ecological and evolutionary significance. Canadian Journal of Botany, 61: 3207-3216. Desrosiers, J. 1991. Patterns of growth and biomass allocation in Saxifraga oppositifolia in the High Arctic. MSe. thesis, Department of Botany, University of Alberta, Edmonton. Diemer, M . and S. Prock. 1993. Estimates of alpine seed bank size in two Central European and one Scandinavian subarctic plant communities. Arctic and Alpine Research, 25(3): 194-200. Drury, W . H . and I.C.T. Nisbet. 1973. Succession. Journal of the Arnold Arboretum, 54(3): 331-368. Ebersole, J.J. 1989. Role of seed bank in providing colonizers on a tundra disturbance in Alaska. Canadian Journal of Botany, 67: 466-471. Edlund, S.A. and B .T . Alt. 1989. Regional congruence of vegetation and summer climate patterns in the Queen Elizabeth Islands, Northwest Territories, Canada. Arctic, 42(1): 3-23. Egler, F . E . 1954. Vegetation science concepts I. Initial floristic composition, a factor in old-field vegetation development. Vegetatio, 4: 412-417. Elven, R. and L . Ryvarden. 1975. Dispersal and primary establishment of vegetation. In Fennoscandian Tundra Ecosystems, Part I. Edited by F . E . Wielgolaski. Springer-Verlag, Berlin, pp 81-85. Falconer, G . 1966. Preservation of vegetation and patterned ground under a thin ice body in northern Baffin Island, N .W.T . Geographical Bulletin, 8(2): 194-200. Fastie, C . L . 1995. Causes and ecosystem consequences of multiple pathways of primary succession at Glacier Bay, Alaska. Ecology, 76(6): 1899-1916. Fenner, M . 1985. Seed Ecology. Chapman and Hall, London. Fenton, J .H .C. 1982. Vegetation re-exposed after burial by ice and its relationship to changing climate in the South Orkney Islands. British Antarctic Survey Bulletin, No. 51: 247-255. I l l Forbes, B . C . 1993. Anthropogenic Tundra Disturbance and Patterns of Response in the eastern Canadian Arctic. Ph.D. thesis, Department of Geography, M c G i l l University, Montreal, Quebec. Fox, J.F. 1983. Germinable seed banks of interior alaskan tundra. Arctic and Alpine Research, 15(3): 405-411. Freedman, B. , N . Hil l , J. Svoboda and G . Henry. 1982. Seed banks and seedling occurrence in a high Arctic oasis at Alexandra Fjord, Ellesmere Island, Canada. Canadian Journal of Botany, 60: 2112-2118. Freedman, B. , J. Svoboda and G.H.R. Henry. 1994. Alexandra Fjord - an ecological oasis in the polar desert. In Ecology of a Polar Oasis, Alexandra Fjord, Ellesmere Island, Canada. Edited by J. Svoboda and B. Freedman. Captus University Publications, Toronto, pp 1-9. Gartner, B . L . 1983 Germination characteristics of arctic plants. In Permafrost: Proceedings of the 4th International Conference on Permafrost, Fairbanks, Alaska, pp 334-338. Gleason, H . A . 1926. The individualistic concept of the plant association. Bulletin of the Torrey Botanical Club, 53: 7-26. Goldthwait, R .D. 1960. Study of an ice-cliff in Nunatarsuag, Greenland. U.S. Army Snow, Ice and Permafrost Research Establishment (SIPRE), Technical Report 39. Grime, J.P. 1977. Evidence for the existence of three primary strategies in plants and its relevance to ecological and evolutionary theory. The American Naturalist, 111: 1169-1194. Grime, J.P. 1979. Plant Strategies and Vegetation Processes. Wiley and Sons, New York. Grulke, N . E . and L . C . Bliss. 1983. A note on winter seed rain in the high arctic. Arctic and Alpine Research, 15(2): 261-265. Harper, J .L. 1977. Population Biology of Plants. Academic Press, London. Havstrom, M . , T . V . Callaghan, S. Jonasson and J. Svoboda. 1995. Little Ice Age temperature reduction measured by reduced growth of an arctic heather. Functional Ecology, 9(4): 650-654. Helm, D.J. and E . B . Allen. 1995. Vegetation chronosequence near Exit Glacier, Kenai Fjords National Park, Alaska, U.S .A. Arctic and Alpine Research, 27(3): 246-257. Henry, G.H.R. and J. Svoboda. 1986. Dinitrogen fixation (acetylene reduction) in high arctic sedge meadows. Arctic and Alpine Research, 18(2): 181-187. 112 Henry, G , B . Freedman and J. Svoboda. 1986. Survey of vegetated areas and muskox populations in east-central Ellesmere Island. Arctic, 39(1): 78-81. Hil l , M . O . 1979. T W I N S P A N - a F O R T R A N Program for Arranging Multivariate Data in an Ordered Two-Way Table by Classification of the Individuals and Attributes. Cornell University, Ithaca, N Y . Huston, M . and T. Smith. 1987. Plant succession: life history and competition. The American naturalist, 130(2): 168-198. Ingersoll, C A . and M . V . Wilson. 1993. Buried propagule bank of a high subalpine site: microsite variation and comparisons with aboveground vegetation. Canadian Journal of Botany, 71: 712-717. Ireland, R.R., C D . Bird, G.R. Brassard, W . B . Schofield and D . H . Vitt. 1980. Checklist of Mosses of Canada. National Museum of Canada, Ottawa. Jeffrey, D .W. Soil-Plant Relationships, A n Ecological Approach. Timber Press, Portland. Johnson, E . A . 1975. Buried seed populations in the subarctic forest east of Great Slave Lake, Northwest Territories. Canadian Journal of Botany, 53: 2933-2941. Kellman, M . C 1970. The viable seed content of some forest soil in coastal British Columbia. Canadian Journal of Botany, 53: 2933-2941. Kohn, L . M . and E . Stasovski. 1990. The mycorrhizal status of plants at Alexandra Fjord, Ellesmere Island, Canada, a high-arctic site. Mycologia, 82(1): 23-35. Johnstone, J. and G.H.R. Henry. 1997. Retrospective analysis of growth and reproduction in Cassiope tetragona and relations to climate in the Canadian High Arctic. Arctic and Alpine Research (submitted). Jongman, R . H . G . , C.J.F. ter Braak and O.F.R. van Tongeren. 1995. Data Analysis in Community and Landscape Ecology. Pudoc, Wageningen. Lawrence, D .B . 1958. Glaciers and vegetation in southeast Alaska. American Scientist, 46: 89-122. Lawson, D . E . 1995. Sedimentary and hydrologic processes within modern terrestrial valley glaciers. In Modern Glacial Environments: Processes, Dynamics and Sediments (Glacial Environments: Volume 1). Edited by J. Menzies. Butterworth-Heinemann, U K . pp 337-363. Leek, M . A . 1980. Germination in Barrow, Alaska, tundra soil cores. Arctic and Alpine Research, 12(3): 343-349. 113 Levesque, E . 1996. Minimum area and cover-abundance scales as applied to polar desert vegetation. Arctic and Alpine Research, 28(2): 156-162. Levesque, E . and J. Svoboda. 1995. Germinable seed bank from polar desert stands, Central Ellesmere Island, Canada. In Global Change and Arctic Terrestrial Ecosystems. Edited by T . V . Callaghan, U . Molau, M . J . Tyson, J.I. Holten, W . C . Oechel, T. Gilmanov, B . Maxwell, B. Sveinbojorrnsson. Proceedings of papers contributed to the International Conference, 21-26 August 1993, Oppdal, Norway. Ecosystems research report 10, European Commission, pp 97-107. Marchand, P.J. and D . A . Roach. 1980. Reproductive strategies of pioneering alpine species: seed production, dispersal and germination. Arctic and Alpine Research, 12(2): 137-146. Matthews, J .A. 1978. Plant colonization patterns on a gletschervorfeld, southern Norway: a meso-scale geographical approach to vegetation change and phytometric dating. Boreas, 7: 155-178. Matthews, J.A. 1979. Refutation of convergence in a vegetation succession. Naturwissenschaften, 66: 47-49. Matthews, J .A. 1992. The Ecology of Recently-Deglaciated Terrain: a Geoecological Approach to Glacier Forelands and Primary Succession. Cambridge University Press, Cambridge, Great Britain. Matthews, J .A. and R.J. Whittaker. 1987. Vegetation succession on the Storbreen glacier foreland, Jotunheimen, Norway: A review. Arctic and Alpine Research, 19(4): 385-395. McGraw, J.B. 1980. Seed bank size and distribution of seeds in cottongrass tussock tundra, Eagle Creek, Alaska. Canadian Journal of Botany, 58: 1607-1611. McGraw, J.B. and M . C . Vavrek. 1989. The role of buried viable seeds in Arctic and alpine plant communities. In Ecology of Soil Seed Banks. Edited by M . A . Leek, V . T . Parker and R . L . Simpson. Academic Press, California, pp 91-105. McKeague, J.A. 1978. Manual on Soil Sampling and Methods of Analysis. Canadian Society of Soil Science. 212 pp. Miller, O . K . 1982. Mycorrhizae, mycorrhizal fungi, and fungal biomass in subalpine tundra near Eagle Summit, Alaska. Ffolarctic Ecology, 5: 125-134. Muc, M . , J, Svoboda and B. Freedman. 1994. Aboveground standing crop in plant communities of a polar desert oasis. In Ecology of a Polar Oasis, Alexandra Fjord, Ellesmere Island, Canada.. Edited by J. Svoboda and B. Freedman. Captus University Publications, Toronto, pp 65-74. 114 Muller, C . H . 1952. Plant succession in arctic heath and tundra in northern Scandinavia. Bulletin of the Torrey Botanical Club, 79: 296-309. Murray, D.F . 1987. Breeding systems in vascular flora in Arctic North America. In Differentiation Patterns in Higher Plants. Edited by K . M . Urbanska. Academic Press, Orlando, Florida, pp 239-262. Murray, D . F . 1995. Causes of arctic plant diversity: origin and evolution. In Arctic and Alpine Biodiversity: Patterns, Causes and Ecosystem Consequences. Edited by F.S. Chapin III and C . Korner. Springer-Verlag, Berlin, pp 21-32. Noble, I.R. and R.O. Slatyer. 1980. The use of vital attributes to predict successional change in plant communities subject to recurrent disturbances. Vegetatio, 43: 5-21. Oberbauer, S. and P.C. Miller. 1982. Effect of water potential on seed germination. Holarctic Ecology, 5: 218-220. Page, A . L . , R . H . Miller and D.R. Keeney. 1982. Methods of Soil Analysis Part 2: Chemical and Microbiological Properties. American Society of Agronomy Inc, Soil Science Society of America Inc., Madison, Wisconsin. Palmer, W . H . and Miller, A . K . 1961. Botanical evidence for the recession of a glacier. Oikos, 12(1): 75-86. Persson, A. 1964. The vegetation at the margin of the receding Glacier Skaftafellsjokull, southeastern Iceland. Botaniska Notiser, 117(4): 323-354. Polunin, N . 1959. Circumpolar Arctic Flora. Oxford University Press, Great Britain. Porsild, A . E . and W.J . Cody. 1980. Vascular plants of continental Northwest Territories, Canada. National Museum of Natural Sciences. Prach, K . 1993. On the rate of succession. Oikos, 66(2): 343-345. Ryvarden, L . 1971. Studies of seed dispersal I. Trapping of diaspores in the alpine zone at Finse, Norway. Norwegian Journal of Botany, 18: 215-226. Ryvarden. L . 1975. Studies of seed dispersal II. Winter-dispersed species at Finse, Norway. Norwegian Journal of Botany, 22: 21-24. Savile, D . B . O . 1960. Limitations of the competitive exclusion principle. Science, 132: 1761. Savile, D . B . O . 1972. Arctic Adaptations in Plants. Monograph No. 6, Canada Department of Agriculture. Information Canada, Ottawa. 115 Scherff, E . J . , C . Galen and M . L . Stanton. 1994. Seed dispersal, seedling survival and habitat affinity in a snowbed plant: limits to the distribution of the snow buttercup, Ranunculus adoneus. Oikos, 69: 405-413. Schyatt, V . 1961. Blue icefields, moraine features and glacier fluctuations. Scientific results, Norwegian-British-Swedish Antarctic Expedition 1949-1952, Glaciology II, v.IV: 183-204. Shaver, G.R. and J. Kummerow. 1992. Phenology, resource allocation, and growth of arctic vascular plants. In Arctic Ecosystems in a Changing Climate, an Ecological Perspective. Edited by F.S. Chapin III, R . L . Jefferies, J.F. Reynolds, G.R. Shaver and J. Svoboda. Academic Press, San Diego, California, pp 193-211. Smith, R.I.L. 1982. Plant succession and re-exposed moss banks on a deglaciated headland in Arthur Harbour, Anvers Island. British Antarctic Survey Bulletin, No. 51: 193-199. Sninzow, G . K . 1962. Investigation of shear zones in the ice sheet margin, Thule area, Greenland, 4(32): 215-229. Sohlberg, E . and L . C . Bliss. 1984. Microscale patterns of vascular plant distribution in two high-arctic communities. Canadian Journal of Botany, 62: 2033-2042. S0renson, T . 1941. Temperature relations and phenology of the north-east Greenland flowering plants. Meddelelser om Gronland, 125: 1-305. Spence, J.R. 1990. Seed rain in grassland, herbfield, snowbank and fellfield in the alpine zone, Craigieburn Range, South Island, New Zealand. New Zealand Journal of Botany, 28: 439-450. Stocklin, J. and E . Baumler. 1996. Seed rain, seedling establishment and clonal growth strategies on a glacier foreland. Journal of Vegetation Science, 7: 45-56. Stork, A . 1963. Plant immigration in front of retreating glaciers, with examples from the Kebnekajse area, northern Sweden. Geografiska Annular, 45: 1-22. Svoboda, J. and G.H.R. Henry. 1987. Succession in marginal arctic environments. Arctic and Alpine Research, 19(4): 373-384. S Y S T A T / o r Windows: Statistics, Version 5 Edition. 1992. Evanston, IL: S Y S T A T , Inc. Syers, J .K. and T . W . Walker. 1969. Phosphorous transformations in a chronosequence of soils developed on a windblown sand in New Zealand. I. Total and organic phosphorous. Journal of Soil Science, 20: 57-64. 116 ter Braak, C.J.F. 1987. C A N O C O - a F O R T R A N Program for Canonical Community Ordination by [Partial] [Detrended] [Canonical] Correspondence Analysis, Principal Components Analysis and Redundancy Analysis (Version 2.1). T N O Institute of Applied Computer Science, Wageningen. ter Braak, C.J.F. 1990. Update notes: C A N O C O Version 3.10. Agricultural Mathematics Group, Wageningen. Thompson, K. and J.P. Grime. 1979. Seasonal variation in the seed banks of herbaceous species in ten contrasting habitats. Journal of Ecology, 67: 893-921. Tilman, D . 1985. The resource-ratio hypothesis of plant succession. The American Naturalist, 125(6): 827-852. Tisdale, E . W . , M . A . Fosberg, C . E . Poulton. 1966. Vegetation and soil development on a recently glaciated area near Mount Robson, British Columbia. Ecology, 47(4): 517-523. Tishkov, R.J. 1986. Primary succession in arctic tundra on the west coast of Spitsbergen (Svalbard). Polar Geography and Geology, 10: 148-156. Urbanska, K . M . and M . Schiitz. 1986. Reproduction by seed in alpine plants and revegetation research above timberline. Botanica Helvetica, 96(l):43-60. Vetaas, O R . 1994. Primary succession of plant assemblages on a glacier foreland-B0dalsbreen, southern Norway. Journal of Biogeography, 21: 297-308. Viereck, L . A . 1966. Plant succession and soil development on gravel outwash of the Muldrow Glacier, Alaska. Ecological Monographs, 36(3): 181-199. Walker, L . R . and F.S. Chapin III. 1986. Physiological controls over seedling growth in primary succession on an Alaskan floodplain. Ecology 67(6): 1508-1523. Walker, L . R . and F.S. Chapin III. 1987. Interactions among processes controlling successional change. Oikos, 50: 131-155. Walker, L . R . , J .C. Zasada and F.S. Chapin III. 1986. The role of life history processes in primary succession on an Alaskan floodplain. Ecology, 67(5): 1243-1253. Wein, R . W . and A . N . Rencz. 1976. Plant cover and standing crop sampling procedures for the Canadian High Arctic. Arctic and Alpine Research, 8(2): 139-150. Whittaker, R.J. 1991. The vegetation of the Storbreen gletschervorfeld, Jotunheimen, Norway. IV Short-term vegetation changes. Journal of Biogeography, 18: 41-52. 117 Whittaker, R.J. 1989. The vegetation of the Storbreen gletschervorfeld, Jotunheimen, Norway. III. Vegetation-environment relationships. Journal of Biogeography, 16: 413-433. Whittaker, R.J. 1991. The vegetation of the Storbreen gletschervorfeld, Jotunheimen, Norway. IV. Short-term vegetation change. Journal of Biogeography, 18: 41-52. Whittaker, R.J. 1993. Plant population patterns in a glacier foreland succession: pioneer herbs and later-colonizing shrubs. Ecography, 16: 117-136. Wijk, S. 1986. Influence of climate and age on annual shoot increment in Salix herbacea. Journal of Ecology, 74: 685-692. Wojciechowski, M . F . and M . E . Heimbrook. 1984. Dinitrogen fixation in alpine tundra, Niwot Ridge, Front Range, Colorado, U.S .A. Arctic and Alpine Research, 16: 1-10. Wood, D . M . and W . F . Morris. 1990. Ecological constraints to seedling establishment on the Pumic Plains, Mount St. Helens, Washington. American Journal of Botany, 77: 1411-1418. Worsley, P. and M . R . Ward. 1974. Plant colonization of recent \"annual\" moraine ridges at Austre Okstindbreen, North Norway. Arctic and Alpine, 6(2): 217-230. Young, S.B. 1989. To the Arctic: A n Introduction to the Far Northern World. John Wiley and Sons, Toronto. 118 Appendices Appendix 1 Interpretation of glacial advances on the Alexandra Fjord foreland from radiocarbon dates In August 1994, three rooted pre-Little Ice Age (-LIA) Salix arctica specimens were collected for radiocarbon analyses on the foreland of the western lobe of the retreating Twin Glacier. These specimens were collected within the zone bounded by the maximum ice advance during the L I A (ca. 1850 ) and the 1994 snout position (Plate 2.2). The Salix specimens were collected at 30, 70 and 240 m from the glacier snout and their corresponding radiocarbon dates were 830+65 (WAT-2940), 240+65 (WAT-2941) and 2 6 0 ± 6 5 (WAT-2942) years before present, respectively. Bergsma et al. (1984) reported radiocarbon dates of 4 0 0 ± 1 4 0 , 4 1 0 ± 4 5 and 4 3 0 ± 9 0 years for plant specimens (Cassiope tetragona and Salix arctica) collected from the edge of the western lobe terminus, between 1981 and 1983 (~ 100-120 m from the 1994 glacier snout position). Taking into account the approximate age of the plant specimens prior to their burial by ice, the authors concluded that the plant community presently being released from the Twin Glacier was entombed by ice between the years 1410 and 1690. The 240 year date is thought to be unreliable as it is spatially situated between the 830 year and 400-430 year dates. Assuming that the Salix specimens were at least 20-25 years old before their burial by ice, the 830 year date suggests that the newly exposed portion of the relict plant community may have been buried as early as A.D. 1120. Furthermore, the date suggests that this glacier and possibly others in the High Arctic were already advancing during or at the end of the Little Climatic Optimum (A.D. 950-1250.). The slightly warmer climate during this period may have triggered more precipitation between autumn and spring (G. Henry and J . Svoboda, pen. comm., 1997); increased snowfall would have fed the source areas abover the permanent snow line on the ice fields and potentially allowed for expansion of the outlet glaciers. Moreover, a shift towards increased snowfall may have resulted in large permanent snow beds at the foot of glaciers that did not completely melt during the warmer summers. 119 A p p e n d i x 2 Plant species list for the Twin Glacier (western lobe) foreland, Alexandra Fjord, Ellesmere Island, Canada. Nomenclature follows Porsild and Cody (1980) for vascular plants and Ireland et al. (1980) for mosses. Relict mosses were identified from a collection within 50 cm of the ice edge. f Passive species in ordination analyses; * Species outside of quadrats. Vascular Plants Mosses Relict Vegetation Cardamine bellidifolia Aulacomnium turgidum Vascular plants: Carex nardina Barbula spp* Cassiope tetragona Cassiope tetragona Bartramia ithyphylla Luzula arctica Cerastium alpinum Bryum argenteum Luzula confusa Draba lactea Bryum dioicous Carex nardina Draba nivalis Ceratodon purpureus Salix arctica Draba oblongata Conostomum tetragonum Dryas integrifolia Draba subcapitata cf. Cynodontium Saxifraga oppositifolia Draba spp Desmatodon cf. latifolius Saxifraga tricuspidata Dryas integrifolia Dicranella crispa Epilobium latifolium Dicranella cf. grevilleana Mosses: Erysimum Pallasii* Dicranowesia cf. crispula Aulacomnium turgidum Festuca brachyphylla cf. Dicranum Bartramia ithyphylla Hierochloe alpina* Didymodon spp Distichium cf. capillaceum Luzula arctica' Distichium capillaceum* Ditrichum flexicaule Luzula confusa Ditrichumflexicaule* Pogonatum urnigerum Melandrium apetalum Encalyptra rhaptocarpa Polytricum piliferum Minuartia rubella Funaria groutiana Racomitrium aciculare Oxyria digyna* Grimmia spp Racomitrium lanuginosum Papaver radicatum Orthotrichum speciosum' Trichostomum cuspidatissimum Poa arctica Pogonatum dentatum Salix arctica Pogonatum urnigerum Saxifraga caespitosa} Pohlia cruda Saxifraga cernua Pohlia drummondii Saxifraga nivalis Pohlia nutans Saxifraga oppositifolia Pohlia proligera Saxifraga rivularisi Pohlia cf. wahlenbergii Saxifraga tricuspidata Polytricum alpinum Silene acaulis^ Polytricum juniperinum Stellaria longipes^ Psilopilum cavifolium Vaccinium uliginosum* Racomitrium canescens Racomitrium heterostichum Racomitrium lanuginosum Racomitrium panschii Saelania glaucescens* Tayloria spp* Tortula ruralis Tricostomum cf. cuspidatissimum' 120 A p p e n d i x 3 Appendix Table 3.1: Summary of one-way analysis of variance to test for a terrain age effect on a) nitrate-nitrogen, b) available phosphorous and c) exchangeable potassium. Phosphorous data were logio -transformed prior to analysis and the nonparametric Kruska l l -Wal l i s A N O V A was used for the nitrogen data. Significance level, a = 0.05. a) Group (age in years) n Rank sum 1(2) 5 61.000 2(9) 5 42.500 3(22) 5 46.000 4(44) 5 60.500 K-W statistic = 1.600; p = 0.659;df = 3 b) Source Sum of Squares df Mean Square F-ratio p-value r 2 Age 0.059 3 0.020 0.221 0.881 0.040 Error 1.434 16 0.090 c) Source Sum of Squares df Mean Square F-ratio p-value r 2 Age 772.000 3 257.333 0.867 0.478 0.140 Error 4747.200 16 296.700 Appendix Table 3.2: Summary of multivariate analysis of covariance testing for a substrate (paleo-soil versus glacio-fluvial sediment) effect on soil parameters (soil p H , organic matter content, soi l moisture content and fine substrate), where terrain age (distance) is the covariable. Data were rank transformed prior to analysis. Significance level, a = 0.05. Multivariate statistic value df F-statistic p-value Wilk's lamda 0.212 4, 93 86.377 0.000 Pillai trace 0.788 4, 93 86.377 0.000 Hotelling-Lawley trace 3.715 4, 93 86.377 0.000 Appendix Table 3.3: Summary of an analysis of covariance testing for a substrate (paleo-soil versus glacio-fluvial sediment) effect on total vegetation cover, where terrain age is the covariable. Data were rank-transformed before analysis. Significance level , a = 0.05. Source Sum of Squares df Mean Square F-ratio p-value r 2 substrate 21941.333 1 21941.333 35.193 0.000 0.302 age 4743.079 1 4743.079 7.608 0.007 error 61723.088 99 623.466 121 Appendix 4 Vegetation survey on a glacier foreland (79°02'N; 78°58'W), Beitstad Fjord, Ellesmere Island The vegetation communities varied greatly on the gneiss-granite based glacier foreland at the head of Beitstad Fjord, Ellesmere Island (Figure 1.1) due to low water-collecting depressions, which were dominated by Eriophorum angustifolium and E. Scheuchzeri. Sampling took place on a slightly elevated portion of the foreland where conditions were more comparable to the other forelands that were sampled during this research. Relict vegetation and paleo-soil were not observed on the study site. A boulder trimline, approximately 40 m from the glacier snout was thought to delineate the maximum extent of the Little Ice Age glacial advance (ca 1850 A.D.) . Species presence/absence data were recorded in 50 cm x 50 cm quadrats that were placed every 3 m along two 57 m transects, leading away from the glacier snout. The two transects were approximately 100 m apart. The graph below shows the combined results of the two transects, which indicates that the mosses were the first colonizers and that there was no direct replacement by the extremely sparse vascular plants. The relatively wide variety of species (Draba cinerea, Epilobium latifolium, Luzula confusa, Papaver radicatum, Minuartia rubella, Stellaria spp, Cassiope tetragona, Saxifraga nivalis, S. cernua, S. tricuspidata, S. caespitosa and Trisetum spicatum) that were observed on the study site, in addition to those actually sampled and graphed, gives some indication that the succession is directional. There appeared to be little change in diversity and cover beyond the boulder line. Didymodon spp Draba corymbosa Poa arctica Draba subcapitata Salix arctica Dryas integrifolia Poa abbreviata Psilopilum cavifolium Saxifraga oppositifolia Bryum spp - -Funaria-Dicranella spp - -Polytrichum-Pogonatum spp - -Pohlia spp - -20 30 40 50 Distance from glacier snout (m) 60 122 Appendix 5 Vegetation survey on a glacier foreland (79°40'N; 74°55'W), Dobbin-Allman Bay Valley, Ellesmere Island The scantily vegetated valley bounded by Dobbin Bay to the north and Allman Bay to the south on Ellesmere Island supports several glaciers. One of these glaciers located on the west side of the valley approximately equidistant from the two subtending bays, was selected as a study site (Figure 1.1). The site was characterized by relatively homogeneous limestone substrate and kame deposits. A boulder trim-line was noted approximately 200 m from the glacier snout, which likely represents the maximum glacial advance of the Little Ice Age (ca 1850 A.D.) . There was no sign of relict vegetation or preserved organic soil. Species presence/absence data were recorded within approximate 2 m x 2 m areas every 10 m for ten 200 m transects, leading away from the glacier snout. Approximately 100 m separated each of the ten transects. The graph below shows the species presence for the ten transects combined. Moss species were evident only upon close examination and were the sole life form for the first 175 m, where scant numbers of vascular plants began to colonize. Although extremely sparse, Salix arctica, Papaver radicatum, Saxifraga oppositifolia and Saxifraga caespitosa were observed at and beyond the boulder line (-200 m). The plants were widely scattered, even beyond the boulder margin, indicating a sluggish, if not stagnant, succession without species replacement. A localized Puccinellia-Nostoc community was also noted approximately 500 m from the ice edge. Saxifraga oppositifolia Draba spp Puccinellia spp Tortella fragilis Encalyptra procera Distichium capillaceum Desmatodon leucostoma Pohlia spp 50 100 150 Distance from glacier snout (m) 200 123 Appendix 6 Vegetation survey on a glacier foreland (79°37'N; 75°15'W), Franklin Pierce Bay, Ellesmere Island Weathered, limestone rocks characterized the substrate on a glacier foreland north of Franklin Pierce Bay, Ellesmere Island (Figure 1.1). There was a substantial change in elevation (-150 m) from the north to the south end of the glacier terminus. A margin of boulders extending approximately 230 m from the ice edge was regarded as the maximum extent of the glacial advance of the Little Ice Age. Beyond this margin, Salix arctica, Dryas integrifolia and Saxifraga oppositifolia were the most abundant vascular plants, though widely scattered. Although paleo-soil was not noted, unrooted scant remnants of relict Salix arctica, Dryas integrifolia, Saxifraga oppositifolia and a moss species were noted within 50 m of the glacial snout. Species presence/absence data were recorded in approximate 2 m x 2 m areas every 10 m along three 230 m transects, leading away from the glacial front. Approximately 300 m separated each transect. The combined results of the three transects are shown in the graph below. Although the succession slowly expanded in a directional manner from a few mosses to a scant Salix-Dryas-Saxifraga community, species replacement was not directly evident. Dryas integrifolia Papaver radicatum Salix arctica Draba subcapitata Cerastium alpinum Didymodon spp -j-Encalyptra procera Saxifraga oppositifolia Puccinellia Andersonii Minuartia rubella Lophozia spp (liverwort) Draba corymbosa Pohlia spp Desmatodon leucostoma Funaria groutiana 50 100 150 200 Distance from glacier snout (m) 250 124 Appendix 7 Vegetation survey on the Tear Drop Glacier* foreland (79°10'N; 79°45'W), Sverdrup Pass, Ellesmere Island Sampling took place on the Tear Drop Glacier foreland located at the drainage divide of Sverdrup Pass, a valley subtended by Flagler Bay to the east and Irene Bay to the west on central Ellesmere Island (Figure 1.1). On this gneiss-granite based foreland, a marked black algal growth began about 15 m from the glacier terminus, possibly due to the extensive amount of available meltwater draining from the glacier. At approximately 170 m from the ice edge, a boulder margin was noted which was believed to delineate the maximum glacial extent during the Little Ice Age. Relict vegetation was noted within the first 10 m of the glacier snout which consisted of Dryas integrifolia, Cassiope tetragona, Saxifraga oppositifolia, Saxifraga tricuspidata, Luzula spp and various mosses. Paleo-soil was present, but very infrequent. Species presence/absence data were recorded using 50 cm x 50 cm quadrats along three 145 m transects leading away from the glacier snout. Quadrats were sampled every 3 m for the first 30m, every 5 m for the next 25 m, every 10 m for the following 30 m and every 20 m for the last 60 m. The transects were separated by intervals of approximately 20 m. The graph below shows the species presence for the three transects combined. The pioneering community was dominated by both vascular plants and bryophytes, while the intermediate and later stages were largely made up of vascular plants only. Beyond the boulder trimline was a community dominated^by Carex aquatilis with Carex atrofusca, Alopecurus alpinus, Eriophorum angustifolium, E. Scheuchzeri, Salix arctica, Dryas integrifolia and Saxifraga oppositifolia. In addition to the diverse list of species shown in the graph, the following three were also observed on the Tear Drop Glacier foreland: Silene acaulis, Saxifraga rivularis and Saxifraga flagellaris. The trend from the graph, as well as, personal observations lead to the conclusion that the succession on this glacier foreland was directional with species replacement. *not an official name. 125 Pedicularis hirsuta Eriophorum angustifolium -J-Aulacomnium turgidum \\ Saxifraga caespitosa Meesia uliginosa Carex aquatilis Polygonum viviparum Carex nardina Lesquerella arctica j-Orthothecium chryseum Cardamine bellidifolia Epilobium latifolium Bryoerythrophylla spp 4-Bartramia ithyphylla f Oxyria digyna Dryas integrifolia Saxifraga tricuspidata -J-Potentilla hyparctica Carex misandra Melandrium spp -[• Luzula arctica Minuartia rubella • • Draba glabella •• Draba subcapitata • • Poa arctica • • Festuca brachyphylla • • Saxifraga foliolosa • • Alopecurus alpinus • • Draba lactea • • Cerastium alpinum • -Saxifraga nivalis • • Stellaria longipes • • Draba corymbosa • -Poa abbreviata • -Draba spp • • Poa spp • • Funaria-Dicranella spp Philonotis tomentella Braya purpurascens - -Salix arctica • -Papaver radicatum • -Saxifraga oppositifolia • -Bryum spp • -Draba oblongata • • Polytrichum-Pogonatum spp • -Luzula confusa Racomitrium panschii Puccinellia Andersonii - -Psilopilum cavifolium - -Pohlia spp • -Saxifraga cernua 0 20 40 60 80 100 120 140 160 Distance from glacier snout (m) 126 A p p e n d i x 8 Appendix Table 8.1: Summary of paired-sample t tests, testing for a substrate (paleo-soil versus glacio-fluvial sediment) effect on germinable seed bank density. Luzula confusa data were logio-transformed before analysis. Significance level, a = 0.05. A l l species Luzula confusa n 18 18 mean difference -45.722 10.444 S D difference 266.302 110.651 t -0.728 0.401 df 17 17 p-value 0.476 0.889 Appendix Table 8.2: Summary of Kruskall Wallis one-way analysis of variance to test for an age effect on the germinable seed bank for a) all species and b) Luzula confusa. Significance level, a = 0.05. a) A l l species b) Luzula confusa Group Terrain age (y) n Rank sum Group Terrain age (y) n Rank sum 1 0.25 10 395 1 0.25 10 510 2 1 10 444 2 1 10 510 3 2 10 463.5 3 2 10 587 4 3 10 602.5 4 3 10 758.5 .5 4 10 683 5 4 10 739 6 6 10 552.5 6 6 10 605 7 7 10 921 7 7 10 1067 8 8 10 1247 8 8 10 1386 9 9 10 1323 9 9 10 1461.5 10 11 10 1261 10 11 10 1387.5 11 13 10 1735 11 13 10 1674 12 16 10 1346 12 16 10 1456 13 19 10 1608 13 19 10 1605.5 14 21 10 1438 14 21 10 1410.5 15 24 10 1504 15 24 10 1450 16 26 10 1685 16 26 10 1446 17 29 . 10 1763 17 29 10 1813.5 . 18 31 10 1590 18 31 10 1469.5 19 34 10 1925 19 34 10 1903 20 35 10 1319.5 20 35 10\" 1306 21 37 10 1400.5 21 37 10 1301 22 39 10 1518 22 39 10 1443 23 41 8 1028.5 23 41 8 898 24 42 8 1054.5 24 42 8 909.5 25 43 6 1007.5 25 43 6 960 26 43 6 1059 26 43 6 819 K-W-stat. = 104.633; p = 0.000; df = 25 K - W stat. = 82.637; p = 0.000; df = 25 127 Appendix 9 Appendix Table 9.1: Summary of Kruskall Wallis one-way analysis of variance to test for an age effect on the fall-winter germinable seed rain for Luzula confusa. Significance level, a = 0.05. Group Distance interval (m) Terrain age (y) n Rank sum 1 0-20 0.25-1 15 2353.5 • 2 20-40 2-3 15 1253 3 40-60 4-6 11 344.5 4 60-80 7-8 12 1207.5 5 80-100 9-12 13 1203 6 100-120 13-16 14 1660.5 7 120-140 19-21 13 1446.5 8 140-160 24-26 14 1838.5 9 160-180 29-31 13 1940.5 10 180-200 34-35 12 1467.5 11 200-220 37-39 15 1732.5 12 220-240 41-42 15 1094.5 13 240-260 43 15 980.5 14 260-280 43+ 19 1377.5 K-W statistic = 56.539; p = 0.000; df = 13 128 "@en ; edm:hasType "Thesis/Dissertation"@en ; vivo:dateIssued "1997-05"@en ; edm:isShownAt "10.14288/1.0099147"@en ; dcterms:language "eng"@en ; ns0:degreeDiscipline "Geography"@en ; edm:provider "Vancouver : University of British Columbia Library"@en ; dcterms:publisher "University of British Columbia"@en ; dcterms:rights "For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use."@en ; ns0:scholarLevel "Graduate"@en ; dcterms:title "Primary plant succession on the Twin Glacier foreland, Alexandra Fjord, Ellesmere Island, Canadian high Arctic"@en ; dcterms:type "Text"@en ; ns0:identifierURI "http://hdl.handle.net/2429/5817"@en .