Ecological Research

, Volume 23, Issue 2, pp 363–370

Changes in the structure and heterogeneity of vegetation and microsite environments with the chronosequence of primary succession on a glacier foreland in Ellesmere Island, high arctic Canada

  • Akira S. Mori
  • Takashi Osono
  • Masaki Uchida
  • Hiroshi Kanda
Original Article

DOI: 10.1007/s11284-007-0388-6

Cite this article as:
Mori, A.S., Osono, T., Uchida, M. et al. Ecol Res (2008) 23: 363. doi:10.1007/s11284-007-0388-6

Abstract

Primary plant succession was investigated on a well-vegetated glacier foreland on Ellesmere Island in high arctic Canada. A field survey was carried out on four glacier moraines differing in time after deglaciation to assess vegetation development and microsite modification in the chronosequence of succession. The results showed evidence of directional succession without species replacement, which is atypical in the high arctic, reflecting the exceptionally long time vegetation development. During this successional process, Salix arctica dominated throughout all moraines. The population structures of S. arctica on these moraines implied the population growth of this species with progressing succession. The population density of S. arctica reflected the abundance of vascular plants, suggesting that development of the plant community might be related to structural changes and the growth of constituting populations. Through such growths of the population and the whole community with progressing succession, the spatial heterogeneity of vegetation gradually declines. Moreover, this vegetation homogenization is accompanied by changes in the spatial heterogeneity of microsite environments, suggesting significant plant effects on the modification of microsite environments. Accordingly, it was concluded that the directional primary succession observed on this glacier foreland is characterized by the initial sporadic colonization of plants, subsequent population growths, and the community assembly of vascular plants, accompanied by microsite modification.

Keywords

Community assembly Microsite modification Polar oasis Population growth Salix arctica 

Introduction

The extreme high arctic environment is characterized by low temperatures, a short growing season, nutrient-poor soil, and low precipitation, all of which contribute to quite low net production and sparse ground cover (e.g., Chapin 1987; Svoboda and Henry 1987; Jones and Henry 2003). It is, therefore, regarded as the marginal zone for plant growth and survival. The high arctic is dominated by polar deserts characterized by vast expanses of barren or sparsely vegetated terrain with no significant land animal populations or by icefields (Freedman et al. 1994). However, areas of well-developed vegetation are occasionally observed in specific sites, and are sometimes referred to as “polar oases” (Muc et al. 1994). Although such sites occupy a quite limited area within the high arctic landmass, they provide a critical habitat for the biota (Bliss and Gold 1994), profoundly emphasizing their importance within this harsh environment. However, despite the fundamental importance of understanding the development and function of arctic ecosystems, primary plant succession following deglaciation remains largely undocumented in the high arctic (Bliss and Gold 1994; Hodkinson et al. 2003), including polar oases (Jones and Henry 2003). This has largely resulted from the fact that steady vegetation development is seldom observed here due to the rarity and slow rate of vegetation change (Jones and Henry 2003). Thus, the clarification of patterns and mechanisms of primary succession in the infrequent but important well-vegetated areas of the high arctic is of paramount importance.

The compositional characteristics of plant species and their shifts during the chronosequence or time series of ecosystem development provide basic information on plant community assembly. However, despite the fact that ecosystem development on newly formed or exposed substrates on glacier forelands has been investigated in various climatic regions (Matthews and Whittaker 1987; Bliss and Gold 1994; Chapin et al. 1994; Jumpponen et al. 1999; Hodkinson et al. 2003), little attention has been paid to plant population structures and their time-dependent changes as principal components of primary succession following deglaciation (Matthews 1992; Crouch 1993). This is especially true with regard to high arctic environments, thus, reflecting our limited knowledge on arctic ecosystems. Crouch (1993) pointed out the necessity for a population ecology approach in addition to analysis of the whole plant community when studying primary succession on glacier forelands. It is, therefore, important to elucidate the structural changes in plant populations during the sequence of community assembly on well-vegetated glacier forelands in the high arctic, since abundant plants are found there compared to much more extensive harsh high arctic areas, such as polar deserts.

The initial stage of primary succession is principally characterized by the sporadic appearance of plants. Subsequently, plant abundance increases, leading to population development and, consequently, a more structured community. As a result, vegetation structure gradually changes from being highly heterogeneous to more homogeneous (del Moral and Jones 2002). Heterogeneity is known to be a ubiquitous feature of ecosystems (Chen et al. 2002), playing a central role in the theoretical framework of ecology (Legendre and Fortin 1989). The most distinct aspect of heterogeneity is its spatiality, reflecting the intrinsic characteristics of ecological processes (Bellehumeur and Legendre 1998). Several studies have suggested that an understanding of the spatial heterogeneity of vegetation and shifts with time can help clarify the mechanisms of plant succession (Dlugosch and del Moral 1999; del Moral and Jones 2002). However, despite its overwhelming significance, little is known about the heterogeneity in primary succession systems within the high arctic. Moreover, during succession on deglaciated forelands, vegetation development is accompanied by microsite modification (Matthews 1992); therefore, the heterogeneity of microsite environments is also expected to change with time after glacier retreat. Accordingly, the quantification of spatial heterogeneity in terms of both plant and microsite environments and the detection of shifts during the chronosequence of succession should contribute further to clarification of the successional sequence following deglaciation.

On a well-vegetated glacier foreland in the lower stream area of Oobloyah Valley, Ellesmere Island (Okitsu et al. 2004; Mori et al. 2006), which is located beyond 80°N in latitude and is supposed to be one of the northernmost developed vegetation on the Earth, this study aims to clarify vegetation development and accompanying microsite modification during primary succession following deglaciation. Specifically, we describe (1) the process of the community assembly of vascular plants and (2) intrinsic population growth of a specific dominating vascular plant species (Salix arctica). S. arctica is a deciduous, dioecious dwarf shrub, which is one of the most common and abundant shrubs in the Canadian high arctic (Tolvanen et al. 2001). Among vascular plant species appearing on deglaciated moraines in this area, this species is expected to have a major role in ecosystem development, including a recently deglaciated, initial successional stage (Mori et al. 2006). Then, we discuss (3) spatial heterogeneity in terms of both vegetation and microsite environments and shifts with the chronosequence of succession on this precious vegetated terrain in the high arctic.

Materials and methods

Study site

The study area (80°50–52′N 82°49–51′W, WGS84) is located within the proglacial field of the southern front of Arklio Glacier in the Krieger Mountains near Oobloyah Bay, Ellesmere Island, Nunavut, Canada (Fig. 1). Although no climatic data was available for the study area, data was obtained from the weather station at Eureka (80°00′N 85°56′W), 130 km south of Oobloyah Bay, showing the extremely harsh climate. The monthly mean temperatures of the warmest and coldest months (July and February, respectively) were about 3.3 and −38.0°C, respectively. The annual mean temperature was about −19.7°C and the annual precipitation was about 64 mm. The study area was located near the mouth of Oobloyah Valley, which runs east to west, entering Oobloyah Bay from the west. This valley is bounded by the steep slopes and valley glaciers of the Krieger Mountains in the north and by the gentle slopes of Neil Peninsula in the south. The geography here is well adapted to the periglacial environment, implying a long development time (Barsch 1981). The geological features of the study area are described in King (1981) and Okitsu et al. (2004).
Fig. 1

The geographical location of the study area (80°50–52′N 82°49–51′W) in Oobloyah Bay, Ellesmere Island, Nunavut, Canada

Arklio Glacier has developed glacial moraines representing different developmental periods since the Last Glacial (King 1981). According to geomorphological observations (Hasegawa and Sawaguchi, unpublished) and relative dating from measurements of weathering rind thickness in angular rock fragments (Hasegawa 2003), combined with related preceding studies (Barsch 1981; King 1981), the developmental periods of these moraines have been estimated (Okitsu et al. 2004). Specifically, King (1981) obtained a total of 25 radiocarbon dates for this valley, covering from 44,000±2,500 to 730±60 years, and the present. These results suggest an ice-free condition of this glacier foreland during the Last Glacial, and, thereby, it seems that vegetation has developed over an exceptionally long time, approximately 25,000–35,000 years (Okitsu et al. 2004). However, the region of the northwest Queen Elizabeth Islands, where this study area is included, is generally considered to be covered by the Innuitian Ice Sheet during the Last Glacial (Atkinson and England 2004). Thus, the actual deglaciated time of this study area is still tentative.

This study focused on four glacial moraines with different establishment periods (Fig. 2), representing different terrain ages. The youngest, Moraine A, is located just below Arklio Glacier and consists of fresh sharp boulders. Its establishment was estimated as being during the Little Ice Age (400–250 years ago). Moraine B was established during the Holocene (3,300–2,400 years ago), while Moraine C was established during the last stage of the Last Glacial (15,000–8,000 years ago). The oldest moraine in the study area, Moraine D, was assumed to have originated in the Last Glacial (35,000–25,000 years ago). Based on the age order of the moraines, we can construct a chronosequence of primary succession following deglaciation. These datings are based on the previous paper of Okitsu et al. (2004). The maximum ice-free period of the study area can, therefore, be estimated as at least 25,000 years, possibly providing a precious opportunity to investigate fairly long time for ecosystem development. Due to such a quite long time of ice-free period, it is difficult to assume that the initial conditions of each moraine investigated was the same and comparative. However, it is worth noting that the order of establishment of these moraines is unchanging and older moraines show more developed successional status. Actually, in spite of its harsh, high-arctic location, this area shows precious highly vegetated physiognomy (Fig. 2) and retains a lot of vascular plant species (Table 1), reflecting the considerably long time vegetation development.
Fig. 2

An overview of the study area from the third mountain in the Krieger Mountain adjacent to Arklio Glacier

Table 1

Frequency occurrence (%) of vascular plants on each moraine (n=20)

Vascular plant species

Moraine

A

B

C

D

Epilobium latifolium

30

0

0

0

Salix arctica

20

70

60

100

Dryas integrifolia

5

85

100

100

Stellaria monantha

5

20

10

25

Cassiope tetragona

0

85

95

100

Saxifraga tricuspidata

0

85

0

0

Saxifraga oppositifolia

0

55

100

100

Luzula confusa

0

25

5

100

Saxifraga cernua

0

15

35

0

Papaver radicatum

0

10

10

55

Carex nardina

0

5

40

0

Carex spp.a

0

5

5

10

Saxifraga nivalis

0

5

5

0

Poa arctica

0

5

0

15

Minuartia rossi

0

5

0

5

Cerastium alpinum

0

5

0

0

Poa spp.a

0

0

0

30

Carex misandra

0

0

0

15

Cardamine bellidifolia

0

0

0

15

Oxyria digyna

0

0

0

15

Juncus biglumis

0

0

0

10

Luzula spp.a

0

0

0

10

Draba spp.a

0

0

0

10

Draba micropetala

0

0

0

5

Unknown spp.

0

0

0

5

a Several plant species cannot be identified to their species, although their genus can be identified

Field research

In July 2004, a 50-m line transect was established on each moraine; to minimize the effects of moraine topography, for example, slope, each transect was set along a moraine ridgeline. In each transect, 1×1 m quadrats were also established at 2.5-m intervals (20 quadrats in total per transect). Within each quadrat, we recorded the microsite percent cover (%) of vascular plants, vascular plants plus cryptogams, and rocks [rocks (≥2 mm) were distinguished from fine sediment (<2 mm), and boulders are also included as rocks here]. Percentage cover was assessed by eye. We then measured the longest diameter of the largest rock (cm) within each quadrat. In this measurement, we did not remove any litters and simply measured an unburied part of the largest rock. So, increasing vegetation cover leads to decreasing rock cover. The maximum accumulation depth of plant organic litter (cm) was also recorded by randomly measuring the litter depth five to ten times in each quadrat. Also, the species names of all vascular plants observed were recorded within each qudrat. Of all dominating vascular plants on the moraines in this glacier foreland, S. arctica is the only species that occurs on the youngest moraine (Mori et al. 2006), which is sparsely vegetated, continuing to dominate even on older moraines (Okitsu et al. 2004), which might have been formed during the Last Glacial. Therefore, we focused on S. arctica as a representative species to infer successional time-dependent changes in the compositional characteristics of vascular plant populations. Within each quadrat, we recorded the diameter at ground level (mm) and the life stage of all S. arctica individuals; life stage was determined by observing reproductive status and size. The sex of reproductive individuals was recorded based on observations of inflorescences. The life stage of non-reproductive individuals was determined from the development status of the stem/shoots. Small individuals were defined as those with tiny stem/shoots and those yet to prostrate, while large individuals had obvious prostrating stems. In identifying each individual of S. arctica, we tracked their prostrating stems developing within organic litters to confirm no connection with any other adjacent individuals.

Statistical analysis

Differences in microsite cover, and microsite (maximum rock size and maximum organic litter depth) and plant properties (species abundance of vascular plants and population density of S. arctica) among moraines were evaluated using the Scheffe’s test. In this analysis, the microsite and plant properties were ln (x+1) transformed to minimize the effect of extreme outliers and to approximate a normal distribution. The coefficient of variation (CV) was also calculated for each measured microsite and plant property to quantify the magnitude of heterogeneity within each moraine (Matthews 1992; Hirobe et al. 2001). The CV may not be a parameter to quantify the heterogeneity, but it is helpful to compare the within-moraine variation of each measured property among moraines because the number of quadrats (n=20) was common for all moraines. Changes in the variation with time can be regarded as indicative of time-dependent changes in the heterogeneity. Thus, we used the CV to simply express the changes in the microsite and plant heterogeneity.

For populations of S. arctica, the percentage of individuals belonging to each life-stage was compared among moraines using a chi-square test. The above statistical analyses were performed with SPSS software, version 10.0.5 (SPSS Inc., Chicago, IL).

Using data from the 20 quadrats within each transect, we assessed the relationship between microsite and plant properties. In this analysis, each property was again ln(x+1) transformed. Further, spatial autocorrelation is known to exist with respect to microsite properties, especially organic accumulation (Hirobe et al. 2001); therefore, since the variables do not satisfy the assumption of independence, correlation analysis could not be performed in a typical fashion. Fisher’s method of randomization allows a modified null hypothesis to be tested when observations are not independent (Mitchell-Olds 1987). In this study, null distributions of the correlation coefficient (r) were, therefore, generated by randomly assigning the observed properties to the quadrats using 5,000 permutations for each randomization test. The observed r value was then compared to the null distributions to determine the significance. The applied null hypothesis was that one property was independent of another property.

Results

Microsite cover on each moraine

The results of microsite coverage by plants and rocks in each investigated transect are shown in Fig. 3. Moraine A was characterized by quite less dominance by plants and high cover by rocks. Then, with increasing moraine age, microsite cover by plants increased. Vascular plant cover significantly increased with the moraine age, while the CV of these coverage values decreased with the age (Table 2). Similarly, vascular plants plus cryptogam cover significantly increased with the moraine age, while the CV decreased with the age (Table 2). In contrast, rock cover decreased with increasing the moraine age, while the CV increased with the age (Table 2).
Fig. 3

Coverage (%) by vascular plants, cryptogams plus vascular plants, and rocks measured within each quadrat of transects on the moraines. Each quadrat was set at 2.5-m intervals (20 quadrats in total per transect)

Table 2

Microsite cover within 1×1-m quadrats (n=20) on each moraine

Coverage

Moraine

A

B

C

D

Vascular plants (%)

 Mean

0.8a

37.0b

41.3b

66.0c

 CV

244

60

42

20

Vascular and cryptogamic plants (%)

 Mean

1.8a

78.5b

81.0b

98.0c

 CV

205

16

12

3

Rocks (%)

 Mean

82.5a

21.3b

16.5b

0.8c

 CV

12

61

71

244

Identical letters indicate no significant difference at P>0.05 based on Scheffe’s test

CV=coefficient of variation (%)

Microsite and plant properties on each moraine

The maximum rock size significantly decreased with increasing moraine age following deglaciation, while the CV of rock size tended to increase (Table 3). Moreover, the accumulation depth of plant organic litter significantly increased with the moraine age, while the CV of litter depth decreased from the highest value on Moraine A to the lowest value on Moraine D (Table 3).
Table 3

Microsite properties within 1×1-m quadrats (n=20) on each moraine

Quadrat properties

Moraine

A

B

C

D

Maximum rock size (cm)

 Mean

36.3a

23.0b

20.7b

9.8c

 CV

41

60

58

67

Maximum organic accumulation (cm)

 Mean

0.3a

4.5b

7.8c

17.1d

 CV

131

45

54

24

Species abundance of vascular plants (species/m2)

 Mean

0.6a

4.2b

4.7b

7.3c

 CV

113

48

23

24

Population density of S. arctica (no./m2)

 Mean

0.2a

3.8b

3.0b

10.1c

 CV

205

97

132

46

Identical letters indicate no significant difference at P>0.05 based on Scheffe’s test

CV=coefficient of variation (%)

Species abundance of vascular plants on each moraine significantly increased with the moraine age, while the CV of species abundance decreased with the age (Table 3). The number of vascular plant species found on this glacier foreland ranged from 4 on Moraine A to 19 on Moraine D (Table 1). Several vascular plant species were observed. Epilobium latifolium occurred only on the youngest moraine, Moraine A, where it was the dominant species. However, on Moraines B, C, and D, Dryas integrifolia, Cassiope tetragona, S. arctica, and Saxifraga oppositifolia were the dominant vascular plant species; the former two were especially dominant on all three moraines. S. tricuspidata was also abundant on Moraine B, but was not found on the older moraines. Luzula confusa appeared on Moraines B, C, and D, showing an especially sharp increase in frequency on Moraine D. In general, the frequency of all vascular plant species was higher in older moraines, and there was no obvious evidence of species replacement in the chronosequence of succession.

The population density of S. arctica likewise increased with the moraine age (Table 3). In all moraines except Moraine A, S. arctica populations showed an inverse J-shaped diameter distribution (Fig. 4). Table 4 shows the within-moraine percentages of S. arctica based on life stage. Although the population density of S. arctica increased with moraine age, the percentage did not differ among Moraines B, C, and D. Furthermore, the percentage of non-reproductive and reproductive individuals did not differ among these three moraines (χ2=0.967, P>0.05).
Fig. 4

Frequency size [diameter at ground level (mm)] distributions of S. arctica on each moraine

Table 4

Proportion (%) of S. arctica individuals within each morainea

Life stage

Moraine

χ2

A (n=4)

B (n=76)

C (n=60)

D (n=202)

Non-reproductive individualsb

4.689NS

Small

0.0

59.2

60.0

52.5

Large

50.0

15.8

21.7

26.7

Reproductive individuals

Male

0.0

9.2

6.7

6.4

Female

50.0

15.8

11.7

14.4

a Differences in the proportion among moraines were tested using the chi-square test. Moraine A was excluded from this analysis, due to the small number of individuals here

NS=P>0.05

b Small individuals were defined as those with tiny stem/shoots and those yet to prostrate, while large individuals had obvious prostrating stems

A summary of the correlations among microsite environments and plant properties within each moraine is shown in Table 5. The maximum rock size was independent of the maximum litter depth on all moraines. The maximum rock size was also unrelated to plant properties, except species abundance on Moraine A, while the maximum litter depth was related to plant properties. Organic litter depth significantly increased with the species abundance of vascular plants on Moraines A and B, and with the population density of S. arctica on all moraines. Among the plant properties examined, the species abundance of vascular plants was significantly positively correlated with the density of S. arctica on all moraines except Moraine D.
Table 5

Correlation matrices between microsite and plant properties

Moraine

Quadrat properties

Maximum organic accumulation (cm)

Species abundance of vascular plants (species/m2)

Population density of S. arctica (no./m2)

A

Maximum rock size (cm)

−0.118NS

−0.493*

−0.344NS

Maximum organic accumulation (cm)

0.649**

0.587**

Species abundance of vascular plants (species/m2)

0.630***

B

Maximum rock size (cm)

−0.184NS

0.032NS

−0.071NS

Maximum organic accumulation (cm)

0.652**

0.632**

Species abundance of vascular plants (species/m2)

0.725***

C

Maximum rock size (cm)

−0.418+

−0.371NS

−0.272NS

Maximum organic accumulation (cm)

0.207NS

0.467*

Species abundance of vascular plants (species/m2)

0.621**

D

Maximum rock size (cm)

0.114NS

−0.249NS

−0.200NS

Maximum organic accumulation (cm)

0.045NS

0.430*

Species abundance of vascular plants (species/m2)

0.359NS

Values indicate correlation coefficients (r)

Significance levels were based on the randomization test. Significance levels: ***P<0.001; **P<0.01;  *P<0.05, +P<0.1, NS P>0.1

Discussion

Community assembly

Among the overwhelming presence of polar deserts in the high arctic Ellesmere Island (Muc et al. 1994), this glacier foreland, in spite of its high latitudinal location beyond 80°N, exhibits an atypical vegetation development. The present study provides evidence of a directional succession pattern without species replacement (Svoboda and Henry 1987), since plant cover (Table 2), species abundance (Table 3), and the frequency of each vascular plant species (Table 1) increased with time after glacier retreat, and there was little evidence of species replacement in the successional chronosequence. In polar deserts dominating in the high arctic, because of overwhelming physical stresses and harsh climate, the typical succession pattern is generally expressed as “non-directional, non-replacement (of species) succession”, which means a very few limited species succeed to survive and to keep their status while fluctuating in cover, abundance, and productivity (Svoboda and Henry 1987). Therefore, directional succession is quite unlikely in the high arctic ecosystem. However, Okitsu et al. (2004) suggested that the vegetation physiognomy of this glacier foreland is gradually approaching a C. tetragona-dominated dwarf shrub heath, which is similar to those principally observed in more southern regions than Ellesmere Island (Bliss 1997). The present results confirmed the dominance of this heath (Table 1). This glacier foreland has never been covered completely by an ice sheet (King 1981), thus, making it one of the postulated refuges for flora in central Ellesmere Island during the Last Glacial (Svoboda and Henry 1987; Pielou 1991). Vegetation in this area could have, thus, developed over an exceptionally long time. Such specific glaciological characteristics may contribute to the development of the southern type of vegetation physiognomy here.

In the high arctic, plant establishment is often strongly restricted by dominating abiotic factors and, thus, the limited availability of safe sites (Lévesque 2001; Cooper et al. 2004; Mori et al. 2006). Actually, the initial stage of succession, Moraine A, is characterized by high cover by large rocks/boulders (Tables 2 and 3), which means limited space for plant colonization (Matthews 1992). Plant cover was, thus, quite low on this youngest moraine (Table 2; Fig. 3), reflecting unfavorable conditions such as substrate instability and poor nutrient conditions (Matthews 1992; Mori et al. 2006). However, the directional succession usually occurs under conditions where the sum of biological driving forces, such as facilitation and competition (e.g., Jones and Henry 2003), is equal to or more than the environmental resistance (Svoboda and Henry 1987). This implies that, although initial plant colonization is strongly regulated by abiotic factors, biotic factors would gradually dominate in the subsequent fairly-long successional process.

Population growth

While pioneer E. latifolium appeared only on Moraine A, S. arctica was dominant even on the oldest moraine, Moraine D (Table 1). Generally, most vascular plants followed a similar trend, increasing in frequency with increasing moraine age (Table 1). S. arctica, therefore, provides an opportunity to elucidate the structure and growth of plant populations during the successional sequence; they are expected to be closely associated with the patterns of community assembly following deglaciation, as seen in the significant positive relationship between the density of this species and vascular species abundance in the chronosequence (Table 5). For S. arctica populations, size structures with numerous small individuals on each moraine (Fig. 4) are expected to be the result of constant juvenile recruitment, implying a relatively steady population structure (He and Duncan 2000; Mori and Takeda 2004). Furthermore, it is interesting to note that, in spite of increasing population density of S. arctica increasing with moraine age, the percentage of individuals according to life stage did not differ among moraines (Table 4). This further indicates stable compositional characteristics in the process of population growth of this species. Such structure and growth of constituting populations would contribute to community assembly on this glacier foreland.

Vegetation and microsite heterogeneity

Vegetation heterogeneity gradually declines with progressing succession (Dlugosch and del Moral 1999; del Moral and Jones 2002). Consistent with previous studies, the CVs obtained here (Tables 2 and 3) indicate that within-moraine variation in plant properties generally decreases with time. Because plant abundances are closely associated with microsite properties (Table 5), changes in vegetation as a source of litter supply is expected to have a significant effect on microsite development. Reflecting the vegetation homogenization, within-moraine variation in organic accumulation thus decreased in older moraines with more organic matter (Table 3). In contrast, the mean surface cover and exposed size of initially dominant sedimentary rocks decreased with moraine age, and, thereby, their within-moraine variation became larger (Tables 2 and 3), resulting from increasing vegetation cover and the resultant accumulation of organic matter. Although microsite modification such as soil formation on deglaciated moraines is known to be driven both by autogenic and allogenic processes, such as loess accumulation (Matthews 1992), the present study suggests that it is largely attributed to the development and homogenization of vegetation. Jones and Henry (2003) pointed out the importance of terrain age in explaining variation in species composition and abundance. Matthews (1992) also mentioned that variability in organic carbon is largely explained by moraine age. Similarly, in this glacier foreland, moraine age explains the variation in both vegetation and microsite properties at the landscape level, because of the gradual changes in heterogeneity of these two factors.

Due to dominating unfavorable factors for vegetation establishment and resultant sparse occurrence of plants (Dlugosch and del Moral 1999), litter input is spatially limited, especially on the recently formed Moraine A. The spatial distribution of primary production directly affects the spatial distribution of organic matter through litter input (Hirobe et al. 2001), leading to the positive correlation between organic litter accumulation and vascular plant abundance on this moraine (Table 5). This plant–microsite association can be further confirmed by focusing on the plant population; in spite of the quite sporadic occurrence of S. arctica on Moraine A, organic litter depth, which was also highly spatially heterogeneous at the early successional stage (Table 3; Matthews 1992), clearly reflected the abundance of S. arctica (Table 5). Because S. arctica showed high dominance from initial to later successional stages, this species is expected to have one of the significant roles in soil development in the long succession process. Accordingly, population growths of plants seemingly contribute to development of not only the plant community itself, but also that of microsite environments. Further, plant-induced soil modification resulting from the establishment of pioneer colonizers may facilitate subsequent plant invasion, consequently contributing to further vegetation development (Chapin et al. 1994). The presence of plant effects on the modification of microsite environments was, thus, confirmed at both the community and population levels.

Reflecting alternation between the heterogeneity and homogeneity (Dutilleul and Legendre 1993) of vegetation and of microsite environments, this glacier foreland has changed and developed over a long period of time. Further, periglacial processes are also important determinants of vegetation patterns (Cannone et al. 2004). Specifically, the older moraines have been affected by substrate disturbances such as geliturbation, but vegetation conspicuously develops and homogenizes on these terrains, further emphasizing the presence of biological forces increasing in the later successional stages. During the directional primary succession observed on the deglaciated moraines, heterogeneous retreating margin gradually develops, eventually becoming a relatively homogeneous C. tetragona-dominated physiognomy. This fairly long process is characterized by the initial colonization of plants, subsequent population growths, and the resultant community assembly, accompanied by microsite modification.

Acknowledgments

The authors thank Mr. Bob Howe, Dr. S. Iwasaki, and the members of the Polar Continental Shelf Project, Natural Resources, Canada, for their assistance in the logistics and field research. We also thank Dr. S. Kojima for the identification of vascular plants. Dr. K. Takahashi and two anonymous reviewers provided us with useful comments in revising the earlier manuscripts. This study was supported by a Grant-in-Aid for Priority Areas Research from the Japanese Ministry of Education, Culture, Sports, Science and Technology, Japan (grant no. 11208204).

Copyright information

© The Ecological Society of Japan 2007

Authors and Affiliations

  • Akira S. Mori
    • 1
    • 2
  • Takashi Osono
    • 1
  • Masaki Uchida
    • 3
  • Hiroshi Kanda
    • 3
  1. 1.Division of Environmental Science and Technology, Graduate School of AgricultureKyoto UniversityKyotoJapan
  2. 2.Forest Ecology Lab, School of Resource and Environmental ScienceSimon Fraser UniversityBurnabyCanada
  3. 3.National Institute of Polar ResearchTokyoJapan

Personalised recommendations