Spatial variables were based on a linear combination of eigenvectors derived by principal coordinates of neighbour matrices (PCNM)Īppendix S6. The four partitions show the adjusted R 2 values for environmental variables, shared variables between environmental and spatial variables, spatial variables, and unexplained variation based on the distance-based redundancy analysis. β-deviation is the standard effect size of pair-wise community dissimilarity (Bray–Curtis dissimilarity). Environmental and spatial influences on β-deviation of all, forest- and tundra-specific, and shared species for each transect in tundra and forest habitats. A linear regression trend line is shown, with significance assessed by the Mantel test ( p < .05), for visual interpretationĪppendix S5. β-deviation is the standard effect of pair-wise community dissimilarity (Bray–Curtis dissimilarity) among transects. Relationships between β-deviation and environmental or geographic distance among transects for all, shared, and forest- and tundra-specific species in tundra and forest habitats. Differences in β-diversity among quadrats, among transects, and among quadrats within each transect for all, forest- and tundra-specific, and shared species between forest and tundra habitatsĪppendix S4. Non-metric multidimensional scaling ordination diagram based on species abundance (coverage) in quadrats and transectsĪppendix S3. The Sørensen dissimilarity index (an index of dissimilarity between communities), Simpson dissimilarity index (an index of compositional turnover between communities), and the nestedness-resultant dissimilarity index are shownĪppendix S2. Compositional differences between forest and tundra communities. Jvs1281-AppendixS1-S7.docxWord document, 733.1 KBĪppendix S1. This positive feedback loop may be a driving process in the expansion of shared species and vegetation change. However, positive relationships in the abundance or occurrence of dominant shrubs may indicate a positive feedback loop between colonizing species and environmental modifications. ConclusionsĬommunity assemblages of initiators of vegetation change in tundra habitats are unrestricted by any external factors, such as dispersal limitation or environmental filtering. However, we found a positive co-occurrence pattern among the dominant shared shrubs to be a characteristic of the tundra habitat, and this relationship explained the patterns of community structure within tundra habitats. Therefore, community assemblages of shared species in tundra habitat were independent from local conditions determined by environment and spatial location. ResultsĪlthough environmental and spatial factors significantly affected the community structure of the specific and shared species in the forest habitat, the variation in community structure within and among tundra habitats was not explained by environmental or spatial factors for any community components. Interspecies interactions were estimated based on the patterns of co-occurrence between all pairs of species. To elucidate the assembly processes of the initiators of vegetation change and other community components of the forest–tundra ecotone, we evaluated β-diversity among and within habitats and its causal factors (i.e.
Vegetation was surveyed in eight 100-m line transects with 16 quadrats (1 m × 1 m) per transect established in both forest and tundra habitats. Kuujjuarapik/Whapmagoostui (55.31 N, 77.75 W), Quebec, Canada. We aimed to evaluate factors mediating the assembly processes for community initiators of vegetation change and determine the ecological processes driving vegetation changes. However, the community-level processes driving vegetation change are poorly understood. Changes in vegetation structure, including shrub expansion, occur in forest–tundra ecotones in sub-arctic regions.
0 kommentar(er)
