Environmental factors explaining the vegetation patterns in a temperate peatland

Although ombrotrophic temperate peatlands are important ecosystems for maintaining biodiversity in eastern North America, the environmental factors influencing their flora are only partly understood. The relationships between plant species distribution and environmental factors were thus studied within the oldest temperate peatland of Québec. Plant assemblages were identified by cluster analysis while CCA was used to related vegetation gradients to environmental factors. Five assemblages were identified; three typical of open bog and two characterized by more minerotrophic vegetation. Thicker peat deposit was encounter underlying the bog assemblages while higher water table level and percentage of free surface water distinguished the minerotrophic assemblages. Overall, the floristic patterns observed were spatially structured along the margins and the expanse. The most important environmental factors explaining this spatial gradient were the percentage of free surface water and the highest water-table level. To cite this article: S. Pellerin et al., C. R. Biologies 332 (2009).     

Notes on the Scottish flora

Background: Abernethy Forest, an area of native Caledonian Scots pine forest, includes areas of open or sparsely forested peatland, largely ombrotrophic, but locally influenced by water from mineral ground.

Aims: To relate vegetation species-composition to the limit of ombrotrophy inferred from water chemistry.

Methods: Species composition was listed in 300 sample quadrats; chemical analyses of water samples were available from 200 of these. The vegetation data were classified using TWINSPAN. Relationships between vegetation and chemical results were analysed statistically using contingency tables.

Results: Eleven ‘vegetation types’ were recognised, five (189 samples) judged as primarily ombrotrophic-bog vegetation. The others showed evidence of flushing, or transition to forest. A previous analysis of major cations in the 200 water samples indicated that a limiting Ca/Mg ratio of 1.0 divided 61 samples showing telluric influence from 139 essentially ombrotrophic samples. Neither division was sharp. Contingency tables showed that the division based on vegetation types was strongly associated with the division of the water samples based on cation ratios.

Conclusions: Large areas of bog south and east of Loch Garten are unequivocally ombrotrophic; the main telluric-influenced sites occupy openings in the forest west of the loch. The limited ‘fen’ flora probably reflects the small size and isolation of the calcium-rich sites and the dominating influence of the ombrotrophic peats.     

Peat and water chemistry at Big Run Bog,a peatland in the Appalachian mountains of West Virginia,USA

At Big Run Bog, aSphagnum-dominated peatland in the unglaciated Appalachian Plateau of West Virginia, significant spatial variation in the physical and chemical properties of the peat and in surface and subsurface (30 cm deep) water chemistry was characterized. The top 40 cm of organic peat at Big Run Bog had average values for bulk density of 0.09 g · cm^–3, organic matter concentration of 77%, and volumetric water content of 88%. Changes in physical and chemical properties within the peat column as a function of depth contributed to different patterns of seasonal variation in the chemistry of surface and subsurface waters. Seasonal variation in water chemistry was related to temporal changes in plant uptake, organic matter decomposition and element mineralization, and to varying redox conditions associated with fluctuating water table levels. On the average, total Ca, Mg, and N concentrations in Big Run Bog peat were 33, 15, and 1050 mol · g^–1, respectively; exchangeable Ca and Mg concentrations were 45 and 14 eq · g^–1 , respectively. Surface water pH averaged 4.0 and Ca⁺⁺ concentrations were less than 50 eq · L^–1 . These chemical variables have all been used to distinguish bogs from fens. Physiographically, Big Run Bog is a minerotrophic fen because it receives inputs of water from the surrounding forested upland areas of its watershed. However, chemically, Big Run Bog is more similar to true ombrotrophic bogs than to minerotrophic fens.     

Extremely short diapause in rotifers and its fitness consequences

Peatland pools usually show a marked gradient regarding their minerotrophic (harder and less acidic waters) versus ombrotrophic (softer and more acidic waters) status, which appears to be crucial in structuring their planktonic food webs. In this study, we analyzed the effect of such habitat diversity in shaping bacterioplankton assemblages. The planktonic bacteria from five pools located in Rancho Hambre peat bog (Tierra del Fuego Island) were studied through morphological and cytometric approaches, over more than one seasonal cycle. The community was always dominated by small cocci with an average cell size of 0.27 × 0.36 µm. Bacterioplankton morphological structure and cytometric fingerprint were correlated (Mantel test: P < 0.001), both methods used to characterize bacterioplankton showed significant differences between minero- and ombrotrophic pools. Variation in bacterial assemblage structure was mainly explained by abiotic variables relevant in peatlands such as pH, total hardness, conductivity, concentration of inorganic nutrients, and concentration and quality of dissolved organic carbon. Notably, these relatively fast approaches detected similar landscape-driven ecological patterns as previous high-throughput sequencing molecular studies of prokaryotes from the same pools, promising to be useful screening tools for limnological surveys as well as for monitoring the response of bacterial assemblages to environmental changes in peatland ecosystems.     

COMPARISON OF INTRASPECIFIC VARIATIONS IN THE REPRODUCTION AND PHOTOSYNTHESIS OF AN UNDERSTORY HERB,ARNICA CORDIFOLIA

Intraspecific variations in the reproduction of individual ramets and monospecific patches of the understory herb Arnica cordifolia Hook. were compared with variations in photosynthesis and understory light conditions. Ramets and patches were compared from three microhabitats (open, intermediate, and shaded) that differed in daily integrated irradiance. Individual ramets from open microhabitats (> 12 MJ m⁻² d⁻¹) had 23% more total dry wt and produced twice as many seeds, when compared to ramets from shaded locations (< 5 MJ m⁻² d⁻¹). In addition, monospecific patches from open locations were 63% more dense, and estimates of seed and vegetative patch reproductive effort were 4 and 2 times greater, respectively, when compared to shaded plant patches. For all measurements, ramets and patches from intermediate understory locations (6–10 MJ m⁻² d⁻¹) were intermediate in reproductive capacity between those of open and shaded locations. In addition, A. cordifolia seeds from open microhabitats germinated significantly better (45%) than either intermediate or shaded location seeds under high light and only seeds from shaded microhabitats germinated (14%) in the dark. Compared to shaded location plants, the greater total dry weight and seed production of individual ramets and the greater estimated reproductive effort of patches from open locations corresponded to a greater maximum photosynthetic rate (16.9 μmoles m⁻² s⁻¹) and daily carbon gain (12.2 g m⁻² d⁻¹). Possibly, a greater photosynthetic capacity may make more photosynthetic resources available for reproduction by A. cordifolia plants in open locations. Thus, intraspecific variation in physiology may contribute to intraspecific variation in reproduction.     

Abundance, biomass, structure, and activity of the microbial complexes of minerotrophic and ombrotrophic peatlands

Very large microbial biomass was revealed in peat bogs by means of fluorescence microscopy. In ombrotrophic peatlands, the pool of the dry-weight microbial biomass in the 1.5-m layer constituted 3-4 t/ha and was twice as high as in the minerotrophic peat bogs. Fungal biomass was predominant (55-99%) in ombrotrophic peatlands, while bacterial biomass predominated in minerotrophic peatlands (55-86%). In ombrotrophic peatlands, the microbial biomass was concentrated in the upper layers, while in minerotrophic peatlands, it was uniformly distributed in the bulk. After drainage, the microbial pool in the ombrotrophic peatlands increased twofold; that in the minerotrophic peatlands remained at the same level. The potential activity of nitrogen fixation and denitrification was revealed across the whole profile of the peatlands. The average values of these potential activities were five times higher in the minerotrophic peatlands, where bacterial biomass predominated.     

Long-term changes in spatial patterns of emergent vegetation in a Mediterranean floodplain: natural versus anthropogenic constraints

This study describes the long-term changes of spatial patterns of cut-sedge (Cladium mariscus) and common reed (Phragmites australis) in the Spanish floodplain wetland Las Tablas de Daimiel. Using seven sets of aerial photographs, we determined changes in their spatial patterns (size of patches) between 1945 and 2001 that resulted from combinations of natural change and anthropogenic stress (irrigated agriculture, waste water discharge, fire, ploughing). Our approach consisted in using 1 ha cells as units of spatial resolution in principal coordinates of neighbour matrices and spatial correlograms to assess the spatial scale of interest and spatial patterns at (1) the whole wetland and (2) two areas of the wetland with locally contrasting biophysical settings and anthropogenic stress history. Results showed that vegetation spatial patterns were influenced by natural variability until the 1970s. Thereafter, anthropogenic perturbation became the main driver of vegetation change, especially in the lower part of the wetland where local impacts were stronger. Natural variability did not fragment cut-sedge patches, and the effects of biological traits were less important for its spatial pattern. By contrast, man-made change resulted in a marked cut-sedge cover decrease and patch fragmentation, and the importance of its biological traits on spatial patterns increased. The trends of reed spatial patterns were generally opposite to, but not as clear as those for, cut-sedge. The trends of spatial patterns were especially evident at very broad (3,500–10,000 m) and broad (1,000–3,400 m) spatial scales. Competition for space and abiotic factors (water quality and water depth) were not strong predictors of cover variability at the 1-ha scale, particularly in the lower area of the wetland. This suggests that other environmental variables need to be considered in spatially explicit modelling of vegetation spatial patterns in wetlands. Consideration of spatial hierarchies and species-specific ecological traits is paramount to the conservation of degraded wetlands.     

Spatial and Temporal Variability in Growing-Season Net Ecosystem Carbon Dioxide Exchange at a Large Peatland in Ontario,Canada

We measured net ecosystem exchange of carbon dioxide (CO2) (NEE) during wet and dry summers (2000 and 2001) across a range of plant communities at Mer Bleue, a large peatland near Ottawa, southern Ontario, Canada. Wetland types included ombrotrophic bog hummocks and hollows, mineral-poor fen, and beaver pond margins. NEE was significantly different among the sites in both years, but rates of gross photosynthesis did not vary spatially even though species composition at the sites was variable. Soil respiration rates were very different across sites and dominated interannual variability in summer NEE within sites. During the dry summer of 2001, net CO2 uptake was significantly smaller, and most locations switched from a net sink to a source of CO2 under a range of levels of photosynthetically active radiation (PAR). The wetter areas--poor fen and beaver pond margin--had the largest rates of CO2 uptake and smallest rates of respiratory loss during the dry summer. Communities dominated by ericaceous shrubs (bog sites) maintained similar rates of gross photosynthesis between years; by contrast, the sedge-dominated areas (fen sites) showed signs of early senescence under drought conditions. Water table position was the strongest control on respiration in the drier summer, whereas surface peat temperature explained most of the variability in the wetter summer. Q 10 temperature-respiration quotients averaged 1.6 to 2.2. The ratio between maximum photosynthesis and respiration ranged from 3.7:1 in the poor fen to 1.2:1 at some bog sites; it declined at all sites in the drier summer owing to greater respiration rates relative to photosynthesis in evergreen shrub sites and a change in both processes in sedge sites. Our ability to predict ecosystem responses to changing climate depends on a more complete understanding of the factors that control NEE across a range of peatland plant communities.     

Abundance,biomass, structure,and activity of the microbial complexes of minerotrophic and ombrotrophic peatlands

A very large microbial biomass was revealed in peat bogs by means of fluorescence microscopy. In ombrotrophic peatlands, the pool of the dry-weight microbial biomass in the 1.5-m layer constituted 3–4 t/ha and was twice as high as in the minerotrophic peat bogs. Fungal biomass was predominant (55–99%) in ombrotrophic peatlands, while bacterial biomass predominated (55–86%) in minerotrophic peatlands. In ombrotrophic peatlands, the microbial biomass was concentrated in the upper layers, while in minerotrophic peatlands, it was uniformly distributed in the bulk. After drainage, the microbial pool in the ombrotrophic peatlands increased twofold; that in the minerotrophic peatlands remained at the same level. The potential activity of nitrogen fixation and denitrification was revealed across the whole profile of the peatlands. The average values of these potential activities were five times higher in the minerotrophic peatlands, where bacterial biomass predominated.     

Woody plant assemblages in isolated forest patches in a semiarid agricultural matrix

Spatial and temporal isolation and environmental variability are important factors explaining variation in plant species composition. The effect of fragmentation and disturbance on woody plant species composition was studied using data from 32 remnant church forest patches in northern Ethiopia. The church forests are remnants of dry Afromontane forest, embedded in a matrix of intensively used crop and grazing lands. We used canonical correspondence analysis and partial canonical correspondence analysis to analyze the effects of fragmented and isolated forest-patch identity, environmental and spatial variables on woody plant species composition in different growth stages. The dominance of late successional species was higher at the adult growth stage than seedlings and saplings growth stages. In the adult stages, late successional species like Olea europaea subsp. cuspidate had high frequency of occurrence. Forest patch identity was more important in explaining woody plant assemblages than environmental and spatial variables. For all growth stages combined, environmental variables explained more of the explained total fraction of variation in species composition than spatial variables. Topographic variables best explained variations in species composition for saplings, adults and all growth stages combined, whereas the management regime was most important for seedlings species composition. Our results show that in a matrix of cultivated and grazing land, fragmented and isolated forest patches differ in woody plant species assemblages. Some species are widely distributed and occurred in many patches while other occurred only in one or a few forest patches. Thus, our results indicate that remnant forest patches are important for preserving rare plant species and therefore management practices should focus on minimizing disturbance to the church forests and if possible increase church forest patch size.     

Relationships between spatial configuration of tropical forest patches and woody plant diversity in northeastern Puerto Rico

The destruction and fragmentation of tropical forests are major sources of global biodiversity loss. A better understanding of anthropogenically altered landscapes and their relationships with species diversity and composition is needed in order to protect biodiversity in these environments. The spatial patterns of a landscape may control the ecological processes that shape species diversity and composition. However, there is little information about how plant diversity varies with the spatial configuration of forest patches especially in fragmented tropical habitats. The northeastern part of Puerto Rico provides the opportunity to study the relationships between species richness and composition of woody plants (shrubs and trees) and spatial variables [i.e., patch area and shape, patch isolation, connectivity, and distance to the Luquillo Experimental Forest (LEF)] in tropical forest patches that have regenerated from pasturelands. The spatial data were obtained from aerial color photographs from year 2000. Each photo interpretation was digitized into a GIS package, and 12 forest patches (24–34 years old) were selected within a study area of 28 km². The woody plant species composition of the patches was determined by a systematic floristic survey. The species diversity (Shannon index) and species richness of woody plants correlated positively with the area and the shape of the forest patch. Larger patches, and patches with more habitat edge or convolution, provided conditions for a higher diversity of woody plants. Moreover, the distance of the forest patches to the LEF, which is a source of propagules, correlated negatively with species richness. Plant species composition was also related to patch size and shape and distance to the LEF. These results indicate that there is a link between landscape structure and species diversity and composition and that patches that have similar area, shape, and distance to the LEF provide similar conditions for the existence of a particular plant community. In addition, forest patches that were closer together had more similarity in woody plant species composition than patches that were farther apart, suggesting that seed dispersal for some species is limited at the scale of 10 km.     

Change in spatial distribution patterns of a biennial plant between growth stages and generations in a patchy habitat

BACKGROUND AND AIMS: The aim of the study was to evaluate factors causing change in spatial distribution patterns of plants between growth stages and generations for a monocarpic biennial plant, Lysimachia rubida. It was assumed that habitat heterogeneity was a primary factor determining spatial patterns of plants, and a randomization procedure was developed for testing the null hypothesis that only spatial association with ground surface conditions determined spatial patterns of plants. METHODS: A 5-year demographic census was conducted on an open dry habitat that was heterogeneous with regard to the ground surface conditions. KEY RESULTS: There was significant habitat association in that plants at vegetative and reproductive stages were denser in areas with smaller gravel than with larger gravel. Point process analyses rejected the null hypothesis of the spatial association with ground surface conditions. CONCLUSIONS: The results suggest that other factors, such as patchy seed dispersal, secondary dispersal of the seeds and life-history variation at various spatial scales, also affected spatial patterns of individuals in a population of L. rubida. Spatial structures and dynamics of a local population in a patchy habitat represent various performances of plants within patches and seed dispersal within a patch and beyond the patch.     

Within‐stand spatial structure and relation of boreal canopy and understorey vegetation

Question: How do spatial patterns and associations of canopy and understorey vegetation vary with spatial scale along a gradient of canopy composition in boreal mixed‐wood forests, from younger Aspen stands dominated by Populus tremuloides and P. balsamifera to older Mixed and Conifer stands dominated by Picea glauca? Do canopy evergreen conifers and broad‐leaved deciduous trees differ in their spatial relationships with understorey vegetation? Location: EMEND experimental site, Alberta, Canada. Methods: Canopy and understorey vegetation were sampled in 28 transects of 100 contiguous 0.5 m × 0.5 m quadrats in three forest stand types. Vegetation spatial patterns and relationships were analysed using wavelets. Results: Boreal mixed‐wood canopy and understorey vegetation are patchily distributed at a range of small spatial scales. The scale of canopy and understorey spatial patterns generally increased with increasing conifer presence in the canopy. Associations between canopy and understorey were highly variable among stand types, transects and spatial scales. Understorey vascular plant cover was generally positively associated with canopy deciduous tree cover and negatively associated with canopy conifer tree cover at spatial scales from 5–15 m. Understorey non‐vascular plant cover and community composition were more variable in their relationships with canopy cover, showing both positive and negative associations at a range of spatial scales. Conclusions: The spatial structure and relation of boreal mixed‐wood canopy and understorey vegetation varied with spatial scale. Differences in understorey spatial structure among stand types were consistent with a nucleation model of patch dynamics during succession in boreal mixed‐wood forests.     

Scale specific determinants of tree diversity in an old growth temperate forest in China

The major processes generating pattern in plant community composition depend upon the spatial scale and resolution of observation, therefore understanding the role played by spatial scale on species patterns is of major concern. In this study, we investigate how well environmental (topography and soil variables) and spatial variables explain variation in species composition in a 25-ha temperate forest in northeastern China. We used new variation partitioning approaches to discover the spatial scale (using multi-scale spatial PCNM variables) at which environmental heterogeneity and other spatially structured processes influence tree community composition. We also test the effect of changing grain of the study (i.e. quadrat size) on the variation partitioning results. Our results indicate that (1) species composition in the Changbai mixed broadleaf-conifer forest was controlled mainly by spatially structured soil variation at broad scales, while at finer scales most of the explained variation was of a spatial nature, pointing to the importance of biotic processes. (2) These results held at all sampling grains. However, reducing quadrat size progressively reduced both spatially and environmentally explained variance. This probably partly reflects increasing stochasticity in species abundances, and the smaller proportion of quadrats occupied by each species, when quadrat size is reduced. The results suggest that environmental heterogeneity (i.e. niche processes) and other biotic processes such as dispersal work together, but at different spatial scales, to build up diversity patterns.     

Vegetation of Sphagnum bogs in highly disturbed landscapes: relative influence of abiotic and anthropogenic factors

Question: Has the vegetation of Sphagnum bogs been affected by more than 200 years of human activities? Location: Bas‐Saint‐Laurent region, southeastern Québec, Canada. Methods: Data (species assemblages, abiotic and spatio‐historical variables) were collected in 16 bogs ranging from 2 to 189 ha, and incorporated in a geographical information system. Major gradients in vegetation composition were identified using DCA. CCA was used to relate vegetation gradients to abiotic and spatio‐historical variables. Results: A clear segregation of species assemblages was observed, from open and undisturbed bogs to forested and highly disturbed sites. Among abiotic factors, tree basal area, water table level and peat thickness had a significant influence on plant species composition. Among spatio‐historical factors, disturbance level, area loss and fire were the most influential factors. Variance partitioning between these groups of factors suggests that spatio‐historical factors had a major influence on peatlands, representing 22% of the variation observed in the plant species assemblages while abiotic factors represent only 17% of the variation. Conclusions: The results highlight the influence of agricultural and other anthropogenic activities on plant assemblages and suggest that even wetlands apparently resistant to disturbances, such as peatlands, can be severely affected by anthropogenic factors. Plant species assemblages of ombrotrophic peatlands of the Bas‐Saint‐Laurent region were, and still are, largely influenced by human activities.     

Multivariate analysis of a fine-scale breeding bird atlas using a geographical information system and partial canonical correspondence analysis: environmental and spatial effects

Aim To assess the relative roles of environment and space in driving bird species distribution and to identify relevant drivers of bird assemblage composition, in the case of a fine‐scale bird atlas data set. Location The study was carried out in southern Belgium using grid cells of 1 × 1 km, based on the distribution maps of the Oiseaux nicheurs de Famenne: Atlas de Lesse et Lomme which contains abundance for 103 bird species. Methods Species found in < 10% or > 90% of the atlas cells were omitted from the bird data set for the analysis. Each cell was characterized by 59 landscape metrics, quantifying its composition and spatial patterns, using a Geographical Information System. Partial canonical correspondence analysis was used to partition the variance of bird species matrix into independent components: (a) ‘pure’ environmental variation, (b) spatially‐structured environmental variation, (c) ‘pure’ spatial variation and (d) unexplained, non‐spatial variation. Results The variance partitioning method shows that the selected landscape metrics explain 27.5% of the variation, whilst ‘pure’ spatial and spatially‐structured environmental variables explain only a weak percentage of the variation in the bird species matrix (2.5% and 4%, respectively). Avian community composition is primarily related to the degree of urbanization and the amount and composition of forested and open areas. These variables explain more than half of the variation for three species and over one‐third of the variation for 12 species. Main conclusions The results seem to indicate that the majority of explained variation in species assemblages is attributable to local environmental factors. At such a fine spatial resolution, however, the method does not seem to be appropriated for detecting and extracting the spatial variation of assemblages. Consequently, the large amount of unexplained variation is probably because of missing spatial structures and ‘noise’ in species abundance data. Furthermore, it is possible that other relevant environmental factors, that were not taken into account in this study and which may operate at different spatial scales, can drive bird assemblage structure. As a large proportion of ecological variation can be shared by environment and space, the applied partitioning method was found to be useful when analysing multispecific atlas data, but it needs improvement to factor out all‐scale spatial components of this variation (the source of ‘false correlation’) and to bring out the ‘pure’ environmental variation for ecological interpretation.     

Mesoscale distribution of understorey plants in temperate forest (Kalø, Denmark): the importance of environment and dispersal

We studied the influence of environmental heterogeneity and dispersalprocesses on mesoscale distribution patterns of understorey plants in a 330ha ancient, managed temperate forest area. Similarity matrixregression analysis showed that overall species similarity was affected bysite(the two disjunct halves of the forest area), spatial autocorrelation at the100m scale, edaphic conditions, stand type and stand structure, andthe occurrence of open areas, but not by stand age or topography. Usingautologistic regression we analysed the influence of clumping, site, edaphicconditions, stand composition, open areas, and stand structure on theindividualdistribution patterns of the 60 most abundant species. Only five species haddistributions that were not significantly related to any of the explanatoryvariables. A large number of species showed significant, and often differing,relationships to at least one of the environmental parameters. At least 20species exhibited clumping at the 100 m scale that was independentof the environmental parameters. Principal coordinate analysis and an analysisusing the Ellenberg ecological indicator values suggested that no importantgradients had been overlooked. Dispersal ability (estimated from dispersalmode)and clumping independent of environmental heterogeneity were related. Poordispersers exhibited stronger clumping at the 100 m scale thangooddispersers. Our results support the joint importance of environmentalheterogeneity and dispersal processes in shaping the mesoscale distributionpatterns of forest herbs. We conclude that the distribution of plant species,and species coexistence and species composition in plant communities, cannot beexplained without simultaneously considering environmental heterogeneity anddispersal.     

Scaling patterns of plankton diversity: a study of ciliates in a tropical coastal lagoon

Although spatial and temporal variation in plankton diversity is regularly investigated in surveys, experiments, and models, there is a lack of methods for predicting spatial patchiness of plankton diversity. We develop and apply a suite of geostatistical and multiple-regression analysis tools to assess ciliate diversity in a tropical coastal lagoon; these methods can predict spatial and temporal patterns of diversity, provide error estimates associated with these predictions, and assess which environmental factors may drive diversity patchiness. Geostatistical analysis was applied to H′ (Shannon diversity index) from 25 to 35 data collected from a sampling grid (40 × 40 m), and 10 dispersed lagoonal sites, in the dry and rainy seasons, on numerous occasions. Conditional simulation and kriging were used to predict diversity at the lagoonal and small scales, respectively. The relationship between diversity (H′) and the environment was examined by multiple-regression analysis. Thirty-six ciliate morphospecies occurred; H′ ranged from 0 to 1.9 and was patchy. Multiple regression indicated ciliate diversity changed seasonally, increasing when the sand bar was open and the lagoon was connected with the sea. Geostatistical analysis extended the recognition that seasonal changes alter diversity: when the rainy season produced a variable environment, relatively small scale patches and diversity gradients occurred; in the dry season, when the lagoon was physically uniform, larger and fewer diversity patches occurred; at the sublagoonal scale, diversity patches were similarly structured in the two seasons. Results indicate environmental variability and immigration can be the main drivers behind ciliate diversity in the lagoon. We recommend the patterns these data reveal and methods we employ be considered when further studies are continued to examine diversity on local and larger scales. Handling editor: S. Declerck     

Spatial ordination of vegetation data using a generalization of Wartenberg's multivariate spatial correlation

Question: Are there spatial structures in the composition of plant communities? Methods: Identification and measurement of spatial structures is a topic of great interest in plant ecology. Univariate measurements of spatial autocorrelation such as Moran's I and Geary's c are widely used, but extensions to the multivariate case (i.e. multi‐species) are rare. Here, we propose a multivariate spatial analysis based on Moran's I (MULTISPATI) by introducing a row‐sum standardized spatial weight matrix in the statistical triplet notation. This analysis, which is a generalization of Wartenberg's approach to multivariate spatial correlation, would imply a compromise between the relations among many variables (multivariate analysis) and their spatial structure (autocorrelation). MULTISPATI approach is very flexible and can handle various kinds of data (quantitative and/or qualitative data, contingency tables). A study is presented to illustrate the method using a spatial version of Correspondence Analysis. Location: Territoire d'Etude et d'Expérimentation de Trois‐Fontaines (eastern France). Results: Ordination of vegetation plots by this spatial analysis is quite robust with reference to rare species and highlights spatial patterns related to soil properties.     

Influence of edaphic factors on the spatial structure of inland halophytic communities: a case study in China

Abstract. In order to understand the influence of edaphic factors on the spatial structure of inland halophytic plant communities, a 2.6 km² study site, located on the lower fringe of the alluvial fan of the Hutubi River, in an arid region of China, was sampled and mapped. 105 patches were found to be homogeneous in species composition. Plant species and their coverage were recorded in each patch. 45 patches were randomly selected for the measurement of edaphic variables. A map with quadrat locations and boundaries of patches was digitized into a GIS and related to the vegetation and edaphic data matrices. CCA was used to evaluate the relative importance of edaphic factors in explaining the variation of the species assemblages and to identify the ecological preferences of species. The spatial structure of the communities and the main edaphic factors were analyzed using correlograms, Mantel correlograms and clustering under constraint of spatial contiguity. Gradient analysis showed that there are two distinct vegetation gradients in the study area, one of which is determined mainly by soil moisture (determined by depth to the water table), and the other by soil salinity (determined by electrical conductivity and hydrolytic alkalinity of the first soil layer). However, spatial analyses showed that at the sampling scale the halophytic communities in the study area are structured along one main spatial gradient determined by the water table level. Similar spatial autocorrelation structures between the factors related to the first soil layer and the communities, given our sampling scale, could not be detected. Our results suggest that the relative importance of the effects of different edaphic factors on the spatial structure of halophytic communities is scale-dependent. The partitioning of species variation indicates that in addition to edaphic factors, other factors, such as biotic interactions, may play an important role in structuring these communities.