Presence‐only versus presence–absence data in species composition determinant analyses

Aim Studying relationships between species and their physical environment requires species distribution data, ideally based on presence–absence (P–A) data derived from surveys. Such data are limited in their spatial extent. Presence‐only (P‐O) data are considered inappropriate for such analyses. Our aim was to evaluate whether such data may be used when considering a multitude of species over a large spatial extent, in order to analyse the relationships between environmental factors and species composition. Location The study was conducted in virtual space. However, geographic origin of the data used is the contiguous USA. Methods We created distribution maps for 50 virtual species based on actual environmental conditions in the study. Sampling locations were based on true observations from the Global Biodiversity Information Facility. We produced P–A data by selecting ∼1000 random locations and recorded the presence/absence of all species. We produced two P‐O data sets. Full P‐O set was produced by sampling the species in locations of true occurrences of species. Partial P‐O was a subset of full P‐O data set matching the size of the P–A data set. For each data set, we recorded the environmental variables at the same locations. We used CCA to evaluate the amount of variance in species composition explained by each variable. We evaluated the bias in the data set by calculating the deviation of average values of the environmental variables in sampled locations compared to the entire area. Results P–A and P‐O data sets were similar in terms of the amount of variance explained by the different environmental variables. We found sizable environmental and spatial bias in the P‐O data set, compared to the entire study area. Main conclusions Our results suggest that although P‐O data from collections contain bias, the multitude of species, and thus the relatively large amount of information in the data, allow the use of P‐O data for analysing environmental determinants of species composition.     

Reinvestigation on species richness and environmental correlates of bryophytes at a regional scale in China

Aims The aim of this article is 4-fold: (i) to update species richness of bryophytes for each of the Chinese provinces based on the most current knowledge on distributions of bryophytes in China, (ii) to provide a set of analyses based on the updated species richness data and the environmental variables used in a recent article on species richness of bryophytes in China, (iii) to expand the analysis presented in the recent article by relating species richness of bryophytes to over 15 additional climatic variables and (iv) to determine the degree to which the relationships between bryophyte species richness and environmental variables that were reported in the recent article might have been biased.Methods Over 180 literatures with national, provincial and local species lists of bryophytes in China were used in this study. Taxonomy and nomenclature of bryophytes in China were standardized according to The Plant List. Correlation and regression analyses were used to examine the relationships between species richness or species density of bryophytes in Chinese provinces and environmental variables.Important findings On average, each Chinese province possesses 700.6 species of bryophytes, which is 112.1 species more than previously reported. With the updated species richness data reported in this study, stronger relationships between species richness of bryophytes and environmental variables have been found, compared with those found in a recently published study for China. When single environmental variables were considered, precipitation-related variables were, on average, more strongly correlated with species richness and species density than were temperature-related variables. Environmental variables were on average correlated more strongly with species density than with species richness of bryophytes at the regional scale in China. Our study showed that measures quantifying the average and variation of environmental conditions within each Chinese province explained 82.7% and 71.1% of the variation in species richness of liverworts and mosses, respectively, and explained 86.5% and 70.7% of the variation in species density of liverworts and mosses, respectively.     

Reconstructing paleoecology and paleoenvironmental variables using factor analysis and regression: some limitations

We test the success of Principal Components, Factor and Regression Analysis at recovering environmental signals using numerical experiments in which we control species environmental responses, the environmental conditions and the sampling scheme used for calibration. We use two general conditions, one in which sampling of a continental margin for benthic foraminiferal assemblages is done in a standard grid and the driving environmental variables are correlated to one another, and the other where sampling is done so that the environmental variables are uncorrelated. The first condition mimics many studies in the literature. We find that where the controlling environmental variables are correlated, Principal Components/Factor Analysis yield factors that reflect the common variance (correlation) of those variables. Since this common variance is largely a product of the sampling scheme, the factors extracted do not reliably present true species ecologic behavior. This behavior cannot be accurately diagnosed and faulty interpretations may lead to substantial error when using factor coefficients to reconstruct conditions in the past. When the sampling scheme is constructed so that the controlling environmental variables for the calibration data set are uncorrelated the factor patterns will reflect these variables more accurately. Species responses can be more successfully interpreted from the Principal Components/Factor Analysis structure matrices. Additionally, regression analysis can successfully extract the independent environmental signals from the biotic data set. However, matrix closure is a confounding effect in all our numerical results as it distorts species' abundances and spatial distribution in the calibration data set. Our results show clearly that a knowledge of the controlling environmental variables, and the correlations among these variables over a study area, is essential for the successful application of multivariate techniques for paleoenvironmental reconstruction.     

Effects of species partition on explanatory variables in direct gradient analysis - a case study from Senegal

Abstract. Species-environment data from Senegal, West Africa, are used to study the effects of partition of a large species data set into subsets corresponding to rare and common species respectively. The original data set contains 129 woody plant species from 909 plots and 60 explanatory variables. By applying Canonical Correspondence Analysis to data subsets, marked differences in the forward-selected variables were detected. The highest resemblance was found between the complete species set and the common species subset. Only one of eight selected variables was common to all species and the rare species groups. These findings were tested with partial ordination, applying the selected variables from the original species group (Vb), as variables and covariables to the analyses of common and rare species. For the common species this application resulted in a constrained ordination with higher eigenvalues as compared to the set of variables selected with reference to the common species group. Using the rare species group, the application of Vb gave a much lower sum of eigenvalues than did the ordination with selected variables based on the rare species group only. Evidently, the set of variables selected on the basis of the rare species data were more significant. Hence, the resulting gradients depend on the frequency of the species. Gradient analysis is apparently only valid for groups of species with closely resembling characteristics. This implies that different functional types of species, with different distributions and abundances, respond individually to environmental variation. Extrapolating deduced gradients from one species group to another maybe risky, particularly when used in vegetation modelling.     

Partitioning the variation among spatial,temporal and environmental components in a multivariate data set

Abstract We propose a method of partitioning the variation in a multivariate set of data according to (i) environmental variables, (ii) variables describing the spatial structure in the data and (iii) temporal variables. This method is an extension of an existing method for partialling out the spatial component of environmental variation, using canonical analysis. Our proposed method extends this approach by including temporal variables in the analysis. Thus, the partitioning of variation for a data matrix of species’abundances or biomass can include, by our methodology, the following components: (1) pure environmental, (2) pure spatial, (3) pure temporal, (4) pure spatial component of environmental, (5) pure temporal component of environmental, (6) pure combined spatial/temporal component, (7) combined spatial/temporal component of environmental and (8) unexplained. In addition, permutation testing accompanying the analyses allows tests of significance for the relationship between the different components and the species data. We illustrate the method with a set of survey data of penaeid species (prawns) obtained on the far northern Great Barrier Reef, Australia. This extension is a useful tool for multivariate analysis of ecological data from surveys, where space, time and environment commonly overlap and are important influences on observed variation.     

Species richness in mammalian herbivores: patterns in the boreal zone

Latitudinal gradients in species diversity are well established for a number of plant and animal taxa. Both historical and present-day environmental factors have been suggested to be responsible for observed patterns. We tested the hypothesis that current environmental features of the environment (primary productivity and regional landscape structure) may explain the longitudinal variation in species richness of mammalian herbivores in the Holarctic boreal zone. Mammalian herbivore species diversity was strongly correlated with a number of environmental variables measured. We reduced the data set by a principal components analysis (PCA), and found that in the Palearctic, species richness is positively related to warm climate (high temperature sum), the number of hardwood species, and the area of boreal forest. In the Nearctic, species richness increases as the length of the growing season and the number of coniferous tree species increase. Thus indirect measures of primary productivity as well as tree species number may accurately predict species richness in mammalian herbivores. In addition, there seems to be a strong species-area effect at the regional level. The larger and more homogeneous in terms of forest coverage a boreal section is, the more species coexist there.     

Patterns of tree species richness in relation to environment in southeastern New South Wales,Australia

Abstract We present regression models of species richness for total tree species, two growth forms, rainforest trees (broadleaf evergreens) and eucalypts (sclerophylls), and two large subgenera of Eucalyptus. The correlative models are based on a data set of 166 tree species from 7208 plots in an area of southeastern New South Wales, Australia. Eight environmental variables are used to model the patterns of species richness, four continuous variables (mean annual temperature, rainfall, radiation and plot size), plus four categorical factors (topographic position, lithology, soil nutrient level and rainfall seasonality). Generalized linear modelling with curvilinear and interaction terms, is used to derive the models. Each model shows a significant and differing response to the environmental predictors. Maximum species richness of eucalypts occurs at high temperatures, and intermediate rainfall and radiation conditions on ridges with aseasonal rainfall and intermediate nutrient levels. Maximum richness of rainforest species occurs at high temperatures, intermediate rainfall and low radiation in gullies with summer rainfall and high nutrient levels. The eucalypt subgenera models differ in ways consistent with experimental studies of habitat preferences of the subgenera. Curvilinear and interaction terms are necessary for adequate modelling. Patterns of richness vary widely with taxonomic rank and growth form. Any theories of species diversity should be consistent with these correlative models. The models are consistent with an available energy hypothesis based on actual evapotranspiration. We conclude that studies of species richness patterns should include local (e.g. soil nutrients, topographic position) and regional (e.g. mean annual temperature, annual rainfall) environmental variables before invoking concepts such as niche saturation.     

Responses of benthic macroinvertebrate communities to altitude and geology in tributaries of the Sepik River (Papua New Guinea): the influence of taxonomic resolution on the detection of environmental gradients

1. Benthic macroinvertebrate communities were sampled in 30 tributary streams at altitudes from sea level to about 3000 m draining three geologically distinct regions within the catchment of the Sepik River, Papua New Guinea. The fauna of this near‐pristine river has been little studied, and the impacts of ongoing and anticipated human impacts on the Sepik are uncertain. 2. Data on community composition were analysed at different levels of taxonomic resolution (species or morphospecies versus family) to compare their responses to environmental variables such as altitude and geology (reflected in water chemistry), and to indicate their potential utility for the detection of environmental change. 3. A total of 183 000 macroinvertebrates representing 250 species were collected, predominantly insects (232 species and >99% of individuals). The fauna was co‐dominated by Diptera (42% of individuals; 32 morphospecies, mainly Orthocladiinae, Simuliidae and Chironominae) and Ephemeroptera (36%; 48 species), although the Trichoptera showed the highest species (67) and family (13) richness, with Coleoptera ranked third (43 species). Naucoridae (Heteroptera) and Crambidae: Acentropinae (Lepidoptera), each represented by 13 species, were distinctive faunal elements. Mayflies were represented by only four families, one consisting of a single species. 4. Multivariate analysis of the species‐level data set revealed that community composition was influenced by geological region, but the effect was largely due to altitude as most streams in one region (the Central Highlands) were at higher elevations (>800 m) than streams in the other two regions (<500 m). However, altitude had no direct effect on species richness. A secondary influence of current speed and a subsidiary effect of water chemistry (pH and N‐NO₃) on community composition were also detected. Naucorid bugs showed evidence of altitudinal zonation and some species replacement, plus a tendency for certain genera to be associated with highland or lowland streams. 5. Analysis of the family‐level data set failed to uncover strong effects of any environmental variable, either individually or in combination, although some sensitivity to altitude plus slope was detected. 6. These findings suggest that attempts to use macroinvertebrates to detect environmental change in New Guinea streams will require species‐level monitoring of community composition.     

Prediction of occurrence of vascular plants in deciduous forests of South Sweden by means of Ellenberg indicator values

Abstract. In this study we present a new method for predicting the occurrences of species using data from deciduous forests in South Sweden. Complete species lists of vascular plants were compiled from 101 stands and from representative sample plots inside the stands. Soil samples from each stand were collected for determination of pH and nitrogen mineralization. Presence-absence data for species were fitted to the values of four environmental variables - soil moisture, soil reaction (pH), soil nitrogen and light - by means of Linear (Multiple) Logistic Regression (LLR), and Gaussian (Multiple) Logistic Regression (GLR). First, these values were estimated by calculating the weighted averages of Ellenberg indicator values. Second, the estimates for reaction and nitrogen were substituted by the real measurements of pH and mineralized NH₄⁺, keeping the Ellenberg estimates for light and moisture. The models were validated by an independent test data set. In general, the models had high predictive abilities. GLR fitted the species occurrences better to the environmental variables than LLR, but had a lower accuracy of prediction of species occurrence in the stands. The use of soil measurements instead of Ellenberg indicator values did not improve the predictive abilities of the models. The environmental conditions in the stand test set were successfully estimated by using species data from the plots. When using the species lists of the stands instead of plot data, a slightly better predictive ability was obtained. The collection of plot data, however, is easier and less time-consuming. The accuracy of prediction differed considerably between species.     

Community‐level vs species‐specific approaches to model selection

A topic of particular current interest is community‐level approaches to species distribution modelling (SDM), i.e. approaches that simultaneously analyse distributional data for multiple species. Previous studies have looked at the advantages of community‐level approaches for parameter estimation, but not for model selection – the process of choosing which model (and in particular, which subset of environmental variables) to fit to data. We compared the predictive performance of models using the same modelling method (generalised linear models) but choosing the subset of variables to include in the model either simultaneously across all species (community‐level model selection) or separately for each species (species‐specific model selection). Our results across two large presence/absence tree community datasets were inconclusive as to whether there was an overall difference in predictive performance between models fitted via species‐specific vs community‐level model selection. However, we found some evidence that a community approach was best suited to modelling rare species, and its performance decayed with increasing prevalence. That is, when data were sparse there was more opportunity for gains from “borrowing strength” across species via a community‐level approach. Interestingly, we also found that the community‐level approach tended to work better when the model selection problem was more difficult, and more reliably detected “noise” variables that should be excluded from the model.     

Biodiversity–productivity relationship in ponds: Community and metacommunity patterns along time and environmental gradients

Primary production correlates with diversity in various ways. These patterns may result from the interaction of various mechanisms related to the environmental context and the spatial and temporal scale of analysis. However, empirical evidence on diversity‐productivity patterns typically considers single temporal and spatial scales, and does not include the effect of environmental variables. In a metacommunity of macrophytes in ephemeral ponds, we analysed the diversity‐productivity relationship patterns in the field, the importance of the environmental variables of pond size and heterogeneity on such relationship, and the variation of these patterns at local (community level) and landscape scales (metacommunity level) across 52 ponds on twelve occasions, over five years (2005–2009). Combining all sampling dates, there were 377 ponds and 1954 sample‐unit observations. Vegetation biomass was used as a proxy for productivity, and biodiversity was represented by species richness, evenness, and their interaction. Environmental variables comprised pond area, depth and internal heterogeneity. Productivity and species richness were not directly related at the metacommunity level, and were positively related at the community level. Taking environmental variables into account revealed positive species richness‐productivity relationships at the metacommunity level and positive quadratic relationships at the community level. Productivity showed both positive and negative linear and nonlinear relationships with the size and heterogeneity of ponds. We found a weak relationship between productivity and evenness. The identity of variables associated with productivity changed between spatial scales and through time. The pattern of relationships between productivity and diversity depends on spatial scale and environmental context, and changes idiosyncratically through time within the same ecosystem. Thus, the diversity‐productivity relationship is not only a property of the study system, but also a consequence of environmental variations and the temporal and spatial scale of analysis.     

Arable weed vegetation of the northeastern part of the Czech Republic: effects of environmental factors on species composition

Factors determining changes in species composition of arable field weed vegetation in the northeastern part of the Czech Republic were studied. Gradsect sampling, i.e. a priori stratified selection of sampling sites, was used for the field research. Using this method, a data set of 174 vegetation plots, covering a whole range of basic environmental characteristics in the study area, was compiled in 2001–2003. A set of environmental variables (altitude, annual precipitation, mean annual temperature, soil type, soil pH and crop type) together with date of sampling was obtained for each plot. Ordination methods were used to determine the effects of variables on arable weed composition. For each variable, the gross and net effect on weed species composition were calculated. All variables considered in this study had a significant effect on weed species composition and explained 7.25% of the total variation in species data. Major changes in weed species composition in the study area were associated with different crop types. The second most important gradient in the variability of weed vegetation in the study area was associated with altitudinal and climatic changes followed by seasonal changes and different soil types and pH. Our results show that on a regional scale, the relative importance of different crop types and their associated management on changes in arable weed species composition is higher than the relative importance of climatic variables. The relative importance of climatic variables decreases with their decreasing length of gradient.     

Monitoring of marine benthic communities and taxonomic resolution: an approach through diverse habitats and substrates along the Southern Iberian coastline

Studies conducted along the southern Iberian coastline validate macrobenthic community analyses at taxonomic levels higher than that of species. Twelve studies on littoral benthic communities, carried out by the same research team, were selected spanning both a variety of sampling strategies (spatial, temporal, spatio-temporal) and substrate/habitat types (sediment, rock, algae). In order to establish differences between the results obtained at the taxonomic levels of species, family and order, similarities among stations were calculated using Spearman’s coefficient for ranges. A subset of three studies was selected to investigate possible differences in ‘best-explaining’ environmental variables with taxonomic level. The environmental variables selected at species level were the same as those found at levels of family and order. It is concluded that studies at the different levels of taxonomic resolution (species, family, order) lead to similar results both with regard to relative community distributions and the environmental variables associated with these. The importance of this result for monitoring similar benthic communities is discussed.     

Distribution of forest dwelling carabids (Coleoptera): spatial scale and the concept of communities

Associations between spatial distribution of ground-beetles (Carabidae) and environmental variables were studied over three hierarchical scales in deciduous forest in central Alberta, Canada We also examined the relationship between species abundance and distribution on several scales ranging from the local scale of our study to that of the North American temperate deciduous forest Understorey plant cover, tree cover, and occurrence of other carabids were associated with distribution of particular species at the smallest ecological scales within populations However, great differences in population sues of carabid species among five distinct sites several kilometres apart were not correlated with variation in the same environmental variables In central Alberta, abundance and extent of distribution were correlated positively among the 30 carabid species collected, and distributions of the ten species classified as 'core' species were generally aggregated at all spatial scales On the continental scale, there was a significant positive correlation between abundance and distribution for the 114 species of the entire data set, and the six species meeting the criteria of 'core' taxa on this scale, were also 'core' elements in central Alberta Further analysis of covariance of core elements of species assemblages across different taxa provides a sound empirical approach for understanding community organization     

Integrating life history traits into predictive phylogeography

Predictive phylogeography seeks to aggregate genetic, environmental and taxonomic data from multiple species in order to make predictions about unsampled taxa using machine‐learning techniques such as Random Forests. To date, organismal trait data have infrequently been incorporated into predictive frameworks due to difficulties inherent to the scoring of trait data across a taxonomically broad set of taxa. We refine predictive frameworks from two North American systems, the inland temperate rainforests of the Pacific Northwest and the Southwestern Arid Lands (SWAL), by incorporating a number of organismal trait variables. Our results indicate that incorporating life history traits as predictor variables improves the performance of the supervised machine‐learning approach to predictive phylogeography, especially for the SWAL system, in which predictions made from only taxonomic and climate variables meets only moderate success. In particular, traits related to reproduction (e.g., reproductive mode; clutch size) and trophic level appear to be particularly informative to the predictive framework. Predictive frameworks offer an important mechanism for integration of organismal trait, environmental data, and genetic data in phylogeographic studies.     

Scale and species-environment relationships in a central Himalayan oak forest,Nepal

A quantitative gradient study in a central Himalayan mixed Quercus forest (Q. semecarpifolia and Q. lamellosa) was made to evaluate the relationships between environmental variables and species composition at different scales.The data (91 taxa × 120 sub-plots) were sampled at three sites, where groups of four sub-plots (2.5 m × 2.5 m) were sampled within 10 m × 10 m. The species data were analysed together with the environmental data (elevation, relative radiation (RI) and soil variables) using Correspondence Analysis (CA) and its constrained version (CCA). The environmental variables used in CCA were chosen by forward selection.Elevation was the over-riding complex gradient (2000–3000 m a.s.l.), with loss-on-ignition, total nitrogen, and RI covarying. The most important factors independent of elevation were available phosphorus and tree canopy cover, whereas pH and nitrogen had minor independent influences.The overall species environment correlation was highest for the largest plot size. The species environment correlation increased with spatial extent for the largest plot size. The field-layer and shrub-tree strata did not consistently differ in their concordance between species and environment. The field-layer species had a stronger relationships with the soil variables, which may relate to rapid changes over a short spatial extent both for the field-layer species and for the soil variables.     

Evolution of termite functional diversity: analysis and synthesis of local ecological and regional influences on local species richness

Aim To (1) describe termite functional diversity patterns across five tropical regions using local species richness sampling of standardized areas of habitat; (2) assess the relative importance of environmental factors operating at different spatial and temporal scales in influencing variation in species representation within feeding groups and functional taxonomic groups across the tropics; (3) achieve a synthesis to explain the observed patterns of convergence and divergence in termite functional diversity that draws on termite ecological and biogeographical evidence to‐date, as well as the latest evidence for the evolutionary and distributional history of tropical rain forests. Location Pantropical. Methods A pantropical termite species richness data set was obtained through sampling of eighty‐seven standardized local termite diversity transects from twenty‐nine locations across five tropical regions. Local‐scale, intermediate‐scale and large‐scale environmental data were collected for each transect. Standardized termite assemblage and environmental data were analysed at the levels of whole assemblages and feeding groups (using components of variance analysis) and at the level of functional taxonomic groups (using correspondence analysis and canonical correspondence analysis). Results Overall species richness of local assemblages showed a greater component of variation attributable to local habitat disturbance level than to region. However, an analysis accounting for species richness across termite feeding groups indicated a much larger component of variation attributable to region. Mean local assemblage body size also showed the greater overall significance of region compared with habitat type in influencing variation. Ordination of functional taxonomic group data revealed a primary gradient of variation corresponding to rank order of species richness within sites and to mean local species richness within regions. The latter was in the order: Africa > south America > south‐east Asia > Madagascar > Australia. This primary gradient of species richness decrease can be explained by a decrease in species richness of less dispersive functional taxonomic groups feeding on more humified food substrates such as soil. Hence, the transects from more depauperate sites/regions were dominated by more dispersive functional taxonomic groups feeding on less humified food substrates such as dead wood. Direct gradient analysis indicated that ‘region’ and other large‐scale factors were the most important in explaining patterns of local termite functional diversity followed by intermediate‐scale geographical and site variables and, finally, local‐scale ecological variables. Synthesis and main conclusions Within regions, centres of termite functional diversity lie in lowland equatorial closed canopy tropical forests. Soil feeding termite evolution further down food substrate humification gradients is therefore more likely to have depended on the long‐term presence of this habitat. Known ecological and energetic constraints upon contemporary soil feeders lend support for this hypothesis. We propose further that the anomalous distribution of termite soil feeder species richness is partly explained by their generally very poor dispersal abilities across oceans. Evolution, radiation and dispersal of soil feeder diversity appears to have been largely restricted to what are now the African and south American regions. The inter‐regional differences in contemporary local patterns of termite species richness revealed by the global data set point to the possibility of large differences in consequent ecosystem processes in apparently similar habitats on different continents.     

Variation in the composition and structure of tropical savannas as a function of rainfall and soil texture along a large-scale climatic gradient in the Northern Territory, Australia

Abstract. Variation in structural and compositional attributes of tropical savannas are described in relation to variation in annual rainfall and soil texture along a subcontinental-scale gradient of rainfall in the wet-dry tropics of the Northern Territory, Australia. Rainfall varies along the gradient from over 1500 mm p.a. in the Darwin region ( c . 12° S) to less than 500 mm in the Tennant Creek region ( c . 18° S). Soils are patchy, and sands, loams and clays may occur in all major districts within the region. We utilized a large data set (1657 quadrats ° 291 woody species; with numerous measured and derived sample variables) covering an area of 0.5 million km². Correlations between floristic composition of woody species and environmental variables were assessed using DCA ordination and vector fitting of environmental variables. Vectors of annual rainfall and soil texture were highly correlated with variation in species composition. Multiple regression analyses incorporating linear and quadratic components of mean annual rainfall and topsoil clay content were performed on three structural attributes (tree height, tree cover, tree basal area) and two compositional attributes (woody species richness, deciduous tree species richness). Tree height declined with decreasing rainfall; cover, basal area, woody species richness and deciduous species richness all declined with decreasing rainfall and increasing soil clay content. Regression models accounted for between 17% and 45% of the variation in the data sets. Variation in other factors such as soil depth, landscape position and recent land-use practices (for which there were no data on an individual quadrat basis) are likely to have contributed to the large residual variation in the data set.