Predictor Of Species Richness Research Articles

Do species distribution models predict species richness in urban and natural green spaces? A case study using amphibians

Urban green spaces are potentially important to biodiversity conservation because they could provide patches of high quality habitat or connectivity to nearby habitat. Presence-only species distribution models (SDMs) represent a potential tool for assessing the biodiversity value of urban green space; however, there is limited research to validate SDM results with field surveys to see if the predictions accurately represent observed species richness. We generated a range of SDMs using multiple suitability thresholds for 23 species of amphibians that occur in southwest, Ohio, USA. The distributions were overlaid to enumerate species richness. We surveyed 20 sites for amphibian species to evaluate model predictions. Our models over-predicted species richness relative to survey data. For example, we observed a mean pairwise difference of 14 species between models of species richness and observed values. Our results suggest either SDMs built with landscape variables we selected did not represent accurately amphibian richness, or the amphibian surveys did not detect all species present. Analyzing sites that had more than three sampling events suggests the explanation of inadequate sampling effort is only partially correct. Differences such as that between predicted and observed values of species richness is a challenge for land managers and conservation biologists that need a tool for modeling biodiversity. Species distribution models did project relative species richness well in urban and non-urban green space, which suggests this technique offers a spatially explicit way to identify more species rich areas and may help managers and conservation biologists manage systems with greater efficiency.

Predicted Effect of Landscape Position on Wildlife Habitat Value of Conservation Reserve Enhancement Program Wetlands in a Tile‐drained Agricultural Region

AbstractJustification for investment in restored or constructed wetland projects are often based on presumed net increases in ecosystem services. However, quantitative assessment of performance metrics is often difficult and restricted to a single objective. More comprehensive performance assessments could help inform decision‐makers about trade‐offs in services provided by alternative restoration program design attributes. The primary goal of the Iowa Conservation Reserve Enhancement Program is to establish wetlands that efficiently remove nitrates from tile‐drained agricultural landscapes. A secondary objective is provision of wildlife habitat. We used existing wildlife habitat models to compare relative net change in potential wildlife habitat value for four alternative landscape positions of wetlands within the watershed. Predicted species richness and habitat value for birds, mammals, amphibians, and reptiles generally increased as the wetland position moved lower in the watershed. However, predicted average net increase between pre‐ and post‐project value was dependent on taxonomic group. The increased average wetland area and changes in surrounding upland habitat composition among landscape positions were responsible for these differences. Net change in predicted densities of several grassland bird species at the four landscape positions was variable and species‐dependent. Predicted waterfowl breeding activity was greater for lower drainage position wetlands. Although our models are simplistic and provide only a predictive index of potential habitat value, we believe such assessment exercises can provide a tool for coarse‐level comparisons of alternative proposed project attributes and a basis for constructing informed hypotheses in auxiliary empirical field studies.

Modeling α- and β-diversity in a tropical forest from remotely sensed and spatial data

Comprehensive information on species distribution and species composition patterns of plant communities is required for effective conservation and management of biodiversity. Remote sensing offers an inexpensive means of attaining complete spatial coverage for large areas, at regular time intervals, and can therefore be extremely useful for estimating both species richness and spatial variation of species composition (α- and β-diversity). An essential step to map such attributes is to identify and understand their main drivers. We used remotely sensed data as a surrogate of plant productivity and habitat structure variables for explaining α- and β-diversity, and evaluated the relative roles of productivity-habitat structure and spatial variables in explaining observed patterns of α- and β-diversity by using a Principal Coordinates of Neighbor Matrices analysis. We also examined the relationship between remotely sensed and field data, in order to map α- and β-diversity at the landscape-level in the Yucatan Peninsula, using a regression kriging procedure. These two procedures integrate the relationship of species richness and spatial species turnover both with remotely sensed data and spatial structure. The empirical models so obtained can be used to predict species richness and variation in species composition, and they can be regarded as valuable tools not only for identifying areas with high local species richness (α-diversity), but also areas with high species turnover (β-diversity). Ultimately, information obtained in this way can help maximize the number of species preserved in a landscape.

Parasitoid wasp diversity in apple orchards along a pest-management gradient

This study examined the impact of pest-management practices on Hymenopteran parasitoid diversity and assemblage composition in six apple orchards in southeastern Michigan, USA. The orchards comprised a gradient of pest-management intensity from organic to IPM to conventional practices. We used a pesticide toxicity index to quantify monthly and seasonal pest-management intensity in each orchard and conducted monthly vacuum-sampling of wasps during summer of 2009.Monthly toxicity scores predicted wasp abundance, setting an upper bound on the number of wasps in the orchard, but did not predict species richness. Total species richness was significantly higher in the organic orchard than in all others, but in August a conventional orchard had the highest wasp abundance and species richness. These results suggest that a toxicity index could guide pesticide choices or application times so as to increase parasitoid wasp populations, which have untapped potential for pest management in apple orchards.

Dissecting NDVI–species richness relationships in Hawaiian dry forests

AbstractAim A growing body of research has used the normalized difference vegetation index (NDVI) as a proxy for productivity to predict species richness. Yet the mechanisms that produce the relationship between NDVI and species richness remain unclear because of correlated biotic and abiotic factors that influence NDVI. In this study we investigated different biotic and abiotic effects that potentially drive plant species richness–productivity relationships.Location Hawaiian Islands, USA.Methods We quantified woody plant species richness, structure (density, basal area and canopy height), and species composition along a precipitation gradient of 14 Hawaiian dry forest plots. We then used structural equation models combined with 10 years of satellite data to disentangle the effects of precipitation, structure and NDVI‐estimated productivity on species richness.Results Underlying the simple correlation between NDVI and species richness was the indirect effect of precipitation and direct effect of forest structure. The best‐fit model showed there was no direct effect of NDVI on species richness.Main conclusions Our results demonstrate that complex relationships drive simple correlations between species richness and productivity. Considering the mechanisms and underlying factors driving NDVI–species richness relationships could improve predictions of species diversity as satellite measures of productivity have an increasingly important role in habitat mapping, species distribution modelling and predictions for global change.

The confounded effects of habitat disturbance at the local, patch and landscape scale on understorey birds of the Atlantic Forest: Implications for the development of landscape-based indicators

The search for surrogates of changes in species richness and community structure in fragmented landscapes involves not only the selection of predictors, such as landscape metrics or environmental variables, but also the identification of the spatial scale that is most relevant to the taxa in question. However the heavily intercorrelated nature of many structural features in fragmented landscapes complicates analyses, and the wide variation in species responses prevents the identification of a general trend. In this study, we used a two-tiered hierarchical variation partitioning to identify the unique and shared effects of: 1 – changes in vegetation structure at the plot scale, patch structure (size and shape), and forest cover at the landscape scale; and 2 – variables within these scales; as predictors of species richness and species’ abundances of birds in a fragmented landscape of Atlantic Forest; with the goal of aiding to the development of biodiversity indicators. Birds were sampled with mist-nets with a constant effort of 680 net-h at each of 23 sites, which resulted in almost 2600 captures. At the community level, regression models showed that changes in plot, patch and landscape scale variables explained a large proportion of the variation in species richness, but results from variation partitioning showed that the intercorrelation among predictors was so high that the unique contribution of each was non-significant. Our results point to a relatively large unique effect of local and landscape scale variables at the community level, but we also show that results vary greatly depending on the trophic guild being analysed. At the species level, multiple scale models also presented high explanatory power, however, species responses were so varied that we could not detect a general trend. We conclude that there is no single ‘best’ scale that could function as a proxy for changes in bird communities because each species and functional guild is uniquely affected by the environment, and suggest that efforts should be focused on finding indicators that encompasses all scales and the needs of different taxa.

Variation in habitat use of coexisting rodent species in a tropical dry deciduous forest

Use of habitat is a critical component related to structure of small-mammal communities, with partitioning occurring primarily along dimensions of microhabitat, although use of microhabitat often does not explain fully use at a macrohabitat level. Through grid studies of small mammals in coastal Colima, Mexico (during January 2003–2005), we appraised influence of available habitat, species richness, abundance, and cumulative abundance of other small mammals on variation in habitat used by species. We evaluated 14 habitat variables (reflecting ground cover, slope, canopy, and vegetation density on vertical and horizontal axes) and developed a composite variable (principal component 1) reflecting general openness of habitat through which we addressed habitat use. For the four most common mammalian species (Sigmodon mascotensis, Heteromys pictus, Baiomys musculus, and Oryzomys couesi), two measures of variation in habitat used were employed to estimate niche breadth, one of which assessed variation in habitat use relative to variation present on a grid. Sigmodon mascotensis and B. musculus preferred areas that were more open, and H. pictus and O. couesi occupied less-open areas; breadth of habitat use did not differ interspecifically. Habitat use was more variable on grids with more variability in habitat, although not greater than chance expectations. Findings do not lend support to the resource-breadth hypothesis as an explanation for population densities of species at a local level or the habitat-heterogeneity hypothesis as a predictor of species richness. Variation in habitat used by S. mascotensis did not proportionally increase when diverse habitat was available but was greater when the species was more abundant. For H. pictus, when cumulative abundance of other small mammals was greater, breadth of habitat used was greater. Intraspecific density-dependent habitat selection may result in S. mascotensis selecting a greater variety of habitats, while greater interspecific abundance is related to a greater range in use of habitats by H. pictus. Baiomys musculus used a higher proportion of habitat relative to that available when more species were present on a grid. Variation in habitat used by O. couesi was unrelated to any factor examined. Overall, the four species responded in notably different ways with respect to availability of habitat, abundance, and presence of other species.

Effects of habitat fragmentation on ant richness and functional composition in Brazilian Atlantic forest

Ants are a dominant faunal group in tropical forests, but their responses to human disturbances remain poorly investigated. Here we examine the relative effects of habitat fragmentation (fragment size and amount of forest cover retained in the sur- rounding landscape) and habitat structure (tree density and richness) on ant species and functional composition in the Atlantic forest of northeast Brazil. Our study was carried out in 19 fragments ranging in size from 10 to 3,500 ha. We recorded 146 ant species from 41 genera. Fragment size and tree density were the most important variables predicting species richness and functional diversity. Additionally, fragment area and tree density also explained the richness of a range of global functional groups (Cryptic Species, Specialists Predators and Tropical-climate Specialists) and Atlantic forest functional groups (Cryptic Predators, Cryptic Omnivores, Epigaeic Predators and Epigaeic Omnivores). Taxonomic composition was influenced by fragment area and tree richness, while functional compo- sition responded to forest cover and tree density. Overall variation of global functional groups was also affected by tree richness. Our results suggest that specialist ants (espe- cially cryptic species, specialist predators and climate specialists) are particularly sensitive to forest fragmentation. Such functional groups would make useful indicators of ant responses to forest fragmentation.

The performance of range maps and species distribution models representing the geographic variation of species richness at different resolutions

ABSTRACTAim The method used to generate hypotheses about species distributions, in addition to spatial scale, may affect the biodiversity patterns that are then observed. We compared the performance of range maps and MaxEnt species distribution models at different spatial resolutions by examining the degree of similarity between predicted species richness and composition against observed values from well‐surveyed cells (WSCs).Location Mexico.Methods We estimated amphibian richness distributions at five spatial resolutions (from 0.083° to 2°) by overlaying 370 individual range maps or MaxEnt predictions, comparing the similarity of the spatial patterns and correlating predicted values with the observed values for WSCs. Additionally, we looked at species composition and assessed commission and omission errors associated with each method.Results MaxEnt predictions reveal greater geographic differences in richness between species rich and species poor regions than the range maps did at the five resolutions assessed. Correlations between species richness values estimated by either of the two procedures and the observed values from the WSCs increased with decreasing resolution. The slopes of the regressions between the predicted and observed values indicate that MaxEnt overpredicts observed species richness at all of the resolutions used, while range maps underpredict them, except at the finest resolution. Prediction errors did not vary significantly between methods at any resolution and tended to decrease with decreasing resolution. The accuracy of both procedures was clearly different when commission and omission errors were examined separately.Main conclusions Despite the congruent increase in the geographic richness patterns obtained from both procedures as resolution decreases, the maps created with these methods cannot be used interchangeably because of notable differences in the species compositions they report.

Predicting species richness in tropical lichenized fungi with ‘modular’ combinations of character states

A numerical approach is presented to predict the number of potentially undiscovered species in groups of species that exhibit a ‘modular’ combination of character states, that is, nearly ‘free’ combination that is not random but allows for a wide range of cross-combination of character states. The method uses a character correlation index (CCI) based on observed versus expected frequencies of pairwise character state combinations in known species of a group to estimate the degree of potentially positive and negative evolutionary constraints in the co-occurrence of character states. For instance, if a character state combination, such as ‘muriform ascospores’ and ‘presence of norstictic acid’, occurs less frequently than expected by chance, it indicates the presence of a negative evolutionary constraint. If no such constraint is detected, but a particular character combination has not yet been observed in a described species, it can be predicted that yet undescribed species will exhibit that combination. The method can therefore be used to detect ‘gaps’ in the taxonomic record, i.e. character combinations that have not been detected in known species but are likely to occur in undiscovered species. Modular combination of character states appears to be particularly common in Fungi, including lichens, and the method was here used to predict species richness in the lichen genus Graphis, specifically the Graphis scripta group. A total of 42 species is known in this group and, out of a total of 240 possible character state combinations, a further 48 were found likely to represent undiscovered species, suggesting that the real species number in this group is close to 90 or about twice the number now known. This agrees well with previous estimates for undiscovered diversity of tropical lichen species and for the first time gives a taxonomic prediction of how such ‘missing’ species are expected to look like.

Drivers of diversity and tree cover in gardens, parks and streetscapes in an Australian city

While we know that urban vegetation is often distributed unequally, most studies have been undertaken in cities with relatively high levels of income inequality, using a single measure of distribution (usually tree cover) and in a single land use. This study explores predictors of both tree cover and species richness in gardens, streetscapes and parks in Ballarat, Australia. Spatial regression models found that education level was a more important predictor of tree cover than household income across all land uses in Ballarat which can be explained by some people with high incomes relative to education level choosing to live in new residential developments with disproportionately low levels of tree cover. Inequality in tree cover was greater in streetscapes than in residential gardens, suggesting that ‘top down’ political factors are more important than individual behaviours in determining tree cover in Ballarat. In contrast, physical rather than socioeconomic factors were better predictors of species richness across all land uses, highlighting that different measures of vegetation distribution are not necessarily correlated.

A Phenological Mid-Domain Analysis of Non-Native and Native Species Recruitment Richness

A mid-domain model was used to examine differences in patterns of species recruitment diversity between native and non-native species groups within shallow sub-tidal epifaunal communities of southern New England. A mid-domain model is a stochastic null model that predicts species richness patterns within a bounded domain. Deviations from the mid-domain model were examined to assess: 1) if each species group’s recruitment diversity differed from the mid-domain model; and 2) if there were differences in the patterns of deviation between groups. In general, the shape of mid-domain null was correlated to empirical patterns of recruitment diversity (i.e., mid-summer peaks in recruitment diversity) although among-year variation occurred in both the pattern of recruitment diversity and the level of fit. Akiake’s information criterion was used to determine if the mid-domain effect contributed to linear regression model predictions of phenological species recruitment patterns. We found distinct differences between the level of fit between native and non-native species groups. Native species did not significantly differ from the mid-domain null in five of seven years examined, while non-native species were significantly different in all years but two. Non-native species also showed a much weaker correlation to the mid-domain model and the seasonal water temperature signal. Collectively, this evidence supports the hypothesis that native and non-native species have different factors controlling the timing of their recruitment. Mid-domain nulls proved to be useful in the analysis of species richness patterns along temporal axes and provide a valuable baseline for further analysis.

Biogeography of pelagic bacterioplankton across an antagonistic temperature–salinity gradient in the Red Sea

The Red Sea is a unique marine ecosystem with contrasting gradients of temperature and salinity along its north-to-south axis. It is an extremely oligotrophic environment that is characterized by perpetual year-round water column stratification, high annual solar irradiation, and negligible riverine and precipitation inputs. In this study, we investigated whether the contemporary environmental conditions shape community assemblages by pyrosequencing 16S rRNA genes of bacteria in surface water samples collected from the northeastern half of this water body. A combined total of 1855 operational taxonomic units (OTUs) were recovered from the 'small-cell' and 'large-cell' fractions. Here, a few major OTUs affiliated with Cyanobacteria and Proteobacteria accounted for ∼93% of all sequences, whereas a tail of 'rare' OTUs represented most of the diversity. OTUs allied to Surface 1a/b SAR11 clades and Prochlorococcus related to the high-light-adapted (HL2) ecotype were the most widespread and predominant sequence types. Interestingly, the frequency of taxa that are typically found in the upper mesopelagic zone was significantly elevated in the northern transects compared with those in the central, presumably as a direct effect of deep convective mixing in the Gulf of Aqaba and water exchange with the northern Red Sea. Although temperature was the best predictor of species richness across all major lineages, both spatial and environmental distances correlated strongly with phylogenetic distances. Our results suggest that the bacterial diversity of the Red Sea is as high as in other tropical seas and provide evidence for fundamental differences in the biogeography of pelagic communities between the northern and central regions.

Insular lake island biogeography: using lake metrics to predict diversity in littoral zone mollusk communities

Island biogeography theory can be used to explain patterns of species richness on various types of habitat islands, including freshwater lake systems. Mollusk production in these systems also has been linked to various water-chemistry variables, such as pH, alkalinity, hardness, and specific conductance. We examined how mollusk diversity patterns were related to geographical and limnological factors in insular lakes of the Beaver and Manitou Archipelagos in Lake Michigan (Laurentian Great Lakes), USA. The strongest correlations observed were with shoreline development (r = 0.80), specific conductance (r = 0.87), and pH (r = 0.87). Principal components analysis revealed that isolation by distance and PO43− concentration also may have affected species richness and abundance. Shoreline length was a better predictor of species richness than surface area, but both measures of habitat size were unable to account for much of the variation in species richness. The data suggest that shoreline length and development represent available habitat area more accurately than lake area for primarily littoral-dwelling mollusks. The relatively weak correlations observed with lake area and isolation from Lake Michigan suggest that application of island biogeography theory to predict mollusk species richness using only lake surface area and isolation by distance is limited for freshwater mollusks.

Relative role of contemporary environment versus history in shaping diversity patterns of China's woody plants

What determines large‐scale patterns of species diversity is a central and controversial topic in biogeography and ecology. In this study, we compared the effects of contemporary environment and historical contingencies on species richness patterns of woody plants in China, using fine‐resolution geographic databases of the distributions of 11 405 woody species and climate, topography, and vegetation information. Residuals of species richness‐environment generalized linear models were significantly different from 0 in the majority of seven biogeographical regions, and also differed significantly between these regions, indicating significant deviation from the predicted species richness based on contemporary environment. Additionally, species richness of a given biogeographical region deviated substantially from the predictions of species richness‐environment models developed for the remaining regions combined. This suggests different richness‐environment relationships among regions. These results indicate important historical signals in the species richness patterns of woody plants across China. The signals are especially pronounced in the eastern Himalayas, the Mongolian Plateau, and the Tibetan Plateau, perhaps reflecting their special geological features and history. Nevertheless, partial regression indicated that historical effects were less important relative to contemporary environment. In conclusion, contemporary environment (notably climate) determines the general trend in woody‐plant species richness across China, while historical contingencies generate regional deviations from this trend. Our findings imply that both species diversity and regional evolutionary and ecological histories should be taken into account for future nature conservation.

Higher taxa as surrogates of species richness of spiders in insect‐resistant transgenic rice

Abstract Biodiversity assessments can often be time‐ and resource‐consuming. Several alternative approaches have been proposed to reduce sampling efforts, including indicator taxa and surrogates. In this study, we examine the reliability of higher taxon surrogates to predict species richness in two experimental rice fields of Fujian Province, southeastern China during 2005 and 2009. Spider samples in transgenic and nontransgenic plots were collected using a suction sampler. Both the genus and family surrogates had significant and positive linear relationships with species richness in the transgenic and nontransgenic rice fields. The rice varieties did not significantly influence the linear relationships. Our findings suggest that higher‐taxon surrogacy could be a useful alternative to complete species inventory for risk assessments of transgenic rice.

The fate of European breeding birds under climate, land‐use and dispersal scenarios

AbstractMany species have already shifted their distributions in response to recent climate change. Here, we aimed at predicting the future breeding distributions of European birds under climate, land‐use, and dispersal scenarios. We predicted current and future distributions of 409 species within an ensemble forecast framework using seven species distribution models (SDMs), five climate scenarios and three emission and land‐use scenarios. We then compared results from SDMs using climate‐only variables, habitat‐only variables or both climate and habitat variables. In order to account for a species’ dispersal abilities, we used natal dispersal estimates and developed a probabilistic method that produced a dispersal scenario intermediate between the null and full dispersal scenarios generally considered in such studies. We then compared results from all scenarios in terms of future predicted range changes, range shifts, and variations in species richness. Modeling accuracy was better with climate‐only variables than with habitat‐only variables, and better with both climate and habitat variables. Habitat models predicted smaller range shifts and smaller variations in range size and species richness than climate models. Using both climate and habitat variables, it was predicted that the range of 71% of the species would decrease by 2050, with a 335 km median shift. Predicted variations in species richness showed large decreases in the southern regions of Europe, as well as increases, mainly in Scandinavia and northern Russia. The partial dispersal scenario was significantly different from the full dispersal scenario for 25% of the species, resulting in the local reduction of the future predicted species richness of up to 10%. We concluded that the breeding range of most European birds will decrease in spite of dispersal abilities close to a full dispersal hypothesis, and that given the contrasted predictions obtained when modeling climate change only and land‐use change only, both scenarios must be taken into consideration.

Complex effects of scale on the relationships of landscape pattern versus avian species richness and community structure in a woodland savanna mosaic

Landscape pattern metrics are widely used for predicting habitat and species diversity. However, the relationship between landscape pattern and species diversity is typically measured at a single spatial scale, even though both landscape pattern, and species occurrence and community composition are scale‐dependent. While the effects of scale on landscape pattern are well documented, the effects of scale on the relationships between spatial pattern and species richness and composition are not well known. Here, our main goal was to quantify the effects of cartographic scale (spatial resolution and extent) on the relationships between spatial pattern and avian richness and community structure in a mosaic of grassland, woodland, and savanna in central Wisconsin. Our secondary goal was to evaluate the effectiveness of a newly developed tool for spatial pattern analysis, multiscale contextual spatial pattern analysis (MCSPA), compared to existing landscape metrics. Landscape metrics and avian species richness had quadratic, exponential, or logarithmic relationships, and these patterns were generally consistent across two spatial resolutions and six spatial extents. However, the magnitude of the relationships was affected by both resolution and extent. At the finer resolution (10‐m), edge density was consistently the best predictor of species richness, followed by an MCSPA metric that measures the standard deviation of woody cover across extents. At the coarser resolution (30‐m), NDVI was the best predictor of species richness by far, regardless of spatial extent. Another MCSPA metric that denotes the average woody cover across extents, together with percent of woody cover, were always the best predictors of variation in avian community structure. Spatial resolution and extent had varying effects on the relationships between spatial pattern and avian community structure. We therefore conclude that cartographic scale not only affects measures of landscape pattern per se, but also the relationships among spatial pattern, species richness, and community structure, often in complex ways, which reduces the efficacy of landscape metrics for predicting the richness and diversity of organisms.

Environmental influences on fish assemblages in irrigation ponds

Ponds are common features of the landscape and are considered important for freshwater biodiversity conservation. Although fish have a significant impact on the lentic ecosystems, the environmental factors that regulate fish assemblages in human-created water bodies, such as irrigation ponds, remain unclear. We evaluated the relationship between environmental factors and the fish assemblage structure in 31 ponds located in northern Japan. Species richness (range: 1–9) was positively correlated with the size of the inflow channel. Multivariate analyses revealed that the size of the inflow channel was a better predictor for species richness than lake morphology (surface area and maximum depth), vegetation coverage, water quality (turbidity, pH, DO, and EC), distance to the main channel, and distance to an adjacent pond. Species richness was significantly different between ponds with and without an inflow channel. Furthermore, three of the four most commonly observed species are thought to be relatively tolerant to low oxygen. Given that ponds have a relatively high local extinction rate resulting from exposure to stressful conditions, such as low oxygen and/or small population sizes, our results suggest that immigration from surrounding water bodies plays an important role in maintaining species richness of pond-dwelling fish.

Diversity and community composition of euglossine bee assemblages (Hymenoptera: Apidae) in western Amazonia

Tropical forests are known for their diverse insect fauna. We aimed to determine the effect and relative importance of latitude, elevation and climatic factors affecting species richness and turnover in euglossine bee assemblages along a gradient of 18° latitude from tropical rainforests to subtropical, deciduous dry forests in Peru and Bolivia. Sixteen forest sites were sampled during the dry season. Variance partitioning techniques were applied to assess the relative effects of the spatial and environmental variables on species richness and composition. Furthermore, we conducted a Species Indicator Analysis to find characteristic species for the biogeographic zones. There was a significant decrease in species richness towards the subtropical area. The best predictors of species richness were precipitation and its consequences on soil properties as well as temperature seasonality. The abundance of euglossines was most closely related to precipitation and soil-pH, but the causal links of abundance to these factors is unclear since soil-pH itself is correlated to a drastic turnover of vegetation structure. Based on the analysis of assemblage composition we propose three different assemblages with a transitional zone at the southern tropical area. The biogeographical distribution of euglossine bees along our study transect appears to be primarily related to climatic conditions and does not reflect the common subdividion of Amazonia into drainage systems.