A large array of species distribution model (SDM) approaches has been developed for explaining and predicting the occurrences of individual species or species assemblages. Given the wealth of existing models, it is unclear which models perform best for interpolation or extrapolation of existing data sets, particularly when one is concerned with species assemblages. We compared the predictive performance of 33 variants of 15 widely applied and recently emerged SDMs in the context of multispecies data, including both joint SDMs that model multiple species together, and stacked SDMs that model each species individually combining the predictions afterward. We offer a comprehensive evaluation of these SDM approaches by examining their performance in predicting withheld empirical validation data of different sizes representing five different taxonomic groups, and for prediction tasks related to both interpolation and extrapolation. We measure predictive performance by 12 measures of accuracy, discrimination power, calibration, and precision of predictions, for the biological levels of species occurrence, species richness, and community composition. Our results show large variation among the models in their predictive performance, especially for communities comprising many species that are rare. The results do not reveal any major trade-offs among measures of model performance; the same models performed generally well in terms of accuracy, discrimination, and calibration, and for the biological levels of individual species, species richness, and community composition. In contrast, the models that gave the most precise predictions were not well calibrated, suggesting that poorly performing models can make overconfident predictions. However, none of the models performed well for all prediction tasks. As a general strategy, we therefore propose that researchers fit a small set of models showing complementary performance, and then apply a cross-validation procedure involving separate data to establish which of these models performs best for the goal of the study.

Vegetated roofs are human-manufactured ecosystems and potentially promising conservation tools for various taxa and habitats. Focussing on arthropods, we conducted a 3 year study on newly constructed vegetated roofs with shallow substrates (up to 10 cm) and vegetation established with pre-grown mats, plug plants and seeds to describe pioneer arthropod communities on roofs and to compare them with ground level communities. We vacuum sampled arthropods from the roofs and nearby ground level sites with low, open vegetation, i.e., potential source habitats. We showed that the roofs and ground sites resembled each other for ordinal species richness but differed in community composition: with time the roofs started to resemble each other rather than their closest ground level habitats. Species richness increased with time on roofs and at ground level, but the roofs had consistently less species than the ground sites and only a few species were unique to the roofs. Also, the proportion of predators increased on roofs, while not at ground level. We conclude that vegetated roofs established with similar substrates and vegetation, filter arthropods in a way that produces novel communities that are different from those at ground level but similar to one another. The role of these insular communities in species networks and ecosystem function remains to be investigated.

Ecological drift can override the effects of deterministic niche selection on small populations and drive the assembly of some ecological communities. We tested this hypothesis with a unique data set sampled identically in 200 streams in two regions (tropical Brazil and boreal Finland) that differ in macroinvertebrate community size by fivefold. Null models allowed us to estimate the magnitude to which beta-diversity deviates from the expectation under a random assembly process while taking differences in richness and relative abundance into account, i.e., beta-deviation. We found that both abundance- and incidence-based beta-diversity was negatively related to community size only in Brazil. Also, beta-diversity of small tropical communities was closer to stochastic expectations compared with beta-diversity of large communities. We suggest that ecological drift may drive variation in some small communities by changing the expected outcome of niche selection, increasing the chances of species with low abundance and narrow distribution to occur in some communities. Habitat destruction, overexploitation, pollution, and reductions in connectivity have been reducing the size of biological communities. These environmental pressures might make smaller communities more vulnerable to novel conditions and render community dynamics more unpredictable. Incorporation of community size into ecological models should provide conceptual and applied insights into a better understanding of the processes driving biodiversity.

A popular idea in ecology is that trait variation among individuals from the same species may promote the coexistence of competing species. However, theoretical and empirical tests of this idea have yielded inconsistent findings. We manipulated intraspecific trait diversity in a ciliate competing with a nematode for bacterial prey in experimental microcosms. We found that intraspecific trait variation inverted the original competitive hierarchy to favour the consumer with variable traits, ultimately resulting in competitive exclusion. This competitive outcome was driven by foraging traits (size, speed and directionality) that increased the ciliate's fitness ratio and niche overlap with the nematode. The interplay between consumer trait variation and competition resulted in non-additive cascading effects-mediated through prey defence traits-on prey community assembly. Our results suggest that predicting consumer competitive population dynamics and the assembly of prey communities will require understanding the complexities of trait variation within consumer species.

Premise of research. Phenotypic traits that consistently mediate species' responses to environmental variation (functional traits) provide a promising approach toward generalizing ecological and evolutionary patterns and thereby gaining insights into the processes generating them. In the plant functional ecology literature, most trait-based studies have focused on traits mediating either resource competition or responses to variation in the abiotic environment, while traits mediating reproductive interactions have often been neglected. Methodology. Here, I discuss the value of herkogamy, the spatial separation of male and female functions in flowers, as a functional trait in plant reproductive biology and review the evidence relevant to the hypothesis that taxa exhibiting greater herkogamy have historically experienced more reliable pollination and more outcrossed mating systems. Pivotal results. A large body of work in the field of plant reproductive biology has identified a set of nearly ubiquitous correlations between average herkogamy and features of plant mating systems, notably, autofertility (seed set in the absence of pollinators) and outcrossing rate. Herkogamy often varies extensively among populations and species, and the adaptive interpretation is that herkogamy exhibits local adaptation to the reliability of the pollination environment. Conclusions. These results underline the value of herkogamy as a functional trait representing variation in mating histories. Many important insights are likely to emerge from studies leveraging herkogamy as an easily measured proxy of plant mating systems, as already demonstrated in comparative studies and studies of reproductive interactions. Greater consideration of herkogamy and other reproductive-function traits in studies of species coexistence may provide a more complete understanding of community assembly processes.

1. Community responses to and recovery from disturbances depend on local (e.g. presence of refuges) and regional (connectivity to recolonization sources) factors. Droughts are becoming more frequent in boreal regions, and are likely to constitute a severe disturbance for boreal stream communities where organisms largely lack adaptations to such hydrological extremes. 2. We conducted an experiment in 24 semi-natural stream flumes to assess the effects of local and regional factors on the responses of benthic invertebrate communities to a short-term drought. We manipulated flow (drought vs. constant-flow), spatial arrangement of leaf litter patches (aggregated vs. evenly distributed) and colonization from regional species pool (enhanced vs. ambient connectivity) to test the combined effects of disturbance, resource arrangement and connectivity on the structural and functional responses of benthic invertebrate communities. 3. We found that a drought as short as 1 week reduced invertebrate taxonomic richness and abundance, mainly through stochastic extinctions. Such changes in richness were not reflected in functional diversity. This suggests that communities were characterized by a high degree of functional redundancy, which allowed maintenance of functional diversity despite species losses. Feeding groups responded differently to drought, with organic matter decomposers responding more than scrapers and predators. 4. Three weeks were insufficient for complete invertebrate community recovery from drought. However, recovery was greater in channels subjected to enhanced connectivity, which increased taxonomic diversity and abundance of certain taxa. Spatial configuration of resources explained the least variation in our response variables, having a significant effect only on invertebrate abundance and evenness (both sampling occasions) and taxonomic richness (end of recovery period). 5. Even a short drought, if occurring late in the season, may not allow communities to recover before the onset of winter, thus having a potentially long-lasting effect on stream communities. For boreal headwaters, extreme dewatering poses a novel disturbance regime that may trigger substantial and potentially irreversible changes. An improved understanding of such changes is needed to underpin adaptive management strategies in these increasingly fragmented and disturbed ecosystems.

Although spatial and temporal variation are both important components structuring microbial communities, the exact quantification of temporal turnover rates of fungi and bacteria has not been performed to date. In this study, we utilised repeated resampling of bacterial and fungal communities at specific locations across multiple years to describe their patterns and rates of temporal turnover. Our results show that microbial communities undergo temporal change at a rate of 0.010-0.025 per year (in units of Sorensen similarity), and the change in soil is slightly faster in fungi than in bacteria, with bacterial communities changing more rapidly in litter than soil. Importantly, temporal development differs across fungal guilds and bacterial phyla with different ecologies. While some microbial guilds show consistent responses across regional locations, others show site-specific development with weak general patterns. These results indicate that guild-level resolution is important for understanding microbial community assembly, dynamics and responses to environmental factors.