Often ecologists collect multivariate abundance (or presence-absence) data simultaneously across many taxa, at many places, with the intention of studying what occurs where, and why. Functional traits can help explain why. This talk reviews recent advances in methods for studying the extent to which functional traits can explain differences in environmental response across different taxa, their performance, and current state-of-the-art. Specifically, these methods essentially incorporate functional traits into a multivariate analysis, acting as additional predictors, but on columns (taxa) rather than on rows (sites) as environmental variables do. The interaction between environmental and trait variables solves the “fourth-corner problem” of finding environment-trait associations. It has been shown that these methods can significantly outperform the use of community-weighted means, and that the analysis method needs to account for: (1) correlation in abundance across taxa; (2) variation in environmental response across taxa not explained by functional traits. Regrettably, few methods are currently available that achieve both objectives.

Widespread use of ecological traits in community analyses might provide a solution to perceived problems of generality in community ecology. For traits to value-add to species-level analyses, they must either improve our ability to generalise across ecosystems with different species or provide a clearer understanding of shared responses within an ecosystem. We used the Global Ants Database, which includes georeferenced information on ant species assemblages and species traits, to test whether detection of differences in assemblages among habitats is improved by including traits. We focussed on three key morphological traits related to body size, vision and locomotion and used a fourth corner approach to modelling the environment-trait interactions on the mvabund package in R. We examined 23 studies separately and in combination, using ANOVA. Across all studies, inclusion of traits provided a moderate improvement of model fit (~15%), although the improvement was significant in 39% of individual studies. Where environmental contrasts were strong, the fourth corner term tended to explain more of the variance, but the value of including the trait-environment interaction term did not improve with species richness. While the mean per-study improvement provided by the inclusion of traits was only moderate, our findings suggest that ecological traits are of significant value in detecting environmental signals across ecosystems without shared species and are therefore likely to contribute to improved generality in community ecology.

Robust predictions of animal responses to climate conditions are vital for understanding community dynamics and ecosystem processes, particularly in response to rapid global changes, such as increased aridity and unpredictable rainfall. In arid ecosystems, the majority of pulse-reserve dynamics are driven by large rainfall events. However, the relative responses of arid-zone organisms to short-term changes in climate are unknown, particularly smaller rainfall events that briefly interrupt longer-term drought. During a dramatic shift from high to low annual rainfall, we used arthropods as model animals to describe the effects of a small summer rainfall event on population dynamics.
Most arid-zone arthropods declined during drought, as predicted by pulse-reserve theory. We present evidence that small pulse rainfall events may play a role in buffering small, quickly reproducing detritivores from population declines during drought. Taxa within this functional group retained high abundances despite the onset of winter temperatures and lack of significant subsequent rainfall. Some maintained a higher abundance over another summer. In contrast, larger herbivores and omnivores did not benefit significantly from the small rainfall event, and continued to decline over the two-year study period. Taxa traits are therefore likely to control which organisms benefit from rainfall inputs that do not improve primary productivity. By using a framework to identify which taxa benefit from small rainfall events, ecologists have an increased ability to predict the vulnerability of different species to increases in rainfall variability in the future.

There is little doubt the increasing power of computer processors and development of more sophisticated modelling tools, means the need for character or trait data is more important than ever. The Atlas of Living Australia (ALA) is Australia’s largest repository of biodiversity occurrence data and has been exploring the potential to incorporate traits into its biodiversity data services. Along with links to DNA repositories, this will allow researchers to explore the wider relationships between phylogenetics, functional characteristics and the environment.
Initial pilot work with the Biodiversity and Climate Change Virtual Laboratory (BCCVL), allowing users to combine species level trait values and ALA occurrence records, then export the data directly to the latest BCCVL species trait model, demonstrates that aggregators will need to deliver this type of data in the future. However, trait and character based data is significantly more complex than occurrence data with: little or no standardisation, limited commonality across lifeforms, inconsistent data collection, and a large array of ‘standard’ traits within lifeforms even if there is community agreement. Delivering large scale trait data merged with genomic and occurrence data will need a community wide strategy that: fosters collaboration; utilises existing ontologies and trait databases; assists communities to build and promote standards; and develops showcase exemplars starting with focused activities such as species identification.

Despite their cold climate and isolation, the Southern Ocean Islands are not immune to biological invasions. These extreme conditions and remoteness also make predicting invasions challenging. Data are scarce and extreme conditions necessitate extrapolation when applying correlative distribution models. Furthermore, cold-tolerance, a key determinant of establishment success, is generally more variable than its upper thermal counterpart.

To overcome these challenges we used remote sensing, downscaled using geophysical modelling validated by in situ data logging, to spatio-temporally map the thermal microclimate of the Southern Ocean Islands. We next measured the thermal performance of two springtail species that have established previously across the region and represent an ongoing threat. Physiological data were used as priors in a Bayesian species distribution model, which was projected across microclimate maps to identify the most suitable areas for establishment. To predict future invasions we used microclimate data that reflect 2100 conditions and adjusted thermal performance responses to account for plasticity and evolutionary potential.

The area of the Southern Ocean Islands that is suitable for the establishment of each species is far larger than current invaded ranges. This suggests that the invasion process is ongoing and, in the absence of strict biosecurity, impacts are likely to extend beyond those currently observed. Incorporating remote sensing, physiological testing, and evolutionary potential into species distribution modelling in extreme environments holds great promise when applied to the study of remote, cold regions, such as polar or montane habitats, where data are scarce and adaptation along an invasion front is likely.

Antarctica’s terrestrial ecosystems are populated by diverse yet understudied invertebrate communities, essential for healthy ecosystem functioning. Contemporary distribution patterns are highly influenced by climatic changes that have occurred since the end of the Last Glacial Maximum (LGM) that led to melting ice-sheets and rising sea-levels. Mite, springtail and nematode samples collected from the islands off the Antarctic Peninsula between 2014- 2016 will be analysed using comparative phylogeographic techniques that combine molecular phylogenies with geographical and climatic datasets. These will be contrasted with temperate and tropical Australian soil communities collected along a similar latitudinal transect, with islands representing similarly isolated ecosystems. Comparing communities with a known period of isolation will assist in the timing of speciation events and formulating dispersal models. Detailed analysis will also reveal the drivers of distribution at the a) local scale; environmental and biotic variables, b) regional scale; climatic influences and gene flow; and, c) temporal scale; evolution and dispersal. Preliminary sequencing data shows reduced mite diversity at lower latitudes, that mirrors observations in microbial communities. This project has the potential of establishing baseline biodiversity for these vital ecosystem operators, and advance our predictions based on current trends in climate change and ecosystem fragmentation.