Ecological Society of Australia 2014 Annual Conference

We have a lot of lab members presenting at the Ecological Society of Australia Annual Conference this year. A showcase of research! Below are the abstracts. More information on the conference can be found here.


Monday 29/9/2014

11:30 am: Nathan Emery (@ecotechnica), Ellery Room B

Experimental evidence confirms the interactive effect of soil and climate on predicted distributions

Bioclimatic models are useful for quantifying predicted changes to geographic distributions by correlating species presence with their associated climatic variables. However, models that use only climatic data fail to fully account for other non-climatic attributes that might significantly influence future geographic occurrence. Caution is warranted, therefore, when interpreting bioclimatic models that do not incorporate a broader set of ecological factors. To further understand the consequence of including non-climatic factors, such as edaphic attributes, in predictive models, we used a series of iterative experimental data sets to compare and assess the predictive precision of our models. We used contemporary and historical geographic records of the plant species Actinotus helianthi, coupled with bioclimatic data, to generate predicted habitat suitability using four IPCC AR5 climate scenarios in MaxEnt. Using these projections, seedling emergence and growth was tested in soil from four local sites, and twelve predicted sites, where suitable climate was calculated to occur in 2070. Seedling emergence was significantly influenced by population and origin of soil. Multivariate analyses produced a model which indentified pH, Na, salinity and PO4 as best explaining the patterns in seedling growth. We added three of these variables as spatial layers to the initial bioclimatic model to produce a new (climate and soil) model. The climate-only model over-predicted geographic areas with high suitability (60-100% suitability) in three of the four climate scenarios. These results illustrate the capacity to improve the precision of predictive bioclimatic models by incorporating experimental evidence to better represent the ecological preferences of species.


3:30 pm: Eveline Rijksen (@evelinerijksen), Ellery Room D

Dungeons and Dragons; does predation drive habitat choice and behaviour in desert reptiles?

Predation pressure, both actual and perceived, is thought to be a key driver of behaviour, habitat choice and population dynamics. It may drive closely related, sympatric species to display strikingly different behaviours to negotiate risks posed by predators. This mechanism helps to facilitate coexistence and high local diversity in diverse animal assemblages. The hyper-diverse lizard communities of central Australia provide an unusually good opportunity to identify the effects of predation on habitat use and behaviour.

In the spinifex covered sand-dunes of the north-eastern Simpson Desert, two sympatric species of agamid lizards, the military dragon (Ctenophorus isolepis) and the central netted dragon (C. nuchalis) occur abundantly.

However, an analysis of home range, habitat use and nocturnal behaviour shows that both species choose contrasting microhabitats within their home range. These contrasting habitats are closely related to the percentage of spinifex cover which in turn is associated with different predation risks. A further experiment imitating actual predation risk confirms that both reptile species incur different predation pressures from a suite of predators.

From this, we propose that predation indeed has a driving effect on divergence of behaviour and habitat use of same-clade species.


4:45 pm: Aaron Greenville (@AarontheEcolog), Ellery Room D

The web of arid life: biotic and abiotic interactions in a changing world.

Interactions are important in driving the composition and functioning of ecological assemblages and in maintaining diversity. How both biotic and abiotic interactions drive assemblages are fundamental questions in ecology, but complex systems with multiple species are difficult to study. Climate is a major abiotic driver for species in central Australia and the central desert regions are not immune from climate change, with higher temperatures and an increase in the frequency and magnitude of extreme rainfall events already recorded over the last 100 years. Wildfire return intervals are also predicted to decrease from climate change, making it imperative that we understand how both biotic and abiotic interactions shape ecological systems.

Here we use structural equation modelling to integrate remote camera trapping and live-trapping of vertebrates with long term (>15 years) vegetation data from the Simpson Desert to investigate interactions between the biota and with rainfall and fire. We then use these models to predict how changes in rainfall and wildfire events, in-line with future climate scenarios, will permeate up the trophic levels and interact with top-down effects from mammalian carnivores during both boom and bust resources periods in central Australia.


Thursday 2/10/2014

9:00 am: Chris Dickman, MacDonnell Room B/C

Keynote Address: Boom-bust cycles in arid environments: times of plenty, times of hazard

Episodic floods or heavy rains in arid environments usually herald pulses of productivity that trigger ‘booms’ in the numbers of consumer organisms and dramatic but short-lived increases in local and regional biodiversity. But these times of plenty can also be times of hazard, providing opportunities for invasive plants and animals to move into arid regions at the expense of native biota. As ‘bust’ times of low resources begin, wildfires often sweep through large areas and provide further opportunities and hazards for biota. In this talk, I describe long-term (20-25 year) trends in vegetation, lizards and small mammals that have been studied at sites in the Simpson Desert, central Australia, and show the responses of these groups to flooding rains, droughts and wildfire. In general, perennial vegetation cover varies from zero in the wake of wildfire to more than fifty percent after flood rains, lizards respond idiosyncratically to these events but show relative stability in their population dynamics; rodents irrupt and dasyurid marsupials decline after heavy rains, and all small mammals disappear after wildfire. Analysis of mammalian boom and bust cycles suggests that rodents act as single, large regional populations, whereas dasyurid dynamics respond to local factors. Populations of all small mammals are driven to low levels by invasive predators and face local extinction over large areas if these predators are present post-fire. Climate projections indicate that central Australia will become hotter in future, and that boom and bust cycles will intensify. These conditions will likely lead to more extensive predator incursions and more frequent and extensive wildfires, exacerbating the hazards faced currently by biota. I discuss management options that may help to mitigate these effects.


10:30 am: Glenda Wardle (@desert_ecology), Ellery Room B

Watching the grass grow: long-term dynamics of spinifex cover and biomass

Arid ecosystems dominated by the perennial hummock-forming grass, spinifex, extend over 60% of the continent. Fauna and flora alike depend on the cover and resources provided by spinifex yet we lack a deep understanding of the ecology of the plants that shape this uniquely Australian biome and how it might respond to increasing climate extremes. To establish a baseline for considering future change, we plot the long-term dynamics in cover of Triodia basedowii, the species that dominates the vegetation in the dune fields of the Simpson Desert and explore the relationship between estimates of percent cover and the size, shape, abundance and biomass of individuals of spinifex. After a wildfire, we mapped and measured 1600 Spinifex individuals in 240 5m x 5m vegetation plots in burnt and unburnt vegetation, in three dune zones and across local to regional scales. As expected, the abundance of Spinifex was significantly higher in unburnt vegetation (mean ±SD 10.6±6.3 per plot) compared to burnt vegetation (mean = 3.2±4.7). Spinifex was three-fold more abundant in swales compared to dune crests, and Spinifex numbers on the crests were highly variable (mean = 2.9±4.3). The biomass of spinifex was roughly equivalent to around 2.5kg for spherical clumps up to 1m in area, but slightly less for larger clumps transitioning into the classic ring shape, n=123. Spinifex cover is important as both habitat and for preventing dust storms, and therefore, retaining these features will be important for this major biome.


10:45 am: Vuong Nguyen, Ellery Room B

Time series analyses of spinifex cover using multivariate state-space models

Spinifex grasslands dominate the landscape of the Simpson Desert, playing a major role in the vegetation composition and providing a majority of the fuel load for fires to spread. We obtained long-term data on spinifex cover in two separate datasets: the first was conducted from 1995-2012 over nine sites and the second was conducted from 2004-2013 over four sites containing plots that were either burnt or unburnt from large-scale wildfires in 2002. Despite being of moderate length, both datasets contain numerous missing values due to the logistical and financial constraints of conducting field work in the Simpson Desert. Here, we use multivariate auto-regressive state-space (MARSS) models to overcome this limitation in the data and investigate long term temporal and spatial trends of spinifex in the Simpson Desert. MARSS models are a class of models that can use multiple time series from different sites and species to allow joint estimation of parameters and testing for spatial relationships between sites whilst accounting for observation error. The models from fixed plot data demonstrate that temporal variation in spinifex cover differs between the burnt and unburnt areas and that the burnt areas across sites all follow similar dynamics showing slow, steady increases in cover, compared to unburnt areas which are much more variable and inconsistent in their behaviour between sites. The implications of these findings are that even with incomplete time series data, having data of sufficient length from multiple sites enables reasonable interpretations of temporal and spatial patterns in population dynamics.


12:00 pm: Alan Kwok (@NaturesTextures), Ellery Room B

Patterns of invertebrate abundance over 6 years of sampling in central Australia

Globally, there is a glaring lack of baseline data on the long-term dynamics of invertebrate populations, despite the overwhelming numerical supremacy of invertebrates in terrestrial systems. Abundance information is crucial to improving our understanding of these biota and the ecosystem processes they are involved in, especially in arid ‘boom and bust’ systems where invertebrate populations could be expected to show marked fluctuations. Here, we present data on patterns of ground-dwelling invertebrate abundance from a 6-year period in the Simpson Desert, central Australia. This period (1999 – 2005) encompassed a range of climatic conditions, allowing investigation of faunal responses to several extreme rainfall events, as well as several years of below or average rainfall. Over 170,000 invertebrates were sampled from 32 invertebrate orders. Roughly 85% of all invertebrates were ants, with Collembola accounting for a further 10%. Acarina and Araneae were also sampled consistently, but in far lower numbers. There was substantial temporal variability in invertebrate abundance, with each taxon exhibiting different patterns over time. This variation was not necessarily related to resource pulses (rainfall and/or vegetative productivity). Indeed, few taxa showed consistent increases or decreases following rainfall events. These data clearly indicate that invertebrates, like other components of the biota, undergo large fluctuations in abundance over long time scales. However, the results also suggest that invertebrates may not respond to resource pulses in a similar fashion to other components of the biota, and hint that suites of interacting processes may play key roles.



Monday 29/9/2014, 5:30 pm, MacDonnell Room A and The Ghan Foyer:

PlantPopNet: a spatially distributed model system for population ecology

Glenda Wardle1, Yvonne Buckley2, and PlantPopNet Steering Group

1Desert Ecology Research Group and School of Biological Sciences, The University of Sydney, NSW 2006, Australia. 2Trinity College Dublin, Ireland

We rely on plant populations for health, wealth and nutrition, through the products and services they supply. Plants are an important basal component of terrestrial food webs and contribute directly and indirectly to biodiversity maintenance. Global change due to climate, invasions and shifting land-use practices is altering these vital systems and therefore knowing the drivers of population change demands our immediate attention.

Ecologists expect population shifts in response to global change; however, the data available for developing and testing movement and persistence models are spatially very limited. We could progress further and faster on this urgent problem if we could study many mapped populations and discern the mechanisms driving population change. Starting with Plantago lanceolata as a model system, we propose a co-ordinated effort to develop theory, supported by an awesome data set, on the abiotic and biotic drivers of population persistence and distribution. Here we invite participation in PlantPopNet a new globally distributed project on spatial plant population dynamics. Twitter @plantpopnet


Raising your research profile by publishing your data

Anita Smyth, Glenda Wardle and SHaRED

Publication of ecological datasets is becoming more common. This is because researchers recognise the benefits and because it is being mandated by a growing number of journals and granting bodies. Open data publication can raise researcher profiles by increasing personal citations thanks to indexing houses now publishing data citation metrics (e.g., Web of Science Data Citation Index). Other benefits include online data sourcing, new collaborations and more publishable units overall. Publishing data is also good science as it supplements data collection, shows multiple data perspectives, provides educational material and enables peer curation. Not all researchers are supportive of open data publication and even those that are have some concerns. These include the possibility of research being scooped, misuse of data and the uncertainty of legal protection.

SHaRED  is a data publication resource built specifically by ecological informaticians for Australian ecologists. It is a convenient and user friendly publication service that provides long term data security, maximises data reuse, utilises standard citation approaches and minimises data misuse within an Australian legal framework. Datasets submitted via SHaRED are stored securely in the Australian Ecological Knowledge and Observation System (ӔKOS) and are freely available online via its data portal. The conditions of dataset use are protected by the Australian Creative Commons 3.0 – By Attribution licence which is enforceable under Australian law. Dataset packages are published with a citable DOI which links researchers to data descriptions (metadata) and optional download.


The era of big data and big demography

Glenda Wardle1, Roberto Salguero-Gómez2,3 and Owen R. Jones4, 5

1Desert Ecology Research Group and School of Biological Sciences, The University of Sydney, NSW 2006, Australia. Email:

2Evolutionary Biodemography Laboratory, Max Planck Institute for Demographic Research. Konrad-Zuse-Straße 1, Rostock, DE-18057, Germany.

3ARC Centre of Excellence for Environmental Decisions, The University of Queensland, School of Biological Sciences, Queensland 4072, Australia.

4Max-Planck Odense Center on the Biodemography of Aging, University of Southern Denmark, Odense, Denmark.

5Department of Biology, University of Southern Denmark, Odense, Denmark.

Demography, the study of population dynamics and its underlying drivers, is central to the understanding of ecology and evolution. As the environment affects populations through its impacts on the vital rates of individuals (e.g., survival, growth, development, reproduction, dispersal), these traits provide the information to understand how populations are expected to change over time. By comparing among species we can determine how biomes shape plant and animal ecology, how populations and communities respond to global change, and how to develop successful management tools for endangered or invasive species.

Here we introduce two open access databases compiled from published matrix population models and new information collected from unpublished sources. The COMPADRE Plant Matrix Database version 3.0, launched mid 2014, is an open-source online repository containing 468 studies from 598 species worldwide with a total of 5,621 matrices. COMPADRE also contains relevant ancillary information (e.g. ecoregion, growth form, taxonomy, phylogeny, etc.) that facilitates interpretation of the numerous demographic metrics that can be derived from the matrices. The COMADRE Animal Matrix Database, scheduled for release in early 2015, contains around 1300 animal species.

Large collections of datasets allow broad questions to be addressed at the global scale, and we expect open access to this information, its frequent updates, and its integration with other online resources will allow researchers to address timely and important ecological and evolutionary questions.

Twitter @CompadreDB

Posters in other languages


ӔKOS – a next generation data resource for 21st century researchers

Anita Smyth, Glenda Wardle and AEKOS

Ecologists interact with data by collecting, collating, curating, analysing, modelling and illustrating it. Some ecologists are creators and users of their own data, some are users of others data (data collators) and some are combinations of the two. Data collators form strong data collaborations with data creators as each can bring substantial benefits to science outcomes. The workflow of these partnerships tend to be a one-to-many collaboration involving the steps of data request, follow-up, delivery, follow-up, assessment, collation, quality assurance and feedback. When working with many collaborators in different research organisations worldwide, unforeseen delays are the norm and it can take up to 70% of the total research effort to obtain suitable data. With advances in informatics and cloud computing, many opportunities now exist to increase research efficiencies. Online data portals that integrate high quality ecological data for thousands of sites from different datasets at a national level and which provide rich data descriptions are urgently needed to enhance data reuse potential and to keep pace with researcher needs.

The Australian Ecological Knowledge and Observation System (ÆKOS) is an online data service that provides free and open access to high quality, well-described, integrated ecological site data. You can currently access vegetation data for nearly 100,000 sites that has been integrated across 1,000 datasets throughout Australia. Animal data with mammals leading the way will be available in Jun-2015. In addition, ÆKOS enables access to datasets submitted via SHaRED.


Linking Long-Term Field Survey & Satellite Measurements of Vegetation Structure to Understand Ecosystem Dynamics

Stuart Phinn1, 2, Peter Scarth1,2, Sabrina (Dan) Wu1, Glenda Wardle 3, 4, Aaron Greenville3, 4, and Chris Dickman3, 4

1The University of Queensland, 2TERN Auscover,3The University of Sydney, 4TERN-Long Term Ecological Research Network

Contact –

This work completed a pilot project analysing 28 years of monthly vegetation cover maps derived from satellite images and concurrent annual ecosystem field survey data. The data were from long term monitoring plots in the hummocky grassland dominated, desert complexes of far western Queensland. Globally, satellite image archives spanning the past 20-30 years are being made publicly available at no cost, in formats that can be easily used for a number of applications. However, these archives lack two critical elements that enable their use in a much larger range of ecological and management activities: (1) direct linkage and validation with ecologically and management relevant environmental parameters; and (2) time-series analysis tools that work across spatially and temporally (collected every month for 30 years) dense data sets. We applied a continentally validated method for mapping the fraction of green or photo-synthetic vegetation, brown or non-photosynthetic vegetation and bare-ground, to derive a time series of image maps from 1988-2013. Time series of each cover fraction were extracted in monthly intervals within polygons corresponding to annual long term vegetation and faunal survey plots. Statistical analyses of the green, brown and bare ground fraction time series were compared to known spatio-temporal variations in field measured vegetation cover and fauna.  Clear patterns were evident in co-variations between satellite measured green/brown cover and field surveys of the same variables. This approach may enable the dynamics of local vertebrate and invertebrates populations to be related to direct environmental controls other than rainfall duration, but more closely related to food availability and habitat properties.

Twitter @desert_ecology