There is a discrepancy between leading ecological theory and the theory being applied in fields such as invasion biology. Given the urgent need for informed and effective ecosystem management, it is important to bridge this gap. Demographic stochasticity (arising from the probabilistic nature of core processes like births and deaths) has typically been dismissed as noise disguising an underlying deterministic skeleton and thus is often ignored in community ecology. Here, we provide a Bayesian framework to incorporate demographic stochasticity into phenomenological competition models and assess potential implications for invasion biology. This study included eleven commonly co-occurring native and exotic annual plant species from the York gum-jam woodlands in south-west Western Australia; a hyper-diverse annual plant community facing threat from exotic plant invasion. In a Bayesian framework, we developed plant fecundity models to estimate competition coefficient distributions (allowing for both competition and facilitation), drawn from Poisson distributions representing the underlying demographic stochasticity. The inclusion of demographic stochasticity provided the same answers as deterministic models, rather than disguising these patterns, but revealed flexibility within species’ responses. While all distributions of the effects of neighbours on focal species’ fecundity included some competitive effects, several distributions for native and invasive species also included facilitatory effects (largely from heterospecific neighbours). These distributions provide a quantification of the realistic variation in intrinsic fecundity and neighbour effects that species experience. Incorporating this variability into invasion biology will help capture the inherit unpredictability in community level response to, and success of biological invasions.

Freshwater ecosystems are valued by people and are one of the most threatened ecosystems on the planet. Environmental flows (e-flows) is a tool commonly used by managers to protect or repair riverine function and biodiversity. To create effective e-flow policy, water planners require information about flow-biota and/or flow-cultural relationships. Taking an eco-cultural approach to policy development is particularly important in systems where indigenous communities maintain strong links to the river. Today, ecologists are increasingly recognizing that functional or trait-based approaches provide a better insight into the risks posed by anthropogenic change compared to traditional species-based approaches. Surprisingly, the e-flow field has been slow to adopt a functional approach. This study advances the discipline of e-flows by transitioning away from flow-biota relationships that describe species-oriented patterns towards flow-eco-cultural relationships that describe functions and values. We use a multi-species Bayesian hierarchical model to link species-specific fish abundance (catch data) to ecological functions (traits) and indigenous values. The Bayesian model also allowed us to account for methodological complexities when studying fish (i.e. variable detection). We operationalise our results by predicting functions and values across gradients of water availability to reveal landscape-scale implications. Our study was undertaken in the Fitzroy River, a relatively pristine system in the Kimberley region of north-western Australia that is highly valued by local Indigenous communities, and is facing imminent water resource development. The findings of our study will feed directly into water planning to ensure sustainable development of this highly valued river.

Changes in natural flow regimes have contributed to declines in many freshwater fish species. Attempts to reverse these declines often focus on environmental watering, that is, the targeted delivery of water to restore fundamental population processes such as recruitment and survival. However, identifying effective water delivery strategies is complicated because we lack a clear understanding of how flow influences population dynamics. We sought to estimate two key vital rates—recruitment and survival—in Murray cod (Maccullochella peelii) and to relate these vital rates to flow. However, we lacked the data typically used to estimate vital rates, namely, data on marked individuals or individual ages. To overcome these limitations, we developed a hierarchical model with three levels. First, we used data on individual lengths to estimate ages, drawing on existing knowledge of Murray cod growth. Second, we used a catch-curve approach to infer recruitment and age-class abundances, accounting for differences in survival and detection among cohorts and river systems. Last, we related age-class abundances to a suite of flow metrics, testing several hypotheses about flow effects on different Murray cod life stages. We found highly variable associations between vital rates and flow, with differences in flow associations among rivers and age classes. These results highlight the importance of intra- and inter-population variation in flow associations in Murray cod, knowledge of which is critically important to the timing and delivery of environmental water.

Brachychiton sp. Ormeau (L.H.Bird AQ435851) is listed as critically endangered under the EPBC Act 1999 and endangered under the Queensland NCA 1992. The Ormeau Bottle Tree is a long-lived perennial that takes approximately 20 years to reach sexual maturity, with sporadic flowering and fruiting related to seasonal weather conditions.

At the time of listing (2013), the population comprised 161 mainly adult trees spread disproportionately across three subpopulations within a 6.5km2 area of the northern Darlington Range in the Upper Pimpama and Albert River catchments. Surveys in early 2019 by Healthy Land and Water as part of the Ormeau Bottle Tree project (funded by the Australian Government’s National Landcare Program) confirmed the persistence of known mature trees and located over 400 juvenile trees.

The distance of juvenile trees from known mature trees (up to 600m) and their placement often upslope and under large perch trees, e.g., Queensland Brush Box (Lophostomen confertus) suggests an aerial seed predator such as those reported for other Brachychiton species, e.g., crows, magpie, currawongs. Dispersal occurred mainly into ecotones or neighbouring dry sclerophyll forest indicating Brachychiton sp. Ormeau acts as a rainforest pioneer under suitable environmental conditions, in particular, the absence of fire.

This creates interesting management questions for land managers responsible for the three subpopulations. This presentation will explore the ecological implications of managing these sites for the persistence of the Ormeau Bottle Tree.

Associations between functional traits and environment are common. Yet putting those associations to use is less common. Predicting where species occur, rather than just explaining their distributions is a challenging problem. Trait-based species distribution models offer a means for incorporating functional information into correlative models. This might provide a middle ground between detailed species-specific mechanistic models and climate envelope models.
Here we fit models of 20 species eucalypt distributions along environmental gradients in Gariwerd (Grampian Mountains) of Victoria and predict to ~85 eucalypts in 18 other bioregions along the Great Dividing Range and surrounds south east of the Hume Fwy. We evaluate predictions within species, between regions. Then we predict only based on traits, first within Gariwerd, and then between the regions. We evaluate predictive performance and test whether predictive performance declines with geographic distance and with environmental distance and with compositional dissimilarity. We found transferability to be remarkably good considering we are doing this without knowledge of the species identity and only a limited set of traits. The transferability of trait only prediction was poor when compared to other studies testing within taxon transfer, but by no means overwhelmingly so. This suggests trait-based SDM can provide good first-order models for species responses along environmental gradients.

Species distribution models are widely used to forecast species responses to climate change and inform conservation decisions. Models are typically based on correlations between species occurrence data and environmental predictors, with underlying mechanisms captured only implicitly. However, there is a growing interest in approaches that explicitly model processes such as physiology, dispersal, demography, and biotic interactions that underpin species range shifts. A wide array of these process-explicit modelling approaches are now available, but it is not clear which methods are best for which types of problems or scenarios. In addition, data to parameterise and build these models are often limited, which could reduce their ability to accurately forecast range dynamics. To evaluate the promise of these emerging methods, we developed novel integrated models that link eco-physiological models with individual-based models, which can simulate responses of real species such as koalas to climate change and landscape dynamics. We coupled these models with a sampling module that recorded data required to fit different methods (e.g. vital rates, dispersal distances, population abundances) under different sampling regimes. Using these simulated datasets, we then tested the ability of different process-explicit distribution models such as occupancy dynamics models, coupled SDM-population models and demographic distribution models to capture and predict range dynamics. Using this simulation approach, we gain insight into what drives performance of different process-explicit methods, mechanisms that they do and do not capture and how to best allocate efforts to improve the reliability of range dynamics predictions under climate change.

Correlative species distribution models are good at describing species' current distributions, and inferring their environmental drivers. However they are pretty bad at predicting what will happen if we change something, like fragmenting habitat, introducing other species, or implementing a control or conservation action. Replacing the statistically convenient (but ecologically meaningless) internal structure of these models with demographic models should enable us to make much better predictions of how distributions will change.

I will present recent work developing demographic species distribution models that extend matrix population models to explicitly consider how vital rates vary through space (like spatial integral projection models) but are fitted to commonly available species distribution data (like dynamic range models). Combining these approaches enables us to fit ecologically-realistic species distribution models without the need for detailed demographic data. We can include density dependence, dispersal, biotic interactions and prior knowledge of species' population dynamics and ecology.

Asking for more information from the same data means we have to deal with a number of potential issues, including poorly-identified parameters and more computationally intensive statistical inference. I'll argue that non-identifiability is actually a good thing (in this context) and show how the computational issues can be resolved using the Bayesian inference package greta, and some new extensions for modelling dynamical systems.

With inadequate resources to manage the threats facing threatened species and ecological communities worldwide, achieving projected management outcomes is critical for efficient resource allocation and species recovery. Despite this, conservation plans to mitigate threats rarely articulate the likelihood of management success. We developed a general value of information approach to quantify the impact of uncertainty on 20 threatening processes affecting 976 listed species and communities in New South Wales, Australia. We discovered that uncertainty is a major impediment to the efficient management of threatened species. On average, removing uncertainty about management effectiveness could triple the gain in persistence achieved by managing under current uncertainty. Not all threatening processes were equally impacted by uncertainty. Management of fire, invasive animals and a plant pathogen were most impeded by uncertainty; management of invasive plants was least impacted. By providing the most comprehensive quantification of the impacts of uncertainty on threat management yet compiled, we emphasise the tremendous importance of reducing uncertainty about species responses to management, and show that failure to consider management effectiveness wastes resources and impedes species recovery.

Long-term pathogen control or eradication in wildlife is rare and represents a major challenge in conservation. Control is particularly difficult for environmentally transmitted pathogens. A treatment program was undertaken in Northern Tasmania that aimed to eradicate the environmentally transmitted Sarcoptes scabiei mite (causative agent of sarcoptic mange) from a local population of bare-nosed wombats (Vombatus ursinus) that was experiencing an epizootic event. Specifically, topical moxidectin (Cydectin) was administered to wombats via the use of remote treatment dosing stations (burrow flaps) that were replenished weekly for 12 weeks. This program temporarily halted the progression of the disease but ultimately the local population of wombats went extinct. A novel host-disease model was developed for this system in order to better understand why the program failed, and importantly, how it could be modified to increase success. The frequency that wombats switch burrows was determined to be an important positive driver of mite persistence. Sensitivity analyses indicated that treatment delivery success and duration of the program were the two most important contributors to treatment success and they act in a synergistic manner. This research emphasises the utility of applying model-guided management techniques in order to achieve practical solutions in the field.

Species distribution models (SDMs) are well-used tools in the study of flora and fauna, but their use for fungi is little-known. Yet, a breadth of SDM applications for fungi exists, thanks to growing availability of suitable data, and increased interest in knowing where to expect fungal species for better handling those that do harm and those that need conservation. In view of this new interest, one worthwhile question is: how should we rethink a baseline SDM, especially adapted for fungi?

Here, I present a review on existing SDM uses for fungi, gathered across a substantial range of topics, including agriculture and forestry, conservation, biogeography, and even human diseases. I show how the challenge of modelling fungi inspired some users to approach model building differently. Further, the special features of fungi – their smallness and their sensitivity to temporal conditions – affect how one best chooses species and environmental data, and how one best employs different modelling techniques. I also discuss the current state of fungal data, their opportunities for modelling and their caveats.

This overview is useful not only to those interested in fungi, but also for any ecologists wondering about how to adapt standard SDMs to specialised cases. I will conclude with a few words about my future PhD plans, for establishing sound methods for modelling distributions of Australian fungi.

Fire influences fauna by changing the distribution and abundance of resources available to them. Many structural resources (such as litter and shrub cover) are reduced by fire, and then re-accumulate over time in the post-fire environment. Consequently, time since fire (TSF) is often used as a surrogate for resource availability and faunal abundance. However, the value of TSF as a surrogate is species-specific, and may vary spatially. Determining the inter-relationships between TSF, structural resources and animal abundance is important for animal conservation in flammable ecosystems.

The heath mouse (Pseudomys shortridgei) is an endangered native Australian mammal which inhabits highly fire-prone heathlands. There are discrepancies in the literature regarding the species’ home range, resource preferences and response to fire. The objectives of my study are to measure the relationship between heath mouse abundance and a) TSF and b) resource availability, and c) to determine whether TSF is a useful surrogate for resources important to heath mice. We used Elliott traps and camera traps to survey heath mice at 40 sites spanning a 72-year chronosequence. Additionally, we measured resource availability by conducting habitat surveys. I will present preliminary results showing the relative influences of TSF and resource availability on heath mouse abundance. Determining the link between fire, resources and heath mouse abundance will reveal how this species interacts with their environment. Understanding this relationship can assist land managers in developing conservation strategies which will alleviate the threat of extinction of the heath mouse.