Biology subjects

Brook, B. W.

Publications and source records attributed to Brook, B. W..

16 recordsLinked to original sources

bioRxiv2025

Evaluating drone-mounted thermal infrared sensors for macropod monitoring in Tasmania

Thermal drones offer significant advantages for monitoring wildlife in low-light conditions; however, detection performance is influenced by technical settings and environmental factors. This study evaluates the use of nocturnal drone surveys to detect macropods in Tasmania, with the aim of determining optimal flight parameters for balancing detection accuracy and survey efficiency. Field surveys were conducted in Narawntapu National Park using varying flight altitudes and image overlap settings, and detection rates were compared across thermal datasets. Manual annotations and basic thresholding method were used to quantify detection success. Results showed that detection rates were highest when surveys were conducted under cooler ambient temperatures and at moderate altitudes (e.g. 60 metres AGL) with 50% image overlap. These findings provide practical guidance for designing nocturnal drone surveys and offers baseline recommendations for using drone-mounted thermal sensors to monitor large-bodied, crepuscular mammals, with broader implications for scalable wildlife monitoring programs.

ecology↗

bioRxiv2025

Automated detection of macropods in Tasmania using drone surveys and convolutional neural networks (CNNs)

Manual annotation of drone imagery is labour-intensive and prone to observer bias, particularly when applied to large datasets across varied environments. To address this, a deep-learning pipeline was developed and evaluated for identifying macropods in visual (RGB) drone imagery, using a convolutional neural network (CNN) adapted from the DeepForest framework. The model was trained on annotated images of Forester kangaroos (Macropus giganteus tasmaniensis) and Bennetts wallabies (Notamacropus rufogriseus) collected across two Tasmanian study sites. Performance was assessed using independent test sets from each site, representing open and forest-edge habitats, as well as a combined multi-site test set. Detection accuracy was quantified using precision, recall, and F1 scores, with further analyses evaluating the effect of solar altitude angle on model performance. The model achieved high recall across sites, indicating strong potential for minimising missed detections under diverse conditions. These results demonstrate the feasibility of applying transfer learning to drone-based wildlife surveys and highlight the promise of deep learning models for reducing manual effort in macropod monitoring, with applications for broader conservation and management workflows.

ecology↗

bioRxiv2023

Tree species richness and evenness affect forest biomass differently across biogeographic regions

The relationship between tree species diversity, measures of forest structure, and forest biomass has long been debated, with local- or continental-scale studies often finding contrasting results. Given the importance of forests as global carbon sinks, understanding the characteristics that underpin biomass accumulation is thus a critical component of mitigating climate change. Here we present a global analysis of 11,400 forest plots, sourced from scientific publications and forest inventories, to investigate the association of forest basal area (used as a proxy for biomass) with stem density and measures of tree species diversity (richness and evenness). We used generalised additive models to account for the confounding effects of climate and spatial signal and we modelled the density, climate, and diversity effects both globally and for each biogeographic region. Stem density showed a strong positive association with basal area across all biogeographic regions, while the effect of species richness varied. In the Palearctic, Nearctic, and Neotropical biogeographic regions, basal area was positively associated with species richness, although this was only detectable for lower values of basal area. In the Ethiopian and Oriental biogeographic regions there was no relationship between richness and basal area, while in the Australian biogeographic regions it was negative. The weak-to-no association between species evenness and basal area in all bioregions other than Australia suggests that the overall correlation emerges from processes operating at more local scales. Our results highlight the importance of accounting for biogeographic processes when evaluating strategies to mitigate climate change and support nature conservation.

ecology↗

bioRxiv2023

Land use change drives major loss of Southeast Asian biodiversity

Southeast Asia is highly biodiverse and currently experiences among the highest rates of tropical deforestation globally, but impacts on biodiversity are not well synthesized. We use Bayesian multi-level modeling to meta-analyse 831 pairwise comparisons of biodiversity in sites subject to human land use change and anthropogenic forest disturbance (for example in plantations or logged forest) versus undisturbed sites. After controlling for hierarchical dependences, we show that biodiversity is a fifth lower in sites with these land-use changes (95% credible interval= 16-28%, mean = 22%). This reduction was greater when land use change/anthropogenic forest disturbances were high-intensity (34% reduction in biodiversity) compared to low-intensity (18% reduction), and effects were consistent across biogeographic regions and taxa. Oil-palm plantations lead to the greatest reduction in biodiversity (39%, CI 27-48%), and agroforests the least (24%, CI 10-37%). We also find that biodiversity is reduced in young secondary forest by 26% (CI 4-42%) compared to undisturbed forest, but there is no reduction in biodiversity for intermediate or mature-aged secondary forest (although species composition is potentially altered). Overall, our study provides the clearest evidence yet of the substantial detrimental impact of land-use change and anthropogenic forest disturbance on the biodiversity of Southeast Asia.

ecology↗

bioRxiv2022

Landscape functional connectivity for butterflies under different scenarios of land-use, land-cover, and climate change in Australia

Pollinating invertebrates are vital to terrestrial ecosystems but are impacted by anthropogenic habitat loss/fragmentation and climate change. Conserving and improving landscape connectivity is important to offset those threats, yet its assessment for invertebrates is lacking. In this study, we evaluated the functional connectivity between protected areas in Australia for 59 butterfly species, under present conditions and different future scenarios (for 2050 and 2090) of land-use, land-cover, and climate change. Using circuit-theory analysis, we found that functional connectivity under present conditions varies widely between species, even when their estimated geographical ranges are similar. Under future scenarios, functional connectivity is predicted to decrease overall, with negative changes worsening from 2050 to 2090, although a few species are positive exceptions. We have made our results available as spatial datasets to allow comparisons with taxa from other studies and can be used to identify priority areas for conservation in terms of establishing ecological corridors or stepping-stone habitat patches. Our study highlights the importance of considering pollinating invertebrates when seeking holistic conservation and restoration of a landscapes functional connectivity, underscoring the need to expand and promote protected areas to facilitate functional connectivity under future scenarios of global change. Research DataThe habitat suitability maps and functional connectivity maps are made available as GeoTiff images via Figshare (10.6084/m9.figshare.19130078).

ecology↗

bioRxiv2021

Dominant carnivore loss benefits native avian and invasive mammalian scavengers

Scavenging by large carnivores is integral for ecosystem functioning by limiting the build-up of carrion and facilitating widespread energy flows. However, top carnivores have declined across the world, triggering trophic shifts within ecosystems. In this study, we use a natural removal experiment of disease-driven decline and island extirpation of native mammalian (marsupial) carnivores to investigate top-down control on utilisation of experimentally placed carcasses by two mesoscavengers - the invasive feral cat and native forest raven. Ravens were the main beneficiary of carnivore loss, scavenging for five times longer in the absence of native mammalian carnivores. Cats scavenged on almost half of all carcasses in the region without dominant native carnivores. This was eight times more than in areas where other carnivores were at high densities. In the absence of native mammalian carnivores, all carcasses persisted in the environment for 3 weeks. Our results reveal the efficiency of carrion consumption by mammalian scavengers. These services are not readily replaced by less-efficient facultative scavengers. This demonstrates the significance of global carnivore conservation and supports novel management approaches, such as rewilding in areas where the natural suite of carnivores is missing.

ecology↗

bioRxiv2021

Robust, data-driven bioregionalizations emerge from diversity concordance

AimDespite the increasing interest in developing new bioregionalizations and assessing the most widely accepted biogeographic frameworks, no study to date has sought to systematically define a system of small bioregions nested within larger ones that better reflect the distribution and patterns of biodiversity. Here, we examine how an algorithmic, data-driven model of diversity patterns can lead to an ecologically interpretable hierarchy of bioregions. LocationAustralia. Time periodPresent. Major taxa studiedTerrestrial vertebrates and vascular plants. MethodsWe compiled information on the biophysical characteristics and species occupancy of Australias geographic conservation units (bioregions). Then, using cluster analysis to identify groupings of bioregions representing optimal discrete-species areas, we evaluated what a hierarchical bioregionalization system would look like when based empirically on the within-and between-site diversity patterns across taxa. Within an information-analytical framework, we then assessed the degree to which the World Wildlife Funds (WWF) biomes and ecoregions and our suite of discrete-species areas are spatially associated and compared those results among bioregionalization scenarios. ResultsInformation on biodiversity patterns captured was moderate for WWFs biomes (50- 58% for birds beta, and plants alpha and beta diversity, of optimal discrete areas, respectively) and ecoregions (additional 4-25%). Our plants and vertebrate optimal areas retained more information on alpha and beta diversity across taxa, with the two algorithmically derived biogeographic scenarios sharing 86.5% of their within- and between-site diversity information. Notably, discrete-species areas for beta diversity were parsimonious with respect to those for alpha diversity. Main conclusionsNested systems of bioregions must systematically account for the variation of species diversity across taxa if biodiversity research and conservation action are to be most effective across multiple spatial or temporal planning scales. By demonstrating an algorithmic rather than subjective method for defining bioregionalizations using species-diversity concordances, which reliably reflects the distributional patterns of multiple taxa, this work offers a valuable new tool for systematic conservation planning.

ecology↗

bioRxiv2021

Predicted the impacts of climate change and extreme-weather events on the future distribution of fruit bats in Australia

AimFruit bats (Megachiroptera) are important pollinators and seed dispersers whose distribution might be affected by climate change and extreme-weather events. We assessed the potential impacts of those changes, particularly more frequent and intense heatwaves, and drought, on the future distribution of fruit bats in Australia. We also focus a case study on Tasmania, the southernmost island state of Australia, which is currently devoid of fruit bats but might serve as a future climate refugium. LocationAustralia (continental-scale study) and Tasmania. MethodsSpecies distribution modelling was used to predict the occurrence of seven species of fruit bats, using an ensemble of machine-learning algorithms. Predictors included extreme-weather events (heatwave and drought), vegetation (as a proxy for habitat) and bioclimatic variables. Predictions were made for the current-day distribution and future (2050 and 2070) scenarios using multiple emission scenarios and global circulation models. ResultsChanges in climate and extreme-weather events are forecasted to impact all fruit-bat species, with the loss and gain of suitable areas being predominantly along the periphery of a species current distribution. A higher emission scenario resulted in a higher loss of areas for Grey-headed flying fox (Pteropus poliocephalus) and Spectacled flying fox (P. conspicillatus) but a higher gain of areas for the Northern blossom bat (Macroglossus minimus). The Grey-headed flying fox (Pteropus poliocephalus) is the only study species predicted to potentially occur in Tasmania under future scenarios. Main conclusionsFruit bats are likely to respond to climate change and extreme weather by migrating to more suitable areas, including regions not historically inhabited by those species such as Tasmania--possibly leading to human-wildlife conflicts. Conservation strategies (e.g., habitat protection) should focus on areas we found to remain suitable under future scenarios, and not be limited by state-political boundaries.

ecology↗

bioRxiv2021

Hot, unpredictable weather interacts with land use to restrict the distribution of the Yellow-tailed Black-Cockatoo

Conserving nomadic species is challenging due to the difficulty in monitoring their characteristically transient populations, and thereby detecting range-wide declines. An example is the Yellow-tailed Black-Cockatoo (YTBC; Zanda funerea), which disperses widely in search of food and is regularly--but sporadically--observed across eastern Australia. Under climate warming, a general southward shift in species distributions is expected in the southern hemisphere, with the extreme southern margins being truncated by an ocean barrier. Given these constraints, we ask whether sufficient refugia will exist for the YTBC in the future, by: (i) modelling habitat relationships within current geographic range of the YTBC based on weather, climate, vegetation, and land use, and (ii) using this framework, coupled with climate-model projections, to forecast 21st century impacts. Intensive land use and high variability in temperature and rainfall seem to most limit YTBC occurrence. In contrast, areas with a cooler, stable climate, and a network of old-growth forests, such as occurs in parts of south-eastern Australia and Tasmania, are most suitable for the species. As Australia becomes progressively hotter under climate change, the preferred bioclimatic envelope of the YTBC is forecast to contract poleward (as a general pattern) and to fragment within the existing range. However, despite an extensive loss of climatically suitable regions, the YTBC might find stable refugia at the southern margins of its geographic range, although continued loss of old-growth forests undermines their nesting potential. Therefore, beyond habitat conservation, creating nesting opportunities within plantation forests would likely be an effective conservation strategy to preserve habitat quality in climate refugia.

ecology↗

bioRxiv2021

Roadkill islands: carnivore extinction shifts seasonal use of roadside carrion by generalist avian scavenger

O_LIGlobal road networks facilitate habitat modification and are integral to human expansion. Many animals, particularly scavengers, use roads as they provide a reliable source of food, such as carrion left after vehicle collisions. Tasmania is often cited as the roadkill capital of Australia, with the isolated offshore islands in the Bass Strait experiencing similar, if not higher, levels of roadkill. However, native mammalian predators on the islands are extirpated, meaning the remaining scavengers are likely to experience lower interference competition. C_LIO_LIIn this study, we use a naturally occurring experiment to examine how the loss of mammalian carnivores within a community impacts roadside foraging behaviour by avian scavengers. C_LIO_LIWe monitored the locations of roadkill and forest ravens (Corvus tasmanicus), an abundant scavenger species, on eight road transects across the Tasmanian mainland (high scavenging competition) and the Bass Strait islands (low scavenging competition). We represented raven observations as one-dimensional point patterns, using hierarchical Bayesian models to investigate the dependence of raven spatial intensity on habitat, season, distance to roadkill and route location. C_LIO_LIWe found that roadkill carcasses were a strong predictor of raven presence along road networks. The effect of roadkill was amplified on roads on the Bass Strait islands, where roadside carrion was a predictor of raven presence across the entire year. In contrast, ravens were more often associated with roadkill on Tasmanian mainland roads in the autumn, when other resources were low. This suggests that in the absence of competing mammalian scavengers, ravens choose to feed on roadside carrion throughout the year, even in seasons when other resources are available. This low interference competition could be disproportionately benefiting forest ravens, leading to augmented raven populations and changes to the vertebrate community structure. C_LIO_LIOur study provides evidence that scavengers modify their behaviour in response to reduced scavenger species diversity, potentially triggering trophic shifts and highlighting the importance of conserving or reintroducing carnivores within ecosystems. C_LI

ecology↗

bioRxiv2021

Humans hastened the range collapse and extinction of woolly mammoth

Pathways to extinction start long before the death of the last individual. However, causes of early-stage population declines and the susceptibility of small residual populations to extirpation are typically studied in isolation. Using validated process-explicit models, we disentangle the ecological mechanisms and threats that were integral in the initial decline and later extinction of the woolly mammoth. We show that reconciling ancient DNA data on woolly mammoth population decline with fossil evidence of location and timing of extinction requires process-explicit models with specific demographic and niche constraints, and a constrained synergy of climatic change and human impacts. Validated models needed humans to hasten climate-driven population declines by many millennia, and to allow woolly mammoths to persist in mainland Arctic refugia until the mid-Holocene. Our results show that the role of humans in the extinction dynamics of woolly mammoth began well before the Holocene, exerting lasting effects on the spatial pattern and timing of its range-wide extinction.

ecology↗

bioRxiv2021

Association between land cover, plant genera and pollinator dynamics in mixed-use landscapes

Pollinators are globally threatened by land-use change, but its effect varies depending on the taxa and the intensity of habitat degradation. However, pollinator-landscape studies typically focus on regions of intensive human activities and on a few focal species. Evaluating pollinator responses in landscapes with moderate land-use changes and on multiple pollinator groups would therefore fill an important knowledge gap. This study aims to determine the predictive capacity and effect of habitat characteristics on the relative abundance of multiple pollinator groups in mixed-use landscapes. To do this, we collected field data on the relative abundance of nectivorous birds, bees, beetles, and butterflies across the Tasman Peninsula (Tasmania, Australia). We then applied Random Forests to resolve the effects of land use (protected areas, plantation, and pasture), land cover at different radii (100 m and 2000 m), and plant genera on pollinator abundance. Overall, land cover and plant genera were more important predictors of pollinator abundance than land use. And the effect of land use, land cover, and plant genera varied depending on the pollinating group. Pollinator groups were associated with a range of plant genera, with the native genera Acacia, Leptospermum, Leucopogon, Melaleuca, Pomaderris, and Pultenaea being among the most important predictors. Our results highlight that one size does not fit all--that is pollinator response to different landscape characteristics vary, emphasise the importance of considering multiple habitat factors to manage and support a dynamic pollinator community, and demonstrates how land management can be informed using predictive modelling.

ecology↗

bioRxiv2021

Characterising the spatio-temporal threats, conservation hotspots, and conservation gaps for the most extinction-prone bird family (Aves: Rallidae)

With thousands of vertebrate species now threatened with extinction, there is an urgent need to understand and mitigate the causes of wildlife collapse. As distinct evolutionary clades can follow different routes to endangerment, there is value in taxon-specific analyses when assessing species vulnerability to threats and identifying gaps in conservation actions. Rails (Aves: Rallidae), being the most extinction-prone bird Family globally, and with one third of extant rail species now threatened or near-threatened, are an emphatic case in point. Yet even for this well-studied group, there is uncertainty in our understanding of what factors might be causing this vulnerability, whether the current threats are consistent with those that led to recent extinctions, and ultimately, what conservation actions might be necessary to mitigate further losses. Here, we undertook a global synthesis of the temporal and spatial threat patterns for Rallidae and determined conservation priorities and gaps. We found two key pathways in the threat pattern for rails. One follows the same trajectory as extinct rails, where island endemic and flightless rails are most threatened, mainly due to invasive predators. The second, created by the recent diversification of anthropogenic activities, involves continental rails (generally in the Neotropics), threatened most commonly by agriculture, natural-system modifications and residential and commercial development. Conservation efforts around most-at-risk species should be adapted according to the most relevant geographic scale (bioregions or countries), and principal locality type of the population (continental or island endemic). Indonesia, the U.S.A., the United Kingdom, New Zealand, and Cuba were the priority countries identified by our classification system incorporating species unique evolutionary features and level of endangerment, but also among the countries that lack conservation actions the most. Future efforts should predominantly target improvements in ecosystem protection and management, as well as ongoing research and monitoring. Forecasting the impacts of climate change on island endemic rails and disentangling the specific roles of extrinsic and intrinsic traits (like flightlessness), will be particularly valuable avenues of research for improving our forecasts of rail vulnerability.

ecology↗

bioRxiv2021

Predicting spatial and seasonal patterns of wildlife-vehicle collisions in high-risk areas

ContextVehicle collisions with wildlife can injure or kill animals, threaten human safety, and threaten the viability of rare species. This has led to a focus in road-ecology research on identifying the key predictors of road-kill risk, with the goal of guiding management to mitigate its impact. However, because of the complex and context-dependent nature of the causes of risk exposure, modelling road-kill data in ways that yield consistent recommendations has proven challenging. AimHere we used a novel multi-model machine-learning approach to identify the spatio-temporal predictors, such as traffic volume, road shape, surrounding vegetation and distance to human settlements, associated with road-kill risk. MethodsWe collected data on the location, identity and size of each road mortality across four seasons along eight roads in southern Tasmania - a road-kill hotspot of management concern. We focused on three large-bodied and frequently impacted crepuscular Australian marsupial herbivore species, the rufous-bellied pademelon (Thylogale billardierii), Bennetts wallaby (Macropus rufogriseus) and the bare-nosed wombat (Vombatus ursinus). We fit the point-location data using lasso-regularization of a logistic generalized linear model (LL-GLM) and out-of-bag optimization of a decision-tree-based random forests (RF) algorithm. ResultsThe RF model, with high-level feature interactions, yielded superior results to the linear additive model, with a RF classification accuracy of 84.8% for the 871 road-kill observations and a true skill statistic of 0.708, compared to 61.2% and 0.205 for the LL-GLM. ConclusionsForested areas with no roadside barrier fence along curved sections of road posed the highest risk to animals. Seasonally, the frequency of wildlife-vehicle collisions increased notably for females during oestrus, when they were more dispersive and so had a higher encounter rate with roads. ImplicationsThese findings illustrate the value of using data-driven approaches to predictive modelling, as well as offering a guide to practical management interventions that can mitigate road-related hazards.

ecology↗

bioRxiv2021

Spatial pattern analysis of line-segment data in ecology

O_LIThe spatial analysis of linear features (lines and curves) is a challenging and rarely attempted problem in ecology. Existing methods are typically expressed in abstract mathematical formalism, making it difficult to assess their relevance and transferability into an ecological setting. A set of concrete and accessible tools is needed. C_LIO_LIWe develop a new method to analyse the spatial patterning of line-segment data. It is based on a generalisation of Ripleys K-function and includes an analogue of the transformed L-function, together with estimators and theoretical expectation values. We introduce a class of line-segment processes, related to the Boolean model, which we use in conjunction with Monte-Carlo methods and information criteria to generate and compare candidate models. We demonstrate the utility of our method using fallen tree (dead log) data collected from two one-hectare Australian tall eucalypt forest plots. C_LIO_LIComparing six line-segment models, we find for both plots that the distribution of fallen logs is best explained by plot-level spatial heterogeneity. The use of non-uniform distributions to model dead-log orientation on the forest floor improves model performance in one of the two sites. Our case study highlights the challenges of model comparison in spatial-pattern analysis, where Monte-Carlo approaches based on the discrepancy of simulated summary functions can generate a different ranking of models than that of information criteria. C_LIO_LIThese methods are of a general nature and are applicable to any line-segment data. In the context of forest ecology, the integration of fallen logs as linear structural features in a landscape with the point locations of living trees, and a quantification of their interactions, will yield new insights into the functional and structural role of tree fall in forest communities and their enduring post-mortem ecological legacy as spatially distributed decomposing logs. C_LI

ecology↗

bioRxiv2021

Extinction of the Thylacine

Like the Dodo and Passenger Pigeon before it, the predatory marsupial Thylacine (Thylacinus cynocephalus), or Tasmanian tiger, has become an iconic symbol of human-caused extinction. The last captive animal died in 1936, but even today reports of the Thylacines possible ongoing survival in remote regions of Tasmania are newsworthy and capture the publics imagination. Extirpated from mainland Australia in the mid-Holocene, the large island of Tasmania became the species final stronghold. Following European settlement in the 1800s, the Thylacine was heavily persecuted and pushed to the margins of its range, although many sightings were reported thereafter--even well beyond the 1930s. To gain a new depth of insight into the extinction of the Thylacine, we assembled an exhaustive database of 1,237 observational records from Tasmania (from 1910 onwards), quantified their uncertainty, and charted the patterns these revealed. We also developed a new method to visualize the species 20th-century spatio-temporal dynamics, to map potential post-bounty refugia and pinpoint the most-likely location of the final persisting subpopulation. A direct reading of the high-quality records (confirmed kills and captures, in combination with sightings by past Thylacine hunters and trappers, wildlife professionals and experienced bushmen) implies a most-likely extinction date within four decades following the last capture (i.e., 1940s to 1970s). However, uncertainty modelling of the entire sighting record, where each observation is assigned a probability and the whole dataset is then subject to a sensitivity analysis, suggests that extinction might have been as recent as the late 1980s to early 2000s, with a small chance of persistence in the remote south-western wilderness areas. Beyond the intrinsically fascinating problem of reconstructing the final fate of the Thylacine, the new spatio-temporal mapping of extirpation developed herein would also be useful for conservation prioritization and search efforts for other rare taxa of uncertain status.

zoology↗