Dr Darren Southwell

Monitoring trends in animal populations in arid regions is challenging due to remoteness and low population densities. However, detecting species' tracks or signs is an effective survey technique for monitoring population trends across large spatial and temporal scales. In this study, we developed a simulation framework to evaluate the performance of alternative track-based monitoring designs at detecting change in species distributions in arid Australia. We collated presence¿absence records from 550 2-ha track-based plots for 11 vertebrates over 13¿years and fitted ensemble species distribution models to predict occupancy in 2018. We simulated plausible changes in species' distributions over the next 15¿years and, with estimates of detectability, simulated monitoring to evaluate the statistical power of three alternative monitoring scenarios: (1) where surveys were restricted to existing 2-ha plots, (2) where surveys were optimized to target all species equally, and (3) where surveys were optimized to target two species of conservation concern. Across all monitoring designs and scenarios, we found that power was higher when detecting increasing occupancy trends compared to decreasing trends owing to the relatively low levels of initial occupancy. Our results suggest that surveying 200 of the existing plots annually (with a small subset resurveyed twice within a year) will have at least an 80% chance of detecting 30% declines in occupancy for four of the five invasive species modeled and one of the six native species. This increased to 10 of the 11 species assuming larger (50%) declines. When plots were positioned to target all species equally, power improved slightly for most compared to the existing survey network. When plots were positioned to target two species of conservation concern (crest-tailed mulgara and dusky hopping mouse), power to detect 30% declines increased by 29% and 31% for these species, respectively, at the cost of reduced power for the remaining species. The effect of varying survey frequency depended on its trade-off with the number of sites sampled and requires further consideration. Nonetheless, our research suggests that track-based surveying is an effective and logistically feasible approach to monitoring broad-scale occupancy trends in desert species with both widespread and restricted distributions.

Catastrophic megafires can increase extinction risks; identifying species priorities for management and policy support is critical for preparing and responding to future fires. However, empirical data on population loss and recovery post-fire, especially megafire, are limited and taxonomically biased. These gaps could be bridged if species' morphological, behavioural, ecological and life history traits indicated their fire responses. Using expert elicitation that estimated population changes following the 2019¿20 Australian megafires for 142 terrestrial and aquatic animal species (from every vertebrate class, one invertebrate group), we examined whether expert estimates of fire-related mortality, mortality in the year post-fire, and recovery trajectories over 10 years/three generations post-fire, were related to species traits. Expert estimates for fire-related mortality were lower for species that could potentially flee or shelter from fire, and that associated with fire-prone habitats. Post-fire mortality estimates were linked to diet, diet specialisation, home range size, and susceptibility to introduced herbivores that damage or compete for resources. Longer-term population recovery estimates were linked to diet/habitat specialisation, susceptibility to introduced species; species with slower life histories and shorter subadult dispersal distances also had lower recovery estimates. Across animal groups, experts estimated that recovery was poorest for species with pre-fire population decline and more threatened conservation status. Sustained management is likely needed to recover species with habitat and diet specialisations, slower life histories, pre-existing declines and threatened conservation statuses. This study shows that traits could help inform management priorities before and after future megafires, but further empirical data on animal fire response is essential.

Aim: After environmental disasters, species with large population losses may need urgent protection to prevent extinction and support recovery. Following the 2019¿2020 Australian megafires, we estimated population losses and recovery in fire-affected fauna, to inform conservation status assessments and management. Location: Temperate and subtropical Australia. Time period: 2019¿2030 and beyond. Major taxa: Australian terrestrial and freshwater vertebrates; one invertebrate group. Methods: From >¿1,050 fire-affected taxa, we selected 173 whose distributions substantially overlapped the fire extent. We estimated the proportion of each taxon¿s distribution affected by fires, using fire severity and aquatic impact mapping, and new distribution mapping. Using expert elicitation informed by evidence of responses to previous wildfires, we estimated local population responses to fires of varying severity. We combined the spatial and elicitation data to estimate overall population loss and recovery trajectories, and thus indicate potential eligibility for listing as threatened, or uplisting, under Australian legislation. Results: We estimate that the 2019¿2020 Australian megafires caused, or contributed to, population declines that make 70¿82 taxa eligible for listing as threatened; and another 21¿27 taxa eligible for uplisting. If so-listed, this represents a 22¿26% increase in Australian statutory lists of threatened terrestrial and freshwater vertebrates and spiny crayfish, and uplisting for 8¿10% of threatened taxa. Such changes would cause an abrupt worsening of underlying trajectories in vertebrates, as measured by Red List Indices. We predict that 54¿88% of 173 assessed taxa will not recover to pre-fire population size within 10¿years/three generations. Main conclusions: We suggest the 2019¿2020 Australian megafires have worsened the conservation prospects for many species. Of the 91 taxa recommended for listing/uplisting consideration, 84 are now under formal review through national processes. Improving predictions about taxon vulnerability with empirical data on population responses, reducing the likelihood of future catastrophic events and mitigating their impacts on biodiversity, are critical.

Aim: Megafire plays a crucial role in driving the distribution of biodiversity around the world. Long-term monitoring is vital for understanding how species are impacted immediately by megafire and subsequently respond over time. However, monitoring should be designed with sufficient statistical power to detect impact and recovery. In this study, we developed a simulation framework for optimizing the design of biodiversity monitoring programmes to detect population recoveries after megafire. Location: Victoria, Australia. Time period: 2019¿2020. Major taxa studied: Vertebrates. Methods: We collated species distribution models for 45 priority vertebrates most likely to respond to management after the 2019¿2020 megafires in Victoria, Australia. We combined these models with fire severity maps to optimize the location of monitoring sites in and around the fire footprint. We simulated the impact of the megafires on species distributions and modelled plausible recoveries over the next 10 years. Using estimates of detectability for a suite of preferred sampling methods, we simulated monitoring at pairs of burnt and unburnt sites to evaluate the statistical power to detect the modelled recoveries. We tested the sensitivity of power to alternative monitoring designs, rates of recovery and monitoring budgets. Results: Priority regions to establish monitoring sites varied by taxonomic group. Power to detect population recoveries increased as the monitoring budget increased, as the recovery rate increased and when the proportion of sites in burnt compared with unburnt habitat increased. According to the optimal monitoring design, an AUD $9M budget could detect 90% of recoveries to pre-fire levels in 40% of species with >80% power. Power was highest for mammals, followed by birds, reptiles and amphibians. Main conclusions: Our simulation approach allowed us to test the relative performance of alternative post-fire monitoring designs ahead of time. Although we focused on megafire, our approach could easily be applied to detect population recoveries after any large-scale catastrophic disturbance.

In the summer of 2019-2020, southern Australia experienced the largest fires on record, detrimentally impacting the habitat of native species, many of which were already threatened by past and current anthropogenic land use. A large-scale restoration effort to improve degraded species habitat would provide fire-affected species with the chance to recover and persist in burnt and unburnt habitat. To facilitate this, decision-makers require information on priority species needs for restoration intervention, the suite of potential restoration interventions, and the priority locations for applying these interventions. We prioritize actions in areas where restoration would most likely provide cost-effective benefits to priority species (defined by each species proportion of habitat burned, threat status, and vulnerability to fires), by integrating current and future species habitat suitability maps with spatially modelled costs of restoration interventions such as replanting, removing invasive species, and implementing ecologically appropriate fire management. We show that restoring the top ~69% (112 million hectares) of the study region (current and future distributions of priority species) accounts for, on average, 95% of current and future habitat for every priority species and costs ~AUD$73 billion yr-1 (AUD$650 hectare-1 yr-1) annualized over 30 years. This effort would include restoration actions over 6 million hectares of fire-impacted habitat, costing ~AUD$8.8 billion/year. Large scale restoration efforts are often costly but can have significant societal co-benefits beyond biodiversity conservation. We also show that up to 291 MtCO2 (~150 Mt DM) of carbon could be sequestered by restoration efforts, resulting in approximately AUD$253 million yr-1 in carbon market revenue if all carbon was remunerated. Our approach highlights the scale, costs, and benefits of targeted restoration activities both inside and outside of the immediate bushfire footprint over vast areas of different land tenures.

Large-scale disturbance events are forecast to increase in severity and frequency due to climate change. On-ground surveys are crucial for assessing the immediate impact of disturbances on biodiversity and for informing management responses. However, there are few examples where quantitative tools have guided post-disturbance survey design. In this study, we integrated species distribution modelling and spatial prioritisation to identify taxonomic and spatial gaps in surveys for 92 priority vertebrates 6 months after the 2019¿20 wildfires in Australia. We predicted the pre-fire distribution of priority species, mapped locations of post-wildfire surveys that were already underway, and integrated this information with remotely-sensed fire severity maps in the tool, Zonation, to prioritise locations for new surveys across three fire severity classes (unburnt, low severity, high severity). Our results suggest that 6 months after the wildfires, surveys by government agencies had targeted 17 of 20 mammals (85%); 11 of 17 birds (65%); 10 of 17 frogs (59%); 10 of 23 reptiles (43%) and 5 of 17 fish (29%). We developed species distribution models for 63 of these species after collating 120,118 occurrence records from 6 data repositories. By predicting their distribution before the wildfires, we most efficiently identified gaps in survey effort while ensuring representation across species and fire severity classes. Our analysis provided an important ¿stocktake¿ of the response effort to the 2019¿20 wildfires in Australia and helped inform the allocation of government-funded wildfire recovery programs. Although we focus on wildfire, our approach could assess gaps in survey effort following any large-scale disturbance.

Several methods of measuring biodiversity in development-offset trades exist. However, there is little consensus on which biodiversity metrics should be used for quantifying development impacts and assigning offsets. We simulated development impacts in a virtual landscape and offset these impacts using six biodiversity metrics: vegetation area, vegetation condition, habitat suitability, species abundance, metapopulation connectivity and rarity-weighted richness. We tested long-term impacts of metric choice during offsetting by combining simulated landscapes with population viability analyses. No net loss or net gains in habitat were achieved using all metrics except vegetation area and condition. Limited habitat and like-for-like requirements resulted in offsets exhausting available habitat in each vegetation class before offset requirements were met when using vegetation-based metrics. We also found that impact avoidance was an important driver in how much compensation offsets could deliver. When impacts avoided high-suitability habitats, all six metrics achieved no net loss or net gains for most species. However, when core habitats were developed, none of the metrics were able to consistently prevent population declines. Synthesis and application. When impacts on high-quality habitat were avoided, and assuming the protection and restoration benefits can occur in practice, vegetation-based metrics may produce offsets which deliver gains in species abundance equivalent to species-specific metrics. However, species-specific metrics outperformed vegetation-based metrics when core habitats were lost. Applying avoidance measures as a first step to minimise biodiversity impacts during development will significantly improve offset outcomes for species and result in greater long-term population benefits delivered through offsetting.

Refuges play a critical role protecting species against the effects of climate change. Managing high priority refuges could improve species resilience and facilitate dispersal during periods of environmental change. In this study, we identified drought refuges in semi-arid Australia for a threatened bird, the malleefowl Leipoa ocellata. Using a Poisson regression model, we quantified the effect of remotely sensed vegetation indices, rainfall, soil moisture and site characteristics on malleefowl breeding activity at 144 sites surveyed from 2000 to 2017 during and after drought. We tested the effect of two vegetation productivity indices on malleefowl breeding activity ¿ the Normalized Difference Vegetation Index (NDVI) and the Enhanced Vegetation Index (EVI) ¿ averaged across three temporal and three spatial scales during the mound building and incubation stage of the breeding cycle. We found that NDVI and EVI were better predictors of malleefowl breeding activity than soil moisture and winter rainfall. The model with the lowest Akaike information criterion value contained NDVI averaged over 3 months (June¿August) and a 1-km radius. Malleefowl breeding activity had a strong positive association with NDVI (0.42 ± se 0.03) and a negative association with slope (-0.34 ± se 0.03) and vegetation patch size (-0.23 ± se 0.02). We found the proportion of refugia (top 20% of predicted breeding activity) in protected areas was highly variable, decreasing from 42% in an extreme wet year (2011) to 14% in an extreme drought year (2007). Expanding the reserve network to include refugia predicted to occur in the south of semi-arid Victoria could improve resilience of malleefowl to climate change. We demonstrate how remotely sensed vegetation indices combined with citizen science data can identify where to protect native vegetation with high, stable productivity. Our approach could be applied to a broad range of species in semi-arid regions vulnerable to climate change.

Aim: Fires can severely impact aquatic fauna, especially when attributes of soil, topography, fire severity and post-fire rainfall interact to cause substantial sedimentation. Such events can cause immediate mortality and longer-term changes in food resources and habitat structure. Approaches for estimating fire impacts on terrestrial species (e.g. intersecting fire extent with species distributions) are inappropriate for aquatic species as sedimentation can carry well downstream of the fire extent, and occur long after fire. Here, we develop an approach for estimating the spatial extent of fire impacts for aquatic systems, across multiple catchments. Location: Southern Australian bioregions affected by the fires in 2019¿2020 that burned >10 million ha of temperate and subtropical forests. Methods: We integrated an existing soil erosion model with fire severity mapping and rainfall data to estimate the spatial extent of post-fire sedimentation threat in waterways and in basins and the potential exposure of aquatic species to this threat. We validated the model against field observations of sedimentation events after the 2019¿20 fires. Results: While fires overlapped with ~27,643 km of waterways, post-fire sedimentation events potentially occurred across ~40,449 km. In total, 55% (n¿= 85) of 154 basins in the study region may have experienced substantial post-fire sedimentation. Ten species¿including six Critically Endangered¿were threatened by post-fire sedimentation events across 100% of their range. The model increased the estimates for potential impact, compared to considering fire extent alone, for >80% of aquatic species. Some species had distributions that did not overlap with the fire extent, but that were entirely exposed to post-fire sedimentation threat. Conclusions: Compared with estimating the overlap of fire extent with species' ranges, our model improves estimates of fire-related threats to aquatic fauna by capturing the complexities of fire impacts on hydrological systems. The model provides a method for quickly estimating post-fire sedimentation threat after future fires in any fire-prone region, thus potentially improving conservation assessments and informing emergency management interventions.

Monitoring of threatened species is a critical part of conserving biodiversity. It is needed to understand population trajectories, threatening processes, and the type and effectiveness of management responses needed to ensure persistence and recovery. Characteristics of some plant species (e.g. immobility) should render them amenable to monitoring, whereas other characteristics (e.g. ephemeral life histories) will make plant monitoring challenging. We evaluated monitoring adequacy and extent for a large sample (839 taxa) of Australia's threatened plants (1336 taxa) and compared this assessment with a similar evaluation for threatened vertebrates. We found 37.2% of threatened plants are monitored, half the rate found for vertebrates. For monitored plants, monitoring quality as assessed using a set of nine criteria was generally low, similar to results for vertebrates. Plants with more imperilled conservation status were more likely to be monitored and tended to have higher quality monitoring. Plants with recovery plans were more likely to be monitored than those without. The likelihood a species was monitored decreased the longer a taxon had been listed under threatened species legislation. Monitoring longevity was poor but inclusion of demographic data and linkages to management were better than for vertebrates. Our assessment highlighted a lack of collated monitoring data for plants, and we recognise there are exemplary programs for threatened plants that can guide improvements in monitoring for other species. Plants are overwhelmingly represented in threatened species lists worldwide and a determined focus to improve the extent and quality of plant monitoring should underpin biodiversity conservation targets.

Rock-wallabies (Petrogale spp.) are one of Australia¿s most speciose genera of mammals, irregularly distributed across much of the continent and its offshore islands. The 25 taxa in the genus Petrogale (17 species and 8 subspecies) have specialised ecological requirements that render them¿vulnerable to numerous threats. Many rock-wallaby populations have declined severely, and most species and subspecies are experiencing ongoing declines in population size, distribution and their conservation status. Despite an explicit recognition of the need for conservation management, some species are not monitored and a consensus on the most appropriate methods for ongoing population monitoring has proven elusive. We reviewed the available literature to understand the conservation issues and threats most relevant to Petrogale spp. We also reviewed rock-wallaby monitoring programs with the aim of identifying which are most informative of population trends, and most suitable for guiding better management responses. Major threats to rock-wallabies include predation by introduced cats and foxes, competition from introduced herbivores and overabundant native herbivores, changed fire regimes and loss of genetic diversity. There are synergisms that exacerbate these threats. While live-trapping gives comprehensive population data, camera traps have proven popular for collecting data over long periods, have minimal animal welfare impacts, and can simultaneously collect data on some significant co-occurring threats (feral predators and herbivores). A variety of rock-wallaby monitoring programs are current in Australia, but few adequately provide the range of data necessary for informed conservation. Monitoring programs should consider incorporating multiple methods to ensure the range of information necessary for successfully conserving rock-wallabies is obtained.

Biodiversity offsets are increasingly employed as an approach to compensate for unavoidable development impacts. Reliance on overly simplistic metrics in assessing the impacts of development, and assigning offset requirements, generally results in offsets which fail to conserve the key ecological values they seek to protect. We conducted a cross-disciplinary quantitative review, based on 255 peer-reviewed publications from three fields of research; offsetting (n = 43), conservation planning (n = 54) and ecology (n = 158), to explore which metrics are commonly used in offsetting compared to the conservation and ecology literature. We recorded the use of biodiversity metrics from 24 categories which captured broad habitat patterns (e.g. habitat area and condition) as well as specific biological and ecological mechanisms (e.g. diversity, population density or landscape connectivity). Our review found that offset studies and programs rely heavily on habitat attributes and area-based metrics, with >70% of the offset literature having used these metrics. Habitat attributes and area-based metrics were less frequently reported in the conservation planning (56 and 59%, respectively) and ecological literature (49 and 15%). Ecological research had a higher frequency of metrics reflecting the biological and ecological processes relevant to biodiversity, such as species¿ population densities and species-specific connectivity. Our results also indicate a notable disconnect in how biodiversity is measured when offsets are planned compared to when their outcomes are evaluated. This demonstrates the need to re-evaluate the way offset policies and programs value, describe and measure biodiversity, so that critical biodiversity values and important ecological processes are appropriately captured, and no net loss is achieved.

Citizen scientists regularly collect monitoring data for threatened species to improve the spatial and temporal resolution of sampling. Such programs should adopt robust data assurance measures and statistical approaches to reduce observer bias and better inform uncertainty estimates while supporting management decisions. In this study, we estimated trends and drivers of malleefowl (Leipoa ocellata) breeding activity within a Bayesian hierarchical modelling framework using 1823 site × years of nest count data collected by volunteers in Australia. Our modelling suggests malleefowl breeding activity decreased 4.8% annually in South Australia (-0.050; 95%CIs -0.062, -0.037), decreased 2.1% annually in Western Australia (-0.022; 95%CI -0.040, -0.004), was stable in Victoria (-0.001; 95%CI -0.010, 0.009) and increased 4.8% annually in New South Wales (0.047; 95%CI 0.009, 0.086). We found strong evidence for positive associations between winter rainfall (0.084; 95%CI 0.004, 0.165), time since fire (0.288; 95%CI 0.179, 0.399) and an interaction between time since fire and the proportion of a site burnt (0.292; 95%CI 0.173, 0.410). Malleefowl breeding activity was negatively associated with patch size (-0.255; 95% CI -0.642, 0.020) and the proportion of a site burnt (-0.191; 95% CI -0.363, -0.030), suggesting small reserves are important for conservation and the extent and frequency of fire should be managed cautiously. While our index of fox abundance decreased as baiting effort increased (-0.484; 95%CI -0.640, -0.317), there was little evidence for this benefiting malleefowl. This study demonstrates how volunteers can play a vital role understanding population trends and informing conservation of a threatened species at a national scale.

Monitoring of threatened species and threatened ecosystems is critical for determining population trends, identifying urgency of management responses, and assessing the efficacy of management interventions. Yet many threatened species and threatened ecosystems are not monitored and for those that are, the quality of the monitoring is often poor. Here we provide a checklist of factors that need to be considered for inclusion in robust monitoring programs for threatened species and threatened ecosystems. These factors can be grouped under four broad themes ¿ the design of monitoring programs, the structure and governance of monitoring programs, data management and reporting, and appropriate funding and legislative support. We briefly discuss key attributes of our checklist under these themes. Key topics in our first theme of the design of monitoring programs include appropriate objective setting, identification of the most appropriate entities to be measured, consistency in methodology and protocols through time, ensuring monitoring is long-term, and embedding monitoring into management. Under our second theme which focuses on the structure and governance of monitoring programs for threatened species and ecosystems, we touch on the importance of adopting monitoring programs that: test the effectiveness of management interventions, produce results that are relevant to management, and engage with (and are accepted by) the community. Under Theme 3, we discuss why data management is critical and highlight that the costs of data curation, analysis and reporting need to be factored into budgets for monitoring programs. This requires that appropriate levels of funding are made available for monitoring programs, beyond just the cost of data collection ¿ a key topic examined in Theme 4. We provide examples, often from Australia, to highlight the importance of each of the four themes. We recognize that these themes and topics in our checklist are often closely inter-related and therefore provide a conceptual model highlighting these linkages. We suggest that our checklist can help identify the parts of existing monitoring programs for threatened species and threatened ecosystems that are adequate for the purpose or may be deficient and need to be improved.

Monitoring is an essential component of adaptive management, and a carefully designed program is needed to ensure high-quality data and inferences over realistic time scales. Co-operation among agencies and incorporating citizen science may help enhance learning while reducing the financial costs of monitoring. We seek to realise this potential while conserving the Australian malleefowl (Leipoa ocellata). An established network of citizen scientists provide low-cost, sustainable annual monitoring data, yet the most effective actions for conserving malleefowl remain highly uncertain. The continent-wide species' distribution presents significant challenges, including multiple environmental strata to sample and numerous management jurisdictions. We outline an adaptive management framework that aims to unify malleefowl conservation priorities nationally, and target monitoring efforts. We elicited a model structure for the drivers of, and threats to, malleefowl persistence in a workshop with land managers and advocates. We parameterised 80 uncertain interactions within this structure using novel ensemble modelling techniques and identified the effectiveness of predator control as a critical uncertainty affecting malleefowl persistence. We developed a classical, spatially replicated experimental design to test whether malleefowl breed more frequently where predators are suppressed. The proposed monitoring design will rely on the contributions of several dozen land managers and 200¿300 citizen scientists annually. We have developed a broad stakeholder base, a proactive communication strategy, and an agile approach to accessing resources to foster resilience and longevity in the monitoring program. If malleefowl conservation successfully adapts in response to monitoring outcomes, it will become one of the largest adaptive management programs on the planet.

Assessing the statistical power to detect changes in wildlife populations is a crucial yet often overlooked step when designing and evaluating monitoring programs. Here, we developed a simulation framework to perform spatially explicit statistical power analysis of biological monitoring programs for detecting temporal trends in occupancy for multiple species. Using raster layers representing the spatial variation in current occupancy and species-level detectability for one or multiple observation methods, our framework simulates changes in occupancy over space and time, with the capacity to explicitly model stochastic disturbances at monitoring sites (i.e., dynamic landscapes). Once users specify the number and location of sites, the frequency and duration of surveys, and the type of detection method(s) for each species, our framework estimates power to detect occupancy trends, both across the landscape and/or within nested management units. As a case study, we evaluated the power of a long-term monitoring program to detect trends in occupancy for 136 species (83 birds, 33 reptiles, and 20 mammals) across and within Kakadu, Litchfield, and Nitmiluk National Parks in northern Australia. We assumed continuation of an original monitoring design implemented since 1996, with the addition of camera trapping. As expected, power to detect trends was sensitive to the direction and magnitude of the change in occupancy, detectability, initial occupancy levels, and the rarity of species. Our simulations suggest that monitoring has at least an 80% chance at detecting a 50% decline in occupancy for 22% of the modeled species across the three parks over the next 15¿yr. Monitoring is more likely to detect increasing occupancy trends, with at least an 80% chance at detecting a 50% increase in 87% of species. The addition of camera-trapping increased average power to detect a 50% decline in mammals compared with using only live trapping by 63%. We provide a flexible tool that can help decision-makers design and evaluate monitoring programs for hundreds of species at a time in a range of ecological settings, while explicitly considering the distribution of species and alternative sampling methods.

Targeted threatened species management is a central component of efforts to prevent species extinction. Despite the development of a range of management frameworks to improve conservation outcomes over the past decade, threatened species management is still commonly characterised as ad hoc. Although there are notable successes, many management programs are ineffective, with relatively few species experiencing improvements in their conservation status. We identify underlying factors that commonly lead to ineffective and inefficient management. Drawing attention to some of the key challenges, and suggesting ways forward, may lead to improved management effectiveness and better conservation outcomes. We highlight six key areas where improvements are needed: 1) stakeholder engagement and communication; 2) fostering strong leadership and the development of achievable long-term goals; 3) knowledge of target species' biology and threats, particularly focusing on filling knowledge gaps that impede management, while noting that in many cases there will be a need for conservation management to proceed initially despite knowledge gaps; 4) setting objectives with measurable outcomes; 5) strategic monitoring to evaluate management effectiveness; and 6) greater accountability for species declines and failure to recover species to ensure timely action and guard against complacency. We demonstrate the importance of these six key areas by providing examples of innovative approaches leading to successful species management. We also discuss overarching factors outside the realm of management influence that can help or impede conservation success. Clear recognition of factors that make species' management more straightforward ¿ or more challenging ¿ is important for setting realistic management objectives, outlining strategic action, and prioritising resources. We also highlight the need to more clearly demonstrate the benefit of current investment, and communicate that the risk of under-investment is species extinctions. Together, improvements in conservation practice, along with increased resource allocation and re-evaluation of the prioritisation of competing interests that threaten species, will help enhance conservation outcomes for threatened species.

Monitoring is essential for effective conservation and management of threatened species and ecological communities. However, more often than not, threatened species monitoring is poorly implemented, meaning that conservation decisions are not informed by the best available knowledge. We outline challenges and provide best-practice guidelines for threatened species monitoring, informed by the diverse perspectives of 26 conservation managers and scientists from a range of organisations with expertise across Australian species and ecosystems. Our collective expertise synthesised five key principles that aim to enhance the design, implementation and outcomes of threatened species monitoring. These principles are (i) integrate monitoring with management; (ii) design fit-for-purpose monitoring programs; (iii) engage people and organisations; (iv) ensure good data management; and (v) communicate the value of monitoring. We describe how to incorporate these principles into existing frameworks to improve current and future monitoring programs. Effective monitoring is essential to inform appropriate management and enable better conservation outcomes for our most vulnerable species and ecological communities.

Biodiversity offsetting schemes permit habitat destruction, provided that losses are compensated by gains elsewhere. While hundreds of offsetting schemes are used around the globe, the optimal timing of habitat creation in such projects is poorly understood. Here, we developed a spatially explicit metapopulation model for a single species subject to a habitat compensation scheme. Managers could compensate for destruction of a patch by creating a new patch either before, at the time of, or after patch loss. Delaying patch creation is intuitively detrimental to species persistence, but allowed managers to invest financial compensation, accrue interest, and create a larger patch at a later date. Using stochastic dynamic programming, we found the optimal timing of patch creation that maximizes the number of patches occupied at the end of a 50-yr habitat compensation scheme when a patch is destroyed after 10 yr. Two case studies were developed for Australian species subject to habitat loss but with very different traits: the endangered growling grass frog (Litoria raniformis) and the critically endangered Mount Lofty Ranges Southern Emu-wren (Spititurus malachurus intermedius). Our results show that adding a patch either before or well after habitat destruction can be optimal, depending on the occupancy state of the metapopulation, the interest rate, the area of the destroyed patch and metapopulation parameters of the focal species. Generally, it was better to delay patch creation when the interest rate was high, when the species had a relatively high colonization rate, when the patch nearest the new patch was occupied, and when the destroyed patch was small. Our framework can be applied to single-species metapopulations subject to habitat loss, and demonstrates that considering the timing of habitat compensation could improve the effectiveness of offsetting schemes.

Understanding where species occur and how difficult they are to detect during surveys is crucial for designing and evaluating monitoring programs, and has broader applications for conservation planning and management. In this study, we modelled occupancy and the effectiveness of six sampling methods at detecting vertebrates across the Top End of northern Australia. We fitted occupancy-detection models to 136 species (83 birds, 33 reptiles, 20 mammals) of 242 recorded during surveys of 333 sites in eight conservation reserves between 2011 and 2016. For modelled species, mean occupancy was highly variable: birds and reptiles ranged from 0.01-0.81 and 0.01-0.49, respectively, whereas mammal occupancy was lower, ranging from 0.02-0.30. Of the 11 environmental covariates considered as potential predictors of occupancy, topographic ruggedness, elevation, maximum temperature, and fire frequency were retained more readily in the top models. Using these models, we predicted species occupancy across the Top End of northern Australia (293,017 km2) and generated species richness maps for each species group. For mammals and reptiles, high richness was associated with rugged terrain, while bird richness was highest in coastal lowland woodlands. On average, detectability of diurnal birds was higher per day of surveys (0.33 ± 0.09) compared with nocturnal birds per night of spotlighting (0.13 ± 0.06). Detectability of reptiles was similar per day/night of pit trapping (0.30 ± 0.09) as per night of spotlighting (0.29 ± 0.11). On average, mammals were highly detectable using motion-sensor cameras for a week (0.36 ± 0.06), with exception of smaller-bodied species. One night of Elliott trapping (0.20 ± 0.06) and spotlighting (0.19 ± 0.06) was more effective at detecting mammals than cage (0.08 ± 0.03) and pit trapping (0.05 ± 0.04). Our estimates of species occupancy and detectability will help inform decisions about how best to redesign a longrunning vertebrate monitoring program in the Top End of northern Australia.

Active engagement with practitioners is a crucial component of model-based decision-making in conservation management; it can assist with data acquisition, improve models and help narrow the ¿knowing¿doing¿ gap. We worked with practitioners of one of the worst invasive species in Australia, the cane toad Rhinella marina, to revise a model that estimates the effectiveness of landscape barriers to contain spread. The original model predicted that the invasion could be contained by managing artificial watering points on pastoral properties, but was initially met with scepticism by practitioners, in part due to a lack of engagement during model development. We held a workshop with practitioners and experts in cane toad biology. Using structured decision-making, we elicited concerns about the original model, revised its structure, updated relevant input data, added an economic component and found the most cost-effective location for a barrier across a range of fixed budgets and management scenarios. We then conducted scenario analyses to test the sensitivity of management decisions to model revisions. We found that toad spread could be contained for all of the scenarios tested. Our modelling suggests a barrier could cost $4·5¿M (2015 AUD) over 50¿years for the most likely landscape scenario. The incorporation of practitioner knowledge into the model was crucial. As well as improving engagement, when we incorporated practitioner concerns (particularly regarding the effects of irrigation and dwellings on toad spread), we found a different location for the optimal barrier compared to a previously published study (Tingley et¿al. 2013). Synthesis and applications. Through engagement with practitioners, we turned an academic modelling exercise into a decision-support tool that integrated local information, and considered more realistic scenarios and constraints. Active engagement with practitioners led to productive revisions of a model that estimates the effectiveness of a landscape barrier to contain spread of the invasive cane toad R.¿marina. Benefits also include greater confidence in model predictions, improving our assessment of the cost and feasibility of containing the spread of toads.

Conservation of endangered species increasingly envisages complex strategies that integrate captive and wild management actions. Management decisions in this context must be made in the face of uncertainty, often with limited capacity to collect information. Adaptive management (AM) combines management and monitoring, with the aim of updating knowledge and improving decision-making over time. We provide a guide for managers who may realize the potential of AM, but are unsure where to start. The urgent need for iterative management decisions, the existence of uncertainty, and the opportunity for learning offered by often highly-controlled captive environments create favorable conditions for AM. However, experiments and monitoring may be complicated by small sample sizes, and the ability to control the system, including stochasticity and observability, may be limited toward the wild end of the spectrum. We illustrate the key steps to implementing AM in threatened species management using four case studies, including the management of captive programs for cheetah (Acinonyx jubatus) and whooping cranes (Grus americana), of a translocation protocol for Arizona cliffroses Purshia subintegra and of ongoing supplementary feeding of reintroduced hihi (Notiomystis cincta) populations. For each case study, we explain (1) how to clarify whether the decision can be improved by learning (i.e. it is iterative and complicated by uncertainty) and what the management objectives are; (2) how to articulate uncertainty via alternative, testable hypotheses such as competing models or parameter distributions; (3) how to formally define how additional information can be collected and incorporated in future management decisions.

Increasing the colonization rate of metapopulations can improve persistence, but can also increase exposure to threats. To make good decisions, managers must understand whether increased colonization is beneficial or detrimental to metapopulation persistence. While a number of studies have examined interactions between metapopulations, colonization, and threats, they have assumed that threat dynamics respond linearly to changes in colonization. Here, we determined when to increase colonization while explicitly accounting for non-linear dependencies between a metapopulation and its threats. We developed patch occupancy metapopulation models for species susceptible to abiotic, generalist, and specialist threats and modeled the total derivative of the equilibrium proportion of patches occupied by each metapopulation with respect to the colonization rate. By using the total derivative, we developed a rule for determining when to increase metapopulation colonization. This rule was applied to a simulated metapopulation where the dynamics of each threat responded to increased colonization following a power function. Before modifying colonization, we show that managers must understand: (1) whether a metapopulation is susceptible to a threat; (2) the type of threat acting on a metapopulation; (3) which component of threat dynamics might depend on colonization, and; (4) the likely response of a threat-dependent variable to changes in colonization. The sensitivity of management decisions to these interactions increases uncertainty in conservation planning decisions.

Adaptive management is a framework for resolving key uncertainties while managing complex ecological systems. Its use has been prominent in fi sheries research and wildlife harvesting; however, its application to other areas of environmental management remains somewhat limited. Indeed, adaptive management has not been used to guide and inform metapopulation restoration, despite considerable uncertainty surrounding such actions. In this study, we determined how best to learn about the colonization rate when managing metapopulations under an adaptive management framework. We developed a mainland-island metapopulation model based on the threatened bay checkerspot butterfl y (Euphydryas editha bayensis) and assessed three management approaches: adding new patches, adding area to existing patches, and doing nothing. Using stochastic dynamic programming, we found the optimal passive and active adaptive management strategies by monitoring colonization of vacant patches. Under a passive adaptive strategy, increasing patch area was best when the expected colonization rate was below a threshold; otherwise, adding new patches was optimal. Under an active adaptive strategy, it was best to add patches only when we were reasonably confi dent that the colonization rate was high. This research provides a framework for managing mainland-island metapopulations in the face of uncertainty while learning about the dynamics of these complex systems.