New incentives and instruments for financing ecosystem restoration require frameworks that support planning, monitoring and reporting, including the identification and use of leading indicators. Leading indicators have the potential to predict the outcomes of restoration interventions before full recovery has occurred. State-and-transition models are a form of ecosystem dynamics modelling that has been widely and systematically applied to classify and describe ecosystem dynamics in North America and Australia, including ecosystem recovery efforts. State-and-transition models provide a framework for the organisation and selection of leading, lagging and coincident indicators for monitoring restoration outcomes. We outline a process for strengthening the application of state-and-transition models to restoration. This includes key considerations that are unique to restoration, such as the importance of starting states and how restoration management interventions can be incorporated into models as drivers of transitions. We demonstrate the approach through a case study and apply our restoration state-and-transition model to help guide the timing of restoration management actions informed by indicators. We then use insights on ecological recovery dynamics provided by the restoration state-and-transition model to generate hypotheses for identifying and selecting leading indicators. For example, the soil nutrient status of retired farmland is predicted to be a leading indicator of potential native forb establishment. Restoration-targeted state-and-transition models play a key role in systematically translating detailed practitioner insights, ecological data and indicators into structured formats that are accessible to a broad audience. To credibly implement the framework and leading indicators in nature markets, complementary scientific tools and governance processes are needed. This includes establishing agreed benchmarks for indicators, designing robust audit mechanisms and conducting further research to develop the evidence base for leading indicators.

Background Frequent fire characterises Australia’s tropical savannas and strongly influences vegetation structure and carbon (C) stocks. Fire management is widely used to reduce greenhouse gas emissions, but long-term effects of fire timing and frequency on woody C accumulation remain uncertain. Aims To quantify stand-level woody C stocks and assess cumulative multi-decadal effects of fire timing and frequency in northern Australian savannas. Methods Field measurements and historical biomass and basal area data from three long-term (17–30-year) fire trials were analysed. Treatments included unburnt controls and burned plots with up to four fire frequencies within a 6-year range and two fire seasons (early and late dry season). Carbon stocks were quantified for standing biomass, coarse woody debris and litter. Key results Total woody C accumulation was greater under early dry season fire than late dry season fire regimes. Longer between-fire intervals (particularly late dry season) may increase woody C stocks, although effect sizes were uncertain. Conclusions Early dry season burning was associated with greater long-term woody C accumulation than late dry season burning. Implications are that early dry season fire management remains important for reducing savanna fire emissions, but optimal fire regimes should consider site-specific conditions and broader ecological objectives.

Fire shapes the biodiversity of tropical savannas, yet challenges exist in accessing and interpreting the breadth of fire-ecology evidence to support appropriate management. We compiled a comprehensive fire–biodiversity database for Australian tropical savannas to improve knowledge discovery and support the iterative development of fire-management principles. Using systematic thematic synthesis, we mapped evidence on fire–biodiversity relationships across contemporary science and Traditional Knowledge (TK). Publication metadata were extracted, standardised and classified using a large language model (LLM)-assisted workflow, then manually reviewed and cross-validated against an independently curated subset. LLM-assisted extraction performed comparably to human extraction, but only with careful parameterisation and quality control. The database spans 382 publications from the Australian savanna biome. The evidence base is dominated by terrestrial flora (≈51%) and studies of mixed-element fire regimes (≈28%), fire frequency (≈17%) and time since fire (≈17%). Partnerships with Indigenous organisations or entities were reported in ≈23% of publications, but only 22 explicitly incorporated TK or Indigenous partners as co-investigators. Evidence is geographically clustered and taxonomically uneven, revealing clear knowledge gaps. Our thematic synthesis map and accompanying database provide a platform to inform culturally grounded, evidence-based fire management and highlight future research priorities.

The world is experiencing a biodiversity crisis. Steep declines in habitat quality and ecosystem services have resulted in interest in markets to help fund ecological restoration. One way that ecological restoration is assessed is through indicators of ecosystem condition, namely, a measurement of how different a landscape is from its preindustrial or reference state. State-and-transition models (STMs) and their quantitative implementation, state-and-transition simulation models (STSMs), can be used to help assess ecosystem condition and to support nature markets. These models can simulate multi-directional change in ecosystems across space and time and are useful tools for combining empirical data and expert knowledge. Here, we provide a brief review of STMs and how they relate to ecosystem condition. We then outline a protocol for building STMs and STSMs and using them to assess condition states at landscape scales. We argue that this is a powerful approach to incorporate many different types of information into an integrated, multi-dimensional assessment of ecosystem condition. With this approach, STMs could support nature markets to drive widespread ecological restoration and reverse declines in biodiversity.

Communicating complex scientific concepts to non-experts is a persistent challenge. The communication of ecological state-and-transition models (STMs) through box-and-arrow diagrams is one example. This paper explores how virtual reality (VR) can make STMs more accessible. Using ecosystem STMs as a case study, we present a proof-of-concept system enabling users to viscerally experience the content of the model. We followed a three-phased participatory design process: first, 2 ecology experts guided the development of a VR prototype. Next, 17 government environmental management professionals evaluated its utility and features. Finally, after refining the system, 12 VR researchers informed design considerations and improvements. Our findings provide practical insights for visualising STMs in VR, and also contribute to the emerging field of "data visceralisation". We found this approach engages users and supports understanding of qualitative aspects of real-world phenomena. However, complex models like ecosystem STMs require the creation of accurate and extensive simulations. We conclude with a discussion for future directions.

Tree hollows are important habitat resources for wildlife globally. In the tropical savannas of northern Australia, the abundance of tree hollows is influenced by both fire and termites. With the regular application of prescribed fire in these ecosystems, it is important to understand the implications of fire management on important habitat resources - especially when applied over the long-term. This study uses a long-term fire experiment (18 years of applied fire treatments) with targeted termite and tree hollow surveys to investigate how the proportion of stem hollowing and abundance of hollow entrances are affected by termites and different long-term fire regimes. We used sonic tomography in a novel application to non-destructively estimate tree stem hollowing. Trunk diameter was identified as the strongest predictor of both stem hollowing and hollow entrance abundance, with larger trees having a greater proportion of the stem hollowing and greater number of hollow entrances. The proportion of stem hollowing tended to be greater closer to the base of the tree, and the number of hollow entrances was greater in Eucalyptus miniata than E. tetrodonta. While the proportion of stem that was hollowed did not influence the number of hollow entrances, the presence of any hollowing at 1.3 m was associated with more hollow entrances. We did not detect an effect of fire activity on stem hollowing or the abundance of hollow entrances at the individual tree level, and it may be that these effects are only detectable at the stand level due to changes in tree demographics with varying fire regimes. As large trees tend to have more hollow entrances, management to promote habitat for wildlife should focus on fire regimes that avoid the loss of large trees in the landscape.

Wetlands are essential for biodiversity conservation and ecosystem services, yet they remain under-monitored in many parts of the world, including wet-dry tropical savannas. In Northern Australia, increasing development pressures and climate change raise concerns about the long-term conservation of the extensive and relatively intact wetlands. This study developed a wetland mapping framework tailored to the Australian tropical savanna, mapping change to wetlands in the Adelaide River Catchment from 1987 to 2024. A multi-index Landsat-based classification approach was implemented in Google Earth Engine, combining spectral indices, topography, and soil properties. To address the strong seasonality of tropical wetlands, the classification included separate wet/ dry season processing. Post-classification refinement using hierarchical rules and auxiliary datasets helped resolve confusion among spectrally similar classes. Wetlands were categorized into twelve classes based on hydrological regime and vegetation structure. Validation showed high accuracy (93 % +/- 1 %), with class-level accuracy above 80 % for most wetland types, including water bodies, mangroves, salt flats, swamps, marshes, and floodplains. Wetlands occupied a substantial portion of the catchment, covering approximately 40.5 % of the area in 1987 and 37.4 % in 2024, with marshes and floodplain woodlands dominating due to the flat terrain. Despite land use changes, 80 % of wetlands retained their class between 1987 and 2024. However, floodplain woodlands declined by 16,044 ha, often transitioning to other wetlands, non-wetland, or agricultural land, which increased by 29,000 ha. Internal transitions were common among estuarine and floodplain wetlands, reflecting natural dynamics. Dry periods reduced open water areas, while human-made wetlands increased. This is the first long-term assessment of wetland distribution in the region and provides essential spatial data for water management and conservation. The framework offers a transferable method for monitoring wetlands in other tropical savannas under environmental and development pressures.

Abstract Tree hollows are critical habitat for many species globally, and fauna studies often include assessments of hollow abundance. However, traditional ground‐based surveys for hollows can be inaccurate, either over‐ or under‐estimating hollow abundance and/or accessibility. In order to address this inaccuracy, ground‐based hollow counts have previously been calibrated using a ‘double‐sampling’ method such as felling or climbing trees. Here we test whether drone‐based surveys can be used to count and assess tree hollow accessibility and discuss the considerations and limitations of using drones for hollow surveys. In this study, we describe a survey of tree hollows in 134 Eucalyptus and Corymbia trees in a tropical savanna south of Darwin, Australia. Tree hollows were first counted from the ground using binoculars, then double‐sampled using drone‐based surveys. Drone‐based surveys detected more hollows than ground‐based surveys, with the latter underestimating potential habitat hollows by at least 15%. Hollows with estimated entrance diameters of 5–10 cm and 10–20 cm were most likely to be missed by ground‐based surveys. Drone‐based surveys also provided more information on hollow accessibility, identifying that 38% of hollows were inaccessible to fauna due to being ‘blind’ or blocked by termite material. Practical implication. Drone‐based surveys potentially offer a more accurate method by which to count and assess tree hollow accessibility for fauna, as well as a less biased means of calibrating ground‐based hollow counts. Important considerations for drone‐based hollow surveys include weather restrictions (e.g. wind and rain), time available, vegetation density and potential impacts on wildlife. Where complete hollow surveys by drone are not possible or there is insufficient time available, we recommend that a subset of ground‐surveyed trees are double‐sampled using drone‐based hollow surveys—particularly for studies where small‐ to medium‐sized hollows are important. Ground‐based hollow surveys alone risk underestimating the abundance of an important habitat resource and overestimating the number of currently accessible hollows. Thus, using a more accurate method such as drone‐based surveys to count or calibrate hollow numbers will likely provide improved estimates of landscape‐scale hollow abundance and accessibility.

Woody plantings are widely promoted to ameliorate biodiversity loss in agricultural landscapes. New market mechanisms are rapidly emerging to expedite such efforts, but limited tools and data to account for benefits achieved hamper their implementation. Using data from 204 primary studies and 1206 paired comparisons, we present a global meta-analysis of the biodiversity benefits of woody plantings in agricultural landscapes, in a biodiversity and ecosystem accounting framework. Consistent with emerging biodiversity and ecosystem accounting methods, we express results as agricultural field:natural reference and planting:natural reference ratios to estimate the biodiversity values of agricultural fields and plantings, respectively. Mean biodiversity abundance and species richness for agricultural fields were 0.40 of those for natural reference sites, compared with 0.62 for plantings averaging 20 years old, indicating a mean biodiversity benefit of 0.22. These values varied significantly among taxonomic groups, with unexpectedly high values for agricultural fields driven by high means for invertebrates. Variation among studies was substantial, and biodiversity values for plantings increased with higher diversity and native dominance of plantings and lower management intensity. Critically, estimates of biodiversity benefits based on abundance versus richness were comparable, but estimates using compositional measures typically implied substantially lower benefits, likely owing to effects of species identity. Our study operationalises approaches for quantifying the benefit of plantings for biodiversity and ecosystem accounting, and emphasises the need to use compositional measures for realistic estimates of biodiversity benefits.

The Full Carbon Accounting Model (FullCAM) simulates carbon (C) pools of live biomass, standing dead mass, debris and soil, the flows among them and the atmosphere, and the influences of fire and harvesting disturbances under Australian conditions. It is regularly used by governments, landowners, companies and researchers, at continental, regional and local scales. Recently, FullCAM was calibrated for seven categories of native tropical savanna vegetation. However, for non-savanna native vegetation, calibrated parameters are available for only two general vegetation categories, based on whether the annual rainfall exceeds or falls below 500 mm. These two categories are too broad to capture the large range of growth conditions, vegetation structures and species assemblages that occur across Australia’s native woody vegetation. Here, our objective was to improve FullCAM’s ability to model variation in C pools and post-disturbance recovery among eight native vegetation categories, from shrublands to rainforests, for which there were differences in biomass allocation, litterfall and/or decomposition. To do this, we calibrated FullCAM for each vegetation type, including 14 parameters that were calculated directly from field observations and 17 that were calibrated using a dataset containing about 9300 field plots with measurements of at least one woody vegetation C stock. New parameters (compared with the two general parameter sets) reduced bias from 77 to 25 % (averaged across C stocks), and root mean square error from 44 to 30 Mg C ha-1. Model accuracy could be further improved (i) by focusing on sites with a known disturbance history, (ii) calibrating as many vegetation categories as possible (instead of eight categories generalising across many species), and (iii) adding more detail to growth calculations to quantify factors that may not be adequately represented by FullCAM’s growth equation.

Ecology is the study of living organisms and their physical environment. Often, in an ecological modelling context, this has meant a lack of explicit consideration of human impacts on ecological systems, despite the clear influence humans have upon the natural environment over millennia. We suggest that this exclusion of the human from ecological modelling has created serious knowledge gaps in solving ecological challenges, especially given that humans are generally the underlying cause for ecosystem state change. Here we propose reframing ecological models to incorporate human (anthropogenic) drivers (including those that maintain ecosystem integrity and those that might alter it) and to view these drivers as interacting and interlinked, using a systems lens. We demonstrate this using the example of a generalised state-and-transition model and a case study example of ecological restoration of agricultural landscapes in southeastern Australia. This method identifies human (anthropogenic) drivers of ecological change that can be targeted for action. To increase the utility of ecological models, an understanding of anthropogenic influences and drivers should be incorporated, embedding them within adaptive management cycles to guide tangible ecological outcomes.

As small (<5.5 kg) native mammals continue to decline throughout northern Australia, there is a critical need to identify the most significant threatening processes and those management actions with the greatest likelihood of alleviating them. Using a structured elicitation process, we sought to identify such management priorities by: (1) reviewing the literature since 2010 to identify a suite of potential drivers of small mammal decline; and (2) estimating the relative impacts of the putative threats and the ability to manage such threats. We reviewed 106 publications, from which we identified 11 threats and 14 threat attributes possibly contributing to the decline of small mammals in the region. From the expert elicitation, we scored and ranked each threat and threat attribute for its severity, geographic scope, potential for mitigation, and uniformity of impact across species. The elicitation suggested that the contemporary threats with the greatest relative impact on native mammals in northwestern Australia are: (1) predation by feral cats; and (2) habitat degradation by feral livestock and by inappropriate fire regimes. Tractable management actions aimed at reducing the density of feral livestock and improving fire regimes are likely to simultaneously reduce predation pressure from feral cats on small mammals and improve the availability of critical resources.

Tropical savannas typically experience high fire frequencies, with prescribed fire commonly used as a management tool. Termites play an important role in the ecological functioning of tropical savannas, yet we have a limited understanding of how fire affects these important ecosystem engineers. To account for the effects of fire management on ecosystem structure and function, we need to understand the links between fire management and termite communities. This study used a long-term (18-year) fire experiment in a tropical savanna near Darwin, northern Australia, to investigate the effects of different fire regimes on termite species composition, abundance and activity. We measured termite abundance and activity using a combination of baiting and reduced transect survey methods and compared these with fire activity (summarised fire frequency and intensity) and woody cover. Termite species richness was similar across all fire treatments, and the level of fire activity had a minimal effect on species composition, which was more strongly influenced by woody cover. Wood-feeding termite abundance and the consumption of wood baits were negatively correlated with fire activity and positively correlated with woody cover. Soil/wood interface-feeding termites showed no correlation with fire activity but a positive correlation with woody cover. Significant negative mediation effects of fire activity through woody cover were detected on the abundance of wood- and soil/wood interface feeders and wood and straw bait consumption. Grass-feeding termites were encountered too infrequently to draw conclusions about their correlation with fire activity and woody cover; however, straw bait consumption was positively correlated with fire activity. Synthesis and applications. The effects of fire on termite abundance and activity are primarily indirect, mediated through changes in vegetation structure. As high fire activity is associated with reduced woody cover, maintaining regimes of frequent, high-intensity fires over the long term has the potential to affect ecosystem function. While minimising the occurrence of high-intensity, late dry season fires is consistent with fire management goals in these savannas, care is still required to avoid the negative consequences of high fire frequencies.

Termites are important ecosystem engineers in many ecosystems globally. Hence, surveys of termite species composition, abundance and activity can be important for understanding ecosystem function-especially in biomes where they tend to be abundant, such as tropical savannas. However, comprehensively surveying termites can be challenging due to their cryptic nature and varied feeding and nesting habits, which strongly influence the effectiveness of different survey methods. Baiting and active searches of reduced transects are two methods commonly used to sample termites, and while these methods have been evaluated in the savannas of South Africa, this has not occurred in the extensive tropical savannas of northern Australia. Thus, this study evaluated the effectiveness of baits and reduced transects to assess termite species richness and activity across 18 x 1 ha experimental plots in a tropical savanna near Darwin, Australia. Surveys in each plot consisted of two 60 x 2 m transects and a 9 x 3 baiting grid of alternating buried wood and paper baits and surface straw baits. Baits were checked three times: at 4-, 7- and 10-week intervals following placement. Upon survey completion, the sampling effort, efficacy and costs of each method were compared. Reduced transects detected all 32 species recorded in this study, representing four feeding groups (from undecayed wood to highly decayed organic material in the soil). Baiting detected 20 species, but failed to detect some of the species that fed on decayed materials. Paper baits, checked only twice (at 4 and 10 weeks following placement), were required to detect all species sampled at both wood and paper baits. Therefore, overall baiting costs could be reduced (without data loss) by using paper baits only and reducing the number of bait checks. Compared with baiting using all three bait types, reduced transects detected the most species and had the lowest per-species cost. Consequently, reduced transect surveys are the most effective method in these northern Australian savannas when assessing species composition. However, if the abundance of species that feed on undecayed wood or levels of termite activity are being assessed, then reduced baiting is a more appropriate method.

Effective satellite-based monitoring of ecosystem integrity or condition needs to address four key challenges: (a) context dependency; (b) alternative ecological states; (c) short-term temporal ecosystem dynamics; and (d) scarcity of reference data where ecosystems retain high levels of integrity. Here we present a typology, and outline strengths and weaknesses, of different approaches to mapping and monitoring ecosystem integrity across entire regions or continents using time series satellite data. We then describe how one of these approaches, the Habitat Condition Assessment System (HCAS), addresses all of the above challenges, and provide an outline of the evolved method which includes annual outputs, and Australian continent applications. HCAS requires three readily available inputs (i.e., representative examples of relatively natural areas as reference sites, remotely sensed ecosystem characteristics, and environmental covariate data) and could be easily adapted and applied by other countries to provide an effective indicator of ecosystem integrity for nature-based decisions.

The global decline in the extent and condition of ecological communities has resulted in an increasing demand for recovery and conservation plans. Conservation plans for ecological communities require a management framework with measurable, time-bound objectives, a targeted management strategy, and indicators that enable actions to be evaluated in relation to objectives. Methods that allow for the transfer of knowledge among similar systems and facilitate consistent and comparable plans are essential, especially when resources are constrained. We describe a process to streamline the development of conservation plans by combining functionally similar community sub-types into a multi-community State and Transition Model. We demonstrate this approach in a case study where we use the combined expertise of Australian ecologists to build a multi-community State and Transition Model for eucalypt woodlands of southern Australia – an ecosystem which occupies a vast geographical range across temperate Australia and includes many distinct vegetation communities, a growing number of which are endangered or threatened. We identify commonalities and differences among three broad woodland sub-types including a set of eight general condition states, a list of drivers of transitions among condition states, and the uncertainties and time-frames associated with each transition. Two key findings across all models are that management is state-dependent, and transition directly to the ‘Exemplar’ state from any other state is considered highly unlikely. Other examples of State and Transition Models in the literature are focused on single communities or a significantly smaller scale, and this is the first attempt to construct a nationally relevant multi-community State and Transition Model via a structured and consultative process. Based on this case study, we propose a repeatable protocol for developing multi-community State and Transition Models. This process could improve and streamline the development of robust conservation plans for threatened ecological communities more broadly.