Substantial declines in biodiversity over the past century demonstrate an immediate need to preserve ecosystems and further mitigate habitat loss. Monitoring changes in ecosystem condition at region thru continental to global scales can provide important information about biodiversity declines and help facilitate targeted intervention. Efforts to use satellite imagery to map ecosystem condition change have experienced challenges with distinguishing observed changes from the natural variation of ecosystems. In this study we use an innovative deep learning architecture to pair time series satellite imagery with locations of known on-ground condition. Our model was developed using 209,041 on-ground records of native species present in the landscape, as a surrogate measure of ecosystem condition, coupled with Landsat time series data and topographic and climatological datasets. We predict ecosystem condition across the Australian continent for several years (2010, 2015, 2020, 2021, 2022) at 100 m. Arid regions in Australia's interior had predicted condition scores close to reference condition (1) for all years. Comparatively, highly modified landscapes in Australia's southeastern and southwestern regions had predicted condition scores closer to fully degraded (0). Mean predicted ecosystem condition across Australia was greater than 0.65 for all years, suggesting greater overall presence of native species rather than absence, however this was spatially variable. Our results demonstrate that using deep learning techniques and time series data can provide quantitative information on ecosystem condition, accounting for temporal variability of vegetation phenology and spatial variability across bioregions. Ongoing efforts to collect essential biodiversity variables from space must consider integrating with deep leaning approaches that have capacity for context driven spatial modelling. This will help ensure mapping products can support policy and inform intervention strategies.

Adoption of the Kunming–Montreal Global Biodiversity Framework (KM‐GBF) has led to further growth in attention being directed to the challenge of bending the curve of biodiversity loss to achieve a nature‐positive world. However, concerns have been raised that, unless progress towards achieving net gain in biodiversity can be measured effectively at a whole‐system level, the term ‘nature‐positive’ risks becoming no more than a convenient label for any action intended to benefit nature. Biodiversity is complex, and initiatives promoting and informing the measurement of biodiversity change usually address this complexity by employing separate metrics to assess different biodiversity levels and entities, and different causal‐framework components (responses, pressures, state and benefits). As a consequence, assessing achievement of nature positive at a whole‐system level is challenged by the reality that these various components do not exist or function in isolation from one another but are instead connected and interact in a myriad of ways. More integrative biodiversity indicators can complement and add significant value to the use of simpler metrics by helping to account for: (i) major sources of non‐additivity in the way that biodiversity itself, and the consequences of actions impacting biodiversity, scale from local to whole‐system level as a function of compositional variation and ecological interactions within and between discrete biological entities (e.g. species and ecosystem types); and (ii) complexities in how multiple pressures (e.g. climate and land‐use change) and multiple management responses (e.g. protection and restoration) combine in shaping outcomes for biodiversity. Integrative indicators incorporating a predictive capability also offer considerable potential to more strongly couple monitoring of actions implemented under the KM‐GBF targets to monitoring of outcomes achieved under that framework's goals. In addition, these same indicators can more effectively link monitoring to planning of further actions, which can most effectively and efficiently advance progress towards bending the curve of biodiversity loss, across both public and private sectors.

Abstract. Model intercomparisons are emerging as a powerful approach to improve our understanding of biodiversity and ecosystem responses to global changes, thereby supporting policy design and decision-making. However, efforts to date have focused on climate change impacts on species richness at a global scale, while regional impacts remain comparatively understudied. To address this gap, we propose a protocol for a biodiversity model intercomparison project focused on regional-to-continental scale patterns of species abundance and occurrence across space and time. This protocol covers a continuum of modeling approaches, ranging from correlative to process-explicit models. We detail the standardized input data used for model calibration –climate, land-use, and biodiversity data– and outline a common framework for model validation and performance assessment. Our protocol is structured in two phases. The first phase involves model specification, calibration, and validation using historical data, together with counterfactual experiments designed to attribute observed changes in population abundance and occurrence to different environmental drivers. In the second phase, model projections are compared under future climate and land-use scenarios. This two-phase approach enables direct comparison of model predictive performance on common benchmark data, providing an empirical basis for assessing confidence in future projections. By combining regional-scale biodiversity time series, common calibration and out-of-sample validation, and explicit representation of ecological processes, this protocol advances regional biodiversity model intercomparisons and sets the stage to reveal fundamental insights into which models, processes, and drivers determine historical and future biodiversity dynamics.

1. Landscape connectivity links local habitat condition to broader ecological processes including dispersal, recolonisation, range shifts, and species persistence, that underpin biodiversity responses to land-use and climate change. Several established biodiversity indicators — habitat connectedness (including as an input to the Biodiversity Habitat Index, BHI), Protected Area Connectedness Index (commonly referred to as PARC-connectedness), and the Bioclimatic Ecosystem Resilience Index (BERI) — formalise these processes as least-cost, condition-weighted representations of connected habitat. Applying these landscape connectivity formulations consistently across large spatial extents, multiple dispersal scales, protected-area configurations, and climate futures remains constrained by the computational cost of fine-resolution connectivity over broad landscapes. 2. We present a multi-resolution analytical framework that addresses this constraint by extending established cost-benefit connectivity logic into a scalable, indicator-oriented architecture. Building on the long-recognised principle that spatial precision matters most for areas close to each focal cell, formalised in earlier implementations through distance-dependent aggregation, we reformulate the calculation around a hierarchical graph derived from globally anchored, precomputed raster overviews. This design preserves fine-grain detail where it most influences connectivity, represents more distant landscape context at progressively coarser resolution, and produces reusable path-level outputs that are computed once and applied across dispersal scales and climate scenarios without repeating graph traversal. 3. From these shared path distances, habitat connectedness, PARC-connectedness, and BERI are each derived as condition-weighted, dispersal-decayed integrals over accessible habitat, differing only in how destination nodes are weighted and what ecological quantity is being assessed. The unified derivation ensures consistency across indicators while eliminating redundant computation. In an example application for Tasmania, Australia (approximately 68,400 km2; more than 300,000 cells at about 900 m resolution), all three indicators were computed across three dispersal scales and six climate scenarios in under a minute on a standard laptop. 4. The framework makes no claim to a new definition of connectivity. Its contribution is analytical and architectural, enabling established and well-validated connectivity concepts to be applied reproducibly and efficiently at the scales demanded by contemporary biodiversity monitoring and conservation planning. It is distributed as a Python package with a highperformance Rust engine, designed for integration into open geospatial workflows and large-scale indicator applications.

ABSTRACT Fragmented systems for monitoring and assessing biodiversity and ecosystem services limit the ability to track progress at local and national scales in international multilateral environmental agreements (MEAs). This greatly challenges coordinated actions to meet agreed‐upon global commitments. Filling this gap requires integrated and concerted design of data‐to‐decision workflows. The Essential Biodiversity Variables (EBVs) and Essential Ecosystem Service Variables (EESVs) are tools that can coordinate structured and consistent monitoring, generate harmonized and scalable data products, and facilitate reporting that is useful for multiple purposes. Specifically, EBV/EESV data products are intended to synthesize information to serve the needs of the Global Biodiversity Framework (GBF), the System of Environmental‐Economic Accounts Ecosystem Accounting (SEEA‐EA), and assessments of the Intergovernmental Science‐Policy Platform on Biodiversity and Ecosystem Services (IPBES) while also informing regional, national, and sub‐national conservation policy. This integrative approach works if local data collection is designed to be interoperable, and it is fundamental to improve models, forecasts, and indicators required in key policy and decision processes. Through application cases, we demonstrate the use of EBVs and EESVs in national assessments and scenario analyses for strategic policy and spatial planning with scalable and repeatable workflows from primary data to indicators for decision support.

Developing ecosystem models has traditionally been limited to a small global community of experts because of the complex skills and resources required. However, the emergence of user-friendly artificial intelligence (AI) tools with powerful generative capabilities could democratize ecosystem modeling, enabling both experts and nonspecialists to build models. We explore a speculative future where AI enables automated end-to-end model development and application. Although such tools could accelerate and enhance modeling tasks, their widespread adoption raises concerns about data integrity, bias, interpretation reliability, and the potential erosion of human expertise. We argue that regardless of AI's technical advancement, human engagement and control remain essential. The global community must respond by identifying key factors that distinguish desirable outcomes and developing infrastructure, standards, and ethical guidelines to ensure AI use in ecosystem modeling remains scientifically robust while supporting sustainable and equitable outcomes.

Aim The condition of terrestrial ecosystems (i.e., structure, function and composition) has been altered by increasing human pressure globally, affecting biodiversity and the sustainability of services that ecosystems provide. Effective conservation and restoration decisions will be supported by a continuous global fine-scale assessment of changes in ecosystem condition over time. Innovation We developed a framework to provide annual ecosystem condition assessment based on remotely sensed ecosystem condition variables. Our conceptually simple approach compares the ecosystem variables of a target site to training sites with known condition. The comparison yields an ecosystem condition score as the similarity to the training sites. This framework provides a way to identify the temporal change in ecosystem condition while accommodating the natural dynamics of ecosystems and the spatial change from alternative natural states of ecosystems. Main Conclusions As a case study, we mapped annual ecosystem condition across Australia from 2013 to 2022, using Landsat-derived condition variables at 100 m resolution. Our approach estimated ecosystem condition with high accuracy (RMSE = 0.23), which remained consistent along large environmental gradients. When evaluated against independent field-based observations and other condition maps, our approach showed high agreement and identified the trajectories of ecosystem condition over time for sites with known changes. The case study suggests our approach is a useful tool to assess and monitor annual changes in ecosystem condition at continental scales. Our maps provide the information needed for assessing long-term anthropogenic impacts on ecosystem condition and to support policy, planning and management decisions.

Scenario building is a powerful tool for evaluating drivers of environmental change and assessing alternative socioecological pathways, helping integrate science-based information into decision-making. Nonetheless, this potential has not been fully embraced by scientists and decision-makers, in part owing to limitations of current scenario frameworks at representing the diversity of values for nature and potential transformative changes to bend the biodiversity loss curve. There is still a need to further develop scientists' capacities to include a transdisciplinary perspective in scenario building to address the drivers of transformative change. This paper addressesthese needs by reflecting on the role of scientists engaged in scenario building in the construction of sustainable futures through the lens of three key concepts: social learning, knowledge co-production and conscientiza & ccedil;& atilde;o (a Portuguese term meaning to build sociopolitical awareness and take action). Drawing on a survey of participants of a Scenario Building School and a literature review, we suggest that scientists require capacity building to leverage these concepts together for the construction of transformative futures. This includes addressing power imbalances, improving inclusive and transdisciplinary participatory methods, reaching consensus and promoting action. We recommend that scientists engaged in scenario building focus on fostering transformative changes, challenging mainstream storylines, embracing diversity and addressing inequalities to pursue sustainable futures.

Current biodiversity metrics derived from remote sensing data are typically applied to small local areas, require significant training data, and are not easily extensible globally. Here we propose the mathematical concept of intrinsic dimensionality (ID) as a method to quantify terrestrial vegetation variability without a need for in situ training data. We apply this technique to airborne imaging spectroscopy data from the Surface Biology and Geology High Frequency Time series (SHIFT) airborne campaign, with weekly overflights from February to May 2022 over a region in California stretching from Figueroa Mountain in the Los Padres National Forest to Point Conception and adjacent coastal areas. ID is considered in both spatial and temporal context—spatial ID represents spectral variability across a geographical region at a single time step, and temporal ID represents spectral variability over time for a single geographical location. Results show an encouraging and significant correlation between spatially calculated ID and in situ vegetation species richness data despite different spatial scales between the two ( p = 0.01). Spatial ID remained largely unchanged at each time step over the course of three months during the spring green‐up period when vegetation characteristics and spectral responses were changing rapidly (number of species remains unchanged even though spectra reflect phenological change over time). The temporal ID remained constant for pseudo‐invariant surfaces such as parking lots, roofs, and rock, but showed increased ID with time for trees, shrubs, and grasses. This robustness of spatial ID to seasonal change is desirable in any measure of species richness because it is insulated from changes in vegetation condition that are unrelated to plant species richness. Even though the spatial ID is consistent across acquisition dates, when considering the full time series (temporal ID), we find that subweekly sampling may be necessary to spectrally capture the full phenological cycle of certain vegetation types.

Habitat retention and restoration are fundamental to minimising species extinctions. Mitigating impacts on threatened species habitat requires knowing where and how much habitat remains, the quality of the habitat, and how it is changing over time. A systematic approach is therefore essential for regular and consistent measurement of threatened species habitat that can be applied across multiple scales.We present a new approach to track progress in providing habitat for threatened species over time and space, and apply it to the Australian continent. This involves: 1) estimating the historical (pre-intensification) habitat distribution for listed threatened species, and 2) deriving a metric of habitat provision for threatened species by combining the historical habitat distributions with remotely sensed annual ecosystem condition data. We demonstrate the method for Australia using 1,518 nationally-listed threatened species and ecosystem condition data for 2017 and 2018, and present summaries by sub-national jurisdictions.Across Australia, intensively developed regions had the greatest number of threatened species based on estimated historical habitat. Between 2017 and 2018 threatened species habitat decreased in six of eight sub-national jurisdictions. Percent losses were greatest in the smallest jurisdictions, which were also estimated to retain the least threatened species habitat. As of 2018, less than 50 % of the historical habitat for threatened species is estimated to have remained across Australia. Improving ecosystem condition could have the largest benefits where multiple threatened species potentially co-occur or where little historical habitat remains.Our approach facilitates high-level assessments of status and trends in habitat provision to support threatened species. The data and metric are spatially explicit and scalable, enabling aggregation over any spatial area such as for biophysical ecosystem accounting. We provide the estimated historical habitat distributions to facilitate ongoing assessments.

The hypothesis that pyrodiversity begets biodiversity is foundational to conservation management in fire-prone ecosystems and has received extensive research attention. However, empirical evidence for the hypothesis remains ambivalent. Moreover, few studies directly assess the key question of how much pyrodiversity is needed to conserve all species within a community. A novel way of addressing this is to use the biodiversity–maximisation approach developed for reserve selection as part of strategic conservation planning. We apply this approach to an ant dataset from a long-term fire experiment in northern Australia to establish how many of the six experimental fire treatments are required to represent all local ant diversity. We identified the treatment combinations required to maximise species richness and geometric mean abundance. We repeated this for six fire-activity classes based on cumulative fire intensity experienced by plots over the course of the experiment. We found that a very limited number of fire treatments or fire activity classes were needed to represent all of the highly diverse ant species and to maximise the geometric mean abundance of ants. We attribute this to the substantial small-scale heterogeneity of fire behaviour and vegetation structure within individual fire treatments. We conclude that high pyrodiversity at larger spatial scales is not required for sustaining ant biodiversity in our study system. We believe that a reserve selection approach is a powerful method for assessing how much pyrodiversity is needed to conserve biodiversity and recommend that it be applied to other taxa and other ecosystems.

Nitrogen cycles control the structure, function, and composition of ecosystems globally. Despite their importance, our understanding of long‐term changes in global nitrogen cycles remains limited. The foliar nitrogen stable isotope ratio (δ 15 N) serves as a valuable metric for assessing changes in nitrogen cycling and potentially in plant nitrogen availability. However, existing observations of δ 15 N suffer from spatial bias and temporal discontinuity with contradictory findings across biomes, hindering our ability to detect and attribute drivers of change. Leveraging ground‐based observations as our calibration source, we derived annual maps of foliar δ 15 N spanning from 1984 to 2022 globally from Landsat spectra. We found that the Landsat‐derived δ 15 N effectively captured the observations, with an R 2 of 0.77 and a normalized root mean square error of 0.15. Globally, we found widespread temporal changes in δ 15 N with significant decreases for 44% and increases for 16% of vegetated ecosystems. Foliar δ 15 N mostly declined in forest ecosystems but increased in non‐forest land cover types. Gross primary productivity and its trend consistently explained spatiotemporal variation of δ 15 N globally, indicating increasing plant demand could lead to decreasing δ 15 N. Our study presents an innovative approach to effectively monitor and track potential changes in global nitrogen cycles over the past four decades, setting the stage for more impactful management and conservation strategies.

There is a rapidly growing need for efficient but rigorous methods for organizations to assess and disclose their biodiversity impacts. We devised a habitat-based analytical approach for estimating the direct impacts of an organization on biodiversity. In our broad approach, we considered the time series of an organization's spatial footprint and assumed its biodiversity position was the accumulated positive and negative impacts over space and time. We demonstrated the approach by assessing the biodiversity position of CSIRO-Australia's national science agency, which has owned or controlled 50 sites across Australia since 1916, covering >460,000 ha. We applied 3 complementary habitat-based biodiversity indicators (effective habitat area, species extinction risk, and threatened species habitat), all with a fine resolution annual (1987-2023) time series of ecosystem condition as their basis. At the end of the most recent observation year, the CSIRO was in a negative biodiversity position in terms of all 3 biodiversity indicators. Over the time series considered, the activities of CSIRO were estimated to have led to an increase in the extinction risk for all native species by 1.0 species; a reduction in effective habitat area of 11,945 ha and a reduction in threatened species habitat of 22,307 species hectares (i.e., condition-weighted amount of habitat available to threatened species). Although the magnitude of the biodiversity position for CSIRO was strongly influenced by a single very large site (Murchison), the vast majority of the CSIRO sites were also in a negative position when considered separately. We demonstrated how future-looking scenario analysis can be linked with this biodiversity assessment approach, with a single natural regeneration action across the large Murchison site estimated to return CSIRO's biodiversity position close to neutral within 50 years.

Biodiversity indicators measure progress toward global biodiversity goals, including the Kunming-Montreal Global Biodiversity Framework of the UN Convention on Biological Diversity. A suite of indicators is typically needed to capture the complexity of biodiversity. For the suite to be effective, it needs to capture the important aspects of the system, without over-representing some at the expense of others. It is therefore important to identify redundancies and contradictions between indicators within a suite. Comparing indicators derived from independent input data can also serve as a validation of indicator accuracy, or alternatively, identify unexpected behaviour and potential flaws or incompatibilities in data or construction. We examined the relationships between eight widely adopted global biodiversity indicators, all proposed for use in the monitoring framework of the Kunming-Montreal global biodiversity framework. We focused on the relationships between indicators driven by landuse data and those driven by species observations, testing for pairwise and multivariate correlations between indicator values for ecoregions. We used hierarchical clustering to identify spatial patterns in relationships between landuse and species response indicators. We found no strong correlation between landuse and species-response indicators in ecoregions with high landuse. Reasons for disagreement among indicators may include inaccurate, or coarse spatial data and mismatched baselines. We did find strong correlations between indicators sharing input data. We suggest that indicators selected for the Kunming-Montreal framework be systematically reviewed for correlations and recommend that suites of indicators should consider the independence of input data sources to minimise the risk of using correlated or biased measures of conservation progress.

Based on an extensive model intercomparison, we assessed trends in biodiversity and ecosystem services from historical reconstructions and future scenarios of land-use and climate change. During the 20th century, biodiversity declined globally by 2 to 11%, as estimated by a range of indicators. Provisioning ecosystem services increased several fold, and regulating services decreased moderately. Going forward, policies toward sustainability have the potential to slow biodiversity loss resulting from land-use change and the demand for provisioning services while reducing or reversing declines in regulating services. However, negative impacts on biodiversity due to climate change appear poised to increase, particularly in the higher-emissions scenarios. Our assessment identifies remaining modeling uncertainties but also robustly shows that renewed policy efforts are needed to meet the goals of the Convention on Biological Diversity.