The Ecological Knowledge System (EKS) for the Nature Repair Market, established by CSIRO in partnership with the Department of Climate Change, Energy, the Environment and Water, supports market integrity and participation by making it easier to access the ecological information needed to inform method development, project planning and investment decisions.
Methods under the market set out requirements for how eligible projects must be carried out. The first ‘method’ for the Nature Repair Market was made in February 2025, the 'Nature Repair (Replanting Native Forest and Woodland Ecosystems) Methodology Determination 2025'. Proponents can now apply to the Clean Energy Regulator to register a project under this method.
Under the method, proponents must use the National Biodiversity Assessment System (NBAS), a key component of the EKS. NBAS provides a nationally consistent approach to assessing biodiversity and forecasting the biodiversity benefits of projects. NBAS works by using information from state and transition models combined with on ground project data and national spatial datasets. This presentation will discuss how ecological knowledge, including vegetation data, from local to national scales is used to inform development and implementation of the first nature repair method.

A growing challenge is determining how to objectively compare the growing multitude of satellite-derived “vegetation” properties. This work demonstrates a comparison tool , able to be used for assessing which satellite-derived-vegetation-product (SDVP) is fit for purpose. It also presents the case for using consistently curated field and drone data for validation. In some areas, SDVP have a long history of established programs, developing, implementing and revising vegetation management. The “vegetation” properties mapped are not consistent, e.g. these maps cover properties that include vegetation: extent (by [in]consistently defined vegetation-types, e.g. woodland, forest, non-native); composition (by structure and/or floristics); canopy cover; fractional cover per pixel; foliage projective cover; leaf area index (LAI); height; and above ground biomass. The SDVP’s assessed in this trial work include statewide (4 products), Australia-wide (11 products) and global (4 products).

Initial comparisons shows that the majority of SDVPs focus on vegetation cover and extent, operate at variable spatial and temporal scales, ranging from daily to annual intervals and from local patches to continental extents. Most SDVPs have published and accessible algorithms, along with some form of validation. However, validation studies are not consistent and are challenging to compare. The reported limitations and assumptions concern factors such as inaccuracies in specific vegetation types, scale-dependent discrepancies, sensor-specific errors, temporal inconsistencies, and variability in methodological approaches. Priorities for future work emphasize standardized validation protocols, improved reporting of limitations, enhancement of spatial resolution, integration of multi-source satellite data, and addressing inconsistencies across vegetation property definitions.

The initial results will be presented in a form able to be communicated via web-page in a consistent, secure and accessible location, potentially www.eoa.org.au and published as a pre-print, with the intention to update on an annual basis. Follow on work is intended to progress this to global scales once the initial Australian work is complete.

Shelterbelts, also known as windbreaks or tree rows, can provide significant environmental benefits alongside gains in crop and pasture productivity. However, the extent and characteristics of shelterbelts across Australia remain poorly understood due to limitations in current national vegetation datasets — particularly the spatial and temporal resolution. Since the benefits of shelterbelts are location specific and highly variable, this data gap constrains our ability to quantify their current value and identify opportunities for expansion.

Recent advances in satellite imagery, including Sentinel-2 and PlanetScope, now allow for high-resolution mapping and monitoring of shelterbelts over time. Our project integrates this imagery with LiDAR from representative biomes, as well as elevation and hydrographic datasets, to produce annual shelterbelt classifications at 10 m resolution across Australian agricultural regions. We apply the latest research on shelterbelt function, along with landscape ecology frameworks, to automatically classify shelterbelts using factors such as the length, width, height, continuity and landscape position. We then combine these classifications with national wind models to estimate the area of sheltered vs. unsheltered farmland, and per pixel degree of shelter.

This presentation outlines our methodology and initial results. By enabling consistent, high-resolution mapping of shelterbelts, we aim to inform targeted adoption in areas where they offer the greatest benefits. This could improve food security, soil biodiversity, drought resilience, climate change mitigation, and other ecosystem services.

Post-disturbance forest recovery is often monitored using spectral indices that measure canopy ‘greenness’. However, this approach can be misleading, as a rapid return of leaf area (spectral recovery) often masks a persistent lag in the accumulation of biomass and structural complexity (structural recovery). This discrepancy hinders accurate assessments of ecosystem resilience and habitat restoration. This research proposes a novel framework to decouple these recovery processes by combining remote sensing data across multiple platforms and scales. Our objective is to develop and validate community-specific models that provide a more comprehensive quantification of regeneration. We will fuse high temporal resolution multispectral archives (Landsat/Sentinel-2) to track leaf area trajectories with high-resolution hyperspectral imagery for vegetation classification, and LiDAR data (GEDI, airborne, and drone-based) to measure canopy structure. This framework will be applied across the diverse landscapes of Kangaroo Island, still recovering from the catastrophic 2019-20 Black Summer fires, to compare recovery trajectories across its distinct vegetation communities, such as woodlands, mallee, and coastal heathlands. We hypothesise that recovery trajectories will not only diverge between spectral and structural metrics but also differ significantly among these communities, reflecting their unique post-fire regeneration strategies. This poster presents our conceptual framework and ongoing data acquisition strategy, establishing the scientific basis for a robust monitoring approach. By moving beyond simple greenness, this work aims to deliver the critical evidence needed to more effectively guide conservation actions and support Australia’s emerging nature repair markets.

Trait-based models are a powerful tool to predict vegetation responses to climate change, but they are currently limited by the lack of information about trait variation across life-stages. Traits sampled from a single life-stage are often used interchangeably to model plant responses. However, traits may differ between stages, with juveniles expected to show traits linked to ruderalism (R), while adults may exhibit competitor (C) or stress-tolerant (S) traits. Therefore, extrapolating traits from one life-stage to another is problematic and may lead to inaccurate model predictions. Here, I use Acacia and Eucalyptus - the two most diverse plant genera in Australia - as a model system to investigate developmental variation in CSR trait-strategies. Species in these genera display varying degrees of heteroblasty (distinct juvenile vs. adult leaf morphology) which may lead to different strategies across life-stages. By extracting information from the Austraits database, I tested whether key leaf traits— area (LA), dry matter content (LDMC) and specific leaf area (SLA) - vary across 214 Acacia and 622 Eucalyptus species. Preliminary results suggest that LMA is often higher in adults, LDMC shows no difference across stages, while LA patterns are species-specific. Only three species (A.aulacocarpa, E.miniata and E.platyphylla) have juvenile/adult information for all three leaf traits, allowing calculation of their CSR strategies. Juveniles exhibited higher R scores, but were not R-strategists per se, while adults combined C- and S-strategies. It remains to be tested whether such strategy differences are more prevalent in more heteroblastic species. This study highlights how juvenile traits are still largely underrepresented in the Austrait database, with only 3-9% of the total number of Acacia and Eucalyptus species having juvenile leaf trait data available. Given the developmental trait differences observed here, further data collection at early life-stages is recommended to improve the predictive capacity of current trait-based models.