Periodic Reporting for period 1 - MAMBO (Modern Approaches to the Monitoring of BiOdiversity)
Berichtszeitraum: 2022-09-01 bis 2024-02-29
MAMBO overall objectives:
1. Develop, evaluate and integrate image and sound recognition-based AI solutions for EU biodiversity monitoring from species to habitats.
2. Develop, test and deliver high spatial resolution regional EU habitat extent maps (satellite remote sensing) and site-specific (e.g. Nature 2000), but EU consistent, habitat condition metrics (airborne LiDAR and drone data).
3. Promote the standardized calculation and automated retrieval of habitat metrics using in-situ observations, deep learning and remote sensing.
4. Co-design MAMBO’s novel ecological monitoring tools with researchers, policy makers, citizens and other stakeholders, evaluate their costs and benefits and make them widely available.
5. Build a new global community of practice for the development and application of these cutting-edge technologies through proof-of-concept implementation across the EU.
6. Test and implement existing and MAMBO’s novel tools for upscaling and contribute to an integrated European biodiversity monitoring system with potential for dynamic adaptations.
The technical developments and research networks formed through MAMBO will pave the way for a revolution in the scope and capacity for monitoring species and habitats, especially those for which knowledge gaps still exist. MAMBO thus supports the implementation of the Green Deal, the EU biodiversity strategy 2030 and the Birds and Habitats Directives and will contribute to the Commission's new governance framework for biodiversity under the future EU Nature Restoration Law. We will make a substantive contribution to understanding climate change and land-use impacts on biodiversity and advise beneficiaries at global, regional and national levels.
We have tested algorithms for fine grained image and acoustic classification. We have expanded the training data by developing a network of European biodiversity portals. Training data for a sound recognition model of European grasshoppers and for image recognition for insects in camera trap images has been published. We have developed a pipeline for analysis of images of entire plant communities from single images of vegetation plots.
We delivered a prototype deep learning framework for habitat extent mapping and created a first set of habitat maps. We completed a review of habitat condition metrics and have identified key metrics suitable for LiDAR and drone remote sensing. We collated and pre-processed existing LiDAR data from four MAMBO sites and developed workflows for deriving tree cover, linear features and dead wood. We completed the evaluation of current challenges for upscaling the use of (sub)national airborne LiDAR or site-specific drone collected data for EU-wide habitat monitoring.
We have initiated a dialogue with managers of demonstration sites, to clarify their needs, and their levels of advances in the appropriation of AI-based IT tools. We have started to test how data from new technologies can be combined with data from traditional sampling methods for species monitoring. A high-throughput workflow, which will be used for upscaling LiDAR vegetation metrics has been published. The suitability of pollinator camera observation data for the quantification of pollinator climatic niches quantified and predictive ability has been tested.
We have worked to ensure that the project’s findings have a meaningful impact on policy and practice. We have established MAMBO's identity and a strategy for communication and dissemination. Our website and engagement in social media has raised awareness of MAMBO's goals and achievements. Through collaboration with EU initiatives, we have worked towards ensuring that MAMBO's contributions have lasting effects. We have also worked towards guidance for the implementation of MAMBO's insights into global assessments such as IPBES and IPCC.
We have ensured the effective running of the MAMBO project and an adequate flow of information between partners. We have developed the data management plan for the MAMBO project in accordance with the FAIR principles. We have appointed an independent ethics advisor for MAMBO.