As quoted in a recent paper in Nature Communications, the growing amount of the collected environmental data is not optimally used: “there is a mismatch between the ever-growing volume of raw measures (videos, images, audio-recordings) acquired for ecological studies and our ability to process and analyse this multi-source data to derive conclusive ecological insights rapidly and at scale”[[ Tuia, D., Kellenberger, B., Beery, S. et al. Perspectives in machine learning for wildlife conservation. Nat Commun13, 792 (2022) https://doi.org/10.1038/s41467-022-27980-y.]]. In the European Union, there is already a range of group of experts monitoring species and habitats, including in the view of reporting under the Birds and Habitats directives. However, the generated datasets are not sufficiently accessible (too many small, isolated communities of practice, different servers, different data access methods, different formats, rarely accessible through web-services) and too often not well known or advertised outside of their original circle of experts: the access to the results (consolidated data, statistics, maps) of these field surveys should be significantly concentrated behind single entry points. Also, the access to modern technologies (e.g. image recognition, sound analysis, high-throughput DNA-based techniques, usage of AI, usage of space, etc.) too often represents an important effort for each group of experts, beyond their environmental expertise. As a result, the technological developments remain an important effort for each group, while the solutions should better be provided as a service (to be configured to the need of each group) and mutualised. The natural domain being very large and sometimes difficult to access, the existing databases are still not dense enough, in terms of spatial and temporal coverage: many species and habitats are insufficiently covered (and sometimes not monitored at all), resulting in information gaps. Also, scarce samplings do not allow to distinguish non-presence from a lack of/insufficient/inadequate fields visit. A massive use of automated, and potentially mobile, sensor technologies (such as, but not limited to, images, video, sounds/ultra-sounds recording, spectral signatures, structure description by lidar, environmental DNA sampling, etc.) the use of remote sensing technologies (e.g. to over large areas, monitor environmental condition) and associated with processing algorithms (in particular, but not limited to, deep learning and AI processing algorithms) is therefore needed. The goal of this topic is to facilitate the access to data, encourage the usage of automated/robotic/space data collection systems for data collection, encourage community approaches for the exchange of data and good practices (in particular for data processing).
Proposals should address Area A or Area B as follows. The Area should be clearly indicated on the application.
- Area A: a project focussing on data harvesting through high-throughput methods (as described in the introduction, e.g. environmental DNA, sound/image analysis, lidar, spectrometry, usage of mobile platforms, remote-sensing, etc.), analysis and interoperability solutions, with the goal of concentrating the information in a single access point, and lowering the technical hurdle for the biologist and managers of natural sites, offering the best solutions in a ready-to-use form;
- Area B: a project focussing on new robotic solutions, including mobile, to improve the efficiency of biodiversity related solutions, allowing to improve the performance of the field campaign, with denser information of species and habitats.
Area A: data harvesting, analysis and interoperability solutions
The successful proposal is expected to address the needs in terms of IT solutions, to increase information density, in terms of species and habitats sampled, territory coverage, timeliness, and accuracy.
As a result, much denser data collections should be available through a common data portal. The successful proposal should demonstrate the feasibility to combine different sources of information, for example to assess the conservation status of habitats or species. In that respect, several approaches could be tested, from data combinations defined by expert rules, and data storage formats, to machine learning or data-mining technologies. Such digital solutions could support the definition of conservation measures and management plans, and the monitoring and forecast (though model ingesting in-situ observations) of their progress to their objectives, at site, regional and national levels. Furthermore, the results could be used by member states for their formal monitoring and reporting obligations, or to check and enhance the performance of Nature Based Solutions.
The successful proposal should:
- Ensure interoperability of available data, enabling EU-scale information systems by developing solutions to connect and harvest data from already existing data bases. This will guarantee information fusion and support third party usage of the data.
- Develop cost-effective and easy-to use tools and software to collect and analyse different existing data sources and formats (in vivo data, photographs, sound recordings, lidar, spectrometry, eDNA, satellite images etc.), to facilitate cost-effective data analysis, map and link existing databases and provide algorithms to better analyse them.
- Develop data hosting and data processing solutions to extract information on populations (such as diversity, counts, trends), habitats (such as identification, area covered, and area change in time), assessment of conservation status and trend, information of species and habitats health conditions, degradations, and destructions (natural or human-driven). The accumulation of information should allow synoptic analysis of species and habitats, allowing to detect hot spot of issues and trends. Innovative solutions, such as data mining, remote-sensing and AI approaches need to be considered.
- Develop a solution to host, process, analyse and search available data in relation to protected habitats and species (including protected sites management information, their conservation objectives and measures, and restoration actions).
- Analyse and define infrastructure solutions, that would let biologists and managers of natural sites quickly create a dedicated working framework, furbished with all data harvesting, processing, sharing solutions. In this approach, the future European Green Deal data space should be considered as a potential common solution, or part of the solution.
- Develop tutorials for practitioners, based on academics and industry knowledge, on how to best use existing databases and data harvesting, data analysis and data sharing solutions. The tutorials should help the users to quickly set up and use their working environment.
- Propose easy-to-use solutions to utilise robotic sensors and Internet of Things (IoT): automated sensors, automated sampler, including mobile sensors (terrestrial, aerial and under-water) and animals tagging solutions, data sharing through wireless communication systems, to support a systematic data collection. Such approach should help better mapping the known/unknown and significantly increase the density of collected data, spatially and temporally.
- Analyse the conditions under which data, raw data acquired from sampling, data coming from existing databases and data resulting from processing can be shared. A clear data sharing framework, accommodating special needs, simple to use in practice, supporting open data policies, and enabling the broadest usage whilst encouraging the largest community to contribute, should be defined. Special attention will be paid to endangered species and sensitive species (in the sense of the Birds and Habitats Directives) for which the shared data needs to be controlled, and methods for effective detection of invasive species by high throughput search would be encouraged.
- Enable EU Member States, Associated Countries, and accession countries to coherently set conservation objectives, preparing management plans, manage shared habitat types and species, deal with similar conflicts and socio-economic dimensions, permitting procedures, spatial planning, with a focus on implementing the Birds and Habitats Directives and their Natura 2000 network.
- Fully exploit and build complementarities with the ongoing work regarding the establishment of the European Open Science Cloud and interact with relevant projects developing metadata standards and added value tools to ensure interoperability within and across fields of study.
- Contribute to a web of FAIR data and supporting services that enable an interconnected disciplinary ecosystem that allows stakeholders to share digital objects and build on them in a seamless fashion.
- The architecture for a unified EU web-GIS with all the data collected from the Directives should be considered. In that matter, the proposed system should allow the member states sharing their habitats and species maps, and in particular the habitats maps used to designate their Natura 2000 sites, as well as subsequent updates. Also, the platform should help collecting information to update habitats and species maps, in order to obtain a common knowledge database about habitats and species, and their evolution, in relation to the Birds and Habitats Directives. The platform should as well foster the implementation of open data best practices at European level and across boundaries.
- Automatic translation functions should be offered by the platform to better connect EU Member States, Associated Countries and Accession Countries to support them in the implementation of the legislation on nature protection (such as the Birds and Habitats directives, the Invasive Alien Species regulation or the Marine Strategic Framework Directive).
Proposals should consider the possibilities offered by the future “Green data spaces” (CNECT). The DEP CSAs on the “preparatory actions for the European Green Deal Data Space” (exploring cloud-to-edge solutions, platforms and initiatives for data storage, exchange, and analysis as good practices for setting up the data spaces) are expected for Q4 2022-Q2 2024 and the “data spaces support centre” will start delivering on architectural blueprints in late 2023 and onward.
Proposals should earmark the necessary resources for cooperation and networking activities. Proposals should link to other relevant Horizon 2020 and Horizon Europe projects and initiatives, such as BiCIKL, EuropaBON, BioDT and connect to existing European Biodiversity data infrastructures including DiSSCo, eLTER and LifeWatch, where relevant. Proposals should also connect with relevant projects under Horizon Europe topics, such as HORIZON-CL6-2021-BIODIV-01-01: European participation in global biodiversity genomics endeavours aimed at identifying all biodiversity on Earth.”, HORIZON-CL6-2021-BIODIV-01-02: Data and technologies for the inventory, fast identification and monitoring of endangered wildlife and other species groups, HORIZON-CL6-2021-BIODIV-01-07: Ecosystems and their services for an evidence-based policy and decision-making and HORIZON-MISS-2021-OCEAN-02: Protect and restore marine and fresh water ecosystems and biodiversity. Projects using satellite data should link to HORIZON-CL6-2021-GOVERNANCE-01-14: User-oriented solutions building on environmental observation to monitor critical ecosystems and biodiversity loss and vulnerability in the European Union.
The possible participation of the JRC would help ensure that the methodologies proposed can support environmental compliance assurance, particularly by leveraging geospatial intelligence.
Collaboration with the European partnership on biodiversity “Biodiversa +” should be explored, as needed.
Area B: new robotic sensors for biodiversity
To increase the density of species and habitats observations across the EU territory, new robotic, and possible mobile, solutions need to be developed.
The proposed innovative solutions should:
- Be ready to use, easy to deploy and operate in natural environment.
- Consider automated solutions, and mobile platforms (land, air, water and under water) carrying sensors (such as, but not limited to, image, sound, lidar, spectrometry, eDNA, etc.) should be designed with fields campaigns in mind, in particular in terms of autonomy (energy, autonomy of moving and sampling decisions). Improvements in terms of species tagging, and species-carried tracking or telemetry devices should also be considered.
- The project should focus on innovative sensors that would allow significantly increasing knowledge in biodiversity, or bringing new information about the species and habitats conservation status, and increase spatial and temporal coverage, and to facilitate access to environments that are difficult to sample.
- Propose a large degree of data collecting automation and compatibility with the system described in project 1.
- The project should generate at least 1 innovative prototype of robotic/automated sensor and 1 innovative prototype of mobile solution, demonstrating improved performances compared to the currently available solutions.
- The project should analyse the conditions and costs of the production of the robotic system, as well as the conditions and costs of its usage and maintenance.
The project “Natural Intelligence for Robotic Monitoring of Habitat” could provide hints about the usage of mobile robotic sensors.
International cooperation is encouraged.