Biodiversity protection through innovative decision tools
While society is heavily dependent on healthy, thriving ecosystems, biodiversity conservation and climate change mitigation are not always at the forefront of policy and decision-making. “Biodiversity has certainly gained in visibility in recent years at all levels,” notes Pierre Bonnet(opens in new window), researcher at the French Agricultural Research Centre for International Development(opens in new window) (CIRAD) in Montpellier. “However, it is still often treated as a secondary consideration, especially when weighed against immediate economic or political concerns.” Indeed, one challenge is making biodiversity information usable at the scale and speed that decisions are made.
Creating biodiversity decision support applications
To address this, Bonnet and his team set out to create support applications designed to guide decision-making and safeguard biodiversity for the future, through the EU-funded GUARDEN(opens in new window) project. The GUARDEN system they developed draws on recent advances in citizen science, deep learning, Earth observation (EO) and hybrid modelling to offer accessible and usable information. The system is based on the observation that biodiversity concerns are not just limited to species conservation, but also underpin ecosystem services essential to agriculture, water regulation and climate change mitigation. GUARDEN decision support applications (DSAs) were developed, comprising multiple analytical components. These include species identification, habitat mapping and scenario modelling. These modules are interoperable and modular, meaning users can combine the tools needed for their unique decision-making and obtain tailored outputs. “These DSAs are designed to support land-use planners, policymakers and other stakeholders in making biodiversity-conscious decisions,” adds Bonnet. The tools incorporate multiple technological layers, including EO, and state-of-the-art artificial intelligence (AI) models for species identification – such as classification of plant communities from georeferenced images and the identification of bird species from audio recordings. “These approaches allow for scalable, semi-automated biodiversity monitoring in both natural and human-dominated ecosystems,” says Bonnet. The DSAs are currently being tested in several sector-specific case studies, notably in the fields of urban planning and infrastructure development.
Implementing AI and Earth observation in decision-making
Among the most significant results is the demonstration that advanced technologies, such as AI and EO, can be implemented in ways that directly support biodiversity-aware decision-making. Some of these tools, such as GeoPl@ntNet(opens in new window), have already demonstrated the potential for both large-scale and fine resolutions. Tools from the project are also available online(opens in new window), contributing to the European knowledge commons and enabling further innovations. “Some of the project’s tools have been tested outside the European context, for example in Madagascar(opens in new window), to confront our approaches with very different environmental, political and economic contexts,” explains Bonnet.
Driving global research collaboration
The development of the tools also successfully brought together a large community across Europe. Several tools and services are being further refined and tested with European and international institutions, including with the Biodiversa+ network(opens in new window), the Joint Research Centre (JRC) and within the context of the World Flora Online consortium. “These interactions are helping shape the system for new real-world operational use and policy integration,” says Bonnet. “The system is being built not as a one-off prototype, but as a flexible, evolving infrastructure that can adapt to emerging needs.”
