Understanding Terrestrial Biodiversity Dynamics in the Anthropocene to Take Better Action

At a time marked by the climate emergency and rapid biodiversity loss, it is crucial to innovate in how we describe, understand, and predict the evolution of the living world. This is precisely the core objective of DynaBIOD, an ambitious French interdisciplinary research program that places plants and terrestrial invertebrates at the center of its work.

Plants make up the vast majority of terrestrial biomass; they are the foundation of life and our ecosystems. Yet, significant gaps remain regarding the monitoring of their temporal dynamics and their complex interactions with fauna (such as pollinators or soil organisms).

DynaBIOD aims to bridge these gaps by leveraging breakthrough technologies (genomics, artificial intelligence, automated sensors) and unlocking the immense wealth of historical natural history collections alongside citizen science data.

The DynaBIOD project strongly resonates with the mission and expertise of the Pl@ntNet platform. Here are the three major points of convergence:

  • Artificial Intelligence and visual recognition at the core of the project:
    DynaBIOD explicitly positions Pl@ntNet as a pioneering and acclaimed initiative in the field of AI for biodiversity. The program will build upon these advances to develop new tools for automated species identification and even the automated detection of ecological interactions.
  • The creation of an unprecedented “multimodal reference database”:
    It is no longer just about matching a name to an image, but about building an extensive database linking every species to its physical characteristics (traits), DNA, photographs, geographic distribution, and interactions with other species. Pl@ntNet will play a key role in populating and validating current image and occurrence databases.
  • Leveraging citizen science:
    The data collected daily by Pl@ntNet users (as well as through other initiatives such as iNaturalist or Vigie-Nature) will be directly integrated by researchers to model current species distributions, understand their responses to global changes, and design future conservation scenarios.

The 4 Major Scientific Challenges of the Project

To achieve its objectives—ranging from data compilation to public decision-making support—the program is structured around four main challenges:

  1. Creating a comprehensive multimodal catalogue:
    The goal is to create a unique catalogue in France for plants and invertebrates. This reference database will combine taxonomic, genetic (DNA sequencing of museum specimens), morphological, photographic, and acoustic data. Image datasets will be labeled by professional and amateur taxonomists, relying on convolutional neural networks to refine machine learning models.
  2. Standardizing innovative monitoring tools:
    To monitor biodiversity on a very large scale, DynaBIOD will standardize innovative methods such as environmental DNA (eDNA, which identifies species from soil or water samples), eco-acoustics, and smart camera traps. The goal is to infer not only the presence of species, but above all their direct and indirect ecological interactions over time.
  3. (Re)constructing long-term time series:
    The project will reconstruct long-term time series (spanning several decades to centuries) by delving into rich natural history collections (herbaria, preserved specimens) and revisiting the field. This will allow researchers to compare past biodiversity with current biodiversity monitored by networks of professional and citizen observers.
  4. Modeling biodiversity trajectories and guiding decision-making:
    By aggregating all these data, the ultimate objective is to model biodiversity trajectories. By cross-referencing trends in flora and fauna with human-induced pressures (climate, land use, pollution), DynaBIOD will provide public decision-makers with action scenarios and optimized conservation strategies for the decades to come.

A “Science-to-Action” Approach

DynaBIOD is not merely a fundamental research project: it is a program driven by public action. Supported by five cross-cutting hubs (Genomics, Collections, Data, AI & Statistics, and Knowledge Transfer), the project will ensure that the generated knowledge and models are directly translated into clear indicators for policymakers, protected area managers, the private sector, and citizens.

To learn more:
https://www.cnrs.fr/fr/nos-recherches/france-2030/pepr/dynamiques-biodiversite-terrestre-dynabiod