Researchers develop Biodiversity Digital Twins to model our planet’s life

Biodiversity is essential for the processes that support all life on Earth. It provides critical resources such as food and energy, and supports ecosystem health. However, climate change, deforestation, and pollution are destroying habitats, altering ecosystems, and eliminating—or introducing—species that are fundamental for the planet’s biosphere.

To tackle the challenges caused by environmental change and human activities on biodiversity, a consortium of 22 partners led by CSC—IT Center for Science, home of the EuroHPC LUMI supercomputer, is developing Biodiversity Digital Twins (BioDT) as a result of the European Commission’s initiative.

The BioDT project aims to revolutionize our understanding of biodiversity dynamics by integrating advanced modeling, simulation, and prediction capabilities. By combining and improving digitally available data and models, BioDT offers approaches for sustainable biodiversity management and ecosystem conservation. BioDT’s combines expertise in biodiversity, ecological modeling, FAIR data, high-performance computing, and artificial intelligence.

BioDT aims to enhance the accuracy and predictive performance of biodiversity models through iterative development and validation against independent data. This approach can be critical for developing decision support tools and policy development.

By continuously updating data, BioDT will provide real-time predictions of biodiversity patterns and processes through interactive maps and summaries. The consortium leverages existing technologies and data from major research infrastructures (GBIF, eLTER, DiSSCo, and LifeWatch ERIC) to achieve this goal.

The project’s impact extends to addressing critical issues, including the impact of environmental change on species and ecosystems, food security, and the implementation of the EU and international policies. The project contributes to the UN Sustainable Development Goals 2 (Zero Hunger), 3 (Good Health and Well-being), 13 (Climate Action), and 15 (Life on Land).

BioDT develops prototype Digital Twins for biodiversity conservation

In order to test its modeling system, BioDT is developing 10 prototype digital twins (pDTs) focused on species and ecosystems of high conservation and policy concern, such as invasive species, pollinators and grasslands.

The pDTs are divided into four main groups:

• Species Response to Environmental Change: focus on the interactions between species and ecosystems. By incorporating temporal dynamics rather than pure space-for-time substitutions, BioDT improves temporal predictions and accuracy. Different sources of uncertainty are quantified using extensive geographic data combined with high-resolution time-series data in a single modeling framework.
• Genetically Detected Biodiversity: addressing food security and challenging environments by integrating genomic methods based on DNA data with traditional biodiversity data. These twins focus on crop wild relatives and other genetic resources for farming and food security, as well as DNA-detected biodiversity in poorly known habitats.
• Dynamics of Species of Policy Concern: applying modeling and high-performance computing to invasive and alien species recognized at EU and national levels. This twin involves using current species occurrence data, and tackling crucial environmental conditions and invasive effects on native taxa and ecosystems.
• Influence of Species Interactions: predicting disease outbreaks using vector species and exploring the patterns and processes of insect pollinators. Work on interaction twins involves further development of data exchange models and establishing temporal historical reference points through digitization of collection specimens.

The pDTs aim to make essential datasets, best practices, expertise, and lessons learned available and ready for use by researchers and research infrastructures in implementing the use cases, while providing.

The pDTs test the models’ predictive performance and data availability scenarios, and apply them to address biodiversity challenges through scenario simulations, predictions, and biomonitoring methods. This iterative approach aims to integrate and compare the predictive performance of various modeling approaches, stimulating the development of next-generation prototypes.

Building Biodiversity Digital Twins: A BioDT collection of scientific papers

To further advance the development and reliability of Biodiversity Digital Twins, the BioDT team has produced 10 scientific papers, compiled in the “Building Biodiversity Digital Twins” issue of the open-science scholarly journal Research Ideas and Outcomes (RIO).

“The collection offers an in-depth understanding of the conceptual and technical advancements achieved towards developing digital twins for a wide range of biodiversity topics. Through the BioDT project, we are enabling a broad audience to interactively understand and predict biodiversity changes across space and time.” says Gabriela Zuquim, Scientific Coordinator at CSC for the BioDT project

The collection serves as a centralized access point to project outputs by the BioDT initiative. Publication of rather unconventional and not traditionally published research outputs is in fact among the unique features of the open-science RIO journal. Another feature is the possibility of individual publications to be mapped to the SDGs they contribute to, thereby further underlining their significance.

In the case of BioDT, RIO has made it possible for the project team to illustrate the process of prototyping Biodiversity Digital Twins in the format of a peer-reviewed scientific article, thereby ensuring its discoverability, credibility, citability, reusability and long-term public availability.

By opting for this transparent approach to sharing their scientific work that has withstood the rigor of formal scientific review, the BioDT project ensures that future scientists can make better and more efficient use of the models developed by the consortium’s researchers, data, and cutting-edge technology.

For example, one publication describes the HONEYBEE Prototype Digital Twin. The prototype will allow, after the ongoing calibration with land use and hive weight data, predictions of honeybee population dynamics, mite infestation and honey production. The model was developed based on a previously developed one, devised to simulate foraging of a single bee colony.

By using the prototype digital twin, users can interactively apply the model on various time and geographic scales ranging from local sites to whole regions or even country level. Thus, it can become an essential tool for the assessment of the viability and productivity of honey bee colonies around Germany, regardless of the specificity of landscapes and management strategies.

Our vision is that the assessment can even be run to take into account different climate-change scenarios. The publication also provides guidelines to potential users of the prototype.

The authors of the paper, led by Dr. Jürgen Groeneveld (Helmholtz Centre for Environmental Research—UFZ, Germany) reminds that despite honey bees “being a managed species, they are severely affected by climate change, emerging parasites and diseases, modern agricultural land use and possibly inappropriate beekeeping practices,” while going on to cite worrying data about the trends in both Europe and the U.S..

Similarly, other publications already available from the collection address equally crucial and pressing issues with impact on a global scale, including disease outbreaks, crop management, invasive species, bird and vegetation dynamics.

“The Building Biodiversity Digital Twins collection of project papers suited our needs perfectly,” said Dmitry Schigel, GBIF Scientific officer and a coordinating editor of the collection. “The project team agreed to capture the project’s iterations and reveal our two-thirds stage prototypes two years into the project with one more to go.

The innovative platform that the Pensoft’s RIO journal provides lets us describe our progress in a less formal but still peer-reviewed setting. Thanks to the efficient work of the author teams, reviewers and co-editors, this special issue came together quickly and now enables our prototype digital twin teams to attract and process feedback from broader audiences.”