Advancing Biodiversity Research and Policy through Innovative Technologies and Collaboration

Group Photo
23 February 2024


BioDT project leader Jesse Harrison took part in the seminar "Towards a Digital Twin for Biodiversity” organised by the BioDT project partner University of Tartu in collaboration with the Estonian Research Council on Thursday, 8th February 2024 at the Permanent Representation of Estonia to the EU in Brussels.

The University of Tartu emphasised in their introduction speech that it is our collective responsibility not to forget about biodiversity, climate change and land use.

Digital Twins on the European Union’s agenda

Signe Ratso, Deputy Director-General for Innovation, Prosperity and International Cooperation at the Directorate-General for Research and Innovation, confirmed that the EU has allocated budget to protect biodiversity and promote biodiversity restoration. The EU recognizes the necessity for transformation, hence its endorsement of ambitious objectives outlined in initiatives such as the European Green Deal. She indicated that the policies must align with scientific findings to truly achieve success.

Signe Ratso pointed out that the utilization of digital twins and AI has opened up a realm of fresh opportunities, expanding the boundaries of scientific exploration. The development of digital twin technologies can benefit from AI as AI algorithms are capable of handling extensive volumes of data. AI is seen as a new ally in the endeavour to protect and restore biodiversity.

Robert Konrad, Adviser for Biodiversity at the Directorate-General for Environment, reminded that digital twins are part of EU sustainability agenda via the Destination Earth (DestinE) initiative, implemented by ECMWF, ESA, and EUMETSAT, with examples such as the Digital Twin on Weather-induced and Geophysical Extremes as well as Climate Change Adaptation Digital Twin.

The digital twins facilitate the transition from hypothetical ‘what if’ scenarios to policy formulation as they help understand what is happening in nature and reconstruct phenomena beyond direct observation. This will enhance our situational awareness of the environmental conditions.

Robert Konrad drew attention to the necessity of having multiple digital twins. By integrating various digital twins, the aim is to have one digital twin that comprehensively analyses and describes Earth's phenomena. For instance, determining appropriate course of action for desired outcomes and assessing real-life impacts. This holistic approach is valuable both to scientists and policymakers for the ultimate benefit of our environment.

Integrating environmental DNA and AI

Urmas Kõljalg, Director of the University of Tartu Natural History Museum and Botanical Gardens, emphasised the importance of communicating environmental DNAs (eDNA). He discussed PlutoF, a virtual research environment which is meant for managing biodiversity data and which, among other services, features an analysis module for molecular sequence identification and species discovery from eDNA samples. He stressed the importance of providing biodiversity related services not only for scientists but also for citizen scientists and industry.

Erik Kristiansson, Professor in biostatistics and bioinformatics at Chalmers University of Technology, presented recent advances in AI development, highlighting the untapped potential of AI in biodiversity research. Unfortunately, biodiversity does currently not drive AI development. For example, large language models for eDNA exists which have the advantage of taking a sequence and making the identification of taxa. The more data we have, the better identification. However, there are wide sources for the data with different data types that can be fragmented and incomplete. Data fusion models can compensate for missing observations, help with errors and inconsistencies. The results have proven to be very promising. Yet, certain data-related challenges persist, such as establishing improved AI systems, which will require substantial efforts as well as combining AI and domain-specific, thus interdisciplinary, knowledge.

Overcoming biodiversity challenges through citizen science

Jana Wäldchen, Leader of the Biod.AI.versity Observation & Integration project at the Max Planck Institute for Biochemistry, highlighted the costliness of biodiversity-related data collection, the decrease of species experts and the necessity of standardized data collection across various regions and ecosystems. Advancement of sensor technology has facilitated data collection, yet manual processing of such vast amounts of data is not possible. AI aids in, for example, monitoring of animals and in observing flowering periods and species interactions. An example of this is the Pl@ntNet app, which identifies plants from pictures, including invasive species, and is part of a citizen science project on plant biodiversity. Active engagement of people is important for encouraging individuals to take an active role in protecting the environment.

Enhancing collaboration and interoperability for biodiversity conservation

The seminar concluded with a panel discussion chaired by Dimitris Koureas, Director of Distributed System of Scientific Collections, Naturalis Biodiversity Center, and including panellists Jesse Harrison, Robert Konrad, Urmas Kõljalg and Erik Kristiansson.

Dimitris Koureas stressed that our capacity to document biodiversity before it diminishes is not advancing quickly enough. He started the discussion with a question: If we are moving towards AI for biodiversity discovery, how can human intelligence and AI work together?

The panellists deliberated on a key challenge: the reliability of results produced by AI methods, for which human expertise for training and validation is needed. Human involvement extends still from input to output stages.

Robert Konrad underscored that AI enables us to gain a better understanding of the drivers behind species loss, thereby facilitating the contribution to legislation such as on pesticide use. Erik Kristiansson highlighted that one important way how can AI assist in interpreting data and aid researchers in making data FAIR is by extracting information from old papers. This is data that would otherwise be unacceptable.

The panellists pointed out the uncertainty regarding biodiversity data gaps, stressing the insufficient attention given to identifying missing data. They noted the scarcity of species absence data, while there is a predominant focus on presence data. It was proposed that to address knowledge gaps, there should be increased collaboration between researchers from different regions, and Horizon Europe and future programs could help in facilitating this. Also, citizen science initiatives can play a pivotal role in bridging knowledge gaps.

It was emphasised that when it comes to digital twins developed as part of various projects, focusing on interoperability, including with DestinE, is crucial.  Also, the disparity between the views of the scientific community and those of policymakers was highlighted. Thereby, exploring strategies to effectively bridge the gap between these two realms for improved policy recommendations was underlined as a priority. All in all, there is a need to see the impact of the work that is being done.

Read more about the topics discussed here.

Many thanks to the University of Tartu for the use of the image accompanying this article.