This session welcomes all submissions presenting DTs or prototype DTs, irrespective of their development stage or maturity level. Presenters are invited to emphasise how their DTs are or can be used by practitioners and decision makers for better informed conservation decision-making. This session also welcomes concepts and ideas about how DTs can foster stronger connection between science, society and practice. Lastly, this session invites ideas or examples of DTs that nurture interaction and engagement of people with the natural environment.
Being a new modelling paradigm for ecology, Digital twins (DTs) have the potential to change the status quo of decision making in nature conservation. By combining process-based predictive models with continuously updating data, DTs offer real-time predictions and give critical up-to-date insights in states and trends of natural systems, which helps decision makers in making more informed pro-active decisions based on current trends. But there is more: ecological DTs have the potential to reach a wide public audience, and offer the ability for people to interact with real-time models of the natural environment. Thereby engaging a large group of people with current phenomena in the natural world.
We present ongoing work of the Crane Radar, which is considered the first fully operational DT in the field of ecology. It runs continuously on a server, tracking and predicting the whereabouts of cranes on their annual migration between their breeding and wintering sites. An interactive map allows birdwatchers and nature enthusiasts to navigate and get a spatial awareness of the crane migration in real-time, which helps them increase their chances of seeing the migration. The Crane Radar reaches over hundred thousand visits per day during peak migration, showing its potential to engage with a wide audience about what is happening in nature.
We present our latest Crane Radar project, which started in October 2025, that goes even further in user engagement; co-creating the radar with users. We will discuss results and insights from two interactive workshops with stakeholders that give us input for further developing the Crane Radar. Addressed in these workshops, specifically, are the users’ understanding of the visualisations of the crane radar, and their understanding of the concept of model uncertainty. By explicitly focusing on the understandings, needs, and requirements of end-users, these topics give us particularly relevant insights for the development of user-oriented digital twins for nature conservation decision making.
How to cite: de Koning, K.: Co-creation with digital twins – an example of the crane radar, World Biodiversity Forum 2026, Davos, Switzerland, 14–19 Jun 2026, WBF2026-35, https://doi.org/10.5194/wbf2026-35, 2026.
It is widely recognized that the management of German forests must be adapted to the challenges of climate change, especially in spruce-dominated forests, which have suffered significantly from drought and bark beetle infestation in recent years. Such a change in management is necessary to preserve the diverse ecosystem services provided by forests, such as wood production, carbon sequestration, and recreational value. In order to find a suitable management approach that is both socially acceptable and feasible, we are pursuing the Living Labs approach. To this end, we are working with stakeholders to develop a series of management scenarios for the forest in the Harz Mountains that meet the needs and expectations of the various stakeholders. We will present how we implement these management scenarios in a digital forest twin that uses the iLand landscape forest model as the virtual component. Although the iLand model is individual-based, it can simulate entire landscapes over centuries. iLand also allows forest management measures and biotic and abiotic disturbances to be taken into account in detail. The model is fed with environmental and forest growth data such as tree height, diameter increment, and weather variables. The model is initialized based on forest plots selected and established by our project partners, with three management modes: 1) natural forest regeneration (no management), 2) extensive management (natural regeneration, native tree species only), and 3) intensive management (removal of deadwood, including non-native tree species). In addition, we are working on initialization workflows based on existing forest classification and forest height products derived from remote sensing data to enable the application of the Digital Twin for every forest in Germany. The results of the Digital Twin are discussed and co-evaluated with stakeholders, and the underlying management strategies are adjusted as necessary. This process is repeated several times if necessary to arrive at management recommendations for a strategy to transform German forests into more climate-resilient ecosystems that can be supported by as many stakeholders as possible.
How to cite: Groeneveld, J., Ehrhardt, S., and Fischer, R.: Developing forest Digital Twins to foster the transition towards climate resilient forests, World Biodiversity Forum 2026, Davos, Switzerland, 14–19 Jun 2026, WBF2026-388, https://doi.org/10.5194/wbf2026-388, 2026.
Biodiversity research increasingly relies on digital twins—high‑fidelity, data‑driven simulations that replicate ecosystems across spatial and temporal scales. While these twins excel at integrating heterogeneous observations (e.g., remote sensing, genomics, citizen science), they often remain limited by static model architectures and manual parameter tuning, constraining rapid hypothesis testing and adaptive management. We propose an inference service supporting agentic Large Language Model (LLM) workflows with High Performance Computing (HPC) resources, enabling a new range of research directions in digital twins, such as accelerating development of data processing pipelines, agentic modelling, information exchange, natural language results interpretation, accelerating adoption, and many more.
Presented service facilitates usage of the HPC resources by exposing an OpenAI API which can be used with any OpenAI API compatible client. This means it exposes a diverse set of possibilities while keeping the computation and data at secure infrastructure, boosting not just research, but also European sovereignty. The most suitable application of this service are ones that are sending large number of requests in short amount of time such as agentic workflows. This would mean that the agents could interact or enhance digital twins, create an interaction layer for general public or non-technical stakeholders either ingesting natural language queries or create interpreted model results in natural language. Additionally, for modellers it allows to combine classical models and LLM agents to create a dynamic semi-autonomous hyper-parameter search or simulation.
We tested this service with agentic simulation for disaster relief in wildfire simulation. This simulation built upon CREW-Wildfire framework simulating a forest wildfire and actions of the rescue team. Our test showed this service works and new agentic framework such as CrewAI and Langchain can provide really good results in this benchmark. Of course the results also depends on the LLM models used.
Technically the service is build upon the same technologies which were leveraged in the BioDT prototype digital twins to interact with the HPC thus providing secure access to these resources. This removes a problem of choosing either an external LLM provider or using small local models with reduced capabilities during the research.
How to cite: Martinovic, T., Matus, A., Bot Goncalves, K., and Martinovic, J.: AI inference service in Digital Twins, World Biodiversity Forum 2026, Davos, Switzerland, 14–19 Jun 2026, WBF2026-553, https://doi.org/10.5194/wbf2026-553, 2026.
