It combines contributions from the JUNEAU and JANUSS projects with the data infrastructure and predictive models developed through DATA and PREDICTION.
These components are integrated within a shared methodological and technical framework, designed to support the development of environmental digital twins that can evolve alongside scientific advances and changing needs for environmental monitoring and analysis.
The result is a demonstration platform showcasing the analysis, simulation and forecasting capabilities developed within JUNON.
Methodological foundations and early prototypes
The initial stages of the project focused on defining a conceptual and methodological architecture to support the programme's future digital twins.
This architecture is based on the ENVRI Reference Model (ENVRI RM), an international reference framework for environmental research infrastructures. It provides a structured approach to integrating the resources required to build digital twins, including environmental data, scientific models, simulation scenarios and digital services.
The framework addresses several key aspects:
- Integrating complex environmental datasets, often involving time-dependent relationships;
- Ensuring interoperability between different systems and components;
- Establishing data exchange mechanisms and communication protocols;
- Documenting data sources, use cases and the functionalities of each prototype.
The first implementation of this architecture led to the development of a demonstrator focused on water-related forecasting. It provided an opportunity to test how environmental data and scientific models could be integrated into a digital twin.
The results demonstrated the suitability of the proposed architecture and established a methodological foundation for further developments.
Towards a multimodal environmental prototype
More recent work has expanded the initial architecture into an operational prototype deployed on server infrastructure.
Originally focused on groundwater, the system is evolving towards a multimodal approach that brings together different environmental domains. Its architecture supports the gradual integration of new datasets from the WATER and SOIL/AIR projects, alongside additional digital twins, including one dedicated to air quality.
The infrastructure now supports:
- Real-time data ingestion;
- Integration of satellite-derived data;
- Spatiotemporal analysis of environmental observations;
- Deployment of predictive models covering multiple environmental domains.
These developments have demonstrated the robustness and scalability of the ENVRI RM-inspired architecture when applied to complex environmental digital twin systems.
Ongoing developments and future directions
Current work focuses on strengthening the technical infrastructure, integrating additional scientific models and improving the platform's usability.
Key developments include:
- Integrating additional hydrological models developed within the programme;
- Enhancing the predictive components produced by the PREDICTION project;
- Automating the scheduling of computational tasks and data processing workflows;
- Further developing and consolidating the application to turn the prototype into a fully operational server-based system.
The project is also expanding its expertise in data science and machine learning to improve the integration of algorithms developed across the programme. This work includes the development of intelligent interfaces designed to make digital twins easier to use, including for non-specialist audiences.
Together, these advances will support the gradual integration of new digital twins and help establish a common methodology for developing environmental digital twins, from data architecture to operational tools that support scientific research and informed decision-making.
Scientific lead:
La Technopole d'Orléans,
LAB’O, 1 avenue du Champ de Mars
45074 Orléans
France