ForestSphere

The ForestSphere project aims to develop a Digital Twin for forest management

Recognising its socioeconomic and ecological importance. Forests contribute significantly to the European Union's GDP and serve as major carbon sinks. However, forest fires pose a substantial threat to forests, leading to environmental damage, loss of life, and economic losses.

Despite advances in Digital Twin technology across various sectors, the current literature has limited exploration of its applications in forest management, particularly in wildfire prediction.

The ForestSphere project intends to address this gap by creating a Forest DT that integrates multimodal and multiscale data from various sensors to improve forest health monitoring and wildfire risk assessment. The main objectives of the project include:

Develop an open, scalable, and modular Digital Twin architecture.
Enable accurate, high-resolution forest mapping using advanced sensing techniques.
Implement real-time data processing for dynamic DT updates.

Utilise Artificial Intelligence for predictive wildfire analyses.
Design user-centred decision-support tools to assist firefighting teams and policymakers.
Conduct pilot studies to demonstrate the applications of the Forest DT in forest and wildfire management.

The project follows a methodology structured in three phases:

Development Phase

Creation of the foundational DT through comprehensive data collection from various sources, including ground sensors, aerial imagery, and satellite remote sensing. Ground data will be collected using handheld devices, meteorological sensors, and robotics to generate detailed 3D models of the forest. Aerial sensors will provide broader coverage to capture canopy structure and vegetation stress, while satellite imagery will enable large-scale monitoring of environmental variables.

Optimisation Phase

Improve the DT’s accuracy and efficiency through AI-driven enhancements and participatory optimisations. New data streams will be integrated, and change-detection algorithms will trigger updates to reflect current forest conditions. Computational optimisations, visualisation tools, and predictive models will allow the DT to simulate fire dynamics and enhance decision-making for emergency responders.

Implementation Phase

Validate the DT through pilot studies, comparing model predictions with real-world measurements. This stage includes integrating the DT into existing decision-support systems to provide real-time insights for wildfire management and ensuring interoperability with various agencies.

The project consortium is led by ADAI, a leading research group on forest fires with over 35 years of work dedicated to various aspects of fire behaviour research. ADAI also possesses extensive experience in leading large national and European projects, as well as defining international policy on best practices for forest and infrastructure resilience against wildfires. ISR-UC is a research centre at the University of Coimbra with decades of expertise in AI, remote sensing, and robotics. Its internationally recognised teams will contribute with their expertise in integrating neural networks and other advanced technologies into the proposed DT. OneSource is a well-established company that develops and applies cutting-edge software and hardware to support civil protection agencies. Its proven expertise in systems integration and validation in European research projects, as well as in deploying its technology, will be vital to the project. REN has been collaborating with ADAI to co-develop and demonstrate technologies with the potential to mitigate wildfire risks across its energy distribution network. In recent years, this collaboration has successfully shown a distributed sensor network capable of detecting and georeferencing fire ignitions, producing automated fire-spread predictions and impact maps, and generating alerts to enable managers to act early and implement corrective measures. Bold Robotics and Sim4Safety are two tech startups that will leverage the developed DT to showcase their products, providing a glimpse into the future of wildfire management. Support from CIM-RC and CML is also crucial in defining the system requirements and specifications, as well as facilitating the execution of pilots and validating the developed technology.

Innovative
features

Combining data from ground sources, drones, and satellites to build a dynamic and accurate digital representation of the forest.

Using AI algorithms and machine learning to predict wildfire risk and fire propagation.

Applying robotic mapping and remote sensing technologies to generate detailed three-dimensional models of forest ecosystems.

Developing interfaces with 3D visualisations and augmented and virtual reality environments, facilitating decision-making by authorities and field teams.

Objectives

Develop a robust Digital Twin

Hable to integrates multiple data sources and tools, with a focus on interoperability, reusability, and future scalability. The Digital Twin architecture will be designed to ensure that data and models can be easily shared across different domains and applications.

Implement remote sensing and robotic mapping

To generate highly detailed, high-fidelity 3D reconstructions of forest ecosystems. This capability will enable more accurate monitoring of vegetation health, terrain structure, and other critical factors for wildfire risk management.

Develop intelligent mechanisms to collect and process data in real time

Allowing the Digital Twin to track environmental changes continuously. This way, the system will always be up to date and prepared to respond quickly and knowledgeably to real-world forest conditions.

Develop predictive models based on artificial intelligence and machine learning

Capable of simulating fire and smoke propagation, reinforcing early-warning systems and improving decision support in forest emergencies.

Design intuitive and engaging interfaces

including 3D visualizations, geospatial dashboards, and augmented or virtual reality environments, that transform the Digital Twin’s data into useful, accessible information, supporting decision-making by authorities, environmental managers, and firefighting teams.

Development Phase

Optimisation Phase

Implementation Phase

Cookie settings