This applied axis addresses the growing need to understand, monitor, and model natural and human-induced hazards—such as floods, wildland fires, pollutant dispersion, and industrial accidents—at multiple scales, from local to global. Driven by climate change, increased urbanization, and the energy transition, this work supports public safety and sustainable industrial development through advanced simulations and innovative modeling approaches.
The key objective is to improve understanding of multi-physics and multi-scale processes that govern these hazards, and to provide predictive and evaluative tools for risk mitigation and decision-making. CERFACS focuses on three main dimensions: (1) analyzing past events (“what now”), (2) forecasting potential future impacts (“what next”), and (3) simulating alternative scenarios (“what if”). This is achieved through high-performance computing (HPC), data assimilation, uncertainty quantification, and the integration of diverse data sources (e.g., in situ, satellite).
Five priority challenges guide this axis:
-
Water-related hazards: Developing a multi-scale simulation framework to model droughts, floods, sediment transport, and biodiversity impacts, using hydrological models and remote sensing.
- Wildland fire modeling: Building a coupled system to simulate fire spread, plume dynamics, and pollutant lifetimes at landscape and atmospheric scales.
- Urban pollutant dispersion: Creating ensemble-based micrometeorological models to assess wind and pollutant behavior in cities during emergencies.
- Industrial accident simulation: Advancing CFD capabilities to model accidental scenarios (leaks, explosions, fires) with high temporal and spatial resolution.
- Emission and long-term environmental impacts: Prototyping modeling chains to assess how local emissions affect global atmospheric and climate systems.
These challenges are tackled using advanced software such as AVBP (LES solver), Meso-NH (atmospheric modeling), lbmpy (Lattice Boltzmann), and tools for hydrological modeling like CEPHEE and T2DiSurf.
Strategically, this research relies on:
- “Numerical algorithms“: Enhancing high-order solvers and alternative methods like LBM to improve simulation accuracy and performance.
-
“Data driven modelling“: Leveraging AI and hybrid strategies to improve assimilation of high-resolution observations, reduce uncertainty, and build surrogate models.
-
“Sustainable programming“: Ensuring scalability and long-term usability of coupled simulations on evolving HPC platforms.
CERFACS collaborates with key shareholders such as Airbus, EDF, Météo-France, ONERA, TotalEnergies, and CNES, with support from regional, national, and European programs. The axis ultimately contributes to building digital twins of complex environmental systems and supporting global initiatives like Destination Earth.
