FireCaster
ForeFire demo site & Firecaster experiments
ForeFire demo site & Firecaster experiments
This is the official demonstration site for ForeFire and the home of Firecaster experiments. ForeFire is an open‑source simulation engine for wildfire spread used across Europe. Firecaster explores methods to compute and deliver forecasts, including use cases, coupled fire–atmosphere simulations, and prototype forecast products.
Front‑tracking resolution, parallel, calibrated for large scale fires (1000ha in less than 10s at meter resolution).
Python bindings, embedded HTTP‑based graphical interface, and an event‑driven scripting language with introspection, making it straightforward to generate ensembles of firefighting scenarios or probabilistic forecasts.
Source code on GitHub, DOIs, example scripts, and peer‑reviewed references for scientific traceability.
Experiments with Meso‑NH (FireCast) for plumes, feedbacks, and advanced diagnostics.
A selection of demonstrators and integrations.
Interactive browser‑based runs with European datasets. This is a demo built on top of the ForeFire API.
Coupled fire–atmosphere simulation system (ForeFire × Meso‑NH).
Augmented‑reality demonstrators that showcase the ForeFire simulator API; not part of the core ForeFire library. Code: github.com/forefireAPI/AR-Weather
Coupled experiments with Meso‑NH (see in MesoNH repository ).
Video demos of ForeFire use cases, including the web simulator and AR experiments; not part of the core ForeFire library.
Python package to generate and compare wildfire Rate‑of‑Spread models, train ANN surrogates, and export C++ code for ForeFire integration.
Note on demos and scripting: the possibility to script ForeFire on real datasets depends heavily on input data flows, whether meteorological or fuel, and is therefore not immediately exportable. Like the EPI, it is constantly evolving and is not contained in a fixed repository, because its configuration depends on the servers where it runs. CNRS may provide access or contract for services, but the automatic scripting platform itself cannot realistically be open source. Regarding Meso‑NH, all developments on the fire side are done within ForeFire, but the atmospheric part remains independent within Meso‑NH. Scripts and parameterized cases for ForeFire can work in coupled mode with Meso‑NH, provided Meso‑NH is correctly installed and configured. Meso‑NH itself evolves independently and cannot be bundled directly, so to run coupled fire–atmosphere cases you need to obtain and install Meso‑NH from its repositories.
You can explore the C++ core and Python bindings on GitHub. The demo web simulator is separate.
Some papers and software references related to ForeFire.
ForeFire is maintained in an academic setting (CNRS — Università di Corsica / UMR SPE). The resources above provide entry points to reproduce the demonstrators and study the methodology.
Useful links around the atmospheric model coupled to ForeFire.