Skip to content

PROPAGATOR Sim

An operational cellular-automata wildfire simulator developed by CIMA Research Foundation. PROPAGATOR couples a Numba-powered propagation core with reusable I/O pipelines and a configurable CLI for fire forecasting.

Get started Programmatic workflow API reference Bibliography

What's Included

  • Simulation engine: the propagator.core package evolves ignition grids, applies stochastic spread models, and handles boundary conditions.
  • Data access: helpers under propagator.io prepare GeoTIFFs or tiled rasters and emit GeoTIFF, GeoJSON, and JSON outputs.
  • Command line tools: the propagator CLI orchestrates runs, handles configuration files, and writes time-stepped products to disk.
  • Documentation + API reference: MkDocs pages provide operator guides, while mkdocstrings renders the public Python API.

Typical Workflow

  1. Prepare a JSON configuration describing ignition geometry, simulation horizon, and model toggles (see example/config*.json).
  2. Supply static data—either as GeoTIFF DEM/vegetation layers or a tileset directory—and optional YAML fuel definitions.
  3. Launch the run via the CLI (uv run propagator …), enabling verbose or recording modes for richer logging.
  4. Inspect the generated rasters, isochrones, and metadata in the configured output directory.
  • Getting started: prerequisites, install, first simulation, and programmatic API tips
  • Programmatic Workflow: end-to-end scripting with loaders and writers
  • CLI: command options, modes, and logging
  • API: Python package and Numba backend reference