Wildfire CA Simulator
Uses the provided map as default to derive DEM and fuel. The CA consumes an ML fire probability layer and geodata to simulate spread over time.
Effective interval: 150 ms (faster with higher wind & threshold)
Status: Idle
Design Notes
- States: 0=Unburnt, 1=Burning, 2=Burnt. Moore 8-neighborhood.
- Base probability from ML per-pixel map.
- Fuel scales ignition; denser fuel → higher chance.
- Slope factor f_slope = exp(k·slope). Uphill spreads faster.
- Wind factor from dot-product with neighbor direction.
- Humidity reduces spread probability.
- Distance: immediate neighbors; ember spotting optional.
- Stochastic CA: RNG decides ignition per neighbor per step.
Simulation Canvas
Green=Unburnt, Orange=Burning, Gray=Burnt
Demo API (optional)
You can also fetch demo frames from the API at /api/simulate (returns small synthetic frames).