Real, validated and ready to use instances available using “instance downloader” algorithm of our Fire Analytics Toolbox QGIS plugin

tl;dr:

  • At least you need to map the fuels of a landscape and the wind speed and direction per hour
  • Supported raster formats: AAIGrid (.asc) and GeoTiff (.tif -partially supported)
  • Wind data goes in a Weather.csv (with other weather parameters)
  • All input data must be put in the same directory with standarized names

Rasters

  • All rasters [and qgis-project] must be projected in the same CRS, in squared meters
  • AAIGrid (.asc) raster format is fully supported unlike GeoTiff (.tif) that only reads a fuels.tif raster (ignoring the rest, still on development)
  • When using QGIS, there’s a sanity check:
    • Using the fuel raster as base, rasters must match
    • In number of pixels in both direction
    • At most one pixel offset in each direction
    • Hence pixel size must be very close [~mm]
rasters purpose units
fuels encode landscape fuel model table
elevation dem data elevation model meters
cbh canopy base height is where the lowest branch is meters
cbd canopy bulk density kg/m3
ccf canopy cover fraction is like cloud cover indicator [0,1]
hm canopy top height meters
py probability density map is to draw ignitions [0,1]

back to top

Fuel model

To encode the landscape in the simulator, you’ll need to map the fuels according to one of the 3 supported models:

These fuel model tables assign a type of fuel model or category to each pixel of the landscape using the integer map defined on the look up tables (.csv)

back to top

Weather

Weather.csv table specification:

  • minimal columns: name,timestamp,wind-speed,wind-direction
  • rows: by default each row lasts one hour
  • fuel model dependant columns:
Canada
Scenario,datetime,APCP,TMP,RH,WS,WD,FFMC,DMC,DC,ISI,BUI,FWI
Kitral
Instance,datetime,WS,WD,TMP,RH
Scott&Burgan
Scenario,datetime,WS,WD,FireScenario
  • FireScenario was deprecated by “Live & Dead Fuel Moisture Content Scenario [1=dry..4=moist]”
  • There’s no need to be consistent with the timestamps, but isoformat YYYY-MM-DDTHH:MM:SS should be followed
  • WIP: unifying weather formats, relying on column names instead of order to read them
  • Cell2Fire wind angle is reversed from normal metereological data, check wind direction for more details.

back to top